Está en la página 1de 149

Cell Stem Cell

Editorial
Five Years, with Thanks
Welcome to the special issue marking the fth anniversary of the launch of Cell Stem Cell. The inaugural issue of the journal, dated July 2007, was published in early June for the ISSCR annual meeting in Cairns, Australia. We have therefore chosen to celebrate the anniversary in June, and to distribute this issue at the ISSCR annual meeting in Yokohama, Japan. We are very grateful to Stemgent, Life Technologies, Lonza, and StemCell Technologies for providing the sponsorship support that enables us to send a copy for every attendee at the meeting. For this special issue, we have focused on progress in the stem cell eld over the past 5 years and outstanding issues that remain through a series of Perspective and Forum articles. For the Perspectives, we asked the authors to summarize key developments in different areas of stem cell research since Cell Stem Cell launched. The approaches they have chosen to take vary some have focused on one or two key points and discuss them in depth, while others have provided more of an overview of major areas of progress. In the Forum section, we gave each author free rein to express their opinion on a topic that they consider important for the eld. Again, the contributions in this section vary, in terms of subject covered and degree of speculation involved. We hope that all of these articles from leaders in the eld will prove thought-provoking and informative, and that together they provide a snapshot of the current status and future directions. In honor of the location of the 2012 ISSCR meeting, we have also chosen to feature three research articles from Japan. These papers, from the groups of Akira Onishi, Yoshiki Sasai, and Jun Yamashita, provide a tting illustration of the strong contributions that Japanese scientists are making to the stem cell eld. The cover of the issue has a connection to Japan as well, because the winning entry from the competition we held earlier this year is from Mai Kimoto, a doctoral student at Hokkaido University. Mais beautiful hand-drawn image depicts a broad range of tissues in which stem cells have been identied and characterized in a design inspired by marine ecosystems. As Leonard Zon discusses in his accompanying editorial, this 5 year anniversary for Cell Stem Cell coincides with the 10 year anniversary of the founding of the ISSCR. Like the journal, the society has evolved signicantly since its rst inception, and has become a major force within the stem cell eld and beyond as society at large grapples with how to support and regulate the stem cell eld. You will see a short montage depicting the societys activities in this issue, and information about the winners of the societys 2012 awards, all of whom we heartily congratulate. We at Cell Stem Cell very much appreciate the mutually supportive relationship that we have with the ISSCR, and look forward to it continuing in the future. In running a journal for the stem cell eld, our primary goal at all times is to make a positive contribution to scientic progress. We are well aware that all scientists, but particularly those working in a fast-moving eld such as stem cells, want the process of publication to be as rapid and straightforward as possible. There are many factors to balance, including author and reviewer time commitment and the desire to ensure that we hold papers to appropriate scientic standards without having unreasonable expectations about experimental content. It has always been our policy to take papers through just one major round of revision to avoid protracted review processes, and the supplemental material guidelines that Cell Press introduced 2 years ago have helped rein in expectations about the amount of information needed for a given publication. We strive to ensure that our consideration processes are timely and thoughtful, and that we both explain our decisions and minimize the time spent for everybody involved. We also aim to be accessible to authors and to provide appropriate advice to help navigate the publication process as smoothly as possible. We have therefore been very happy about the amount of positive feedback we have received from the community about the quality of the editorial service provided by Cell Stem Cell and the role that the journal has played in the stem cell eld, both anecdotally and through a broader survey that Cell Press sent out earlier this year. Our ability to make this type of positive contribution depends on support and assistance from many different members of the scientic community, and we are very grateful to everyone who has helped Cell Stem Cell over the past 5 years. The journal would be nothing without the articles it contains, so we appreciate all of the authors who entrusted us with presenting their work to the broader community. The members of the editorial board and ISSCR advisory committee have provided valuable support and guidance as we have developed Cell Stem Cell and its coverage. Peer reviewers are also a vital component of any journals operation. Over 1,300 individuals have acted as referees for research and review articles submitted to Cell Stem Cell since early 2007, and we are grateful to each and every one of them for the time and effort they put in. Because they are anonymous, for good reason, reviewers are often unsung heroes in the publication process. In an effort to redress that balance, Id like to take this opportunity to recognize their contribution. Rapid, discerning, thoughtful, and constructive reviews help us as editors decide which papers are appropriate for publication and help the authors involved improve their work. The scientic enterprise benets as a result. I am often asked where Cell Stem Cell, and the eld more generally, will go in the future. The honest answer is that I dont know, but that is part of what makes it so exciting! All of the articles in this issue serve to illustrate the rapid pace of progress in the stem cell eld, and the energy that surrounds it as it continues to grow. I personally have felt and continue to feel privileged to be part of this vibrant and fascinating area of science. I am sure I speak for Christina, Sheila, and everyone else involved in Cell Stem Cell when I say that we very much look forward to continuing to work with the research community and to doing our part in helping the stem cell eld continue to succeed. Please come and visit us during the ISSCR meeting in Yokohama at booth 208 to learn more about Cell Stem Cell and the many exciting initiatives going on at Cell Press.

Deborah J. Sweet

Editor, Cell Stem Cell DOI 10.1016/j.stem.2012.05.017 Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc. 637

Cell Stem Cell

Editorial
A Great Match
It is hard to believe that the International Society for Stem Cell Research (ISSCR) is 10 years old, and that coincident with this is the fth-year anniversary of Cell Stem Cell. Having participated in the creation and development of the society and the journal, the success of both has made me appreciate my matchmaking skills. When I founded the ISSCR, I was responding to a call from stem cell researchers to bring investigators together and form a community. The annual meeting served this purpose from its edgling initial meeting in Washington, D.C., with about 550 people, to this years meeting in Yokohama with about 4,000 individuals. There is a tremendous need for scientic interactions, and the community is vibrant and operates at a very high level. One of my initial thoughts for the ISSCR was that it needed a journal. A key driver was the stem cell research communitys call for a dedicated forum to publish stem cell research and to discuss issues faced by the elda high-quality and high-visibility journal to give the eld credibility and capture the highest prole papers. In this endeavor, after a long negotiation, we afliated with Cell Press to start Cell Stem Cell. I am pleased to have been integral in the discussions that resulted in the strong partnership between ISSCR and Cell Stem Cell. Cell Stem Cell has accomplished all of its goals admirably from its initiation, and Debbie Sweet has been a pioneering editor with great vision. With a 2010 impact factor of 25.9, it is clearly a highend journal that publishes impactful papers. The reviews are fantastic and comprehensive. The ISSCR section of the journal is one of its unique features, presenting key undertakings of the society such as committee activities, task force initiatives, and professional guidance. The ISSCR meeting reviews are greatly welcomed. When Cell Stem Cell needed to become nimble, decisions were made to create new sections that reect the progress in the eld. I am particularly interested in the Clinical Progress section that allows data to be put into the journal that bridges basic research and clinical application. There is also a diversity of articles in Cell Stem Cell. This includes the tremendous literature on induced pluripotent stem cells (a eld in itself and one that will greatly contribute to science and medicine in the future). The journal, in short, has been a great success, has had wonderful interactions with the ISSCR, and has moved the eld forward in a palpable and self-renewing manner. The ISSCR has tackled many issues in its rst 10 years. One of the most important aspects was mentoring. It is clear that young investigators have found a home in the ISSCR. The society has a special committee for junior investigators and there is a formal mentoring session at each annual meeting. I remember the time period before induced pluripotent stem cells were available when young investigators were unclear if any funding would be available for human pluripotent stem cell studies on the only available technology, human embryonic stem cells. It was hard to have the courage to start a lab working on pluripotent cells, and now with many advances, the eld is vibrant with many junior investigators.
638 Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc.

The ISSCR has highlighted ethical questions in various ways, including having Laurie Zoloth make a presidential session speech at one of the annual meetings. The society has also made a strong effort to reach out to patient advocates and policy-makers, and few will forget the powerful speech that one such advocate, Charles Sabine, gave at the 2011 meeting. The ethics of working with stem cells were brought to the forefront because of a Presidential executive order. Many of us from the leadership of the ISSCR have appeared on radio, on TV, and in other contexts to inform and educate the public about stem cell research. The stem cell eld has also had to deal with some difcult issues over the past 10 years. We found out that human cloning was not in fact performed in Korea despite prior publication. In Europe, a landmark legal decision about patenting of human embryonic stem cells changed the landscape for future research and commercial development. U.S. researchers weathered considerable uncertainty when a judicial process based on a lawsuit brought by investigators who do not support research on embryonic stem cells shut down federal funding for a period of time. The society responded by dening its role in embryonic stem cell research (a role that not everyone in the eld embraces, but one that is practical and informative). At this point, it seems as though embryonic stem cells are becoming less of a big issue and there is broader recognition that they are important for the future of the eld. In vitro fertilization was once a huge source of ethical disagreement, but is now mainstream. I hope that embryonic stem cells will achieve a similar status. The ISSCR has also provided information and expert opinions to help many governments address their stem cell issues. As part of that effort, the society produced an international document on the practical use and ethical use of stem cells. In addition, counseling of individual governments facilitated the set-up of stem cell institutes in various countries. Government policy has been strongly inuenced by the advice of several ISSCR presidents and leaders. The public efforts of the ISSCR revolve around a single topic chosen every year by the president. In my case, I focused on education of the public. More recently, there has been a decision to examine questionable clinics that do stem cell research. Such snake oil salesmen often promise major clinical success for therapies that are not scientically based. The society is in the process of evaluating how to call out companies that are involved in such efforts and how to inform the public of such behavior. In the interim, we have created a web page to provide information to patients who seek to evaluate new stem cell therapies (http://www.closerlookatstemcells.org). It has been interesting to watch the leadership transitions that occur through the ISSCR. Each year a new president takes the helm and brings their unique perspective to the role. Greg Schultz, the founding executive director, was absolutely critical in the process in the early days, and Nancy Witty, our current executive director, has done a wonderful job over the past 6 years. It is clearly through Nancys leadership that the mission of the ISSCR has been truly executed.

Cell Stem Cell

Editorial
I see a number of future opportunities for the eld of stem cell research and regenerative medicine in which the ISSCR and Cell Stem Cell can play a vital role. One of those is clinical translation. Existing successes include limbal stem cells being transplanted to treat corneal eye diseases and the expansion of blood stem cells. The study of patient-specic induced pluripotent stem cells or diseases in a dish will undoubtedly yield information that will be brought to the clinic. The need for a facilitation of this process is very important and a journal is critical. I feel that Cell Stem Cell could have a pioneering role in promoting translation of stem cell research. There are many issues yet to be overcome, including ones related to safety, manufacturing, and access to stem cell treatments. Another issue for the eld is to maintain the quality of the research it produces. With the increasing number of researchers getting involved, it is critical to remain rigorous, and to ensure that published conclusions are well supported by data. The eld has already seen its share of retractions and controversy. I hope that everyone working in this area will continue to produce fantastic research that builds to a crescendo of clinical relevance. Here is what I said at the very rst meeting at the end of my presidential speech: I am very enthusiastic about the Society, and I hope that 20 years from now we will be utilizing stem cells in the clinic as routine medical therapy for many genetic and degenerative disorders, that the basic genetics and molecular biology and cell biology of the stem cell will be largely understood, and that the Society will be lled with wonderful young trainees and old investigators such as myself. We are now at the halfway mark, celebrating the ISSCRs 10th Annual Meeting where we are hearing a wealth of progress towards this goal, presented by a truly global community of all ages. I want to raise a toast to ISSCR and Cell Stem Cell: to a happy birthday and continued success.

Leonard I. Zon
Founding President, ISSCR DOI 10.1016/j.stem.2012.05.002

Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc. 639

Forum
The Impact of Cell Culture on Stem Cell Research
Elaine Fuchs1,*
1Howard Hughes Medical Institute, Laboratory of Mammalian Cell Biology & Development, The Rockefeller University, New York, NY 10065, USA *Correspondence: fuchslb@rockefeller.edu DOI 10.1016/j.stem.2012.03.010

Cell Stem Cell

Induced pluripotent stem cell research has broadened possibilities for regenerative medicine and captured the worlds attention in a way that science rarely does. However, clinical applications utilizing cultured stem cells have existed for >30 years and can assist benchers and bedsiders in identifying and expediting promising avenues for future therapies.
The stem cell community generally credits Till and McCullochs transplantation experiments of the 1960s on the hematopoietic system for the demonstration that adult tissues contain stem cells (Till and McCulloch, 1961). In those experiments, so-called colony forming units (CFUs) were produced by short-term culture of isolated murine bone marrow cells and then individually transplanted into irradiated recipients to reconstitute the hematopoietic system. From these early in vivo experiments came the concept that stem cells are often rare cells that fuel homeostasis and wound repair through their ability to generate both the proliferating and the differentiating cells of our tissues. The notion that stem cells can do so long term also came from hematopoietic studies as researchers began to perform long-term experiments on hematopoietic stem cells that were isolated and puried directly from the bone marrow and serially transplanted through many generations. Although the hematopoietic system led the way in devising concepts for stem cell biology, the ability to passage stem cells long term in vitro began in the 1970s with the pioneering work of Howard Green on human epithelial stem cells. At the time, most researchers resorted to immortalized, transformed cell lines, which seemed to be the only cells that could proliferate and differentiate and yet easily be passaged long term. Cloned teratocarcinoma cells were particularly interesting to developmental biologists, who were examining the ability of these embryonic-like cells to differentiate along a variety of lineages in vitro. Green noticed that epithelial colonies were among the cell types present within teratocarcinoma cultures and discovered that he could clone and propagate them on a layer of lethally irradiated, diploid mouse 3T3 broblasts, a line that he had developed previously. When the strategy of using a broblast feeder layer was subsequently applied to human epidermal cells, large colonies of diploid epithelial keratinocytes grew that retained their ability to self renew and terminally differentiate long term (Rheinwald and Green, 1975). Green and his colleagues continued to improve upon the culture conditions (reviewed by Green, 1991). They added epidermal growth factor puried from rat submaxillary glands to the culture conditions, as well as insulin and dexamethasone. They also spiked the media with cholera toxin, a constitutive activator of cyclic AMP, which also aided cell growth substantially. As the keratinocytes enormous proliferative powers became increasingly exposed, so did the promise to generate sufciently large sheets of cultured epidermal cells from a small piece of healthy skin to cover the damaged regions of a badly burned patient. These early studies represent the birth of what is now a 30 year successful application of puried human stem cells for regenerative medicine (reviewed by Green, 1991). Therapeutic uses of cultured stem cells are often viewed by the public as futuristic, if not science ction. We need to work harder to educate our society of the already fullled wonders of stem cells for medical applications. The clinical applications for epithelial cells have not ended at epidermal cells. Using very similar culture conditions, other stratied squamous epithelial stem cells, including corneal cells, can be cultured long term in vitro. This ability led to the subsequent application of cultured corneal progenitors to treat patients suffering from corneal blindness, and a 10 year study of successfully treating 100 such patients was recently published by Michele De Luca, Grazia Pellegrini, and colleagues (reviewed by Rama et al., 2010). These early culture methods did much more than ever imagined at the time. Moreover, they not only paved the way for these impressive clinical applications, but in addition they opened the door for embryonic stem cell (ESC) research as we know it today. The concept of coculturing epithelial cells with a broblast feeder layer was quickly adapted to mouse ESCs and worked particularly well when used in conjunction with conditioned medium from teratocarcinoma cells (Evans and Kaufman, 1981; Martin, 1981). Once the hurdle of culturing primary ESCs was overcome, germline transmission was soon achieved, providing graphic illustration that ESCs are truly pluripotent, able to generate the 210 cell types of the mouse (Robertson et al., 1986). Although the early mammalian cell culture studies were instrumental in bringing stem cells to a clinical setting, a two decade gap separated those advances from the ones of Yamanaka and colleagues that have captivated the interests of scientists and public alike (Takahashi and Yamanaka, 2006). What accounted for this gap and how can we expedite progress in regenerative medicine in the years to come? Below are a few ideas based upon the history of the eld and my own experiences as a stem cell biologist. As a postdoctoral fellow and biochemist in Greens laboratory in the late 1970s, I was far more fascinated by

640 Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc.

Cell Stem Cell

Forum
studying how epidermal stem cells balance growth and differentiation at a molecular level than I was in the scientically more mundane, albeit more immediately relevant, task of optimizing culture conditions for burn therapy. Although placing stem cells into a molecular framework has taken three additional decades to evolve, there are several recent indications that continuing to build upon this foundation will be the engine that drives regenerative medicine in the future. In this regard, it is intriguing that cyclic AMP, one of the culture additives that advanced epithelial regenerative medicine in the 1970s, was recently found to act downstream in the Wnt signaling pathway (Goessling et al., 2011), known to have a powerful impact on stem cells and tissue regeneration. This new knowledge has now been translated into an FDAapproved phase 1 clinical trial that could signicantly improve human cord blood transplantations (Goessling et al., 2011). On a similar note, in converting skin broblasts to iPSCs, Yamanaka and coworkers achieved their success by exploiting not only mammalian cell culture technology but also knowledge of the molecular differences between human ESCs and broblasts. By employing an amazingly simple screen to weed out nonessential genes in these differential patterns of gene expression, the researchers honed in on the key transcription factors that promote an embryonic-like fate (Takahashi and Yamanaka, 2006). This knowledge, based squarely upon the molecular foundations of ESC biology, is now fueling efforts to identify small molecules that will optimize iPSC derivation and culture. In other words, while the eld began with cell culture to fuel molecular insights into stem cell biology (as discussed in, for example, Blanpain and Fuchs, 2009), the molecular insights are now fueling approaches to expand the repertoire and populations of cultured stem cells that can be applied in the clinics. Another relevant lesson stemming from the dawn of epidermal stem cell biology is that keratinocytes alter their program of gene expression in culture. Based upon what we currently understand about the extraordinary complexities of skin stem cell niches (reviewed in Blanpain and Fuchs, 2009), it seems unlikely that this will be fully rectied no matter what we may do to optimize culture conditions. That said, the long-standing success of cultured human keratinocytes for burn therapy and the absence of skin cancers in patients engrafted decades ago suggests that epidermal progenitors do not lose their stemness when passaged in vitro, nor do they become transformed. Analogously, when engrafted onto the backs of hairless mice, murine hair follicle stem cells passaged in vitro can generate epidermis, sebaceous glands, and hair follicles, and they even seem to be able to collaborate with dermal cells in recreating a new stem cell niche (reviewed in Blanpain and Fuchs, 2009). These ndings suggest that as long as their capacity for long-term self-renewal and differentiation can be faithfully maintained in vitro, cultured stem cells should continue to hold promise for regenerative medicine even if they transiently adopt a new molecular program when propagated outside their normal microenvironment. As exciting as this concept is, with few exceptions, stem cell populations are in limited supply, posing signicant hurdles even for stem cells such as corneal or hematopoietic stem cells where clinical applications are well established. Taking a page from developmental biology, it would seem that one good way to overcome these barriers would be to funnel our collective research energies into enhancing our knowledge of stem cell activation and self-renewal. Such strategies are ones that are already being actively pursued by a number of stem cell researchers. Indeed, the aim to enhance hematopoietic stem cell self-renewal in vitro was behind the Zon groups efforts to conduct their cleverly devised zebrash screen for FDA-approved small molecules that could enhance the process in vivo. This in turn resulted in the identication of dimethylprostaglandin E2 (dmPGE2), which has not only provided a possible link between cAMP and Wnt stem cell signaling pathways but has also led to the phase 1 clinical trials described above (Goessling et al., 2011). And in a recent RNAi screen for self-renewing hair follicle stem cell genes, clues have surfaced that might be exploited to improve epithelial stem cell expansion in vitro (Chen et al., 2012). If so, this could aid current treatments for corneal blindness, where stem cell numbers are often the rate-limiting step to success (reviewed by Rama et al., 2010). An alternative route to overcoming the limited supplies of cultured stem cells is to transdifferentiate closely related cells that are either in greater supply or which can be propagated more easily in vitro than the desired stem cell. For instance, if we can gain an understanding of the transcriptional differences between corneal and skin stem cells, can we exploit this information to transdifferentiate skin stem cells into corneal stem cells? Given the close relation between these two types of stratied squamous epithelial progenitors, such an approach seems like a baby step relative to Yamanakas giant step that launched the game of transdifferentiation hopscotch. Although the borders between dreaming and thinking beyond the box can sometimes be quite blurred, the recent successes of Wernig and others (reviewed by Chambers and Studer, 2011) suggest that this may be the most accessible route to future clinical applications.
ACKNOWLEDGMENTS E.F. is an Investigator of the Howard Hughes Medical Institute and receives grant support from the National Institutes of Health and the New York State Stem Cell Initiative. REFERENCES Blanpain, C., and Fuchs, E. (2009). Nat. Rev. Mol. Cell Biol. 10, 207217. Chambers, S.M., and Studer, L. (2011). Cell 145, 827830. Chen, T., Heller, E., Beronja, S., Oshimori, N., Stokes, N., and Fuchs, E. (2012). Nature, in press. Evans, M.J., and Kaufman, M.H. (1981). Nature 292, 154156. Goessling, W., Allen, R.S., Guan, X., Jin, P., Uchida, N., Dovey, M., Harris, J.M., Metzger, M.E., Bonifacino, A.C., Stroncek, D., et al. (2011). Cell Stem Cell 8, 445458. Green, H. (1991). Sci. Am. 265, 96102. Martin, G.R. (1981). Proc. Natl. Acad. Sci. USA 78, 76347638. Rama, P., Matuska, S., Paganoni, G., Spinelli, A., De Luca, M., and Pellegrini, G. (2010). N. Engl. J. Med. 363, 147155. Rheinwald, J.G., and Green, H. (1975). Cell 6, 331343. Robertson, E., Bradley, A., Kuehn, M., and Evans, M. (1986). Nature 323, 445448. Takahashi, K., and Yamanaka, S. (2006). Cell 126, 663676. Till, J.E., and McCulloch, E.A. (1961). Radiat. Res. 14, 213222.

Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc. 641

Forum
Synthetic Organs for Regenerative Medicine
Roger A. Pedersen,1,2,* Victoria Mascetti,1,2 and Sasha Mendjan1
Anne McLaren Laboratory for Regenerative Medicine, Stem Cell Institute, University of Cambridge, Cambridge, CB2 0SZ, UK of Surgery, University of Cambridge, Cambridge, CB2 0QQ, UK *Correspondence: roger@stemcells.cam.ac.uk DOI 10.1016/j.stem.2012.04.003
2Department 1The

Cell Stem Cell

Differentiating tissue stem cells can self-assemble into structures that strikingly resemble functional organ subunits. Translating this insight to regenerative medicine presents several challenges.
The concept of regenerative medicine encompasses cell therapy, tissue engineering, and improving the regenerative capacity of endogenous organs. Three major remaining obstacles to the clinical application of stem cells are scaling up production of progenitors, obtaining mature (adult) human phenotypes in terminally differentiated cells, and developing complex tissues. In this article we discuss how knowledge about the mechanisms of cell proliferation, maturation, and self-assembly could be used to address these challenges and help realize the therapeutic potential of stem cells. Adult stem cells (more accurately called tissue stem cells, as they underlie tissue self-renewal also during fetal, neonatal, childhood, and adolescent development) have been extensively studied and characterized in vivo. However, with a few exceptions the controlled in vitro development of tissue stem cells has so far proved elusive. Although there have been impressive clinical successes with transplantation of hematopoietic stem cells (even without in vitro expansion), scale-up will be essential for most other stem cell clinical applications. The difculty of growing tissue stem cells contrasts markedly with long-term proliferation of cultured pluripotent stem cells. Mouse embryonic stem cells (mESCs), epiblast stem cells (EpiSCs), and human embryonic stem cells (hESCs) have been isolated from the embryos own pluripotent cells, which proliferate actively both before and after they undergo further in vivo specialization. Despite their proliferative capacity, however, the quantities of pluripotent stem cells that can be obtained in routine culture (108) are still not sufcient for most clinical applications if the stem cells are differentiated without further cell proliferation. Moreover, in vitro proliferation of pluripotent stem cell progeny typically ceases soon after the induction of differentiation. The requirement for scale-up focuses attention on the possibility of expanding cells in intermediate states between stem cells and mature differentiation. The in vitro counterpart of highly proliferative embryonic primary tissue layers (the so-called germ layersectoderm, endoderm, and mesoderm) would be a valuable asset for achieving scale-up. These could multiply the number of stem cell progeny geometrically, just as transit-amplifying or progenitor cells do during normal tissue development and homeostasis in vivo. The feasibility of inducing mouse broblasts to become neural stem cells has recently been demonstrated (see Zhou and Tripathi, 2012), and endodermal progenitor cells capable of differentiating into lung and other tissues have now been derived from mouse and human ESCs (see Kadzik and Morrisey, 2012). Further studies will be needed to see if mesodermal tissue layer stem cells can also be captured and to learn whether the differentiative capacity of induced neural stem cells and endodermal progenitor cells spans the entire repertoire of their respective primary tissue layers. In addition, studies of the mechanisms regulating proliferation in stem cells and their differentiating progeny will be important both for achieving scale-up and for ensuring that expanded cell populations do not continue to proliferate inappropriately once they are used for cellular therapies. Although numerous protocols now exist for inducing the differentiation of mESCs, EpiSCs, hESCs, and induced pluripotent stem cells (iPSCs) into specic tissue fates, these differentiated cells almost invariably correspond to an immature (fetal) rather than a fully mature (adult) stage of development. This relative immaturity of pluripotent-stem-cell-derived differentiated cells may compromise their function and present a barrier to engraftment and structural integration into adult human tissues. Some evidence may suggest that development of the differentiated progeny of pluripotent cells is accelerated as compared with mouse or human fetal development, but their ultimate phenotype is nonetheless immature. In addition, normal organs contain not only epithelial and mesenchymal cells, but also vasculature, nerves, and fascia, and the activity of all of these components needs to be coordinated appropriately to achieve organ function in vivo. Maturation of all these different elements in vitro might require months of differentiationpromoting culture to generate fully mature, organ-specic cell types from pluripotent stem cells. Recent work by van Vliet et al. (2010) suggests an alternative strategy for achieving mature phenotypes. They identied a disparity between fetal and adult progenitor cells in potential for differentiation into mature cardiomyocyte cells: fetal progenitors differentiated into immature cardiomyocytes, whereas adult progenitors formed mature ones. This result indicates that the capacity for maturation is acquired during in vivo development from a pluripotent cell to a tissue-specic progenitor. In addition, recent studies that have demonstrated reprogramming of differentiated mouse cells directly from one phenotype to another; these studies also provide evidence that such direct reprogramming can be achieved without transiting a less mature state (see Vierbuchen and Wernig, 2011). It may similarly be possible to achieve mature human phenotypes by direct reprogramming of mature cells. However, it remains to be determined what fundamentally distinguishes mature

646 Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc.

Cell Stem Cell

Forum
from immature phenotypes. Understanding the transciptomic, proteomic, and metabolomic networks that underlie cell maturity should enable direct reprogramming from highly proliferative pluripotent or intermediate stem cells into mature adult phenotypes, thus addressing both the scale-up and maturity challenges. The use of stem cells to generate entire organs for transplantation to replace diseased or damaged ones remains the holy grail of regenerative medicine. It may be possible in principle to generate humanized organs in domestic species using pluripotent stem cells and interspecies chimeras, as has been done with rats and mice. However, the ethical and xenotransplantation obstacles inherent in this approach make it very challenging. Growing entire organs from embryonic rudiments completely in vitrorecapitulating the process of organ formation and growthseems equally if not more difcult. An alternative to generating a whole organ may be to generate the functional subunits of an organ in vitro. Most organs are composites of functional subunits (the smallest elements of an organ that retain the essential activity performed by the organ). For the kidney, this subunit is the nephron; for the lung, the alveolus; for the liver, the hepatic lobule; and for the small intestine, the villus and crypt. Recent research shows that it is possible to generate functional organ subunits from the differentiated progeny of tissue stem cells. In these studies, self-organizing small intestine, stomach, and colon epithelial structures (termed organoids) were generated from single, isolated Lgr5+ endoderm cells. Importantly, a recent study provides proof of principle that organoids cultured from intestinal stem cells can be used to repair damaged colon epithelium (Yui et al., 2012). This capacity for self-organization is not limited to endodermal derivatives but extends to both complex neuronal structures resembling hypophysis, retinal, and cortical tissues and to mammary gland epithelium (Eiraku and Sasai, 2012; Chanson et al., 2011). Thus, at least some tissue stem cells possess the inherent capacity to achieve 3D complexity when provided with relatively simple biochemical information, even without the multiple inputs that may occur during embryonic organogenesis. This dramatic progress in achieving self-assembly of functional organ subunits from isolated stem cells addresses unresolved questions about the biochemical nature of the inductive interactions between tissue layers during organogenesis. Information essential for in vivo organogenesis comes from the interactions between the epithelial and mesenchymal components of the organ rudiment. Classical studies of organogenesis regarded these interactions as being either permissive or instructive depending on whether they could be replaced with other tissues (e.g., induction of stomach endoderm development by the mesenchyme layer from another organ) (see Grapin-Botton, 2005). In this light, the successful generation of intestinal organoids implies that growth factors, Matrigel, and the epithelial stem cell progeny themselves can substitute for both permissive and instructive mesenchymal contributions once the tissue stem cells have become specied. Moreover, intestinal subunits can also be generated from hESCs and human iPSCs (Spence et al., 2011; see also Kadzik and Morrisey, 2012), demonstrating that the entire process of development from pluripotency to organotypic function can be synthesized in vitro. The advent of stem-cell-derived synthetic organs provides a great leap forward for regenerative medicine. Tissue engineers have previously assembled dispersed cell populations onto manmade biomaterials or decellularized natural matrices as scaffolds to mimic particular body structures (see Badylak et al., 2012). The recent developments in generating organ functional subunits from tissue stem cells suggests focusing attention instead on assembling functional organ subunits into physiologically relevant structures for use in extracorporeal support or for transplantation. It is important to note that for practical applications such assemblies of functional organ units need not have the appearance of the normal organ if they nevertheless recapitulate the organs functions. Finally, recent studies have revolutionized our concept of the origins of phenotype, showing that pluripotent stem cells, intermediate stem cells, or mature differentiated cells can be induced by expression of key components of the transcription factor network governing state-specic gene expression. This suggests that responsiveness to transcription-factor-driven reprogramming is inherent to many developmental states perhaps all of them. Could the mechanisms underlying cellular self-assembly into functional organ subunits be shared across multiple organ systems? By analogy with transcription-factor-driven reprogramming, perhaps a relatively small number of bioactive components provided in vitro will be able to drive selfassembly of different types of tissue stem cells into their respective organ subunits. Only time will tell whether this scenario is in fact feasible, but, if it is, it would imply that cellular phenotype and organ subunit self-assembly programs are actually embedded in the genetic legacy of each species. Decoding these programs would be a boon not only for regenerative medicine, but also for understanding mechanisms of organogenesis, particularly in humans.

REFERENCES Badylak, S.F., Weiss, D.J., Caplan, A., and Macchiarini, P. (2012). Lancet 379, 943952. Chanson, L., Browneld, D., Garbe, J.C., Kuhn, I., Stampfer, M.R., Bissell, M.J., and LaBarge, M.A. (2011). Proc. Natl. Acad. Sci. USA 108, 3264 3269. Eiraku, M., and Sasai, Y. (2012). Curr. Opin. Neurobiol., in press. Published online, March 9, 2012. 10.1016/j.conb.2012.02.005. Grapin-Botton, A. (2005). Int. J. Dev. Biol. 49, 335347. Kadzik, R.S., and Morrisey, E.E. (2012). Cell Stem Cell 10, 355361. Spence, J.R., Mayhew, C.N., Rankin, S.A., Kuhar, M.F., Vallance, J.E., Tolle, K., Hoskins, E.E., Kalinichenko, V.V., Wells, S.I., Zorn, A.M., et al. (2011). Nature 470, 105109. van Vliet, P., Smits, A.M., de Boer, T.P., Korfage, T.H., Metz, C.H.G., Roccio, M., van der Heyden, M.A.G., van Veen, T.A.B., Sluijter, J.P.G., Doevendans, P.A., and Goumans, M.-J. (2010). J. Cell. Mol. Med. 14, 861870. Vierbuchen, T., and Wernig, M. (2011). Nat. Biotechnol. 29, 892907. Yui, S., Nakamura, T., Sato, T., Nemoto, Y., Mizutani, T., Zheng, X., Ichinose, S., Nagaishi, T., Okamoto, R., Tsuchiya, K., et al. (2012). Nat. Med. 18, 618623. Zhou, Q., and Tripathi, P. (2012). Cell Stem Cell 10, 347348.

Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc. 647

Forum
Rethinking Stroma: Lessons from the Blood
David T. Scadden1,*
1Center for Regenerative Medicine and Cancer Center, Massachusetts General Hospital, Harvard Stem Cell Institute, Department of Stem Cell and Regenerative Biology, Harvard University, 185 Cambridge Street, Boston, MA 02114, USA *Correspondence: david_scadden@harvard.edu DOI 10.1016/j.stem.2012.05.011

Cell Stem Cell

Stroma is a largely understudied component of all organs that contributes to stem cell niches. Studies to dene stromal components in the bone marrow have led to some unexpected ndings that prompt further research.
Search stroma and until last year or so, what popped up rst in Wikipedia was an island off the northern coast of Scotland. But who can blame the Wiki writers when most biologists would have trouble getting beyond connective tissue and vessels? It is a grab-bag term that reects a lack of precise determination of components and functions because there has been little detailed investigation. Here, I argue that in this era when we know the sequence of every human gene, it is unreasonable to accept persistent ambiguity of stromal cellular and extracellular constituents: not just because stroma is present in virtually every human organ, but because stroma is central to tissue homeostasis, repair, and disease. Stroma entered the biologic vernacular in the 19th century as microscopists viewed tissues and saw parenchymal cells embedded in a supportive framework. The framework received the name in Latin for a mattress, stroma. It is often interchangeably used with mesenchyma, which is the structural part of a tissue in support of the functional parenchyma. Parenchyma in Greek is literally the visceral esh that as a verb is poured in to mesenchyma. That mesenchyma would be viewed as relatively inert support for early microscopists is not surprising, yet it has long been known to be critical for developing tissues. Mesenchymal interactions with epithelial parenchyma are essential for organogenesis. Mesenchymal cells emerge during gastrulation and become a part of virtually every tissue of metaozoans. They participate in key patterning events determining with precision the identity, number, and organization of cells comprising developing organs and appendages. For example, epithelialmesenchymal signaling feedback loops involving sonic hedgehog (Shh) and FGF nazet are critical for limb development (Be et al., 2009). Specialized regions of mesenchymal cells form the dermal papillae that regulate hair follicle morphogenesis by b-catenin signaling altering FGF and IGF production (Enshell-Seijffers et al., 2010). Branching morphogenesis, important in multiple organ types, is in part controlled by mesenchymal cell islands producing FGF10 in a Shh-regulated manner (Affolter et al., 2003). Yet, in the homeostasis of adult tissues, these critical functional aspects of mesenchymal cells are generally regarded as vestigial and it has only recently been made clear that mesenchyma and stroma are more than architectural support elements. Mesenchymal stromal cells are increasingly appreciated to be heterogeneous, dynamic, and play a regulatory role in parenchymal cell function in adult tissues. This is evident in the regulatory environment for stem/progenitor cells, particularly in hematopoiesis, and hematopoiesis will be the sole focus hereafter for sake of brevity, though other tissues have been studied by others. The stroma of bone marrow has historically been a focus in hematopoiesis research, at least in part due to the limited ability to maintain or grow hematopoietic stem cells outside the body. Michael Dexter rst demonstrated the importance of stroma in coculture experiments that established the now classic method for in vitro hematopoietic stem/progenitor cell (HSPC) support (Dexter et al., 1977). His laboratory neighbor at the University of Manchester, Raymond Schoeld, observed the variable stem cell properties of the spleen colony-forming unit when cultured in isolation and proposed that stroma was also of critical importance in vivo, providing a stem cell niche (Schoeld, 1978), a term he coined in his landmark paper. While the hematopoietic cells that represent the parenchyma of bone marrow are the primary interest, it is the stroma that has become a highly prominent focus for trying to understand the behavior of the hematopoietic cells in both health and disease. The stroma is seen as the critical piece, still veiled, that drives the physiology of the hematopoietic stem cell. In 2003, both my laboratory and that of Liheng Li rst reported the presence of regulatory cells in the stroma using in vivo genetic models (reviewed in Mercier et al., 2012). There are now more than 1,000 papers on the topic according to Scopus. What that work has shown is that the participating parts of the stroma are highly complex, far more complex than the rst naive reports suggested or than invertebrate models suggested. The invertebrate model of a single cell type governing a single stem cell type is not the case in the bone marrow and likely not in other tissue niches as well. Neural and nonmyelinating Schwann cells, endothelial cells, and mature hematopoietic cells like macrophages and possibly osteoclasts all participate in regulating HSPCs (reviewed in Mercier et al., 2012). Even within the mesenchymal cell pool, multiple candidate populations have evidence of modifying HSPC number, quiescence, and/or localization including those expressing CXCL12, osterix, Nestin, leptin receptor, adipocyte markers, and, perhaps, N-cadherin (reviewed in Mercier et al., 2012). This level of complexity in cell type has created ongoing efforts to discern the overlap or distinction among mesenchymal cells and to organize participants hierarchically, for example, which cell type matters in terms of kit

648 Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc.

Cell Stem Cell

Forum
ligand expression (Ding et al., 2012). In addition, it is becoming clear that there is heterogeneity in stem cells and the pairing of stem cell type and niche cell type is undened. Rationally dening the relationships between cell types and the molecules they produce will require a cataloging process that is likely to take years but offers the potential of giving at least a snapshot view of participants in the orchestra. What will we have at the end? It may well be an assembly without sufcient annotation to know what nodal point can be tweaked to modify how the system behaves as a whole. Returning to the orchestra analogy, we will know who is in the chairs and maybe even what instruments they can play, but we will not know how to make music. That would take more of a systems biology approach, investigating multiple components simultaneously over time to dene just how the components are hierarchically related and integrate their activity to provide physiologic outputs. Such a systems-like approach would be enormously demanding in time and resources that could only be worth it if highly distinctive applications might result. The possibility of distinctive applications is suggested by unexpected outcomes of experimental work noted below that was initially intended to address other questions. Dexter taught that stroma was needed to keep cells happy ex vivo and investigation into the niche that followed generally focused on assessing what is needed for homeostasis. However, perturbing stromal cells resulted in odd hematopoietic phenotypes in at least two instances. For example, genetic perturbations of primitive mesenchymal cell subsets (altering RNA processing or ribosomal genes) or more undened populations (deleting RARg) disordered the regulatory environment of the hematopoietic system sufciently to cause a dysplastic and frank malignant state or a hyperproliferative one, respectively (Raaijmakers et al., 2010; Walkley et al., 2007). This disorganization of the hematopoietic parenchyma by mesenchymal dysfunction reveals how critical the relationships are. Those relationships do not just maintain the localization or number of parenchymal stem cells, but also the integrity of parenchymal organization and function. Moreover, the effect was driven by particular mesenchymal cells. Cells expressing osterix altered hematopoietic function, while those expressing osteocalcin did not (Raaijmakers et al., 2010). Perhaps most interesting about the parenchymal malignancy (acute myeloid leukemia) arising in one of the studies (Raaijmakers et al., 2010) is that molecular characterization of the few leukemias that could be studied indicated that they had multiple, new genetic lesions. These data argue that the alteration of stromal function imposes a new set of rules on the parenchymal cells. It changes the way they function, as evident in dysplasia, and it changes what cells thrive. The emergence of malignant cells was presumably due to selection: a selection that seemed to be persistent when the cells were secondarily transplanted. Altered stroma may impose new determinants of tness on the parenchymal cells with which it interacts. It could then enable and perhaps facilitate outgrowth of a neoplastic clone. This model would argue that the multihit hypothesis of cancer could include a hit outside of the cancer cell itself: a cell nonautonomous participant in the emergence of malignancy. Stroma participating in the invasiveness or growth characteristics of malignancy is a longstanding concept about which much has been published, but a model whereby it can be a primary driver of cancer emergence adds a different dimension and enhances the rationale for learning more about just what stroma is and how it works. This is further fueled by the recent recognition that osteolineage cells participating in bone marrow stroma are highly dynamic (Park et al., 2012). They turn over with rapidity not unlike the hematopoietic system and they are replaced by a stem/progenitor pool much like other tissues of rapid cell turnover. Further, the cells can translocate both interstitially and intravascularly so they could theoretically be a population that could acquire a genetic lesion, and propagate that lesion to a large number of descendent cells and distant sites. They could create a eld of stroma imposing different selection pressures on parenchymal cells. Such a scenario may not be a commonplace basis for neoplasia emerging, but it raises the issue of whether stroma might change with age in a way in which acquired genetic lesions within parenchymal cells might be given a competitive edge, providing them a niche in which previously they would not have thrived. It is a hypothesis worth testing particularly as it might provide insight into new vulnerabilities of cancer cells. If stroma is a part of the ecosystem in which cancer emerges, understanding how its dysfunction could lead to the selection of malignant cells raises another possibility. There are few settings where cancer biology focused on the cancer cell as an autonomous unit have provided opportunities for prevention. Perhaps dening how targeted alterations in specic stromal cells select for dysplastic and neoplastic cells will offer such an opportunity. Even the somewhat odd models such as those cited above may give us insight into the signals at work between mesenchymal and parenchymal cells that enable a malignant prone condition to proceed. Stroma has shown itself in the bone marrow to be far more than just the stuff that holds tissues together and it may hold greater secrets still. Teasing stroma apart in other tissues as well as bone marrow, testing its components as disease participants, and dening whether it can be therapeutically targeted is worth our attention, particularly in the context of tissue stem cells.
REFERENCES Affolter, M., Bellusci, S., Itoh, N., Shilo, B., Thiery, J.P., and Werb, Z. (2003). Dev. Cell 4, 1118. nazet, J.D., Bischofberger, M., Tiecke, E., Be Gonc alves, A., Martin, J.F., Zuniga, A., Naef, F., and Zeller, R. (2009). Science 323, 10501053. Dexter, T.M., Allen, T.D., and Lajtha, L.G. (1977). J. Cell. Physiol. 91, 335344. Ding, L., Saunders, T.L., Enikolopov, G., and Morrison, S.J. (2012). Nature 481, 457462. Enshell-Seijffers, D., Lindon, C., Kashiwagi, M., and Morgan, B.A. (2010). Dev. Cell 18, 633642. Mercier, F.E., Ragu, C., and Scadden, D.T. (2012). Nat. Rev. Immunol. 12, 4960. Park, D., Spencer, J.A., Koh, B.I., Kobayashi, T., Fujisaki, J., Clemens, T.L., Lin, C.P., Kronenberg, H.M., and Scadden, D.T. (2012). Cell Stem Cell 10, 259272. Raaijmakers, M.H., Mukherjee, S., Guo, S., Zhang, S., Kobayashi, T., Schoonmaker, J.A., Ebert, B.L., Al-Shahrour, F., Hasserjian, R.P., Scadden, E.O., et al. (2010). Nature 464, 852857. Schoeld, R. (1978). Blood Cells 4, 725. Walkley, C.R., Olsen, G.H., Dworkin, S., Fabb, S.A., Swann, J., McArthur, G.A., Westmoreland, S.V., Chambon, P., Scadden, D.T., and Purton, L.E. (2007). Cell 129, 10971110.

Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc. 649

Forum
Mobilizing Endogenous Stem Cells for Repair and Regeneration: Are We There Yet?
Freda D. Miller1,3,4,5,* and David R. Kaplan2,4
and Stem Cell Biology Program Biology Program Hospital for Sick Children, Toronto M5G 1L7, Canada 3McEwen Center for Regenerative Medicine 4Department of Molecular Genetics 5Department of Physiology University of Toronto, Toronto M5G 1X5, Canada *Correspondence: fredam@sickkids.ca DOI 10.1016/j.stem.2012.05.004
2Cell 1Developmental

Cell Stem Cell

Harnessing endogenous repair mechanisms to promote tissue regeneration in situations in which it does not normally occur has long been a goal in biomedical science. Recent advances in tissue stem cells indicate that this goal may now be achievable. Here we consider both the promise and the hurdles we still have to overcome.
The great hope of the stem cell eld is that we will be able to translate our growing knowledge of stem cell biology and function into therapeutic breakthroughs and applications. In that regard, most of the therapeutic attention has focused upon stem-cell-based transplantation approaches to disorders that range from type I diabetes to spinal cord injury. However, a second potential therapeutic approach has been suggested by the evolving body of work showing that many adult tissues contain resident stem cell populations. It is now clear that these resident stem cells function to maintain and, in some cases, repair tissues and that when they are depleted or dysfunctional, this can cause premature tissue aging and aberrant repair (for examples, see Su et al., 2009; Wagers, 2012). These ndings have led to the idea that if we could somehow recruit endogenous stem cells, then perhaps we could enhance tissue repair or regeneration. On the surface, this is an extremely attractive idea that, if brought to fruition, would in some conditions obviate the need for cell-based transplantation and all of its associated difculties. Here, we will evaluate the reality of this concept by considering where we are right now, and we will discuss both the promise and the remaining hurdles. The idea that resident stem cell populations play essential roles in tissue maintenance is an old one, originating with the early work on hematopoietic stem cells. Moreover, even the idea that precursors such as muscle satellite cells contribute to tissue repair has been around for many decades (the term precursors will be used here to refer to both stem cells and progenitors). What is new, however, is the recognition that these tissue stem cell populations are relatively widespread and that many of them are essential for maintenance of tissue integrity over the life span of the animal. Examples of this range from the skin, where the resident stem cells are responsible for tissue homeostasis, hair growth, and wound healing, to the brain, where endogenous neural precursors support adult neurogenesis but mount an injury response that is unfortunately not particularly successful. However, while it is only a small conceptual step from discovering tissue stem cells to imagining that we could recruit them to enhance repair, this idea has only recently begun to gain momentum. What, then, is the recent experimental evidence that has encouraged many of us to consider this a feasible therapeutic option rather than simply a great idea? Perhaps the strongest data come from studies showing that tissue stem cell behavior can be enhanced in positive ways by the physiology of the animal. One compelling example involves adult rodent central nervous system neural precursors, in which it has been shown that exercise and pregnancy enhance, and stress and aging suppress, precursor proliferation, survival, neurogenesis, and/or oligodendrogenesis in functionally important ways (reviewed in Ming and Song, 2011). Even injury itself can promote differentiation of adult neural precursors into appropriate neural cell types (reviewed in Mitchell et al., 2004). Evidence that some of these ndings may generalize to humans comes from recent imaging studies that show that aerobic exercise training enhances hippocampal volume in elderly humans (Erickson et al., 2011). A second example comes from rodent parabiosis studies that showed that old muscle and neural stem cells can be rejuvenated by exposure to a young circulation and, conversely, that an old circulation ages young stem cells (reviewed in Wagers, 2012). Together, these studies raised the intriguing possibility that if we could understand how physiological cues enhance stem cell function then perhaps we could exogenously manipulate these same mechanisms to promote repair. So, what are these cues? While such studies are still in the early stage, we know that in some cases they are hormones or growth factors that are well studied in other contexts. With specic regard to the previous examples, in rodents the hormone prolactin is a key player in pregnancyinduced neural cell genesis (Shingo et al., 2003), the growth factor BDNF is required

650 Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc.

Cell Stem Cell

Forum
for exercise-induced neurogenesis (reviewed in Ming and Song, 2011), and the chemokine CCL11 is a potent suppressor of neurogenesis in the aged brain (reviewed in Wagers, 2012). Other important growth factor regulators are being dened in work focused on understanding how stem cell niches and/or tissue injury regulate repair. Wnt7a, for example, promotes rodent satellite cell expansion and thus enhances muscle regeneration (Le Grand et al., 2009), and oncostatin M causes dedifferentiation of cardiomyocytes into a more precursor-like state as a necessary prelude to their participation in cardiac remodeling (Kubin et al., 2011). And, nally, a number of growth factors that were originally identied in stem cell culture studies have been shown to activate stem cells in vivo, with one early example being EGF and neural stem cells (reviewed in Mitchell et al., 2004). This body of work provides proof of concept for the idea that physiologically relevant growth factors and hormones recruit endogenous stem cells in positive ways. So why not just use these same growth factors and hormones as we move to the clinic? In fact, in the hematopoietic system, this has been a very successful approach; erythropoietin and G-CSF are used clinically to enhance red blood cell formation and mobilize hematopoietic precursors, respectively. A second success story involves parathyroid hormone, which regulates osteoblast function to promote bone formation and is now used successfully in humans for osteoporosis and bone regeneration (both reviewed in Wagers, 2012). These examples show that growth factors and hormones can work and provide proof of principle that recruitment of endogenous stem cells is a realistic therapeutic strategy in humans. However, in spite of these highly encouraging results, there are also problems associated with growth factors that can make success uncertain. In particular, growth factors are difcult to deliver specically to the desired site of action and are highly bioactive molecules that could have signicant actions on other cell types. As one example, consider BMP2, which promotes osteogenesis from mesenchymal precursor populations. BMP2 has recently been used in spinal fusion surgeries to promote bone healing, but its use is now being questioned in the face of increasing reports of adverse events, including ectopic bone formation and off-target effects on peripheral sensory neurons and the spinal cord (Carragee et al., 2011). As a second example, a little over a decade ago there was considerable excitement about the possibility of using neurotrophic factors to promote neuronal survival and axonal integrity in the injured and degenerating nervous system. Unfortunately, although a signicant number of clinical trials was conducted, none of these growth factors achieved clinical use, in part because of problems with unexpected physiological responses and difculties in obtaining concentrations that were sufcient to activate their cognate Trk receptors in the central nervous system. This is not to say that we shouldnt move ahead with growth factors to promote endogenous stem-cell-mediated repair. Instead, these examples simply provide a cautionary note that such highly bioactive molecules may have unanticipated consequences in man, particularly if we attempt to use them for chronic conditions. As an alternative to growth factors and hormones, many in the eld have asked whether they can instead identify pharmacological candidates that modulate stem cell behavior. Pharmacological agents have a number of potential advantages. They can be selected or designed to target one aspect of stem cell physiology or function or a signaling pathway, can be administered by different routes, and can be modied to reduce toxicity and offtarget effects. In this regard, two distinct strategies are being pursued. One is a direct spinoff of work dening stem cell regulatory pathways that takes advantage of the many drugs that target relevant signaling pathways that are already in use in humans for other conditions. The second strategy is an unbiased one that involves screening human stem cell populations with large compound libraries, looking for readouts of bioactivity such as enhanced proliferation or differentiation. In this regard, a number of highprole chemical stem cell screens have been published in the recent past (reviewed in Wagers, 2012), primarily looking for molecules that can modulate stem cell behavior in culture as a prelude to transplantation-based approaches or that can promote the survival and/or integrity of pluripotent stem-cell-derived progeny. While this work was largely not focused upon nding chemical modulators of endogenous repair, it nonetheless establishes the feasibility of such screens. Assuming, then, that as a community we are, and will continue to be, moving ahead with such screens, then what are we looking for? What do we mean if we say that we would like to identify compounds that recruit endogenous tissue stem cells to promote repair and regeneration? Do we mean that we want to activate quiescent stem cell populations to enhance proliferation and/or survival of the stem cell and more biased progenitor populations, or do we want to direct differentiation? These questions might best be answered by considering how tissue stem cells normally repair tissues, but unfortunately we still know relatively little about this issue. In one scenario, upon receiving an injury signal, the parent stem cells would increase their symmetric, self-renewing divisions to enhance the size of the stem cell pool, and these would then respond appropriately to environmental cues to generate an increased number of the right kinds of progenitors and differentiated progeny. A second scenario is that the injury environment acts directly upon the appropriate progenitor population, leading to an expansion of that population and a resultant increase in the number of differentiated progeny. Numerous variations upon these scenarios are possible (and even likely), in which the environment regulates everything from proliferation to survival to lineage bias. What, then, should we ask for and measure in our screens to enhance our chances of in vivo success? The answer to this question will likely differ depending upon whether the tissue of interest can normally repair itself or not. For example, skin and bone can repair themselves under optimal conditions, and when skin-derived precursors (SKPs), the dermal precursors we have characterized, are transplanted into injured skin or bone, they appropriately generate dermal broblasts, myobroblasts and adipocytes or chondrocytes and osteocytes, respectively (Biernaskie et al., 2009). Thus, in these tissues, the injury signals are sufcient to faithfully direct either endogenous or exogenous mesenchymal precursors to produce the right functional cell types. In these tissues it may therefore be sufcient to simply expand precursor pools. In contrast, in

Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc. 651

Forum
the injured central nervous system, where repair is very limited, it may be necessary to use compounds that both enhance the stem cell and/or progenitor pools and promote differentiation into benecial cell lineages such as oligodendrocytes. These considerations suggest that, as a rst pass, we should identify pharmacological agents that mobilize and expand the endogenous stem or progenitor pools and that we should do so for tissues in which we already know that physiological cues can recruit endogenous precursors to promote repair and/or in which transplanted precursors will differentiate and integrate appropriately. To ensure success, we should also probably choose systems such as bone, skin, or eye, where we can realistically deliver molecules and assess outcomes. Alternatively, if we do use such approaches for less-accessible organs like the brain, then we need to ensure that we can measure statistically signicant positive outcomes using imaging approaches and/or behavioral assessments. In this regard, conditions like acute spinal cord injury or stroke may not be the best test cases, because short-term outcomes are highly variable in these populations. So, are we there yet with regard to recruiting endogenous tissue stem cells for repair and regeneration? The aforementioned successes in the hematopoietic system argue that, if we are thoughtful and careful in our choice of clinical targets, there is a high probability of success as long as we can identify compounds that can appropriately regulate tissue stem cell biology. Moreover, identication of appropriate bioactive molecules, whether they be growth factors or chemical entities, will be greatly facilitated by our ever-increasing knowledge of endogenous stem cell regulatory mechanisms and by the technology that is allowing high-throughput stem-cell-based screening. Thus, while we are not yet there, we are certainly getting close. Nonetheless, there are still a number of signicant considerations that we must address as we move forward. First, because adult tissue stem cells are long lived but not immortal, will long-term exogenous activation lead to stem cell depletion? For example, genetic depletion of dermal precursors in mice resulted in premature skin aging and wound healing (Su et al., 2009). Perhaps pharmacological mobilizers of endogenous stem cells will have to be used at doses that induce stem cell proliferation but not depletion. Alternatively, perhaps it would be better to promote expansion of progenitors rather than the stem cells themselves. Indeed, this is the strategy that is used by the hematopoietic system, in which large numbers of differentiated progeny must be generated every day. Second, will pharmacologically activated stem cells generate the appropriate progeny, and will these progeny survive and integrate? As discussed above, it is likely that in at least some situations the tissue environment will do this for us. However, it is unclear whether this will also be true for chronic injuries and/or degenerative conditions in which stem cells are exposed to a hostile environment. For example, the parabiosis experiments (Wagers, 2012) suggest that during aging, tissuederived signals that promote stem-cellmediated repair are depleted or nonfunctional. How will we overcome these environmental barriers? One idea is that by reprogramming endogenous precursors genetically, we can induce their differentiation into a specic lineage. This is certainly an exciting possibility, but it is somewhat further from than the clinic than are pharmacological manipulations. Third, if tissue stem cells acquire genetic mutations over the course of an individuals lifetime, and if we promote self-renewal of these perturbed stem cells in aged individuals, then do we increase the risk that they will take a second genetic hit, leading to unregulated growth or tumorigenesis? Arguing against this possibility is the fact that adult tissue stem cells seem to have very strong senescence mechanisms that would likely come into play in this scenario (for example, see Su et al., 2009). Moreover, there is little evidence in animal models and/or even in humans that enhanced tissue stem cell activation causes increased tumorigenesis. Nonetheless, we will have to be particularly alert to this possibility when we consider chronic treatment of aged individuals. The recruitment of endogenous repair mechanisms to promote tissue regeneration in situations in which it does not normally occur has long been a goal in biomedical science. Our recent appreciation that one key repair mechanism involves adult tissue stem cells has opened up a new way of thinking about this goal and a new way of moving forward. The approach envisioned here will not replace stem-cell-based transplantation, because there are many situations in which the endogenous repair capacity will just not be sufcient and/or the endogenous stem cells will themselves be impaired. Nonetheless, such an approach provides great hope for many devastating disorders that are currently untreatable.
ACKNOWLEDGMENTS We would like to apologize to all the authors of primary publications whose work is alluded to in this commentary but not directly referenced because of length limitations. We would also like to thank Drs. Michael Rudnicki, Fabio Rossi, Derek van der Kooy, Ben Alman, and Peter Zandstra for their input while we were writing this article. REFERENCES Biernaskie, J., Paris, M., Morozova, O., Fagan, B.M., Marra, M., Pevny, L., and Miller, F.D. (2009). Cell Stem Cell 5, 610623. Carragee, E.J., Hurwitz, E.L., and Weiner, B.K. (2011). Spine J. 11, 471491. Erickson, K.I., Voss, M.W., Prakash, R.S., Basak, C., Szabo, A., Chaddock, L., Kim, J.S., Heo, S., Alves, H., White, S.M., et al. (2011). Proc. Natl. Acad. Sci. USA 108, 30173022. ling, J., Kostin, S., Gajawada, P., Hein, Kubin, T., Po S., Rees, W., Wietelmann, A., Tanaka, M., rchner, H., Schimanski, S., et al. (2011). Cell Lo Stem Cell 9, 420432. ` , A., and Le Grand, F., Jones, A.E., Seale, V., Scime Rudnicki, M.A. (2009). Cell Stem Cell 4, 535547. Ming, G.L., and Song, H. (2011). Neuron 70, 687702. Mitchell, B.D., Emsley, J.G., Magavi, S.S., Arlotta, P., and Macklis, J.D. (2004). Dev. Neurosci. 26, 101117. Shingo, T., Gregg, C., Enwere, E., Fujikawa, H., Hassam, R., Geary, C., Cross, J.C., and Weiss, S. (2003). Science 299, 117120. Su, X., Paris, M., Gi, Y.J., Tsai, K.Y., Cho, M.S., Lin, Y.L., Biernaskie, J.A., Sinha, S., Prives, C., Pevny, L.H., et al. (2009). Cell Stem Cell 5, 6475. Wagers, A.J. (2012). Cell Stem Cell 10, 362369.

Cell Stem Cell

652 Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc.

Cell Stem Cell

Forum
Why Is It Taking So Long to Develop Clinically Competitive Stem Cell Therapies for CNS Disorders?
Olle Lindvall1,*
1Lund Stem Cell Center, University Hospital, SE-221 84 Lund, Sweden *Correspondence: olle.lindvall@med.lu.se DOI 10.1016/j.stem.2012.04.004

The remarkable advancements in basic stem cell research with implications for several central nervous system disorders have so far not been translated into clinically effective therapies. Here I discuss some of the underlying problems and how they could be overcome.
Introduction The rst attempt to treat a central nervous system (CNS) disorder with cell transplantation took place three decades ago (Backlund et al., 1985). In this study, autologous adrenal medulla cells were implanted into the striatum of Parkinsons disease (PD) patients to provide a local catecholamine source, but the benecial effects were minimal. A few years later, human fetal mesencephalic tissue rich in dopaminergic neuroblasts was transplanted to the striatum in PD patients. These clinical trials established some important basic principles of cell therapy for CNS disorders: grafted neurons can replace dead host neurons in the diseased, 50- to 60-year-old human brain, reinnervate denervated areas, release transmitter, and, in some patients, give rise to therapeutically valuable effects (Lindvall and Kokaia, 2010). Based on these ndings, stem-cell-based therapy for PD has been regarded as a lowhanging fruit, with the requirement for successful treatment being seemingly simple, namely to generate large numbers of standardized dopaminergic neurons for transplantation from stem cells. However, despite major efforts in basic and clinical research, there is still no clinically competitive cell therapy for PD or any other CNS disorder. Clinical trials with stem cells, often of bone marrow origin, are ongoing in, e.g., stroke, amyotrophic lateral sclerosis (ALS), and spinal cord injury (http:// www.clinicaltrials.gov), but whether they will show efcacy is unclear. From my perspective, there are several major problems that explain why the clinical translation of stem cells for neurological disease is so difcult, as outlined below. The Problem of Generating the Right Cells and Understanding Their Mechanisms of Action Stem cells can act in brain diseases by replacing those cells that have died, but they can also restore function through other mechanisms (Lindvall and Kokaia, 2010). In the case of cell replacement, disease pathology determines which cells have to be generated from the stem cells. Different cells will be needed for different diseases. Substantial improvement in PD and ALS will require cells with the properties of dopaminergic and motor neurons, respectively. The situation for cell replacement in Alzheimers disease (AD) is much more complex because the stem cells would have to be predifferentiated in vitro into many different types of neuroblasts for subsequent implantation into a large number of brain areas. Similarly, in stroke there is a loss of several different types of neuron, glial cells, endothelial cells, and parenchyma. These broad defects raise the question of whether it is realistic to expect that clinically valuable improvement in disorders like AD or stroke could be achieved through cell replacement. Importantly, efcacious cell replacement will require the generation of the correct neuronal phenotype. For example, in PD it is not sufcient to generate just any type of dopaminergic neuron. Rather, to induce substantial clinical benet, the human stem-cell-derived dopaminergic neurons must exhibit the specic properties of the neurons that have died, i.e., the substantia nigra neurons (Lindvall et al., 2012). A recent study did succeed in showing efcient conversion of human embryonic stem cells into bona de substantia nigra dopaminergic neurons using a differentiation protocol guided by developmental principles (Kriks et al., 2011). These cells ameliorated PD symptoms after transplantation in animal models without forming tumors. For optimum recovery in many CNS diseases, neuronal replacement and at least partial reconstruction of circuitry should probably be the long-term goal. However, a large number of experimental studies in animal models of these disorders have demonstrated that stem cell delivery gives rise to functional improvements that cannot be explained by neuronal replacement. These benecial effects may also be relevant in clinical settings. For example, systemic or intracerebral delivery of neural and other stem cells in stroke models has been reported to lead to improvements by trophic actions, modulation of inammation, promotion of angiogenesis, remyelination and axonal plasticity, and neuroprotection (Lindvall and Kokaia, 2010). The functional effects can be enhanced if the stem cells have been genetically modied to secrete various factors such as trophic molecules. For clinical competitiveness, it is necessary, though, that the efcacy and safety of the stem-cell-based approach is superior to that of available treatments (e.g., drugs) acting on the same targets. Clinical trials are ongoing in stroke and ALS with delivery of stem cells, which are intended to act not by neuronal replacement but instead through one or more of the other presumed mechanisms. However, it is conceivable that effective therapies will not be developed until the mechanisms of action of the

660 Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc.

Cell Stem Cell

Forum
stem cells are much better understood and can therefore be optimized. The Problem of Using the Right Animal Model and Behavioral Tests Available animal models of CNS diseases do not mimic all aspects of the pathology of the human condition, which may explain lack of efcacy of cell therapy when it is translated to the clinical setting (Lindvall et al., 2012). For example, animal models of PD are mostly based on lesions of the nigrostriatal dopaminergic pathway, induced by toxins, and studies of sensorimotor functions. These models do not imitate the clinical disorder, which has many nonmotor and motor features with nondopaminergic pathology outside the substantia nigra. Attempts to develop transgenic models of PD have been pursued in recent years, but these represent only partial models of the core pathologies. For efcient clinical translation, better animal models that reect the complex pathology and pathogenesis of CNS disorders accurately have to be developed through collaboration between basic scientists and clinicians. Many current models use otherwise healthy, young animals, which again is distinct from the clinical situation in many neurodegenerative diseases, where patients are often older, with concurrent diseases and chronic medication. For example, stroke patients frequently also suffer from hypertension and diabetes. The animal models may not be able to fully predict the adverse events, toxicity of the cell product, immune and other biological responses, and risk for tumor formation that would occur after implantation of cells into patients. A lesson can be learned from the clinical trials with fetal dopaminergic cell therapy in PD. When troublesome graft-induced involuntary movements (so-called dyskinesias) were observed in patients (Freed et al., 2001), this side effect came as a surprise because none of the preclinical studies in rodent and primate models of PD had observed any adverse responses of this type. The risk of tumor formation from cells derived from pluripotent cells also makes clinical translation difcult. For example, life expectancy is virtually normal in PD patients, and therefore even a minor risk of tumor formation associated with stem cell therapy would be unacceptable. It is difcult to assess the clinical tumor risk with human embryonic stem cell derivates using preclinical et al., 2003). xenograft studies (Erdo Thus, for clinical translation, there will need to be rigorous mechanisms for determining the tumorigenicity of stem cells and their derivatives. A prerequisite for application in patients must be a demonstration in an animal model that a given cell-based approach induces substantial improvement of clinically relevant functional decits (Lindvall et al., 2012). For example, in rodent models of PD, behavioral improvement after stem cell therapy is often reported as a reversal of rotational asymmetry in animals with unilateral lesions of the nigrostriatal dopaminergic system. While this test gives a good measure of the dopamine-releasing capacity of the grafts, the decit does not reect any symptom seen in PD patients. Other behavioral tests are available but have only been used in few studies. Basic scientists and clinicians together have to develop functional and behavioral tests that assess decits in animals resembling the impairments in patients with CNS disorders. The Problem of Distribution and Progression of Pathology Even if stem cells improve function in a specic area by neuronal replacement or other mechanisms, effective therapy is hindered if there is concurrent degeneration in other brain regions or if such changes develop after transplantation. For example, dopaminergic denervation in areas not reached by the intraputaminal grafts, such as the ventral striatum, in PD patients with fetal dopaminergic grafts counteracts the symptomatic relief following transplantation (Piccini et al., 2005). Similarly, even if replacement of motor neurons in the spinal cord of ALS patients did work, central motor neurons such as corticospinal neurons, which also degenerate in ALS, would most likely have to be replaced for effective, life-saving restoration of function. For successful, long-term clinical efcacy of stem cells in chronic neurodegenerative disorders, patient selection will be crucial, and neuronal replacement probably has to be combined with a neuroprotective therapy to hinder disease progression. In chronic neurodegenerative disorders, host pathology may also affect the cells derived from the transplanted stem cells, as has been observed in fetal grafts after implantation in PD and Huntingtons disease patients (Kordower et al., 2008; Cicchetti et al., 2009). This consideration may be particularly relevant when patient-specic cells for transplantation are produced by therapeutic cloning, from induced pluripotent stem cells, or by direct conversion of somatic cells. Such cells could exhibit increased susceptibility to the neurodegenerative disease process. In the case of PD, this problem may not be a serious one, because with fetal grafts the propagation of disease pathology is slow, the majority of grafted neurons are unaffected after a decade, and the patients can experience long-term improvement. The Problem of Translating Basic Research Findings to Patients A major problem hindering effective translation is, in my view, insufcient communication between basic scientists and clinicians. My own experience as a clinical neurologist is that the clinic and the basic research laboratory are often completely different worlds. For basic stem cell research to have more impact on the clinical challenges, clinicians have to be involved from an early stage and not just immediately before application in patients. Basic scientists should be educated in the clinical features of CNS disease and the problems related to diagnosis and therapy. The critical scientic steps from basic research to patient application should be dened through cooperation between basic scientists and clinicians. This partnership must function throughout all stages of clinical translation if basic research ndings are to be efciently converted to novel treatments for CNS disorders. The new imaging techniques for monitoring brain and spinal cord in vivo in animals and humans will create golden opportunities for fruitful interaction between basic scientists and clinicians. It is important to emphasize that successful clinical application of stem cells will depend not just on the generation of the right type of cell but also on several other factors, such as appropriate site of delivery of the cells and selecting the suitable patient.

Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc. 661

Cell Stem Cell

Forum
The Problem of Competing Therapeutic Approaches There is considerable variation in terms of the availability of existing therapeutic options for different CNS disorders, and these differences will inuence how quickly stem cells can be translated to the clinic. For example, to be clinically competitive in PD, grafts must give rise to major recovery (at least 70%) of motor function. Motor symptoms in PD patients can already be treated quite well with L-dopa, DA agonists, enzyme inhibitors, and deep brain stimulation. Thus, the efcacy of stem cell grafts in relieving disease symptoms would need to be high. If transplantation of stem-cellderived dopaminergic neurons in a PD patient gave only a 30% reduction in motor symptoms, it would be regarded as scientically exciting but clinically useless. The efcacy of currently available human stem-cell-derived dopaminergic neurons and predictions for the clinical setting are unclear, presenting a problem for clinical translation. As a rst step toward patient application, a cell-potency assay should be used to compare the efcacy of the stem-cell-derived neurons versus equivalent fetal dopaminergic neurons (which can be regarded as the gold standard) in appropriate animal models of PD. Many brain diseases, however, lack effective current treatments. Several such diseases are progressive and ultimately fatal, such as ALS or Huntingtons disease. In these conditions, even a minor improvement induced by stem cells would be clinically useful. If efcacious therapy is lacking, the severity of a disease such as ALS or Huntingtons disease might justify the risks of a stem-cell-based experimental intervention in patients. It should be emphasized, however, that even when there is no effective alternative therapy, no application in patients can be justied if it does not have proven efcacy in the laboratory and scientic understanding of the mechanism of action. For these CNS disorders, careful, laborious, and time-consuming preclinical studies are also required. Clinical trials showing safety alone, without any scientic grounding for their use, are unethical. The Problem of Costs Stem-cell-based treatments for CNS disorders should not only relieve human suffering but also be cost-effective compared to other therapies. To promote clinical translation, scientists should perform health economics studies at an early stage to estimate the potential value of further research in stem cell therapy for various disorders in order to ensure that society makes the best use of research investments. Using health economics modeling and a range of assumptions, it is possible to determine which patients should be targeted with stem cell therapy. Moreover, such modeling will give a price at which the intervention would be cost neutral, i.e., the stem cell therapy would bear its own cost from a societal perspective. This estimated price for stem cell therapy will be important for companies manufacturing the stem-cell-based product to be delivered to the patient. Translation of discoveries in basic stem cell research into safe and effective clinical products for CNS disorders will be very expensive. The European Court of Justice recently decided that no patents can be granted for inventions based on human embryonic stem cells, even if the cell lines were established in the laboratory many years ago and the invention itself does not involve obtaining new embryonic stem cells. This decision may well cause companies in Europe to be reluctant to invest in translational stem cell research because they would be unable to protect their procedures via the patent system. The end result will unfortunately be further delay in the development of clinically effective stem cell therapies for CNS disorders. Conclusions Many CNS disorders in humans currently lack effective treatments, but there is now reason to be optimistic. Experimental studies have clearly indicated that stem cells have the potential to give rise to radical new therapies for these diseases. However, there is no fast track for stem cells to the clinic. Strong investigative basic research remains fundamental for clinical advancement of stem-cell-based approaches. For efcient clinical translation, a road map to the clinic, taking into account the critical scientic, clinical, regulatory, and ethical issues, should be dened and continuously revised by basic scientists and clinicians together. The commitment must be long term, and the aims must be realistic. The biological problems that will be encountered along the way are complex and should not be underestimated.

REFERENCES Backlund, E.O., Granberg, P.O., Hamberger, B., rtensson, A., Sedvall, G., Seiger, Knutsson, E., Ma ., and Olson, L. (1985). J. Neurosurg. 62, A 169173. Cicchetti, F., Saporta, S., Hauser, R.A., Parent, M., Saint-Pierre, M., Sanberg, P.R., Li, X.J., Parker, J.R., Chu, Y., Mufson, E.J., et al. (2009). Proc. Natl. Acad. Sci. USA 106, 1248312488. , F., Bu hrle, C., Blunk, J., Hoehn, M., Xia, Y., Erdo cking, M., Ku stermann, E., Fleischmann, B., Fo Kolossov, E., Hescheler, J., et al. (2003). J. Cereb. Blood Flow Metab. 23, 780785. Freed, C.R., Greene, P.E., Breeze, R.E., Tsai, W.Y., DuMouchel, W., Kao, R., Dillon, S., Wineld, H., Culver, S., Trojanowski, J.Q., et al. (2001). N. Engl. J. Med. 344, 710719. Kordower, J.H., Chu, Y., Hauser, R.A., Freeman, T.B., and Olanow, C.W. (2008). Nat. Med. 14, 504506. Kriks, S., Shim, J.W., Piao, J., Ganat, Y.M., Wakeman, D.R., Xie, Z., Carrillo-Reid, L., Auyeung, G., Antonacci, C., Buch, A., et al. (2011). Nature 480, 547551. Lindvall, O., and Kokaia, Z. (2010). J. Clin. Invest. 120, 2940. stle, O., Isacson, O., Lindvall, O., Barker, R.A., Bru and Svendsen, C.N. (2012). Cell Stem Cell 10, 151155. Piccini, P., Pavese, N., Hagell, P., Reimer, J., rklund, A., Oertel, W.H., Quinn, N.P., Brooks, Bjo D.J., and Lindvall, O. (2005). Brain 128, 2977 2986.

662 Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc.

Cell Stem Cell

Forum
Modeling Brain Disease in a Dish: Really?
Maria C. Marchetto1 and Fred H. Gage1,*
1Salk Institute, 10010 North Torrey Pines Road, La Jolla, CA 92037, USA *Correspondence: gage@salk.edu DOI 10.1016/j.stem.2012.05.008

Cellular programming and reprogramming technology (CPART) presents a novel approach for understanding disease progression and mechanism. In addition, CPART provides an innovative opportunity for developing diagnostic tools and novel drug candidates for therapy. In this Forum, we will discuss obstacles and solutions for modeling brain disease using CPART.
Introduction Cellular programming and reprogramming technology (CPART) has provided a new way to investigate human development and disease. This technology is particularly useful for diseases in which the affected tissue is not available for cell purication and in which aspects of cell development are crucial for the pathology. The central nervous system (CNS) is a good example of tissue that falls into this category. Modeling human brain diseases using induced pluripotent stem cells (iPSC) or induced neural cells (iN) has remarkable potential to generate insights into understanding disease mechanisms and opening new avenues for clinical intervention. Researchers now have the opportunity to study human disease in living, developing neural cells that carry the disease-specic genetic variants that are present in the patient. In addition, CPART represents a fresh approach for developing original diagnostic tools and obtaining novel drug candidates for CNS therapy. Importantly, candidate compounds for treating CNS defects fail in clinical trials in over 90% of cases due to poor targeting, lack of efcacy, and unacceptable side effects (Kola and Landis, 2004). We rmly believe that CPART can offer a valuable additional tool for screening and validating CNS compounds for pharmaceutical companies in the future. In order for CPART to be successful and useful, our underlying and somewhat provocative assumption is that cellular deciencies can be measured in vitro, recapitulating the phenotype that is relevant to human brain disease. These cellular deciencies can be broken down into levels of relevance, ranging from highly relevant recapitulations of clinical observations to phenotypes not previously associated with the disease and to lack of in vitro phenotype all together. Ideally, the modeled in vitro phenotype is the same as the in vivo disease-causing phenotype that allows for the study of the mechanism and development of the disease. A less-pertinent model would result in an in vitro phenotype that is similar mechanistically but is clearly not the same as the in vivo phenotype and likely develops differently. This less-stringent level of relevance could still be used to screen compounds for reversal of phenotypes and could be used as a diagnostic tool. The least-stringent level of relevance would involve a phenotype that is unrelated in any known mechanistic way to the in vivo human disease but that could still potentially be used as a tool for the diagnosis of the disease. All these levels of relevance require the basic assumption that in vitro modeling is robust enough to detect a reliable and statistically meaningful difference between phenotypically normal and abnormal cells derived using CPART (Figure 1). This is particularly important in our view because there could be diseases in which the variability of phenotypes will be too big to achieve reliable statistical value. In these specic cases, more individuals could be included and/or more sensitive assays should be employed. Nonetheless, as promising as this technology is, we are still in its early days. There are a number of roadblocks to be considered, and eventually overcome, before signicant progress on disease mechanisms can be made and meaningful therapies can be proposed. In this Forum, we will discuss some of the current obstacles, beginning with reprogramming a patient biopsy and moving toward modeling disease-relevant phenotypes and proposing effective therapies (Figure 2). Methods for Generating Reprogrammed Cells Ever since the original procedure for producing iPSC was described by Takahashi and Yamanaka using four factors separately in retroviruses (Takahashi and Yamanaka, 2006), the exact genes used and the delivery methods have been changing. While the mechanisms by which the reprogramming occurs remain incompletely understood, the robustness of the general approach is so clear that similar (though not exact) results can be achieved using a variety of methodologies. Interestingly, the vast majority of literature still employs viral integrating strategies. Random viral integration and incomplete silencing of reprogramming factors are likely among the major generators of variability between cell lines. We favor the idea that nonintegrative methods of reprogramming will be the best way to derive new cell lines in the future. The eld will likely reach a consensus on some combination of soluble molecules and transient nonintegrating vectors that carry a small cluster of genes. The limitations on types of starting cells have also evolved from the original broblast to include many other somatic tissues of the body. One of the important developments has been direct programming of somatic biopsy cells (usually broblasts) to neural cells (reviewed in Zhou and Tripathi, 2012). The initial reports of direct cellular programming revealed the disadvantage of not generating a renewable source of programmed cells, but recently several labs have shown that programming can be achieved to a proliferating population of neural precursor cells (iNPC) that can then be propagated

642 Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc.

Cell Stem Cell

Forum
readily available nowwill reduce the variability of the cultures. Partial reprogramming and incomplete erasure of epigenetic marks has also been proposed as a source of variability (Hayden, 2011), though some propose that this specic concern could be addressed by developing more stringent methodologies of selecting populations of REPRO cells that have undergone complete reprogramming. In the literature, few attempts have been made to implement general criteria for dening hallmarks for reprogramming and differentiation, but, given the fast pace of discovery in the eld, more research is denitely needed to precisely dissect and deal with many sources of variability. An example of the dynamic nature of the iPSC characterization is the use of teratoma formation assays as the gold standard criterion for pluripotency validation. We support the view that teratoma generation is primarily qualitative, is difcult to standardize, and has questionable value for in vitro disease ller et al., 2010). modeling (discussed in Mu While broadly used as a reprogramming hallmark, many researchers are now questioning its value, particularly in light of reports showing that partially reprogrammed iPSC lines can also form teratomas (Chan et al., 2009). Another importantand yet not fully understoodsource of variability is the naturally occurring intrinsic variability between individuals. Specically relevant to modeling mental illnesses will be the ability to dissect the phenotypic variability present in neuronal cells derived from nonaffected individuals from the clinically relevant differences present in neurons from affected patients. One obvious way of addressing variability is to increase the number of controls and patients analyzed in combination with specic cohorts of patients who present common clinical histories and/or respond similarly to drugs. When possible, the use of genetically identical individuals (i.e., monogenetic twins) who are concordant or discordant for a mental disorder would also contribute to reducing variability and generating relevant disease hypotheses. In addition, generation of disease team consortiums would greatly facilitate CPART by enhancing the number of individuals analyzed (and new lines produced), promoting exchanges of cells and methodologies, and improving discussion.

Figure 1. Different Possible Outcomes in Disease Modeling using Cellular Programming and Reprogramming Technology
Different possible outcomes in disease modeling using cellular programming and reprogramming technology (CPART) and its relevance for effectively understanding disease mechanism, developing diagnostic tools, and nding novel therapies.

and subsequently differentiated into mature neurons and glia (Zhou and Tripathi, 2012). For the purpose of this review, we will use the term REPRO cells to represent both reprogrammed (iPSC) and programmed (directly differentiated) cells such as iNPC and iN. Accessing the Fate of REPRO Cells While the choice of donor cells and the method for generating the REPRO cells are critical considerations, another challenge is their differentiation into homogeneous populations of the cell type implicated in the disease phenotype. This is a rich eld, and there is a steady ow of protocols for generating specic cell types, but much more work is required. Most strategies for cell fate specication are based on the achievements of in vivo developmental neurobiologists in the last 50 years. While the molecular pathways for generating specic cell types in the mouse do not always recapitulate a human in vitro setting, many of the pathways involved in cell fate decisions are conserved. Moreover, there is a surge in the drive to differentiate human REPRO cells to specic cell lineages, which will, in turn, provide the foundation for new knowledge about human neural development. An important cautionary note is that, with our existing technology, it is particularly challenging to generate mature human neurons in the dish. Most protocols currently available produce neurons that are comparable to cells at

an immature fetal stage, although a lot of research is now dedicated to improving the neuronal outcome in terms of maturity and percentage of electrophysiologically active cells. Finally, much of the effort in vitro has been to establish protocols for differentiating cells into specic cell types through soluble molecules, including growth factors and activators or inhibitors of specic signaling pathways. For more complete fate determination, we will likely need to pay more attention to combinations of cell types and structures, including vasculature, which are critical in vivo for fate specication and maintenance. Variability between Lines As more iPSC lines are produced and shared, it has become increasingly clear to stem cell biologists that there is variability between cell lines that emerged after reprogramming. Differences between clonal lines have been detected at multiple levels, including variable expression prole levels, X-inactivation status on female lines, genetic instability, and differentiation potential (discussed in Hayden, 2011). This observed variation could be attributed in part to the current protocols of reprogramming that include viral integration into the host DNA, as well as the intense cloning involving multiple passages of the cells that is required to apply CPART. It remains to be seen whether the widespread use of nonintegrating technologywhich is

Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc. 643

Cell Stem Cell

Forum
to establish the basic tools for culturing functional human neurons, and revealed meaningful neuronal phenotypes (Marchetto et al., 2010). Importantly, monogenic disorder modeling presents the opportunity to perform gain- and loss-offunction studies to conrm that the neuronal phenotypes observed are disease specic, as opposed to a general, nonspecic effect. The advantages of studying monogenic diseases are many; not the least important is proving, in an isogenic nondiseased cell, that the homologous recombination of the diseased gene is necessary and sufcient for recapitulating the in vitro disease phenotype. Nonetheless, a large percentage of brain disorders (i.e., autism and schizophrenia) have a complex nature and are likely multifactorial: a combination of mutations in several genes (as opposed to mutations in only one gene) and extrinsic factors (inuence of neighbor cells and environment) are involved in the disease pathology. Crucial for the future of modeling complex CNS diseases will be developing novel, more-sensitive tools and improving methodology and analysis such that they will consistently pick up subtle but important differences between controls and affected cell lines (reviewed in Marchetto et al., 2011). New neuronal differentiation protocols can obtain particular subtypes of neurons that are relevant to particular diseases (e.g., dopaminergic neurons for Parkinsons disease, hippocampal and cholinergic neurons for Alzheimers disease, motor neurons for amyotrophic lateral sclerosis). In addition, improving the protocols for evaluating neuronal connectivity properties, synaptic plasticity, and electrophysiological functional outcomes will denitely be of great value in detecting disease-related phenotypes. Examples of techniques that are already available but not yet widely employed in the eld are calcium imaging, light-activated channelrhodopsins, uncaged glutamate, transynaptic labeling using virus, multielectrode arrays, single-cell expression analysis, high-resolution live imaging for spine motility and maturation, synaptic protein recruitment, and axonal transporting dynamic visualization. Combining ideas and technologies from established elds will be highly benecial to this nascent eld. A practical

Figure 2. Obstacles to Be Overcome before Cellular Programming and Reprogramming Technology Can Be Widely Used for Disease Modeling
Obstacles to be overcome before CPART can be widely used for understanding the mechanisms of, and proposing meaningful therapies for patients with, brain diseases. REPRO cells is a term used in this forum to reference both reprogrammed and programmed (direct differentiated) cells such as induced pluripotent stem cells (iPSC), induced neural progenitor cells (iNPC), and induced neurons (iN).

A very important methodology for this eld is statistics. Given the aforementioned variability in the eld, we support the use of multiple cell lines from each patient and ideally multiple patients with variable phenotypes. Careful analysis of the variance present in the data will indicate the minimal number of REPRO patient cell lines needed to obtain a statistically relevant measurement. It is likely that the number of cell lines needed will vary accordingly with the robustness of the assay performed. Sources of error with low numbers of subjects are plentiful in the eld and increase the risk of both Type 1 and Type 2 errors. The success of this eld depends on the ability to replicate ndings from one laboratory to another. One of the great advantages of this eld is that the principle reagent (the REPRO cell) can be banked and made available to other laboratories, meaning that we have an ever-expanding and -renewable source of research material that we can use to go back and compare, contrast, and add numbers to our analysis. Thus, while variability in methods of CPART and differentiation may mask or misrepresent some initial ndings, the

robust and meaningful phenotypes will emerge and persist, and the less-robust observations will not. A further note of caution about variability relates to the variability that exists between individual cells from the same clone. This variability includes expression differences based on variable states of differentiation between, for example, two related neurons next to each other in the culture dish; other variance may be genomic and may result from intrinsic mechanisms of genomic mosaicism, attributed in part to mobile elements or driven by extrinsic factors such as the number of passages, which increases the risk of differential genome instability. It is likely that this is a source of variance that will not be resolved until more advanced methods of genomic analysis are available. Modeling Brain Disease Most successful reports involving disease modeling for neurological diseases to date utilize cells from patients with monogenetic disorders in which the gene mutation is characterized. Modeling monogenic brain disorders advanced the eld, helped

644 Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc.

Cell Stem Cell

Forum
example is the recent incorporation of biomaterials and bioengineering techniques for improved differentiation of iPSC cultures. New alternative methods for better compartmentalization and isolation of neuronal processes using microuidic chambers have been explored and implemented in primary neural cultures. Compartmentalization of neurons using engineered devices would allow comparisons of the dynamic behavior and molecular anatomy of control versus diseased neurons in culture. Additionally, using engineered tridimensional biomatrices to simulate tissue structures may more authentically recapitulate in vivo neuronal branching and connectivity and potentiate a more complete in vitro maturation. For many complex CNS diseases, extrinsic factors play a crucial role in the neuronal pathology. Injury to neurons can be conferred by immunological response from neighbor cells (microglia) and/or from a decline in the health of supporting glial cells (astrocytes and oligodendrocytes). The incorporation of niche cells into in vitro models could improve our understanding of the non-cell-autonomous effects on diseased neurons. New technology for the generation of homogeneous populations of glial cells from iPSCs is now available (Krencik and Zhang, 2011; Yuan et al., 2011). Another alternative for teasing out non-cell-autonomous effects on more mature and integrated neurons is in vivo grafting of REPRO cells in rodent brains. Studying the anatomy and function of transplanted neurons over time informs studies of neurodevelopmental aspects of the disease and of cell-autonomous versus nonautonomous elements. Developmental hallmarks such as neuronal pruning, dendritic branching, spine formation, and maturation could be dynamically observed as transplanted neural progenitors differentiate into neurons over time. State-ofthe-art intracranial live-imaging techniques coupled with electrophysiological studies could facilitate studies of functional integration properties from neurons in real time. Transplantation may also lead to the generation of specic subtypes of neurons that are difcult to produce through in vitro differentiation protocols. These transplantation studies ideally recapitulate the in vivo characteristics of human cells more authentically. Because cells need to be implanted at an early enough time to allow host factors to more accurately act on the REPRO cells, this process may lead to chimerism, so both technical and ethical challenges remain. Advanced whole-genome sequencing technology facilitates detection of new mutations and relevant variations in complex genetic diseases leading to common clinical outcomes. Combining in-depth DNA sequencing analysis with in vitro disease-relevant neuronal phenotypes, high-throughput proteomics/ metabolomics data, and longitudinal clinical studies of predened cohorts of patients will certainly provide a very comprehensive volume of information about disease features. What is clear is that while CPART is evolving, so is whole-genome technology; its depth and accuracy are changing and improving so rapidly that each new analysis is almost outdated by the time it is published. There are certainly opportunities available for CPART-based preclinical studies at all stages of drug development. Drug screening success for brain diseases would depend on the relevance, robustness, and scalability of the detected in vitro phenotype. New initiatives for drug screening using patient-reprogrammed cells are already in place, with the idea of taking advantage of assays developed in academia combined with Food and Drug Administration-approved drugs that have been repurposed. Conclusion As with many new marriages that bring together leading-edge technology with potential clinical applications, we are in the honeymoon stage. Evidence is pouring in every day through publications and opinion pieces that reveal new models, extraordinary results, and promising new applications. But of course every honeymoon comes to an end, and the staying power of this marriage will depend on its ability to stand the test of time and scrutiny. Many of the basic assumptions will be challenged, and many more technical hurdles will need to be overcome, especially the convergence on the best and most-reliable methods for programming and reprogramming cells. In addition, optimal ways to selectively specify cell fates that accurately reect the in vivo behavior of cells will need to be developed. Also challenging will be the development of a variety of tools that most sensitively and reliably measure functional phenotypes that are relevant to disease. While many aspects of this marriage between basic and clinical research will change and likely improve, we predict that this hybrid research approach will not go away but rather will become an essential tool for basic biology and translational medicine.
ACKNOWLEDGMENTS F.H.G. and M.C.M. are supported by the G. Harold and Leila Y. Mathers Charitable Foundation, Annette Merle-Smith, the Leona M. and Harry B. Helmsley Charitable Trust, the JPB Foundation, and the California Institute for Regenerative Medicine (grant TR2-01778). We thank M.L. Gage for editorial assistance and J. Simon for graphics. REFERENCES Chan, E.M., Ratanasirintrawoot, S., Park, I.H., Manos, P.D., Loh, Y.H., Huo, H., Miller, J.D., Hartung, O., Rho, J., Ince, T.A., et al. (2009). Nat. Biotechnol. 27, 10331037. Hayden, E.C. (2011). Nature 473, 272274. Kola, I., and Landis, J. (2004). Nat. Rev. Drug Discov. 3, 711715. Krencik, R., and Zhang, S.C. (2011). Nat. Protoc. 6, 17101717. Marchetto, M.C., Carromeu, C., Acab, A., Yu, D., Yeo, G.W., Mu, Y., Chen, G., Gage, F.H., and Muotri, A.R. (2010). Cell 143, 527539. Marchetto, M.C., Brennand, K.J., Boyer, L.F., and Gage, F.H. (2011). Hum. Mol. Genet. 20 (R2), R109R115. ser, P., and Loring, ller, F.J., Goldmann, J., Lo Mu J.F. (2010). Cell Stem Cell 6, 412414. Takahashi, K., and Yamanaka, S. (2006). Cell 126, 663676. Yuan, S.H., Martin, J., Elia, J., Flippin, J., Paramban, R.I., Hefferan, M.P., Vidal, J.G., Mu, Y., Killian, R.L., Israel, M.A., et al. (2011). PLoS ONE 6, e17540. Zhou, Q., and Tripathi, P. (2012). Cell Stem Cell 10, 347348.

Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc. 645

Cell Stem Cell

Forum
Stem Cell Therapies Could Change Medicine. If They Get the Chance
Irving Weissman1,*
1Institute of Stem Cell Biology and Regenerative Medicine, Stanford University, Stanford, CA 94305 USA *Correspondence: irv@stanford.edu DOI 10.1016/j.stem.2012.05.014

Stem cell therapies have the potential to revolutionize the way we practice medicine. However, in the current climate several barriers and false assumptions stand in the way of achieving that goal.
The rst two precepts of the modied Hippocratic Oath, which all M.D. graduates pledge are, in paraphrase: rst, do no harm; and second, the primary obligation of a physician is to the health of the patient (to which I add and future patients), and a physician will not let issues of race, creed, religion, politics, or personal ethics to stand between the patients health and his/her actions. The stem cell eld, probably more than any I know of in medical science, is plagued by failures to act responsibly on both precepts. While I am usually an optimist, I must admit that there is a possibility that we will continue to be in the Dark Ages of medicine for quite some time. I fear that therapies using puried tissue and organ-specic stem cellsthe only selfrenewing cells in a tissue or that can regenerate that tissue or organ for life will remain elusive. Before I go further, just think about that statement: regenerate that tissue or organ for life. No pharmaceutical, no biotech-developed protein, and no other transplanted cells can do that. If we can deliver puried stem cells safely and effectively as a one-time therapy, we can change medicine, especially for diseases that drugs and proteins cant touch. Moreover, if we manage the costs and charges carefully, this form of therapy could lower overall health care costs dramatically. This vision is based on solid scientic evidence that stem cells regularly maintain, and, if necessary, regenerate tissues in a homeostatically controlled process. So its worth the extra effort to nd a way to make it happen. Doing Harm One of the barriers to practicing stemcell-based regenerative medicine is the existence of fraudulent clinics and individuals who claim unproven therapies without underlying scientic backing. In many cases, they use cells that have never been tested experimentally for their stemness, have not been through IRBapproved protocols that demand experimental evidence to justify the human experiment, and lack both independent medical monitoring of patient safety and oversight by a state or country regulatory system such as the FDA. It is critical that, as the community that speaks for stem cell biology and stem cell medicine, we nd ways to warn patients and caregivers effectively about these concerns (Taylor et al., 2010). There is also a ne line between these clearly fraudulent practices and questionable ones that use the stem cell label, but are not in fact stem cell therapies. For example, cultures of adherent cells from bone marrow, cord blood, or adipose tissue are regularly claimed to be mesenchymal stem cells (MSCs), but in such cultures true stem cells that both selfrenew and differentiate to mesenchymal fates such as bone, cartilage, broblasts, and adipocytes are rare. Mesenchymal stromal cells, as a population, may contain cells that produce immunomodulatory and/or angiogenic factors, but are not sufciently puried or dened to be a characterized entity for research or clinical transplantation. Finding markers that help dene these populations was an important step (Dominici et al., 2006), but until there is a better understanding of how many of these cells can self-renew and give robust regeneration, I do not think they should be called stem cells. There are also many claims that mesenchymal and/or hematopoietic cells can transdifferentiate without gene modication to make brain, liver, heart, skeletal muscle, or other tissues. However, these claims lack rigorous scientic support (Wagers and Weissman, 2004). Highly visible athletes and politicians are among the many patients who have received such treatments. Recently, the Texas Medical Board approved a policy that allows licensed physicians to transplant investigational agents, including MSCs, with IRB approval but without a requirement for FDA approval of safety and efcacy. In my view, this lack of a requirement for FDA oversight and approval for both safety and efcacy is a giant step backward. Another example of questionable stem cell practices comes from some commercial private cord blood banks. Cord blood does contain both HSCs and mesenchymal progenitors. The number of HSCs in each cord is sufcient to give rapid generation of blood only in infants and very small children, and above the age of 7, several HLA-matched cords are needed. The development of public cord blood banks is an important, lifesaving advance for patients needing hematopoietic cell transplants but lacking matched donors. However, this activity is very different from the private cord blood banks that charge signicant amounts to initiate freezing of cord blood cells and then maintain them in case the child from whom the cord is obtained needs therapy. These companies often list a broad range of diseases that now or someday will be treated with stem cells without warning the patients or caregivers that the evidence that cord blood cells will be useful for treating such diseases is still very limited, and in any case the stored cord blood has the same genetic background as the child from whom the cord was obtained. The overall cause of legitimate stem cell therapy would be greatly

Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc. 663

Cell Stem Cell

Forum
advanced by greater control and oversight of these and other organizations making unsupported claims about the potential of stem-cell-based treatments. The Therapeutic Entity Is the Stem Cell Itself Very few adult stem cells have been prospectively isolated, and only prospectively isolated blood-forming stem cells (HSCs) and brain-forming stem cells (NSCs) have been transplanted in clinical trials (Baum et al., 1992; Uchida et al., 2000). Grafts of other tissues, such as skin and bone marrow, depend on the stem cells in that tissue, but prospectively isolated stem cells are usually not used. Instead of using cells as the therapy, as a general rule, large drug companies are approaching the use of disease specic iPSCs or adult stem cells as tools for chemical or protein screens to nd compounds that can be taken as conventional drugs to treat diseases. Some of these efforts are focused on differentiated cells derived from stem cells, but others aim to address diseases where altered or insufcient numbers of stem cells are central to the disease. The principal property of stem cells that makes them special is their ability to self-renew and reconstitute cell populations. Inducing self-renewal in vivo could be difcult to achieve because many factors affect stem cell regulation. It seems unlikely that single molecules will be able to activate all of the necessary pathway genes appropriately to expand a stem cell pool and allow robust and physiologically signicant regeneration. Thus, I think this approach is likely to fall short as a method to replace tissue stem cells in vivo, and efforts will need to focus more on transplanting the cells themselves. However, stem-cell-regulating agents derived from screening could still be used as adjuvants for transplanted stem cells. At a broader level, HSCs themselves form a foundation on which the rest of the regenerative medicine eld could be built. When engrafted, puried HSCs can replace the hematopoietic system. By doing so, they also render the host permanently tolerant to other organs, tissues, or tissue stem cells from the same donor without further immune suppression (Weissman and Shizuru, 2008). In the future, the isolation of HSCs and other tissue stem cells (e.g., NSCs) from the same donor could come from pluripotent stem cell lines, and not living or recently deceased donors. Pluripotent ESC or iPSC line production of HSCs is still not practical, and working out the pathways to achieve that objective remains a critical roadblock to expanding the eld of regenerative medicine. In Vivo Veritas The experiments that validated human, puried HSCs for hematopoietic transplants and human brain-stem-cellderived neurospheres for neural disease transplants used immune-decient mice that were crucial in testing the potential therapeutic effectiveness of these cells in vivo (Weissman, 2002). Although the derivation of patient- and diseasespecic iPSCs can allow experiments in a petri dish, the disease pathogenesis caused by inherited mutations would be more completely understood if the cells could mature in a more physiological setting. One way to study them would be to develop blastocyst chimeras that are implanted and allowed to develop. Mouse ESCs and iPSCs can already be studied using this type of approach. Currently, human ESCs/iPSCs do not form chimeras if placed in mouse blastocysts and implanted. However, human pluripotent stem cell lines are mainly at the epiblast stage, and not the preimplantation blastocyst, and even mouse epiblast cells cannot form long-term blastocyst chimeras. If the substantial practical and ethical issues could be overcome, blastocyst chimeras with human iPSCs might provide insights into the cellular and molecular mechanisms of human disease pathogenesis, and the gene expression programs that allow embryonic tissue stem cells to mature. An Unexpected but Potent Barrier: Business Development Growing up in America, it is obvious to all of us that the transition from discovery to therapy almost always involves for-prot entities. Ingenuity and innovation are hallmarks of our society, and so it is natural that the prospective identication and isolation of adult or tissue stem cells leads to business enterprises. I myself have cofounded several companies that have done discovery, preclinical proof of principle, and even phase I/II clinical trials in the stem cell eld. Each has succeeded in the discovery and preclinical phases, but found that the results of the clinical trials can take a back seat to business decisions. For example, SyStemix Inc. was a 1988 Palo Alto startup that identied a method to prospectively isolate and transplant clinically relevant numbers of human HSCs. The company entered a relationship with Sandoz, Inc. to explore autologous and allogeneic HSC therapies. Purication of mobilized peripheral blood HSCs resulted in depletion of various metastatic cancer cells by 115,000- to 245,000-fold (Prohaska and Weissman, 2009), and thus could be used to reconstitute the hematopoietic system after therapy with a reduced risk of reintroducing tumor cells. This nding led to clinical trials. Twenty-two patients with metastatic breast cancer underwent transplantation of previously mobilized HSCs after veryhigh-dose chemotherapy. Although the trials were small, two hypotheses were tested: (1) can one improve the outcome of patients with chemoresistant metastases? And (2) can one improve the outcomes of relapse patients with both metastases and chemoresponsive cancers? The therapy did not help the patients with chemoresistant breast cancers. However, at 3 years the chemoresponsive cohort who received cancer-depleted HSCs appeared to be doing better than patients with standard mobilized peripheral blood transplants. At that point, Sandoz merged with CIBA to form Novartis, and within a few years the stem cell program was cancelled. Last year Antonia ller and Judy Shizuru published the Mu follow-up of the patients 1315 years later ller et al., 2012). One-third of the (Mu patients who received puried HSCs were still alive, contrasting with the 7% overall survival of 78 contemporaneous Stanford patients with stage IV breast cancer who received standard, unpuried, mobilized peripheral blood transplant therapy. Of the ve long-term surviving patients who had received puried HSCs, four had no recurrence of their breast cancers. Attempts to reinitiate the program in another startup, helped by Novartis management, were halted when consultant oncologists advised investors that stem cell therapies in breast cancer had failed, citing a study indicating that stem cell rescue of high-dose chemotherapy patients with metastatic

664 Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc.

Cell Stem Cell

Forum
breast cancer was no better than chemotherapy alone, that is, only 6% diseasefree survival at 2 years (Stadtmauer et al., 2000). However, Stadtmauer et al. transplanted unfractionated mobilized blood, not puried HSCs, and no amount of evidence about the difference could counter the words stem cell in the title of the NEJM article. This particular problem could have been avoided by more rigorous editorial standards regarding the use of the term stem cell, and I would argue that improved accuracy in this respect would benet many areas of the eld. How can we resolve this conict of goals, that of a company to make a prot, and that of the biomedical researcher to advance medical science for the benet of patients? The largest and best funding experiment I have seen so far comes from the California Institute of Regenerative Medicine. CIRMs charter allows it to fund promising stem-cell-based discoveries to and through phase I trials, taking out the risk that leaves our eld bereft of suitable funds and in the valley of death. However, to overcome the types of problems that the SyStemix trial encountered, this funding would need to be taken beyond initial trials to a point at which the evidence for clinical efcacy was irrefutable. In Closing.... So, whom have I failed to annoy here? In one way or another, I have called out almost all of the different stakeholder groups involved in developing stem cell therapies. I wish I had a better story to tell, but I am convinced that we need to identify and reveal those who directly or indirectly do harm with phony medicines, and those who generate barriers to nding and transplanting adult tissue/ organ stem cells for nancial, religious, political, or other reasons. Unless we do, it will be difcult to usher in the era of stem cell regenerative medicine. Remember, right now our patients, friends, and families are contracting diseases that have a very short window of opportunity in which regenerative therapies can save them, and each delay removes a cohort of them from possible cures. We should not fail them.
REFERENCES Baum, C.M., Weissman, I.L., Tsukamoto, A.S., Buckle, A.M., and Peault, B. (1992). Proc. Natl. Acad. Sci. USA 89, 28042808. Dominici, M., Le Blanc, K., Mueller, I., SlaperCortenbach, I., Marini, F.C., Krause, D.S., Deans, R.J., Keating, A., Prockop, D.J., and Horwitz, E.M. (2006). Cytotherapy 8, 315317. ller, A.M., Kohrt, H.E., Cha, S., Laport, G., Klein, Mu J., Guardino, A.E., Johnston, L.J., StockerlGoldstein, K.E., Hanania, E., Juttner, C., et al. (2012). Biol. Blood Marrow Transplant. 18, 125133. Prohaska, S.S., and Weissman, I.L. (2009). Biology of Hematopoietic Stem and Progenitor Cells: Thomas Hematopoietic Cell Transplantation, F. Applebaum, S. Forman, R.S. Negrin, and K. Blume, eds. (Oxford: Wiley-Blackwell), pp. 3663. Stadtmauer, E.A., ONeill, A., Goldstein, L.J., Crilley, P.A., Mangan, K.F., Ingle, J.N., Brodsky, I., Martino, S., Lazarus, H.M., Erban, J.K., et al; Philadelphia Bone Marrow Transplant Group. (2000). N. Engl. J. Med. 342, 10691076. Taylor, P.L., Barker, R.A., Blume, K.G., Cattaneo, E., Colman, A., Deng, H., Edgar, H., Fox, I.J., Gerstle, C., Goldstein, L.S., et al. (2010). Cell Stem Cell 7, 4349. Uchida, N., Buck, D.W., He, D., Reitsma, M.J., Masek, M., Phan, T.V., Tsukamoto, A.S., Gage, F.H., and Weissman, I.L. (2000). Proc. Natl. Acad. Sci. USA 97, 1472014725. Wagers, A.J., and Weissman, I.L. (2004). Cell 116, 639648. Weissman, I.L. (2002). N. Engl. J. Med. 346, 1576 1579. Weissman, I.L., and Shizuru, J.A. (2008). Blood 112, 35433553.

Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc. 665

Cell Stem Cell

Forum
Restoring Stem Cell Function in Aged Tissues by Direct Reprogramming?
ler1,3,* Natalia Tapia,1 Dong Wook Han,2 and Hans R. Scho
1Department of Cell and Developmental Biology, Max Planck Institute for Molecular Biomedicine, Ro ntgenstrae 20, nster, Germany 48149 Mu 2Department of Stem Cell Biology, School of Medicine, SMART Institute of Advanced Biomedical Science, Konkuk University, 1 Hwayang-dong, Gwangjin-gu, Seoul 143-701, Republic of Korea 3University of Mu nster, Medical Faculty, Domagkstrasse 3, 48149 Mu nster, Germany *Correspondence: ofce@mpi-muenster.mpg.de DOI 10.1016/j.stem.2012.04.010

Adult stem cells are responsible for the cellular turnover of many organs, and an impairment in their function leads to aging and disease. In efforts to reverse the process of tissue stem cell aging, we speculate on the promise and challenges of in vivo direct reprogramming strategies.

Aging correlates with a decline in the function of the adult stem cells (i.e., somatic stem cells) present in many organs of an adult mammal. A decrease in the stem cell pool and/or a restriction in the somatic stem cell differentiation potential affects tissue homeostasis and regeneration. Here, we discuss the possibility of replacing or rejuvenating dysfunctional adult stem cells from an aged tissue or organ using a direct reprogramming approach in situ. Limitations of Generating Differentiated Somatic Cells for Anti-Aging Purposes The ability to induce pluripotency in somatic cells has heralded a new era in the eld of regenerative medicine, because induced pluripotent stem cells (iPSCs) specic to individual patients could provide an unlimited source of specialized cell types for replacing diseased or aged tissues. However, several technical problems must be resolved before safely translating the iPSC technology to the clinic. For therapeutic purposes, the iPSCs must be differentiated into the required cell types. Currently available differentiation protocols aim to recapitulate in vitro the embryonic events that occur in the early embryo in vivo. However, the underlying developmental mechanisms are not well understood, making their faithful modeling in vitro a difcult undertaking and preventing the derivation of sufcient quantities of transplantable cells. In addition to transplantation and proper engraftment, the risk for teratoma formation

remains a major issue with somatic cells differentiated from iPSCs (Cohen and Melton, 2011). The reprogramming of somatic cells into iPSCs entails the use of a set of specic transcription factors that can establish a pluripotent program in a differentiated cell. This method of reprogramming suggests that other cell fates could be induced if the right transcription factor cocktail is used. Recent studies have indeed shown that specic transcription factor combinations can be used to convert broblasts directly into neurons, cardiomyocytes, and hepatocytes, bypassing a pluripotent intermediate and essentially removing the risk for teratoma formation. The cell types generated by such direct lineage reprogramming typically exhibit an immature phenotype resembling that of fetal or neonatal cells. Nevertheless, these cells might acquire a fully functional and mature prole after engraftment into the host and subsequent exposure to the appropriate environmental cues (Cohen and Melton, 2011). However, terminally differentiated cells, such as postmitotic neurons, do not proliferate and cannot be expanded prior to transplantation. Therefore, the generation of a sufcient number of cells for ensuring a successful transplantation remains a challenge. A strategy that might circumvent this obstacle is the direct reprogramming of somatic cells into self-renewing somatic stem cells. The low reprogramming efciency associated with this approach would not limit the clinical applicability of this strategy because the generated

cells would be able to proliferate. In addition, the somatic stem cells could be transplanted into the host niche, where endogenous stimuli could promote acquisition of a fully mature phenotype. Moreover, transplanted somatic stem cells could self-renew and/or differentiate into cells that would respond to inammatory signals and migrate toward damaged tissues, thus facilitating proper engraftment for regenerative purposes. Recently, we and others reported the direct conversion of mouse broblasts into self-renewing induced neural stem cells (iNSCs) that could differentiate into neurons, astrocytes, and oligodendrocytes, as reviewed in Zhou and Tripathi (2012). A successful direct reprogramming strategy requires the use of culture conditions that are optimal for the desired cell type. However, such conditions have not yet been established for all of the specialized cell types of the adult organism. To overcome this problem, one could perform direct lineage conversion in vivo, where the right environment already exists to promote the survival of the induced cells. As a proof of principle, exocrine pancreatic cells have been directly reprogrammed in vivo into b cells (Zhou et al., 2008). Reprogramming in situ circumvents the need to determine culture conditions for the maintenance of the induced cells in vitro, provides the right niche for achieving a fully mature and functional phenotype, and eliminates the transplantation step. However, as in the in vitro scenario, in vivo conversion into somatic cells is limited by the

Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc. 653

Cell Stem Cell

Forum

Figure 1. Potential Strategies for Restoring Stem Cell Function in Aged Organs and Tissues
(A) Stem cells contain a genetic modication that is transmitted to their progenitor and daughter cells. In this scenario, somatic cells located close to the niche and not derived from the mutated adult stem cell would be reprogrammed by factors (e.g., expressed via viruses) into de novo stem cells. (B) Stem cells contain an epigenetic modication that is transmitted to the progenitors. In this case, progenitor or differentiated cells would be dedifferentiated into stem cells, and the detrimental epigenetic marks would be erased. (C) The stem cell transcriptional network is missing specic transcription factors that would be delivered specically to the stem cells. (D) Malfunction of the stem cells is due to an aged niche, which would be the target for the in vivo reprogramming.

inefciency of the reprogramming process and by the nonproliferative nature of many terminally differentiated cells. Thus, the number of somatic cells directly converted in vivo might not sufce to rescue the tissue deciency. This problem could again, in principle, be overcome by the direct conversion in vivo of differentiated cells into proliferative tissuespecic somatic stem cells. Therefore, direct reprogramming in vivo into somatic stem cells might potentially be able to circumvent many of the obstacles that stand in the way of being able to use direct lineage reprogramming in a therapeutic setting. Reprogramming In Vivo into Somatic Stem Cells Could Compensate for Tissue Aging In theory, if tissue stem cells could be generated from somatic cells in vivo using

the right combination of factors, then the in situ reprogrammed adult stem cells could be used to ameliorate the effects of an age-related decline in the stem cell pool and/or a restriction in the stem cell differentiation potential. Depending on the mechanisms underlying the agerelated stem cell dysfunction, we have divided the different strategies that could be explored into genetic abnormalities, epigenetic modications, dysregulated gene expression, and niche defects. In the case of intrinsic genetic modications in the adult stem cells (Figure 1A), such as chromosomal translocations or mutations, de novo somatic stem cells would need to be generated. Because daughter cells could also have inherited the genetic alterations that are responsible for the dysfunction, the progeny of stem cells that need to be replaced could not be the cells subjected to the direct

conversion. Therefore, somatic cells located close to the stem cell niche that are not derived from the damaged tissue stem cells would be the target cell of choice. Some cell types can present advantages for the reprogramming process that should be taken into consideration. For instance, epithelial cells (e.g., keratinocytes) can be reprogrammed into iPSCs more rapidly than cells from a mesenchymal origin because they do not need to undergo a mesenchymal-toepithelial transition step (Stadtfeld and Hochedlinger, 2010). However, the potential impact of these parameters in the direct conversion into somatic stem cells needs to be investigated. In another scenario (Figure 1B), adult stem cell dysfunction could be caused by epigenetic modications (Pollina and Brunet, 2011). In this case, progenitor or daughter cells differentiated from the

654 Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc.

Cell Stem Cell

Forum
aged adult stem cell would be suitable for conversion, because the reprogramming process would most likely reset the epigenome, correcting the detrimental modications. Progenitor cells lack the unlimited self-renewal activity of somatic stem cells and are committed to differentiation. However, somatic stem cells and the progenitors derived from them have very similar transcriptional networks. For instance, neural stem cells and neural progenitor stem cells both express Sox2 (Zhou and Tripathi, 2012). During the reprogramming of NSCs into iPSCs, Sox2 does not need to be exogenously introduced, because it is already expressed by the NSCs, reducing the number of reprogramming factors needed (Stadtfeld and Hochedlinger, 2010). Thus, we would suggest that the conversion of progenitor or early-committed cells into somatic stem cells might require fewer transcription factors than conversion of more differentiated cells because it could take advantage of genes that are already being endogenously expressed. Finally, whereas the epigenetic memory in somatic stem cells might be substantial when reprogrammed from cells that originated from a different germ layer (Han et al., 2012), using progenitor or earlycommitted cells from the same lineage as the initial target cell would minimize its potential impact. Another scenario would involve alterations in the expression levels of specic genes leading to an adult stem cell malfunction (Figure 1C). For example, deletion of the transcription factors FoxO1, FoxO3, and FoxO4 reduces the number of NSCs in the mouse adult brain, impairing neurogenesis. Similar effects are seen with deletion of the transcription factor Tlx and the enzyme Telomerase. In these situations, it is possible that forced expression of single genes could improve the performance of the aged adult stem cells. Consistent with this idea, the overexpression of Telomerase could restore NSC self-renewal and neurogenesis in aged telomerase-decient mice (Pollina and Brunet, 2011). The overexpression of Tlx in NSCs also delayed the ageassociated NSC decline, but, in addition, it led to the formation of glioma-like lesions (Pollina and Brunet, 2011). As in this case, many genes involved in stem cell maintenance have also been related to tumor formation. Considering that, to date, in vivo transplantation of somatic stem cells generated by direct reprogramming has not been associated with any tumor formation, it appears that induction of a de novo cell phenotype may present a lower risk for tumorigenesis than the overexpression of single genes in aged somatic stem cells (e.g., Tlx). In the generation of de novo somatic stem cells, the old (i.e., aged) and young stem cells will cohabitate in the same niche. If the aged stem cells exhibit a decline in self-renewal activity, we would expect a dilution of the old stem cell pool due to the higher proliferative ability of the de novo-induced tissue stem cells. In addition, the new stem cells would be able to compensate for the restricted differentiation potential of the aged stem cells. However, the aged stem cells exhibiting genetic alterations cannot be eliminated from the stem cell niche and would continue to accumulate mutations, presenting a tumorigenic risk. In the scenarios described above, the decline in tissue homeostasis was caused by intrinsic modications in the stem cells. But adult stem cells could also be affected by extrinsic factors (Figure 1D). The environment or stem cell niche plays an essential role in regulating the selfrenewal and differentiation potential of somatic stem cells. For instance, a reduction in the number of spermatogonial stem cells (SSCs) with age is associated with reduced fertility of male mice, suggesting a decline in SSC self-renewal ability. However, SSCs isolated from young mouse testes could be serially transplanted multiple times into young recipients and could self-renew for a longer time than the life span of a normal mouse (Ryu et al., 2006). In this case, the proliferation decay is not intrinsic to the SSCs but rather is induced by the aged environment. This phenomenon has been also observed in other adult stem cells (e.g., hematopoietic stem cells). Therefore, the tissue stem cell niche can be considered to be one of the main players in preventing, reverting, and/or delaying stem cell aging. At this point, we would like to speculate on the possibility of reprogramming or rejuvenating the aged niche through the overexpression of a minimum set of genes. Unfortunately, very little is known about the transcriptional networks governing the different cell types that form a stem cell niche. Glial cell-line-derived neurotrophic factor (GDNF) is a ligand secreted by Sertoli cells that ensures the self-renewal of SSCs in mouse testes. Interestingly, aged testes show reduced GDNF levels (Ryu et al., 2006), establishing a connection between an alteration in the SSC niche and a decline in SSC proliferation ability. Although mice that express GDNF ubiquitously are infertile due to an inability of the SSCs to differentiate (Meng et al., 2000), it would be interesting to assess the effect of GDNF overexpression solely by Sertoli cells on aged SSC. Nevertheless, it seems unlikely that only one gene could restore the function of an aged stem cell niche. In addition, the aging of the stem cell niche is also affected by systemic factors, as is evident from parabiosis experiments in which circulating factors from the blood of young mice can modify the stem cell niche of old mice and can revert some aging effects (studies that have been recently reviewed in Wagers, 2012). In those cases in which the main cause for the dysfunctional niche relies on the alteration of systemic factors, in situ transcription-mediated reprogramming of the aged adult stem cell niche would not be the strategy of choice. Overall, further knowledge in niche regulation is needed to ascertain better strategies for rejuvenating the stem cell environment and, indirectly, restoring the somatic stem cell functiona promising approach, yet one fraught with many challenges. Methods for Specically Delivering the Reprogramming Factors For any transcription-factor-mediated reprogramming approach, a robust technique for efciently delivering reprogramming factors to a specic cell type of choice would be required but is currently still lacking. The recent demonstration that lentiviral vectors are able to recognize and infect, in vivo, cells presenting with specic cell-surface antigens is promising (Anliker et al., 2010). However, lentiviral vectors integrate into the host genome, precluding their use in medical applications. Unfortunately, the efciency of the conversion process is signicantly compromised when using integrationfree methods, as demonstrated in iPSC reprogramming experiments (Stadtfeld and Hochedlinger, 2010). Therefore, further efforts are needed to overcome

Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc. 655

Cell Stem Cell

Forum
the technical challenges associated with the current reprogramming techniques. Progress in the area of nanotechnology may help develop systems that allow efcient directed delivery of reprogramming factors to regions within tissue and organs. Future Perspectives From graying hair to a decline in neurogenesis, aging affects everyone. Though an in vivo transcription-factor-mediated approach cannot be considered as an elixir for immortality, it could serve as the basis for further discussion about tissue stem cell rejuvenation. Future efforts should be directed toward the characterization of the transcriptional networks of tissue stem cells and their niches and the development of delivery methods for specic cell types. In addition, identication of the molecular modications that stimulate adult stem cell longevity in response to environmental modications (e.g., dietary restrictions) would provide crucial information for improving antiaging reprogramming strategies. The challenges that lie ahead should be an impetus for uncovering the missing links in direct reprogramming and for restoring stem cell function in aged cells.
Cohen, D.E., and Melton, D. (2011). Nat. Rev. Genet. 12, 243252. Han, D.W., Tapia, N., Hermann, A., Hemmer, K., ing, S., Arau zo-Bravo, M.J., Zaehres, H., Wu, Ho G., Frank, S., Moritz, S., et al. (2012). Cell Stem Cell 10, 465472. nen, M.E., Parvinen, Meng, X., Lindahl, M., Hyvo M., de Rooij, D.G., Hess, M.W., Raatikainen-Ahokas, A., Sainio, K., Rauvala, H., Lakso, M., et al. (2000). Science 287, 14891493. Pollina, E.A., and Brunet, A. (2011). Oncogene 30, 31053126. Ryu, B.Y., Orwig, K.E., Oatley, J.M., Avarbock, M.R., and Brinster, R.L. (2006). Stem Cells 24, 15051511. Stadtfeld, M., and Hochedlinger, K. (2010). Genes Dev. 24, 22392263. Wagers, A.J. (2012). Cell Stem Cell 10, 362369. Zhou, Q., and Tripathi, P. (2012). Cell Stem Cell 10, 347348. Zhou, Q., Brown, J., Kanarek, A., Rajagopal, J., and Melton, D.A. (2008). Nature 455, 627632.

ACKNOWLEDGMENTS Thanks to A. Mota for preparing the gure, B. Greber for critically reading the manuscript, and A. Malapetsas for editing the manuscript. We apologize to those colleagues whose work could not be cited due to the restricted number of references.

REFERENCES Anliker, B., Abel, T., Kneissl, S., Hlavaty, J., Caputi, nch, R.C., PetzA., Brynza, J., Schneider, I.C., Mu nek, H., Kontermann, R.E., et al. (2010). Nat. Methods 7, 929935.

656 Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc.

Cell Stem Cell

Forum
Neural Stem Cells and Neurogenesis in the Adult
n1,* ritz1 and Jonas Frise Christian Go
1Department of Cell and Molecular Biology, Karolinska Institute, SE-171 77 Stockholm, Sweden *Correspondence: jonas.frisen@ki.se DOI 10.1016/j.stem.2012.04.005

Research in the eld of adult neurogenesis has seen substantial progress over recent years. Here we discuss some of the major focus areas for future investigation: neural stem cell heterogeneity, the role of latent stem cells, and the extent of neurogenesis in the adult human brain.
New neurons are supplied throughout life in most mammals in two regions of the brain: the dentate gyrus of the hippocampus, by locally residing neural stem cells, and the olfactory bulb, by neural stem cells present in the lateral ventricle wall. The new neurons confer plasticity to the circuitry and increasing evidence has established a role for adult neurogenesis in specic brain functions. The discovery of neural stem cells and neurogenesis in the adult mammalian central nervous system changed our view of the plasticity and function of the brain and spurred enthusiasm for trying to harness their regenerative potential in new therapies for conditions such as depression, stroke, spinal cord injury, and Parkinsons disease. Our understanding of neural stem cells has increased dramatically over the past few years, but there are still many major gaps, some of which are the focus of our discussion here. Neural Stem Cell Heterogeneity Stem cells are notoriously difcult to identify, and we have witnessed a gradual homing in on bona de stem cells in many tissues. This progress has been accompanied by a realization that stem cell populations are often heterogeneous within a tissue and that several distinct stem cells for the same lineage can coexist. For example, in the skin, intestine, and hematopoietic system, there are stem cells that may be in different states of activity or dormancy or may have different roles in homeostasis and regeneration. It will be necessary to develop a detailed characterization of the full repertoire of neurogenic cells in the adult brain to understand the process of adult neurogenesis and to potentially modulate it in therapeutic strategies. In most studies, the properties of neural stem cells are interrogated at a population level rather than at the clonal level, resulting in the analysis of potentially mixed populations, making it difcult to detect heterogeneity (Figure 1). Nevertheless, it is already clear that heterogeneity does exist. For example, in the hippocampus, there are at least two morphologically distinguishable types of astrocytes that give rise to neurons, radial and horizontal, but their lineage relationship is still subject of debate. There is also conicting evidence about whether hippocampal astrocytes are self-renewing or exhausted during the process of generating new neurons (Bonaguidi et al., 2011; Encinas et al., 2011). These apparently contradictory results might reect the coexistence of several stem cell populations with different properties. In the ventricle wall there are indications that astrocytes may have different activity states marked by EGFR expression. In several organs, there are indications that progenitor cells may be able to revert to a stem cell state, and it has been suggested that EGF may induce dedifferentiation of transit amplifying cells in the neural lineage to stem cells. Moreover, neural stem cells located at different positions along the ventricle wall and rostral migratory stream give rise to different repertoires of olfactory bulb neurons (Lledo et al., 2008). Detecting heterogeneity within a population is virtually impossible without using clonal labeling or fate mapping strategies to mark distinct subpopulations. Thus, although there are clear indications of neural stem cell heterogeneity and we have already learned about some aspects of it using such approaches, it seems likely that there is much more diversity still to discover. How do we identify and characterize subpopulations that we do not even know exist? Neuroscience may perhaps be able to learn from genetic fate mapping strategies used in other organs, which have been extremely helpful in identifying diversity. As our marker and genetic reagent tool boxes grow, we are likely to continue nding new features of neural stem cells. Latent Neural Stem Cell Activity It became clear during early studies of neural stem cells not only that cells with in vitro neural stem cell properties are found in neurogenic niches, but also that they are in fact present in most major subdivisions of the adult central nervous system. Moreover, injury endows additional cells with in vitro neural stem cell -Heider et al., 2010; properties (Barnabe Buffo et al., 2008). Thus, in addition to the neural stem cells that are responsible for physiological adult neurogenesis, there are other cells that may have similar potential but are largely dormant (Figure 1). Which are these latent neural stem cells, what is their in vivo function, and can they be induced to replace lost cells? Cortical astrocytes gain in vitro neural stem cell properties after injury (Buffo et al., 2008). However, there is no evidence that they produce cells other than astrocytes outside the neurogenic niches in vivo, and their main function after injury may be to participate in the formation of a glial scar. We do not know today what distinguishes a neurogenic astrocytic stem cell from an astrocyte with latent stem cell or neurogenic potential. It is probably in part regulated cell intrinsically, as neurogenic and nonneurogenic astrocytes are intermingled in the neurogenic niches. Spinal cord ependymal cells have in vitro neural stem cell properties but are largely dormant and do not generate other cell types in vivo in the absence of injury (Figure 1). In response to injury, their proliferation is dramatically increased, a much larger number of them display in vitro neural

Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc. 657

Cell Stem Cell

Forum

Figure 1. Actual and Possible Neural Stem Cells


Active neural stem cell niches in the mouse dentate gyrus and lateral ventricle (LV) wall, showing astrocyte-like stem cells labeled for glial brillary acidic protein (GFAP) and the SRY-BOX 2 (Sox2) transcription factor, and neuroblasts expressing doublecortin (DCX) (Bonaguidi et al., 2011; Lledo et al., 2008). Latent neural stem cell niches in the adult spinal cord and cortex showing ependymal cells lining the central canal (CC) labeled for Sox2 but negative for GFAP and parenchymal -Heider et al., 2010; Buffo et al., 2008). Scale bar represents 20 mm. astrocytes expressing GFAP and Sox2, respectively (Barnabe

stem cell properties, and they constitute inuencing gliogenesis. Glial cells support pect of developing new therapies for the main source of new glial cells in the neuronal function in many ways, and in neurological and psychiatric diseases. -Heider some neural stem cell transplantation Neuronal replacement by cell transplantainjured spinal cord (Barnabe et al., 2010). Forebrain ependymal cells experiments the benecial effect on tion has been explored intensively over lining the lateral walls are completely functional recovery is mediated by glial recent years, and cell replacement from quiescent but can give rise to astrocytes differentiation, rather than by neurons. endogenous sources is an attractive alterand neuroblasts in response to stroke. Demyelination in multiple sclerosis and native. However, the extent to which However, unlike spinal cord ependymal many other types of injuries results in ndings made in the mouse and other cells, they do not self-renew, and instead neuronal malfunction and, over time, mammals will translate to human settings are consumed in the process of gener- neuronal degeneration if the axon is not still remains largely unclear. It will there n et al., 2009). Thus, remyelinated. Endogenous neural stem fore be important to establish the extent ating progeny (Carle among these heterogeneous latent cells, cells give rise mainly to scar-forming of neurogenesis in the adult human so far there is only evidence that the cells astrocytes and few remyelinating oligo- brain in normal and pathological situain the neurogenic lateral ventricle wall dendrocytes after spinal cord injury tions to understand whether this process -Heider et al., 2010). It is ap- could contribute to normal brain function niche can generate neuronal lineage cells (Barnabe in vivo. pealing to look at inuencing gliogenesis, and whether alterations in it may be Heterotopic transplantation experi- for example, to promote differentiation to related to pathology. A seminal study by ments have indicated that the neurogenic remyelinating oligodendrocytes at the Eriksson, Gage, and colleagues demoncapacity of adult neural stem cells is deter- expense of scar forming astrocytes. strated BrdU labeling in hippocampal mined, at least in part, by their extracelneurons in a group of cancer patients lular molecular environment. Unleashing Understanding and Mending that received the labeled nucleotide for latent neurogenic potential outside of the Human Brain disease staging purposes, establishing normal neurogenic regions is a tantalizing From the outset, enthusiasm in the neural the presence of adult neurogenesis in proposition, but a major limiting factor stem cell eld was buoyed by the pros- humans (Eriksson et al., 1998). However, may be the capacity of new neurons this analysis did not allow quantitato integrate in regions outside of tive assessment of the extent of areas that are normally neurogenic. neurogenesis or insights as to In lower vertebrates, such as the whether the process may be newt, neurons can be regenerated affected by pathology. after lesions by latent stem/progenIndications of the extent of neuroitor cells even in normally nonneurogenesis can be inferred from the genic regions. Understanding the number of neuroblasts in a given regenerative mechanisms in lower region. The number of cells with vertebrates could potentially shed neuroblast markers drops precipilight on, and suggest ways to overtously in the rst few months after come, the restrictions that exist in birth in humans, with similar kinetics the mammalian nervous system. and to a similar extent in the hippoAlthough most attention is campus and subventricular zone Figure 2. Neuroblast Kinetics in the Human Brain focused on potential therapeutic (Figure 2), leaving only miniscule Quantication of neuroblasts in the human subventricular zone strategies that involve the generaapparent numbers of neuroblasts (SVZ) and dentate gyrus at different ages based on doublecortion of new neurons, it is also attracafter the rst year (Knoth et al., tin (DCX) expression. Note the logarithmic scale on the y axis. Data are from Knoth et al. (2010) and Sanai et al. (2011). tive to consider the possibility of 2010; Sanai et al., 2011). These low
658 Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc.

Cell Stem Cell

Forum
numbers could suggest that adult neurogenesis is unlikely to be functionally relevant in the adult human brain. However, one should interpret these data with some caution, as they could represent either an overestimate of the number of functionally active neural stem cells or an underestimate if there are other cells that do not express the tested markers. Moreover, computational modeling has suggested that even small numbers of new neurons can have a signicant inuence on overall circuitry. In addition, even if neurogenesis in adult humans is normally very limited, it could still be activated in pathological situations. Retrospective 14C birth dating allows a direct measurement of the age of a cell population, and this has established that there is minimal, if any, adult olfactory bulb neurogenesis in humans (Bergmann et al., 2012). One may argue that humans depend to a small degree on olfaction, and that hippocampal neurogenesis may be more relevant, but the similar dynamics of neuroblast generation in both neurogenic niches in humans raises the question of whether there may be any functionally relevant degree of neurogenesis in the hippocampus. Adult neurogenesis has advanced from a controversial suggestion to having established roles in brain function in rodents. As we are delineating the normal extent and understanding the mechanisms regulating this process, one of the main goals for the future remains to assess whether it is possible to inuence neural stem cells to replace cells in disease.
ACKNOWLEDGMENTS We apologize to the authors whose work the space limitation precluded us from citing. We thank Gerd Kempermann and Arturo Alvarez-Buylla for providing the data shown in Figure 2. Work in the authors laboratory was supported by the Swedish rnfonden, Research Council, Tobias Stiftelsen, Hja Knut och Alice Wallenbergs Stiftelse, AFA rsa kringar, and the ERC. Fo Bergmann, O., Liebl, J., Bernard, S., Alkass, K., Yeung, M.S.Y., Steier, P., Kutschera, W., Johnson, n, M., Druid, H., et al. (2012). Neuron 74, L., Lande 634639. Bonaguidi, M.A., Wheeler, M.A., Shapiro, J.S., Stadel, R.P., Sun, G.J., Ming, G.L., and Song, H. (2011). Cell 145, 11421155. Buffo, A., Rite, I., Tripathi, P., Lepier, A., Colak, D., tz, M. (2008). Proc. Horn, A.P., Mori, T., and Go Natl. Acad. Sci. USA 105, 35813586. n, M., Meletis, K., Go ritz, C., Darsalia, V., Carle Evergren, E., Tanigaki, K., Amendola, M., -Heider, F., Yeung, M.S., Naldini, L., Barnabe et al. (2009). Nat. Neurosci. 12, 259267. Encinas, J.M., Michurina, T.V., Peunova, N., Park, J.H., Tordo, J., Peterson, D.A., Fishell, G., Koulakov, A., and Enikolopov, G. (2011). Cell Stem Cell 8, 566579. rk-Eriksson, T., Eriksson, P.S., Perlieva, E., Bjo Alborn, A.M., Nordborg, C., Peterson, D.A., and Gage, F.H. (1998). Nat. Med. 4, 13131317. Knoth, R., Singec, I., Ditter, M., Pantazis, G., Capetian, P., Meyer, R.P., Horvat, V., Volk, B., and Kempermann, G. (2010). PLoS ONE 5, e8809. Lledo, P.M., Merkle, F.T., and Alvarez-Buylla, A. (2008). Trends Neurosci. 31, 392400. Sanai, N., Nguyen, T., Ihrie, R.A., Mirzadeh, Z., Tsai, H.H., Wong, M., Gupta, N., Berger, M.S., Huang, E., Garcia-Verdugo, J.M., et al. (2011). Nature 478, 382386.

REFERENCES -Heider, F., Go ritz, C., Sabelstro m, H., Barnabe Takebayashi, H., Pfrieger, F.W., Meletis, K., and n, J. (2010). Cell Stem Cell 7, 470482. Frise

Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc. 659

Cell Stem Cell

Perspective
Tissue Stem Cells: New Tools and Functional Diversity
Markus Grompe1,*
1Pape Family Pediatric Research Institute, Oregon Health & Science University, 3181 S.W. Sam Jackson Park Road, Mail Code L321, Portland, OR 97239-3098, USA *Correspondence: grompem@ohsu.edu DOI 10.1016/j.stem.2012.04.006

The detailed understanding of adult tissue stem cells has signicance for both regenerative medicine and oncology. This perspective will discuss how major advances in our ability to identify and monitor these cells, which include genetic lineage tracing, FACS purication, and robust in vitro clonogenic assays, have changed our view of their roles in many organs. Label retention and quiescence are no longer considered obligatory stem cell features. Furthermore, some tissues have more than one type of stem cell, each used in only particular situations of regenerative stress. Thus, there is no one size ts all adult tissue stem cell paradigm.
Spurred by the excitement surrounding the potential of pluripotent stem cells in regenerative medicine (Takahashi et al., 2007), interest in tissue-specic adult stem cells also increased signicantly since the inception of this journal. A detailed understanding of how organs maintain and repair themselves in the postnatal organism is obviously required in order to properly utilize the derivatives of pluripotent cells in therapy. In addition, it is conceivable that the ability to control and manipulate tissue stem cells in situ may obviate the need for exogenous cell transplantation altogether. There is much evidence that tissueresident stem cells may be the origin of malignant tumors (Barker et al., 2009). Therefore, knowing the mechanisms that turn stem cells into cancer stem cells or understanding how cancers derived from differentiated cells acquire stem cell characteristics would probably yield targets for therapy. Although it would be difcult to comprehensively cover all of the advances recently made in our understanding of the many different kinds of tissue stem cells, there are some common themes that apply broadly and these will be discussed herein. Stem Cells versus Progenitors: Not Simply a Semantic Distinction Perhaps because of its popularity with the lay public and funding agencies, the term stem cell is used liberally and often indiscriminately. However, despite much progress, the existence of true stem cells has to date been shown for only few adult tissues when the strict criterion of extensive clonal self-renewal (van der Kooy and Weiss, 2000; Weissman et al., 2001) is used. This widely accepted stem cell denition was derived from the hematopoietic eld where in vivo transplantation assays are available to assess replicative potential. Similar in vivo reconstitution assays do not yet exist for many adult tissues and thus it is not possible to clearly prove self-renewal in those organs by this classic approach. In some rapidly regenerating tissuesmost notably the intestineclonal marking techniques can prove life-long self-renewal in the absence of transplantation and rmly establish the identity of stem cells (Figure 1). However, in many tissues the existence of true stem cells remains yet to be fully proven and the term progenitor is more appropriate to describe the cells that have been found to date. Like true stem cells, progenitors often are multipotential and may give rise to several different mature lineages. However, they can be distinguished from true stem cells by the fact that clonal labeling is transient and does not last for the life of the animal (Figure 1C). In contrast, true stem cells will produce labeled offspring permanently (Figure 1A). Extensive self-renewal in culture may represent an in vitro artifact and does not conclusively prove that the cell in question truly is a stem cell in vivo. The distinction is not just semantic because the limited self-renewal of progenitors may signify restricted regenerative capacity of a particular tissue, particularly in the setting of chronic injury or aging. Although the existence of self-renewing stem cells in highly regenerative organs has been established for some time now (Mackenzie and Bickenbach, 1985; Potten and Hendry, 1975), the precise identity, molecular regulation, and anatomic location of these cells in organs such as colon (Barker et al., 2007), small intestine (Barker et al., 2007), stomach (Barker et al., 2010), breast (Shackleton et al., 2006; Stingl et al., 2006), and skin (Snippert et al., 2010a) were described only fairly recently. In the intestine, these breakthroughs were made by using genetic lineage tracing, a strategy originally devised by developmental biologists (Kretzschmar and Watt, 2012). Although it has been generally challenging to nd highly specic markers for adult tissue stem cells, methods such as genome-wide assessment of gene expression in highly puried cell populations have suggested candidate genes (Barker et al., 2007). These loci can be used to restrict marker gene expression to the cells of interest and their progeny, thereby allowing analysis of self-renewal and potency in vivo. The results of these kinds of studies have been surprising in some cases, most notably in the small intestine, where the cell located in the crypt at the +4 position was long thought to be the true and only stem cell (Potten et al., 2009; Potten and Hendry, 1975). However, lineage tracing of cells expressing the Wnt target Lgr5 unambiguously showed that slender cells at the bottom of the crypt, which are clearly distinct from the +4 cells, also act as stem cells (Barker et al., 2007; Snippert et al., 2010b). Lgr5+ cells give rise to all of the differentiated cells types of the intestinal villus and divide hundreds of times during the life of the animal. Highly specic stem cell markers such as Lgr5 are still missing for many other tissues and their identication should be a high priority for the eld. Markers specic for stem cells and progenitors in organs such as lung,
Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc. 685

Cell Stem Cell

Perspective
At labeling
A

Short chase

Long chase

reserve stem cell

active stem cell

transient amplifying cells

Mature cells
Figure 2. Schematic of a System with Two Stem Cells
Both the active and reserve stem cells are capable of self-renewal. In most situations, including normal tissue maintenance, only the active stem cell contributes to renewal of the tissue, dividing continuously and giving rise to an amplifying compartment and mature cells. When the active stem cell is lost, a normally dormant (label-retaining) reserve stem cell is activated and replenishes the active stem cell compartment.

Figure 1. Lineage Tracing in Adult Tissue Stem Cell Systems


(A) A self-renewing stem cell is labeled with a specic marker (green). All cells in the hierarchy are permanently marked in long-term follow up. (B) Specic labeling of mature, differentiated cells. If these cells are replaced by a mature stem/progenitor, the label will disappear from the mature population with time. (C) A progenitor, not capable of indenite self-renewal, is labeled. After short follow-up, multiple lineages are marked. However, the marked cells disappear after longer times.

heart, and liver would facilitate studies of the ontogeny of cells formed during adult organ maintenance and repair and would probably resolve some of the current controversies. Specic stem cell markers would also allow the determination of whether cancers originate from that population, as shown for Lgr5+ colon stem cells (Barker et al., 2009). Label Retention and Quiescence: Stem Cell Markers of the Past? Perhaps the most generally signicant consequence of the discovery of the Lgr5+ intestinal stem cell is our changed view of quiescence as an essential property of tissue stem cells. Label retention was once considered a hallmark of tissue stem cells et al., 2003; Mackenzie and Bicken(Clarke et al., 2003; Duvillie bach, 1985; Morris and Potten, 1994; Terskikh et al., 2011). In order to identify label-retaining cells (LRCs), replicating DNA was labeled by administration of radioactive thymidine or BrdU for a prolonged period, followed by an even longer washout
686 Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc.

time. Tissues were then stained and cells that still retained their labels months or years after labeling were considered to be stem cells (Potten and Hendry, 1975). This paradigm was based on the notion that stem cells should persist in tissues and divide only rarely to protect their genome. Work in the epidermis had already clearly shown that LRC were not the only stem cells in skin (Claudinot et al., 2005; Li and Clevers, 2010), but the very rapid rate of division in Lgr5+ intestinal stem cells plainly demonstrated that quiescence and label retention are not inherent properties of all tissue stem cells. What then is the role of the label-retaining +4 intestinal stem cell (Potten et al., 2009)? Further investigation of this topic has revealed another important new concept in tissue stem cell biology: there can be more than one type of stem cell (Figure 2; Li and Clevers, 2010). Although it is clear that Lgr5+ stem cells are responsible for self-renewal during normal homeostasis, certain injuries, particularly irradiation, result in the activation of the label-retaining +4 stem cell (Takeda et al., 2011; Tian et al., 2011; Yan et al., 2012). Activated +4 cells are able to produce new Lgr5+ cells, both in vivo and in vitro. Hence, the +4 position intestinal stem cell is the prototype of a facultative tissue stem cell, called upon only in very specic circumstances (Figure 2). Although the +4 cells are normally quiescent, lineage tracing with Hopx as a specic marker for this population has shown that the relationship between the two stem cells is bidirectional (Takeda et al., 2011). Depending on the situation, Lgr5+ intestinal stem cells can produce Hopx+ +4 cells and vice versa (Figure 2).

Cell Stem Cell

Perspective
Multiple distinct stem cell populations are also found in the skin. Here the classic LRConce thought to be the only stem cellresides in the bulge of the hair follicle. Normally quiescent, these cells can spring into action after treatment with phorbol esters (Cotsarelis et al., 1990). Although bulge LRCs clearly can act as stem cells, an independent population, labeled by Lgr6, cycles continuously and renews all skin lineages for the life of the animal (Jaks et al., 2010; Snippert et al., 2010a). Evidence for facultative stem cells also exists in several other tissues, including those with slow turnover in the adult such as the liver (Dorrell et al., 2011; Duncan et al., 2009) or muscle (Tedesco et al., 2010). In the liver, replication of mature hepatocytes is responsible for tissue maintenance and also many injury responses (Michalopoulos, 2010). However, normally dormant facultative adult stem cells can be activated by certain injuries (Furuyama et al., 2011; Malato et al., 2011), proliferate, and give rise to multiple differentiated lineages until the tissue is restored. By denition, facultative stem cells are quiescent and thus it appears that the label retention methodology is particularly well suited to identify this kind of stem cell. Because the utilization of more than one progenitor cell compartment during normal tissue maintenance and injury has now been shown in multiple examples, it will be important to carefully evaluate multiple kinds of clinically relevant injury paradigms in each tissue. Different physiological and pathological stressors might very well give signicantly different answers regarding the identity and molecular regulation of the resident stem cell. In Vitro Clonogenic Assays: Faithful or Misleading? When lineage tracing tools and/or transplantation assays are not available, clonal self-renewal in vitro coupled with in vitro differentiation can be used as a surrogate to identify putative stem/ progenitor cells. If these assays are combined with cell fractionation methods such as FACS, it becomes possible to identify the most clonogenic cells and interrogate their properties, including the signals required for self-renewal. After clonogenic growth, differentiation protocols can be tested to determine the signals that govern maturation. Much progress has been made in this area for many different tissue stem cell systems. New surface markers have been developed and colony-forming assays now exist for most tissues. Because genetic lineage tracing is not possible in humans and cell transplantation assays are not available for many tissues, clonogenic growth is also used as an assay for human tissue stem cells, even when their murine counterparts are tractable by the other methods. In many tissues, for example the liver and intestine, there is excellent congruence between the clonogenic assays and lineage tracing (Dorrell et al., 2011; Sato et al., 2009). However, there are also several examples where in vitro assays produce different answers than the more denitive in vivo genetic lineage tracing. Normal prostate gland epithelium contains two main differentiated cell types: luminal columnar epithelium and basal cells below the luminal layer (Wang et al., 2009). In the mouse prostate, Lin/Sca-1+/ CD49f+ cells are most clonogenic in sphere-forming assays and most capable of forming prostate tissue in renal grafts (Lawson et al., 2007), indicating that prostate stem cells reside in the basal layer. Using similar assays, human prostate stem cells are also thought to be basal in nature (Garraway et al., 2010). Interestingly, however, genetic lineage tracing experiments using the transcription factor Nkx3.1 as a marker indicate that a completely distinct, rare luminal epithelial cell is responsible for regenerating the prostate after androgen deprivation (Wang et al., 2009). Once again, these ndings suggest that there may be more than one type of stem cell in this tissue. However, it is also possible that the clonogenic basal cells are not stem cells in vivo at all and that the in vitro assay does not faithfully reect tissue biology. Similarly, in vitro experiments had long suggested that the ducts of the adult pancreas might be able to produce endocrine b cells by neogenesis. Several publications agree that ducts are the progenitors for the acinar exocrine pancreatic cells (Criscimanna et al., 2011; Furuyama et al., 2011; Kopp et al., 2011b; Xu et al., 2008), but there is discordance in regards to the origin of b cells. When denitive genetic lineage tracing tools were used to specically identify the progeny of pancreatic ducts in vivo (Furuyama et al., 2011; Kopp et al., 2011b), it was clearly shown that they do not ever produce new b cells, even after extensive injury such as duct ligation or b cell ablation (Kopp et al., 2011a). These cautionary tales illustrate that in vitro clonogenic assays can be misleading and that cells identied in this fashion need to be validated as true tissue stem cells by more denitive methodology. Tissues with No Need for Stem Cells? One of the most interesting recent ndings is that some adult tissues do not appear to have a stem cell at all, which may explain their vulnerability to cell loss. Most notable in this regard is the endocrine pancreas. This tissue also exemplies an interesting variation of the genetic lineage tracing methodology. If a specic stem cell marker is not available, one can determine the existence or assess the importance of stem cells by using a marker that is highly specic for their mature, differentiated progeny (Figure 1B). In the case of the endocrine pancreas such a marker is insulin, which is expressed only in b cells. Adult b cells can be genetically labeled and their fate, and that of their derivatives, followed over time (Dor et al., 2004; Malato et al., 2011). This kind of lineage tracing experiment allows asking the question whether unlabeled cells contribute to the lineage over an extended period. Emergence of unlabeled b cells would indicate the presence of a stem cell or progenitor. Unexpectedly, such experiments based on labeling of insulin+ cells showed that b cells did not derive from insulin stem cells or progenitors in postnatal life during normal tissue maintenance (Dor et al., 2004). Instead, newly born b cells were generated only by division of pre-existing b cells. This observation has challenged the previous notion that the adult pancreas harbors b cell stem cells or progenitors. Similar experiments were done for the liver utilizing transthyretin as a marker of mature hepatocytes (Malato et al., 2011). The experiments showed that liver homeostasis in adult mice is achieved by hepatocyte replication similar to the b cell situation (Dor et al., 2004). Unlike the pancreas, however, specic injuries revealed the emergence of unlabeled hepatocytes, which must be derived from facultative stem cells or progenitors. Lineage tracing with mature markers nicely complements genetic marking studies using stem cell-specic genes. The use of both approaches in the same tissue increases the condence in models of tissue stem cell behavior if the results are congruent (Dorrell et al., 2011;
Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc. 687

Cell Stem Cell

Perspective
Furuyama et al., 2011; Iverson et al., 2011; Malato et al., 2011; Shin et al., 2011). Thus, the liver has now been rmly added to the list of organs that harbors an adult tissue stem cell. Interestingly, similar certainty is still lacking for some very important tissues. For example, there currently is no convincing evidence that the podocytes of the renal glomerulus have an adult progenitor. Niches Once a stem cell is identied and localized, it becomes possible to study the extrinsic regulatory signals produced by the local microenvironment, the niche (reviewed in Lander et al., 2012). This particular aspect of tissue stem cell biology is especially important for cancer. Many of the factors that impact healthy adult tissue stem cells also play a role in tumor self-renewal (Burness and Sipkins, 2010; Cabarcas et al., 2011). Each tissue and each stem cell uses different niche signals, but the Wnt/ b-catenin pathway has received the most attention overall and has been found to play a key role in the intestine, skin (Haegebarth and Clevers, 2009), and breast (Tanos and Brisken, 2008). The full elucidation of interactions between niches and stem cells will undoubtedly be one of the main topics of tissue stem cell research in the future. Therapeutic opportunities abound in this area because molecules to block or activate extrinsic signaling pathways could be used to affect the behavior of particular stem cells, both in terms of ablation (cancer) and expansion (tissue repair). An intriguing application of the knowledge garnered from the study of the intestinal stem cell niche is the bacterial delivery of b-catenin-lowering transkingdom RNAi to lower the risk of cancer in familial adenomatosis of the colon (FAP) (Xiang et al., 2009). miRNAs: Regulators of Stem Cell Quiescence and Self-Renewal Our newfound ability to isolate and highly purify tissue stem cells in some organs has also opened up the study of intrinsic signals. In addition to transcription factors, microRNAs (miRNAs) have emerged as key players in determining the fate of tissue stem cells and their progeny. Improved methods for proling miRNAs in very small numbers of cells (Smith and Murray, 2012) have facilitated identication of major changes in specic miRNAs as stem cells are activated and embark on differentiation (Yi and Fuchs, 2011). A single microRNA, miR-489, maintains satellite cells (the facultative stem cells of skeletal muscle) in quiescence and its downregulation results in satellite cell activation (Cheung et al., 2012). Conversely, miR-1 and miR-206 are upregulated when satellite cells differentiate and suppress Pax7 message, a transcription factor that is important for maintaining satellite cell quiescence (Chen et al., 2010). In addition to their important roles in somatic stem cell activation and differentiation, miRNAs can also be important for self-renewal. Recent examples include miR-205, which is highly expressed in mammary stem cells and can drive self-renewal when overexpressed (Greene et al., 2010; Ibarra et al., 2007). In the skin, inducible overexpression of miR-125b in adult stem cells produces an increased stem cell pool (Zhang et al., 2011). miRNAs also regulate the regenerative abilities of mature cells, such as recently shown for miR-21 in hepatocytes (Ng et al., 2012).
688 Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc.

Conclusions and Future Directions The past 5 years have seen the transition from fairly nonspecic methods for the identication of tissue stem cells such as label retention to more denitive methodology, especially genetic lineage tracing. The ability to prospectively isolate tissue stem cells with high purity and to visualize them within their niches is producing incisive insights into the intrinsic and extrinsic (niche) regulation of their behavior. Recent progress in intestinal stem cell biology has been particularly impressive and it can be hoped that other important organ systems such as the lung, liver, and kidney will follow suit in the near future.
ACKNOWLEDGMENTS This work was supported by NIH grant DK51592 (M.G.)

REFERENCES Barker, N., van Es, J.H., Kuipers, J., Kujala, P., van den Born, M., Cozijnsen, M., Haegebarth, A., Korving, J., Begthel, H., Peters, P.J., and Clevers, H. (2007). Identication of stem cells in small intestine and colon by marker gene Lgr5. Nature 449, 10031007. Barker, N., Ridgway, R.A., van Es, J.H., van de Wetering, M., Begthel, H., van den Born, M., Danenberg, E., Clarke, A.R., Sansom, O.J., and Clevers, H. (2009). Crypt stem cells as the cells-of-origin of intestinal cancer. Nature 457, 608611. Barker, N., Huch, M., Kujala, P., van de Wetering, M., Snippert, H.J., van Es, J.H., Sato, T., Stange, D.E., Begthel, H., van den Born, M., et al. (2010). Lgr5(+ve) stem cells drive self-renewal in the stomach and build long-lived gastric units in vitro. Cell Stem Cell 6, 2536. Burness, M.L., and Sipkins, D.A. (2010). The stem cell niche in health and malignancy. Semin. Cancer Biol. 20, 107115. Cabarcas, S.M., Mathews, L.A., and Farrar, W.L. (2011). The cancer stem cell nichethere goes the neighborhood? Int. J. Cancer 129, 23152327. Chen, J.F., Tao, Y., Li, J., Deng, Z., Yan, Z., Xiao, X., and Wang, D.Z. (2010). microRNA-1 and microRNA-206 regulate skeletal muscle satellite cell proliferation and differentiation by repressing Pax7. J. Cell Biol. 190, 867879. Cheung, T.H., Quach, N.L., Charville, G.W., Liu, L., Park, L., Edalati, A., Yoo, B., Hoang, P., and Rando, T.A. (2012). Maintenance of muscle stem-cell quiescence by microRNA-489. Nature 482, 524528. Clarke, R.B., Anderson, E., Howell, A., and Potten, C.S. (2003). Regulation of human breast epithelial stem cells. Cell Prolif. 36 (Suppl 1 ), 4558. Claudinot, S., Nicolas, M., Oshima, H., Rochat, A., and Barrandon, Y. (2005). Long-term renewal of hair follicles from clonogenic multipotent stem cells. Proc. Natl. Acad. Sci. USA 102, 1467714682. Cotsarelis, G., Sun, T.T., and Lavker, R.M. (1990). Label-retaining cells reside in the bulge area of pilosebaceous unit: implications for follicular stem cells, hair cycle, and skin carcinogenesis. Cell 61, 13291337. Criscimanna, A., Speicher, J.A., Houshmand, G., Shiota, C., Prasadan, K., Ji, B., Logsdon, C.D., Gittes, G.K., and Esni, F. (2011). Duct cells contribute to regeneration of endocrine and acinar cells following pancreatic damage in adult mice. Gastroenterology 141, 14511462, 1462, e1e6. Dor, Y., Brown, J., Martinez, O.I., and Melton, D.A. (2004). Adult pancreatic beta-cells are formed by self-duplication rather than stem-cell differentiation. Nature 429, 4146. Dorrell, C., Erker, L., Schug, J., Kopp, J.L., Canaday, P.S., Fox, A.J., Smirnova, O., Duncan, A.W., Finegold, M.J., Sander, M., et al. (2011). Prospective isolation of a bipotential clonogenic liver progenitor cell in adult mice. Genes Dev. 25, 11931203. Duncan, A.W., Dorrell, C., and Grompe, M. (2009). Stem cells and liver regeneration. Gastroenterology 137, 466481.

Cell Stem Cell

Perspective
, B., Attali, M., Aiello, V., Quemeneur, E., and Scharfmann, R. (2003). Duvillie Label-retaining cells in the rat pancreas: location and differentiation potential in vitro. Diabetes 52, 20352042. Furuyama, K., Kawaguchi, Y., Akiyama, H., Horiguchi, M., Kodama, S., Kuhara, T., Hosokawa, S., Elbahrawy, A., Soeda, T., Koizumi, M., et al. (2011). Continuous cell supply from a Sox9-expressing progenitor zone in adult liver, exocrine pancreas and intestine. Nat. Genet. 43, 3441. Garraway, I.P., Sun, W., Tran, C.P., Perner, S., Zhang, B., Goldstein, A.S., Hahm, S.A., Haider, M., Head, C.S., Reiter, R.E., et al. (2010). Human prostate sphere-forming cells represent a subset of basal epithelial cells capable of glandular regeneration in vivo. Prostate 70, 491501. Greene, S.B., Herschkowitz, J.I., and Rosen, J.M. (2010). The ups and downs of miR-205: identifying the roles of miR-205 in mammary gland development and breast cancer. RNA Biol. 7, 300304. Haegebarth, A., and Clevers, H. (2009). Wnt signaling, lgr5, and stem cells in the intestine and skin. Am. J. Pathol. 174, 715721. Ibarra, I., Erlich, Y., Muthuswamy, S.K., Sachidanandam, R., and Hannon, G.J. (2007). A role for microRNAs in maintenance of mouse mammary epithelial progenitor cells. Genes Dev. 21, 32383243. Iverson, S.V., Comstock, K.M., Kundert, J.A., and Schmidt, E.E. (2011). Contributions of new hepatocyte lineages to liver growth, maintenance, and regeneration in mice. Hepatology 54, 655663. rd, R. (2010). The hair follicle-a stem cell zoo. Jaks, V., Kasper, M., and Toftga Exp. Cell Res. 316, 14221428. Kopp, J.L., Dubois, C.L., Hao, E., Thorel, F., Herrera, P.L., and Sander, M. (2011a). Progenitor cell domains in the developing and adult pancreas. Cell Cycle 10, 19211927. Kopp, J.L., Dubois, C.L., Schaffer, A.E., Hao, E., Shih, H.P., Seymour, P.A., Ma, J., and Sander, M. (2011b). Sox9+ ductal cells are multipotent progenitors throughout development but do not produce new endocrine cells in the normal or injured adult pancreas. Development 138, 653665. Kretzschmar, K., and Watt, F.M. (2012). Lineage tracing. Cell 148, 3345. Lander, A.D., Kimble, J., Clevers, H., Fuchs, E., Montarras, D., Buckingham, M., Calof, A.L., Trumpp, A., and Oskarsson, T. (2012). What does the concept of the stem cell niche really mean today? BMC Biol. 10, 19. Lawson, D.A., Xin, L., Lukacs, R.U., Cheng, D., and Witte, O.N. (2007). Isolation and functional characterization of murine prostate stem cells. Proc. Natl. Acad. Sci. USA 104, 181186. Li, L., and Clevers, H. (2010). Coexistence of quiescent and active adult stem cells in mammals. Science 327, 542545. Mackenzie, I.C., and Bickenbach, J.R. (1985). Label-retaining keratinocytes and Langerhans cells in mouse epithelia. Cell Tissue Res. 242, 551556. rmann, N., Ng, R., Wang, B., Zape, J., Kay, M.A., Malato, Y., Naqvi, S., Schu Grimm, D., and Willenbring, H. (2011). Fate tracing of mature hepatocytes in mouse liver homeostasis and regeneration. J. Clin. Invest. 121, 48504860. Michalopoulos, G.K. (2010). Liver regeneration after partial hepatectomy: critical analysis of mechanistic dilemmas. Am. J. Pathol. 176, 213. Morris, R.J., and Potten, C.S. (1994). Slowly cycling (label-retaining) epidermal cells behave like clonogenic stem cells in vitro. Cell Prolif. 27, 279289. Ng, R., Song, G., Roll, G.R., Frandsen, N.M., and Willenbring, H. (2012). A microRNA-21 surge facilitates rapid cyclin D1 translation and cell cycle progression in mouse liver regeneration. J. Clin. Invest. 122, 10971108. Potten, C.S., and Hendry, J.H. (1975). Differential regeneration of intestinal proliferative cells and cryptogenic cells after irradiation. Int. J. Radiat. Biol. Relat. Stud. Phys. Chem. Med. 27, 413424. Potten, C.S., Gandara, R., Mahida, Y.R., Loefer, M., and Wright, N.A. (2009). The stem cells of small intestinal crypts: where are they? Cell Prolif. 42, 731750. Sato, T., Vries, R.G., Snippert, H.J., van de Wetering, M., Barker, N., Stange, D.E., van Es, J.H., Abo, A., Kujala, P., Peters, P.J., and Clevers, H. (2009). Single Lgr5 stem cells build crypt-villus structures in vitro without a mesenchymal niche. Nature 459, 262265. Shackleton, M., Vaillant, F., Simpson, K.J., Stingl, J., Smyth, G.K., AsselinLabat, M.L., Wu, L., Lindeman, G.J., and Visvader, J.E. (2006). Generation of a functional mammary gland from a single stem cell. Nature 439, 8488. Shin, S., Walton, G., Aoki, R., Brondell, K., Schug, J., Fox, A., Smirnova, O., Dorrell, C., Erker, L., Chu, A.S., et al. (2011). Foxl1-Cre-marked adult hepatic progenitors have clonogenic and bilineage differentiation potential. Genes Dev. 25, 11851192. Smith, S.M., and Murray, D.W. (2012). An overview of microRNA methods: expression proling and target identication. Methods Mol. Biol. 823, 119138. Snippert, H.J., Haegebarth, A., Kasper, M., Jaks, V., van Es, J.H., Barker, N., van de Wetering, M., van den Born, M., Begthel, H., Vries, R.G., et al. (2010a). Lgr6 marks stem cells in the hair follicle that generate all cell lineages of the skin. Science 327, 13851389. Snippert, H.J., van der Flier, L.G., Sato, T., van Es, J.H., van den Born, M., Kroon-Veenboer, C., Barker, N., Klein, A.M., van Rheenen, J., Simons, B.D., and Clevers, H. (2010b). Intestinal crypt homeostasis results from neutral competition between symmetrically dividing Lgr5 stem cells. Cell 143, 134144. Stingl, J., Eirew, P., Ricketson, I., Shackleton, M., Vaillant, F., Choi, D., Li, H.I., and Eaves, C.J. (2006). Purication and unique properties of mammary epithelial stem cells. Nature 439, 993997. Takahashi, K., Tanabe, K., Ohnuki, M., Narita, M., Ichisaka, T., Tomoda, K., and Yamanaka, S. (2007). Induction of pluripotent stem cells from adult human broblasts by dened factors. Cell 131, 861872. Takeda, N., Jain, R., LeBoeuf, M.R., Wang, Q., Lu, M.M., and Epstein, J.A. (2011). Interconversion between intestinal stem cell populations in distinct niches. Science 334, 14201424. Tanos, T., and Brisken, C. (2008). What signals operate in the mammary niche? Breast Dis. 29, 6982. Tedesco, F.S., Dellavalle, A., Diaz-Manera, J., Messina, G., and Cossu, G. (2010). Repairing skeletal muscle: regenerative potential of skeletal muscle stem cells. J. Clin. Invest. 120, 1119. Terskikh, V.V., Vasiliev, A.V., and Vorotelyak, E.A. (2011). Label retaining cells and cutaneous stem cells. Stem Cell Rev., in press. Published online July 9, 2011. 10.1007/s12015-011-9299-6. Tian, H., Biehs, B., Warming, S., Leong, K.G., Rangell, L., Klein, O.D., and de Sauvage, F.J. (2011). A reserve stem cell population in small intestine renders Lgr5-positive cells dispensable. Nature 478, 255259. van der Kooy, D., and Weiss, S. (2000). Why stem cells? Science 287, 1439 1441. Wang, X., Kruithof-de Julio, M., Economides, K.D., Walker, D., Yu, H., Halili, M.V., Hu, Y.P., Price, S.M., Abate-Shen, C., and Shen, M.M. (2009). A luminal epithelial stem cell that is a cell of origin for prostate cancer. Nature 461, 495500. Weissman, I.L., Anderson, D.J., and Gage, F. (2001). Stem and progenitor cells: origins, phenotypes, lineage commitments, and transdifferentiations. Annu. Rev. Cell Dev. Biol. 17, 387403. Xiang, S., Keates, A.C., Fruehauf, J., Yang, Y., Guo, H., Nguyen, T., and Li, C.J. (2009). In vitro and in vivo gene silencing by TransKingdom RNAi (tkRNAi). Methods Mol. Biol. 487, 147160. , G., Bonne , S., De Leu, N., Xiao, X., Van de CasXu, X., DHoker, J., Stange teele, M., Mellitzer, G., Ling, Z., Pipeleers, D., et al. (2008). Beta cells can be generated from endogenous progenitors in injured adult mouse pancreas. Cell 132, 197207. Yan, K.S., Chia, L.A., Li, X., Ootani, A., Su, J., Lee, J.Y., Su, N., Luo, Y., Heilshorn, S.C., Amieva, M.R., et al. (2012). The intestinal stem cell markers Bmi1 and Lgr5 identify two functionally distinct populations. Proc. Natl. Acad. Sci. USA 109, 466471. Yi, R., and Fuchs, E. (2011). MicroRNAs and their roles in mammalian stem cells. J. Cell Sci. 124, 17751783. Zhang, L., Stokes, N., Polak, L., and Fuchs, E. (2011). Specic microRNAs are preferentially expressed by skin stem cells to balance self-renewal and early lineage commitment. Cell Stem Cell 8, 294308.

Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc. 689

Cell Stem Cell

Perspective
Mesenchymal Stromal Cells: New Directions
Armand Keating1,*
1Cell Therapy Program, Princess Margaret Hospital, Toronto, ON M5G 2M9, Canada *Correspondence: armand.keating@uhn.on.ca DOI 10.1016/j.stem.2012.05.015

Research into mesenchymal stromal/stem cells (MSCs) has been particularly exciting in the past ve years. Our understanding of mechanisms of MSC-mediated tissue regeneration has undergone considerable evolution. Recent investigation of the primary in situ counterpart of cultured MSCs has led to fresh insights into MSC physiology and its role in the immune system. At the same time, the clinical application of MSCs continues to increase markedly. Taken together, a reappraisal of the denition of MSCs, a review of current research directions, and a reassessment of the approach to clinical investigation are timely and prudent.
Introduction Few cell types have captivated so many biomedical researchers over the last 10 years as have mesenchymal stromal/stem cells (MSCs). PubMed, in 2012, identies over 17,000 references for mesenchymal stem cells and more than 4,500 for mesenchymal stromal cells. There have been several comprehensive recent reviews on MSCs (Uccelli et al., 2008; Bianco et al., 2008; Tolar et al., 2010; Ranganath et al., 2012). Hence, rather than cover all of the work in this eld, in this perspective I will focus on some areas that have seen notable advance in the past 5 years and others that warrant further investigation to improve our insight into the properties and potential of this intriguing cell population. First, a case can be made to revisit the nomenclature and denition of MSCs, not as a semantic exercise, but to better dene the direction of research. The availability of new molecular tools makes the need for rigorous denitions increasingly important. Moreover, the differences between MSC populations derived from different tissues are becoming more apparent, presenting an additional challenge to devising a universal denition. MSCs as currently dened are a phenomenon of in vitro culture, suggesting that extrapolating the function of these cells to activity in vivo must be done with caution. This limitation highlights the need for direct in vivo studies with endogenous MSCs or an equivalent physiological population as an essential next step in establishing their true biological role. It is encouraging in this regard that recent studies have employed transgenic animal models to enable the tracking and assessment of MSC-like cells in vivo. The mechanisms underlying tissue regeneration and immune modulation by therapeutic doses of MSCs also require further elucidation, particularly the extent to which the two processes intersect. The more recent appreciation that MSCs may not mediate tissue regeneration by direct cell replacement is also likely to redirect investigation into more fruitful directions. Finally, in view of the extraordinarily rapid and extensive use of MSCs clinically, a reappraisal of the approach to the development of clinical protocols based on conrmed laboratory and preclinical observations would be timely and helpful. Background MSCs were initially identied as a subpopulation of bone marrow cells with osteogenic potential as shown by heterotopic transplantation and subsequently were conrmed to contain clonal, plastic adherent bone-marrow derived nonhematopoietic cells in the mouse and guinea pig (Friedenstein et al., 1968, 1970, 1976). An in vitro colony assay developed by Friedenstein and coworkers to detect the clonogenic cell among this population (the colony-forming unit-broblast [CFU-F]) was also adapted for human marrow (Castro-Malaspina et al., 1980). Subsequent studies in the 1980s focused on the role of a similar population of bone marrow stromal cells derived from the adherent layers of long-term bone marrow cultures in supporting hematopoiesis (Dexter et al., 1977, reviewed in Clark and Keating, 1995). Caplans proposal that these cells were mesenchymal stem cells (Caplan, 1991) capable of differentiation to all cells of mesodermal lineage stimulated investigation into their role in mediating tissue regeneration. Although the multilineage differentiation potential of MSCs was later shown (Pittenger et al., 1999), in vivo demonstration that these cells possess the hallmark stem cell characteristics of self-renewal and differentiation had not been accomplished. Confusion arising from the denition of the MSC population made comparisons among published studies in the 1990s and 2000s problematic and led to the proposal of new terminology and criteria by the International Society for Cellular Therapy (ISCT) (Horwitz et al., 2005; Dominici et al., 2006). According to these widely adopted proposals, the cells were more appropriately considered mesenchymal stromal cells given that not all were stem cells (Horwitz et al., 2005). The minimum criteria for MSCs included plastic adherence and in vitro trilineage differentiation to adipogenic, chondrogenic, and osteogenic cells (Dominici et al., 2006). Additional requirements included cell surface expression of CD105 (endoglin, SH2), CD73 (ecto-50 nucleotidase), and CD90 (Thy1) and the absence of the hematopoietic markers, CD45, CD19, CD19 or CD79, CD14 or CD11b, and HLA-DR. A particular challenge for the eld has been the absence of a specic marker to dene MSCs, although a large number of different determinants have been associated, albeit not exclusively, with them (reviewed by Lindner et al., 2010 for human MSCs), including CD271 (low-afnity nerve growth factor receptor) (Jones et al., 2002) and CD146 (Sacchetti et al., 2007). MSCs are also highly active metabolically, secreting not only components of the extracellular matrix (Wight et al., 1986) but also a vast array of cytokines (reviewed by Horwitz and Dominici, 2008). More recent work has documented
Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc. 709

Cell Stem Cell

Perspective
extensively the secretome and proteome of MSCs (Ranganath et al., 2012). In addition to bone marrow, MSC populations can be obtained readily from adipose tissue (Zuk et al., 2002) and also from a variety of tissues including placenta (In t Anker et al., 2004), skin (Shih et al., 2005), umbilical cord blood (Erices et al., 2000), umbilical cord perivascular cells (Sarugaser et al., 2005), umbilical cord Whartons jelly (Wang et al., 2004), dental pulp (Gronthos et al., 2000), amniotic uid (Nadri and Soleimani, 2007), synovial membrane (De Bari et al., 2001), and breast milk (Patki et al., 2010). Revisiting the Denition of MSCs The minimum criteria for dening MSCs established earlier (Horwitz et al., 2005; Dominici et al., 2006) may now be unduly constraining for a number of reasons. First, the characteristics of MSCs may vary according to the source of tissue. In an effort to dene an MSC-like product, scientic entrepreneurs and biotechnology companies have focused on differences in surface marker prole to optimize intellectual property protection of relatively similar cell types. The recognition of speciesspecic differences in cell characteristics and generation of a variety of transcriptional and secretomic signatures for the cells also indicate diversity. Moreover, panels of reagents (especially antibodies) equivalent to those available for characterizing human MSCs are still not in place for a number of other species, so the criteria recommended by the ISCT (Dominici et al., 2006) may be difcult to meet. The challenge is to devise an appropriate denition without losing the benet that the current criteria provide in enabling evaluation of different studies of similar, if not identical, cell populations. A major hurdle is the absence of a single characteristic or marker with which to dene MSCs. Nonetheless, a re-evaluation is timely and will require consensus among leading investigators in the eld. In addition to standard methods of cell characterization of which surface marker prole and differentiation potential are the mainstays, the relative benets of more advanced molecular tools including assessments of the cell transcriptome, proteome, and secretome (Ranganath et al., 2012) should be evaluated in creating this new denition. Moreover, the need to demonstrate trilineage differentiation, especially toward the chondrogenic lineage by MSCs derived from tissues other than bone marrow, also requires reassessment. It is possible that a global denition of MSCs may now be overly simplistic or unnecessary. Specic denitions of particular MSC subsets may sufce, provided that they accurately and reproducibly dene the cells under study. For example, the socalled stromal vascular fraction (SVF) of adipose-derived cells represents a highly heterogenous cell population and contains cells that express CD90 but not CD105 until they become plastic adherent (Yoshimura et al., 2006). Nonetheless, the cells have been considered to be MSC like. This issue is of additional signicance because SVF cells have been extensively applied in clinical settings, despite a paucity of reported trials. It is unclear whether these cell products are uniformly dened prior to clinical administration. Some general concepts of a new approach to the nomenclature, denition, and characterization of MSCs may provide
710 Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc.

a framework for discussion. The rationale is to help inform the investigation of these cells rather than to serve merely as a classication: (1) The general population of MSCs should continue to be identied as mesenchymal stromal cells, although this is not an ideal term. (2) The term mesenchymal stem cell should be used to specically describe a cell with documented self-renewal and differentiation characteristics. (3) MSCs should be categorized as cultured or primarythis is an important distinction (see below) because the characteristics are likely to be different and should avoid confusion when comparisons are made between studies. (4) The source of MSCs should be specied (e.g., adipose, BM, cord blood, etc.); differences in cell characteristics are likely to be encountered. (5) Species should be identiedthis information is not always explicitly stated in the text of publications (except in the Methods section) and has led to confusion in the past. (6) Minimum criteria for a surface marker prole need to be revisited and are likely to vary among species. (7) The need to document the in vitro differentiation potential of the cells should be re-examined. (8) The in vitro clonogenic capacity of MSCs should be enumerated. (9) The reproducible representation of transcriptome, proteome, and secretome of MSCs should be evaluated and the major factors inuencing the signatures should be identied and specied. (10) Consideration should be given to characterizing the cells according to tissue specicity (e.g., the differentiation potential of human umbilical cord perivascular cells is more extensive than for BM MSCs). Stem Cell Properties of Cultured MSCs Despite numerous reviews attesting to the stem cell nature of MSCs from their ability to undergo differentiation along at least three lineages, there appear to be only three studies that can lay claim to identifying stem cells among human cultured MSCs, on the basis of rigorous clonal analysis. Muraglia et al. (2000) showed by limiting cell dilution that clones arising from single cells of bone marrow stromal cultures displayed multilineage differentiation potential and exhibited self-renewal. These authors proposed a hierarchical model in which there was sequential loss of lineage potential from the most primitive osteo-chondroadipogenic to osteo-chondrogenic, and nally to osteogenic precursors. Notably, osteo-adipogenic and chondro-adipogenic precusors were not detected, nor were purely chondrogenic or adipogenic clones. Lee et al. (2010) conducted single-cell studies of GFP-marked human MSCs (using irradiated stromal feeder layers to facilitate growth) and demonstrated that a minor subpopulation with high proliferative potential exhibited differentiation along osteogenic, chondrogenic, and adipogenic lineages and could self-renew from colony replating assays. Analyzing the clonogenic differentiation capacity of another MSC population, human umbilical cord perivascular cells

Cell Stem Cell

Perspective
(HUCPVCs), Sarugaser et al. (2009) documented the selfrenewal and multipotent capacity of an infrequent mesenchymal stem cell able to differentiate to myogenic, osteogenic, chondrogenic, adipogenic, and broblastic lineages and proposed a hierarchical stem cell lineage relationship for these cells. These examples highlight the differences in differentiation potential between cells obtained from different tissues. This is an important area of investigation because as in the case of hematopoietic stem cell lineage relationships, much can be learned from studies of MSC clones that may be lost by an investigation of a heterogeneous MSC population, even one enriched for clonogenic cells. Immunomodulatory Properties of Cultured MSCs At this point, there is a considerable body of literature documenting the pleotropic effects of MSCs on the immune system. MSCs act on both the adaptive and innate immune systems by suppressing T cells, suppressing dendritic cell maturation, reducing B cell activation and proliferation, inhibiting proliferation and cytotoxicity of NK cells, and promoting the generation of regulatory T cells via an IL-10 mechanism. The role of MSCs in mediating these processes by affecting the expression of inammatory cytokines is well established. This topic has been covered extensively in several reviews (Nauta and Fibbe, 2007; n, 2007; Uccelli et al., 2008; Tolar et al., Le Blanc and Ringde 2010; Chen et al., 2011, among others), and I will therefore focus on drawing attention to a few key issues. One major area of MSC-mediated activity is T cell suppression (Yang et al., 2009). Several recent studies have identied pathways that are involved, including downregulation of NF-kB signaling and cell cycle arrest at G0/G1 (Jones et al., 2007; Choi et al., 2011). However, it is still somewhat unclear to what extent these pathways will have physiological signicance. Some of the confusion in the literature in this area may be alleviated by the appreciation that there are major differences in the mechanisms of T cell suppression among species. For example, in humans and Rhesus monkeys, indoleamine 2,3- dioxygenase (IDO) is predominantly involved in T cell suppression, whereas nitric oxide is the main mediator in mice (Ren et al., 2008; Ren et al., 2009). One emerging area of investigation involves studies of Toll-like receptors (TLRs) on MSCs and their contribution to immune modulation. These receptors respond to so-called danger signals consisting of molecules released by injured tissue or microbial invasion (e.g., endotoxin, LPS, dsRNA, and heat shock proteins). At least ten human TLRs are known and are expressed on innate immune effector cells (Kawai and Akira, 2011). Surprisingly, functional TLR3 and TLR4 are abundantly expressed on human BM-derived MSCs. Ligation of these TLRs induces activation of proinammatory signals and prevents the suppression of T cell proliferation, possibly by MSC-mediated downregulation of Notch ligand (Liotta et al., 2008; Tomchuck et al., 2008). MSC-associated TLR signaling appears to not only involve a direct immune stress response but also the promotion of MSC migration (with TLR3 ligation). Interestingly, TLR3 and TLR4 stimulation does not appear to suppress IDO activity or PGE2 levels that decrease inammatory responses (Liotta et al., 2008) and raises important implications for the role of MSCs in host defense. These observations suggest that activation of the TLRs on MSCs may maintain antiviral host defense. TLR-mediated proinammatory responses by MSCs could potentially have additional functional implications. On the basis of the divergent patterns of TLR3 versus TLR4 ligation in a short-term assay with respect to cytokine and chemokine secretion, cell migration capabilities, TGF-b secretion, and expression of the downstream effectors, SMAD3/SMAD7, Waterman et al. (2010) proposed a novel paradigm for MSC action. In their model, MSCs can polarize to a proinammatory MSC1 type (TLR4-primed) or an immunosuppressive MSC2 (TLR3-primed) phenotype, analogous to the action of M1 versus M2 monocyte/macrophages (Dayan et al., 2011). Thus, the classical monocyte/macrophage responses to injury are reprised with the MSC1 (response to acute injury)/MSC2 (anti-inammatory/healing) model (Figure 1). It is also possible that through TLR signaling, MSCs play a pivotal role in both initiating the clearance of pathogens and promoting the repair of injured tissue, raising the possibility that MSCs could be employed clinically to augment host defense (Auletta et al., 2012). For future applications, the challenge will be to discover the key factors that contribute to achieving a balance that functions effectively in the best interests of the host. The next steps include conrmatory studies using different assays for further testing in animal models. In that regard, investigators will need to deal with the additional level of complexity from MSC-mediated augmentation of IL-10 production by macro meth et al., 2009). phages via TLR4 ligation (Ne As is true for most studies of MSCs, the bulk of these immune modulation experiments were conducted with cultured MSCs. Data generated in vivo from putatively equivalent primary MSCs (MSCs in situ) remain lacking. Unfortunately, an assessment of immune interactions of uncultured MSCs in vivo has the same limitations as those for other MSC studies: the low frequency of primary MSCs in vivo, a lack of appropriate animal models, and interspecies variation in mechanisms of action (Ren et al., 2009). Differing results may also be reconciled by taking into account opposing mechanisms to maintain immune homeostasis. Alternative explanations include differences in cell dose, assay methodology, and MSC source. Given these limitations, an attempt to extrapolate in vitro data by Uccelli et al. (2008) is laudable and possibly amenable to testing. These authors provide several intriguing potential explanations for effects in vivo. For example, the effect of infused cultured MSCs on NK and dendritic cells may result in potentially opposite interactions that eventually will be resolved by predominant microenvironmental cytokine levels. Evolving Concepts of Tissue Regeneration by MSCs Over the past decade, there has been considerable evolution in our understanding of the mechanisms underlying tissue regeneration from MSCs. Progress may have been limited to some extent by the concept of the mesenchymal stem cell and the implicit idea that the objective was cell replacement therapy. For example, the concept of transdifferentiation of hematopoietic progenitors into cardiac cells was difcult to dislodge, despite rigorous studies failing to support the idea (Murry et al., 2004; Balsam et al., 2004). It was interesting that this notion was displaced by the phenomenon of cell fusion, another
Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc. 711

Cell Stem Cell

Perspective

Figure 1. Proposed Immunomodulatory Mechanisms of Cultured MSCs


MSC-mediated immune interactions shown here include a proposed polarization of MSCs into MSC1 and MSC2 cells as a result of activation of Toll-like receptors (based on work by Waterman et al., 2010). Activation of MSC-resident TLR4 leads to a MSC1 or M1 type cell with a proinammatory response, whereas activation of TLR3 gives rise to a M2 type MSC with an anti-inammatory/immunosuppressive response. Overall outcome will depend on the balance between the cytokines/chemokines released into the microenvironment.

biological process also unlikely to account for documented improvements in preclinical models of cell treatment of injured tissue (if only because of its very low frequency). However, the possibility that partial cellular reprogramming, leading to the acquisition of some characteristics of the desired lineage, could contribute to the tissue regeneration capacity of MSCs (Rose et al., 2008) remains to be investigated. A recent example of high throughput screening using human MSCs to identify small molecules that promote chondrogenic differentiation suggests an approach that may be more fruitful (Johnson et al., 2012). These investigators showed that the small molecule kartogenin induces chondrocyte differentiation of MSCs, protects articular cartilage in vitro, and promotes cartilage repair after intra-articular injection in an osteoarthritis animal model. Whether the administration of exogenous MSC-derived chrondrocytic cells will be superior to local treatment with the heterocyclic molecule alone is not yet known. Nonetheless, a more extensive drug discovery approach to identify molecules that mediate the differentiation/reprogramming of MSCs along mesodermal lineages is an exciting prospect. Other explanations for the varying degrees of efcacy mediated by MSCs have been extensively reviewed elsewhere and are often characterized as paracrine effects. The cells are perceived to exert their effects by the release of factors that stimulate tissue recovery on many potential levels, including stimulation of endogenous stem/progenitor cells, suppression of apoptosis of vulnerable cells, remodeling of extracellular matrix, and stimulation of new blood vessel formation. Investigating MSCs as cytokine factories will likely uncover new mechanisms and identify compounds that may in some cases supplant the cells themselves (Ranganath et al., 2012). For example, tumor necrosis factor-inducible gene 6 protein (TSG6) is an immunosuppressive molecule produced by MSCs
712 Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc.

that partially recapitulates the hemodynamic improvement after intravenous infusion of the cells following experimental acute myocardial infarction in mice (Lee et al., 2009). This study serves to further underscore the shift toward the importance of the immunomodulatory properties of MSCs in regenerating injured tissue. Another example is the association between cardiac improvement and an MSC-mediated switch in macrophages/ monocytes inltrating ischemic tissue from the M1 to M2 phenotype (Dayan et al., 2011). Of interest, the switch was observed among circulating monocytes but not in the bone marrow, raising the possibility of a potentially useful distinction between more commonly accepted paracrine phenomena versus an allocrine effect produced by exogenous cells in a remote location. How MSCs communicate with endogenous cells requires further study and the contribution by which cell-cell contact mediates the biological effects needs further clarication. In this regard, exploring the role of exosomes, secreted vesicles potentially involved in intercellular communication may provide novel insights (Lai et al., 2011). Physiological Role of Primary In Situ MSCs The study of culture-expanded MSCs is unlikely to help establish the physiological role of native in vivo cells. Progress in dealing with this limitation has initially been slow, partly because potentially useful experimental tools have been employed only recently and the frequency of putative native MSCs is very low. However, momentum is growing as the importance of these studies becomes more evident. McGonagle and others have shown that the in vivo counterpart of MSCs has the following immunophenotype: CD45 or low, CD271+ (Jones et al., 2002). More recent data show that the cells within this population have greater transcriptional activity than cultured MSCs or dermal broblasts, reecting

Cell Stem Cell

Perspective
broader differentiation potential and a marked increase in the transcription of osteogenic and Wnt-related genes (Churchman et al., 2012). CD105+ cells can also be isolated in situ from human bone marrow and exhibit high levels of CFU-F activity, generate CD105+ CD90+, and CD106+ cells that undergo trilineage differentiation (adipogenic, chondorgenic and osteogenic lineages) after culture, and differentiate into osteoblasts in vivo in response to BMP-2 (Aslan et al., 2006). Other evidence indicates that the human in situ MSC in vivo is CD146+, gives rise to CFU-F, and exhibits self-renewal in vivo. These cells are also capable of forming both bone and heterotopic hematopoiesis-associated MSCs from single clones in immune-decient murine experiments. The CD146+CD45 cells are subendothelial and localize in vivo as adventitial reticular cells (Sacchetti et al., 2007). More recent work from another group has conrmed that CFU-F activity resides exclusively in the CD271+ cell population enriched directly from human marrow cells and shown that both CD271+CD146+ or CD271+CD146() cells can give rise to stromal clones that form bone ossicles and hematopoiesis-associated stromal cells (Tormin et al., 2011). The Frenette group has shown that a small proportion of MSCs are Nestin+, can self-renew in vivo, contain all the CFU-F activity of the bone marrow, and undergo osteogenic, chondrogenic, and adipogenic differentiation ndez-Ferrer et al., 2010). The relationship of these mesen(Me chymal stem cells and CXCL12-abundant reticular (CAR) cells (Sugiyama et al., 2006), which also have osteoprogenitor capacity, requires further investigation. However, short-term ablation of CAR cells in vivo impaired the ability of BM cells to undergo adipogenic and osteogenic differentiation (Omatsu et al., 2010). The Scadden group has further examined osteolineage progenitors in the MSC pool. Their recent elegant study of bone maintenance and repair (Park et al., 2012) highlights the importance of genetic tools that better dene the in vivo role of BM MSCs. They showed that a subset of Nestin+ osteolineage-restricted MSCs present in vivo are able to replace shortlived mature osteoblasts to maintain homeostasis and respond to bone injury (Park et al., 2012). Taking an innovative approach involving phage display and cell sorting, Daquinag et al. (2011) screened combinatorial libraries for peptides that target adipose stromal cells in vivo in the mouse based on the immunophenotype prole, CD34CD31CD45. They found a cell surface marker, the N-terminally truncated proteoglycan, d-decorin highly expressed on the cells in vivo and identied resistin, a known protein adipokine, as its endogenous ligand. They hypothesized that signaling by resistin via the d-decorin receptor regulates the fate of adipose stromal cells. Although observed almost in passing, the authors note that the d-decorin is localized on the cell surface that faces away from blood vessels, suggesting an opportunity to interact with extracellular matrix components. In addition, they found that culturing the stromal vascular fraction (SVF) of adipose cells under standard conditions for generating MSCs led to loss of cell surface d-decorin. These data underscore the challenges associated with identifying unique cell markers on cultured MSCs. Nonetheless, a similar approach for identifying an analogous receptor/ligand on bone marrowderived MSCs may also yield valuable information regarding the nature and biology of the native MSC in vivo. Clinical Application At this point, there is extensive clinical activity involving MSCs, and many available treatments are outside the oversight of national regulatory bodies or clinical trial sites such as ClinicalTrials.gov. Moreover, the outcomes of a large number of these treatments are not documented in peer-reviewed journals. Unfortunately, the rationale for the clinical application of MSCs, particularly in regenerative medicine, has lagged behind laboratory observations. It is important to optimize the design of MSC trials based on the most current preclinical observations to maximize their scientic rigor. Several protocols involving systemic administration of MSCs to treat injured tissue are still in progress because of the notion of cell replacement therapy rather than on the more recently accepted paracrine and anti-inammatory effects of these cells. The study outcomes are unlikely to be optimal if the major effect is actually an antiinammatory one and may arise from number of factors including inappropriate dose, scheduling, or route of administration. Furthermore, the coadministration of anti-inammatory agents may be a confounding factor. A second issue is the difculty in fully evaluating completed clinical trials for which the results have not been formally published in international peer-reviewed journals. Valuable insights into trials design, patient selection, underlying rationale, and potential improvements would be gained by rigorous peer review. Nevertheless, the results of several phase II trials with MSCs show promise. Le Blancs phase II trial using MSCs to treat steroid-resistant aGvHD (Le Blanc et al., 2008) indicates that a multicenter randomized controlled trial should be conducted. Because several transplant centers already routinely employ MSCs for that indication on the basis of only the phase II data, the need for a randomized controlled trial seems quite urgent. MSCs were also tested for their ability to support kidney transplantation on the basis of the promising data treating aGvHD. In an open label prospective trial, 159 patients undergoing a living related donor kidney transplant were registered for randomization to receive IL-2 receptor antibody induction therapy versus autologous BM-MSCs to assess rejection rate (Tan et al., 2012). Although patient and graft survival were similar, patients receiving MSCs had a lower incidence of acute rejection, decreased probability of opportunistic infections, and better kidney function 1 year later. In addition, preclinical data have suggested that MSCs may have a role in the management of acute myocardial infarction. An industry-sponsored randomized double-blind placebocontrolled dose escalation study of systemically administered MSCs after acute myocardial infarction in 53 patients provided safety and preliminary efcacy data (Hare et al., 2009). Adverse events were similar between the test and placebo groups over a 6 month period. Ventricular arrhythmias were reduced (p = 0.025) and pulmonary function improved (p = 0.003) in patients receiving MSCs. In a subset analysis, patients with an anterior acute myocardial infarct had improved ventricular function (ejection fraction) compared with the placebo cohort. These are encouraging data and a prospective randomized trial with clinically signicant endpoints is awaited with interest. Well-designed clinical trials will be critical for determining whether MSCs can be effective in treating tissue injury or
Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc. 713

Cell Stem Cell

Perspective
immune disorders. Success is more likely if clinician investigators work very closely with laboratory researchers to design better clinical trials. Particular attention should be paid to factors that may be overlooked but could affect the efcacy of MSCs, including culture conditions/medium, oxygen tension, time from thawing after cryopreservation to administration, the tissue source of the MSCs, priming or activation prior to administration, the use of gene engineered versus unmodied cells, MSC subsets, and autologous versus allogeneic cells. Given the serious limitations associated with testing human MSCs in animal models, detailed analysis of immune and other perturbations in patients participating in prospective trials should be undertaken to help optimize subsequent protocols. Two other areas also require further attention. The importance of cell tracking and persistence of the exogenous MSCs in vivo in affecting clinical outcome can only be addressed when the cells are safely and effectively labeled and monitored. Although several studies have looked at the persistence of MSCs in animal models, similar correlative studies are required in human study recipients. Currently, the only viable option is superparamagnetic iron nanoparticles and magnetic resonance imaging, but even this approach has very limited availability. Finally, given the underreporting of MSC treatments and the paucity of publications describing long-term follow-up after MSC administration, a convincing case can be made for establishing a database registry of cell therapy recipients to track treatment outcomes and monitor for long-term adverse events. Two additional aspects suggest that clinical correlative studies are warranted. Although the recent observation of the acquisition of chromosomal aberrations in cultured human adult stem cells (Ben-David et al., 2011) is of uncertain signicance et al., 2012; Prockop and Keating, 2012), it will (Sensebe be important to correlate assays of genetic instability prospectively with clinical outcome. The other aspect relates to the interaction of MSCs with cancer. Although there is a growing body of literature in this area (Djouad et al., 2003), the outcomes of experimental models appear to be conicting. A spectrum of responses has been observed with different tumor models, from tumor suppression to stimulation. Clinical correlation with studies of the signaling pathways involved in stromal-tumor interactions is an important goal that should also accompany the establishment of a cell therapy patient registry. Conclusions In summary, the past 5 years have been remarkably active for MSC studies. Several initiatives can be undertaken to further accelerate the process of enhancing our understanding of MSC biology and improve access to well-designed clinical trials. Current denitions of this cell population need to be revisited given the wide range of tissue sources and the recognition of subpopulations with specic properties. Extending the availability of an international reference MSC repository for access to all investigators is also a priority. Additional animal models need to be developed to better identify and study the biology of primary in situ MSCs. The design of optimal clinical trials requires the close cooperation of laboratory and clinical investigators. Future studies need to be designed that also include assaying perturbations in patients in vivo and are therefore
714 Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc.

best positioned to overcome some of the inherent limitations of xenogeneic animal models with human MSCs.
ACKNOWLEDGMENTS A.K. is a consultant for the biotech start-ups TRT (Toronto, ON), involved in HUCPVC technology, and Temple Therapeutics (Temple, TX), involved in TSG6. REFERENCES Aslan, H., Zilberman, Y., Kandel, L., Liebergall, M., Oskouian, R.J., Gazit, D., and Gazit, Z. (2006). Osteogenic differentiation of noncultured immunoisolated bone marrow-derived CD105+ cells. Stem Cells 24, 17281737. Auletta, J.J., Deans, R.J., and Bartholomew, A.M. (2012). Emerging roles for multipotent, bone marrow-derived stromal cells in host defense. Blood 119, 18011809. Balsam, L.B., Wagers, A.J., Christensen, J.L., Kodis, T., Weissman, I.L., and Robbins, R.C. (2004). Haematopoietic stem cells adopt mature haematopoietic fates in ischaemic myocardium. Nature 428, 668673. Ben-David, U., Mayshar, Y., and Benvenisty, N. (2011). Large-scale analysis reveals acquisition of lineage-specic chromosomal aberrations in human adult stem cells. Cell Stem Cell 9, 97102. Bianco, P., Robey, P.G., and Simmons, P.J. (2008). Mesenchymal stem cells: revisiting history, concepts, and assays. Cell Stem Cell 2, 313319. Caplan, A.I. (1991). Mesenchymal stem cells. J. Orthop. Res. 9, 641650. Castro-Malaspina, H., Gay, R.E., Resnick, G., Kapoor, N., Meyers, P., Chiarieri, D., McKenzie, S., Broxmeyer, H.E., and Moore, M.A. (1980). Characterization of human bone marrow broblast colony-forming cells (CFU-F) and their progeny. Blood 56, 289301. Chen, P.-M., Yen, M.-L., Liu, K.-J., Sytwu, H.-K., and Yen, B.-L. (2011). Immunomodulatory properties of human adult and fetal multipotent mesenchymal stem cells. J. Biomed. Sci. 18, 49. Choi, H., Lee, R.H., Bazhanov, N., Oh, J.Y., and Prockop, D.J. (2011). Antiinammatory protein TSG-6 secreted by activated MSCs attenuates zymosan-induced mouse peritonitis by decreasing TLR2/NF-kB signaling in resident macrophages. Blood 118, 330338. Churchman, S.M., Ponchel, F., Boxall, S.A., Cuthbert, R., Kouroupis, D., Roshdy, T., Giannoudis, P.V., Emery, P., McGonagle, D., and Jones, E.A. (2012). Native CD271(+) multipotential stromal cells (MSCs) have a transcript prole indicative of multiple fates with prominent osteogenic and Wnt pathway signalling activity. Arthritis Rheum. Published online February 29, 2012. 10.1002/art.34434. Clark, B.R., and Keating, A. (1995). Biology of bone marrow stroma. Ann. N Y Acad. Sci. 770, 7078. Daquinag, A.C., Zhang, Y., Amaya-Manzanares, F., Simmons, P.J., and Kolonin, M.G. (2011). An isoform of decorin is a resistin receptor on the surface of adipose progenitor cells. Cell Stem Cell 9, 7486. Dayan, V., Yannarelli, G., Billia, F., Filomeno, P., Wang, X.-H., Davies, J.E., and Keating, A. (2011). Mesenchymal stromal cells mediate a switch to alternatively activated monocytes/macrophages after acute myocardial infarction. Basic Res. Cardiol. 106, 12991310. De Bari, C., DellAccio, F., Tylzanowski, P., and Luyten, F.P. (2001). Multipotent mesenchymal stem cells from adult human synovial membrane. Arthritis Rheum. 44, 19281942. Dexter, T.M., Allen, T.D., and Lajtha, L.G. (1977). Conditions controlling the proliferation of haemopoietic stem cells in vitro. J. Cell. Physiol. 91, 335344. l, D., Djouad, F., Plence, P., Bony, C., Tropel, P., Apparailly, F., Sany, J., Noe and Jorgensen, C. (2003). Immunosuppressive effect of mesenchymal stem cells favors tumor growth in allogeneic animals. Blood 102, 38373844. Dominici, M., Le Blanc, K., Mueller, I., Slaper-Cortenbach, I., Marini, F.C., Krause, D.S., Deans, R.J., Keating, A., Prockop, Dj., and Horwitz, E.M. (2006). Minimal criteria for dening multipotent mesenchymal stromal cells.

Cell Stem Cell

Perspective
The International Society for Cellular Therapy position statement. Cytotherapy 8, 315317. Erices, A., Conget, P., and Minguell, J.J. (2000). Mesenchymal progenitor cells in human umbilical cord blood. Br. J. Haematol. 109, 235242. Friedenstein, A.J., Petrakova, K.V., Kurolesova, A.I., and Frolova, G.P. (1968). Heterotopic of bone marrow. Analysis of precursor cells for osteogenic and hematopoietic tissues. Transplantation 6, 230247. Friedenstein, A.J., Chailakhjan, R.K., and Lalykina, K.S. (1970). The development of broblast colonies in monolayer cultures of guinea-pig bone marrow and spleen cells. Cell Tissue Kinet. 3, 393403. Friedenstein, A.J., Gorskaja, J.F., and Kulagina, N.N. (1976). Fibroblast precursors in normal and irradiated mouse hematopoietic organs. Exp. Hematol. 4, 267274. Gronthos, S., Mankani, M., Brahim, J., Robey, P.G., and Shi, S. (2000). Postnatal human dental pulp stem cells (DPSCs) in vitro and in vivo. Proc. Natl. Acad. Sci. USA 97, 1362513630. Hare, J.M., Traverse, J.H., Henry, T.D., Dib, N., Strumpf, R.K., Schulman, S.P., Gerstenblith, G., DeMaria, A.N., Denktas, A.E., Gammon, R.S., et al. (2009). A randomized, double-blind, placebo-controlled, dose-escalation study of intravenous adult human mesenchymal stem cells (prochymal) after acute myocardial infarction. J. Am. Coll. Cardiol. 54, 22772286. Horwitz, E.M., and Dominici, M. (2008). How do mesenchymal stromal cells exert their therapeutic benet? Cytotherapy 10, 771774. Horwitz, E.M., Le Blanc, K., Dominici, M., Mueller, I., Slaper-Cortenbach, I., Marini, F.C., Deans, R.J., Krause, D.S., and Keating, A.; International Society for Cellular Therapy. (2005). Clarication of the nomenclature for MSC: The International Society for Cellular Therapy position statement. Cytotherapy 7, 393395. In t Anker, P.S., Scherjon, S.A., Kleijburg-van der Keur, C., de Groot-Swings, G.M., Claas, F.H., Fibbe, W.E., and Kanhai, H.H. (2004). Isolation of mesenchymal stem cells of fetal or maternal origin from human placenta. Stem Cells 22, 13381345. Johnson, K., Zhu, S., Tremblay, M.S., Payette, J.N., Wang, J., Bouchez, L.C., Meeusen, S., Althage, A., Cho, C.Y., Wu, X., and Schultz, P.G. (2012). A Stem Cell-Based Approach to Cartilage Repair. Science. Published online April 5, 2012. Jones, E.A., Kinsey, S.E., English, A., Jones, R.A., Straszynski, L., Meredith, D.M., Markham, A.F., Jack, A., Emery, P., and McGonagle, D. (2002). Isolation and characterization of bone marrow multipotential mesenchymal progenitor cells. Arthritis Rheum. 46, 33493360. Jones, S., Horwood, N., Cope, A., and Dazzi, F. (2007). The antiproliferative effect of mesenchymal stem cells is a fundamental property shared by all stromal cells. J. Immunol. 179, 28242831. Kawai, T., and Akira, S. (2011). Toll-like receptors and their crosstalk with other innate receptors in infection and immunity. Immunity 34, 637650. Lai, R.C., Chen, T.S., and Lim, S.K. (2011). Mesenchymal stem cell exosome: a novel stem cell-based therapy for cardiovascular disease. Regen. Med. 6, 481492. n, O. (2007). Immunomodulation by mesenchymal Le Blanc, K., and Ringde stem cells and clinical experience. J. Intern. Med. 262, 509525. Le Blanc, K., Frassoni, F., Ball, L., Locatelli, F., Roelofs, H., Lewis, I., Lanino, E., Sundberg, B., Bernardo, M.E., Remberger, M., et al; Developmental Committee of the European Group for Blood and Marrow Transplantation. (2008). Mesenchymal stem cells for treatment of steroid-resistant, severe, acute graft-versus-host disease: a phase II study. Lancet 371, 15791586. Lee, R.H., Pulin, A.A., Seo, M.J., Kota, D.J., Ylostalo, J., Larson, B.L., Semprun-Prieto, L., Delafontaine, P., and Prockop, D.J. (2009). Intravenous hMSCs improve myocardial infarction in mice because cells embolized in lung are activated to secrete the anti-inammatory protein TSG-6. Cell Stem Cell 5, 5463. Lee, C.C., Christensen, J.E., Yoder, M.C., and Tarantal, A.F. (2010). Clonal analysis and hierarchy of human bone marrow mesenchymal stem and progenitor cells. Exp. Hematol. 38, 4654. Lindner, U., Kramer, J., Rohwedel, J., and Schlenke, P. (2010). Mesenchymal Stem or Stromal Cells: Toward a Better Understanding of Their Biology? Transfus. Med. Hemother. 37, 7583. Published online Mar 15, 2010. `, L., Manuelli, C., Frosali, F., Mazzinghi, B., Liotta, F., Angeli, R., Cosmi, L., Fil Maggi, L., Pasini, A., Lisi, V., et al. (2008). Toll-like receptors 3 and 4 are expressed by human bone marrow-derived mesenchymal stem cells and can inhibit their T-cell modulatory activity by impairing Notch signaling. Stem Cells 26, 279289. ndez-Ferrer, S., Michurina, T.V., Ferraro, F., Mazloom, A.R., Macarthur, Me B.D., Lira, S.A., Scadden, D.T., Maayan, A., Enikolopov, G.N., and Frenette, P.S. (2010). Mesenchymal and haematopoietic stem cells form a unique bone marrow niche. Nature 466, 829834. Muraglia, A., Cancedda, R., and Quarto, R. (2000). Clonal mesenchymal progenitors from human bone marrow differentiate in vitro according to a hierarchical model. J. Cell Sci. 113, 11611166. Murry, C.E., Soonpaa, M.H., Reinecke, H., Nakajima, H., Nakajima, H.O., Rubart, M., Pasumarthi, K.B., Virag, J.I., Bartelmez, S.H., Poppa, V., et al. (2004). Haematopoietic stem cells do not transdifferentiate into cardiac myocytes in myocardial infarcts. Nature 428, 664668. Nadri, S., and Soleimani, M. (2007). Comparative analysis of mesenchymal stromal cells from murine bone marrow and amniotic uid. Cytotherapy 9, 729737. Nauta, A.J., and Fibbe, W.E. (2007). Immunomodulatory properties of mesenchymal stromal cells. Blood 110, 34993506. meth, K., Leelahavanichkul, A., Yuen, P.S.T., Mayer, B., Parmelee, A., Doi, Ne K., Robey, P.G., Leelahavanichkul, K., Koller, B.H., Brown, J.M., et al. (2009). Bone marrow stromal cells attenuate sepsis via prostaglandin E(2)-dependent reprogramming of host macrophages to increase their interleukin-10 production. Nat. Med. 15, 4249. Omatsu, Y., Sugiyama, T., Kohara, H., Kondoh, G., Fujii, N., Kohno, K., and Nagasawa, T. (2010). The essential functions of adipo-osteogenic progenitors as the hematopoietic stem and progenitor cell niche. Immunity 33, 387399. Park, D., Spencer, J.A., Koh, B.I., Kobayashi, T., Fujisaki, J., Clemens, T.L., Lin, C.P., Kronenberg, H.M., and Scadden, D.T. (2012). Endogenous bone marrow MSCs are dynamic, fate-restricted participants in bone maintenance and regeneration. Cell Stem Cell 10, 259272. Patki, S., Kadam, S., Chandra, V., and Bhonde, R. (2010). Human breast milk is a rich source of multipotent mesenchymal stem cells. Hum. Cell 23, 3540. Pittenger, M.F., Mackay, A.M., Beck, S.C., Jaiswal, R.K., Douglas, R., Mosca, J.D., Moorman, M.A., Simonetti, D.W., Craig, S., and Marshak, D.R. (1999). Multilineage potential of adult human mesenchymal stem cells. Science 284, 143147. Prockop, D.J., and Keating, A. (2012). Relearning the lessons of genomic stability of human cells during expansion in culture: Implications for clinical research. Stem Cells. Published online April 11, 2012. 10.1002/stem.1103. Ranganath, S.H., Levy, O., Inamdar, M.S., and Karp, J.M. (2012). Harnessing the mesenchymal stem cell secretome for the treatment of cardiovascular disease. Cell Stem Cell 10, 244258. Ren, G., Zhang, L., Zhao, X., Xu, G., Zhang, Y., Roberts, A.I., Zhao, R.C., and Shi, Y. (2008). Mesenchymal stem cell-mediated immunosuppression occurs via concerted action of chemokines and nitric oxide. Cell Stem Cell 2, 141150. Ren, G., Su, J., Zhang, L., Zhao, X., Ling, W., Lhuillie, A., Zhang, J., Lu, Y., Roberts, A.I., Ji, W., et al. (2009). Species variation in the mechanisms of mesenchymal stem cell-mediated immunosuppression. Stem Cells 27, 19541962. Rose, R.A., Jiang, H., Wang, X.H., Helke, S., Tsoporis, J.N., Gong, N., Keating, S.C., Parker, T.G., Backx, P.H., and Keating, A. (2008). Bone marrow-derived mesenchymal stromal cells express cardiac-specic markers, retain the stromal phenotype, and do not become functional cardiomyocytes in vitro. Stem Cells 26, 28842892. Sacchetti, B., Funari, A., Michienzi, S., Di Cesare, S., Piersanti, S., Saggio, I., Tagliaco, E., Ferrari, S., Robey, P.G., Riminucci, M., and Bianco, P. (2007). Self-renewing osteoprogenitors in bone marrow sinusoids can organize a hematopoietic microenvironment. Cell 131, 324336.

Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc. 715

Cell Stem Cell

Perspective
Sarugaser, R., Lickorish, D., Baksh, D., Hosseini, M.M., and Davies, J.E. (2005). Human umbilical cord perivascular (HUCPV) cells: a source of mesenchymal progenitors. Stem Cells 23, 220229. Sarugaser, R., Hanoun, L., Keating, A., Stanford, W.L., and Davies, J.E. (2009). Human mesenchymal stem cells self-renew and differentiate according to a deterministic hierarchy. PLoS ONE 4, e6498. , L., Tarte, K., Galipeau, J., Krampera, M., Martin, I., Phinney, D.G., Sensebe and Shi, Y.; MSC Committee of the International Society for Cellular Therapy. (2012). Limited acquisition of chromosomal aberrations in human adult mesenchymal stromal cells. Cell Stem Cell 10, 910, author reply 1011. Shih, D.T., Lee, D.C., Chen, S.C., Tsai, R.Y., Huang, C.T., Tsai, C.C., Shen, E.Y., and Chiu, W.T. (2005). Isolation and characterization of neurogenic mesenchymal stem cells in human scalp tissue. Stem Cells 23, 10121020. Sugiyama, T., Kohara, H., Noda, M., and Nagasawa, T. (2006). Maintenance of the hematopoietic stem cell pool by CXCL12-CXCR4 chemokine signaling in bone marrow stromal cell niches. Immunity 25, 977988. Tan, J., Wu, W., Xu, X., Liao, L., Zheng, F., Messinger, S., Sun, X., Chen, J., Yang, S., Cai, J., et al. (2012). Induction therapy with autologous mesenchymal stem cells in living-related kidney transplants: a randomized controlled trial. JAMA 307, 11691177. Tolar, J., LeBlanc, K., Keating, A., and Blazar, B.R. (2010). Hitting the right spot with mesenchymal stromal cells (MSCs). Stem Cells 28, 14461455. Tomchuck, S.L., Zwezdaryk, K.J., Coffelt, S.B., Waterman, R.S., Danka, E.S., and Scandurro, A.B. (2008). Toll-like receptors on human mesenchymal stem cells drive their migration and immunomodulating responses. Stem Cells 26, 99107. tz, B., Ehinger, M., Ditzel, N., Tormin, A., Li, O., Brune, J.C., Walsh, S., Schu Kassem, M., and Scheding, S. (2011). CD146 expression on primary nonhematopoietic bone marrow stem cells is correlated with in situ localization. Blood 117, 50675077. Uccelli, A., Moretta, L., and Pistoia, V. (2008). Mesenchymal stem cells in health and disease. Nat. Rev. Immunol. 8, 726736. Wang, H.S., Hung, S.C., Peng, S.T., Huang, C.C., Wei, H.M., Guo, Y.J., Fu, Y.S., Lai, M.C., and Chen, C.C. (2004). Mesenchymal stem cells in the Whartons jelly of the human umbilical cord. Stem Cells 22, 13301337. Waterman, R.S., Tomchuck, S.L., Henkle, S.L., and Betancourt, A.M. (2010). A new mesenchymal stem cell (MSC) paradigm: polarization into a pro-inammatory MSC1 or an Immunosuppressive MSC2 phenotype. PLoS ONE 5, e10088. Wight, T.N., Kinsella, M.G., Keating, A., and Singer, J.W. (1986). Proteoglycans in human long-term bone marrow cultures: biochemical and ultrastructural analyses. Blood 67, 13331343. Yang, S.H., Park, M.J., Yoon, I.H., Kim, S.Y., Hong, S.H., Shin, J.Y., Nam, H.Y., Kim, Y.H., Kim, B., and Park, C.G. (2009). Soluble mediators from mesenchymal stem cells suppress T cell proliferation by inducing IL-10. Exp. Mol. Med. 41, 315324. Yoshimura, K., Shigeura, T., Matsumoto, D., Sato, T., Takaki, Y., Aiba-Kojima, E., Sato, K., Inoue, K., Nagase, T., Koshima, I., and Gonda, K. (2006). Characterization of freshly isolated and cultured cells derived from the fatty and uid portions of liposuction aspirates. J. Cell. Physiol. 208, 6476. Zuk, P.A., Zhu, M., Ashjian, P., De Ugarte, D.A., Huang, J.I., Mizuno, H., Alfonso, Z.C., Fraser, J.K., Benhaim, P., and Hedrick, M.H. (2002). Human adipose tissue is a source of multipotent stem cells. Mol. Biol. Cell 13, 42794295.

716 Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc.

Cell Stem Cell

Perspective
Induced Pluripotent Stem Cells: Past, Present, and Future
Shinya Yamanaka1,2,*
for iPS Cell Research and Application, Kyoto University, Kyoto 606-8507, Japan Institute of Cardiovascular Disease, San Francisco, CA 94158, USA *Correspondence: yamanaka@cira.kyoto-u.ac.jp DOI 10.1016/j.stem.2012.05.005
2Gladstone 1Center

The development of iPSCs reected the merging of three major scientic streams and has in turn led to additional new branches of investigation. However, there is still debate about whether iPSCs are functionally equivalent to ESCs. This question should be answered only by science, not by politics or business.
Introduction In 2006, we showed that stem cells with properties similar to ESCs could be generated from mouse broblasts by simultaneously introducing four genes (Takahashi and Yamanaka, 2006). We designated these cells iPSCs. In 2007, we reported that a similar approach was applicable for human broblasts and that by introducing a handful of factors, human iPSCs can be generated (Takahashi et al., 2007). On the same day, James Thomsons group also reported the generation of human iPSC using a different combination of factors (Yu et al., 2007). The Merging of Three Scientic Streams Led to the Production of iPSCs Like any other scientic advance, iPSC technology was established on the basis of numerous ndings by past and current scientists in related elds. There were three major streams of research that led us to the production of iPSCs (Figure 1). The rst stream was reprogramming by nuclear transfer. In 1962, John Gurdon reported that his laboratory had generated tadpoles from unfertilized eggs that had received a nucleus from the intestinal cells of adult frogs (Gurdon, 1962). More than three decades later, Ian Wilmut and colleagues reported the birth of Dolly, the rst mammal generated by somatic cloning of mammary epithelial cells (Wilmut et al., 1997). These successes in somatic cloning demonstrated that even differentiated cells contain all of the genetic information that is required for the development of entire organisms, and that oocytes contain factors that can reprogram somatic cell nuclei. In 2001, Takashi Tadas group showed that ESCs also contain factors that can reprogram somatic cells (Tada et al., 2001). The second stream was the discovery of master transcription factors. In 1987, a Drosophila transcription factor, Antennapedia, was shown to induce the formation of legs instead of antennae when ectopically expressed (Schneuwly et al., 1987). In the same year, a mammalian transcription factor, MyoD, was shown to convert broblasts into myocytes (Davis et al., 1987). These results led to the concept of a master regulator, a transcription factor that determines and induces the fate of a given lineage. Many researchers began to search for single master regulators for various lineages. The attempts failed, with a few exceptions (Yamanaka and Blau, 2010). The third, and equally important, stream of research is that involving ESCs. Since the rst generation of mouse ESCs in
678 Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc.

1981 (Evans and Kaufman, 1981; Martin, 1981), Austin Smith and others have established culture conditions that enable the long-term maintenance of pluripotency (Smith et al., 1988). A key factor for maintenance of mouse ESCs was leukemia inhibitory factor (LIF). Likewise, since the rst generation of human ESCs (Thomson et al., 1998), optimal culture conditions with basic broblast growth factor (bFGF) have been established. Combining the rst two streams of research led us to hypothesize that it is a combination of multiple factors in oocytes or ESCs that reprogram somatic cells back into the embryonic state and to design experiments to identify that combination. Using information about the culture conditions that are needed to culture pluripotent cells, we were then able to identify four factors that can generate iPSCs. Maturation and Understanding of iPSC Technology Soon after our initial report of mouse iPSCs, other groups recapitulated the factor-based reprogramming both in mice (Maherali et al., 2007; Wernig et al., 2007) and humans (Lowry et al., 2008; Park et al., 2008b). One of the advantages of iPSC technology is its simplicity and reproducibility. Many laboratories began to explore the underlying mechanisms and to modify the procedures. Although iPSCs can be generated reproducibly, the efciency of the process remains low: typically less than 1% of transfected broblasts become iPSCs. This low efciency initially raised the possibility that iPSCs are derived from rare stem or undifferentiated cells coexisting in broblast cultures (Yamanaka, 2009a). Subsequent studies showed, however, that iPSCs can be derived from terminally differentiated lymphocytes (Loh et al., 2009) and postmitotic neurons (Kim et al., 2011a). Thus, most, if not all, somatic cells have a potential to become iPSCs, albeit with different efciencies. How then can just a small set of factors induce reprogramming of somatic cells? It is beyond the scope of this essay to provide an overview of the many studies that have addressed this important question. From my perspective, the consensus of many scientists seems to be that the reprogramming factors initiate the reprogramming process in many more than 1% of transfected cells but that the process is not completed in most of the cells. Poorly understood stochastic events seem to be required for full reprogramming to take place (Hanna et al., 2009; Yamanaka, 2009a). As I discuss below, culture conditions

Cell Stem Cell

Perspective

Figure 1. Three Scientic Streams that Led to the Development of iPSCs

Figure 2. New Scientic Streams that Emerged from the Development of iPSCs

seem to function as a driving force that can help promote full reprogramming. Initially iPSCs were generated using either retroviruses or lentiviruses, which might cause insertional mutagenesis and thus would pose a risk for translational application and could perhaps even lead to adverse effects like those seen in some attempts at gene therapy (Hacein-Bey-Abina et al., 2003). Mice derived from retrovirally derived iPSCs are apparently normal, as long as expression of the c-Myc transgene is repressed (Aoi et al., 2008; Nakagawa et al., 2008). However, the long-term safety of human iPSCs cannot be guaranteed through mouse studies alone. In addition, retroviruses may make iPSCs immunogenic (Zhao et al., 2011). Thus, for the purpose of cell transplantation therapy, we will need to avoid induction methods that involve vector integration into the host genome. Many ways to generate integration-free iPSCs have been reported. These methods include plasmid (Okita et al., 2011a; Okita et al., 2008), Sendai virus (Fusaki et al., 2009), adenovirus (Stadtfeld et al., 2008), synthesized RNAs (Warren et al., 2010), and proteins (Kim et al., 2009). In addition, attempts have been made to induce reprogramming by small molecules. Among these, plasmids and Sendai viruses are now routinely used in many laboratories. In the Center for iPS Cell Research and Application, Kyoto University, our favored methods are to use episomal plasmids for regenerative medicine and either retroviruses or episomal plasmids for in vitro studies. We prefer these methods because of their simplicity and reproducibility. Scientists are now largely shifting their efforts from technology development per se to applications. New Scientic Streams Have Emerged from iPSC Technology Streams in science never cease (Figure 2). After the seminal work in mice by Rudolf Jaenischs laboratory (Hanna et al., 2007), scientists are now making progress toward using iPSCs in regenerative medicine, for example for the treatment of Parkinsons disease (Kriks et al., 2011), platelet deciency (Takayama et al., 2010), spinal cord injury (Nori et al., 2011; Tsuji et al., 2010), and macular degeneration (Okamoto and Takahashi, 2011). Patient-derived iPSCs have been shown to be useful for

modeling diseases and screening drug candidate libraries. Starting with the seminal studies by the groups led by George Daley (Park et al., 2008a), and Kevin Eggan (Dimos et al., 2008), more than 100 reports published in the past three years use diseasespecic iPSCs. I was surprised that patient-specic iPSCs can be used to recapitulate phenotypes of not only monogenic diseases but also late-onset polygenic diseases, such as Parkinsons disease (Devine et al., 2011), Alzheimers disease (Israel et al., 2012; Yagi et al., 2011; Yahata et al., 2011), and schizophrenia (Brennand et al., 2011). Excitement surrounds the potential for application of these cells to both analysis of disease mechanisms and investigation of potential new treatments. Somatic cells derived from iPSCs, particularly cardiac myocytes and hepatocytes, could also be used for toxicology testing as an alternative to existing approaches (Yamanaka, 2009b). In addition, to these medical applications, iPSCs can be used in animal biotechnology. Monkey (Liu et al., 2008), porcine (West et al., 2010), and canine (Shimada et al., 2010) iPSCs can be used for genetic engineering in these animals, allowing for the generation of disease models and the production in larger animals of useful substances, such as enzymes, that are decient in patients with genetic diseases. The technology might potentially be useful in the future for preserving endangered animals as well (Ben-Nun et al., 2011), although many challenges would need to be overcome. One of the most striking applications of iPSCs was reported by Nakauchi and colleagues, who generated a rat pancreas in a mouse, by microinjecting rat iPSCs into mouse blastocysts decient in a gene essential for pancreas development (Kobayashi et al., 2010). In the future, it might become possible to generate organs for human transplantation using a similar strategy. Another scientic stream that emerged from iPSC technology is direct reprogramming from one somatic lineage to another. As mentioned above, attempts to identify a single master transcription factor have failed for most somatic lineages. However, in light of the success of iPSC reprogramming, scientists switched from searching for a single factor to looking for a combination. Melton and colleagues reported the conversion of exocrine cells to endocrine cells in the mouse pancreas by using a combination of three transcription factors (Zhou et al., 2008).
Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc. 679

Cell Stem Cell

Perspective
Table 1. Number of ESC and iPSC Clones Analyzed in Published Studies Conclusion about the Relationship between ESCs and iPSCs It is difcult to distinguish between them Clone Numbers First Author A.M. Newman M.G. Guenther C. Bock There are notable differences M. Chin C.M. Marchetto J. Deng Z. Ghosh A. Doi Y. Ohi K. Kim R. Lister Year 2010 2010 2011 2009 2009 2009 2010 2011 2011 2011 2011 ESC 23 36 20 3 2 3 6 3 3 6 2 iPSC 68 54 12 5 2 4 4 9 9 12 5 Figure 3. Overlapping Variations Present in iPSC and ESC Clones
Measurement of a range of properties of iPSCs and ESCs, including gene expression, DNA methylation, differentiation propensity, and (for mouse cells) complementation activity in embryos has led to the realization that the properties of both ESC and iPSC lines vary. However, as analysis of signicant numbers of clones from multiple laboratories has accumulated, it has become clear that there is considerable overlap in terms of the properties of ESC and iPSC lines and, at a general level, these two cell types are difcult to distinguish.

Their seminal work was soon followed by many in vitro examples of converting broblasts to various somatic cells, such as neural cells (Vierbuchen et al., 2010), hepatocytes (Huang et al., 2011), cardiac myocytes (Ieda et al., 2010), and hematopoietic progenitor cells (Szabo et al., 2010). Direct reprogramming is straightforward and rapid. One hurdle that remains is how to obtain a sufcient amount of target cells for downstream applications. The best usage of this new technology may be in situ direct reprogramming (Qian et al., 2012). The Big Question: Are iPSCs Different from ESCs? One of the most important questions regarding iPSCs is whether they are different from ESCs and, if so, whether any differences that do exist are functionally relevant. During the rst few years of our studies of iPSCs, we were amazed by their remarkable similarity to ESCs. Starting in 2009, however, scientists started reporting differences between iPSCs and ESCs. For example, Chin et al. (2009) compared three human ESC lines and ve iPSC lines by expression microarrays and identied hundreds of genes that were differentially expressed (Chin et al., 2009). They concluded that iPSCs should be considered a unique subtype of pluripotent cells. Two other studies also compared the global gene expression between ESCs and iPSCs and identied persistent donor cell gene expression in iPSCs (Ghosh et al., 2010; Marchetto et al., 2009). It was Deng et al. (2009) who rst reported that there were differences in DNA methylation between the two types of pluripotent stem cell lines after they performed the targeted bisulte sequencing of three human ESC clones and four iPSCs lines. Doi et al. (2009) also reported that there were differentially methylated genes, such as BMP3, between ESCs and iPSCs. Subsequently, three studies reported epigenetic memories of donor cells in human induced pluripotent cells (Kim et al., 2011b; Lister et al., 2011; Ohi et al., 2011). However, other studies have concluded that it is difcult to distinguish iPSCs from ESCs by gene expression or DNA methylation. Two reports showed that both iPSC clones and ESC clones have overlapping variations in gene expression and thus that the two types of pluripotent stem cells are clustered
680 Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc.

together by these analyses (Guenther et al., 2010; Newman and Cooper, 2010). They argued that these variations are, at least in part, derived from the different induction and culture conditions used by each laboratory. Bock et al. (2011) demonstrated that iPSCs and ESCs are very similar in their gene expression and DNA methylation and that some iPSC clones cannot be distinguished from ESCs. By examining how many iPS and ES clones were compared, we observed a clear tendency (Table 1). Studies that reported differences either in gene expression or DNA methylation compared relatively small numbers (generally fewer than 10) for each group), whereas those that found it difcult to distinguish iPSCs from ESCs analyzed many more clones, and clones from multiple laboratories. Another major point of discussion has been the ability of the cells to differentiate and whether iPSCs are functionally different from ESCs in this respect. Hu et al. (2010) performed in vitro directed neural differentiation of ve human ESC clones and 12 iPSCs clones. They showed that all of the ESC clones differentiated into Pax6 positive cells, with more than 90% efcacy, but the iPSC clones showed poorer differentiation, with 10% to 50% efcacy. However, Boulting et al. (2011) examined 16 human iPSC clones for their ability to differentiate into motor neurons and found that 13 of these iPSC clones differentiated with comparable efcacies to ESCs. So, again, there are conicting conclusions regarding the similarity between iPSCs and ESCs. Taken together, these studies showed that iPSC clones and ESC clones have overlapping degrees of variation (Figure 3). It should be noted that variations among ESC clones have been well documented (Osafune et al., 2008; Ward et al., 2004). Although it is possible that iPSC clones show greater variation, and that some clones differ from ESCs in their gene expression, DNA methylation, or differentiation ability (Miura et al., 2009), it appears that at least some iPSC clones are indistinguishable from ESC clones.

Cell Stem Cell

Perspective
It is interesting to consider what brings about such variation between the iPSC clones. We learned an important lesson from two related reports regarding mouse iPSCs (Carey et al., 2011; Stadtfeld et al., 2010). These two studies, conducted in the Hochedlinger lab and the Jaenisch lab, used very similar secondary induction systems to generate mouse iPSCs. However, the properties of the iPSC clones were very different between the two laboratories. Most of the iPSC clones generated in Hochedlinger lab could not be successfully used to generate germline competent chimeras by microinjection or all-iPSC mice by tetraploid complementation, the most stringent criterion to evaluate pluripotency. They showed that the loss of imprinting of the Dlk1-Dio3 gene cluster predicts these poor iPSC abilities. In sharp contrast, most of the iPSC clones generated in the Jaenisch laboratory had normal imprinting of the Dl1-Dios3 cluster and were capable of generating high quality chimeras and viable all-iPSC mice. The only notable difference between the two laboratories methods was the order of the reprogramming factors in the expression cassettes, and this difference resulted in higher expression levels of Oct4 and Klf4 in the cells generated by the Jaenisch laboratory. By increasing the expression of Oct4 and Klf4 (Carey et al., 2011), or by supplementing with ascorbic acid (Stadtfeld et al., 2012), the quality of the iPSCs generated by an otherwise very similar method was enhanced. Thus, the level and stoichiometry of the reprogramming factors, as well as culture conditions, during iPSC generation can contribute signicantly to the variation seen in the epigenetic state and pluripotent potential of the resulting iPSCs. These data demonstrated that incomplete or imperfect reprogramming is not a fundamental problem associated with iPSCs. Instead, differences in the quality of iPSC clones seem to be largely due to technical variables, such as the factor combinations, gene delivery methods, and culture conditions. In addition, some variation between iPSC clones can be attributed to stochastic events during reprogramming, which cannot be controlled. Thus, evaluation and selection will be essential for identifying iPSC clones that are suitable for medical applications. Is There a Dark Side to Induced Pluripotency? Several reports have suggested that, in addition to variation in gene expression, DNA methylation, and pluripotent potential, there are other potential abnormalities in iPSCs, including somatic mutations (Gore et al., 2011), copy number variations (Hussein et al., 2011), and immunogenicity (Zhao et al., 2011). In some of these reports, the negative aspects of iPSCs were, in my opinion, overstated. The media overreacted, as did accompanying scientic commentaries with alarming words in their titles, such as dark side, under attack, aw, troublesome, and growing pains (Apostolou and Hochedlinger, 2011; Dolgin, 2011; Hayden, 2011; Pera, 2011; Zwaka, 2010). However, despite these doomsday headlines, subsequent analyses have indicated that many of the genetic differences found in iPSCs seem to have pre-existed in the original somatic cells, and therefore arose independently of the reprogramming process itself (Cheng et al., 2012; Young et al., 2012). Reprogramming to form iPSCs is inherently clonal, and therefore variations that exist at a low frequency within the starting cell population can become more apparent when analyzing individual clones derived from it and comparing them to the parental cell population as a whole. Another study showed that a set of iPSC clones that are capable of generating all-iPSC mice have very few genetic alterations relative to their parental cells (Quinlan et al., 2011). The chimeric and progeny mice derived from iPSCs that are devoid of the Myc transgene appear to be normal, indicating that these iPSCs do not contain detrimental genetic alterations that have a negative impact on function (Nakagawa et al., 2008; Nakagawa et al., 2010). With regard to immunogenicity, it is not clear whether the reported weak immune reaction to transgene-free iPSCs is signicant (Okita et al., 2011b) because the most prominent study that reported the immunogenicity of the cells examined undifferentiated iPSCs (Zhao et al., 2011), which will never be used in cell transplantation therapy. We have to understand all of these results and consider them in context to have a balanced view of iPSCs. Why Are ESCs and iPSCs So Remarkably Similar? Although there may be some differences between iPSCs and ESCs, they are, nevertheless, remarkably similar. If anything, we should perhaps be wondering why iPSCs and ESCs are in fact so similar despite their different origins and generation methods. No other examples of this level of similarity between man-made cells and naturally-existing cells exist. Several types of somatic cells, such as neural cells and cardiac myocytes, have been generated from ESCs/iPSCs or directly from broblasts. These man-made somatic cells have some of the characteristics of their normal counterparts that exist in vivo, but they are still very different from natural neural cells and cardiac myocytes. The similarity between ESCs and iPSCs is therefore in many ways exceptional. One potential explanation is that ESCs are in fact also manmade. It is possible that ESCs do not exist under physiological conditions and instead are selected and established by cultivating the cells of the inner cell mass (ICM) under specic culture conditions. ESCs are different from the majority of cells in the ICM in many respects. For example, although cells in the ICM possess a low degree of global DNA methylation (Reik et al., 2001), ESCs have a higher level of methylation (Li et al., 1992). A Ras family gene, ERas, is highly expressed in mouse ESCs but not in embryos (Takahashi et al., 2003). ESCs also have longer telomeres than are seen in embryos (Varela et al., 2011). Thus, we may be discussing the relationship between two types of man-made cells rather than between man-made cells and naturally existing cells. Through many researchers efforts, the eld has established culture conditions that enable the generation and long-term maintenance of both mouse and human ESCs. It is likely that these culture conditions select for cells with certain properties, and this selection would also contribute to making ESCs and iPSCs appear as similar as they do. Concluding Thoughts If we accept the idea that ESCs and iPSCs are both articial cells types generated in the laboratory, we move on to another important question: do ESCs truly represent an ultimate control or gold standard for iPSCs? I think the answer is probably no. Instead, future studies should focus on the capacity of iPSCs themselves
Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc. 681

Cell Stem Cell

Perspective
to form new tissues, organs, and model organisms, as a stream that exists in parallel to that of ESCs as a branch of the same overall experimental river. I believe that iPSC technology is now ready for many applications, including stem cell therapies. From each induction procedure, multiple iPSC clones of various qualities emerged (Figure 3). It is thus essential to select good clones for medical applications. We may be able to narrow down candidates for good clones by marker gene expressions. However, we have to conrm in vitro differentiation propensities and genome and epigenome integrities. For wide-spread use, it might be necessary to establish in advance stocks of qualied iPSC clones from healthy volunteers or from cord-blood stocks. Immunorejection could be decreased by generating iPSCs from HLA homozygous donors (Okita et al., 2011a). iPSC technology will likely have a substantial impact not only on science but also on business and politics. However, iPSCs should be evaluated based strictly on the scientic data, and all such data should be thoroughly considered for its relevance to potential clinical applications of the cells. Scientists should focus on research, and politicians and businesses should rely on the hard evidence generated from scientic studies to inform future directions rather than on the opinions of those who do not fully understand the eld.
ACKNOWLEDGMENTS S.Y. would like to thank Mari Ohnuki and Michiyo Koyanagi, and all the members of his laboratories in Kyoto and San Francisco for scientic discussion and supports. S.Y. is a member without salary of the scientic advisory boards of iPierian, iPS Academia Japan, and Megakaryon Corporation. REFERENCES Aoi, T., Yae, K., Nakagawa, M., Ichisaka, T., Okita, K., Takahashi, K., Chiba, T., and Yamanaka, S. (2008). Generation of pluripotent stem cells from adult mouse liver and stomach cells. Science 321, 699702. Apostolou, E., and Hochedlinger, K. (2011). Stem cells: iPS cells under attack. Nature 474, 165166. Ben-Nun, I.F., Montague, S.C., Houck, M.L., Tran, H.T., Garitaonandia, I., Leonardo, T.R., Wang, Y.C., Charter, S.J., Laurent, L.C., Ryder, O.A., and Loring, J.F. (2011). Induced pluripotent stem cells from highly endangered species. Nat. Methods 8, 829831. Bock, C., Kiskinis, E., Verstappen, G., Gu, H., Boulting, G., Smith, Z.D., Ziller, M., Croft, G.F., Amoroso, M.W., Oakley, D.H., et al. (2011). Reference Maps of human ES and iPS cell variation enable high-throughput characterization of pluripotent cell lines. Cell 144, 439452. Boulting, G.L., Kiskinis, E., Croft, G.F., Amoroso, M.W., Oakley, D.H., Wainger, B.J., Williams, D.J., Kahler, D.J., Yamaki, M., Davidow, L., et al. (2011). A functionally characterized test set of human induced pluripotent stem cells. Nat. Biotechnol. 29, 279286. Brennand, K.J., Simone, A., Jou, J., Gelboin-Burkhart, C., Tran, N., Sangar, S., Li, Y., Mu, Y., Chen, G., Yu, D., et al. (2011). Modelling schizophrenia using human induced pluripotent stem cells. Nature 473, 221225. Carey, B.W., Markoulaki, S., Hanna, J.H., Faddah, D.A., Buganim, Y., Kim, J., Ganz, K., Steine, E.J., Cassady, J.P., Creyghton, M.P., et al. (2011). Reprogramming factor stoichiometry inuences the epigenetic state and biological properties of induced pluripotent stem cells. Cell Stem Cell 9, 588598. Cheng, L., Hansen, N.F., Zhao, L., Du, Y., Zou, C., Donovan, F.X., Chou, B.K., Zhou, G., Li, S., Dowey, S.N., et al; NISC Comparative Sequencing Program. (2012). Low incidence of DNA sequence variation in human induced pluripotent stem cells generated by nonintegrating plasmid expression. Cell Stem Cell 10, 337344. Chin, M.H., Mason, M.J., Xie, W., Volinia, S., Singer, M., Peterson, C., Ambartsumyan, G., Aimiuwu, O., Richter, L., Zhang, J., et al. (2009). Induced pluripotent stem cells and embryonic stem cells are distinguished by gene expression signatures. Cell Stem Cell 5, 111123. Davis, R.L., Weintraub, H., and Lassar, A.B. (1987). Expression of a single transfected cDNA converts broblasts to myoblasts. Cell 51, 9871000. Deng, J., Shoemaker, R., Xie, B., Gore, A., LeProust, E.M., Antosiewicz-Bourget, J., Egli, D., Maherali, N., Park, I.H., Yu, J., et al. (2009). Targeted bisulte sequencing reveals changes in DNA methylation associated with nuclear reprogramming. Nat. Biotechnol. 27, 353360. Devine, M.J., Ryten, M., Vodicka, P., Thomson, A.J., Burdon, T., Houlden, H., Cavaleri, F., Nagano, M., Drummond, N.J., Taanman, J.W., et al. (2011). Parkinsons disease induced pluripotent stem cells with triplication of the a-synuclein locus. Nat. Commun. 2, 440. Dimos, J.T., Rodolfa, K.T., Niakan, K.K., Weisenthal, L.M., Mitsumoto, H., Chung, W., Croft, G.F., Saphier, G., Leibel, R., Goland, R., et al. (2008). Induced pluripotent stem cells generated from patients with ALS can be differentiated into motor neurons. Science 321, 12181221. Doi, A., Park, I.H., Wen, B., Murakami, P., Aryee, M.J., Irizarry, R., Herb, B., Ladd-Acosta, C., Rho, J., Loewer, S., et al. (2009). Differential methylation of tissue- and cancer-specic CpG island shores distinguishes human induced pluripotent stem cells, embryonic stem cells and broblasts. Nat. Genet. 41, 13501353. Dolgin, E. (2011). Flaw in induced-stem-cell model. Nature 470, 13. Evans, M.J., and Kaufman, M.H. (1981). Establishment in culture of pluripotential cells from mouse embryos. Nature 292, 154156. Fusaki, N., Ban, H., Nishiyama, A., Saeki, K., and Hasegawa, M. (2009). Efcient induction of transgene-free human pluripotent stem cells using a vector based on Sendai virus, an RNA virus that does not integrate into the host genome. Proc. Jpn. Acad., Ser. B, Phys. Biol. Sci. 85, 348362. Ghosh, Z., Wilson, K.D., Wu, Y., Hu, S., Quertermous, T., and Wu, J.C. (2010). Persistent donor cell gene expression among human induced pluripotent stem cells contributes to differences with human embryonic stem cells. PLoS ONE 5, e8975. Gore, A., Li, Z., Fung, H.L., Young, J.E., Agarwal, S., Antosiewicz-Bourget, J., Canto, I., Giorgetti, A., Israel, M.A., Kiskinis, E., et al. (2011). Somatic coding mutations in human induced pluripotent stem cells. Nature 471, 6367. Guenther, M.G., Frampton, G.M., Soldner, F., Hockemeyer, D., Mitalipova, M., Jaenisch, R., and Young, R.A. (2010). Chromatin structure and gene expression programs of human embryonic and induced pluripotent stem cells. Cell Stem Cell 7, 249257. Gurdon, J.B. (1962). The developmental capacity of nuclei taken from intestinal epithelium cells of feeding tadpoles. J. Embryol. Exp. Morphol. 10, 622640. Hacein-Bey-Abina, S., Von Kalle, C., Schmidt, M., McCormack, M.P., Wulffraat, N., Leboulch, P., Lim, A., Osborne, C.S., Pawliuk, R., Morillon, E., et al. (2003). LMO2-associated clonal T cell proliferation in two patients after gene therapy for SCID-X1. Science 302, 415419. Hanna, J., Saha, K., Pando, B., van Zon, J., Lengner, C.J., Creyghton, M.P., van Oudenaarden, A., and Jaenisch, R. (2009). Direct cell reprogramming is a stochastic process amenable to acceleration. Nature 462, 595601. Hanna, J., Wernig, M., Markoulaki, S., Sun, C.W., Meissner, A., Cassady, J.P., Beard, C., Brambrink, T., Wu, L.C., Townes, T.M., and Jaenisch, R. (2007). Treatment of sickle cell anemia mouse model with iPS cells generated from autologous skin. Science 318, 19201923. Hayden, E.C. (2011). Stem cells: The growing pains of pluripotency. Nature 473, 272274. Hu, B.Y., Weick, J.P., Yu, J., Ma, L.X., Zhang, X.Q., Thomson, J.A., and Zhang, S.C. (2010). Neural differentiation of human induced pluripotent stem cells follows developmental principles but with variable potency. Proc. Natl. Acad. Sci. USA 107, 43354340. Huang, P., He, Z., Ji, S., Sun, H., Xiang, D., Liu, C., Hu, Y., Wang, X., and Hui, L. (2011). Induction of functional hepatocyte-like cells from mouse broblasts by dened factors. Nature 475, 386389.

682 Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc.

Cell Stem Cell

Perspective
rva , E., Ng, Hussein, S.M., Batada, N.N., Vuoristo, S., Ching, R.W., Autio, R., Na ma la inen, R., Olsson, C., et al. (2011). Copy number variaS., Sourour, M., Ha tion and selection during reprogramming to pluripotency. Nature 471, 5862. Ieda, M., Fu, J.D., Delgado-Olguin, P., Vedantham, V., Hayashi, Y., Bruneau, B.G., and Srivastava, D. (2010). Direct reprogramming of broblasts into functional cardiomyocytes by dened factors. Cell 142, 375386. Israel, M.A., Yuan, S.H., Bardy, C., Reyna, S.M., Mu, Y., Herrera, C., Hefferan, M.P., Van Gorp, S., Nazor, K.L., Boscolo, F.S., et al. (2012). Probing sporadic and familial Alzheimers disease using induced pluripotent stem cells. Nature 482, 216220. Kim, D., Kim, C.H., Moon, J.I., Chung, Y.G., Chang, M.Y., Han, B.S., Ko, S., Yang, E., Cha, K.Y., Lanza, R., and Kim, K.S. (2009). Generation of human induced pluripotent stem cells by direct delivery of reprogramming proteins. Cell Stem Cell 4, 472476. Kim, J., Lengner, C.J., Kirak, O., Hanna, J., Cassady, J.P., Lodato, M.A., Wu, S., Faddah, D.A., Steine, E.J., Gao, Q., et al. (2011a). Reprogramming of postnatal neurons into induced pluripotent stem cells by dened factors. Stem Cells 29, 9921000. Kim, K., Zhao, R., Doi, A., Ng, K., Unternaehrer, J., Cahan, P., Huo, H., Loh, Y.H., Aryee, M.J., Lensch, M.W., et al. (2011b). Donor cell type can inuence the epigenome and differentiation potential of human induced pluripotent stem cells. Nat. Biotechnol. 29, 11171119. Kobayashi, T., Yamaguchi, T., Hamanaka, S., Kato-Itoh, M., Yamazaki, Y., Ibata, M., Sato, H., Lee, Y.S., Usui, J., Knisely, A.S., et al. (2010). Generation of rat pancreas in mouse by interspecic blastocyst injection of pluripotent stem cells. Cell 142, 787799. Kriks, S., Shim, J.W., Piao, J., Ganat, Y.M., Wakeman, D.R., Xie, Z., CarrilloReid, L., Auyeung, G., Antonacci, C., Buch, A., et al. (2011). Dopamine neurons derived from human ES cells efciently engraft in animal models of Parkinsons disease. Nature 480, 547551. Li, E., Bestor, T.H., and Jaenisch, R. (1992). Targeted mutation of the DNA methyltransferase gene results in embryonic lethality. Cell 69, 915926. Lister, R., Pelizzola, M., Kida, Y.S., Hawkins, R.D., Nery, J.R., Hon, G., Antosiewicz-Bourget, J., OMalley, R., Castanon, R., Klugman, S., et al. (2011). Hotspots of aberrant epigenomic reprogramming in human induced pluripotent stem cells. Nature 471, 6873. Liu, H., Zhu, F., Yong, J., Zhang, P., Hou, P., Li, H., Jiang, W., Cai, J., Liu, M., Cui, K., et al. (2008). Generation of induced pluripotent stem cells from adult rhesus monkey broblasts. Cell Stem Cell 3, 587590. Loh, Y.H., Agarwal, S., Park, I.H., Urbach, A., Huo, H., Heffner, G.C., Kim, K., Miller, J.D., Ng, K., and Daley, G.Q. (2009). Generation of induced pluripotent stem cells from human blood. Blood 113, 54765479. Lowry, W.E., Richter, L., Yachechko, R., Pyle, A.D., Tchieu, J., Sridharan, R., Clark, A.T., and Plath, K. (2008). Generation of human induced pluripotent stem cells from dermal broblasts. Proc. Natl. Acad. Sci. USA 105, 28832888. Maherali, N., Sridharan, R., Xie, W., Utikal, J., Eminli, S., Arnold, K., Stadtfeld, M., Yachechko, R., Tchieu, J., Jaenisch, R., et al. (2007). Directly reprogrammed broblasts show global epigenetic remodeling and widespread tissue contribution. Cell Stem Cell 1, 5570. Marchetto, M.C., Yeo, G.W., Kainohana, O., Marsala, M., Gage, F.H., and Muotri, A.R. (2009). Transcriptional signature and memory retention of human-induced pluripotent stem cells. PLoS ONE 4, e7076. Martin, G.R. (1981). Isolation of a pluripotent cell line from early mouse embryos cultured in medium conditioned by teratocarcinoma stem cells. Proc. Natl. Acad. Sci. USA 78, 76347638. Miura, K., Okada, Y., Aoi, T., Okada, A., Takahashi, K., Okita, K., Nakagawa, M., Koyanagi, M., Tanabe, K., Ohnuki, M., et al. (2009). Variation in the safety of induced pluripotent stem cell lines. Nat. Biotechnol. 27, 743745. Nakagawa, M., Koyanagi, M., Tanabe, K., Takahashi, K., Ichisaka, T., Aoi, T., Okita, K., Mochiduki, Y., Takizawa, N., and Yamanaka, S. (2008). Generation of induced pluripotent stem cells without Myc from mouse and human broblasts. Nat. Biotechnol. 26, 101106. Nakagawa, M., Takizawa, N., Narita, M., Ichisaka, T., and Yamanaka, S. (2010). Promotion of direct reprogramming by transformation-decient Myc. Proc. Natl. Acad. Sci. USA 107, 1415214157. Newman, A.M., and Cooper, J.B. (2010). Lab-specic gene expression signatures in pluripotent stem cells. Cell Stem Cell 7, 258262. Nori, S., Okada, Y., Yasuda, A., Tsuji, O., Takahashi, Y., Kobayashi, Y., Fujiyoshi, K., Koike, M., Uchiyama, Y., Ikeda, E., et al. (2011). Grafted humaninduced pluripotent stem-cell-derived neurospheres promote motor functional recovery after spinal cord injury in mice. Proc. Natl. Acad. Sci. USA 108, 1682516830. Ohi, Y., Qin, H., Hong, C., Blouin, L., Polo, J.M., Guo, T., Qi, Z., Downey, S.L., Manos, P.D., Rossi, D.J., et al. (2011). Incomplete DNA methylation underlies a transcriptional memory of somatic cells in human iPS cells. Nat. Cell Biol. 13, 541549. Okamoto, S., and Takahashi, M. (2011). Induction of retinal pigment epithelial cells from monkey iPS cells. Invest. Ophthalmol. Vis. Sci. 52, 87858790. Okita, K., Matsumura, Y., Sato, Y., Okada, A., Morizane, A., Okamoto, S., Hong, H., Nakagawa, M., Tanabe, K., Tezuka, K., et al. (2011a). A more efcient method to generate integration-free human iPS cells. Nat. Methods 8, 409412. Okita, K., Nagata, N., and Yamanaka, S. (2011b). Immunogenicity of induced pluripotent stem cells. Circ. Res. 109, 720721. Okita, K., Nakagawa, M., Hyenjong, H., Ichisaka, T., and Yamanaka, S. (2008). Generation of mouse induced pluripotent stem cells without viral vectors. Science 322, 949953. Osafune, K., Caron, L., Borowiak, M., Martinez, R.J., Fitz-Gerald, C.S., Sato, Y., Cowan, C.A., Chien, K.R., and Melton, D.A. (2008). Marked differences in differentiation propensity among human embryonic stem cell lines. Nat. Biotechnol. 26, 313315. Park, I.H., Arora, N., Huo, H., Maherali, N., Ahfeldt, T., Shimamura, A., Lensch, M.W., Cowan, C., Hochedlinger, K., and Daley, G.Q. (2008a). Disease-specic induced pluripotent stem cells. Cell 134, 877886. Park, I.H., Zhao, R., West, J.A., Yabuuchi, A., Huo, H., Ince, T.A., Lerou, P.H., Lensch, M.W., and Daley, G.Q. (2008b). Reprogramming of human somatic cells to pluripotency with dened factors. Nature 451, 141146. Pera, M.F. (2011). Stem cells: The dark side of induced pluripotency. Nature 471, 4647. Qian, L., Huang, Y., Spencer, C.I., Foley, A., Vedantham, V., Liu, L., Conway, S.J., Fu, J.D., and Srivastava, D. (2012). In vivo reprogramming of murine cardiac broblasts into induced cardiomyocytes. Nature. Published online April 18, 2012. Quinlan, A.R., Boland, M.J., Leibowitz, M.L., Shumilina, S., Pehrson, S.M., Baldwin, K.K., and Hall, I.M. (2011). Genome sequencing of mouse induced pluripotent stem cells reveals retroelement stability and infrequent DNA rearrangement during reprogramming. Cell Stem Cell 9, 366373. Reik, W., Dean, W., and Walter, J. (2001). Epigenetic reprogramming in mammalian development. Science 293, 10891093. Schneuwly, S., Klemenz, R., and Gehring, W.J. (1987). Redesigning the body plan of Drosophila by ectopic expression of the homoeotic gene Antennapedia. Nature 325, 816818. Shimada, H., Nakada, A., Hashimoto, Y., Shigeno, K., Shionoya, Y., and Nakamura, T. (2010). Generation of canine induced pluripotent stem cells by retroviral transduction and chemical inhibitors. Mol. Reprod. Dev. 77, 2. Smith, A.G., Heath, J.K., Donaldson, D.D., Wong, G.G., Moreau, J., Stahl, M., and Rogers, D. (1988). Inhibition of pluripotential embryonic stem cell differentiation by puried polypeptides. Nature 336, 688690. Stadtfeld, M., Apostolou, E., Akutsu, H., Fukuda, A., Follett, P., Natesan, S., Kono, T., Shioda, T., and Hochedlinger, K. (2010). Aberrant silencing of imprinted genes on chromosome 12qF1 in mouse induced pluripotent stem cells. Nature 465, 175181. Stadtfeld, M., Apostolou, E., Ferrari, F., Choi, J., Walsh, R.M., Chen, T., Ooi, S.S., Kim, S.Y., Bestor, T.H., Shioda, T., et al. (2012). Ascorbic acid prevents loss of Dlk1-Dio3 imprinting and facilitates generation of all-iPS cell mice from terminally differentiated B cells. Nat. Genet. 44, 398405. Stadtfeld, M., Nagaya, M., Utikal, J., Weir, G., and Hochedlinger, K. (2008). Induced pluripotent stem cells generated without viral integration. Science 322, 945949.

Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc. 683

Cell Stem Cell

Perspective
o, R.M., Schnerch, A., Mitchell, R., FiebigSzabo, E., Rampalli, S., Risuen Comyn, A., Levadoux-Martin, M., and Bhatia, M. (2010). Direct conversion of human broblasts to multilineage blood progenitors. Nature 468, 521526. Tada, M., Takahama, Y., Abe, K., Nakatsuji, N., and Tada, T. (2001). Nuclear reprogramming of somatic cells by in vitro hybridization with ES cells. Curr. Biol. 11, 15531558. Takahashi, K., Mitsui, K., and Yamanaka, S. (2003). Role of ERas in promoting tumour-like properties in mouse embryonic stem cells. Nature 423, 541545. Takahashi, K., Tanabe, K., Ohnuki, M., Narita, M., Ichisaka, T., Tomoda, K., and Yamanaka, S. (2007). Induction of pluripotent stem cells from adult human broblasts by dened factors. Cell 131, 861872. Takahashi, K., and Yamanaka, S. (2006). Induction of pluripotent stem cells from mouse embryonic and adult broblast cultures by dened factors. Cell 126, 663676. Takayama, N., Nishimura, S., Nakamura, S., Shimizu, T., Ohnishi, R., Endo, H., Yamaguchi, T., Otsu, M., Nishimura, K., Nakanishi, M., et al. (2010). Transient activation of c-MYC expression is critical for efcient platelet generation from human induced pluripotent stem cells. J. Exp. Med. 207, 28172830. Thomson, J.A., Itskovitz-Eldor, J., Shapiro, S.S., Waknitz, M.A., Swiergiel, J.J., Marshall, V.S., and Jones, J.M. (1998). Embryonic stem cell lines derived from human blastocysts. Science 282, 11451147. Tsuji, O., Miura, K., Okada, Y., Fujiyoshi, K., Mukaino, M., Nagoshi, N., Kitamura, K., Kumagai, G., Nishino, M., Tomisato, S., et al. (2010). Therapeutic potential of appropriately evaluated safe-induced pluripotent stem cells for spinal cord injury. Proc. Natl. Acad. Sci. USA 107, 1270412709. Varela, E., Schneider, R.P., Ortega, S., and Blasco, M.A. (2011). Different telomere-length dynamics at the inner cell mass versus established embryonic stem (ES) cells. Proc. Natl. Acad. Sci. USA 108, 1520715212. dhof, T.C., and Vierbuchen, T., Ostermeier, A., Pang, Z.P., Kokubu, Y., Su Wernig, M. (2010). Direct conversion of broblasts to functional neurons by dened factors. Nature 463, 10351041. Ward, C.M., Barrow, K.M., and Stern, P.L. (2004). Signicant variations in differentiation properties between independent mouse ES cell lines cultured under dened conditions. Exp. Cell Res. 293, 229238. Warren, L., Manos, P.D., Ahfeldt, T., Loh, Y.H., Li, H., Lau, F., Ebina, W., Mandal, P.K., Smith, Z.D., Meissner, A., et al. (2010). Highly efcient reprogramming to pluripotency and directed differentiation of human cells with synthetic modied mRNA. Cell Stem Cell 7, 618630. Wernig, M., Meissner, A., Foreman, R., Brambrink, T., Ku, M., Hochedlinger, K., Bernstein, B.E., and Jaenisch, R. (2007). In vitro reprogramming of broblasts into a pluripotent ES-cell-like state. Nature 448, 318324. West, F.D., Terlouw, S.L., Kwon, D.J., Mumaw, J.L., Dhara, S.K., Hasneen, K., Dobrinsky, J.R., and Stice, S.L. (2010). Porcine induced pluripotent stem cells produce chimeric offspring. Stem Cells Dev. 19, 12111220. Wilmut, I., Schnieke, A.E., McWhir, J., Kind, A.J., and Campbell, K.H. (1997). Viable offspring derived from fetal and adult mammalian cells. Nature 385, 810813. Yagi, T., Ito, D., Okada, Y., Akamatsu, W., Nihei, Y., Yoshizaki, T., Yamanaka, S., Okano, H., and Suzuki, N. (2011). Modeling familial Alzheimers disease with induced pluripotent stem cells. Hum. Mol. Genet. 20, 45304539. Yahata, N., Asai, M., Kitaoka, S., Takahashi, K., Asaka, I., Hioki, H., Kaneko, T., Maruyama, K., Saido, T.C., Nakahata, T., et al. (2011). Anti-Ab drug screening platform using human iPS cell-derived neurons for the treatment of Alzheimers disease. PLoS ONE 6, e25788. Yamanaka, S. (2009a). Elite and stochastic models for induced pluripotent stem cell generation. Nature 460, 4952. Yamanaka, S. (2009b). A fresh look at iPS cells. Cell 137, 1317. Yamanaka, S., and Blau, H.M. (2010). Nuclear reprogramming to a pluripotent state by three approaches. Nature 465, 704712. Young, M.A., Larson, D.E., Sun, C.W., George, D.R., Ding, L., Miller, C.A., Lin, L., Pawlik, K.M., Chen, K., Fan, X., et al. (2012). Background mutations in parental cells account for most of the genetic heterogeneity of Induced Pluripotent Stem Cells. Cell Stem Cell, in press. Yu, J., Vodyanik, M.A., Smuga-Otto, K., Antosiewicz-Bourget, J., Frane, J.L., Tian, S., Nie, J., Jonsdottir, G.A., Ruotti, V., Stewart, R., et al. (2007). Induced pluripotent stem cell lines derived from human somatic cells. Science 318, 19171920. Zhao, T., Zhang, Z.N., Rong, Z., and Xu, Y. (2011). Immunogenicity of induced pluripotent stem cells. Nature 474, 212215. Zhou, Q., Brown, J., Kanarek, A., Rajagopal, J., and Melton, D.A. (2008). In vivo reprogramming of adult pancreatic exocrine cells to beta-cells. Nature 455, 627632. Zwaka, T.P. (2010). Stem cells: Troublesome memories. Nature 467, 280281.

684 Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc.

Cell Stem Cell

Perspective
Hematopoietic Stem Cell Heterogeneity Takes Center Stage
Michael R. Copley,1 Philip A. Beer,1 and Connie J. Eaves1,*
1Terry Fox Laboratory, British Columbia Cancer Agency, Vancouver, V5Z 1L3 BC, Canada *Correspondence: ceaves@bccrc.ca DOI 10.1016/j.stem.2012.05.006

Over the past 10 years, increasing evidence has accumulated that heterogeneity is a feature of hematopoietic stem cell (HSC) proliferation, self-renewal, and differentiation based on examination of these properties at a clonal level. The heterogeneous behavior of HSCs reects the operation of a complex interplay of intrinsic and extrinsic variables. In this review, we discuss key ndings from the last 5 years that reveal new insights into the mechanisms involved.
The hematopoietic system has long served as an important model for delineating both tissue-specic and more generalized mechanisms that balance cell loss and replacement. Fifty years ago, the introduction of clonal transplant assays established the existence of multipotent hematopoietic cells and laid the conceptual foundation of a hematopoietic stem cell (HSC) compartment. In the past 10 years, a multitude of pathways and extrinsic factors that play critical roles in regulating HSC numbers in vivo have been identied. More recently, evidence of intrinsic and extrinsic mechanisms regulating the very properties that dene HSCs have emerged. These ndings introduce a new level of complexity into understanding how HSC behavior is controlled. They thus raise many new challenges, but also suggest new possibilities for obtaining previously obscured insights into diseases caused by perturbed HSC function and into mechanisms of relevance to development, aging, and regenerative medicine. Background Many tissues depend on mechanisms that balance the loss and renewal of their specialized cellular elements. Early microscopy of solid tissues with high rates of cell turnover (such as the skin and gut) suggested that this balance is achieved through a multistep, unidirectional developmental process in which mature end cells are ultimately generated from an initial population of morphologically undifferentiated precursor (stem) cells. Such morphological studies led to the concept of tissue-specic stem cells and intermediate transit-amplifying cells, but were not well suited to investigations of mechanisms that regulate their numbers and differentiation behavior. Not surprisingly, therefore, it was studies of the hematopoietic system, where the precise location of the stem cells remains controversial, that led to a new, functional approach for identifying stem cells based on an analysis of the types of progeny they produce. This functional approach to hematopoietic stem cell (HSC) identication was launched by the discovery that the regenerative potential of these cells can be activated upon their transplantation into a host whose blood cell output has been compromised, e.g., by a lethal dose of irradiation (Ford et al., 1956; Jacobson et al., 1951). The fact that single HSCs can produce millions of mature blood cells led to the rst efforts to use clonal analyses for their detection (Becker et al., 1963; Till and McCulloch,
690 Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc.

1961). The rst such cells were called colony-forming units spleen (CFU-S) based on their ability to produce macroscopically visible colonies in the spleen in a 12 week period, but were subsequently found to contain very few cells with longer-term hematopoietic repopulating activity (Jones et al., 1996; Ploemacher and Brons, 1989). Nevertheless, many of the results obtained from analyzing the behavior of CFU-S have remained relevant to understanding more primitive subsets of hematopoietic cells. A key observation that rapidly emerged from the study of CFU-S was the high degree of variability in the numbers and types of daughter cells produced in these transplant assays (Siminovitch et al., 1963; Till et al., 1964). This nding was later strengthened by analyses of multilineage colonies generated in vitro, where external stimuli can be kept the same (Humphries et al., 1979, 1981; Ogawa et al., 1983), and further reinforced by experiments that used retroviral marking to track clones generated in vivo over more prolonged periods (Dick et al., 1985; Lemischka et al., 1986). These ndings stimulated subsequent efforts to elucidate both the local external regulators (niche control) of HSC behavior and the internal regulators of their various fate choices (i.e., their viability, proliferative state, selfrenewal probability, and differentiation lineage options). In the past few decades, most experiments aimed at understanding mechanisms regulating HSC functionality have been designed and the results interpreted based on the assumption that HSCs can be quantied and characterized using a singular measurement of their continuing production at a clonal level of lymphoid and myeloid cells for at least 16 weeks in suitably compromised hosts (Bryder et al., 2006). This assumption is based on the concept that the self-renewal potential and differentiation (lineage) potential of HSCs are largely controlled by a common network of intrinsically specied mechanisms that establish a common ground state in these cells. This idea is not to be confused with the extent to which these potentialities are expressed, because these are known to be subject to microenvironmental cues. Recent work has now thrown new light onto these issues and suggests a greater level of heterogeneity in HSC populations than previously envisaged. Additionally, alterations in HSC properties occur during ontogeny and aging. In this review, we highlight work from the past 5 years that has contributed to this growing theme of functional heterogeneity in HSC biology.

Cell Stem Cell

Perspective
Figure 1. Lymphopoietic Differences between HSC Subtypes Are Manifested at Multiple Levels of Lymphoid Differentiation
HSCs with durable self-renewal consist of 2 subtypes (a-HSCs and b-HSCs). These have equivalent robust myeloid differentiation ability but they differ in their ability to produce mature lymphoid cells as shown. The lymphoid deciency of a-HSCs results from a multistep reduction in lymphoid cell output (Benz et al., 2012).

HSC Differentiation Programs A major technological development in the last decade was the introduction of methods for purifying mouse HSCs with longterm repopulating activity (at least 16 weeks) to near homogeneity (Kiel et al., 2005; Osawa et al., 1996). This approach made it feasible to examine the in vivo clonal outputs of these HSCs with the certainty that the activity being observed could be attributed to one as opposed to several cellsa caveat inherent in limiting dilution approaches. The mature cell outputs of hundreds of single HSC transplants have now been analyzed, and for many of these, serial transplants have also been performed to determine the stability of the primary patterns observed. The results have cemented the controversial idea that differences in HSC behavior are due to their possession of different self-renewal and differentiation properties, as rst ller-Sieburg and coworkers using proposed by studies of Mu ller-Sieburg et al., 2002). limiting dilution HSC transplants (Mu Different HSC subtypes have since been dened using a variety of criteria to distinguish their mature blood cell outputs. ller-Sieburg group has identied different HSC subsets The Mu as myeloid-biased (My-bi), lymphoid-biased (Ly-bi), or balanced (Bala), based on a measurement of the predominant lineages within the total output of donor-derived blood cells (i.e., exclusive of the recipients contribution to, or the total number of, ller-Sieburg et al., 2012). We have mature blood cells) (Mu distinguished the different differentiation behaviors of individual HSCs by measuring the relative contributions of individual HSC outputs to the total circulating myeloid and lymphoid cell numbers (i.e., those derived from both donor and recipient HSCs). This different method distinguishes HSC subtypes as lymphoid-decient (a), balanced (b), or myeloid-decient (g and d) and both a- and b-HSCs are found to contribute similarly to the circulating pool of myeloid cells (Dykstra et al., 2007). Serial transplantation experiments have been important in establishing the concept that the different clonal patterns of differentiation represent intrinsically stable HSC programs, as opposed to temporally or stochastically variable states. Such experiments have shown that only a- and b-HSCs can be serially transplanted, and their unique differentiation behaviors are usually stably maintained over periods of years in primary transplant recipients (Benz et al., 2012; Dykstra et al., 2007). Likewise, the initial pattern of differentiation by which they are dened is frequently expressed by their clonally generated daughter HSCs when these are transplanted into secondary mice (Benz

lleret al., 2012; Dykstra et al., 2007, 2011; Mu Sieburg et al., 2002, 2012). However, it is also apparent that the original HSC program is not always maintained with complete delity as demonstrated by the generation of secondary b clones from primary a clones, and vice versa (Benz et al., 2012). The mechanisms determining the lineage outputs that distinguish different HSC subtypes are complex. For example, HSCs with different lymphoid outputs appear to produce lymphoid precursors that display differences in their IL-7 responsiveness (Sieburg et al., 2006). In addition, our group (Benz et al., 2012) has found that the reduced production of lymphoid cells that is characteristic of a-HSCs is due to a deciency that operates at multiple levels as their progeny attempt to move down the lymphoid pathway (Figure 1). At present, methods to prospectively separate different HSC subtypes are not sufciently discriminating to enable their differences to be examined at a molecular level. However, microuidics-based analyses of the transcriptomes of individual cells in a highly puried HSC population have suggested that additional subsets may exist based on their nonrandom pattern of expression of certain genes (Glotzbach et al., 2011). In addition, several other potential clues have been uncovered. One is the evidence that TGF-b may inuence HSC subtypes differently (Challen et al., 2010). A second is the nding that the microenvironment of the bone marrow is a preferential site of enrichment for a- versus b-HSCs (Benz et al., 2012). A third is the recent observation that Btaf, a gene found to regulate DNA damage responses, may differentially inuence the maintenance of the a- versus b-HSC states (Wang et al., 2012). Taken together, these ndings suggest that differences will be found in the molecular machinery that determines the differentiation potential of individual HSCs, and that such differences will be maintained in the clonally derived daughter HSCs produced in consecutive self-renewal divisions. The impressive stability of the differentiation behavior that these cells and their progeny display is consistent with an underlying epigenetic mechanism, and the recent observation of perturbations in both human and mouse blood cell lineage outputs associated with alterations in the DNA methylation machinery provides further support for ske et al., 2009; Challen et al., 2012; Hodges this idea (Bro et al., 2011; Ji et al., 2010; Trowbridge et al., 2009). Nevertheless, the inherently articial nature of the single-cell transplantation conditions used to characterize the potentialities of individual HSCs necessarily limits extrapolating such data to understanding their activity during normal development, growth, and aging. In addition, it is still difcult to devise an enrichment protocol that gives the purities of HSCs required to make
Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc. 691

Cell Stem Cell

Perspective
single-cell transplants of HSCs practical while ensuring all HSCs have been captured. Cellular barcoding using viruses to introduce unique sequences into individual HSCs (Gerrits et al., 2010) offers an important new clonal tracking strategy and also has the potential to circumvent limitations of single-cell approaches. This technology uses lentiviral gene transfer to uniquely label the genomes of individual HSCs, and hence all of their progeny. Its application to studies of HSC outputs has already enabled these to be assessed in single mice transplanted with large numbers of highly puried and individually barcoded HSCs (Lu et al., 2011). Interestingly, although much greater competitive inuences are expected in this situation compared to previous single HSC transplantation experiments, the patterns of lineage outputs were similar in both cases. These ndings are signicant because they demonstrate that the mechanisms responsible for the heterogeneous differentiation behavior of individual HSCs are not strongly inuenced by the number of HSC clones active in a given (irradiated) mouse. Cellular barcoding, like the single HSC transplantation strategy, also has limitations. Because the barcode is delivered by a viral vector, the HSCs need to be cultured for several hours to achieve a desirable gene transfer efciency, which necessarily alters the properties of the HSCs. In addition, the shorter the transduction protocol, the less likely the most primitive HSC subsets will be transduced. Insertion of the viral sequence can also be mutagenic or at least have functional consequences that may affect the biology of the cell transduced and/or its progeny (Cavazzana-Calvo and Fischer, 2007; Cavazzana-Calvo et al., 2010). Other caveats include the dependence on statistical assumptions in deriving clone size information from sequencing barcoded DNA. Analysis of viral integration sites in engrafted patients from gene therapy trials has served as an important source of evidence that long-term clones produced in humans typically produce both myeloid and lymphoid cells (Cavazzana-Calvo et al., 2011). However, because these data are derived from transplanted patients who have severe immune deciency syndromes, they might have been inuenced by the fact that the regenerated cells were produced in a setting with a bias toward the detection of donor-derived lymphoid cells. Of interest, in this regard, is a recent b-thalassemia patient who was transplanted with genetically corrected autologous cells and subsequently found to be reconstituted with a long-lived lymphoid-decient clone (Cavazzana-Calvo et al., 2010). Although alternate interpretations exist, this nding could suggest the existence of lymphoid-decient HSC subtypes in humans. In summary, the clones produced by individual HSCs in irradiated mice display a large range of different lymphoid and myeloid cell outputs. However, this heterogeneity in differentiation behavior does not appear to be the exclusive result of random uctuations in the expression of a multilineage program. Rather, the heterogeneity seems to reect the operation of an intrinsically determined control mechanism that establishes a unique program in each HSC. Although little is known about this mechanism, it is clear that it can be perpetuated through many HSC self-renewal divisions with relatively high, but not rigidly xed, delity, and that it targets multiple downstream stages of the lymphoid commitment and differentiation process.
692 Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc.

HSC Proliferative State Regulation In adult mice and humans, most of the HSCs are in a proliferatively inactive (quiescent, G0) state. The most recent studies of the HSC compartment in adult mice indicate that it is composed of two kinetically distinct subpopulations, with turnover rates of 5 weeks and 21 weeks (termed activated and dormant HSC, respectively) (Wilson et al., 2008). This conclusion contrasts with earlier reports that suggested turnover rates of 45 weeks (Bradford et al., 1997; Cheshier et al., 1999), although these latter rates were based on analyses of a phenotypically dened compartment that we now understand is predominantly composed of cells that do not have durable self-renewal activity. Similar information is not available for human HSCs, but a prolonged HSC turnover rate (40 weeks) has also been suggested (Catlin et al., 2011). These observations raise a number of questions regarding the differences between activated and dormant HSCs, including their mechanisms of regulation, whether they are hierarchically structured or correlate with different lineage programs, and how their numbers change during ontogeny and aging. TGF-b is well known for its ability to arrest the cycling of primitive hematopoietic cells, but evidence that this contributes to the homeostatic regulation of HSCs in vivo has been inconclusive. In a recent series of experiments, freshly isolated HSCs were found to display active TGF-b signaling (Yamazaki et al., 2009), indicating a role for this pathway in vivo. Further corroboration comes from studies in which HSCs unable to express a TGF-b receptor demonstrated impaired long-term hematopoiesis and heightened proliferative activity in transplanted recipients (Yamazaki et al., 2011). The same group also identied nonmyelinating Schwann cells as the cells responsible for the activation of TGF-b in the bone marrow, thereby uncovering a potential novel role for this cell type as an upstream regulator of TGFb-mediated control of HSC quiescence (Yamazaki et al., 2011). The signaling components responsible for HSC responses to both proliferative and antiproliferative signals have recently been further elucidated. Once of these is FOXO3A, which has been shown to act as an HSC dormancy factor via its ability to regulate p27 and p57 (Miyamoto et al., 2007). This effect is likely facilitated by a p27/p57-mediated blockade of the nuclear transport of HSC70/cyclin D1, because a knockout of both of these genes leads to increased cycling and an associated nuclear import of this complex (Zou et al., 2011). Changes in the levels of reactive oxygen species in HSCs and alterations in oxygensensing molecules also appear to be associated with changes in HSC turnover rates (Miyamoto et al., 2007; Takubo et al., 2010). However, it remains unclear whether these are a cause or consequence of changes in HSC cycling. Finally, the tumor suppressor Lkb1 was recently found to be important for the maintenance of HSC quiescence because its conditional deletion leads to loss of quiescence followed by rapid HSC depletion and pancytopenia (Nakada et al., 2010). HSC Self-Renewal Activity The concept of distinct subsets of repopulating cells dates back several decades (Hodgson and Bradley, 1984), and the fact that they can be prospectively isolated as phenotypically separable subsets has been recognized for over a decade (Bryder et al., 2006). However, evidence of different subsets of HSCs with

Cell Stem Cell

Perspective
Figure 2. Differences between Fetal and Adult Mouse HSCs
Certain key properties of HSCs in fetal and adult tissues differ. These affect their self-renewal activity, their cycling control, and their generation of mature myeloid cells. In addition, the production of B cells from fetal HSCs can proceed in the absence of IL-7 and thymic stromal-derived lymphopoietin (TSLP), whereas, in the adult, IL-7 is an absolute requirement.

more prolonged self-renewal abilities has only recently been obtained (Benveniste et al., 2010; Dykstra et al., 2007, 2011; Iscove and Nawa, 1997; Morita et al., 2010; Sieburg et al., 2011). This evidence is based on two types of measurements. One is the lifetime of individual HSC-derived clones of mature blood cells produced in transplanted mice, which are observed to range from a minimum of 4 months (by denition) to >2 years. The second more stringent endpoint is the detection of clonally perpetuated blood cell production that is sustained through at least three serial transplants; i.e., beyond the lifetime of a normal mouse. The prolonged maintenance of daughter HSC activity for at least 6 months following secondary transplantation appears to be associated with the maintenance of robust myeloid differentiation activity that is shared by a- and b-HSCs. In contrast, in HSCs that have more time-restricted myeloid outputs (intermediate HSCs) and/or have become largely restricted to producing lymphoid cells (g and d subtypes), self-renewal activity is not usually maintained beyond 4 months after initial transplantation (Dykstra et al., 2007). Importantly, mouse HSCs with more limited self-renewal activities can be prospectively distinguished from those that exhibit durable activity, the latter being most enriched within a subpopulation that does not express CD49b and shows the highest expression of CD150 (encoded by Slamf1) (Benveniste et al., 2010; Kent et al., 2009; Kiel et al., 2005; Morita et al., 2010). Steel factor (SF, also known as Kit ligand or Stem Cell Factor [SCF]), has long been known to play a key role in the control and maintenance of primitive hematopoietic cells (McCulloch et al., 1964). New insights obtained in the last 5 years include the demonstration that SF can rapidly modulate self-renewal decisions of mouse HSCs, even when other variables, including cell death and cell cycle progression, are kept constant (Kent et al., 2008). The cellular origins of SF important for HSC regulation in vivo have also recently been claried. This was achieved by generating different types of transgenic mice in which the Kitl gene was deleted in a cell-type-specic manner, followed by analysis of HSC activity. These experiments showed that SF produced by endothelial and perivascular stromal cells, but not from osteoblasts, is necessary for mouse HSC regeneration after

transplant (Ding et al., 2012). These ndings reinforce previous studies indicating a role for external signals in regulating the extent to which adult HSCs express their innate self-renewal potential in vivo (Iscove and Nawa, 1997). In addition, cellintrinsic mechanisms appear to instruct distinct heritable self-renewal probabilities in different HSCs, as suggested by similarities in the longevity of daughter clones generated from the same clonally expanded primary HSCs (Sieburg et al., 2011). Over the last 2 decades, the many transcription factors, chromatin modiers, and cell cycle regulators that have been found to affect HSC self-renewal behavior suggest that a number of different pathways regulate this process. The recently demonstrated role of epigenetic regulators in affecting HSC selfrenewal adds another layer of complexity (Challen et al., 2012). However, whether these factors inuence self-renewal independently of cell survival or proliferation, and how their effects are ultimately integrated, remains largely obscure. In summary, our current understanding of mouse HSC selfrenewal control suggests that the probability of undergoing a self-renewal division and the durability of this behavior is an intrinsically determined phenomenon that can be modied by the presence or absence of extrinsic factors. Additionally, the durability of this activity appears to be tightly linked to the maintenance of an active myeloid differentiation program. HSC Property Changes during Development and Aging It has been known for decades that primitive hematopoietic cells within the developing fetus show a much higher turnover than adult bone marrow HSCs (Becker et al., 1965; Fleming et al., 1993; Morrison et al., 1995). A pronounced difference between the proliferative activity of fetal and adult HSCs is now appreciated to be one of several properties of HSCs that change during development. Others include HSC self-renewal activity in transplanted adult recipients (Bowie et al., 2007b; Pawliuk et al., 1996), the growth-factor dependence of their self-renewal activity (Bowie et al., 2007a; Ikuta and Weissman, 1992), the numbers and types of mature cells they produce (Bowie et al., 2006; Dorshkind and Montecino-Rodriguez, 2007; Ikuta et al., 1990; Micklem et al., 1972), and the growth-factor requirements of their lymphoid-restricted progeny (Kikuchi and Kondo, 2006) (Figure 2). These distinguishing features seem to also be associated with unique programs because several genesBmi1 (Park et al., 2003), G1 (Hock et al., 2004a), and Tel/Etv6 (Hock et al., 2004b)are intrinsically required for the in vivo generation and maintenance of adult, but not fetal, HSCs. An important new element has been the revelation that several of these properties change abruptly and in a highly coordinated
Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc. 693

Cell Stem Cell

Perspective
fashion within a short period 14 weeks after birth in mice. The properties of HSCs that are known to be targeted by this developmental switch include their cycling status, their speed of expansion after transplantation, and their granulopoietic activity (Bowie et al., 2006, 2007b). In addition, the cytokine requirements of their lymphopoietic progeny are altered at this time (Kikuchi and Kondo, 2006). Remarkably, when fetal HSCs are transplanted into adult mice, their progeny HSC acquire adult-like properties (in terms of their granulopoietic output and posttransplant rate of expansion in secondary hosts) in the same time frame as they would have done if left unperturbed in the developing mouse. Taken together, these results suggest the operation of an intrinsically determined and timed process responsible for regulating the transition from a fetal to an adult HSC program. Changes in the representation of the HSC subtypes (as dened by their lineage outputs: see discussion of a, b, and g and d, above) in different tissues also occur during ontogeny. This change, however, does not share a similar timing with the changes that alter HSC proliferative activity and self-renewal control (Benz et al., 2012). In fact, no change in HSC subtype distribution is seen between 3 and 4 weeks of age in the mouse; i.e., the interval during which other HSC properties coordinately switch from a fetal to an adult program. However, a change in HSC subtype distribution does occur, but much earlier, starting in the late embryo and correlating with the relocation of HSCs from the liver to the bone marrow. These ndings suggest that the mechanism regulating the changes in the subtype composition of the HSC pool involves extrinsic factors emanating from the bone marrow microenvironment and is uncoupled from the mechanism that regulates the fetal-to-adult switch in cycling and self-renewal control, which appears to be intrinsically regulated (Benz et al., 2012). Evidence for a similar fetal-to-adult switch in humans was initially implied by the marked decrease in the rate of granulocyte telomere shortening seen in children between the age of 1 and 2 years, suggesting a change in the cycling status of the HSC compartment (Rufer et al., 1999). One difculty of these studies is the wide variability in normal granulocyte telomere length, which makes rm conclusions in humans difcult to draw. Subsequently, however, longitudinal measurements of telomere length in the blood cells of baboons have shown similar results to those from the human studies, further corroborating the idea that primates also show an abrupt change in HSC cycling during the early period of their development (Baerlocher et al., 2007). A possible regulatory component of the fetal-to-adult HSC switch is the transcription factor SOX17. The Sox17 gene was recently found to be expressed at a higher level in fetal HSCs than that in adult HSCs, as determined by both direct measurements of gene expression and an analysis of HSCs isolated from a Sox17 knockin reporter mouse (Kim et al., 2007). Inducible deletion of the Sox17 gene was then used to demonstrate an essential role for SOX17 in maintaining HSCs in fetal/neonatal mice. Importantly, Sox17 deletion in adult mice had no effect on HSC numbers or functionality (Kim et al., 2007). These ndings provided the rst example of a gene product required specically for the survival of fetal, but not adult, HSCs, and therefore evidence for a fetal-specic HSC transcriptional program. This idea is further supported by the demonstration
694 Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc.

that Ezh2 is similarly required for the maintenance of fetal, but not adult, HSCs (Mochizuki-Kashio et al., 2011). Interestingly, forced expression of Sox17 in adult murine hematopoietic cells was found to enhance the HSC-derived output of myeloid cells (a fetal HSC property) (He et al., 2011), consistent with a partial reactivation of the fetal program. Another recently identied potential regulator of fetal versus adult HSC identity is Lin28b. Lin28b is expressed at higher levels in fetal lymphoid progenitor populations than those seen in adult populations, and forced overexpression of Lin28 in adult hematopoietic progenitors reactivates a fetal-like lymphoid differentiation program (Yuan et al., 2012). These authors also found that Lin28b is differentially expressed in fetal and adult HSC-containing populations (LinSca-1+c-Kit+ cells), suggesting that LIN28B may also regulate differences between their developmental states. Since LIN28 acts as a regulator of developmental timing ssing in C. elegans by inhibiting let-7 microRNA processing (Bu et al., 2008), this pathway may serve as a conserved regulator of developmental changes in cell properties as animals transition from a fetal to an adult stage. Recent studies have also identied changes affecting the HSC compartment during the aging process. One of these relates to the decreased level of lymphoid cells (and hence increased proportion of myeloid cells) seen in the peripheral blood of older mice and humans (Berkahn and Keating, 2004). This is known to be associated with a concomitant decrease in committed lymphoid progenitor numbers (Miller and Allman, 2003; Min et al., 2006; Rossi et al., 2005), but a gradual ascendency within the HSC compartment of lymphoid-decient HSCs has also been reported (Beerman et al., 2010; Benz et al., 2012; Cho et al., 2008). In addition, it has been shown that serial transplantation of young adult HSCs in young adult recipients selectively expands/produces lymphoid-decient HSCs, demonstrating that these changes do not depend on an aging microenvironment (Dykstra et al., 2011). Thus, multiple HSC-related mechanisms of how aging may lead to a decline in lymphoid cells are likely relevant (Figure 3). These mechanisms may include a gradual erosion of lymphopoietic activity across a homogeneous population of HSCs (Model 1), a gradual dominance of a lymphoid-decient HSC subtype (Model 2), or a conversion of lymphoid-competent to lymphoid-decient HSC subtypes (Model 3). Evidence of an age-related increase in lymphoid-decient HSCs in older people (Park et al., 2003) suggests that the ndings in mice will likely have human correlates. The gradual dominance of the murine HSC population by lymphoid-decient HSCs has been postulated to reect a proliferative or selfrenewal advantage of these cells (Beerman et al., 2010; Cho et al., 2008). Another possible mechanism would be the recently demonstrated ability of lymphoid-competent HSCs to generate lymphoid-decient HSC progeny (Benz et al., 2012). Accumulation of DNA damage within HSCs has been proposed as a mechanism that contributes to age-associated decreases in their functionality. Although there has been little experimental data to support or refute this hypothesis, a recent study found that mice defective in mediators of DNA-damage regulators do not manifest changes in the size or composition of their HSC pool under homeostatic conditions. However, when transplanted, their HSCs did show increased apoptosis, reduced reconstituting activity, and decreased self-renewal

Cell Stem Cell

Perspective
Some of these behaviors appear to be coordinately regulated, whereas others have been clearly shown to be temporally separable. Interestingly, most involve an integration of both intrinsically established states and variable cues from the environment. Current reliance on transplantation strategies to identify HSCs, although powerful and highly informative, may be less useful for elucidating processes that control blood cell production under physiological conditions. In addition, recent ndings are largely based on studies of HSCs in mice or other model organisms and thus still await validation in humans. Advances in human HSC purication (Majeti et al., 2007; Notta et al., 2011) have extended earlier evidence of multiple human hematopoietic cell types with different self-renewal properties, and an increasing panoply of immunodecient xenograft hosts (Doulatov et al., 2012) can be expected to facilitate their further characterization in the near future. With the ongoing explosion of new technologies to analyze single HSC responses, their molecular composition, and their location in situ, the future of HSC biology looks both challenging and rosy for those prepared to let go of old dogma and search for new mechanisms.
Figure 3. Theoretical Models of HSC-Mediated Mechanisms that May Contribute to Aging-Associated Decreases in the Output of Mature Lymphoid Cells
In Model 1, hematopoiesis is sustained by the progeny of a homogeneous population of HSCs (red) that undergo a gradual erosion of their intrinsic lymphoid differentiation ability over time (blue). In Model 2, aging is associated with a clonal expansion of lymphoid-decient HSCs (a-subtype, pale blue) with a concomitant depletion of lymphoid-competent HSCs (b-subtype, pink). In Model 3, there is a gradual ascendancy of a-HSCs with aging as a result of their derivation from b-HSCs. Current experimental data indicate that all of these models could be operative but are also insufcient to discriminate their relative importance. ACKNOWLEDGMENTS We thank C. Benz and M. Kardel for helpful comments. This review was prepared with support from grants funded by the Terry Fox Run, the Stem Cell Network, and the Canadian Institutes of Health Research (CIHR). M.R.C. holds a Vanier Scholarship from CIHR and a Studentship from the Michael Smith Foundation for Health Research. P.A.B. holds a Kay Kendall Leukemia Fund Intermediate Fellowship from the UK. REFERENCES Baerlocher, G.M., Rice, K., Vulto, I., and Lansdorp, P.M. (2007). Longitudinal data on telomere length in leukocytes from newborn baboons support a marked drop in stem cell turnover around 1 year of age. Aging Cell 1, 121123. Becker, A.J., McCulloch, E.A., and Till, J.E. (1963). Cytological demonstration of the clonal nature of spleen colonies derived from transplanted mouse marrow cells. Nature 197, 452454. Becker, A.J., McCulloch, E.A., Siminovitch, L., and Till, J.E. (1965). The Effect of Differing Demands for Blood Cell Production on DNA Synthesis by Hemopoietic Colony-Forming Cells of Mice. Blood 26, 296308. Beerman, I., Bhattacharya, D., Zandi, S., Sigvardsson, M., Weissman, I.L., Bryder, D., and Rossi, D.J. (2010). Functionally distinct hematopoietic stem cells modulate hematopoietic lineage potential during aging by a mechanism of clonal expansion. Proc. Natl. Acad. Sci. USA 107, 54655470. Benveniste, P., Frelin, C., Janmohamed, S., Barbara, M., Herrington, R., Hyam, D., and Iscove, N.N. (2010). Intermediate-term hematopoietic stem cells with extended but time-limited reconstitution potential. Cell Stem Cell 6, 4858. Benz, C., Copley, M.R., Kent, D.G., Wohrer, S., Cortes, A., Aghaeepour, N., Ma, E., Mader, H., Rowe, K., Day, C., et al. (2012). Hematopoietic stem cell subtypes expand differentially during development and display distinct lymphopoietic programs. Cell Stem Cell 10, 273283. Berkahn, L., and Keating, A. (2004). Hematopoiesis in the elderly. Hematology 9, 159163. Bowie, M.B., McKnight, K.D., Kent, D.G., McCaffrey, L., Hoodless, P.A., and Eaves, C.J. (2006). Hematopoietic stem cells proliferate until after birth and show a reversible phase-specic engraftment defect. J. Clin. Invest. 116, 28082816. Bowie, M.B., Kent, D.G., Copley, M.R., and Eaves, C.J. (2007a). Steel factor responsiveness regulates the high self-renewal phenotype of fetal hematopoietic stem cells. Blood 109, 50435048.

(Rossi et al., 2007). These ndings, coupled with accruing evidence that normal HSCs do accumulate DNA damage with age, now provide formal support for a mechanistic connection. Accumulation of oxidative DNA damage appears to have a similar inuence on human HSCs (Yahata et al., 2011). In contrast to genomic DNA damage, mitochondrial DNA mutations do not change HSCs functions but profoundly alter the downstream, differentiating progenitors the HSCs generate (Norddahl et al., 2011). Many of the changes in HSCs that occur during early development and aging are likely governed by shared as well as distinct mechanisms. For example, the fetal-to-adult HSC transition seems to inuence most HSCs independently of their differentiation subtype, whereas the age-related changes in the peripheral blood composition in aging mice and humans appear to involve alterations in the relative prevalence of different HSC subtypes in the total HSC compartment. Additionally, other changes are likely to be caused by general mechanisms of senescence (Dykstra et al., 2011), which are overlaid on the aging-related mechanisms that affect the subtype make-up of the HSC compartment. Conclusion The body of work outlined in this review highlights the reality of heterogeneity in multiple biological properties of HSCs. These include their self-renewal, proliferation, and lineage output.

Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc. 695

Cell Stem Cell

Perspective
Bowie, M.B., Kent, D.G., Dykstra, B., McKnight, K.D., McCaffrey, L., Hoodless, P.A., and Eaves, C.J. (2007b). Identication of a new intrinsically timed developmental checkpoint that reprograms key hematopoietic stem cell properties. Proc. Natl. Acad. Sci. USA 104, 58785882. Bradford, G.B., Williams, B., Rossi, R., and Bertoncello, I. (1997). Quiescence, cycling, and turnover in the primitive hematopoietic stem cell compartment. Exp. Hematol. 25, 445453. ske, A.M., Vockentanz, L., Kharazi, S., Huska, M.R., Mancini, E., Scheller, Bro M., Kuhl, C., Enns, A., Prinz, M., Jaenisch, R., et al. (2009). DNA methylation protects hematopoietic stem cell multipotency from myeloerythroid restriction. Nat. Genet. 41, 12071215. Bryder, D., Rossi, D.J., and Weissman, I.L. (2006). Hematopoietic stem cells: the paradigmatic tissue-specic stem cell. Am. J. Pathol. 169, 338346. ssing, I., Slack, F.J., and Grosshans, H. (2008). let-7 microRNAs in developBu ment, stem cells and cancer. Trends Mol. Med. 14, 400409. Catlin, S.N., Busque, L., Gale, R.E., Guttorp, P., and Abkowitz, J.L. (2011). The replication rate of human hematopoietic stem cells in vivo. Blood 117, 44604466. Cavazzana-Calvo, M., and Fischer, A. (2007). Gene therapy for severe combined immunodeciency: are we there yet? J. Clin. Invest. 117, 14561465. Cavazzana-Calvo, M., Payen, E., Negre, O., Wang, G., Hehir, K., Fusil, F., Down, J., Denaro, M., Brady, T., Westerman, K., et al. (2010). Transfusion independence and HMGA2 activation after gene therapy of human b-thalassaemia. Nature 467, 318322. Cavazzana-Calvo, M., Fischer, A., Bushman, F.D., Payen, E., HaceinBey-Abina, S., and Leboulch, P. (2011). Is normal hematopoiesis maintained solely by long-term multipotent stem cells? Blood 117, 44204424. Challen, G.A., Boles, N.C., Chambers, S.M., and Goodell, M.A. (2010). Distinct hematopoietic stem cell subtypes are differentially regulated by TGF-beta1. Cell Stem Cell 6, 265278. Challen, G.A., Sun, D., Jeong, M., Luo, M., Jelinek, J., Berg, J.S., Bock, C., Vasanthakumar, A., Gu, H., Xi, Y., et al. (2012). Dnmt3a is essential for hematopoietic stem cell differentiation. Nat. Genet. 44, 2331. Cheshier, S.H., Morrison, S.J., Liao, X., and Weissman, I.L. (1999). In vivo proliferation and cell cycle kinetics of long-term self-renewing hematopoietic stem cells. Proc. Natl. Acad. Sci. USA 96, 31203125. ller-Sieburg, C.E. (2008). A new mechanism Cho, R.H., Sieburg, H.B., and Mu for the aging of hematopoietic stem cells: aging changes the clonal composition of the stem cell compartment but not individual stem cells. Blood 111, 55535561. Dick, J.E., Magli, M.C., Huszar, D., Phillips, R.A., and Bernstein, A. (1985). Introduction of a selectable gene into primitive stem cells capable of longterm reconstitution of the hemopoietic system of W/Wv mice. Cell 42, 7179. Ding, L., Saunders, T.L., Enikolopov, G., and Morrison, S.J. (2012). Endothelial and perivascular cells maintain haematopoietic stem cells. Nature 481, 457462. Dorshkind, K., and Montecino-Rodriguez, E. (2007). Fetal B-cell lymphopoiesis and the emergence of B-1-cell potential. Nat. Rev. Immunol. 7, 213219. Doulatov, S., Notta, F., Laurenti, E., and Dick, J.E. (2012). Hematopoiesis: a human perspective. Cell Stem Cell 10, 120136. Dykstra, B., Kent, D., Bowie, M., McCaffrey, L., Hamilton, M., Lyons, K., Lee, S.J., Brinkman, R., and Eaves, C. (2007). Long-term propagation of distinct hematopoietic differentiation programs in vivo. Cell Stem Cell 1, 218229. Dykstra, B., Olthof, S., Schreuder, J., Ritsema, M., and de Haan, G. (2011). Clonal analysis reveals multiple functional defects of aged murine hematopoietic stem cells. J. Exp. Med. 208, 26912703. Fleming, W.H., Alpern, E.J., Uchida, N., Ikuta, K., Spangrude, G.J., and Weissman, I.L. (1993). Functional heterogeneity is associated with the cell cycle status of murine hematopoietic stem cells. J. Cell Biol. 122, 897902. Ford, C.E., Hamerton, J.L., Barnes, D.W., and Loutit, J.F. (1956). Cytological identication of radiation-chimaeras. Nature 177, 452454. Gerrits, A., Dykstra, B., Kalmykowa, O.J., Klauke, K., Verovskaya, E., Broekhuis, M.J., de Haan, G., and Bystrykh, L.V. (2010). Cellular barcoding tool for clonal analysis in the hematopoietic system. Blood 115, 26102618. Glotzbach, J.P., Januszyk, M., Vial, I.N., Wong, V.W., Gelbard, A., Kalisky, T., Thangarajah, H., Longaker, M.T., Quake, S.R., Chu, G., and Gurtner, G.C. (2011). An information theoretic, microuidic-based single cell analysis permits identication of subpopulations among putatively homogeneous stem cells. PLoS ONE 6, e21211. He, S., Kim, I., Lim, M.S., and Morrison, S.J. (2011). Sox17 expression confers self-renewal potential and fetal stem cell characteristics upon adult hematopoietic progenitors. Genes Dev. 25, 16131627. Hock, H., Hamblen, M.J., Rooke, H.M., Schindler, J.W., Saleque, S., Fujiwara, Y., and Orkin, S.H. (2004a). G-1 restricts proliferation and preserves functional integrity of haematopoietic stem cells. Nature 431, 10021007. Hock, H., Meade, E., Medeiros, S., Schindler, J.W., Valk, P.J., Fujiwara, Y., and Orkin, S.H. (2004b). Tel/Etv6 is an essential and selective regulator of adult hematopoietic stem cell survival. Genes Dev. 18, 23362341. Hodges, E., Molaro, A., Dos Santos, C.O., Thekkat, P., Song, Q., Uren, P.J., Park, J., Butler, J., Rai, S., McCombie, W.R., et al. (2011). Directional DNA methylation changes and complex intermediate states accompany lineage specicity in the adult hematopoietic compartment. Mol. Cell 44, 1728. Hodgson, G.S., and Bradley, T.R. (1984). In vivo kinetic status of hematopoietic stem and progenitor cells as inferred from labeling with bromodeoxyuridine. Exp. Hematol. 12, 683687. Humphries, R.K., Jacky, P.B., Dill, F.J., Eaves, A.C., and Eaves, C.J. (1979). CFU-S in individual erythroid colonies derived in vitro from adult mouse marrow. Nature 279, 718720. Humphries, R.K., Eaves, A.C., and Eaves, C.J. (1981). Self-renewal of hemopoietic stem cells during mixed colony formation in vitro. Proc. Natl. Acad. Sci. USA 78, 36293633. Ikuta, K., and Weissman, I.L. (1992). Evidence that hematopoietic stem cells express mouse c-kit but do not depend on steel factor for their generation. Proc. Natl. Acad. Sci. USA 89, 15021506. Ikuta, K., Kina, T., MacNeil, I., Uchida, N., Peault, B., Chien, Y.H., and Weissman, I.L. (1990). A developmental switch in thymic lymphocyte maturation potential occurs at the level of hematopoietic stem cells. Cell 62, 863874. Iscove, N.N., and Nawa, K. (1997). Hematopoietic stem cells expand during serial transplantation in vivo without apparent exhaustion. Curr. Bio. 7, 805808. Jacobson, L.O., Simmons, E.L., Marks, E.K., and Eldredge, J.H. (1951). Recovery from radiation injury. Science 113, 510511. Ji, H., Ehrlich, L.I., Seita, J., Murakami, P., Doi, A., Lindau, P., Lee, H., Aryee, M.J., Irizarry, R.A., Kim, K., et al. (2010). Comprehensive methylome map of lineage commitment from haematopoietic progenitors. Nature 467, 338342. Jones, R.J., Collector, M.I., Barber, J.P., Vala, M.S., Fackler, M.J., May, W.S., Grifn, C.A., Hawkins, A.L., Zehnbauer, B.A., Hilton, J., et al. (1996). Characterization of mouse lymphohematopoietic stem cells lacking spleen colonyforming activity. Blood 88, 487491. Kent, D.G., Dykstra, B.J., Cheyne, J., Ma, E., and Eaves, C.J. (2008). Steel factor coordinately regulates the molecular signature and biologic function of hematopoietic stem cells. Blood 112, 560567. hrer, S., Dykstra, B.J., Ma, E., Cheyne, Kent, D.G., Copley, M.R., Benz, C., Wo J., Zhao, Y., Bowie, M.B., Zhao, Y., et al. (2009). Prospective isolation and molecular characterization of hematopoietic stem cells with durable selfrenewal potential. Blood 113, 63426350. Kiel, M.J., Yilmaz, O.H., Iwashita, T., Yilmaz, O.H., Terhorst, C., and Morrison, S.J. (2005). SLAM family receptors distinguish hematopoietic stem and progenitor cells and reveal endothelial niches for stem cells. Cell 121, 1109 1121. Kikuchi, K., and Kondo, M. (2006). Developmental switch of mouse hematopoietic stem cells from fetal to adult type occurs in bone marrow after birth. Proc. Natl. Acad. Sci. USA 103, 1785217857.

696 Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc.

Cell Stem Cell

Perspective
Kim, I., Saunders, T.L., and Morrison, S.J. (2007). Sox17 dependence distinguishes the transcriptional regulation of fetal from adult hematopoietic stem cells. Cell 130, 470483. Lemischka, I.R., Raulet, D.H., and Mulligan, R.C. (1986). Developmental potential and dynamic behavior of hematopoietic stem cells. Cell 45, 917927. Lu, R., Neff, N.F., Quake, S.R., and Weissman, I.L. (2011). Tracking single hematopoietic stem cells in vivo using high-throughput sequencing in conjunction with viral genetic barcoding. Nat. Biotechnol. 29, 928933. Majeti, R., Park, C.Y., and Weissman, I.L. (2007). Identication of a hierarchy of multipotent hematopoietic progenitors in human cord blood. Cell Stem Cell 1, 635645. McCulloch, E.A., Siminovitch, L., and Till, J.E. (1964). Spleen-Colony Formation in Anemic Mice of Genotype Ww. Science 144, 844846. Micklem, H.S., Ford, C.E., Evans, E.P., Ogden, D.A., and Papworth, D.S. (1972). Competitive in vivo proliferation of foetal and adult haematopoietic cells in lethally irradiated mice. J. Cell. Physiol. 79, 293298. Miller, J.P., and Allman, D. (2003). The decline in B lymphopoiesis in aged mice reects loss of very early B-lineage precursors. J. Immunol. 171, 23262330. Min, H., Montecino-Rodriguez, E., and Dorshkind, K. (2006). Effects of aging on the common lymphoid progenitor to pro-B cell transition. J. Immunol. 176, 10071012. Miyamoto, K., Araki, K.Y., Naka, K., Arai, F., Takubo, K., Yamazaki, S., Matsuoka, S., Miyamoto, T., Ito, K., Ohmura, M., et al. (2007). Foxo3a is essential for maintenance of the hematopoietic stem cell pool. Cell Stem Cell 1, 101112. Mochizuki-Kashio, M., Mishima, Y., Miyagi, S., Negishi, M., Saraya, A., Konuma, T., Shinga, J., Koseki, H., and Iwama, A. (2011). Dependency on the polycomb gene Ezh2 distinguishes fetal from adult hematopoietic stem cells. Blood 118, 65536561. Morita, Y., Ema, H., and Nakauchi, H. (2010). Heterogeneity and hierarchy within the most primitive hematopoietic stem cell compartment. J. Exp. Med. 207, 11731182. Morrison, S.J., Hemmati, H.D., Wandycz, A.M., and Weissman, I.L. (1995). The purication and characterization of fetal liver hematopoietic stem cells. Proc. Natl. Acad. Sci. USA 92, 1030210306. ller-Sieburg, C.E., Cho, R.H., Thoman, M., Adkins, B., and Sieburg, H.B. Mu (2002). Deterministic regulation of hematopoietic stem cell self-renewal and differentiation. Blood 100, 13021309. ller-Sieburg, C., Sieburg, H.B., Bernitz, J.M., and Cattarossi, G. (2012). Mu Stem cell heterogeneity: implications for aging and regenerative medicine. Blood 119, 39003907. Nakada, D., Saunders, T.L., and Morrison, S.J. (2010). Lkb1 regulates cell cycle and energy metabolism in haematopoietic stem cells. Nature 468, 653658. Norddahl, G.L., Pronk, C.J., Wahlestedt, M., Sten, G., Nygren, J.M., Ugale, A., Sigvardsson, M., and Bryder, D. (2011). Accumulating mitochondrial DNA mutations drive premature hematopoietic aging phenotypes distinct from physiological stem cell aging. Cell Stem Cell 8, 499510. Notta, F., Doulatov, S., Laurenti, E., Poeppl, A., Jurisica, I., and Dick, J.E. (2011). Isolation of single human hematopoietic stem cells capable of longterm multilineage engraftment. Science 333, 218221. Ogawa, M., Porter, P.N., and Nakahata, T. (1983). Renewal and commitment to differentiation of hemopoietic stem cells (an interpretive review). Blood 61, 823829. Osawa, M., Hanada, K., Hamada, H., and Nakauchi, H. (1996). Long-term lymphohematopoietic reconstitution by a single CD34-low/negative hematopoietic stem cell. Science 273, 242245. Park, I.K., Qian, D., Kiel, M., Becker, M.W., Pihalja, M., Weissman, I.L., Morrison, S.J., and Clarke, M.F. (2003). Bmi-1 is required for maintenance of adult self-renewing haematopoietic stem cells. Nature 423, 302305. Pawliuk, R., Eaves, C., and Humphries, R.K. (1996). Evidence of both ontogeny and transplant dose-regulated expansion of hematopoietic stem cells in vivo. Blood 88, 28522858. Ploemacher, R.E., and Brons, R.H. (1989). Separation of CFU-S from primitive cells responsible for reconstitution of the bone marrow hemopoietic stem cell compartment following irradiation: evidence for a pre-CFU-S cell. Exp. Hematol. 17, 263266. Rossi, D.J., Bryder, D., Zahn, J.M., Ahlenius, H., Sonu, R., Wagers, A.J., and Weissman, I.L. (2005). Cell intrinsic alterations underlie hematopoietic stem cell aging. Proc. Natl. Acad. Sci. USA 102, 91949199. Rossi, D.J., Bryder, D., Seita, J., Nussenzweig, A., Hoeijmakers, J., and Weissman, I.L. (2007). Deciencies in DNA damage repair limit the function of haematopoietic stem cells with age. Nature 447, 725729. mmendorf, T.H., Kolvraa, S., Bischoff, C., Christensen, K., Rufer, N., Bru Wadsworth, L., Schulzer, M., and Lansdorp, P.M. (1999). Telomere uorescence measurements in granulocytes and T lymphocyte subsets point to a high turnover of hematopoietic stem cells and memory T cells in early childhood. J. Exp. Med. 190, 157167. llerSieburg, H.B., Cho, R.H., Dykstra, B., Uchida, N., Eaves, C.J., and Mu Sieburg, C.E. (2006). The hematopoietic stem compartment consists of a limited number of discrete stem cell subsets. Blood 107, 23112316. ller-Sieburg, C.E. (2011). Predicting clonal Sieburg, H.B., Rezner, B.D., and Mu self-renewal and extinction of hematopoietic stem cells. Proc. Natl. Acad. Sci. USA 108, 43704375. Siminovitch, L., McCulloch, E.A., and Till, J.E. (1963). The Distribution of Colony-Forming Cells among Spleen Colonies. J. Cell. Physiol. 62, 327336. Takubo, K., Goda, N., Yamada, W., Iriuchishima, H., Ikeda, E., Kubota, Y., Shima, H., Johnson, R.S., Hirao, A., Suematsu, M., and Suda, T. (2010). Regulation of the HIF-1alpha level is essential for hematopoietic stem cells. Cell Stem Cell 7, 391402. Till, J.E., and McCulloch, E.A. (1961). A direct measurement of the radiation sensitivity of normal mouse bone marrow cells. Radiat. Res. 14, 213222. Till, J.E., McCulloch, E.A., and Siminovitch, L. (1964). A Stochastic Model of Stem Cell Proliferation, Based on the Growth of Spleen Colony-Forming Cells. Proc. Natl. Acad. Sci. USA 51, 2936. Trowbridge, J.J., Snow, J.W., Kim, J., and Orkin, S.H. (2009). DNA methyltransferase 1 is essential for and uniquely regulates hematopoietic stem and progenitor cells. Cell Stem Cell 5, 442449. Wang, J., Sun, Q., Morita, Y., Jiang, H., Gross, A., Lechel, A., Hildner, K., Guachalla, L.M., Gompf, A., Hartmann, D., et al. (2012). A differentiation checkpoint limits hematopoietic stem cell self-renewal in response to DNA damage. Cell 148, 10011014. Wilson, A., Laurenti, E., Oser, G., van der Wath, R.C., Blanco-Bose, W., Jaworski, M., Offner, S., Dunant, C.F., Eshkind, L., Bockamp, E., et al. (2008). Hematopoietic stem cells reversibly switch from dormancy to selfrenewal during homeostasis and repair. Cell 135, 11181129. Yahata, T., Takanashi, T., Muguruma, Y., Ibrahim, A.A., Matsuzawa, H., Uno, T., Sheng, Y., Onizuka, M., Ito, M., Kato, S., and Ando, K. (2011). Accumulation of oxidative DNA damage restricts the self-renewal capacity of human hematopoietic stem cells. Blood 118, 29412950. Yamazaki, S., Iwama, A., Takayanagi, S., Eto, K., Ema, H., and Nakauchi, H. (2009). TGF-beta as a candidate bone marrow niche signal to induce hematopoietic stem cell hibernation. Blood 113, 12501256. Yamazaki, S., Ema, H., Karlsson, G., Yamaguchi, T., Miyoshi, H., Shioda, S., Taketo, M.M., Karlsson, S., Iwama, A., and Nakauchi, H. (2011). Nonmyelinating Schwann cells maintain hematopoietic stem cell hibernation in the bone marrow niche. Cell 147, 11461158. Yuan, J., Nguyen, C.K., Liu, X., Kanellopoulou, C., and Muljo, S.A. (2012). Lin28b Reprograms Adult Bone Marrow Hematopoietic Progenitors to Mediate Fetal-like Lymphopoiesis. Science 335, 11951200. Zou, P., Yoshihara, H., Hosokawa, K., Tai, I., Shinmyozu, K., Tsukahara, F., Maru, Y., Nakayama, K., Nakayama, K.I., and Suda, T. (2011). p57(Kip2) and p27(Kip1) cooperate to maintain hematopoietic stem cell quiescence through interactions with Hsc70. Cell Stem Cell 9, 247261.

Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc. 697

Cell Stem Cell

ISSCR: Committee Forum


dric Blanpain: Ce ISSCRs Outstanding Young Investigator for 2012
dric Blanpain from the Interdisciplinary Research Institute, Universite libre de Bruxelles earned the 2012 Ce ISSCR-University of Pittsburgh Outstanding Young Investigator Award for his exceptional achievements as an early-career stem cell researcher.
The International Society for Stem Cell Research (ISSCR) was delighted to highlight the remarkable caliber of an early-career stem cell biologist with the fourth annual ISSCR-University of Pittsburgh Outstanding Young Investigator Award. This years dric Blanpain, M.D., Ph.D., a tenured Assistant award went to Ce Professor of the Belgian Research National Scientic Fund (FNRS) at the Interdisciplinary Research Institute (IRIBHM), libre de Bruxelles (ULB) in Belgium. Since starting Universite dric has distinguished his own laboratory at FNRS in 2006, Ce himself as a leader in stem cell research, dening stem cell contributions to development, tissue homeostasis, and cancer, focusing on the skin, the mammary gland, and the heart. dric received his M.D. at the ULB, in June 1995, with summa Ce cum laude and was board certied in internal medicine in 2002. He demonstrated his talent as a researcher from the start, excelling in his Ph.D. studies. In the laboratory of Marc dric studied how CCR5, Parmentier, M.D., Ph.D., at the ULB, Ce a receptor critical for HIV infection, interacts with its natural ligands and the viral envelope, and how CCR5 trafcking inuences this critical step of viral infection. This research resulted in a number of highly cited publications and was recognized by several awards, including the Galien Award of Pharmacology in 2002, a prestigious award for young scientists in Belgium. dric moved to the Rockefeller University in After graduation, Ce New York City, where he launched his career in stem cell biology dric began as a postdoctoral fellow in my laboratory. In my group, Ce a longstanding and remarkably fruitful collaboration with another of my postdoctoral fellows, William (Bill) Lowry (now at University of California, Los Angeles), publishing three important co-rstauthored papers during their postdoctoral studies at Rockefeller. When the rst and most important of these studies was launched, Doina Tumbar in my group had just developed a novel method to isolate and purify hair follicle bulge stem cells by using transgenic mice that uorescently marked infrequently cycling, label-retaining cells of the skin epidermis. Exploiting our stem cell dric and Bill devised a new purication gene expression prole, Ce scheme to isolate follicle stem cells by uorescence-activated cell sorting for surface markers. By optimizing conditions that permit clonal analyses in vitro and engraftments in vivo, they showed that the progenies of a single bulge stem cell were able to differentiate into all lineages of the epidermis, including hair follicles and sebaceous glands, when transplanted into recipient hairless mice. This landmark work provided the rst demonstration that the infrequently cycling hair follicle bulge cells can function as multipotent stem cells in tissue regeneration (Blanpain et al., 2004). In the second major study, the two pursued another of our prior observations, namely that excessive b-catenin stabilization in the developing epidermis resulted in de novo hair follicle morphogen-

dric and esis. Using a regulateable gain-of-function strategy, Ce Bill demonstrated that in the adult follicle, b-catenin stabilization results in precocious activation of the stem cells (Lowry et al., 2005). Finally, the two used a loss-of-function approach to show that Notch signaling functions in the skin epidermis in vivo by promoting the transition of proliferative basal epidermal cells to differentiating suprabasal (spinous) cells (Blanpain et al., 2006). All three of these papers were highlighted by other journals and have received considerable attention in the eld. When dric departed from my lab to begin his independent career, Ce he did so leaving a legacy as a brilliant young scientist with an extraordinary capacity to ask interesting questions, collaborate, andthrough his leadershipmove science forward. What also dric special is his gift for delivering incisive, construcmakes Ce tive, and fair criticism with a smile. Importantly, the high standards he applies to others are also the ones he applies to himself. Such people are both rare and exceptional in science, and they make the research of those around them better. dric has As an independent researcher in Brussels, Ce continued to shine in pursuing his interests in epithelial stem cell biology. His research group now studies the role of stem cells during homeostasis and in disease, particularly cancer. He has applied his knowledge and skills beyond the skin, and has made groundbreaking contributions to the eld of stem cell research. drics group Here, I highlight only some of the seminal work that Ce has accomplished in the short time since he left my laboratory. For the vast majority of cancers, the origins of the tumor drics initiating cell are still unknown. To address this problem, Ce lab focused on basal cell and squamous cell carcinomas of the skin, providing key insights into which cells within the skin give rise to these two common epithelial cancers. His team rst reported that basal cell carcinomas arise from long-lived stem cells of the interfollicular epidermis (Youssef et al., 2010). He
Cell Stem Cell 10, 751752, June 14, 2012 2012 ISSCR 751

ISSCR: Committee Forum


then showed that, in contrast, squamous cell carcinomas can derive from hair follicle stem cells and possibly other cell types within the skin, a nding supported by similar ndings from his former collaborator, Bill Lowry (Lapouge et al., 2011; White et al., 2011). These results were quite unexpected, since basal cell carcinomas more closely resemble hair follicle cell differentiation whereas squamous cell carcinomas show a differentiation program more closely parallel with the epidermis. Another key question in the cancer eld is how adult stem cells maintain their genomic integrity, and sense and respond to DNA drics group was one of damage within their natural niche. Ce two that showed that hair follicle stem cells do not protect their genome by asymmetrical chromosome segregation (Sotiropoulou et al., 2008; Waghmare et al., 2008). His group also demonstrated that adult bulge stem cells are resistant to DNAdamage-induced cell death due to an elevated expression of Bcl2 and an accelerated DNA repair activity (Sotiropoulou et al., 2010). This work has important implications for our understanding of how skin stem cells protect themselves against sun damage and how these cells may get transformed. Cancer stem cells (CSCs) have been described in many different mouse and human cancers, including squamous tumors of the skin drics group showed that one of (Schober and Fuchs, 2011). Ce these CSC factors, VEGF, plays a critical dual role in regulating skin tumor stemness (Beck et al., 2011). The secretion of VEGF by tumor cells stimulates neoangiogenesis, which creates a vascular niche. Intriguingly, VEGF also acts directly on tumor cells in an autocrine loop through an Nrp1-dependent mechanism to promote CSC renewal and tumor growth. These elegant studies give new insights and add a concrete mechanism to our understanding of the intricate crosstalk happening between CSCs and their microenvironment. Illustrating his versatility and ability to address important ques dric used novel lineage tracing and clonal tions in the eld, Ce analysis in mice to decipher the cellular hierarchy of the mammary epithelium during development, homeostasis, and lactation. His group discovered that under physiological conditions, mammary epithelial lineages originate from multipotent embryonic progenitors that are replaced soon after birth by distinct, lineage-restricted unipotent stem cells (Van Keymeulen et al., 2011). This work has been widely discussed and highlighted and is likely to be a paradigm-shifting landmark in the eld. drics ability to transition from one stem In a nal example of Ce cell type to another, he identied the earliest cardiovascular progenitor cell and described stem cell hierarchy and functionality in the mammary gland. He used mouse embryonic stem cells (ESCs) in which gene expression can be temporally regulated, demonstrating that Mesp1 acts as a key regulatory switch during cardiovascular specication (Bondue et al., 2008). Since then he has engineered a Mesp1-GFP reporter in ESCs and shown that Mesp1-GFP cells are strongly enriched for multipotent cardiovascular progenitors, which at the clonal level can differentiate into all cardiovascular cell types (Bondue et al., 2011; van den Ameele et al., 2012). drics broad expertise and diversity of accomplishments in Ce stem cell biology has earned him wide respect, and his work has been praised for its thoroughness and conceptual depth. He is widely commended for his active participation in the stem cell community, his contributions to editorial boards, and the authoritative reviews that he has continued to write on his own as an established independent investigator. His recognition by the eld is illustrated by the numerous talks he has been invited to give and by grant support and awards. It is only tting that he is now the recipient of the 2012 ISSCR-University of Pittsburgh Outstanding Young Investigator Award. Within the international dric has emerged as a community of stem cell biologists, Ce collegial, versatile, and thoughtful young scientist with a spectacular ability to conduct stem cell science at the highest level. dric Those of us who have had the pleasure of interacting with Ce have witnessed his scholarship and productivity rst-hand. dric this honor, the stem cell community applauds In awarding Ce his shining trajectory and eagerly anticipates his next discovery.
REFERENCES Beck, B., Driessens, G., Goossens, S., Youssef, K.K., Kuchnio, A., Caauwe, A., Sotiropoulou, P.A., Loges, S., Lapouge, G., Candi, A., et al. (2011). Nature 478, 399403. Blanpain, C., Lowry, W.E., Geoghegan, A., Polak, L., and Fuchs, E. (2004). Cell 118, 635648. Blanpain, C., Lowry, W.E., Pasolli, H.A., and Fuchs, E. (2006). Genes Dev. 20, 30223035. Bondue, A., Lapouge, G., Paulissen, C., Semeraro, C., Iacovino, M., Kyba, M., and Blanpain, C. (2008). Cell Stem Cell 3, 6984. nnler, S., Chiapparo, G., Chabab, S., Ramialison, M., Paulissen, Bondue, A., Ta C., Beck, B., Harvey, R., and Blanpain, C. (2011). J. Cell Biol. 192, 751765. Lapouge, G., Youssef, K.K., Vokaer, B., Achouri, Y., Michaux, C., Sotiropoulou, P.A., and Blanpain, C. (2011). Proc. Natl. Acad. Sci. USA 108, 74317436. Lowry, W.E., Blanpain, C., Nowak, J.A., Guasch, G., Lewis, L., and Fuchs, E. (2005). Genes Dev. 19, 15961611. Schober, M., and Fuchs, E. (2011). Proc. Natl. Acad. Sci. USA 108, 1054410549. Sotiropoulou, P.A., Candi, A., and Blanpain, C. (2008). Stem Cells 26, 29642973. , G., De Clercq, S., Youssef, K.K., Sotiropoulou, P.A., Candi, A., Mascre Lapouge, G., Dahl, E., Semeraro, C., Denecker, G., Marine, J.C., and Blanpain, C. (2010). Nat. Cell Biol. 12, 572582. van den Ameele, J., Tiberi, L., Bondue, A., Paulissen, C., Herpoel, A., Iacovino, M., Kyba, M., Blanpain, C., and Vanderhaeghen, P. (2012). EMBO Rep. 13, 355362. Van Keymeulen, A., Rocha, A.S., Ousset, M., Beck, B., Bouvencourt, G., Rock, J., Sharma, N., Dekoninck, S., and Blanpain, C. (2011). Nature 479, 189193. Waghmare, S.K., Bansal, R., Lee, J., Zhang, Y.V., McDermitt, D.J., and Tumbar, T. (2008). EMBO J. 27, 13091320. White, A.C., Tran, K., Khuu, J., Dang, C., Cui, Y., Binder, S.W., and Lowry, W.E. (2011). Proc. Natl. Acad. Sci. USA 108, 74257430. Youssef, K.K., Van Keymeulen, A., Lapouge, G., Beck, B., Michaux, C., Achouri, Y., Sotiropoulou, P.A., and Blanpain, C. (2010). Nat. Cell Biol. 12, 299305.

Cell Stem Cell

Elaine Fuchs1,*
1Laboratory of Mammalian Cell Biology and Development, Howard Hughes Medical Institute, The Rockefeller University, New York, NY 10065, USA *Correspondence: fuchslb@rockefeller.edu DOI 10.1016/j.stem.2012.05.001

752 Cell Stem Cell 10, 751752, June 14, 2012 2012 ISSCR

Cell Stem Cell

Perspective
Adult Neural Stem Cells Bridge Their Niche
Luis C. Fuentealba,1 Kirsten Obernier,1 and Arturo Alvarez-Buylla1,*
1Department of Neurological Surgery and Institute for Regeneration Medicine, University of California, San Francisco, San Francisco, CA 94143, USA *Correspondence: abuylla@stemcell.ucsf.edu DOI 10.1016/j.stem.2012.05.012

Major developments in the neural stem cell (NSC) eld in recent years provide new insights into the nature of the NSC niche. In this perspective, we integrate recent anatomical data on the organization of the two main neurogenic niches in the adult brain, the ventricular-subventricular zone (V-SVZ) and the subgranular zone (SGZ), with signaling pathways that control the behavior of NSCs. NSCs in the adult brain stretch into physiologically distinct compartments of their niche. We propose how adult NSCs morphology may allow these cells to integrate multiple signaling pathways arising from unique locations of their niche.
The fascinating process of developmental tissue growth and morphogenesis is orchestrated by stem cells that contribute to organ maintenance (tissue homeostasis) and repair in the adult organism. For many years, the brain, with its extraordinary structure, connectivity, complexity, and diversity of cell types, was considered an exception; neural stem cells (NSCs) were thought to be present only during development when this amazing organ is put together. This view began to change with the discovery of adult neurogenesis (for historical perspective, see [Altman, 2011; Nottebohm, 2011]), followed by the identication of cells that in vitro (Reynolds and Weiss, 1992) and in vivo (Doetsch et al., 1999; Seri et al., 2001, 2004) can function as NSCs generating neurons, glial cells, or both. These discoveries led to a shift in concepts, not only about the potential of the postnatal brain to engage in processes only thought possible in development, but also about the nature of NSCs themselves. In the adult mammalian brain, NSCs are retained in two regions, the ventricular-subventricular zone (V-SVZ) and the subgranular zone (SGZ). The V-SVZ, in the walls of the lateral ventricles (Figure 1), contains a subpopulation of cells with astroglial properties (B1 cells) that function as NSCs, giving rise to intermediate progenitors (IPCs or transient amplifying progenitors, also known as C cells), which in rodents predominantly generate neurons destined for the olfactory bulb (OB) (for review see [Kriegstein and Alvarez-Buylla, 2009]). In the SGZ at the interface of the hilus and dentate gyrus (Figure 2), NSCs also correspond to astroglial cells, which have a radial process that traverses the granule cell layer. These cells, known by multiple namesradial astrocytes (Seri et al., 2001, 2004), type-1 progenitors (Filippov et al., 2003), or radial glia-like cells (Bonaguidi et al., 2011) generate new dentate granule neurons via IPC1 and IPC2 (also known as type 2a and 2b cells, respectively) (for review see [Bonaguidi et al., 2012; Zhao et al., 2008]). In both germinal zones, NSCs and IPCs correspond to the primary and secondary progenitors, respectively. Many studies analyzing proliferation do not distinguish between primary or secondary progenitors, and in these cases we will refer to both collectively as progenitor cells. The adult V-SVZ NSCs in the walls of the lateral ventricles differ signicantly in location and structure from those in the adult hippocampal SGZ. Unlike B1 cells in the V-SVZ, which like many embryonic neural stem cells lie next to the ventricle and have
698 Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc.

processes that contact the cerebrospinal uid (CSF), dentate radial astrocytes are found deeper in the brain parenchyma, away from the walls of the ventricle and surrounded by neurons and other glial cells. Yet, B1 cells and radial astrocytes share some key features. Both express astroglial markers and have ultrastructural characteristics of astrocytes (Kriegstein and Alvarez-Buylla, 2009). Most importantly, they both have long processes that allow them to reach into compartments of the niche far away from where the cell bodies reside (Figures 1 and 2). Through some of these processes, NSCs make contact with the vasculature, which in both germinal regions plays key roles in their regulation. Therefore, NSCs are in contact with unique compartments of their niche, and this may determine whether they remain quiescent or are induced to proliferate. A large number of studies during the last decade provide important new insights about the nature of adult NSCs and IPCs and about the microenvironment that surrounds them. In this perspective, we aim to connect the recent literature providing insights into the anatomy of the adult neurogenic niches with the emerging knowledge of factors that control the behavior of NSCs (Table 1). Organization of adult V-SVZ and SGZ NSCs into distinct domains may ultimately provide an integrative perspective on how adult NSCs are regulated (Figures 1 and 2). Domains of Adult Ventricular-Subventricular Zone B1 Cells V-SVZ B1 cells are immersed in a remarkably diverse microenvironment. The highly specialized architecture within this niche implicates both cell-cell interactions and soluble factors as important regulators of NSC behavior. B1 cells retain the basic apical-basal polarity of their predecessors, radial glia. Similar to radial glia and neuroepithelial cells, most, if not all, B1 cells contact the ventricle through small, specialized apical processes that contain a single primary cilium (Mirzadeh et al., 2008; Shen et al., 2008). They also have long basal processes with specialized endings contacting blood vessels (BV). Therefore, adult B1 cells are also part of a VZ and not only a SVZ, hence the new descriptor: V-SVZ (Ihrie and Alvarez-Buylla, 2011). A VZ is retained in many, if not all, adult vertebrates including birds, reptiles, amphibians, and sh (Alvarez-Buylla et al., 1998; Byrd a-Verdugo and Brunjes, 2001; Chapouton et al., 2007; Garc et al., 2002; Goldman and Nottebohm, 1983; Polenov and

Cell Stem Cell

Perspective
Figure 1. Schematic of the Different Domains of B1 Cells within the Adult V-SVZ
The upper left panel shows a frontal cross-section of the adult mouse brain showing the location of the V-SVZ, where neurogenesis in walls of the lateral ventricles (V) continues throughout life. The lower panel shows cellular composition of the adult V-SVZ niche and domains of B1 cells. NSCs correspond to type B1 cells (blue). B1 cells are surrounded by multiciliated ependymal cells (E) forming pinwheel-like structures on the ventricular surface. B1 cells give rise to IPCs (or C cells, green), which correspond to transit-amplifying cells that divide to generate neuroblasts (type A cells, red). B1 cells retain epithelial properties, with a thin apical process (containing a primary cilium) that contacts the lateral ventricle (V) and a long basal process ending on blood vessels (BV, purple). Therefore, B1 cells can be subdivided into three domains. Domain I (proximal or apical, dark blue) contains the primary cilium and is in direct contact with the CSF; within this domain, B1 cells can access soluble factors within the CSF and signaling molecules from neighboring ependymal cells. Domain II (intermediate, medium blue) is in close proximity to IPCs, neuroblasts, neuronal terminals, and other supporting cells; cell-cell interactions between B1 cells and their progeny could occur within this compartment. Domain III (distal, light blue) comprises a basal process ending in a specialized end-foot that contacts blood vessels; bloodborne factors and endothelial-derived factors may act on B1 cells in this domain.

Chetverukhin, 1993). To better understand how signals within the V-SVZ may be compartmentalized, we propose that V-SVZ B1 cells can be subdivided into three domains: proximal (apical, I), intermediate (II), and distal (basal, III) (Figure 1). A Periscope in the VentricleB1 Cells Proximal Domain The proximal domain of B1 cells is in direct contact with the ventricle. When viewed en face, from the ventricular side, the rodent V-SVZ is organized as pinwheels; the small apical endings of B1 cells in the center are surrounded by a rosette of ependymal cells with large apical surfaces (Figure 1). Intercellular junctions are found between B1 cells, at B1-ependymal boundaries, and between ependymal cells, and each type has unique ultrastructural characteristics (Mirzadeh et al., 2008). Expression of ankyrin3, an adaptor protein known to regulate the attachment of membrane proteins (including N-cadherin) to the cytoskeleton, is specically found in the apical-lateral borders of ependymal cells, but not in B1 cells. Its expression is controlled by the ependymal-specic transcription factor FOXJ1. Inactivation of FoxJ1 in the postnatal brain results in decreased ankyrin3 and reduced neurogenesis (Paez-Gonzalez et al., 2011). Ependymal cells also help maintain the molecular composition of the apical compartment by propelling the CSF with their multiple motile cilia (Sawamoto et al., 2006). B1 cells, with their small apical surface, are in direct contact with the CSF, which contains soluble factors that could modulate NSC behavior (Lehtinen et al., 2011; Zappaterra et al., 2007). Insulin-like growth factor 2 (IGF2) in the adult CSF has been shown to regulate V-SVZ progenitor proliferation (Lehtinen et al., 2011). Bone morphogenic proteins (BMPs), Wnts, Sonic hedgehog (SHH), and retinoic acid are also present in the CSF and may modulate the behavior of B1 cells (Huang et al., 2010; Lehtinen et al., 2011).

Ependymal cells secrete the BMP antagonist noggin, which promotes V-SVZ progenitor proliferation and neuroblast generation in vitro and in vivo (Lim et al., 2000; Peretto et al., 2004). Consistently, ependymal expression of LRP2, a receptor that sequesters BMP4, is required for progenitor proliferation and neurogenesis in vivo (Gajera et al., 2010). Yet, there is other evidence indicating that inhibition of BMPs increases oligodendrogenesis at the expense of neurogenesis (Colak et al., 2008) and might be involved in increased glial differentiation upon demyelination (Jablonska et al., 2010). The proximal domain of B1 cells also harbors a primary cilium that may directly integrate signaling of CSF factors, although such integration remains to be demonstrated. Disruption of a gene encoding for a protein required for cilia assembly intraagellar transport 88 (Ift88) in glutamate aspartate transporter (GLAST)-positive cells results in decreased numbers of adult bromodeoxyuridine (BrdU)-label retaining cells (Beckervordersandforth et al., 2010). In contrast, another recent study suggests that depletion of Ift20 in cells expressing glial brillary acidic protein (GFAP) at early postnatal stages had only minor effects on neurogenesis in the adult V-SVZ (Amador-Arjona et al., 2011). The primary cilium is essential for transduction of Shh signaling during neural tube development (Wong and Reiter, 2008). It is therefore possible that this organelle may be required for the transduction of Shh signaling observed in specic subregions of the V-SVZ (Ihrie et al., 2011). Necking with the NeighborsB1 Cells Intermediate Domain Sustained neurogenesis throughout life requires a tight balance between NSC proliferation and the number of differentiated progeny produced. It has been hypothesized that feedback
Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc. 699

Cell Stem Cell

Perspective
Figure 2. Schematic of the Different Domains of SGZ Radial Astrocytes
The upper panel shows a frontal cross-section of the adult mouse brain showing the hippocampal formation (left). The insert shows a higher magnication indicating the location of the dentate gyrus (right). The lower panel shows cellular composition of the adult dentate gyrus and domains of SGZ radial astrocytes. Radial astrocytes (RA, also known as type 1 cells, blue) give rise to IPCs (green), which progressively (via IPC1 and IPC2 [type 2a and type 2b cells]) differentiate into immature granule cells (IGCs [type 3 cells], red). Mature granule cells (GCs, brown) send an axon parallel to the SGZ into the hilus, whereas their dendrites branch into the molecular layer (ML). Radial astrocytes are polarized cells with their cell body in the SGZ, a long main shaft that traverses through the granule cell layer (GCL) and then branches diffusely in the inner molecular layer (IML). Here we subdivide radial astrocytes into three domains. Domain I (proximal, dark blue) faces the hilus, harbors a primary cilium, and contacts with blood vessels (BV, purple) and, through lateral processes, neighboring radial astrocytes. Factors derived from blood, endothelial cells, and neighboring radial astrocytes act on NSCs within this domain. Domain II (intermediate, medium blue) contains the cell body and the main shaft. This part of the cell interacts closely with IPCs and GCs; this domain allow specic cell-cell interactions of NSCs with their progeny and detection of local neural activity and signaling from GCs. Domain III (distal, light blue) contacts other glial cells, axons, and synaptic terminals in the IML; NSCs may detect levels of neural activity from Mossy cells and other neurons within this compartment.

mechanisms must exist to inform NSCs of the number of new neurons already generated. The intermediate domain of B1 cells is in intimate contact with IPCs and neuroblasts, perhaps allowing direct feedback from progeny to NSCs. Canonical Notch signaling is highly active in V-SVZ NSCs and regulates their maintenance (Imayoshi et al., 2010). Conditional depletion of the downstream effector, recombining binding protein suppressor of hairless (RbpJk), in adult nestin+ cells leads to a transient increase in IPCs and newborn OB neurons followed by a drastic reduction in NSC numbers that ultimately results in reduced neurogenesis. Thus, Notch signaling might maintain B1 cells by inhibiting the production of IPCs. IPCs express high levels of ASCL1 (MASH1), which is repressed by HES1, a downstream effector of Notch signaling. ASCL1 in turn is known to promote the expression of Notch ligands (Kopan and Ilagan, 2009), suggesting a possible feedback mechanism via lateral inhibition between IPCs and NSCs by direct cell-cell contact. Feedback mechanisms may also occur among NSCs as shown in zebrash (Chapouton et al., 2010). In this vertebrate model, expression of the Notch ligand DeltaA in dividing NSCs, activates Notch in neighboring NSCs maintaining quiescence. Although it is unclear how Notch signaling activity is regulated
700 Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc.

in the murine V-SVZ, Notch ligands are expressed throughout the V-SVZ and Delta1 and Jagged1 expression has been observed in IPCs and neuroblasts (Aguirre et al., 2010; Irvin et al., 2004). This is in agreement with earlier observations in which ablation of IPCs and neuroblasts by the antimitotic drug AraC activates NSCs that leads to V-SVZ regeneration (Doetsch et al., 1999). On the other hand, it has been suggested that increased numbers of IPCs due to enhanced epidermal growth factor receptor (EGFR) signaling may suppress Notch signaling in NSCs, although the exact mechanism remains unclear (Aguirre et al., 2010). During development, Notch signaling oscillates during interkinetic nuclear migration of radial glia (Shimojo et al., 2008). It is unknown whether Notch signaling oscillates in the adult V-SVZ during cell-cycle progression. However, it has been shown that mitotic B1 cells tend to be closer to the ventricular surface (Mirzadeh et al., 2008). Whether changes in the position of B1 cells nuclei are associated with different phases of the cell cycle remains to be determined (Shen et al., 2008; Tavazoie et al., 2008). Neurotransmitters in the V-SVZ also appear to regulate NSC behavior, and this regulation is likely to occur in the intermediate domain where newly generated neuroblasts closely interact with both of their predecessors, B1 cells and IPCs. Neuroblasts spontaneously release the neurotransmitter gamma-aminobutyric acid (GABA) and induce depolarization of progenitors by activation of functional GABAA-receptors. This, in turn, inhibits progenitor cell-cycle progression and neuronal production via epigenetic mechanisms that involve phosphorylation of H2AX (Fernando et al., 2011; Liu et al., 2005). The diazepam binding

Cell Stem Cell

Perspective
Table 1. Summary of Factors Involved in the Regulation of Progenitor Cell Behavior and Putative Domains Where They May Act B1 cell domain I. Proximal (apical) Factor Igf2 Noggin/LRP2 Noggin/Chordin Primary cilium Shh II. Intermediate Notch GABA DBI Dopamine Serotonin III. Distal (basal) SDF1 PEDF BTC Radial astrocyte domain I. Proximal Factor VEGF Primary cilium II. Intermediate Notch Noggin/FXR2 BMPs Long-term potentiation Seizures III. Distal GABA Glutamate/NMDAR Effect on progenitor cells Proliferation [ Proliferation [ Oligodendrogenesis [ Proliferation [ Ventral specication Maintenance Proliferation Y Proliferation [ Proliferation [ Proliferation/Neurogenesis [ Recruitment to vasculature Proliferation [ Proliferation/Neurogenesis [ Effect on progenitor cells Proliferation/Neurogenesis [ Proliferation [ Maintenance Maintenance Proliferation [ Quiescence Proliferation/Neurogenesis [ Proliferation/Neurogenesis [ Proliferation Y Differentiation [ Differentiation [ References Lehtinen et al. (2011) Lim et al. (2000), Peretto et al. (2004), Gajera et al. (2010) Colak et al. (2008), Jablonska et al. (2010) Amador-Arjona et al. (2011), Beckervordersandforth et al. (2010) Ihrie et al. (2011) Imayoshi et al. (2010), Aguirre et al. (2010), Chapouton et al. (2010) Liu et al. (2005), Fernando et al. (2011) Alfonso et al. (2012) glinger et al. (2004), Ho OKeeffe et al. (2009), Kim et al. (2010) Banasr et al. (2004) Shen et al. (2008) rez-Castillejo et al. (2006), Ram et al. (2009) Andreu-Agullo mez-Gaviro et al. (2012) Go References Licht et al. (2011), Cao et al. (2004) Amador-Arjona et al. (2011) Breunig et al. (2008), Han et al. (2008) Lugert et al. (2010), Ables et al. (2010), Ehm et al. (2010) Bonaguidi et al. (2008), Guo et al. (2011) Mira et al. (2010) Bruel-Jungerman et al. (2006) Parent et al. (1997), Ma et al. (2009) Wang et al. (2005), Tozuka et al. (2005) Deisseroth et al. (2004)

This table summarizes some of the known factors that affect progenitor behavior, although for many factors it is still unclear whether they act directly on NSCs, IPCs, or both. We here suggest possible sites of action within the three proposed putative domains of B1 cells (Figure 1) and radial astrocytes (Figure 2), the NSCs of V-SVZ and SGZ, respectively. The upper half of the table lists factors that may act on V-SVZ B1 cells. Domain I shows factors released from ependymal cells and present in the CSF. Domain II shows signaling between B1 cells and their progeny or neuronal terminals from other parts of the brain. Domain III shows blood-borne factors and those released from endothelial cells. The lower half of the table lists factors that may act on SGZ radial astrocytes. Domain I shows vascular-derived factors and signaling acting through the primary cilium. Domain II shows factors that could facilitate interaction between radial astrocytes and surrounding cells. Domain III shows neuronal input within the IML.

inhibitor protein (DBI) is secreted into the extracellular space by B1 cells and IPCs, but not neuroblasts, and competes with GABA for binding to its receptor. This leads to decreased inward Cl currents resulting in increased proliferation of the progenitor population (Alfonso et al., 2012). V-SVZ progenitors also receive inputs from nonneurogenic regions of the brain. The neurotransmitter dopamine is released into the V-SVZ from terminals of neuronal projections from the substantia nigra. Activation of dopamine D2-like receptors on IPCs increases their proliferation via an EGF-dependent mecha glinger et al., 2004; OKeeffe et al., 2009). Dopamine nism (Ho may also induce IPC proliferation and OB neurogenesis via D3 receptors (Kim et al., 2010). In addition, serotonin release from

raphe nuclei neurons into the V-SVZ has been shown to positively modulate neurogenesis, although it remains to be determined whether this effect is direct (Banasr et al., 2004). Stretching Out to the VasculatureB1 Cells Distal Domain An extensive vascular plexus runs parallel to the V-SVZ. A specialized end-foot in the basal process conforming to the distal domain of B1 cells (Figure 1) allows these NSCs to interact closely with endothelial cells (ECs). Indeed, ECs, and possibly factors derived from the circulation, support proliferation and self-renewal of V-SVZ progenitors in vitro (Shen et al., 2004). Clusters of dividing B1 cells and IPCs are associated with BVs at regions where the blood-brain barrier appears to be leaky
Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc. 701

Cell Stem Cell

Perspective
(Tavazoie et al., 2008). Blood-derived factors may directly access B1 cells and IPCs and regulate their proliferation. The chemokine SDF1 is expressed by endothelial and ependymal cells forming a gradient within the V-SVZ (Kokovay et al., 2010). High levels of SDF1 secreted by endothelial cells induce the recruitment of activated B1 cells and IPCs by chemotaxis into the vascular plexus. This effect is mediated by the induction of a6b1-integrin expression in these cell populations. Blockage of a6-integrins in progenitor cells results in the loss of adhesion to the vasculature and proliferation defects in vivo. Interestingly, the transition of IPCs into neuroblasts is accompanied by a decrease in the expression of b1-integrin, enabling the differentiating progeny to migrate away from their niche. Pigment epithelium-derived factor (PEDF) is also secreted by endothelial and ependymal cells in the V-SVZ. PEDF infusion into the lateral ventricles results in a signicant increase in the number of BrdU-label retaining cells and in the number of rez-Castillejo et al., 2006). PEDF interdividing GFAP+ cells (Ram acts synergistically with the Notch pathway to regulate self renewal in vitro by increasing EGFR expression (Andreu-Agullo et al., 2009). More recently, endothelial-derived betacellulin (BTC), an EGF-like growth factor, has been shown to stimulate progenitor proliferation in vitro. BTC infusion in vivo induces a signicant increase in V-SVZ proliferation and neurogenesis mez-Gaviro et al., 2012). Interestingly, the effects of BTC (Go are different from those of EGF, which like BTC induces a burst in progenitor proliferation but results in reduced neurogenesis and increased oligodendrogenesis (Doetsch et al., 2002; Gonzalez-Perez et al., 2009; Kuhn et al., 1997). This could be explained in part by the observation that BTC can bind and activate ErbB-4 and EGFR (ErbB-1) receptors present in neuroblasts and progenitor cells, respectively, whereas EGF mostly acts on IPCs. Notably, V-SVZ NSCs fail to regenerate IPCs and neuroblasts in Btc null mice following antimitotic treatment with mez-Gaviro et al., 2012). AraC (Go Taken together, the striking morphology of B1 cells allows them to interact with multiple environments. The proximal domain allows B1 cells with their apical surface, including the primary cilium, to receive signals from the CSF and from neighboring ependymal cells. The intermediate domain is likely a site for feedback signaling with IPCs and neuroblasts. B1 cells further access local (endothelial-derived) but also distant (blood-derived) secreted factors with their distal domain. However, some signaling pathways may function in multiple domains of B1 cells. For example, it is likely that Notch signaling, in addition to its function in the intermediate domain mediating interaction of IPCs and NSCs, also plays an important role in the proximal domain of these primary progenitors (e.g., between B1 cells and ependymal cells). It will be interesting to determine whether the response of NSCs is tuned to the domain, to the signaling pathway, or to both. Real-time imaging of B1 cells may allow future studies to monitor responses within different domains of NSCs and provide a more precise context under which the different domains contribute to specic NSC behavior. Domains of Adult Subgranular Zone Radial Astrocytes NSCs in the hippocampal SGZ, unlike those in the V-SVZ, are not in contact with the ventricular system. Nevertheless, radial astrocytes are regularly arrayed in the SGZ and along the dentate
702 Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc.

gyrus and are highly polarized, similar to the apical-basal organization observed in radial glia and B1 cells (Kempermann et al., 2003; Seri et al., 2001, 2004). Radial astrocytes also span at least three putative domains. We dene the proximal domain as the side of radial astrocytes that faces the hilus and includes contacts with blood vessels, a primary cilium, and lateral processes that frequently contact other radial astrocytes. The intermediate domain includes the cell body and the main shaft through the granule cell layer, where the cells have thin appendages intercalated among mature granule neurons. The distal domain is highly branched and contacts neuronal processes, synapses, and other glial cells in the inner molecular layer. Radial astrocytes may contact blood vessels in multiple compartments, but interactions with the vasculature have been only studied in their proximal domain (Figure 2). Snooping into the HilusRadial Astrocytes Proximal Domain The SGZ is intimately associated with a rich bed of endothelial cells. Interestingly, active angiogenesis and vascular remodeling occurs in parallel with neurogenesis (Palmer et al., 2000). This is in sharp contrast to the V-SVZ, where endothelial cell division is rare or undetectable (Shen et al., 2008; Tavazoie et al., 2008). Increased angiogenesis is associated with expression of the vascular endothelial growth factor (VEGF), which is also associated with increased progenitor proliferation and neurogenesis (Cao et al., 2004; Licht et al., 2011). However, whether VEGF acts directly on radial astrocytes remains unknown. Notably, physical exercise, a paradigm that induces proliferation of SOX2+ radial astrocytes (Suh et al., 2007), increases VEGF expression (Cao et al., 2004). The vasculature also secretes IGF1 and brain-derived neurotrophic factor (BDNF), which promote proliferation and differentiation of progenitor cells, n et al., 2009). respectively (Chen et al., 2005; Llorens-Mart The proximal domain of radial astrocytes also harbors a primary cilium, which has been shown to be essential for Shh signaling. Conditional deletion of Ift20 in GFAP+ radial astrocytes results in a signicant decrease in SGZ IPC proliferation and a concomitant impairment in spatial learning (Amador-Arjona et al., 2011). Notably, disruption of ciliogenesis during development results in decreased dentate gyrus Shh signaling and almost a complete absence of radial astrocytes during postnatal life (Breunig et al., 2008; Han et al., 2008). Similar defects in the establishment of radial astrocytes occur upon mutation of the Smoothened receptor, which is essential for Shh signaling (Han et al., 2008). These results indicate that Shh signaling through the primary cilium is essential for the transition from embryonic to adult NSCs in the SGZ. Shoulder to Shoulder with the ProgenyRadial Astrocytes Intermediate Domain The intermediate domain of SGZ radial astrocytes, which contains most of the cell body and the main shaft of the radial astrocyte process, contacts IPCs, populations of mature granule neurons, and possibly other neuronal cell types. The majority of IPCs are found next to radial astrocytes cell bodies (Kempermann et al., 2003; Seri et al., 2004), in what we here dene as part of the intermediate domain. However, there is some evidence that suggests that early IPCs (IPC1) are generated by asymmetric division of the radial astrocytes with a horizontal mitotic plane parallel to the SGZ (Encinas et al., 2011; Seri

Cell Stem Cell

Perspective
et al., 2004). Therefore, these initial IPCs may transiently interact with the proximal domain. However, proliferating IPC1 cells are observed in the intermediate domain suggesting that after they are produced, they rapidly translocate to the intermediate domain next to the radial astrocytes cell body. Recent work using lineage-tracing experiments of Hes5-expressing NSCs conrms that ASCL1+ IPC1 proliferate, but they apparently only divide once before converting into IPC2 cells that no longer express ASCL1 and have turned on the early neuronal marker doublecortin (DCX) (Lugert et al., 2012). Interestingly, this study suggests that further amplication of the lineage occurs by division of the DCX+/TBR2+ IPC2 cells and not by IPC1, as previously thought. The high expression of Hes5 (and also RBPJk) in radial astrocytes indicates active canonical Notch signaling in these NSCs (Ehm et al., 2010; Lugert et al., 2010). IPCs in turn are thought to express the Notch ligand Jagged-1 (Breunig et al., 2007; Lavado et al., 2010). Maintenance of quiescence among radial astrocytes might be controlled through a feedback mechanism by Jagged-1/Notch. Consistently, conditional deletion of RbpJk in adult GLAST+ SGZ radial astrocytes results in short-term expansion of the IPC pool and reduction in the number of radial astrocytes (Ehm et al., 2010). In contrast, conditional ablation of Notch1 in nestin+ cells results in a strong reduction in the number of radial astrocytes but without a transient increase in IPCs (Ables et al., 2010). This difference might be explained by compensatory effects of other Notch receptors or by noncanonical Notch signaling. Notably, the expression of SOX2, a transcription factor essential for maintenance of radial astrocytes (Favaro et al., 2009), is regulated by Notch signaling via RBPJk, which directly targets the Sox2 promoter (Ehm et al., 2010). SOX2, in turn, inhibits Wnt-mediated activation of the neuronal fate determinant NeuroD (Kuwabara et al., 2009), and it may directly target Shh expression (Favaro et al., 2009). Thus, reduction in Notch signaling and SOX2 expression may be required for the induction of proneural genes like Ascl1 and NeuroD to stimulate the transition from radial astrocytes to IPCs. Noggin is expressed in dentate gyrus granule cells, the hilus, and SGZ radial astrocytes, and its expression is regulated in a cell-autonomous manner by the RNA-binding protein FXR2 (Bonaguidi et al., 2008; Guo et al., 2011). Physical exercise reduces BMP4 and increases noggin expression (Gobeske et al., 2009), and overexpression of noggin increases the number of dividing GFAP+ radial astrocytes (Bonaguidi et al., 2008). This suggests that BMPs inhibit NSC proliferation. Another recent study indicates that BMP controls NSC quiescence in the SGZ (Mira et al., 2010). Nondividing radial SOX2+ cells show high levels of activated SMAD1, a specic mediator of BMP signaling, whereas dividing progenitors do not. Consistent with a role for BMPs in NSC quiescence, noggin infusion results in an initial increase in neuronal production and a concomitant reduction of SOX2+ progenitors and neuronal production at later time points. Conditional ablation of the receptor BmprIa or of Smad4 also results in decreased quiescence among NSCs. The shaft of radial astrocytes in the intermediate domain is also intimately associated with mature granule neurons. Direct contact or exchange of signals with neurons may contribute to radial astrocyte regulation. It is possible that radial astrocytes may be able to sense neighboring network activity associated with the column of granule neurons surrounding their radial shaft, possibly through neurotransmitter-mediated spillover or extracellular potassium. Neuronal activityinduced for example by learning and memory, or seizures (Inokuchi, 2011)may directly activate proliferation of NSCs in the dentate gyrus. Hippocampal induction of long-term potentiation (LTP), a widely studied mechanism associated with memory formation (Lynch, 2004), results in increased SGZ proliferation and neuronal differentiation (Bruel-Jungerman et al., 2006). General activation of granule neurons following seizures results in increased SGZ neurogenesis, although the precise mechanisms of this effect remain elusive (reviewed in [Kokaia, 2011]). One study suggests that Shh signaling rapidly increases following electroconvulsive treatment (ECT) and that it may be required for seizure-induced SGZ proliferation (Banerjee et al., 2005). Interestingly, after ECT, granule neurons transiently express high levels of the DNA repair protein GADD45b, a process that requires NMDA-type glutamate receptors (NMDARs) (Ma et al., 2009). Loss of Gadd45b expression partially blocks ECT-induced SGZ progenitor proliferation. This study also shows that Gadd45b induces the expression of genes known to modulate neurogenesis such as Bdnf and broblast growth factor 1 (Fgf1) by DNA-demethylation at promoter regions. Seizures result in nonphysiological high levels of neuronal activity. Whether physiological levels of neuronal ring, possibly in a localized manner, affect individual radial astrocyte proliferation remains to be determined. Neuronal activity may also exert effects on neurogenesis through the distal or proximal domains of radial astrocytes. Hanging on WiresRadial Astrocytes Distal Domain In addition to a shaft that contacts granule neurons, radial astrocytes have an elaborate and extensive set of thin branches and lamellae in the inner molecular layer. Little is known about the exchange of signals that take place here, but it would be surprising for such an elaborate terminal arbor of the radial process not to function in the regulation of these NSCs. The inner molecular layer of the dentate gyrus receives GABAergic and glutamatergic inputs from hilar interneurons and Mossy cells, rster et al., 2006). respectively (Fo Several studies have investigated the role of neurotransmitters in the regulation of SGZ neurogenesis. There is evidence that GABA promotes SGZ progenitor differentiation, but it remains unknown whether GABAergic inputs have a direct effect on the proliferation of SGZ progenitor cells, as shown for the V-SVZ. GABA does not seem to stimulate radial astrocyte responses directly, but it induces depolarization of IPCs resulting in increased Ca2+ inux and enhanced NeuroD expression. This, in turn, inhibits progenitor proliferation and promotes neuronal differentiation (Tozuka et al., 2005; Wang et al., 2005). Electrical stimulation of the perforant pathway activates granule cells and GABAergic interneurons and directly induces inward GABAergic currents in IPCs. GABAergic terminals, containing vesicular GABA transporter (VGAT), are closely associated with IPCs (Tozuka et al., 2005). Recent studies have shown that corelease with SDF1 facilitates GABA transmission from local GABAergic basket interneurons (Bhattacharyya et al., 2008; Kolodziej et al., 2008). These ndings suggest that SGZ progenitors receive functional inputs from hippocampal GABAergic interneurons that promote neuronal differentiation, but it remains unknown whether this is localized uniquely to the distal domain.
Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc. 703

Cell Stem Cell

Perspective
Glutamate also induces depolarization of IPCs, inhibits expression of Hes1 and Id2, and increases expression of NeuroD (Deisseroth et al., 2004). This process is regulated by voltage-gated Ca2+ channels and is mediated through NMDARs. However, functional NMDARs are not detectable in SGZ progenitor cells in tissue slices (Tozuka et al., 2005). It will be important to determine whether the response of radial astrocytes to glutamate vary depending on the domains of these cells exposed to this neurotransmitter. In summary, the three domains of radial astrocytes span three anatomical layers: the SGZ, the granule cell layer, and the inner molecular layer. The proximal domain in the SGZ has a primary cilium and interacts with the vasculature. Similarly to V-SVZ B1 cells, the intermediate domain of radial astrocytes directly interacts with the progeny. The intermediate and distal domains, spanning a long appendage and branches in the inner molecular layer, may expose radial astrocytes to neuronal networks and their level of activity. The domain organization of NSCs may help explain how radial astrocytes within the dentate gyrus might be able to integrate local activity from granule neurons right next to the primary shaft of these cells versus more widespread inputs arising from parallel bers in the inner molecular layer and the hilus. There likely exists a local topographic organization to regulate neurogenesis at the level of individual radial astrocytes. This remains an interesting question for future research. Self-Renewal of Adult Neural Stem Cells and Age-Related Changes The above suggests that domains in NSCs are tuned to specic compartments within their niche. In order to determine how NSCs integrate this information, the behavior of these cells in vivo needs to be understood. However, the patterns of proliferation of adult NSCs remain unknown or are highly controversial. Selfrenewal is considered a dening property of stem cells and important for their long-term retention in adults. Thus, a major question is whether adult NSCs have this classical property of somatic stem cells. NSCs were initially identied in vitro by their ability to generate free-oating aggregates (neurospheres) in response to growth factors (Weiss et al., 1996). So far, most evidence for self-renewal is based on these in vitro assays. A major limiting factor to study NSC behavior in vivo is the lack of markers that exclusively identify these primary progenitors (Kriegstein and Alvarez-Buylla, 2009). Thus, although BrdU-label retaining experiments in vivo suggest the presence of cells with stem cell characteristics (Bonaguidi et al., 2008; Chiasson et al., 1999; Doetsch et al., 1999), self-renewal has not been directly demonstrated in vivo in the V-SVZ, and this issue remains controversial in the SGZ (Bonaguidi et al., 2011; Encinas et al., 2011; Lugert et al., 2012). Encinas and collaborators (Encinas et al., 2011) suggest that SGZ NSCs do not self-renew but are consumed with age (the disposable stem cell hypothesis). In this scenario, activated radial astrocytes undergo up to three sequential asymmetric divisions that generate IPCs. The NSCs then terminally differentiate into mature astrocytes. This study also suggests that IPCs divide multiple times, which is in contrast to earlier estimations of one to two rounds (Seri et al., 2004). Contrary to the disposable stem cell hypothesis and the depletion of NSCs with time, two other recent studies indicate that stable populations of radial
704 Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc.

astrocytes are maintained in the SGZ for extended periods of time. Lugert et al. (2012) nds that a cohort of Hes5 expressing NSCs, which contributes continually to neurogenesis over this time, remains in the SGZ for up to 100 days, suggesting some level of self-renewal. Contrary to the nding in Encinas et al. (2011), lineage tracing of these Hes5+ NSCs does not reveal an increase in the generation of mature astrocytes during this period, and the stoichiometry of labeled NSCs, IPCs, and neurons produced also suggests limited IPC amplication. Bonaguidi et al. (2011) have followed the lineage of individual nestin+ radial astrocytes in the SGZ and observed that a small number of these NSCs self-renew symmetrically with a larger population self-renewing asymmetrically. Some SGZ NSCs are multipotent, generating both astrocytes and neurons, whereas others are unipotent and generate only neurons or astrocytes, therefore suggesting that SGZ NSCs are heterogeneous in their behavior (Bonaguidi et al., 2011). This study also suggests that SGZ IPCs divide up to ve times before differentiating into young neurons. Interestingly, asymmetrically dividing radial astrocytes generate a highly proliferative IPC and a radial astrocyte, which, in contrast to the disposable model, returns to a quiescent state (Bonaguidi et al., 2011). Thus, the quiescent state of radial astrocytes may be reversible, with multiple cycles of activation and quiescence. Interestingly, upon conditional Pten deletion, radial astrocytes differentiate into postmitotic astrocytes after a transient increase in symmetric self-renewing divisions. Irrespective of whether adult NSCs self-renew or are consumed, it is clear that neurogenesis in the adult V-SVZ and SGZ decrease with age (Conover and Shook, 2011; Jessberger and Gage, 2008), and it will be interesting to see how this controversy resolves to understand how the behavior of NSCs is regulated and changes with aging. Aging is associated with a decline in cognitive function, impairments in learning and memory, and higher susceptibility to neurodegenerative disorders, a process that is paralleled by the reduction in SGZ neurogenesis (Jessberger and Gage, 2008). Interestingly, this decline may in part be due to changes in factors present in the circulation (Villeda et al., 2011). In heterochronic parabionts, where the vasculature of young and aged mice is surgically joined, the aged milieu decreases neurogenesis in young adults, whereas the young milieu has some rejuvenating effects on the aged SGZ. The chemokine CCL-11 is increased in the plasma and CSF of elderly humans and has been suggested as one of the molecules that could decrease SGZ proliferation. Wnt3, which is secreted from hippocampal astrocytes and promotes neuronal differentiation by activation of NeuroD and Dcx expression, decreases with age (Kuwabara et al., 2009; Okamoto et al., 2011). Recent data further suggest that the age-related decline in SGZ neurogenesis is due to transition of radial astrocytes into quiescence rather than a loss of the stem cell population. In line with this, and contrary to the depletion of NSCs during aging, physical exercise and seizures, both known to promote neurogenesis in the young adult hippocampus, can (partially) reverse age-induced changes (Kronenberg et al., 2006; Lugert et al., 2010; Okamoto et al., 2011). Thus, the quiescent state might remain reversible in the aged SGZ. However, whether activation of quiescent stem cells in the aged SGZ is only transient and whether it might lead to subsequent stem cell exhaustion is unknown. On the other

Cell Stem Cell

Perspective
hand, activation of quiescent cells itself might not be sufcient to restore neurogenesis (Kronenberg et al., 2006; Rao et al., 2008). Neurogenesis also declines in an age-dependent manner in the V-SVZ. It has been suggested that aging leads to impairments of ne odor discrimination in mice due to a decrease in the number of newly generated interneurons between 2 and 24 months of age (Enwere et al., 2004). This is accompanied by a diminution in EGFR signaling and proliferation and thinning of the V-SVZ. Some astrocytes acquire ependymal features like 9 + 2 motile cilia and apical placement of mitochondria and are incorporated into the ependymal layer (Conover and Shook, 2011; Luo et al., 2006). Interestingly, only the dorsolateral aspect of the V-SVZ remains neurogenic in aged mice. This might affect the generation of specic neuronal subtypes in the aged brain, because V-SVZ NSCs are heterogeneous, generating at least six different subtypes of OB interneurons depending on their position along dorsal-ventral and anterior-posterior axes (Merkle et al., 2007). Although the number of label-retaining and proliferating cells in vivo seems to decline during aging, the percentage of proliferating cells does not (Bouab et al., 2011). Similarly, the number of V-SVZ neurosphere-forming cells declines with age, but nestin+ cells appear to be equally competent of producing neurospheres (Ahlenius et al., 2009; Enwere et al., 2004). Cell intrinsic factors like mTERT, p16, and p21 have been shown to contribute to the decrease in neurogenesis associated with aging (Conover and Shook, 2011). Deciency of p21 increases the numbers of BrdU-label retaining cells in young adult V-SVZ but leads to exhaustion of these cells in old mice (Kippin et al., 2005). It is not clear whether the age-related decline in V-SVZ neurogenesis is due to terminal differentiation, a limited intrinsic capacity of NSCs to self-renew and/or molecular and cellular changes in the microenvironment suppressing proliferation and/or promoting senescence or quiescence. Cell-autonomous quiescence might be a mechanism to prevent tumor formation by minimizing the risk of accumulated mutations and DNA damage. This requires a tight balance of expression of proto-oncogenes and tumor suppressors. For example, the proto-oncogene BMI-1 promotes self-renewal by repressing the tumor suppressors p21, p16, and p19 (Fasano et al., 2009; Favaro et al., 2009; Molofsky et al., 2006). In the aged SVZ, increased levels of p16 suppress stem cell activity but possibly protect from tumor formation. Indeed, it has been shown that adult V-SVZ NSCs could serve as cells of origin for some gliomas (Jacques et al., 2010). Interestingly, p16 is frequently lost in glioblastoma (Furnari et al., 2007; Wiedemeyer et al., 2008). In sum, it remains unclear whether NSCs self-renew in vivo or are consumed with age and whether extrinsic factorswithin the niche or systemicdecrease NSC proliferation in the aged brain. These are important questions to understand how different compartments in the niche affect the behavior of NSCs. Unlike other tissues where robust regenerative responses can be mounted even in advanced age, neurogenesis in the V-SVZ and SGZ decreases with age. Little is known about how the different compartments of the niche are altered with aging. The integrity of B1 cell proximal domain is likely affected during aging in the ventral V-SVZ. Ventral stenosis, with loss of ependymal cells and fusion of the lateral and medial walls in old mice, has been associated to decrease neurogenesis in this ventral region (Luo et al., 2006). How NSC morphology and signaling within NSC domains is affected with age remains an interesting area for future research. Conclusions Adult NSCs (B1 astrocytes in V-SVZ and radial astrocytes in SGZ) have processes that most likely allow them to sample specic compartments of their niche. Therefore, these NSCs can respond to both local factors next to their cell body as well as signals derived farther away such as those from blood vessels, the ventricles, or neuronal plexus. Moreover, NSC domains allow them to receive signals that are not accessible to other cells within the niche (e.g., CSF-borne factors may not be available to IPCs or neuroblasts within the V-SVZ niche). Although it has been shown that NSCs respond to multiple factors, complete insights into NSC behavior will require a better understanding of the compartmentalized signaling that occurs in vivo. How B1 cells and radial astrocytes are induced to proliferate or to remain quiescent likely depends on the subcellular integration of multiple signaling pathways arising from the primary cilium, the end-feet around blood vessels, the cell body, and their main or lateral processes. Thus, NSCs bridge multiple locations of the niche, allowing them to receive a broad spectrum of signals, which, combined, may orchestrate their behavior. Although we have reviewed emerging evidence suggesting how some major signaling pathways may act primarily through one of the proposed domains, as mentioned above, it is entirely possible that the same signaling pathway could also act on other parts of the same cell. The response could differ depending on the region receiving the signal, highlighting the relevance of studying NSCs within their niche. Real-time imaging of intracellular signaling and simultaneous observation of the resulting NSC behavior may provide important new insights on the basic regulation of the primary progenitors, which make adult neurogenesis possible.
ACKNOWLEDGMENTS The authors thank Robert A. Lindquist, Timothy R. Stowe, Shawn F. Sorrells, and Cheuk Ka Tong for critical discussion of the review. Special thanks also to Kenneth X. Probst for preparation of the illustrations. Due to limited space of this review, we could not cover all the exciting contributions to the eld. The Alvarez-Buylla laboratory is funded by the US National Institutes of Health (NS28478 and HD32116) and the John G. Bowes Research Fund. A.A.-B. is the Heather and Melanie Muss Endowed Chair of Neurological Surgery at UCSF. L.C.F. is supported by the Helen Hay Whitney Foundation, and K.O. by the German Research Foundation (Deutsche Forschungsgemeinschaft). REFERENCES Ables, J.L., Decarolis, N.A., Johnson, M.A., Rivera, P.D., Gao, Z., Cooper, D.C., Radtke, F., Hsieh, J., and Eisch, A.J. (2010). Notch1 is required for maintenance of the reservoir of adult hippocampal stem cells. J. Neurosci. 30, 1048410492. Aguirre, A., Rubio, M.E., and Gallo, V. (2010). Notch and EGFR pathway interaction regulates neural stem cell number and self-renewal. Nature 467, 323327. Ahlenius, H., Visan, V., Kokaia, M., Lindvall, O., and Kokaia, Z. (2009). Neural stem and progenitor cells retain their potential for proliferation and differentiation into functional neurons despite lower number in aged brain. J. Neurosci. 29, 44084419.

Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc. 705

Cell Stem Cell

Perspective
Alfonso, J., Le Magueresse, C., Zuccotti, A., Khodosevich, K., and Monyer, H. (2012). Diazepam binding inhibitor promotes progenitor proliferation in the postnatal SVZ by reducing GABA signaling. Cell Stem Cell 10, 7687. Altman, J. (2011). The Discovery of Adult Mammalian Neurogenesis. In Neurogenesis in the Adult Brain I: Neurobiology, T. Seki, K. Sawamoto, J.M. Parent, and A. Alvarez-Buylla, eds. (Tokyo: Springer), pp. 346. a-Verdugo, J.M., Mateo, A.S., and Merchant-Larios, Alvarez-Buylla, A., Garc H. (1998). Primary neural precursors and intermitotic nuclear migration in the ventricular zone of adult canaries. J. Neurosci. 18, 10201037. Amador-Arjona, A., Elliott, J., Miller, A., Ginbey, A., Pazour, G.J., Enikolopov, G., Roberts, A.J., and Terskikh, A.V. (2011). Primary cilia regulate proliferation of amplifying progenitors in adult hippocampus: implications for learning and memory. J. Neurosci. 31, 99339944. , C., Morante-Redolat, J.M., Delgado, A.C., and Farin as, I. Andreu-Agullo (2009). Vascular niche factor PEDF modulates Notch-dependent stemness in the adult subependymal zone. Nat. Neurosci. 12, 15141523. Banasr, M., Hery, M., Printemps, R., and Daszuta, A. (2004). Serotonininduced increases in adult cell proliferation and neurogenesis are mediated through different and common 5-HT receptor subtypes in the dentate gyrus and the subventricular zone. Neuropsychopharmacology 29, 450460. Banerjee, S.B., Rajendran, R., Dias, B.G., Ladiwala, U., Tole, S., and Vaidya, V.A. (2005). Recruitment of the Sonic hedgehog signalling cascade in electroconvulsive seizure-mediated regulation of adult rat hippocampal neurogenesis. Eur. J. Neurosci. 22, 15701580. Beckervordersandforth, R., Tripathi, P., Ninkovic, J., Bayam, E., Lepier, A., Stempfhuber, B., Kirchhoff, F., Hirrlinger, J., Haslinger, A., Lie, D.C., et al. (2010). In vivo fate mapping and expression analysis reveals molecular hallmarks of prospectively isolated adult neural stem cells. Cell Stem Cell 7, 744758. Bhattacharyya, B.J., Banisadr, G., Jung, H., Ren, D., Cronshaw, D.G., Zou, Y., and Miller, R.J. (2008). The chemokine stromal cell-derived factor-1 regulates GABAergic inputs to neural progenitors in the postnatal dentate gyrus. J. Neurosci. 28, 67206730. Bonaguidi, M.A., Peng, C.Y., McGuire, T., Falciglia, G., Gobeske, K.T., Czeisler, C., and Kessler, J.A. (2008). Noggin expands neural stem cells in the adult hippocampus. J. Neurosci. 28, 91949204. Bonaguidi, M.A., Wheeler, M.A., Shapiro, J.S., Stadel, R.P., Sun, G.J., Ming, G.L., and Song, H. (2011). In vivo clonal analysis reveals self-renewing and multipotent adult neural stem cell characteristics. Cell 145, 11421155. Bonaguidi, M.A., Song, J., Ming, G.L., and Song, H. (2012). A unifying hypothesis on mammalian neural stem cell properties in the adult hippocampus. Curr Opin Neurobiol. Published online April 13, 2012. 10.1016/j.conb.2012.03.013. rard, K., and Fernandes, Bouab, M., Paliouras, G.N., Aumont, A., Forest-Be K.J. (2011). Aging of the subventricular zone neural stem cell niche: evidence for quiescence-associated changes between early and mid-adulthood. Neuroscience 173, 135149. Breunig, J.J., Silbereis, J., Vaccarino, F.M., Sestan, N., and Rakic, P. (2007). Notch regulates cell fate and dendrite morphology of newborn neurons in the postnatal dentate gyrus. Proc. Natl. Acad. Sci. USA 104, 2055820563. Breunig, J.J., Sarkisian, M.R., Arellano, J.I., Morozov, Y.M., Ayoub, A.E., Sojitra, S., Wang, B., Flavell, R.A., Rakic, P., and Town, T. (2008). Primary cilia regulate hippocampal neurogenesis by mediating sonic hedgehog signaling. Proc Natl Acad Sci 105, 1312713132. Bruel-Jungerman, E., Davis, S., Rampon, C., and Laroche, S. (2006). Long-term potentiation enhances neurogenesis in the adult dentate gyrus. J. Neurosci. 26, 58885893. Byrd, C.A., and Brunjes, P.C. (2001). Neurogenesis in the olfactory bulb of adult zebrash. Neuroscience 105, 793801. Cao, L., Jiao, X., Zuzga, D.S., Liu, Y., Fong, D.M., Young, D., and During, M.J. (2004). VEGF links hippocampal activity with neurogenesis, learning and memory. Nat. Genet. 36, 827835. Chapouton, P., Jagasia, R., and Bally-Cuif, L. (2007). Adult neurogenesis in non-mammalian vertebrates. Bioessays 29, 745757. Chapouton, P., Skupien, P., Hesl, B., Coolen, M., Moore, J.C., Madelaine, R., Kremmer, E., Faus-Kessler, T., Blader, P., Lawson, N.D., and Bally-Cuif, L. (2010). Notch activity levels control the balance between quiescence and recruitment of adult neural stem cells. J. Neurosci. 30, 79617974. Chen, J., Zacharek, A., Zhang, C., Jiang, H., Li, Y., Roberts, C., Lu, M., Kapke, A., and Chopp, M. (2005). Endothelial nitric oxide synthase regulates brainderived neurotrophic factor expression and neurogenesis after stroke in mice. J. Neurosci. 25, 23662375. Chiasson, B.J., Tropepe, V., Morshead, C.M., and van der Kooy, D. (1999). Adult mammalian forebrain ependymal and subependymal cells demonstrate proliferative potential, but only subependymal cells have neural stem cell characteristics. J. Neurosci. 19, 44624471. Colak, D., Mori, T., Brill, M.S., Pfeifer, A., Falk, S., Deng, C., Monteiro, R., tz, M. (2008). Adult neurogenesis Mummery, C., Sommer, L., and Go requires Smad4-mediated bone morphogenic protein signaling in stem cells. J. Neurosci. 28, 434446. Conover, J.C., and Shook, B.A. (2011). Aging of the Subventricular Zone Neural. Stem Cell Niche. Aging Dis. 2, 149163. Deisseroth, K., Singla, S., Toda, H., Monje, M., Palmer, T.D., and Malenka, R.C. (2004). Excitation-neurogenesis coupling in adult neural stem/progenitor cells. Neuron 42, 535552. , I., Lim, D.A., Garc a-Verdugo, J.M., and Alvarez-Buylla, A. Doetsch, F., Caille (1999). Subventricular zone astrocytes are neural stem cells in the adult mammalian brain. Cell 97, 703716. Doetsch, F., Petreanu, L., Caille, I., Garcia-Verdugo, J.M., and Alvarez-Buylla, A. (2002). EGF converts transit-amplifying neurogenic precursors in the adult brain into multipotent stem cells. Neuron 36, 10211034. ritz, C., Covic, M., Scha ffner, I., Schwarz, T.J., Karaca, E., Ehm, O., Go Kempkes, B., Kremmer, E., Pfrieger, F.W., Espinosa, L., et al. (2010). RBPJkappa-dependent signaling is essential for long-term maintenance of neural stem cells in the adult hippocampus. J. Neurosci. 30, 1379413807. Encinas, J.M., Michurina, T.V., Peunova, N., Park, J.H., Tordo, J., Peterson, D.A., Fishell, G., Koulakov, A., and Enikolopov, G. (2011). Division-coupled astrocytic differentiation and age-related depletion of neural stem cells in the adult hippocampus. Cell Stem Cell 8, 566579. Enwere, E., Shingo, T., Gregg, C., Fujikawa, H., Ohta, S., and Weiss, S. (2004). Aging results in reduced epidermal growth factor receptor signaling, diminished olfactory neurogenesis, and decits in ne olfactory discrimination. J. Neurosci. 24, 83548365. Fasano, C.A., Phoenix, T.N., Kokovay, E., Lowry, N., Elkabetz, Y., Dimos, J.T., Lemischka, I.R., Studer, L., and Temple, S. (2009). Bmi-1 cooperates with Foxg1 to maintain neural stem cell self-renewal in the forebrain. Genes Dev. 23, 561574. Favaro, R., Valotta, M., Ferri, A.L., Latorre, E., Mariani, J., Giachino, C., Lancini, C., Tosetti, V., Ottolenghi, S., Taylor, V., and Nicolis, S.K. (2009). Hippocampal development and neural stem cell maintenance require Sox2-dependent regulation of Shh. Nat. Neurosci. 12, 12481256. ng, M., and Fernando, R.N., Eleuteri, B., Abdelhady, S., Nussenzweig, A., Anda Ernfors, P. (2011). Cell cycle restriction by histone H2AX limits proliferation of adult neural stem cells. Proc. Natl. Acad. Sci. USA 108, 58375842. Filippov, V., Kronenberg, G., Pivneva, T., Reuter, K., Steiner, B., Wang, L.P., Yamaguchi, M., Kettenmann, H., and Kempermann, G. (2003). Subpopulation of nestin-expressing progenitor cells in the adult murine hippocampus shows electrophysiological and morphological characteristics of astrocytes. Mol. Cell. Neurosci. 23, 373382. rster, E., Zhao, S., and Frotscher, M. (2006). Laminating the hippocampus. Fo Nat. Rev. Neurosci. 7, 259267. Furnari, F.B., Fenton, T., Bachoo, R.M., Mukasa, A., Stommel, J.M., Stegh, A., Hahn, W.C., Ligon, K.L., Louis, D.N., Brennan, C., et al. (2007). Malignant astrocytic glioma: genetics, biology, and paths to treatment. Genes Dev. 21, 26832710. Gajera, C.R., Emich, H., Lioubinski, O., Christ, A., Beckervordersandforth tz, M., KemperBonk, R., Yoshikawa, K., Bachmann, S., Christensen, E.I., Go mann, G., et al. (2010). LRP2 in ependymal cells regulates BMP signaling in the adult neurogenic niche. J. Cell Sci. 123, 19221930.

706 Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc.

Cell Stem Cell

Perspective
n, S., Flames, N., Collado, L., Deslis, E., and a-Verdugo, J.M., Ferro Garc Font, E. (2002). The proliferative ventricular zone in adult vertebrates: a comparative study using reptiles, birds, and mammals. Brain Res. Bull. 57, 765775. Gobeske, K.T., Das, S., Bonaguidi, M.A., Weiss, C., Radulovic, J., Disterhoft, J.F., and Kessler, J.A. (2009). BMP signaling mediates effects of exercise on hippocampal neurogenesis and cognition in mice. PLoS ONE 4, e7506. Goldman, S.A., and Nottebohm, F. (1983). Neuronal production, migration, and differentiation in a vocal control nucleus of the adult female canary brain. Proc. Natl. Acad. Sci. USA 80, 23902394. mez-Gaviro, M.V., Scott, C.E., Sesay, A.K., Matheu, A., Booth, S., Galichet, Go C., and Lovell-Badge, R. (2012). Betacellulin promotes cell proliferation in the neural stem cell niche and stimulates neurogenesis. Proc. Natl. Acad. Sci. USA 109, 13171322. Gonzalez-Perez, O., Romero-Rodriguez, R., Soriano-Navarro, M., GarciaVerdugo, J.M., and Alvarez-Buylla, A. (2009). Epidermal growth factor induces the progeny of subventricular zone type B cells to migrate and differentiate into oligodendrocytes. Stem Cells 27, 20322043. Guo, W., Zhang, L., Christopher, D.M., Teng, Z.Q., Fausett, S.R., Liu, C., George, O.L., Klingensmith, J., Jin, P., and Zhao, X. (2011). RNA-binding protein FXR2 regulates adult hippocampal neurogenesis by reducing Noggin expression. Neuron 70, 924938. Han, Y.G., Spassky, N., Romaguera-Ros, M., Garcia-Verdugo, J.M., Aguilar, A., Schneider-Maunoury, S., and Alvarez-Buylla, A. (2008). Hedgehog signaling and primary cilia are required for the formation of adult neural stem cells. Nat. Neurosci. 11, 277284. glinger, G.U., Rizk, P., Muriel, M.P., Duyckaerts, C., Oertel, W.H., Caille, I., Ho and Hirsch, E.C. (2004). Dopamine depletion impairs precursor cell proliferation in Parkinson disease. Nat. Neurosci. 7, 726735. Huang, X., Liu, J., Ketova, T., Fleming, J.T., Grover, V.K., Cooper, M.K., Litingtung, Y., and Chiang, C. (2010). Transventricular delivery of Sonic hedgehog is essential to cerebellar ventricular zone development. Proc. Natl. Acad. Sci. USA 107, 84228427. Ihrie, R.A., and Alvarez-Buylla, A. (2011). Lake-front property: a unique germinal niche by the lateral ventricles of the adult brain. Neuron 70, 674686. Ihrie, R.A., Shah, J.K., Harwell, C.C., Levine, J.H., Guinto, C.D., Lezameta, M., Kriegstein, A.R., and Alvarez-Buylla, A. (2011). Persistent sonic hedgehog signaling in adult brain determines neural stem cell positional identity. Neuron 71, 250262. Imayoshi, I., Sakamoto, M., Yamaguchi, M., Mori, K., and Kageyama, R. (2010). Essential roles of Notch signaling in maintenance of neural stem cells in developing and adult brains. J. Neurosci. 30, 34893498. Inokuchi, K. (2011). Adult neurogenesis and modulation of neural circuit function. Curr. Opin. Neurobiol. 21, 360364. Irvin, D.K., Nakano, I., Paucar, A., and Kornblum, H.I. (2004). Patterns of Jagged1, Jagged2, Delta-like 1 and Delta-like 3 expression during late embryonic and postnatal brain development suggest multiple functional roles in progenitors and differentiated cells. J. Neurosci. Res. 75, 330343. Jablonska, B., Aguirre, A., Raymond, M., Szabo, G., Kitabatake, Y., Sailor, K.A., Ming, G.L., Song, H., and Gallo, V. (2010). Chordin-induced lineage plasticity of adult SVZ neuroblasts after demyelination. Nat. Neurosci. 13, 541550. Jacques, T.S., Swales, A., Brzozowski, M.J., Henriquez, N.V., Linehan, J.M., Mirzadeh, Z., O Malley, C., Naumann, H., Alvarez-Buylla, A., and Brandner, S. (2010). Combinations of genetic mutations in the adult neural stem cell compartment determine brain tumour phenotypes. EMBO J. 29, 222235. Jessberger, S., and Gage, F.H. (2008). Stem-cell-associated structural and functional plasticity in the aging hippocampus. Psychol. Aging 23, 684691. Kempermann, G., Gast, D., Kronenberg, G., Yamaguchi, M., and Gage, F.H. (2003). Early determination and long-term persistence of adult-generated new neurons in the hippocampus of mice. Development 130, 391399. Kim, Y., Wang, W.Z., Comte, I., Pastrana, E., Tran, P.B., Brown, J., Miller, R.J., r, Z., and Szele, F.G. (2010). Dopamine stimulation of Doetsch, F., Molna postnatal murine subventricular zone neurogenesis via the D3 receptor. J. Neurochem. 114, 750760. Kippin, T.E., Martens, D.J., and van der Kooy, D. (2005). p21 loss compromises the relative quiescence of forebrain stem cell proliferation leading to exhaustion of their proliferation capacity. Genes Dev. 19, 756767. Kokaia, M. (2011). Seizure-induced neurogenesis in the adult brain. Eur. J. Neurosci. 33, 11331138. Kokovay, E., Goderie, S., Wang, Y., Lotz, S., Lin, G., Sun, Y., Roysam, B., Shen, Q., and Temple, S. (2010). Adult SVZ lineage cells home to and leave the vascular niche via differential responses to SDF1/CXCR4 signaling. Cell Stem Cell 7, 163173. llt, V., Kolodziej, A., Schulz, S., Guyon, A., Wu, D.F., Pfeiffer, M., Odemis, V., Ho and Stumm, R. (2008). Tonic activation of CXC chemokine receptor 4 in immature granule cells supports neurogenesis in the adult dentate gyrus. J. Neurosci. 28, 44884500. Kopan, R., and Ilagan, M.X. (2009). The canonical Notch signaling pathway: unfolding the activation mechanism. Cell 137, 216233. Kriegstein, A., and Alvarez-Buylla, A. (2009). The glial nature of embryonic and adult neural stem cells. Annu. Rev. Neurosci. 32, 149184. Kronenberg, G., Bick-Sander, A., Bunk, E., Wolf, C., Ehninger, D., and Kempermann, G. (2006). Physical exercise prevents age-related decline in precursor cell activity in the mouse dentate gyrus. Neurobiol. Aging 27, 15051513. Kuhn, H.G., Winkler, J., Kempermann, G., Thal, L.J., and Gage, F.H. (1997). Epidermal growth factor and broblast growth factor-2 have different effects on neural progenitors in the adult rat brain. J. Neurosci. 17, 58205829. Kuwabara, T., Hsieh, J., Muotri, A., Yeo, G., Warashina, M., Lie, D.C., Moore, L., Nakashima, K., Asashima, M., and Gage, F.H. (2009). Wnt-mediated activation of NeuroD1 and retro-elements during adult neurogenesis. Nat. Neurosci. 12, 10971105. Lavado, A., Lagutin, O.V., Chow, L.M., Baker, S.J., and Oliver, G. (2010). Prox1 is required for granule cell maturation and intermediate progenitor maintenance during brain neurogenesis. PLoS Biol. 8, 8. Lehtinen, M.K., Zappaterra, M.W., Chen, X., Yang, Y.J., Hill, A.D., Lun, M., Maynard, T., Gonzalez, D., Kim, S., Ye, P., et al. (2011). The cerebrospinal uid provides a proliferative niche for neural progenitor cells. Neuron 69, 893905. Licht, T., Goshen, I., Avital, A., Kreisel, T., Zubedat, S., Eavri, R., Segal, M., Yirmiya, R., and Keshet, E. (2011). Reversible modulations of neuronal plasticity by VEGF. Proc. Natl. Acad. Sci. USA 108, 50815086. a-Verdugo, J.M., Lim, D.A., Tramontin, A.D., Trevejo, J.M., Herrera, D.G., Garc and Alvarez-Buylla, A. (2000). Noggin antagonizes BMP signaling to create a niche for adult neurogenesis. Neuron 28, 713726. Liu, X., Wang, Q., Haydar, T.F., and Bordey, A. (2005). Nonsynaptic GABA signaling in postnatal subventricular zone controls proliferation of GFAPexpressing progenitors. Nat. Neurosci. 8, 11791187. n, I., and Trejo, J.L. (2009). Mechanisms n, M., Torres-Alema Llorens-Mart mediating brain plasticity: IGF1 and adult hippocampal neurogenesis. Neuroscientist 15, 134148. tz, M., Haas, Lugert, S., Basak, O., Knuckles, P., Haussler, U., Fabel, K., Go C.A., Kempermann, G., Taylor, V., and Giachino, C. (2010). Quiescent and active hippocampal neural stem cells with distinct morphologies respond selectively to physiological and pathological stimuli and aging. Cell Stem Cell 6, 445456. ller, M., Giachino, C., and Taylor, V. Lugert, S., Vogt, M., Tchorz, J.S., Mu (2012). Homeostatic neurogenesis in the adult hippocampus does not involve amplication of Ascl1(high) intermediate progenitors. Nat. Commun. 3, 670. Luo, J., Daniels, S.B., Lennington, J.B., Notti, R.Q., and Conover, J.C. (2006). The aging neurogenic subventricular zone. Aging Cell 5, 139152. Lynch, M.A. (2004). Long-term potentiation and memory. Physiol. Rev. 84, 87136. Ma, D.K., Jang, M.H., Guo, J.U., Kitabatake, Y., Chang, M.L., Pow-Anpongkul, N., Flavell, R.A., Lu, B., Ming, G.L., and Song, H. (2009). Neuronal activityinduced Gadd45b promotes epigenetic DNA demethylation and adult neurogenesis. Science 323, 10741077.

Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc. 707

Cell Stem Cell

Perspective
Merkle, F.T., Mirzadeh, Z., and Alvarez-Buylla, A. (2007). Mosaic organization of neural stem cells in the adult brain. Science 317, 381384. Mira, H., Andreu, Z., Suh, H., Lie, D.C., Jessberger, S., Consiglio, A., San Em s-Torrejo n, M.A., Nakashima, K., et al. (2010). ela, R., Marque eterio, J., Hortigu Signaling through BMPR-IA regulates quiescence and long-term activity of neural stem cells in the adult hippocampus. Cell Stem Cell 7, 7889. Mirzadeh, Z., Merkle, F.T., Soriano-Navarro, M., Garcia-Verdugo, J.M., and Alvarez-Buylla, A. (2008). Neural Stem Cells Confer Unique Pinwheel Architecture to the Ventricular Surface in Neurogenic Regions of the Adult Brain. Cell Stem Cell 3, 265278. Molofsky, A.V., Slutsky, S.G., Joseph, N.M., He, S., Pardal, R., Krishnamurthy, J., Sharpless, N.E., and Morrison, S.J. (2006). Increasing p16INK4a expression decreases forebrain progenitors and neurogenesis during ageing. Nature 443, 448452. Nottebohm, F. (2011). Song Learning in Birds Offers a Model for Neuronal Replacement in Adult Brain. In Neurogenesis in the Adult Brain I: Neurobiology, T. Seki, K. Sawamoto, J.M. Parent, and A. Alvarez-Buylla, eds. (Tokyo: Springer), pp. 4784. OKeeffe, G.C., Tyers, P., Aarsland, D., Dalley, J.W., Barker, R.A., and Caldwell, M.A. (2009). Dopamine-induced proliferation of adult neural precursor cells in the mammalian subventricular zone is mediated through EGF. Proc. Natl. Acad. Sci. USA 106, 87548759. Okamoto, M., Inoue, K., Iwamura, H., Terashima, K., Soya, H., Asashima, M., and Kuwabara, T. (2011). Reduction in paracrine Wnt3 factors during aging causes impaired adult neurogenesis. FASEB J. 25, 35703582. Paez-Gonzalez, P., Abdi, K., Luciano, D., Liu, Y., Soriano-Navarro, M., Rawlins, E., Bennett, V., Garcia-Verdugo, J.M., and Kuo, C.T. (2011). Ank3dependent SVZ niche assembly is required for the continued production of new neurons. Neuron 71, 6175. Palmer, T.D., Willhoite, A.R., and Gage, F.H. (2000). Vascular niche for adult hippocampal neurogenesis. J. Comp. Neurol. 425, 479494. Parent, J.M., Yu, T.W., Leibowitz, R.T., Geschwind, D.H., Sloviter, R.S., and Lowenstein, D.H. (1997). Dentate granule cell neurogenesis is increased by seizures and contributes to aberrant network reorganization in the adult rat hippocampus. JNeurosci 17, 37273738. Peretto, P., Dati, C., De Marchis, S., Kim, H.H., Ukhanova, M., Fasolo, A., and Margolis, F.L. (2004). Expression of the secreted factors noggin and bone morphogenetic proteins in the subependymal layer and olfactory bulb of the adult mouse brain. Neuroscience 128, 685696. Polenov, A.L., and Chetverukhin, V.K. (1993). Ultrastructural radioautographic analysis of neurogenesis in the hypothalamus of the adult frog, Rana temporaria, with special reference to physiological regeneration of the preoptic nucleus. II. Types of neuronal cells produced. Cell Tissue Res. 271, 351362. nchez-Sa nchez, F., Andreu-Agullo , C., Ferro n, S.R., rez-Castillejo, C., Sa Ram nchez, P., Mira, H., Escribano, J., and Farin as, I. (2006). Aroca-Aguilar, J.D., Sa Pigment epithelium-derived factor is a niche signal for neural stem cell renewal. Nat. Neurosci. 9, 331339. Rao, M.S., Hattiangady, B., and Shetty, A.K. (2008). Status epilepticus during old age is not associated with enhanced hippocampal neurogenesis. Hippocampus 18, 931944. Reynolds, B.A., and Weiss, S. (1992). Generation of neurons and astrocytes from isolated cells of the adult mammalian central nervous system. Science 255, 17071710. Sawamoto, K., Wichterle, H., Gonzalez-Perez, O., Choln, J.A., Yamada, M., Spassky, N., Murcia, N.S., Garcia-Verdugo, J.M., Marin, O., Rubenstein, J.L., et al. (2006). New neurons follow the ow of cerebrospinal uid in the adult brain. Science 311, 629632. a-Verdugo, J.M., McEwen, B.S., and Alvarez-Buylla, A. (2001). Seri, B., Garc Astrocytes give rise to new neurons in the adult mammalian hippocampus. J. Neurosci. 21, 71537160. a-Verdugo, J.M., Collado-Morente, L., McEwen, B.S., and Seri, B., Garc Alvarez-Buylla, A. (2004). Cell types, lineage, and architecture of the germinal zone in the adult dentate gyrus. J. Comp. Neurol. 478, 359378. Shen, Q., Goderie, S.K., Jin, L., Karanth, N., Sun, Y., Abramova, N., Vincent, P., Pumiglia, K., and Temple, S. (2004). Endothelial cells stimulate self-renewal and expand neurogenesis of neural stem cells. Science 304, 13381340. Shen, Q., Wang, Y., Kokovay, E., Lin, G., Chuang, S.M., Goderie, S.K., Roysam, B., and Temple, S. (2008). Adult SVZ Stem Cells Lie in a Vascular Niche: A Quantitative Analysis of Niche Cell-Cell Interactions. Cell Stem Cell 3, 289300. Shimojo, H., Ohtsuka, T., and Kageyama, R. (2008). Oscillations in notch signaling regulate maintenance of neural progenitors. Neuron 58, 5264. Suh, H., Consiglio, A., Ray, J., Sawai, T., DAmour, K.A., and Gage, F.H. (2007). In vivo fate analysis reveals the multipotent and self-renewal capacities of Sox2+ neural stem cells in the adult hippocampus. Cell Stem Cell 1, 515528. Tavazoie, M., Van der Veken, L., Silva-Vargas, V., Louissaint, M., Colonna, L., Zaidi, B., Garcia-Verdugo, J.M., and Doetsch, F. (2008). A Specialized Vascular Niche for Adult Neural Stem Cells. Cell Stem Cell 3, 279288. Tozuka, Y., Fukuda, S., Namba, T., Seki, T., and Hisatsune, T. (2005). GABAergic excitation promotes neuronal differentiation in adult hippocampal progenitor cells. Neuron 47, 803815. Villeda, S.A., Luo, J., Mosher, K.I., Zou, B., Britschgi, M., Bieri, G., Stan, T.M., Fainberg, N., Ding, Z., Eggel, A., et al. (2011). The ageing systemic milieu negatively regulates neurogenesis and cognitive function. Nature 477, 9094. Wang, L.P., Kempermann, G., and Kettenmann, H. (2005). A subpopulation of precursor cells in the mouse dentate gyrus receives synaptic GABAergic input. Mol. Cell. Neurosci. 29, 181189. Weiss, S., Reynolds, B.A., Vescovi, A.L., Morshead, C., Craig, C.G., and van der Kooy, D. (1996). Is there a neural stem cell in the mammalian forebrain? Trends Neurosci. 19, 387393. Wiedemeyer, R., Brennan, C., Heffernan, T.P., Xiao, Y., Mahoney, J., Protopopov, A., Zheng, H., Bignell, G., Furnari, F., Cavenee, W.K., et al. (2008). Feedback circuit among INK4 tumor suppressors constrains human glioblastoma development. Cancer Cell 13, 355364. Wong, S.Y., and Reiter, J.F. (2008). The primary cilium at the crossroads of mammalian hedgehog signaling. Curr. Top. Dev. Biol. 85, 225260. Zappaterra, M.D., Lisgo, S.N., Lindsay, S., Gygi, S.P., Walsh, C.A., and Ballif, B.A. (2007). A comparative proteomic analysis of human and rat embryonic cerebrospinal uid. J. Proteome Res. 6, 35373548. Zhao, C., Deng, W., and Gage, F.H. (2008). Mechanisms and functional implications of adult neurogenesis. Cell 132, 645660.

708 Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc.

Cell Stem Cell

Perspective
The Promise and Perils of Stem Cell Therapeutics
George Q. Daley1,2,3,*
1Stem Cell Transplantation Program, Division of Pediatric Hematology/Oncology, Manton Center for Orphan Disease Research, Howard Hughes Medical Institute, Childrens Hospital Boston and Dana Farber Cancer Institute; Division of Hematology, Brigham and Womens Hospital; and Department of Biological Chemistry and Molecular Pharmacology, Harvard Medical School, Boston, MA 02115, USA 2Broad Institute, Cambridge, MA 02142, USA 3Harvard Stem Cell Institute; Boston, MA 02138, USA *Correspondence: george.daley@childrens.harvard.edu DOI 10.1016/j.stem.2012.05.010

Stem cells are the seeds of tissue repair and regeneration and a promising source for novel therapies. However, apart from hematopoietic stem cell (HSC) transplantation, essentially all other stem cell treatments remain experimental. High hopes have inspired numerous clinical trials, but it has been difcult to obtain unequivocal evidence for robust clinical benet. In recent years, unproven therapies have been widely practiced outside the standard clinical trial network, threatening the cause of legitimate clinical investigation. Numerous challenges and technical barriers must be overcome before novel stem cell therapies can achieve meaningful clinical impact.
Cell Therapeutics: The Current Standard of Care In the twentieth century small molecule and protein drugs proved remarkably successful in restoring health and extending life span, but in the twenty-rst century our aging population will face an increasing burden of organ failure and neurodegenerative disease. Such conditions are unlikely to be cured by drugs alone and instead call for restoration of tissue function through novel therapeutic approaches. Transplantation of whole organs heart, lung, liver, kidney, small bowel, and pancreashas become routine in modern medicine and has saved countless lives, while grafts of the skin and cornea for burns or ocular injury and transfusions of red blood cells and platelets for diseaserelated or chemotherapy-induced cytopenias are likewise widely employed tissue and cell therapies. However, current therapeutic strategies either are limited by donor availability and immunologic barriers or pertain to only a minor range of conditions. For the many diseases and disorders of aging for which there is no cure, innovative applications of tissue engineering and novel cell therapies derived from pluripotent and tissue-restricted stem cells represent major frontiers for the future. Hematopoietic stem cells (HSCs), the therapeutic constituents of whole bone marrow and umbilical cord blood, have been the most widely employed stem cell therapy. When successful, HSC transplantation can be curative for scores of genetic blood disorders like thalassemia and immune deciency and for malignancies like leukemia and lymphoma. HSC transplantation is undoubtedly the most successful application of stem cells in medicine, yet for many conditions success rates remain frustratingly low and morbidity and mortality unacceptably high. The need for precise molecular matching of donor and recipient means that many patients lack a suitable donor, either within their own family or in the public at large, even when databases list many millions of potential unrelated donors. When a match can be found, minor mismatches between donor and recipient frequently incite graft versus host disease (GVHD), an attack of the donor immune effector T cells against host tissues that results in skin rash, mucositis, diarrhea, and liver and lung destruction. GVHD is a major cause of treatment associated
740 Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc.

morbidity and mortality. Finally, grafts can fail, and disease can relapse. Although it is difcult to give a precise gure for the overall success rate for HSC transplantation, even an optimist would acknowledge that some 50% of patients are left without a cure or with a permanent disability. Thus, even our most successful form of stem cell therapy remains a heroic effort, reserved only for the sickest patients who have no better alternative. Lessons from the Historical Development of HSC Transplantation The evolution of HSC transplantation from its experimental origins to its acceptance as a standard of care in medicine is a tale that is both inspiring and cautionary. E. Donnall Thomas and colleagues were the rst to perform marrow transplantation for otherwise fatal leukemia in the 1950s (Thomas et al., 1957). The rationale was predicated upon the known capacity for radiation to suppress leukemic hematopoiesis and studies demonstrating that injections of marrow rescued mice from otherwise lethal radiation exposure (Jacobson et al., 1951; Lorenz et al., 1951). Thomas wrote in a memoir in 2005, These patients inspired us to speculate that it might be possible to destroy leukemic cells and normal marrow by lethal whole body irradiation, with reconstitution of marrow by marrow transplantation. Arguably, the rst studies in humans were founded upon rather minimal evidence of efcacy in rodent models, and Thomas further noted, We recognized that it would be important to do similar studies in an animal model . [and] decided to move forward with studies of man and dog at the same time (Thomas, 2005). Indeed, Thomas and colleagues suffered considerable failure in preclinical canine models and witnessed the deaths of many scores of patients, which prompted great skepticism about whether the human experiments should continue. Nevertheless, Thomas and his intrepid team of investigators forged ahead. It took almost two decades before advances in research on tissue matching to dene compatible donor-recipient pairs, and improved treatment of graft versus host disease and the infectious complications of marrow transplant allowed marrow transplantation to achieve consistent success in the late 1970s.

Cell Stem Cell

Perspective
Some important principles emerge from this lesson in the history of HSC transplantation. First, the risk of the intervention should be commensurate with the severity of the underlying condition to be treated. The aggressively malignant nature of the conditions being treatedfatal leukemia and marrow aplasiameant that the rst practitioners of marrow transplantation were justied and even compelled to attempt heroic and potentially highly toxic interventions for invariably fatal diseases. Second, although human biology is only partially predictable from animal models, preclinical animal models remain a key element in the scientic development of novel therapies. At the beginning of human marrow transplantation, it was understood that identical twins accepted skin and solid organ grafts, but only a minority of the time did siblings. Experiments in the murine and canine marrow transplantation models reected similar transplantation barriers. Notwithstanding these sobering limitations, the early practice of marrow transplant in patients proceeded despite a lack of robust evidence in animal models for graft acceptance between unrelated individuals. Only later were methods for lymphocyte matching developed (the antecedent to HLA typing), which was the key development in advancing the success of marrow transplantation. Finally, important and fundamental insights into therapeutic mechanisms were required before the eventual success of clinical translation of HSC transplantation therapies. With the benet of hindsight, one could argue that the earliest human transplants were premature and doomed to fail. One might question whether a therapy as toxic as marrow transplant, with so little evidence for success in animal models prior to testing in humans, could emerge in the current era. Under todays more rigorous regulatory climate, institutional review boards weigh risks and potential benet on behalf of patients, insist on an impartial process of informed consent to minimize misconceptions about therapeutic potential, and monitor adverse events in the course of clinical trials. Indeed, one might reasonably conclude that todays IRBs might not have approved the early studies of Thomas and colleagues, but if they had, would have interceded to stop the experiments when the high incidence of treatment-related mortality became apparent. The conjecture that modern-day IRBs might not approve the early experiments in HSC transplant does not imply that HSC transplant would not emerge under the current regulatory climate. On the contrary, I believe that bone marrow transplant could be developed within todays environment of strict clinical research regulation, although by a more conservative path that would spare considerable patient morbidity and mortality. As we learned from premature attempts at gene therapy in the early 1990s, new therapeutic technologies require considerable understanding of fundamental mechanisms before they can be delivered with condence. Indeed, roughly 70% of early phase clinical trials of pharmaceuticals fail and over 50% at phase III (Ledford, 2011), and thus it stands to reason that signicant resources are squandered because of the imprecision of early stage clinical research. Yet, especially with novel technologies, clinical experimentation proceeds energetically, because hope triumphs over experience. From this authors perspective, a conservative approach to clinical translation of stem cell therapies is warranted at this time, not because stem cell treatments are excessively risky (though some may yet prove to be), but

Figure 1. Clinical Trials of Major Stem Cell Types


Pie chart indicating the relative numbers of open trials testing clinical interventions for hematopoietic, neural, mesenchymal, adipose, and embryonic stem cells, as listed on the U.S. NIH website clinicaltrials.gov.

rather because our understanding of the mechanisms by which stem cells might prove useful, and in which diseases, remains primitive. In a climate where government and philanthropic funds for fundamental research are increasingly scarce, and investment capital from the private sector for biotechnology has dried up, purely empirical attempts at stem cell therapy are difcult to justify, given the high probability of failure. In a 1995 report assessing the investment in gene therapy by the U.S. National Institutes of Health, a panel chaired by Stuart Orkin and Arno Motulsky recommended increased emphasis on research dealing with the mechanisms of disease pathogenesis, further development of animal models of disease, enhanced use of preclinical gene therapy approaches in these models, and greater study of stem cell biology in diverse organ systems (http://oba.od.nih.gov/oba/rac/panelrep.pdf). Similar recommendations regarding the need for proper investments in fundamental aspects of stem cell therapeutics seems warranted and prudent at this time. Stem Cell Therapeutics: Frontline Clinical Trials and Medical Innovations A search of the Unites States government-sponsored website www.clinicaltrials.gov with the term stem cells lists over 4,000 past, current, and anticipated trials, with over 1,750 now open (Figure 1). The vast majority of open trials aim to build upon decades of research and clinical experience in hematopoietic transplantation (>1,200), and include strategies to expand the suboptimal dose of HSCs within umbilical cord blood, to complement gene defects in HSCs through viral transgene delivery (gene therapy), and to engineer T cells to attack malignancy via adoptive immunotherapy. Despite the relatively primitive understanding of therapeutic mechanisms for other stem cells, hundreds more trials are testing mesenchymal (115), adipose-derived (36), and neural stem cells (280), sometimes in quite bold and unconventional ways that bear little resemblance to the known differentiation potential or modes of tissue
Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc. 741

Cell Stem Cell

Perspective

Figure 2. Worldwide Experimental Trials of Stem Cell-Based Therapies


World map showing locations of open, closed, and pending clinical trials of stem cell-based interventions as listed on clinicaltrials.gov. The relative numbers of trials performed outside of the U.S. may indeed be markedly understated because of reporting bias at the U.S. government clinical trials website.

regeneration or repair associated with these classes of stem cells. As of this writing, three trials pertain to products derived from ESCs. A wide array of stem cell studies are being carried out on a global basis on all continents, suggesting widespread clinical interest (Figure 2). Mesenchymal stem cells (MSCs) are dened by their broblast-like morphology, adherence to plastic, expression of a specic set of surface antigens (CD105+, CD90+, CD73+), and capacity for osteogenic, chondrogenic, and adipogenic fates in vitro. MSCs are most often derived from bone marrow but can also be isolated from adipose tissue; adipose-derived stem cells may also consist of pericytes or endothelial progenitors that may differ somewhat in their properties from MSCs. Easy access to large quantities is an advantage for adiposederived stem cells, which are being tested for soft-tissue repair and regeneration (Tobita et al., 2011). Both autologous (self) and allogeneic (foreign) MSCs are being tested in vivo to enhance healing that reects their in vitro potential to form bone or cartilage, as in bone fracture and joint cartilage repair (Grifn et al., 2011). Although such studies are founded on strong preclinical evidence and sound scientic and clinical hypotheses, evidence for robust clinical efcacy of MSCs for orthopedic indications has been challenging to conrm, and to date no therapy based on MSCs has yet won approval by the U.S. Food and Drug Administration (FDA). The difculty in proving the efcacy of regenerative treatments based on the well-char742 Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc.

acterized cellular potentials of MSCs suggests that our understanding of how even familiar stem cells can be exploited therapeutically in vivo remains primitive. MSCs are being tested in a wide range of clinical indications where the clinical hypotheses are more speculative, the therapeutic mechanisms are incompletely dened, and in some instances the preclinical evidence is highly contentious. For example, from a scientic foundation that can be traced to a highly controversial report that whole bone marrow would regenerate cardiac muscle following transplantation into injured hearts (Orlic et al., 2001), an observation later disproven (Balsam et al., 2004), thousands of patients have been treated in trials worldwide with various cell preparations of bone marrow or MSCs, with the scientic community debating the signicance of the results (Choi et al., 2011). Subsequent studies have invoked a variety of contingent mechanisms including salutary paracrine effects on resident cardiomyocytes and putative cardiac stem cells, neoangiogenesis, and biomechanical alterations due to scarring (Gnecchi et al., 2008; Menasche, 2011; Williams et al., 2011). The questions about underlying mechanism notwithstanding, combined meta-analyses of numerous trials has argued for measureable yet quite modest therapeutic effects, which has left practitioners unsure of the signicance and robustness of these therapeutic approaches (Tongers et al., 2011). MSCs have also been widely tested for their capacity to mitigate autoimmunity, following somewhat serendipitous

Cell Stem Cell

Perspective
observations that MSCs can interfere with in vitro immunological assays such as mixed lymphocyte reactions and modulate production and function of the major classes of immune cells (Kode et al., 2009; Shi et al., 2011). Although it is unclear whether immune antagonism reects any native function of MSCs in vivo, ex vivo expanded preparations have been infused in patients in hopes of mitigating transplant-related graft versus host disease and autoimmune conditions like Crohns disease, multiple sclerosis, and systemic lupus (Kebriaei and Robinson, 2011; Shi et al., 2011). One can nd numerous reports of efcacy in the literature, but these are mixed with negative data (Kebriaei and Robinson, 2011). The precise role of MSCs as agents for immune modulation remains to be proven. When clinical indications stray yet further from the presumptive core functions of MSCs, and therapeutic mechanisms become increasingly speculative, clinical translation is a largely empirical rather than a rational effort. Likewise, while umbilical cord blood (UCB) has emerged as a viable alternative to other sources of HSCs (e.g., mobilized peripheral blood or bone marrow) for the treatment of leukemia and nonmalignant hematologic conditions (Rocha et al., 2004), it has also become a common source for experimental interventions in a wide variety of nonhematologic indications as disparate as myocardial infarction, multiple sclerosis, amyotrophic lateral sclerosis, cerebral palsy, traumatic brain injury, stroke, and inherited metabolic disorders (Copeland et al., 2009; Harris, 2009; McKenna and Sheth, 2011; Prasad and Kurtzberg, 2009). Evidence exists that a number of distinct cell types can be cultured from UCB, gler et al., 2004; Pelosi including multipotential stem cells (Ko et al., 2012), but it is unclear whether such expandable cell populations exist at appreciable levels in unmanipulated samples. While in theory such cells could mediate therapeutic effects, nonhematologic indications for UCB transplantation have not been widely accepted into standard practice. When clinical investigation proceeds largely empirically, and without a deeper understanding of the basic therapeutic mechanisms, it is difcult to reformulate therapeutic strategies after clinical failures. Neural stem cells (NSC) can be cultured from fetal and adult brain and demonstrated to differentiate into neurons, oligodendrocytes, and astrocytes in vitro. Given the wide array of neurologic conditions that have devastating clinical consequences, there is considerable interest in the therapeutic potential of neural regeneration therapies. However, neurodegenerative diseases, catastrophic stroke, traumatic brain injury, and spinal paralysis are among the most daunting challenges for regenerative medicine. The development of the brain and peripheral nerves and their interconnectedness with tissues throughout the body requires a remarkably complex choreography during fetal development. The proper milieu for directing the formation of highly specied neuronal subtypes and guiding their projection to and interconnectedness with critical targets is highly unlikely to exist in the adult body. But faced with compelling unmet medical need and desperation on the part of patients, there are hundreds of investigator-initiated clinical trials occurring in academic settings (Figure 1), and several companies have forged efforts to develop novel therapies through intracerebral or spinal transplantation of neural stem cells (Trounson et al., 2011). StemCells Inc (California, USA) has tested NSCs in Battens disease (neuronal ceroid lipofuscinosis) and was able to document safe delivery but discontinued the trial because of the inability to accrue an adequate number of patients. Their current focus is Pelizaeus-Merzbacher disease, a myelin disorder, and chronic spinal cord injury. Other companies are testing NSC transplant for stroke (ReNeuron, United Kingdom), amyotropic lateral sclerosis (Neuralstem, Inc, Maryland, USA), and Parkinsons disease (NeuroGeneration, California, USA). In most of these cases, the clinical hypotheses being tested do not depend upon the generation of neurons de novo, but instead on complementation of enzyme deciencies, remyelination, or modulation of endogenous repair through neoangiogenesis or neuroprotection. Although widely publicized, there are comparatively few clinical trials of products derived from human embryonic stem cells (hESCs). The rst trial conducted in humans delivered oligodendrocyte progenitors for the remyelination of spinal cord axons damaged through crush injury. These studies were based on extensive preclinical experience with the derivation and characterization of oligodendrocytes and their delivery in animal models that showed remyelination and restoration of motor function (Keirstead et al., 2005; Liu et al., 2000; McDonald and Belegu, 2006; McDonald and Howard, 2002; McDonald et al., 1999; Nistor et al., 2005). Moreover, this rst trial required a herculean effort to satisfy FDA regulatory oversight, by report entailing the submission of over 20,000 pages of data and documentation. The trial, sponsored by the Geron Corporation (California, USA), enrolled and treated its rst four patients before being discontinued due to a decision by company management to focus on alternative corporate priorities (Baker, 2011). No formal results have yet been released regarding the phase 1 clinical trial in this rst small cohort of patients, but the primary endpoints were safety of the cells, and at the very least one hopes that some evidence will be gleaned that products of ESCs can be delivered without risk of teratoma, although long-term followup of all treated patients will be necessary. The only other current clinical trials involve transplantation of hESC-derived cells to treat retinal blindness. This condition takes many forms, both genetic and age-related, and as a group of disorders has many appealing features for stem cell-based interventions. The retina is accessible for local delivery of cells, which can then be monitored via direct visualization. The retina may also provide some degree of immune privilege. Very preliminary results of a trial involving the subretinal injection of hESCderived retinal pigment epithelial cells for Stargardts macular degeneration and another for age-related macular degeneration sponsored by the company Advanced Cell Technologies (ACT) were recently reported, despite experience on only one patient in each trail (Schwartz et al., 2012). Only one of the two patients showed evidence of persistent cells but both were reported to show some restoration of visual perception. While it is difcult to draw conclusions from these early trials due to the limited numbers of patients involved and the very brief 4 month period of follow-up, the trials represent milestones in that the investigators succeeded in clearing considerable regulatory hurdles and met very high standards of preclinical cell characterization and quality control prior to exposing patients to the risk of ESCbased products. The experience alone, for both investigators and regulators, is an essential albeit small step forward in the long path to establishing ESC-based therapeutics.
Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc. 743

Cell Stem Cell

Perspective
While MSCs, NSCs, and products from ESCs are being tested in the context of numerous clinical trials, yet another arm of regenerative medicinetissue engineeringis comingling MSCs or a variety of other cultured cell types with biocompatible materials to solve surgical challenges. Reconstruction of bladders (Aboushwareb and Atala, 2008; Atala, 2011; Tian et al., 2010), tendons (Sun et al., 2011), and complex structures like the trachea (Macchiarini et al., 2008) represent solutions to highly personal needs of specic patients and are acceptably performed as highly innovative and individualized surgical therapies, part of the long tradition of surgical innovation. The mechanisms for developing such novel interventions and gaining acceptance by the surgical and biomedical communities involve the same core principles required for medical interventions sound scientic rationale and methods, institutional and practitioner accountability, thorough and rigorous informed consent, patient follow-up, timely reporting of adverse events, peer review of therapeutic claims, and publication in the medical literature. The potential for therapeutic innovation at the interface of stem cell biology and tissue engineering is particularly appealing but beyond the scope of this review. I refer the reader instead to excellent recent reviews (Grifn et al., 2011; Peck et al., 2012; Sun et al., 2011). Anticipated Future Interventions and Opportunities Among the many disparate conditions, disorders, and diseases for which stem cells have offered promise, a few stand out as particularly compelling. In general, they are conditions where defects are largely cell autonomous and entail the loss or dysfunction of a single class of cells or a monocellular component of a complex tissue, such that restoration of function through cell replacement would be curative or signicantly ameliorate symptoms. Those conditions most amenable to treatment present the least anatomic complexity and affect tissues that do not typically regenerate spontaneously because they lack endogenous pools of tissue stem cells. We can predict ultimate success with most condence if some clinical evidence already exists that cell replacement might indeed be therapeutic, for instance through prior assessments of cadaveric or fetal tissue transplantation. For conditions previously treated with cadaveric or fetal material, efcacy may be limited by the inadequate supply or quality of the cells, making pluripotent or reprogrammed cell sources advantageous. Parkinsons Disease. Although neurologists recognize that Parkinsons disease (PD) has systemic features, the chief decit remains the loss of a specic subtype of midbrain dopaminergic neurons located in a deep brain structure, the substantia nigra, whose many connections to the striatum are responsible for regulating movements, such that PD patients suffer from immobility, rigidity, and tremor. Drug replacement with precursors of dopamine (DA), dopamine agonists, or antagonists of dopamine metabolism serves to ameliorate symptoms but cannot stem the inexorable decline in most patients. Based on decades of experience from several groups with transplantation of fetal tissue sources of DA neurons, deep brain transplantation can indeed restore local DA production and ameliorate symptoms, with some patients showing durable improvement and graft integrity after two decades (Freed et al., 1992; Lindvall et al., 1990; Lindvall et al., 1994; Piccini et al., 1999, 2005). Functional imaging and postmortem analysis support the stable integration and
744 Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc.

persistence of grafts in some patients, prompting continued enthusiasm for this approach among some practitioners, provided that a suitable source of DA neurons can be dened (Freed et al., 1992; Lindvall et al., 1990, 1994; Ma et al., 2010; Nakamura et al., 2001; Piccini et al., 1999, 2000). Others, however, remain skeptical, in part because a trial of fetal grafts randomized against sham surgery was inconclusive, with some patients sustaining functional decline postsurgery due to dyskinesias as a result of excessive graft function (Freed et al., 2001). Supporters of cell therapy for PD point out that a more reliable, consistent, and dened source of DA neurons would justify further testing of transplantation strategies. Many groups have differentiated DA neurons from both neural stem cell and pluripotent stem cell sources and proven functional in rodent models (Hargus et al., 2010; Sanchez-Pernaute et al., 2008; Tabar et al., 2008; Wernig et al., 2008). Analysis of this DA neuron production has not always distinguished among the many different classes of neurons that produce DA throughout the neuraxis, but recent advances have made possible the differentiation from pluripotent cell sources of regionally specic midbrain DA neuronal subtypes whose deciency is most affected in PD is possible, and such cells have been documented to function in rodent and primate models (Chambers et al., 2009; Fasano et al., 2010; Kriks et al., 2011). Moreover, techniques for producing personalized autologous stem cells via somatic cell reprogramming now exist, and it has been shown that autologous cells function better than cells derived from unrelated donors in rodent models of PD transplant (Tabar et al., 2008). The availability of highly specied, dened, autologous DA neuron preparations creates legitimate opportunities for testing in PD patients, including the testing of specic doses to establish a dose-response curve. Nevertheless, even optimistic accounts identify the signicant hurdles that remain (Lindvall and Kokaia, 2010). Notably, any cell therapy must ultimately be superior in safety and efcacy to any drug therapy, and establishing such utility will require large-scale and painstaking prospective trials to be conducted over many years. Thus, despite promise, cell therapy as the standard of care for PD is but a distant horizon. Cell therapy for PD will need to be efcacious and safe to compete with the highly effective drug treatments that currently exist (Hjelmgren et al., 2006). In contrast, a condition like Huntingtons disease, which has no viable drug therapy and is invariably fatal, is an appealing alternative therapeutic target for cell transplantation therapies derived from NSCs and ESCs. Intrastriatal transplantation of homotypic fetal tissues has shown graft durability and reports of amelioration of symptoms in HD patients (Gallina et al., 2010; Nicoleau et al., 2011). As for PD, an improved cell source would facilitate the necessary studies to optimize the dose and target region for cell transplantation. Techniques for directed differentiation of ESCs into relevant medium spiny neurons and amelioration of rodent models of HD have been reported and bode well for future translational clinical studies (Benraiss and Goldman, 2011). Autoimmune Diabetes Mellitus. Type 1 diabetes (T1D; insulindependent, juvenile onset) is an autoimmune condition that involves active immune destruction of the beta cells of the islets of Langerhans of the pancreas, leaving the patient with inadequate supplies of insulin and susceptibility to hyperglycemic

Cell Stem Cell

Perspective
crises characterized by life-threatening ketoacidosis. At diagnosis, patients harbor depleted pools of beta cells and are unable to mount a regenerative response to restore beta cell mass, even if their autoimmune response can be controlled. Whether beta cells regenerate after injury in the adult pancreas has been vigorously debated (Bonner-Weir and Weir, 2005; Dor et al., 2004; Dor and Melton, 2008), but endogenous regeneration under pathologic conditions is not robust, and alternative sources of beta cells would therefore be required. Deriving fully functional beta cells in vitro from pluripotent stem cells has proved challenging, but a group from the biotechnology company Novocell did report successful derivation of precursors in vitro that appear to fully differentiate and mature after transplantation in vivo (DAmour et al., 2006; Kroon et al., 2008). In a more recent advance, Gadue and colleagues have derived a stably expandable endodermal progenitor that is more efcient at producing beta cells than if one proceeds directly from ESC (Cheng et al., 2012). If a reliable source of beta cells can be produced in vitro, a credible path toward clinical development could be envisioned. We know that transplantation of whole pancreas, or infusion of islet preparations from cadaveric sources in the context of a corticosteroid-sparing regimen of immune suppression (the Edmondton Protocol), can restore glycemic control for extended time periods (Shapiro et al., 2000, 2006). Although patients later relapse, the potential for repeated cell infusions would be greatly facilitated by a more abundant source of beta cells, and deriving puried beta cells from pluripotent stem cell sources thus remains a much sought after goal in stem cell biology. As T1D is an autoimmune disorder, it seems unlikely that autologous cells would be a preferable source of material to allogeneic cells, as immune suppression to protect the beta cells would still be required in either scenario. Attempts to convert exocrine pancreatic tissue into beta-like endocrine cells through ectopic expression of transcription factors, a type of direct reprogramming of cell fates in situ, is a new therapeutic concept with provocative appeal (Zhou et al., 2008). Other Treatable Conditions on the Horizon. Corneal injury that leads to scarring and blindness has prompted efforts to culture and expand limbal stem cells into corneal patches in vitro, followed by corneal grafting. Recent reports conrm several independent studies that corneal grafting using alternative sources of epithelial cells can restore vision, and appears to be a promising novel stem cell-based treatment for a grave but rare human condition (Nishida et al., 2004; Rama et al., 2010; Tsai et al., 2000; Tsubota et al., 1999). Liver transplantation cannot meet the demands of patients suffering from liver failure around the globe, and production of hepatocyte-like cells from pluripotent stem cells sources has been reported by several groups. Despite considerable similarity to native hepatocytes, the in vitro derived cells have not yet been reported to be fully functional in animal models, and considerable challenges remain for achieving functional integration of in vitro derived hepatocytes, especially for conditions like cirrhosis that already entail markedly altered liver anatomy and compromised circulation. Similarly, production of cardiomyocytes appears to be robust in the petri dish, but achieving engraftment in the damaged heart of a clinically meaningful dose of cells, together with integration in a manner that restores pump function, remains a major challenge. In this case, clever engineering of biomaterials might enable the creation of contractile cardiac patches that could be sewn onto the heart. Finally, producing HSCs from personalized pluripotent stem cells, coupled to gene repair, is an appealing strategy for dozens of genetic disorders of the bone marrow including immune deciency, hemoglobinopathy, and genetic marrow failure syndromes. Still other potential indications for tissue replacement therapies involve in vitro production of endothelial cells and potentially even human gametes, but none appear to have imminent clinical application. All cell replacement therapies face similar challenges of graft integration into the host environment, which entails trafcking, homing, and integration into native niches or microenvironments, connection to a host blood supply, immune compatibility, and graft durability. Solving such challenges will engage the research community for decades to come. Who Will Translate Stem Cell Science into Regenerative Medicine? Scientic advances in stem cell biology are being driven by the current intellectual ferment and excitement of the eld, but when and how these advances will be translated into successful treatments remain fertile questions for debate. Will cell therapies remain a highly patient-focused endeavor performed solely in academic medical centers, akin to bone marrow or solid organ transplantation? Or will stem cells ever become commercial, pharmaceutical grade off-the-shelf products? One might imagine a future in which medical centers offer highly customized, patient-focused approaches to stem cell treatments, perhaps utilizing the products of personalized induced pluripotent stem cells (see Yamanaka, 2012, this issue). IPS cells have enormous theoretical appeal as vehicles for combined gene repair and cell replacement therapy for genetic disease (Daley and Scadden, 2008). Newer forms of stem cell transplant could replicate the current status of bone marrow transplantation, which has developed into a remarkably complex infrastructure for capturing cellular and molecular information in international registries for literally millions of potential donors and entails lengthy, costly, and risky interventions in intensive clinical care settings. Given the imperative of treating patients in need, stem cell transplants for genetic and acquired diseases will emerge from academic centers because clinician investigators will develop them and patients will demand them. Like gene repair (gene therapy), cell replacement therapies will probably serve rare conditions rst and pertain to small numbers of patients receiving highly individualized treatments, perhaps coupling gene repair with autologous cell replacement approaches, for example for blood diseases. Such small-scale applications will dominate until and unless generic interventions and off-the-shelf approaches prove feasible. The prospects for more widespread stem cell-based treatments depends on either solving the immune rejection barrier, through advances in promoting immune tolerance to allogeneic tissues, or accepting the use of immune suppressioneven lifelongto facilitate allogeneic cell therapies. Immune suppression is already standard for organ transplantation, so we know that its use to facilitate life-sustaining cell therapy is feasible. Because cell manufacture is likely to be the most costly and timeconsuming aspect limiting cell therapies, the prospects for realizing economies of scale would seem to call for the establishment of master cell banks that could be the source of cells
Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc. 745

Cell Stem Cell

Perspective
off-the-shelf. The polymorphism of histocompatibility genes and the resulting variety of tissue types is far too great in human populations to expect banks to be able to supply perfect tissue matches for all potential patients. Instead, one might envision banks of cells derived from donors with highly common genotypes of the histocompatibility genes. This type of approach would be greatly facilitated by cell strains with homozygosity of histocompatibility loci. Past approximations of the number of cell lines that would be needed in such a repository or master cell bank, based on modeling data from pools of kidney transplant patients and recipients in the United Kingdom and Japan, have suggested that a bank comprised on the order of 1050 cell lines might effectively provide a single HLA antigen match (deemed a minimal requirement for acceptable solid organ transplantation) for approximately 80% of the local population (Gourraud et al., 2012; Nakajima et al., 2007; Taylor et al., 2011, 2005). While encouraging, these numbers suggest that some kind of dual system might well be needed in which the vast majority of individuals can benet from off-the-shelf therapies, but personalized autologous cells derived via reprogramming would be needed for those with difcult-to-match tissue types. Alternatives to Cell Therapy Because of the signicant hurdles that remain in terms of cell manufacture, delivery, anatomical integration, and immune suppression for all but highly personalized therapies, it is entirely possible that more traditional modes of treatment will evolve from stem cell research and ultimately prove the most feasible. Indeed, the generation of patient-derived stem cells holds the most immediate promise for advancing traditional drug discovery paradigms (for a recent review, see Grskovic et al., 2011). Capturing diseases in a dish promises to enable cellbased phenotypic assays that could yield new drugs that repair cell and tissue defects, or perhaps act on endogenous pools of stem cells, stimulating repair and regeneration. For tissues that do not readily regenerate from endogenous pools of stem cells, such as the majority of the brain, the heart, and the kidney, another provocative possibility is the direct conversion of one cell or tissue identity to another that has been depleted by disease or injury. A host of such conversions have been realized in vitro, converting broblasts into cells that resemble and exhibit some functions like neurons, cardiomyocytes, and hepatocytes (Vierbuchen and Wernig, 2011). Cell conversion has considerable theoretical advantages, but whether this new cellular alchemy can be harnessed for therapeutic end remains almost science ction at present, although it is clearly worthy of deeper exploration. Threats to Clinical Translation and to the Integrity of Regenerative Medicine Translating the basic discoveries of stem cell biology into robust, effective, and safe new modalities of care will mean solving new challenges; before success, regenerative medicine will suffer many setbacks. While translating too timidly might deprive needy patients of precious time and life quality, testing cells in patients before a deeper understanding of how stem cells work is risky, too. We need to be condent that we understand the full spectrum of safety concerns and can therefore avoid placing patients at undue risk. We also need to design rigorous, blinded, and when possible randomized trials where evidence for clinical efcacy can be dened precisely, rather than depend
746 Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc.

upon anecdote and clinical observation alone. Given that patients and practitioners may carry unrealistic expectations of clinical efcacy, there is a high likelihood for a robust placebo effect as well as interpretive bias in reporting of clinical results. We also need to be conscious of not exhausting resources that would be better spent on more practical health care needs. Premature application runs the risk of high-prole failure that would sully the credibility of this still-developing eld. With the goal of advancing clinical investigation while preserving rigor, promoting medical innovation while protecting patients, and ensuring integrity in regenerative medicine while respecting autonomy of individual practitioners and patients, the International Society for Stem Cell Research (ISSCR) assembled an international group of scientists, surgeons, gene therapists, bioethicists, patient advocates, and attorneys and composed The ISSCR Guidelines for the Clinical Translation of Stem Cells (Hyun et al., 2008). These guidelines articulated principles and standards as a roadmap for practitioners and regulatory bodies when considering if, when, and how to allow tests of experimental stem cell therapies in actual patients. The guidelines call for independent and rigorous analysis of the decision to test novel treatments in patients, by reviewers with relevant area-specic expertise, who are free of conicts of interest that might lead to positive or negative bias. Expert judgment about the reliability and rigor of the preclinical evidence for efcacy and safety of cellular products is essential for weighing the potential risks against the potential benets before launching a clinical trial. Because no preclinical animal or cellular model is entirely predictive of outcomes in patients, a credible and rigorous process of informed consent is essential to protecting the autonomy of patients and their thoughtful engagement in the research process, where they consent to participate without heightened expectations or therapeutic misconception; such wishful thinking renders patients vulnerable to exploitation and contaminates interpretations of therapeutic efcacy. Medical Innovations outside of Clinical Trials Many in the medical eld recognize the value of innovation outside the context of a clinical trial. However, especially if incorporating the use of highly manipulated cell preparations, such innovative attempts at therapy in the United States should fall under the jurisdiction of the Food and Drug Administration. To comply with accepted professional standards governing the practice of medicine, highly novel uses of any cellular product should not be performed on more than a small number of patients before such use is subject to independent review of the scientic rationale, informed consent, close patient followup, and reporting of adverse events. Any attempt to extend the innovative therapy to a larger group of patients should be preceded by a standard clinical trial. Although some may contend that requiring approval for the practice of novel clinical treatments from an independent body undermines the autonomy of practitioners to provide care to their patients, independent peer review ensures that the rationale for treatment is sound and represents a defensible community standard of medical practice. Premature Clinical Translation The traditional strategy for proving that a medical intervention works and is safe requires rigorous clinical trial design, can be

Cell Stem Cell

Perspective
frustratingly slow and costly, and is generally best suited to highly organized medical settings. However, the history of even legitimate medical practice is rife with examples of instances whereby trust in medical intuition alone, or reliance on uncontrolled retrospective or purely observational studies, has led to mistaken presumptions about medical efcacy, only to be corrected when rigorous blinded, randomized trials proved our presumptions to be false (for example, high-dose chemotherapy and autologous marrow rescue for metastatic breast cancer, postmenopausal hormone replacement therapy and cardiovascular risk, to name just two). The edgling eld of stem cells is already suffering from the taint of illegitimate clinical translation. A quick Google search for stem cell treatments returns a plethora of sponsored websites peddling cures for ailments as diverse as Alzheimers disease and autism. As documented by Cauleld and colleagues, such websites systematically overpromise the potential efcacy of stem cells and trivialize the potential risks (Lau et al., 2008). Sadly, even sophisticated patients or their families can be misled by the veneer of scientic credibility on such websites. As stated previously, apart from treatments using HSCs for blood diseases, and various dermal and corneal indications, essentially all other treatments based on stem cells must be considered experimental medical research and should be administered exclusively in organized clinical trials. Subjects in medical research are generally not required to pay for unproven interventions. Administering interventions outside of controlled clinical trials threatens patients and jeopardizes the integrity of and public trust in medical research, compromising legitimate efforts to advance knowledge. Because of the particular vulnerabilities of patients, many governments have enacted laws to protect patients from exploitation and risk, but some practitioners see such regulation as burdensome and unwarranted restraints on their trade. The threat of litigation for medical malpractice serves as an additional constraint on unwarranted medical practice. Recently, the German government shut down the Xcell Clinic when a child died after receiving intracranial injections of cord blood in an unproven intervention. A recent report documented the development of glioneural masses in the brain and spinal cord of a child who was treated with intrathecal infusions of what were reportedly neural stem cells for ataxia telangectasia, a genetic movement disorder (Amariglio et al., 2009). While one hopes that most stem cell interventions are benign, the safety data are still rudimentary. The history of gene therapy was shaped in a deleterious way by the untimely death of a young man, Jesse Gelsinger, in an FDA-approved clinical study. James Wilson, the physician responsible for the gene therapy clinical trial in question, has written a compelling admonition to practitioners of stem cell therapies, warning that much of the history that prompted premature clinical translation of gene therapy is being repeated by the practitioners of stem cell therapy (Wilson, 2009). He sees the same assumptions of a simplistic, theoretical model indicating that the approach ought to work; a large population of patients with disabling or lethal diseases . harboring fervent hopes; and unbridled enthusiasm of some scientists in the eld, fueled by uncritical media coverage. He ends with, I am concerned that expectations for the timeline and scope of clinical utility of hESCs have outpaced the elds actual state of development and threaten to undermine its success. The warning is just as appropriate for all kinds of stem cells umbilical cord blood, neural stem cell, mesenchymal stem cells. Conclusions The maturation of new therapeutics takes decades. If one examines the history of any of the recent new thrusts in biomedicine recombinant DNA, monoclonal antibodies, gene therapy, or RNAithe vanguard treatments were introduced within a decade but 20 years passed before the full impact of the new form of medicine was felt widely in clinical medicine; for RNAi, we are still waiting for clinical success. Fifty years after the rst attempts at HSC transplantation, and even with all the improved understanding we now have of both HSCs and immunological mismatch, our success rates are still woefully inadequate. Although the development of novel stem cell-based therapies will benet greatly from the collective failures and acquired experience of marrow transplantation, our ignorance of the challenges of applying stem cells in distinct tissues with far greater anatomic complexity than the blood should give us pause as practitioners and inspire humility. Realistically, we should anticipate that new therapies based on stem cells for other tissues will likewise take decades to mature. In the short term, there will probably be more failures than successes, and one can only hope that the new eld of regenerative medicine can learn the lessons of the past and proceed with prudence and caution.
ACKNOWLEDGMENTS G.Q.D. is supported by grants from the NIH (R24DK092760, UO1-HL100001, RC4-DK090913, P50HG005550, and special funds from the ARRA stimulus package- RC2-HL102815), the Roche Foundation for Anemia Research, Alexs Lemonade Stand, Ellison Medical Foundation, Doris Duke Medical Foundation, and the Harvard Stem Cell Institute. G.Q.D. is an afliate member of the Broad Institute and an investigator of the Howard Hughes Medical Institute and the Manton Center for Orphan Disease Research. Disclosure: GQD is a member of the scientic advisory board and receives consulting fees and holds equity in the following companies that work with stem cells: Johnson & Johnson, Verastem, iPierian, and MPM Capital. REFERENCES Aboushwareb, T., and Atala, A. (2008). Stem cells in urology. Nat. Clin. Pract. Urol. 5, 621631. Amariglio, N., Hirshberg, A., Scheithauer, B.W., Cohen, Y., Loewenthal, R., Trakhtenbrot, L., Paz, N., Koren-Michowitz, M., Waldman, D., Leider-Trejo, L., et al. (2009). Donor-derived brain tumor following neural stem cell transplantation in an ataxia telangiectasia patient. PLoS Med. 6, e1000029. Atala, A. (2011). Tissue engineering of human bladder. Br. Med. Bull. 97, 81104. Baker, M. (2011). Stem-cell pioneer bows out. Nature 479, 459. Balsam, L.B., Wagers, A.J., Christensen, J.L., Kodis, T., Weissman, I.L., and Robbins, R.C. (2004). Haematopoietic stem cells adopt mature haematopoietic fates in ischaemic myocardium. Nature 428, 668673. Benraiss, A., and Goldman, S.A. (2011). Cellular therapy and induced neuronal replacement for Huntingtons disease. Neurotherapeutics 8, 577590. Bonner-Weir, S., and Weir, G.C. (2005). New sources of pancreatic beta-cells. Nat. Biotechnol. 23, 857861. Chambers, S.M., Fasano, C.A., Papapetrou, E.P., Tomishima, M., Sadelain, M., and Studer, L. (2009). Highly efcient neural conversion of human ES

Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc. 747

Cell Stem Cell

Perspective
and iPS cells by dual inhibition of SMAD signaling. Nat. Biotechnol. 27, 275280. o, A.M., Mills, J.A., Lin, H.C., Kotton, D.N., Cheng, X., Ying, L., Lu, L., Galva Shen, S.S., Nostro, M.C., Choi, J.K., et al. (2012). Self-renewing endodermal progenitor lines generated from human pluripotent stem cells. Cell Stem Cell 10, 371384. Choi, Y.H., Kurtz, A., and Stamm, C. (2011). Mesenchymal stem cells for cardiac cell therapy. Hum. Gene Ther. 22, 317. Copeland, N., Harris, D., and Gaballa, M.A. (2009). Human umbilical cord blood stem cells, myocardial infarction and stroke. Clin. Med. 9, 342345. DAmour, K.A., Bang, A.G., Eliazer, S., Kelly, O.G., Agulnick, A.D., Smart, N.G., Moorman, M.A., Kroon, E., Carpenter, M.K., and Baetge, E.E. (2006). Production of pancreatic hormone-expressing endocrine cells from human embryonic stem cells. Nat. Biotechnol. 24, 13921401. Daley, G.Q., and Scadden, D.T. (2008). Prospects for stem cell-based therapy. Cell 132, 544548. Dor, Y., and Melton, D.A. (2008). Facultative endocrine progenitor cells in the adult pancreas. Cell 132, 183184. Dor, Y., Brown, J., Martinez, O.I., and Melton, D.A. (2004). Adult pancreatic beta-cells are formed by self-duplication rather than stem-cell differentiation. Nature 429, 4146. Fasano, C.A., Chambers, S.M., Lee, G., Tomishima, M.J., and Studer, L. (2010). Efcient derivation of functional oor plate tissue from human embryonic stem cells. Cell Stem Cell 6, 336347. Freed, C.R., Breeze, R.E., Rosenberg, N.L., Schneck, S.A., Kriek, E., Qi, J.X., Lone, T., Zhang, Y.B., Snyder, J.A., Wells, T.H., et al. (1992). Survival of implanted fetal dopamine cells and neurologic improvement 12 to 46 months after transplantation for Parkinsons disease. N. Engl. J. Med. 327, 15491555. Freed, C.R., Greene, P.E., Breeze, R.E., Tsai, W.Y., DuMouchel, W., Kao, R., Dillon, S., Wineld, H., Culver, S., Trojanowski, J.Q., et al. (2001). Transplantation of embryonic dopamine neurons for severe Parkinsons disease. N. Engl. J. Med. 344, 710719. Gallina, P., Paganini, M., Lombardini, L., Mascalchi, M., Porrio, B., Gadda, D., Marini, M., Pinzani, P., Salvianti, F., Crescioli, C., et al. (2010). Human striatal neuroblasts develop and build a striatal-like structure into the brain of Huntingtons disease patients after transplantation. Exp. Neurol. 222, 3041. Gnecchi, M., Zhang, Z., Ni, A., and Dzau, V.J. (2008). Paracrine mechanisms in adult stem cell signaling and therapy. Circ. Res. 103, 12041219. Gourraud, P.A., Gilson, L., Girard, M., and Peschanski, M. (2012). The role of human leukocyte antigen matching in the development of multiethnic haplobank of induced pluripotent stem cell lines. Stem Cells 30, 180186. Grifn, M., Iqbal, S.A., and Bayat, A. (2011). Exploring the application of mesenchymal stem cells in bone repair and regeneration. J. Bone Joint Surg. Br. 93, 427434. Grskovic, M., Javaherian, A., Strulovici, B., and Daley, G.Q. (2011). Induced pluripotent stem cellsopportunities for disease modelling and drug discovery. Nat. Rev. Drug Discov. 10, 915929. Hargus, G., Cooper, O., Deleidi, M., Levy, A., Lee, K., Marlow, E., Yow, A., Soldner, F., Hockemeyer, D., Hallett, P.J., et al. (2010). Differentiated Parkinson patient-derived induced pluripotent stem cells grow in the adult rodent brain and reduce motor asymmetry in Parkinsonian rats. Proc. Natl. Acad. Sci. USA 107, 1592115926. Harris, D.T. (2009). Non-haematological uses of cord blood stem cells. Br. J. Haematol. 147, 177184. Hjelmgren, J., Ghatnekar, O., Reimer, J., Grabowski, M., Lindvall, O., Persson, U., and Hagell, P. (2006). Estimating the value of novel interventions for Parkinsons disease: an early decision-making model with application to dopamine cell replacement. Parkinsonism Relat. Disord. 12, 443452. Hyun, I., Lindvall, O., Ahrlund-Richter, L., Cattaneo, E., Cavazzana-Calvo, M., Cossu, G., De Luca, M., Fox, I.J., Gerstle, C., Goldstein, R.A., et al. (2008). New ISSCR guidelines underscore major principles for responsible translational stem cell research. Cell Stem Cell 3, 607609. Jacobson, L.O., Simmons, E.L., Marks, E.K., and Eldredge, J.H. (1951). Recovery from radiation injury. Science 113, 510511. Kebriaei, P., and Robinson, S. (2011). Treatment of graft-versus-host-disease with mesenchymal stromal cells. Cytotherapy 13, 262268. Keirstead, H.S., Nistor, G., Bernal, G., Totoiu, M., Cloutier, F., Sharp, K., and Steward, O. (2005). Human embryonic stem cell-derived oligodendrocyte progenitor cell transplants remyelinate and restore locomotion after spinal cord injury. J. Neurosci. 25, 46944705. Kode, J.A., Mukherjee, S., Joglekar, M.V., and Hardikar, A.A. (2009). Mesenchymal stem cells: immunobiology and role in immunomodulation and tissue regeneration. Cytotherapy 11, 377391. gler, G., Sensken, S., Airey, J.A., Trapp, T., Mu schen, M., Feldhahn, N., Ko Liedtke, S., Sorg, R.V., Fischer, J., Rosenbaum, C., et al. (2004). A new human somatic stem cell from placental cord blood with intrinsic pluripotent differentiation potential. J. Exp. Med. 200, 123135. Kriks, S., Shim, J.W., Piao, J., Ganat, Y.M., Wakeman, D.R., Xie, Z., CarrilloReid, L., Auyeung, G., Antonacci, C., Buch, A., et al. (2011). Dopamine neurons derived from human ES cells efciently engraft in animal models of Parkinsons disease. Nature 480, 547551. Kroon, E., Martinson, L.A., Kadoya, K., Bang, A.G., Kelly, O.G., Eliazer, S., Young, H., Richardson, M., Smart, N.G., Cunningham, J., et al. (2008). Pancreatic endoderm derived from human embryonic stem cells generates glucoseresponsive insulin-secreting cells in vivo. Nat. Biotechnol. 26, 443452. Lau, D., Ogbogu, U., Taylor, B., Stanski, T., Menon, D., and Cauleld, T. (2008). Stem cell clinics online: the direct-to-consumer portrayal of stem cell medicine. Cell Stem Cell 3, 591594. Ledford, H. (2011). Translational research: 4 ways to x the clinical trial. Nature 477, 526528. Lindvall, O., and Kokaia, Z. (2010). Stem cells in human neurodegenerative disorderstime for clinical translation? J. Clin. Invest. 120, 2940. Lindvall, O., Brundin, P., Widner, H., Rehncrona, S., Gustavii, B., Frackowiak, R., Leenders, K.L., Sawle, G., Rothwell, J.C., Marsden, C.D., et al. (1990). Grafts of fetal dopamine neurons survive and improve motor function in Parkinsons disease. Science 247, 574577. rklund, A., Brooks, D., Lindvall, O., Sawle, G., Widner, H., Rothwell, J.C., Bjo Brundin, P., Frackowiak, R., Marsden, C.D., Odin, P., et al. (1994). Evidence for long-term survival and function of dopaminergic grafts in progressive Parkinsons disease. Ann. Neurol. 35, 172180. Liu, S., Qu, Y., Stewart, T.J., Howard, M.J., Chakrabortty, S., Holekamp, T.F., and McDonald, J.W. (2000). Embryonic stem cells differentiate into oligodendrocytes and myelinate in culture and after spinal cord transplantation. Proc. Natl. Acad. Sci. USA 97, 61266131. Lorenz, E., Uphoff, D., Reid, T.R., and Shelton, E. (1951). Modication of irradiation injury in mice and guinea pigs by bone marrow injections. J. Natl. Cancer Inst. 12, 197201. Ma, Y., Tang, C., Chaly, T., Greene, P., Breeze, R., Fahn, S., Freed, C., Dhawan, V., and Eidelberg, D. (2010). Dopamine cell implantation in Parkinsons disease: long-term clinical and (18)F-FDOPA PET outcomes. J. Nucl. Med. 51, 715. Macchiarini, P., Jungebluth, P., Go, T., Asnaghi, M.A., Rees, L.E., Cogan, T.A., Dodson, A., Martorell, J., Bellini, S., Parnigotto, P.P., et al. (2008). Clinical transplantation of a tissue-engineered airway. Lancet 372, 20232030. McDonald, J.W., and Belegu, V. (2006). Demyelination and remyelination after spinal cord injury. J. Neurotrauma 23, 345359. McDonald, J.W., and Howard, M.J. (2002). Repairing the damaged spinal cord: a summary of our early success with embryonic stem cell transplantation and remyelination. Prog. Brain Res. 137, 299309. McDonald, J.W., Liu, X.Z., Qu, Y., Liu, S., Mickey, S.K., Turetsky, D., Gottlieb, D.I., and Choi, D.W. (1999). Transplanted embryonic stem cells survive, differentiate and promote recovery in injured rat spinal cord. Nat. Med. 5, 1410 1412. McKenna, D., and Sheth, J. (2011). Umbilical cord blood: current status & promise for the future. Indian J. Med. Res. 134, 261269.

748 Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc.

Cell Stem Cell

Perspective
Menasche, P. (2011). Cardiac cell therapy: lessons from clinical trials. J. Mol. Cell. Cardiol. 50, 258265. Nakajima, F., Tokunaga, K., and Nakatsuji, N. (2007). Human leukocyte antigen matching estimations in a hypothetical bank of human embryonic stem cell lines in the Japanese population for use in cell transplantation therapy. Stem Cells 25, 983985. Nakamura, T., Dhawan, V., Chaly, T., Fukuda, M., Ma, Y., Breeze, R., Greene, P., Fahn, S., Freed, C., and Eidelberg, D. (2001). Blinded positron emission tomography study of dopamine cell implantation for Parkinsons disease. Ann. Neurol. 50, 181187. Nicoleau, C., Viegas, P., Peschanski, M., and Perrier, A.L. (2011). Human pluripotent stem cell therapy for Huntingtons disease: technical, immunological, and safety challenges human pluripotent stem cell therapy for Huntingtons disease: technical, immunological, and safety challenges. Neurotherapeutics 8, 562576. Nishida, K., Yamato, M., Hayashida, Y., Watanabe, K., Yamamoto, K., Adachi, E., Nagai, S., Kikuchi, A., Maeda, N., Watanabe, H., et al. (2004). Corneal reconstruction with tissue-engineered cell sheets composed of autologous oral mucosal epithelium. N. Engl. J. Med. 351, 11871196. Nistor, G.I., Totoiu, M.O., Haque, N., Carpenter, M.K., and Keirstead, H.S. (2005). Human embryonic stem cells differentiate into oligodendrocytes in high purity and myelinate after spinal cord transplantation. Glia 49, 385396. Orlic, D., Kajstura, J., Chimenti, S., Jakoniuk, I., Anderson, S.M., Li, B., Pickel, J., McKay, R., Nadal-Ginard, B., Bodine, D.M., et al. (2001). Bone marrow cells regenerate infarcted myocardium. Nature 410, 701705. Peck, M., Gebhart, D., Dusserre, N., McAllister, T.N., and LHeureux, N. (2012). The evolution of vascular tissue engineering and current state of the art. Cells Tissues Organs (Print) 195, 144158. Pelosi, E., Castelli, G., and Testa, U. (2012). Human umbilical cord is a unique and safe source of various types of stem cells suitable for treatment of hematological diseases and for regenerative medicine. Blood Cells Mol. Dis., in press. Published online March 23, 2012. 10.1016/j.bcmd.2012.02.007. rklund, A., Gunn, R.N., Grasby, P.M., Rimoldi, O., Piccini, P., Brooks, D.J., Bjo Brundin, P., Hagell, P., Rehncrona, S., Widner, H., and Lindvall, O. (1999). Dopamine release from nigral transplants visualized in vivo in a Parkinsons patient. Nat. Neurosci. 2, 11371140. rklund, A., Brundin, P., Hagell, P., Ceravolo, R., Piccini, P., Lindvall, O., Bjo Oertel, W., Quinn, N., Samuel, M., Rehncrona, S., et al. (2000). Delayed recovery of movement-related cortical function in Parkinsons disease after striatal dopaminergic grafts. Ann. Neurol. 48, 689695. rklund, A., Oertel, W.H., Piccini, P., Pavese, N., Hagell, P., Reimer, J., Bjo Quinn, N.P., Brooks, D.J., and Lindvall, O. (2005). Factors affecting the clinical outcome after neural transplantation in Parkinsons disease. Brain 128, 2977 2986. Prasad, V.K., and Kurtzberg, J. (2009). Umbilical cord blood transplantation for non-malignant diseases. Bone Marrow Transplant. 44, 643651. Rama, P., Matuska, S., Paganoni, G., Spinelli, A., De Luca, M., and Pellegrini, G. (2010). Limbal stem-cell therapy and long-term corneal regeneration. N. Engl. J. Med. 363, 147155. Rocha, V., Labopin, M., Sanz, G., Arcese, W., Schwerdtfeger, R., Bosi, A., Jacobsen, N., Ruutu, T., de Lima, M., Finke, J., et al; Acute Leukemia Working Party of European Blood and Marrow Transplant Group; Eurocord-Netcord Registry. (2004). Transplants of umbilical-cord blood or bone marrow from unrelated donors in adults with acute leukemia. N. Engl. J. Med. 351, 2276 2285. Sanchez-Pernaute, R., Lee, H., Patterson, M., Reske-Nielsen, C., Yoshizaki, T., Sonntag, K.C., Studer, L., and Isacson, O. (2008). Parthenogenetic dopamine neurons from primate embryonic stem cells restore function in experimental Parkinsons disease. Brain 131, 21272139. Schwartz, S.D., Hubschman, J.P., Heilwell, G., Franco-Cardenas, V., Pan, C.K., Ostrick, R.M., Mickunas, E., Gay, R., Klimanskaya, I., and Lanza, R. (2012). Embryonic stem cell trials for macular degeneration: a preliminary report. Lancet 379, 713720. Shapiro, A.M., Lakey, J.R., Ryan, E.A., Korbutt, G.S., Toth, E., Warnock, G.L., Kneteman, N.M., and Rajotte, R.V. (2000). Islet transplantation in seven patients with type 1 diabetes mellitus using a glucocorticoid-free immunosuppressive regimen. N. Engl. J. Med. 343, 230238. Shapiro, A.M., Ricordi, C., Hering, B.J., Auchincloss, H., Lindblad, R., Robertson, R.P., Secchi, A., Brendel, M.D., Berney, T., Brennan, D.C., et al. (2006). International trial of the Edmonton protocol for islet transplantation. N. Engl. J. Med. 355, 13181330. Shi, M., Liu, Z.W., and Wang, F.S. (2011). Immunomodulatory properties and therapeutic application of mesenchymal stem cells. Clin. Exp. Immunol. 164, 18. Sun, H., Liu, W., Zhou, G., Zhang, W., Cui, L., and Cao, Y. (2011). Tissue engineering of cartilage, tendon and bone. Front. Med. 5, 6169. Tabar, V., Tomishima, M., Panagiotakos, G., Wakayama, S., Menon, J., Chan, B., Mizutani, E., Al-Shamy, G., Ohta, H., Wakayama, T., and Studer, L. (2008). Therapeutic cloning in individual parkinsonian mice. Nat. Med. 14, 379381. Taylor, C.J., Bolton, E.M., Pocock, S., Sharples, L.D., Pedersen, R.A., and Bradley, J.A. (2005). Banking on human embryonic stem cells: estimating the number of donor cell lines needed for HLA matching. Lancet 366, 2019 2025. Taylor, C.J., Bolton, E.M., and Bradley, J.A. (2011). Immunological considerations for embryonic and induced pluripotent stem cell banking. Philos. Trans. R. Soc. Lond. B Biol. Sci. 366, 23122322. Thomas, E.D. (2005). Bone marrow transplantation from the personal viewpoint. Int. J. Hematol. 81, 8993. Thomas, E.D., Lochte, H.L., Jr., Lu, W.C., and Ferrebee, J.W. (1957). Intravenous infusion of bone marrow in patients receiving radiation and chemotherapy. N. Engl. J. Med. 257, 491496. Tian, H., Bharadwaj, S., Liu, Y., Ma, P.X., Atala, A., and Zhang, Y. (2010). Differentiation of human bone marrow mesenchymal stem cells into bladder cells: potential for urological tissue engineering. Tissue Eng. Part A 16, 17691779. Tobita, M., Orbay, H., and Mizuno, H. (2011). Adipose-derived stem cells: current ndings and future perspectives. Discov. Med. 11, 160170. Tongers, J., Losordo, D.W., and Landmesser, U. (2011). Stem and progenitor cell-based therapy in ischaemic heart disease: promise, uncertainties, and challenges. Eur. Heart J. 32, 11971206. Trounson, A., Thakar, R.G., Lomax, G., and Gibbons, D. (2011). Clinical trials for stem cell therapies. BMC Med. 9, 52. Tsai, R.J., Li, L.M., and Chen, J.K. (2000). Reconstruction of damaged corneas by transplantation of autologous limbal epithelial cells. N. Engl. J. Med. 343, 8693. Tsubota, K., Satake, Y., Kaido, M., Shinozaki, N., Shimmura, S., BissenMiyajima, H., and Shimazaki, J. (1999). Treatment of severe ocular-surface disorders with corneal epithelial stem-cell transplantation. N. Engl. J. Med. 340, 16971703. Vierbuchen, T., and Wernig, M. (2011). Direct lineage conversions: unnatural but useful? Nat. Biotechnol. 29, 892907. Wernig, M., Zhao, J.P., Pruszak, J., Hedlund, E., Fu, D., Soldner, F., Broccoli, V., Constantine-Paton, M., Isacson, O., and Jaenisch, R. (2008). Neurons derived from reprogrammed broblasts functionally integrate into the fetal brain and improve symptoms of rats with Parkinsons disease. Proc. Natl. Acad. Sci. USA 105, 58565861. Williams, A.R., Trachtenberg, B., Velazquez, D.L., McNiece, I., Altman, P., Rouy, D., Mendizabal, A.M., Pattany, P.M., Lopera, G.A., Fishman, J., et al. (2011). Intramyocardial stem cell injection in patients with ischemic cardiomyopathy: functional recovery and reverse remodeling. Circ. Res. 108, 792796. Wilson, J.M. (2009). Medicine. A history lesson for stem cells. Science 324, 727728. Yamanaka, S. (2012). Induced pluripotent stem cells: past, present, and future. Cell Stem Cell 10, this issue, 675681. Zhou, Q., Brown, J., Kanarek, A., Rajagopal, J., and Melton, D.A. (2008). In vivo reprogramming of adult pancreatic exocrine cells to beta-cells. Nature 455, 627632.

Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc. 749

Cell Stem Cell

Perspective
Expanding the Boundaries of Embryonic Stem Cells
Uri Ben-David,1 Oded Kopper,1 and Nissim Benvenisty1,*
1Stem Cell Unit, Department of Genetics, Silberman Institute of Life Sciences, The Hebrew University, Jerusalem 91904, Israel *Correspondence: nissimb@cc.huji.ac.il DOI 10.1016/j.stem.2012.05.003

The boundaries of embryonic stem cell (ESC) research have extended considerably in recent years in several important ways. Alongside a deeper understanding of the pluripotent state, ESCs have been successfully integrated into various elds, such as genomics, epigenetics, and disease modeling. Signicant progress in cell fate control has pushed directed differentiation and tissue engineering further than ever before and promoted clinical trials. The geographical distribution of research activity has also expanded, especially for human ESCs. This review outlines these developments and future challenges that remain.
Introduction The isolation of mouse embryonic stem cells (mESCs) from mouse blastocysts three decades ago dramatically advanced the eld of mouse genetics, resulting in the groundbreaking technology of gene targeting. The impact of the derivation of human ESCs (hESCs; Thomson et al., 1998) almost two decades later was just as dramatic, placing the study of pluripotent stem cells at the forefront of biomedical research. Indeed, in recent years as the ethical constraints associated with hESC research have become a less prominent topic of debate, the scientic boundaries of this eld have expanded considerably. In this Perspective, we cover several aspects of this expansion and discuss the major issues that have occupied the eld in recent years. In the past 5 years, core research on ESCs, i.e., research into their self-renewal and differentiation capacities, took advantage of state-of-the-art genome-wide technologies to extend our understanding of the pluripotent state. This increasing understanding has allowed the ESC eld to reach beyond the boundaries of the laboratory, toward the fulllment of its promise for regenerative medicine, with increasing numbers of preclinical and, more recently, clinical trials performed with ESC-derived cells. This maturation, in turn, facilitated the entry of new players into the ESC eld. A growing number of physicians, regulatory agencies, and industrial companies are joining the academically driven journey of ESCs toward the clinic. During the past few years, ESCs have also gone beyond their traditional role as a tool for studying pluripotency and have become a fundamental player in various domains of molecular biology; more and more studies make use of mESCs and hESCs for answering general questions in genetics, epigenetics, and cell biology and for developing novel technologies whose applications may go beyond pluripotent cells. The boundaries of ESC research have also spread in the literal geographic sense across political borders, with laboratories from all over the world making signicant contributions to the eld. The Global Village: ESC Research Worldwide Recent years have seen increasing interest in ESCs throughout the world. In the past 5 years, laboratories from 50 different countries published papers about ESCs, more than doubling the total output of original research papers in the eld, relative
666 Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc.

to the prior 5 years. This increase is even more impressive when considering research on hESCs: between 2007 and 2011, laboratories from 41 countries published papers using hESCs, compared to only 27 countries between 2002 and 2006 (Figure 1), more than tripling the number of papers on these cells. This expansion can probably be attributed, at least in part, to the establishment of clear guidelines for hESC research in many countries. Although the increase in the total number of publications and in the number of countries involved is especially remarkable for hESCs, there is a similar trend for nonhuman ESC research, mostly on mESCs, which also expanded considerably over the same time frame (Figure 1). Importantly, this growth cannot be credited merely to the breakthrough of induced pluripotent stem cells (iPSCs), as papers that involve iPSCs were not included in this analysis. The growing global interest in ESC research is also reected by the relative contribution coming from various parts of the world. While the United States is still the most prolic country in the eld, the relative share of papers published by laboratories from Europe and Asia has become much more signicant (Figure 1). For example, China has doubled its share in the total publication count, in both human and nonhuman ESC-related research. In summary, the increasing number of articles that come out every year, the number of contributing countries, and the relative contribution of these countries all suggest that ESC research is on the rise. Inside the Network: Understanding Pluripotency When examining the pluripotency literature from the last few years, one is overwhelmed by the quantity and quality of genome-wide studies performed in attempts to deconstruct the pluripotent state. It seems that no cutting-edge technology has gone unnoticed by the ESC eld, which harnessed these state-of-the-art tools to uncover the global state of pluripotency (e.g., its genome, transcriptome, proteome, methylome, etc.), in what could be aptly described as the Omics era of ESC research (Loh et al., 2011). The large-scale genome-wide studies exposed complex and dynamic multilayered regulation involving transcriptional networks, chromatin modications, and posttranscriptional regulation (Ng and Surani, 2011; Orkin and Hochedlinger, 2011; Young, 2011) (Figure 2).

Cell Stem Cell

Perspective

Figure 1. ESC Research Distribution throughout the World


World maps comparing the distribution of stem cell research throughout the world between two 5 year periods: 20022006 and 20072011. The numbers of publications involving human and nonhuman ESCs were assessed separately and are thus presented in separate maps. Nonhuman ESCs are mostly, but not exclusively, mouse ESCs. The maps are color-coded by the absolute number of articles published by laboratories from each country. The total number of contributing countries during the examined years appears in the upper right side of each map. Articles dealing with iPSCs were removed from the analysis. Quantication of articles was carried out using ISI Web of Science (http://apps.isiknowledge.com).

Transcriptional Networks The core transcriptional regulatory network of pluripotency was investigated in a series of studies using various genome-wide chromatin-IP based technologies (ChIP-on-chip, ChIP-PET, ChiP-seq, and biochip). Whereas the rst studies identied the key players of the network and its general architecture (Boyer et al., 2005; Loh et al., 2006), recent analyses rened our understanding of this network, revealing a set of distinct-yetintimately-connected modules that cooperate and regulate each other (Chen et al., 2008; Kim et al., 2008, 2010). These studies divided the pluripotent network to its two main compartments, the Oct4-centric and the Myc-centric modules; revealed the interactions within and between each of these modules; highlighted the importance of coregulation and autoregulation for the proper function of the network; and integrated novel signaling pathways into it. More recently, the core pluripotency genes were also shown to control germ layer fate choice, extending the original role of the pluripotent network beyond the maintenance of self-renewal (Thomson et al., 2011). Comparison of the mouse and human pluripotent networks revealed, quite surprisingly, that species-specic transposable elements have considerably altered the transcriptional pluripotent circuitry, so that in each species the same core factors bind a distinct set of TF-binding sites and play distinct roles in pluripotency regulation (Kunarso et al., 2010; Wang et al., 2012). Noncoding RNAs Another recently identied layer of pluripotency regulation is that of noncoding RNAs: micro RNAs (miRNAs) and large intergenic

noncoding RNAs (lincRNAs). The importance of miRNAs in ESCs was demonstrated in several studies. Like their unique characteristic mRNA signature, ESCs exhibit a dened characteristic miRNA signature (Marson et al., 2008). Global loss of miRNAs resulted in defects in both self-renewal and differentiation, whereas specic miRNAs were found to regulate ESC cell cycle, expression of pluripotency factors, and differentiation (Martinez and Gregory, 2010). The miRNAs themselves are, in turn, regulated by pluripotency factors such as Lin28 (Viswanathan et al., 2008), demonstrating the crosstalk between layers of pluripotency regulation. More recently, hundreds of lincRNAs that are involved in the control of the pluripotent state were also discovered (Guttman et al., 2009). Many of them were reported to be bound by Oct4 and Nanog in their promoter regions, directly integrating them into the core pluripotency circuitry (Guttman et al., 2009). Specic lincRNAs were already reported to be essential for pluripotency (Sheik Mohamed et al., 2010), but much is left to be discovered regarding their role. Proteomics Proteomic approaches have recently revealed interactions between the core pluripotency proteins, contributing to the growing understanding of pluripotency (Pardo et al., 2010; Wang et al., 2006). Further studies added phosphorylation dynamics as another layer of regulation in ESCs, identifying signicant changes in the phosphoproteome of ESCs during their differentiation (Brill et al., 2009; Van Hoof et al., 2009). The study of protein interactions also shed light on the important
Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc. 667

Cell Stem Cell

Perspective
Figure 2. Major Achievements and Challenges of ESC Research
A schematic representation of the main topics in ESC research in recent years. The major advancements in each of these topics, and the challenges that lie ahead, are elaborated in the text.

role of chromatin in pluripotency. Direct and indirect interactions between core transcription factors and chromatin modiers/ remodelers were shown to play a crucial part in maintaining the open chromatin state, which is unique for pluripotent cells (Gaspar-Maia et al., 2011). Moreover, there are direct regulatory interactions between these transcription factors and chromatin remodelers (Ang et al., 2011; Gaspar-Maia et al., 2009), further stressing their importance for the pluripotent state. Epigenetic Control One of the most remarkable achievements in recent pluripotency research is its intimate involvement in groundbreaking discoveries in epigenetics, elucidating novel layers of regulation in the complex control of the pluripotent state (Meissner, 2010). At the level of DNA methylation, various studies characterized ESC-specic methylation proles and linked them directly to the core transcriptional networks of ESCs (Fouse et al., 2008; Meissner et al., 2008). A recent study applied Methyl-seq technology to map the ESC methylome at a single-base resolution, revealing a novel class of DNA methylation at non-CpG sites (Lister et al., 2009). These unique non-CpG methylations are enriched in exons of highly expressed genes (Lister et al., 2009) but appear to be dispensable for pluripotency (Ziller et al., 2011), so their role in regulation of gene expression is still unclear. In another major discovery, a novel type of DNA methylcytosine modication was recently discovered in ESCs in which the Tet-family proteins Tet1 and Tet2 transform methylated cytosines into 50 -hydroximethylcytosines (5hmC) (Ito et al., 2010; Koh et al., 2011; Pastor et al., 2011; Tahiliani et al., 2009). This series of studies demonstrated that Tet1 and Tet2 affect self-renewal and differentiation of mESCs. Another layer of epigenetic regulation that has been studied extensively in ESCs is histone modications. ESCs are characterized by bivalent domains generated by the co-occupation of the transcription start sites of genes that control cell fate decisions by the activating mark H3K4me3 and the repressive mark H3K27me3 (Bernstein et al., 2006). This phenomenon has drawn much interest in the eld, as the bivalent domains are considered
668 Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc.

to poise genes for their rapid activation upon differentiation. Recent studies have improved our understanding of the molecular mechanism that underlies these unique domains, demonstrating the important role of the polycomb group (PcG) proteins PRC1 and PRC2 (Margueron and Reinberg, 2011), and introducing new players that participate in the generation and maintenance of this delicate balance of modications (Margaritis and Holstege, 2008; Pasini et al., 2010; Peng et al., 2009). Other histone modications such as H3K9 methylation have also been studied in ESCs (Wen et al., 2009), and together these studies begin to decipher what seems to be an ESC-specic histone code. Having discovered elements that participate in regulating pluripotency, the main challenge that lies ahead seems to be the integration of all these layers, modules and subnetworks into a consolidated regulatory circuitry. Attempts to describe the crosstalk between different layers of regulation have already been reported, connecting, for example, gene expression with DNA methylation (Bock et al., 2011), transcription with histone modications and protein levels (Lu et al., 2009), or DNA methyl et al., 2006). With the ation with histone modication (Vire increasing focus on combinatorial regulation and on crosstalk between network elements, we expect many more exciting connections to be revealed in the future. One but Not the Same: Emerging Variability of the Pluripotent State The increasing understanding of the molecular mechanisms that govern pluripotency inspired fruitful discussions regarding the nature of the pluripotent state and helped rene the very denition of the term pluripotency. The similar yet distinct pluripotent states of various types of ESCs were coined different avors of pluripotency (Buecker and Geijsen, 2010), and as our understanding of these pluripotent states is becoming more solid, so does our control of the transitions between them. Developmental Identity of PSCs The focus of investigation in this area is the difference between mouse embryonic and epiblast stem cells (ESCs and EpiSCs, respectively), and between mESCs and hESCs. MESCs are derived from the inner cell mass of blastocysts, while mouse EpiSCs are derived from postimplantation epiblasts (Brons et al., 2007; Tesar et al., 2007); consequently, these cell types differ in their morphology, culture requirements, developmental potential, expression prole, and amenability to homologous recombination. These differences led to the emergence of the

Cell Stem Cell

Perspective
ve (or ground-state) versus primed pluripoconcepts na tent cells (Nichols and Smith, 2009). Recent work showed that mouse cells can acquire metastable pluripotent states that could be interconverted by endogenous genetic determinants or by exogenous factors (Hanna et al., 2009). HESCs share many features with mESCs but, intriguingly, they also share some of their characteristics with mouse EpiSCs, suggesting that they may represent the primed ESC state and therefore may not harbor the full developmental potenve ESCs. Recently, several groups reported the derivatial of na tion of hESCs and hiPSCS with biological properties similar to those of mESCs (Buecker et al., 2010; Hanna et al., 2010; Li et al., 2009; Wang et al., 2011a). These hESCs exhibited morphology, growth properties, expression proles and signaling dependence that were comparable to those of mESCs, but they were not stable in the absence of genetic manipulave hESCs are indeed superior to tions. Whether these na their primed counterparts in terms of developmental potential remains to be determined; however, the fact that culture conditions are sufcient to interconvert between pluripotent states, both in mESCs and in hESCs, indicates that plasticity in the pluripotent state is more widespread than was previously appreciated. Heterogeneity of PSCs The variability between ESCs has received attention from other directions as well; in recent years, ESCs were shown to be more heterogeneous than previously thought. Both intraculture and interculture heterogeneity exist: within undifferentiated cultures of mESCs and hESCs and of mouse EpiSCs, distinct subpopulations were identied, differing in their expression of molecular markers and in their differentiation potential and therefore presumed to correspond to distinct developmental stages (Canham et al., 2010; Han et al., 2010; Martinez Arias and Brickman, 2011; Stewart et al., 2010; Toyooka et al., 2008). Between undifferentiated hESC lines, large-scale comparisons revealed differences in gene expression and in differentiation, suggesting that not all ESCs lines are equally suitable for any given purpose (Adewumi et al., 2007; Bock et al., 2011). ESC to Every Lab: Advances in ESC Derivation and Propagation The techniques for deriving, propagating, and banking ESCs have signicantly improved in recent years, and these types of advances are key for moving ESCs toward the clinic (Figure 2). Much progress has been made in adapting culture conditions to enable rapid and efcient thawing, passaging, and cryopreservation of hESCs. The most notable discovery in this regard was probably that the Rho-associated kinase (ROCK)-inhibitor Y-27632 permits the survival of dissociated hESCs (Watanabe et al., 2007). Follow-up studies uncovered the molecular mechanism that underlies the high sensitivity of hESCs to dissociation (Chen et al., 2010; Ohgushi et al., 2010) and also utilized this inhibitor for deriving and propagating hESCs in suspension (Amit et al., 2010; Steiner et al., 2010). Several groups have also directed much effort at determining the components of dened media that would enable feederfree growth of hESCs (Akopian et al., 2010) and eliminating animal products from such media, thus making it xeno-free (Lei et al., 2007; Valamehr et al., 2011). These efforts culminated in the generation of good manufacturing practice (GMP) clinical-grade hESCs (Unger et al., 2008). In order to standardize the use of hESCs in biomedical research and, eventually, in the clinic, consensus guidelines for banking and supply of hESCs were proposed (International Stem Cell Banking Initiative, 2009). As human ESC lines are now derived on a weekly basis, the ethnical diversity within the human ESC pool has greatly expanded so that it currently represents dozens of different ethnic backgrounds (Amps et al., 2011). This diversity will be important for the study of ethnically relevant diseases, for the removal of confounding effects due to specic genetic backgrounds, and for the banking of hESCs that would be compatible with as large a population as possible. Apart from the abovementioned advances in the culture of hESCs, derivation and culture techniques were also developed in recent years for ESCs of various species, expanding the repertoire of pluripotent stem cells available for research. In addition to mouse, monkey, and human ESCs, in the past 5 years ESC lines were derived from multiple species including rabbit, canine, andmost importantlyrat (Martins-Taylor and Xu, 2010). These stem cell types should enhance our understanding of the pluripotent state and, especially in the case of the rat, enable the generation of novel model animals for studying human disease. Eyes on the Target: ESCs in Regenerative Medicine Many clinical conditions such as neurodegenerative disorders, diabetes, and some forms of heart and hepatic failure are caused by loss of functionality or insufcient quantity of a particular cell type. The potential of hESCs to differentiate into any cell type of the human body raised the hope for treatment of these clinical conditions and has thus drawn hESCs into the public spotlight. Indeed, exciting recent progress is paving the hESC path into the practice of regenerative medicine (Figure 2). Directing Differentiation For hESCs to live up to expectations, it will be essential to control their differentiation course. One of the most efcient strategies designed to control a pluripotent cell fate is the recapitulation of developmental steps through which cells assume a specic fate during normal development (Murry and Keller, 2008). The rst step in the differentiation of a pluripotent stem cell is transition into one of the three embryonic germ layers: the ectoderm, mesoderm, and endoderm. Multiple studies that applied knowledge of developmental biology to ESC differentiation demonstrated that despite the known differences between hESCs and mESCs (Nichols and Smith, 2009), the signaling pathways that control primary differentiation are very similar (Cohen and Melton, 2011; Murry and Keller, 2008). After acquiring their initial lineage identity, application of various growth factors and culture conditions can continue to direct the cells along multiple differentiation paths. Recently, small molecules have become more useful in differentiation protocols, presenting an appealing alternative to recombinant growth factors, especially when considering the potential for mass production of cells for clinical use (Rubin, 2008). Small molecules are less expensive and more stable than recombinant proteins and can contribute to the development of fully dened
Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc. 669

Cell Stem Cell

Perspective
media. They also allow for more straightforward minimization of differences between batches of reagents, which can reduce experimental variance and help establish reproducible differentiation protocols. However, in some cases there are no known chemical compounds that modulate the desired signaling pathway. In such cases, unbiased screening offered a potential solution. In a landmark study, thousands of chemical compounds were screened in order to identify molecules that can replace Activin A in the induction of denitive endoderm (DE) (Borowiak et al., 2009). Two such small molecules were uncovered in both mESCs and hESCs. Additional studies used the unbiased screening approach to efciently produce desired ESC derivatives such as pancreatic progenitor cells and cardiomyocytes (Chen et al., 2009; Takahashi et al., 2003). Advances in the genetic manipulation of ESCs (Giudice and Trounson, 2008) have enabled the successful generation of several uorescent reporter hESC lines that are extremely useful for a variety of applications. They can be used to distinguish undifferentiated hESCs from their differentiated derivatives (Eiges et al., 2001) or to visualize the appearance of desired differentiated cells in culture (Lavon et al., 2004; Singh Roy et al., 2005). Moreover, reporter cell lines enable the isolation and analysis of various cell populations using uorescence-activated cell sorting (FACS) even without previous knowledge of specic cell surface markers or available antibodies. The reporter cell lines are also valuable for microscopy-based high-throughput screening and can therefore assist with the optimization of differentiation protocols. The engraftment, survival and integration of transplanted cells can be tracked more easily using such reporter hESCs. Although they are mostly useful for in vitro analyses and preclinical studies, reporter cell lines may be benecial for determining the fate and function of transplanted cells in clinical trials as well (Ellis et al., 2010). The tremendous progress in applying understanding of embryognesis to differentiation protocols has enabled the generation of diverse cell types in vitro, including highly specied cells such as retinal pigment epithelium (RPE) (Idelson et al., 2009), mechanosensitive hair cells (Oshima et al., 2010), and primordial germ cells (Hayashi et al., 2011). However, there are still many challenges on the way to the ultimate goal of cells on demand. Differentiation is a stepwise process, passing through several intermediate progenitor cells on the way to a fully-differentiated cell of interest. The rst differentiation step into the desired germ layer is usually the most efcient step, and the differentiation efciency often decreases with each step of the protocol (Cohen and Melton, 2011). As a result, the end product is usually a heterogeneous cell population that contains only low percentage of the specic cell type. A major challenge will therefore be to improve the differentiation efcacy and design reliable methods for isolating the desired cell populations. An appealing alternative can be to propagate intermediate progenitor cells, as was recently demonstrated with denitive endoderm progenitor cells (Cheng et al., 2012). Another related challenge is the reproducibility of differentiation protocols. Different batches of reagents and slight differences in cell culture techniques sometime make it very difcult to recapitulate differentiation protocols successfully and with comparable efciency rates to those orig670 Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc.

inally reported. The distinct differentiation propensity of different hESC lines (Bock et al., 2011) adds to this complexity, because it compromises the generalization of some differentiation protocols. Of note, most ESC-derived differentiated cells are not fully mature, and their in vitro maturation is another obstacle that awaits a solution (though, in some cases, maturation does take place in vivo after cell transplantation (Hayashi et al., 2011; Kriks et al., 2011; Kroon et al., 2008)). Generation of Complex Differentiated Cell Types The main focus of most directed differentiation experiments is to maximize the derivation of one desired cell type for cell replacement therapy. The replacement of complex tissues, however, presents a greater challenge that involves differentiation into several cell types with a three-dimensional (3D) organization. Recently, two different approaches have made progress toward meeting this challenge. The rst approach makes use of the potential of ESCs to respond to extrinsic signals and recapitulate developmental cell fate decisions to generate tissues in a dish. Several recent studies demonstrated that ESCs can not only differentiate to all cell types, but also generate organizer cells that may affect the fate of adjacent cells during embryogenesis (Sharon et al., 2011). Moreover, the cells possess in vitro self-organization capacity and can therefore generate organized and complex 3D tissues, such as cortical structures (Eiraku et al., 2008), the optic cup (Eiraku et al., 2011), adenohypophysis (Suga et al., 2011), and intestinal tissue (Spence et al., 2011). The second approach is the in vivo generation of tissues and organs using chimeric animals. In a remarkable experiment, xenogeneic organ complementation was achieved when rat pluripotent stem cells were injected into mouse blastocysts that lacked the Pdx1 gene. As a result, the rat-mouse chimeras pancreas was composed exclusively of rat cells, demonstrating the feasibility of organ generation through interspecies chimeras (Kobayashi et al., 2010). As the generation of viable chimeras from a nonhuman primate was recently demonstrated (Tachibana et al., 2012), these breakthroughs raise fascinating possibilities regarding organ generation but also raise signicant technical, legal, and ethical questions. Preclinical Evaluation Before transplantation of differentiated cells into patients, it is essential to conduct pre-clinical trials to demonstrate the integration capacity and functionality of the cells in animal models. Finding an appropriate animal model and analyzing the mechanism that underlies the observed improvements can be a challenging task. Demonstrating the functionality of differentiated cells in an animal model that entirely lacks the relevant cell type is the most stringent and straightforward approach; for example, several groups demonstrated the potential of hESCderived b cells (Kroon et al., 2008) and RPE cells (Idelson et al., 2009) to functionally replace their in vivo equivalents. Importantly, most in vivo transplantation experiments are currently conducted in murine models, and the scalability of these assays thus remains an open question. Tackling this issue, a recent study demonstrated the in vivo survival, integration, and function of hESC-derived dopaminergic neurons in rat and mouse Parkinsons disease models and went on to show their survival and integration in parkinsonian adult rhesus monkeys (Kriks et al., 2011).

Cell Stem Cell

Perspective
Clinical Studies The progress in various aspects of ESC technology, discussed above, has established a solid platform for therapeutic implementation. Although some safety issues as well as differentiation and pre-clinical challenges are still unresolved, it seems as though hESCs are starting to fulll their promise, as reected by a couple of ongoing clinical trials. In these pioneering clinical trials, differentiated derivatives of hESCs were transplanted into patients suffering from various clinical conditions: spinal cord injury, dry age-related macular degeneration (AMD), and Stargardts disease (Goldring et al., 2011; Trounson et al., 2011). This encouraging progress was hindered by the surprising decision of Geron, the rst company that took ESCs to the clinic, to terminate all ESC-related experiments and ongoing clinical trials. This blow, however, was somewhat balanced by the positive preliminary report of Advanced Cell Technology (ACT), describing the results of their rst hESC-based clinical trial in human patients (Schwartz et al., 2012). In this report, hESCderived RPE cells were transplanted into patients with either AMD or Stargardts disease. Four months after the implantation, the survival and engraftment of the cells, together with functional visual improvement, were identied, whereas no signs of tumorigenicity or immune rejection were observed (Schwartz et al., 2012). These preliminary clinical results, together with the ever-improving differentiation protocols into highly specied, complex and mature cell types, suggest that a new era of hESC-based therapy might not be very far away. Better Safe Than Sorry: The Tumorigenicity and Immunogenicity of ESCs As human ESC products get closer to the bedside, safety issues have become a serious hurdle that must be overcome before ESC-derived cells can be routinely injected into patients (Goldring et al., 2011). The major safety problem the eld is currently facing is the potential tumorigenicity of the cells, mainly due to residual undifferentiated cells. Several approaches to selectively remove undifferentiated cells from culture have been suggested in order to solve this problem, including the use of genetic labeling, ablation of teratoma-specic genes, sorting out pluripotent stem cells based on antibodies against pluripotent-specic molecules, and specic cytotoxic antibodies (Ben-David and Benvenisty, 2011). Most recently, biomarkers unique to human pluripotent stem cells were used to eliminate pluripotent stem cells from mixed populations (Tang et al., 2011; Wang et al., 2011b). Concomitantly, efforts are being made to characterize the most likely tumors, i.e., teratomas and teratocarcinomas, in an attempt to prevent their formation (Blum et al., 2009; for a discussion of the appropriate terminology for these tumors, see Damjanov and Andrews, 2007; Lensch and Ince, 2007). Although no tumor formation was reported in the preliminary report of the rst clinical trial with ESC-derived cells (Schwartz et al., 2012), the tumorigenicity risk has not been resolved yet and remains a concern that limits the number of cells injected into human patients. Another concern that may affect the safety of ESC-based treatments is their genomic stability in culture. In recent years, large-scale comparison studies revealed recurrent genomic aberrations in hESCs and began to pinpoint the genes that drive these frequent aberrations (Amps et al., 2011; Baker et al., 2007; Mayshar et al., 2010). Some of these aberrations may be associated with oncogenic transformation, and thus would increase the tumorigenicity of the cells (Baker et al., 2007; Ben-David et al., 2011; Lefort et al., 2008). Until strategies to prevent the accumulation of such genomic alterations in ESC cultures are developed, the genomic integrity of the cells needs to be monitored carefully prior to their clinical application. It is worth noting that the genomic instability of ESCs is deleterious for additional reasons: it may compromise their differentiation propensity, the functionality of the differentiated cells, and their usefulness for disease modeling and drug screening. The immunogenicity of ESCs and of ESC-derived cells is a third issue related to the safetyand, obviously, to the successof ESC-based treatments (Kadereit and Trounson, 2011). HLA matching is a major hurdle for hESC-based therapies, especially for treatments of tissues that are not immune privileged. Recent preclinical and clinical trials applied pharmacological immunosuppression to avoid graft rejection (Kriks et al., 2011; Schwartz et al., 2012); however, in one case the immunosuppression was shown to be incomplete (Kriks et al., 2011) and in another the patient did not comply with the immunosuppression regimen (Schwartz et al., 2012), indicating that this important issue is not fully resolved. Although some strategies for tolerance induction have been suggested (Robertson et al., 2007), the main strategy for circumventing this obstacle remains the assembly of ESCs with diverse major histocompatability complex (MHC) haplotypes in the ESC banks that are founded these days throughout the world. ESCaping Drug Attrition: ESCs in the Service of Toxicology Assuring the safe use of new drugs requires the analysis of their safety in the developing embryo and in the adult. In recent years, hESCs have begun to play a major role in toxicology assays (Laustriat et al., 2010). The pluripotent capacity of the cells and their ability to differentiate into many cell types make them a valuable pharmacological tool in three main ways (Figure 2). First, hESCs can be used for screening of teratogens, compounds that are selectively detrimental for the embryo or the fetus, based on the ability of hESCs to mimic early stages of human development (Mayshar et al., 2011; West et al., 2010). The second aspect is drug metabolism, which primarily takes place in the liver. The capability of hESCs to differentiate into fairly mature hepatocytes (Agarwal et al., 2008; Basma et al., 2009) may make them a suitable tool for testing the hepatic metabolism of potential drugs. The third aspect is tissue toxicity, which is based on the growing ability to obtain rather pure populations of clinically relevant cells, such as cardiomyocytes and neurons, from hESCs. Using these differentiated cells for cardioand neurotoxicity screens will allow tissue-specic assessment of drug toxicity (Mandenius et al., 2011). During the past few years, good progress has been made in all three fronts, and it is predicted that in the future hESCs will be routinely used by the pharmacological industry (Wobus and Loser, 2011). Diseases in a Dish: Modeling Human Genetic Disorders with ESCs Mouse ESCs are the major tool for generating mouse models of human diseases, and the contribution of transgenic mouse to
Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc. 671

Cell Stem Cell

Perspective
understanding human disease has been tremendous. However, in many disorders the mouse models fail to recapitulate the human phenotypes. Therefore, hESCs represent an alternative tool for modeling human diseases, by introducing mutations to normal ESCs (Urbach et al., 2004) (Figure 2). In the past 5 years, many different methodologies have been used to manipulate the genome of hESCs, including homologous recombination by plasmids and BAC constructs (Song et al., 2010) or by zinc nger and TALE nucleases (Hockemeyer et al., 2009; Hockemeyer et al., 2011; Leavitt and Hamlett, 2011), and the use of RNAi to knock down specic genes (Tulpule et al., 2010). While both the efciency and the specicity of these methods have improved in recent years, off-target activity remains a concern, as it may introduce collateral damage that may jeopardize their applicability for disease modeling. An alternative to induction of mutations in normal hESCs is the isolation of genetically aberrant hESCs from blastocysts carrying genetic diseases. Screening for diseased embryos by the analysis of single blastomeres at the preimplantation stage is becoming a more common methodology. Thus, preimplantation genetic screening (PGS) is used to identify chromosomal aberrations in human embryos, and hESCs can be derived from such aneuploid embryos, generating in-vitro models for chromosomal disorders such as Down syndrome (Biancotti et al., 2010). In addition, preimplantation genetic diagnosis (PGD) is conducted to screen for embryos that carry monogenic disorders, and multiple ESC lines were derived from such mutated blastocysts (reviewed in Ben-Yosef et al., 2008). Some of the derived disease model hESCs were analyzed to identify disease characteristic phenotypes, either in the undifferentiated state or after their differentiation in culture (Colman and Dreesen, 2009). In developmental disorders, these models also enabled the characterization of phenotypes that are unique to the embryonic stage of the cells. Once dening a phenotype of interest, these powerful models can potentially serve to identify novel drugs that would enable treatment of currently untreatable disorders. The abundance of available cells could be exploited either for testing drugs that target known candidate genes or for performing unbiased high-throughput screens with libraries of varied molecular entities. It is clear that research on hESCs paved the way to the analysis of human disorders using human iPSCs (Robinton and Daley, 2012). It is also evident that the availability of somatic cells from practically any human disorder has made the generation of such models in human iPSCs very accessible for most labs. There is some indication, however, that iPSC models may be affected by an epigenetic memory from the somatic cells and thus might be inferior to ESCs in reecting developmental aspects of the disease (Urbach et al., 2010). Nonetheless, iPSCs are clearly becoming the system of choice for disease modeling, and most disease models are already generated using this methodology (Robinton and Daley, 2012). The New Kids on the Block: Novel Types of ESCs Traditionally, hESCs are derived from blastocysts of IVF embryos and thus represent the outcome of the natural fertilization process. Recently, however, new types of ESCs were intro672 Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc.

duced, diversifying the ESC toolbox with ESCs derived from articially generated blastocysts (Figure 2). Somatic cell nuclear transfer (SCNT) experiments, culminating in the cloning of Dolly the sheep, paved the way for cloning other mammals (Wilmut et al., 1997; Wakayama et al., 1998). In humans, there were various attempts to generate nuclear transfer (NT)-derived hESCs, but the initial successful report was found to be fraudulent (Normile et al., 2006), leading the eld to stagnation. Successful SCNT with human cells was nally reported last year, resulting in nuclear transfer (NT)-derived hESCs (Noggle et al., 2011). Although these ESCs were triploid, as the oocyte genome could not be removed, this study has revitalized interest in using SCNT for deriving personalized ESCs. Comparing these NT-ESCs to normal ESCs and to iPSCs would eventually be necessary to determine whether important differences exist between these pluripotent cell types. Another type of ESCs known for a long time in mouse was nally generated in humans as well: human parthenogenetic ESCs derived through parthenogenetic blastocysts (Kim et al., 2007b). Such blastocysts are generated by the activation of unfertilized oocytes, which undergo duplication of their genomic content and thus harbor two copies of the maternal genome. Human parthenogenetic pluripotent stem cells may serve for the generation of MHC-matched cells for transplantation, as was demonstrated in mouse (Kim et al., 2007a), and may also be used for the study of imprinting (Stelzer et al., 2011). A third striking type of ESCs was recently reported in mouse. Using the same technique of activating haploid oocytes, mESCs were derived from parthenogenetic embryos grown under specic culture conditions, and were than FACS-sorted for low DNA content, resulting in haploid mESC lines (Elling et al., 2011; Leeb and Wutz, 2011). These haploid mESCs may become an invaluable tool for forward and reverse genetics, as was elegantly demonstrated in these two groundbreaking papers. Similar attempts to generate haploid hESCs are currently underway, hopefully to be crowned with success. The Pluripurpose Cell: Using ESCs beyond Pluripotency Research One of the most interesting developments in the ESC eld in recent years is the way it has been integrated into afliated research elds and inuenced all areas of genetics, epigenetics, and cell biology. Indeed, the boundaries between ESC research and other research elds often become blurry, as studies with ESCs, rather than of ESCs, are already rapidly accumulating into a signicant body of work. ESCs have been used as a tool for the investigation of basic questions in various areas of biology, and by now these cells are responsible for major advancements outside the traditional borders of ESC research. To name just a few examples, the identication of an active DNA demethylation enzyme (Bhutani et al., 2010), lincRNAs (Guttman et al., 2009), and extensive transcription initiation (Guenther et al., 2007), as well as improvements in genomic techniques such as zinc nger nucleases (Collin and Lako, 2011) and high-resolution methylation mapping (Jeddeloh et al., 2008), are all discoveries made in ESCs but with implications that go far beyond the biology or the manipulation of pluripotent cells.

Cell Stem Cell

Perspective
It seems, therefore, that ESCs have recently made great impact on a variety of research arenas, serving as a sort of pluripurpose cells, a well characterized in vitro system of normal human proliferating cells that might possibly even replace HeLa cells as default cells of choice. ESCs have thus gone beyond what is usually perceived as ESC research and are extensively used in the biological and biomedical sciences (a fact that is unfortunately overlooked in some public discussions, when the necessity and utility of these cells is drawn into question). Concluding Remarks: The Next 5 Years Soon after Yamanaka and Thomson rst reported the generation of human iPSCs, President George W. Bush referred to this achievement as a scientic advancement within ethical boundaries (Kolata, 2007); however, the expectation that iPSCs would replace embryonic stem cells (ESCs) in the study of pluripotency has proven wrong. On the contrary, the vibrant and ourishing ESC research community received a substantial boost from the induced pluripotency breakthrough, and since 2006 these two pluripotent cell types have complemented and promoted each other (Scott et al., 2011). The rapid discovery rate in the ESC eld, with the surprising twists and turns it sometimes takes, makes it very difcult to predict where ESC research will be 5 years from now. Nonetheless, the achievements of recent years do seem to suggest that ESC research is far from reaching its full capacity and is predicted to continue expanding. Geographically speaking, the gap that still exists between mESCs and hESCs in terms of countries involved suggests that more countries are probably about to get actively engaged in hESC research. From a basic research point of view, there is no doubt that much is left to be discovered regarding the pluripotent state and its control, with an emphasis on combining the complex multilayered regulation into a coherent regulation circuitry. From the perspective of regenerative medicine, data regarding the safety and efcacy of current and future clinical trials will undoubtedly determine when ESC research is mature enough to fulll its promise in the clinic, and success in this arena is bound to draw more private companies into the eld. Novel ESC types and reprogramming methods, possible advancements in the generation of human-animal chimeras, and increasing numbers of disease models are further reasons to believe that exciting years lie before us.
ACKNOWLEDGMENTS The authors thank Dr. Daniel Ronen for critically reading this manuscript and Tamar Golan-Lev for her assistance with the graphical design of the gures. N.B. is the Herbert Cohn Chair in Cancer Research. This research was partially supported by the Legacy Heritage Biomedical Science Partnership Program of the Israel Science Foundation (grant 943/09), the Centers of Excellence Legacy Heritage Biomedical Science Partnership (grant 1801/10), and the Juvenile Diabetes Research Foundation (grant 1-2011-555). REFERENCES Adewumi, O., Aatoonian, B., Ahrlund-Richter, L., Amit, M., Andrews, P.W., Beighton, G., Bello, P.A., Benvenisty, N., Berry, L.S., Bevan, S., et al. (2007). Characterization of human embryonic stem cell lines by the International Stem Cell Initiative. Nat. Biotechnol. 25, 803816. Agarwal, S., Holton, K.L., and Lanza, R. (2008). Efcient differentiation of functional hepatocytes from human embryonic stem cells. Stem Cells 26, 11171127. Akopian, V., Andrews, P.W., Beil, S., Benvenisty, N., Brehm, J., Christie, M., Ford, A., Fox, V., Gokhale, P.J., Healy, L., et al. (2010). Comparison of dened culture systems for feeder cell free propagation of human embryonic stem cells. In Vitro Cell. Dev. Biol. Anim. 46, 247258. Amit, M., Chebath, J., Margulets, V., Laevsky, I., Miropolsky, Y., Shariki, K., Peri, M., Blais, I., Slutsky, G., Revel, M., et al. (2010). Suspension culture of undifferentiated human embryonic and induced pluripotent stem cells. Stem Cell Rev. 6, 248259. Amps, K., Andrews, P.W., Anyfantis, G., Armstrong, L., Avery, S., Baharvand, H., Baker, J., Baker, D., Munoz, M.B., Beil, S., et al. (2011). Screening ethnically diverse human embryonic stem cells identies a chromosome 20 minimal amplicon conferring growth advantage. Nat. Biotechnol. 29, 11321144. Ang, Y.S., Tsai, S.Y., Lee, D.F., Monk, J., Su, J., Ratnakumar, K., Ding, J., Ge, Y., Darr, H., Chang, B., et al. (2011). Wdr5 mediates self-renewal and reprogramming via the embryonic stem cell core transcriptional network. Cell 145, 183197. Baker, D.E., Harrison, N.J., Maltby, E., Smith, K., Moore, H.D., Shaw, P.J., Heath, P.R., Holden, H., and Andrews, P.W. (2007). Adaptation to culture of human embryonic stem cells and oncogenesis in vivo. Nat. Biotechnol. 25, 207215. Basma, H., Soto-Gutierrez, A., Yannam, G.R., Liu, L., Ito, R., Yamamoto, T., Ellis, E., Carson, S.D., Sato, S., Chen, Y., et al. (2009). Differentiation and transplantation of human embryonic stem cell-derived hepatocytes. Gastroenterology 136, 990999. Ben-David, U., and Benvenisty, N. (2011). The tumorigenicity of human embryonic and induced pluripotent stem cells. Nat. Rev. Cancer 11, 268277. Ben-David, U., Mayshar, Y., and Benvenisty, N. (2011). Large-scale analysis reveals acquisition of lineage-specic chromosomal aberrations in human adult stem cells. Cell Stem Cell 9, 97102. Ben-Yosef, D., Malcov, M., and Eiges, R. (2008). PGD-derived human embryonic stem cell lines as a powerful tool for the study of human genetic disorders. Mol. Cell. Endocrinol. 282, 153158. Bernstein, B.E., Mikkelsen, T.S., Xie, X., Kamal, M., Huebert, D.J., Cuff, J., Fry, B., Meissner, A., Wernig, M., Plath, K., et al. (2006). A bivalent chromatin structure marks key developmental genes in embryonic stem cells. Cell 125, 315326. Bhutani, N., Brady, J.J., Damian, M., Sacco, A., Corbel, S.Y., and Blau, H.M. (2010). Reprogramming towards pluripotency requires AID-dependent DNA demethylation. Nature 463, 10421047. Biancotti, J.C., Narwani, K., Buehler, N., Mandefro, B., Golan-Lev, T., Yanuka, O., Clark, A., Hill, D., Benvenisty, N., and Lavon, N. (2010). Human embryonic stem cells as models for aneuploid chromosomal syndromes. Stem Cells 28, 15301540. Blum, B., Bar-Nur, O., Golan-Lev, T., and Benvenisty, N. (2009). The antiapoptotic gene survivin contributes to teratoma formation by human embryonic stem cells. Nat. Biotechnol. 27, 281287. Bock, C., Kiskinis, E., Verstappen, G., Gu, H., Boulting, G., Smith, Z.D., Ziller, M., Croft, G.F., Amoroso, M.W., Oakley, D.H., et al. (2011). Reference Maps of human ES and iPS cell variation enable high-throughput characterization of pluripotent cell lines. Cell 144, 439452. Borowiak, M., Maehr, R., Chen, S., Chen, A.E., Tang, W., Fox, J.L., Schreiber, S.L., and Melton, D.A. (2009). Small molecules efciently direct endodermal differentiation of mouse and human embryonic stem cells. Cell Stem Cell 4, 348358. Boyer, L.A., Lee, T.I., Cole, M.F., Johnstone, S.E., Levine, S.S., Zucker, J.P., Guenther, M.G., Kumar, R.M., Murray, H.L., Jenner, R.G., et al. (2005). Core transcriptional regulatory circuitry in human embryonic stem cells. Cell 122, 947956. Brill, L.M., Xiong, W., Lee, K.B., Ficarro, S.B., Crain, A., Xu, Y., Terskikh, A., Snyder, E.Y., and Ding, S. (2009). Phosphoproteomic analysis of human embryonic stem cells. Cell Stem Cell 5, 204213. Brons, I.G., Smithers, L.E., Trotter, M.W., Rugg-Gunn, P., Sun, B., Chuva de Sousa Lopes, S.M., Howlett, S.K., Clarkson, A., Ahrlund-Richter, L., Pedersen, R.A., et al. (2007). Derivation of pluripotent epiblast stem cells from mammalian embryos. Nature 448, 191195.

Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc. 673

Cell Stem Cell

Perspective
Buecker, C., and Geijsen, N. (2010). Different avors of pluripotency, molecular mechanisms, and practical implications. Cell Stem Cell 7, 559564. Buecker, C., Chen, H.H., Polo, J.M., Daheron, L., Bu, L., Barakat, T.S., Okwieka, P., Porter, A., Gribnau, J., Hochedlinger, K., et al. (2010). A murine ESC-like state facilitates transgenesis and homologous recombination in human pluripotent stem cells. Cell Stem Cell 6, 535546. Canham, M.A., Sharov, A.A., Ko, M.S., and Brickman, J.M. (2010). Functional heterogeneity of embryonic stem cells revealed through translational amplication of an early endodermal transcript. PLoS Biol. 8, e1000379. Chen, X., Xu, H., Yuan, P., Fang, F., Huss, M., Vega, V.B., Wong, E., Orlov, Y.L., Zhang, W., Jiang, J., et al. (2008). Integration of external signaling pathways with the core transcriptional network in embryonic stem cells. Cell 133, 11061117. Chen, S., Borowiak, M., Fox, J.L., Maehr, R., Osafune, K., Davidow, L., Lam, K., Peng, L.F., Schreiber, S.L., Rubin, L.L., et al. (2009). A small molecule that directs differentiation of human ESCs into the pancreatic lineage. Nat. Chem. Biol. 5, 258265. Chen, G., Hou, Z., Gulbranson, D.R., and Thomson, J.A. (2010). Actin-myosin contractility is responsible for the reduced viability of dissociated human embryonic stem cells. Cell Stem Cell 7, 240248. Cheng, X., Ying, L., Lu, L., Galvao, A.M., Mills, J.A., Lin, H.C., Kotton, D.N., Shen, S.S., Nostro, M.C., Choi, J.K., et al. (2012). Self-renewing endodermal progenitor lines generated from human pluripotent stem cells. Cell Stem Cell 10, 371384. Cohen, D.E., and Melton, D. (2011). Turning straw into gold: directing cell fate for regenerative medicine. Nat. Rev. Genet. 12, 243252. Collin, J., and Lako, M. (2011). Concise review: putting a nger on stem cell biology: zinc nger nuclease-driven targeted genetic editing in human pluripotent stem cells. Stem Cells 29, 10211033. Colman, A., and Dreesen, O. (2009). Pluripotent stem cells and disease modeling. Cell Stem Cell 5, 244247. Damjanov, I., and Andrews, P.W. (2007). The terminology of teratocarcinomas and teratomas. Nat. Biotechnol. 25, 1212, discussion 1212. Eiges, R., Schuldiner, M., Drukker, M., Yanuka, O., Itskovitz-Eldor, J., and Benvenisty, N. (2001). Establishment of human embryonic stem celltransfected clones carrying a marker for undifferentiated cells. Curr. Biol. 11, 514518. Eiraku, M., Watanabe, K., Matsuo-Takasaki, M., Kawada, M., Yonemura, S., Matsumura, M., Wataya, T., Nishiyama, A., Muguruma, K., and Sasai, Y. (2008). Self-organized formation of polarized cortical tissues from ESCs and its active manipulation by extrinsic signals. Cell Stem Cell 3, 519532. Eiraku, M., Takata, N., Ishibashi, H., Kawada, M., Sakakura, E., Okuda, S., Sekiguchi, K., Adachi, T., and Sasai, Y. (2011). Self-organizing optic-cup morphogenesis in three-dimensional culture. Nature 472, 5156. Elling, U., Taubenschmid, J., Wirnsberger, G., OMalley, R., Demers, S.P., Vanhaelen, Q., Shukalyuk, A.I., Schmauss, G., Schramek, D., Schnuetgen, F., et al. (2011). Forward and reverse genetics through derivation of haploid mouse embryonic stem cells. Cell Stem Cell 9, 563574. Ellis, J., Baum, C., Benvenisty, N., Mostoslavsky, G., Okano, H., Stanford, W.L., Porteus, M., and Sadelain, M. (2010). Benets of utilizing gene-modied iPSCs for clinical applications. Cell Stem Cell 7, 429430. Fouse, S.D., Shen, Y., Pellegrini, M., Cole, S., Meissner, A., Van Neste, L., Jaenisch, R., and Fan, G. (2008). Promoter CpG methylation contributes to ES cell gene regulation in parallel with Oct4/Nanog, PcG complex, and histone H3 K4/K27 trimethylation. Cell Stem Cell 2, 160169. Gaspar-Maia, A., Alajem, A., Polesso, F., Sridharan, R., Mason, M.J., Heidersbach, A., Ramalho-Santos, J., McManus, M.T., Plath, K., Meshorer, E., et al. (2009). Chd1 regulates open chromatin and pluripotency of embryonic stem cells. Nature 460, 863868. Gaspar-Maia, A., Alajem, A., Meshorer, E., and Ramalho-Santos, M. (2011). Open chromatin in pluripotency and reprogramming. Nat. Rev. Mol. Cell Biol. 12, 3647. Giudice, A., and Trounson, A. (2008). Genetic modication of human embryonic stem cells for derivation of target cells. Cell Stem Cell 2, 422433. Goldring, C.E., Duffy, P.A., Benvenisty, N., Andrews, P.W., Ben-David, U., Eakins, R., French, N., Hanley, N.A., Kelly, L., Kitteringham, N.R., et al. (2011). Assessing the safety of stem cell therapeutics. Cell Stem Cell 8, 618628. Guenther, M.G., Levine, S.S., Boyer, L.A., Jaenisch, R., and Young, R.A. (2007). A chromatin landmark and transcription initiation at most promoters in human cells. Cell 130, 7788. Guttman, M., Amit, I., Garber, M., French, C., Lin, M.F., Feldser, D., Huarte, M., Zuk, O., Carey, B.W., Cassady, J.P., et al. (2009). Chromatin signature reveals over a thousand highly conserved large non-coding RNAs in mammals. Nature 458, 223227. Han, D.W., Tapia, N., Joo, J.Y., Greber, B., Arauzo-Bravo, M.J., Bernemann, C., Ko, K., Wu, G., Stehling, M., Do, J.T., et al. (2010). Epiblast stem cell subpopulations represent mouse embryos of distinct pregastrulation stages. Cell 143, 617627. Hanna, J., Markoulaki, S., Mitalipova, M., Cheng, A.W., Cassady, J.P., Staerk, J., Carey, B.W., Lengner, C.J., Foreman, R., Love, J., et al. (2009). Metastable pluripotent states in NOD-mouse-derived ESCs. Cell Stem Cell 4, 513524. Hanna, J., Cheng, A.W., Saha, K., Kim, J., Lengner, C.J., Soldner, F., Cassady, J.P., Muffat, J., Carey, B.W., and Jaenisch, R. (2010). Human embryonic stem cells with biological and epigenetic characteristics similar to those of mouse ESCs. Proc. Natl. Acad. Sci. USA 107, 92229227. Hayashi, K., Ohta, H., Kurimoto, K., Aramaki, S., and Saitou, M. (2011). Reconstitution of the mouse germ cell specication pathway in culture by pluripotent stem cells. Cell 146, 519532. Hockemeyer, D., Soldner, F., Beard, C., Gao, Q., Mitalipova, M., DeKelver, R.C., Katibah, G.E., Amora, R., Boydston, E.A., Zeitler, B., et al. (2009). Efcient targeting of expressed and silent genes in human ESCs and iPSCs using zinc-nger nucleases. Nat. Biotechnol. 27, 851857. Hockemeyer, D., Wang, H., Kiani, S., Lai, C.S., Gao, Q., Cassady, J.P., Cost, G.J., Zhang, L., Santiago, Y., Miller, J.C., et al. (2011). Genetic engineering of human pluripotent cells using TALE nucleases. Nat. Biotechnol. 29, 731734. Idelson, M., Alper, R., Obolensky, A., Ben-Shushan, E., Hemo, I., Yachimovich-Cohen, N., Khaner, H., Smith, Y., Wiser, O., Gropp, M., et al. (2009). Directed differentiation of human embryonic stem cells into functional retinal pigment epithelium cells. Cell Stem Cell 5, 396408. International Stem Cell Banking Initiative. (2009). Consensus guidance for banking and supply of human embryonic stem cell lines for research purposes. Stem Cell Rev. 5, 301314. Ito, S., DAlessio, A.C., Taranova, O.V., Hong, K., Sowers, L.C., and Zhang, Y. (2010). Role of Tet proteins in 5mC to 5hmC conversion, ES-cell self-renewal and inner cell mass specication. Nature 466, 11291133. Jeddeloh, J.A., Greally, J.M., and Rando, O.J. (2008). Reduced-representation methylation mapping. Genome Biol. 9, 231. Kadereit, S., and Trounson, A. (2011). In vitro immunogenicity of undifferentiated pluripotent stem cells (PSC) and derived lineages. Semin. Immunopathol. 33, 551562. Kim, K., Lerou, P., Yabuuchi, A., Lengerke, C., Ng, K., West, J., Kirby, A., Daly, M.J., and Daley, G.Q. (2007a). Histocompatible embryonic stem cells by parthenogenesis. Science 315, 482486. Kim, K., Ng, K., Rugg-Gunn, P.J., Shieh, J.H., Kirak, O., Jaenisch, R., Wakayama, T., Moore, M.A., Pedersen, R.A., and Daley, G.Q. (2007b). Recombination signatures distinguish embryonic stem cells derived by parthenogenesis and somatic cell nuclear transfer. Cell Stem Cell 1, 346352. Kim, J., Chu, J., Shen, X., Wang, J., and Orkin, S.H. (2008). An extended transcriptional network for pluripotency of embryonic stem cells. Cell 132, 10491061. Kim, J., Woo, A.J., Chu, J., Snow, J.W., Fujiwara, Y., Kim, C.G., Cantor, A.B., and Orkin, S.H. (2010). A Myc network accounts for similarities between embryonic stem and cancer cell transcription programs. Cell 143, 313324. Kobayashi, T., Yamaguchi, T., Hamanaka, S., Kato-Itoh, M., Yamazaki, Y., Ibata, M., Sato, H., Lee, Y.S., Usui, J., Knisely, A.S., et al. (2010). Generation of rat pancreas in mouse by interspecic blastocyst injection of pluripotent stem cells. Cell 142, 787799.

674 Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc.

Cell Stem Cell

Perspective
Koh, K.P., Yabuuchi, A., Rao, S., Huang, Y., Cunniff, K., Nardone, J., Laiho, A., Tahiliani, M., Sommer, C.A., Mostoslavsky, G., et al. (2011). Tet1 and Tet2 regulate 5-hydroxymethylcytosine production and cell lineage specication in mouse embryonic stem cells. Cell Stem Cell 8, 200213. Kolata, G. (2007). Scientists Bypass Need for Embryo to Get Stem Cells. The New York Times: http://www.nytimes.com/2007/11/21/science/21stem.html. Kriks, S., Shim, J.W., Piao, J., Ganat, Y.M., Wakeman, D.R., Xie, Z., CarrilloReid, L., Auyeung, G., Antonacci, C., Buch, A., et al. (2011). Dopamine neurons derived from human ES cells efciently engraft in animal models of Parkinsons disease. Nature 480, 547551. Kroon, E., Martinson, L.A., Kadoya, K., Bang, A.G., Kelly, O.G., Eliazer, S., Young, H., Richardson, M., Smart, N.G., Cunningham, J., et al. (2008). Pancreatic endoderm derived from human embryonic stem cells generates glucose-responsive insulin-secreting cells in vivo. Nat. Biotechnol. 26, 443 452. Kunarso, G., Chia, N.Y., Jeyakani, J., Hwang, C., Lu, X., Chan, Y.S., Ng, H.H., and Bourque, G. (2010). Transposable elements have rewired the core regulatory network of human embryonic stem cells. Nat. Genet. 42, 631634. Laustriat, D., Gide, J., and Peschanski, M. (2010). Human pluripotent stem cells in drug discovery and predictive toxicology. Biochem. Soc. Trans. 38, 10511057. Lavon, N., Yanuka, O., and Benvenisty, N. (2004). Differentiation and isolation of hepatic-like cells from human embryonic stem cells. Differentiation 72, 230238. Leavitt, A.D., and Hamlett, I. (2011). Homologous recombination in human embryonic stem cells: a tool for advancing cell therapy and understanding and treating human disease. Clin. Transl. Sci. 4, 298305. Leeb, M., and Wutz, A. (2011). Derivation of haploid embryonic stem cells from mouse embryos. Nature 479, 131134. Lefort, N., Feyeux, M., Bas, C., Feraud, O., Bennaceur-Griscelli, A., Tachdjian, G., Peschanski, M., and Perrier, A.L. (2008). Human embryonic stem cells reveal recurrent genomic instability at 20q11.21. Nat. Biotechnol. 26, 13641366. Lei, T., Jacob, S., Ajil-Zaraa, I., Dubuisson, J.B., Irion, O., Jaconi, M., and Feki, A. (2007). Xeno-free derivation and culture of human embryonic stem cells: current status, problems and challenges. Cell Res. 17, 682688. Lensch, M.W., and Ince, T.A. (2007). The terminology of teratocarcinomas and teratomas. Nat. Biotechnol. 25, 1211, author reply 12111212. Li, W., Wei, W., Zhu, S., Zhu, J., Shi, Y., Lin, T., Hao, E., Hayek, A., Deng, H., and Ding, S. (2009). Generation of rat and human induced pluripotent stem cells by combining genetic reprogramming and chemical inhibitors. Cell Stem Cell 4, 1619. Lister, R., Pelizzola, M., Dowen, R.H., Hawkins, R.D., Hon, G., Tonti-Filippini, J., Nery, J.R., Lee, L., Ye, Z., Ngo, Q.M., et al. (2009). Human DNA methylomes at base resolution show widespread epigenomic differences. Nature 462, 315322. Loh, Y.H., Wu, Q., Chew, J.L., Vega, V.B., Zhang, W., Chen, X., Bourque, G., George, J., Leong, B., Liu, J., et al. (2006). The Oct4 and Nanog transcription network regulates pluripotency in mouse embryonic stem cells. Nat. Genet. 38, 431440. Loh, Y.H., Yang, L., Yang, J.C., Li, H., Collins, J.J., and Daley, G.Q. (2011). Genomic approaches to deconstruct pluripotency. Annu. Rev. Genomics Hum. Genet. 12, 165185. Lu, R., Markowetz, F., Unwin, R.D., Leek, J.T., Airoldi, E.M., MacArthur, B.D., Lachmann, A., Rozov, R., Maayan, A., Boyer, L.A., et al. (2009). Systems-level dynamic analyses of fate change in murine embryonic stem cells. Nature 462, 358362. Mandenius, C.F., Steel, D., Noor, F., Meyer, T., Heinzle, E., Asp, J., Arain, S., Kraushaar, U., Bremer, S., Class, R., et al. (2011). Cardiotoxicity testing using pluripotent stem cell-derived human cardiomyocytes and state-of-the-art bioanalytics: a review. J. Appl. Toxicol. 31, 191205. Margaritis, T., and Holstege, F.C. (2008). Poised RNA polymerase II gives pause for thought. Cell 133, 581584. Margueron, R., and Reinberg, D. (2011). The Polycomb complex PRC2 and its mark in life. Nature 469, 343349. Marson, A., Levine, S.S., Cole, M.F., Frampton, G.M., Brambrink, T., Johnstone, S., Guenther, M.G., Johnston, W.K., Wernig, M., Newman, J., et al. (2008). Connecting microRNA genes to the core transcriptional regulatory circuitry of embryonic stem cells. Cell 134, 521533. Martinez Arias, A., and Brickman, J.M. (2011). Gene expression heterogeneities in embryonic stem cell populations: origin and function. Curr. Opin. Cell Biol. 23, 650656. Martinez, N.J., and Gregory, R.I. (2010). MicroRNA gene regulatory pathways in the establishment and maintenance of ESC identity. Cell Stem Cell 7, 3135. Martins-Taylor, K., and Xu, R.H. (2010). Determinants of pluripotency: from avian, rodents, to primates. J. Cell. Biochem. 109, 1625. Mayshar, Y., Ben-David, U., Lavon, N., Biancotti, J.C., Yakir, B., Clark, A.T., Plath, K., Lowry, W.E., and Benvenisty, N. (2010). Identication and classication of chromosomal aberrations in human induced pluripotent stem cells. Cell Stem Cell 7, 521531. Mayshar, Y., Yanuka, O., and Benvenisty, N. (2011). Teratogen screening using transcriptome proling of differentiating human embryonic stem cells. J. Cell. Mol. Med. 15, 13931401. Meissner, A. (2010). Epigenetic modications in pluripotent and differentiated cells. Nat. Biotechnol. 28, 10791088. Meissner, A., Mikkelsen, T.S., Gu, H., Wernig, M., Hanna, J., Sivachenko, A., Zhang, X., Bernstein, B.E., Nusbaum, C., Jaffe, D.B., et al. (2008). Genomescale DNA methylation maps of pluripotent and differentiated cells. Nature 454, 766770. Murry, C.E., and Keller, G. (2008). Differentiation of embryonic stem cells to clinically relevant populations: lessons from embryonic development. Cell 132, 661680. Ng, H.H., and Surani, M.A. (2011). The transcriptional and signalling networks of pluripotency. Nat. Cell Biol. 13, 490496. Nichols, J., and Smith, A. (2009). Naive and primed pluripotent states. Cell Stem Cell 4, 487492. Noggle, S., Fung, H.L., Gore, A., Martinez, H., Satriani, K.C., Prosser, R., Oum, K., Paull, D., Druckenmiller, S., Freeby, M., et al. (2011). Human oocytes reprogram somatic cells to a pluripotent state. Nature 478, 7075. Normile, D., Vogel, G., and Couzin, J. (2006). Cloning. South Korean teams remaining human stem cell claim demolished. Science 311, 156157. Ohgushi, M., Matsumura, M., Eiraku, M., Murakami, K., Aramaki, T., Nishiyama, A., Muguruma, K., Nakano, T., Suga, H., Ueno, M., et al. (2010). Molecular pathway and cell state responsible for dissociation-induced apoptosis in human pluripotent stem cells. Cell Stem Cell 7, 225239. Orkin, S.H., and Hochedlinger, K. (2011). Chromatin connections to pluripotency and cellular reprogramming. Cell 145, 835850. Oshima, K., Shin, K., Diensthuber, M., Peng, A.W., Ricci, A.J., and Heller, S. (2010). Mechanosensitive hair cell-like cells from embryonic and induced pluripotent stem cells. Cell 141, 704716. Pardo, M., Lang, B., Yu, L., Prosser, H., Bradley, A., Babu, M.M., and Choudhary, J. (2010). An expanded Oct4 interaction network: implications for stem cell biology, development, and disease. Cell Stem Cell 6, 382395. Pasini, D., Cloos, P.A., Walfridsson, J., Olsson, L., Bukowski, J.P., Johansen, J.V., Bak, M., Tommerup, N., Rappsilber, J., and Helin, K. (2010). JARID2 regulates binding of the Polycomb repressive complex 2 to target genes in ES cells. Nature 464, 306310. Pastor, W.A., Pape, U.J., Huang, Y., Henderson, H.R., Lister, R., Ko, M., McLoughlin, E.M., Brudno, Y., Mahapatra, S., Kapranov, P., et al. (2011). Genome-wide mapping of 5-hydroxymethylcytosine in embryonic stem cells. Nature 473, 394397. Peng, J.C., Valouev, A., Swigut, T., Zhang, J., Zhao, Y., Sidow, A., and Wysocka, J. (2009). Jarid2/Jumonji coordinates control of PRC2 enzymatic activity and target gene occupancy in pluripotent cells. Cell 139, 12901302.

Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc. 675

Cell Stem Cell

Perspective
Robertson, N.J., Brook, F.A., Gardner, R.L., Cobbold, S.P., Waldmann, H., and Fairchild, P.J. (2007). Embryonic stem cell-derived tissues are immunogenic but their inherent immune privilege promotes the induction of tolerance. Proc. Natl. Acad. Sci. USA 104, 2092020925. Robinton, D.A., and Daley, G.Q. (2012). The promise of induced pluripotent stem cells in research and therapy. Nature 481, 295305. Rubin, L.L. (2008). Stem cells and drug discovery: the beginning of a new era? Cell 132, 549552. Schwartz, S.D., Hubschman, J.P., Heilwell, G., Franco-Cardenas, V., Pan, C.K., Ostrick, R.M., Mickunas, E., Gay, R., Klimanskaya, I., and Lanza, R. (2012). Embryonic stem cell trials for macular degeneration: a preliminary report. Lancet 379, 713720. Scott, C.T., McCormick, J.B., DeRouen, M.C., and Owen-Smith, J. (2011). Democracy derived? New trajectories in pluripotent stem cell research. Cell 145, 820826. Sharon, N., Mor, I., Golan-lev, T., Fainsod, A., and Benvenisty, N. (2011). Molecular and functional characterizations of gastrula organizer cells derived from human embryonic stem cells. Stem Cells 29, 600608. Sheik Mohamed, J., Gaughwin, P.M., Lim, B., Robson, P., and Lipovich, L. (2010). Conserved long noncoding RNAs transcriptionally regulated by Oct4 and Nanog modulate pluripotency in mouse embryonic stem cells. RNA 16, 324337. Singh Roy, N., Nakano, T., Xuing, L., Kang, J., Nedergaard, M., and Goldman, S.A. (2005). Enhancer-specied GFP-based FACS purication of human spinal motor neurons from embryonic stem cells. Exp. Neurol. 196, 224234. Song, H., Chung, S.K., and Xu, Y. (2010). Modeling disease in human ESCs using an efcient BAC-based homologous recombination system. Cell Stem Cell 6, 8089. Spence, J.R., Mayhew, C.N., Rankin, S.A., Kuhar, M.F., Vallance, J.E., Tolle, K., Hoskins, E.E., Kalinichenko, V.V., Wells, S.I., Zorn, A.M., et al. (2011). Directed differentiation of human pluripotent stem cells into intestinal tissue in vitro. Nature 470, 105109. Steiner, D., Khaner, H., Cohen, M., Even-Ram, S., Gil, Y., Itsykson, P., Turetsky, T., Idelson, M., Aizenman, E., Ram, R., et al. (2010). Derivation, propagation and controlled differentiation of human embryonic stem cells in suspension. Nat. Biotechnol. 28, 361364. Stelzer, Y., Yanuka, O., and Benvenisty, N. (2011). Global analysis of parental imprinting in human parthenogenetic induced pluripotent stem cells. Nat. Struct. Mol. Biol. 18, 735741. Stewart, M.H., Bendall, S.C., Levadoux-Martin, M., and Bhatia, M. (2010). Clonal tracking of hESCs reveals differential contribution to functional assays. Nat. Methods 7, 917922. Suga, H., Kadoshima, T., Minaguchi, M., Ohgushi, M., Soen, M., Nakano, T., Takata, N., Wataya, T., Muguruma, K., Miyoshi, H., et al. (2011). Self-formation of functional adenohypophysis in three-dimensional culture. Nature 480, 5762. Tachibana, M., Sparman, M., Ramsey, C., Ma, H., Lee, H.S., Penedo, M.C., and Mitalipov, S. (2012). Generation of chimeric rhesus monkeys. Cell 148, 285295. Tahiliani, M., Koh, K.P., Shen, Y., Pastor, W.A., Bandukwala, H., Brudno, Y., Agarwal, S., Iyer, L.M., Liu, D.R., Aravind, L., et al. (2009). Conversion of 5-methylcytosine to 5-hydroxymethylcytosine in mammalian DNA by MLL partner TET1. Science 324, 930935. Takahashi, T., Lord, B., Schulze, P.C., Fryer, R.M., Sarang, S.S., Gullans, S.R., and Lee, R.T. (2003). Ascorbic acid enhances differentiation of embryonic stem cells into cardiac myocytes. Circulation 107, 19121916. Tang, C., Lee, A.S., Volkmer, J.P., Sahoo, D., Nag, D., Mosley, A.R., Inlay, M.A., Ardehali, R., Chavez, S.L., Pera, R.R., et al. (2011). An antibody against SSEA-5 glycan on human pluripotent stem cells enables removal of teratomaforming cells. Nat. Biotechnol. 29, 829834. Tesar, P.J., Chenoweth, J.G., Brook, F.A., Davies, T.J., Evans, E.P., Mack, D.L., Gardner, R.L., and McKay, R.D. (2007). New cell lines from mouse epiblast share dening features with human embryonic stem cells. Nature 448, 196199. Thomson, J.A., Itskovitz-Eldor, J., Shapiro, S.S., Waknitz, M.A., Swiergiel, J.J., Marshall, V.S., and Jones, J.M. (1998). Embryonic stem cell lines derived from human blastocysts. Science 282, 11451147. Thomson, M., Liu, S.J., Zou, L.N., Smith, Z., Meissner, A., and Ramanathan, S. (2011). Pluripotency factors in embryonic stem cells regulate differentiation into germ layers. Cell 145, 875889. Toyooka, Y., Shimosato, D., Murakami, K., Takahashi, K., and Niwa, H. (2008). Identication and characterization of subpopulations in undifferentiated ES cell culture. Development 135, 909918. Trounson, A., Thakar, R.G., Lomax, G., and Gibbons, D. (2011). Clinical trials for stem cell therapies. BMC Med. 9, 52. Tulpule, A., Lensch, M.W., Miller, J.D., Austin, K., DAndrea, A., Schlaeger, T.M., Shimamura, A., and Daley, G.Q. (2010). Knockdown of Fanconi anemia genes in human embryonic stem cells reveals early developmental defects in the hematopoietic lineage. Blood 115, 34533462. Unger, C., Skottman, H., Blomberg, P., Dilber, M.S., and Hovatta, O. (2008). Good manufacturing practice and clinical-grade human embryonic stem cell lines. Hum. Mol. Genet. 17, R48R53. Urbach, A., Bar-Nur, O., Daley, G.Q., and Benvenisty, N. (2010). Differential modeling of fragile X syndrome by human embryonic stem cells and induced pluripotent stem cells. Cell Stem Cell 6, 407411. Urbach, A., Schuldiner, M., and Benvenisty, N. (2004). Modeling for LeschNyhan disease by gene targeting in human embryonic stem cells. Stem Cells 22, 635641. Valamehr, B., Tsutsui, H., Ho, C.M., and Wu, H. (2011). Developing dened culture systems for human pluripotent stem cells. Regen. Med. 6, 623634. Van Hoof, D., Munoz, J., Braam, S.R., Pinkse, M.W., Linding, R., Heck, A.J., Mummery, C.L., and Krijgsveld, J. (2009). Phosphorylation dynamics during early differentiation of human embryonic stem cells. Cell Stem Cell 5, 214226. , E., Brenner, C., Deplus, R., Blanchon, L., Fraga, M., Didelot, C., Morey, L., Vire Van Eynde, A., Bernard, D., Vanderwinden, J.M., et al. (2006). The Polycomb group protein EZH2 directly controls DNA methylation. Nature 439, 871874. Viswanathan, S.R., Daley, G.Q., and Gregory, R.I. (2008). Selective blockade of microRNA processing by Lin28. Science 320, 97100. Wakayama, T., Perry, A.C., Zuccotti, M., Johnson, K.R., and Yanagimachi, R. (1998). Full-term development of mice from enucleated oocytes injected with cumulus cell nuclei. Nature 394, 369374. Wang, J., Rao, S., Chu, J., Shen, X., Levasseur, D.N., Theunissen, T.W., and Orkin, S.H. (2006). A protein interaction network for pluripotency of embryonic stem cells. Nature 444, 364368. Wang, W., Yang, J., Liu, H., Lu, D., Chen, X., Zenonos, Z., Campos, L.S., Rad, R., Guo, G., Zhang, S., et al. (2011a). Rapid and efcient reprogramming of somatic cells to induced pluripotent stem cells by retinoic acid receptor gamma and liver receptor homolog 1. Proc. Natl. Acad. Sci. USA 108, 1828318288. Wang, Y.C., Nakagawa, M., Garitaonandia, I., Slavin, I., Altun, G., Lacharite, R.M., Nazor, K.L., Tran, H.T., Lynch, C.L., Leonardo, T.R., et al. (2011b). Specic lectin biomarkers for isolation of human pluripotent stem cells identied through array-based glycomic analysis. Cell Res. 21, 15511563. Wang, Z., Oron, E., Nelson, B., Razis, S., and Ivanova, N. (2012). Distinct Lineage Specication Roles for NANOG, OCT4, and SOX2 in Human Embryonic Stem Cells. Cell Stem Cell 10, 440454. Watanabe, K., Ueno, M., Kamiya, D., Nishiyama, A., Matsumura, M., Wataya, T., Takahashi, J.B., Nishikawa, S., Muguruma, K., and Sasai, Y. (2007). A ROCK inhibitor permits survival of dissociated human embryonic stem cells. Nat. Biotechnol. 25, 681686. Wen, B., Wu, H., Shinkai, Y., Irizarry, R.A., and Feinberg, A.P. (2009). Large histone H3 lysine 9 dimethylated chromatin blocks distinguish differentiated from embryonic stem cells. Nat. Genet. 41, 246250. West, P.R., Weir, A.M., Smith, A.M., Donley, E.L., and Cezar, G.G. (2010). Predicting human developmental toxicity of pharmaceuticals using human embryonic stem cells and metabolomics. Toxicol. Appl. Pharmacol. 247, 1827.

676 Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc.

Cell Stem Cell

Perspective
Wilmut, I., Schnieke, A.E., McWhir, J., Kind, A.J., and Campbell, K.H. (1997). Viable offspring derived from fetal and adult mammalian cells. Nature 385, 810813. Wobus, A.M., and Loser, P. (2011). Present state and future perspectives of using pluripotent stem cells in toxicology research. Arch. Toxicol. 85, 79117. Young, R.A. (2011). Control of the embryonic stem cell state. Cell 144, 940954. Ziller, M.J., Muller, F., Liao, J., Zhang, Y., Gu, H., Bock, C., Boyle, P., Epstein, C.B., Bernstein, B.E., Lengauer, T., et al. (2011). Genomic distribution and inter-sample variation of non-CpG methylation across human cell types. PLoS Genet. 7, e1002389.

Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc. 677

Cell Stem Cell

Perspective
Cancer Stem Cells: Current Status and Evolving Complexities
Jane E. Visvader1,2,* and Geoffrey J. Lindeman1,2,3
Laboratory, The Walter and Eliza Hall Institute of Medical Research, Melbourne, VIC 3052, Australia of Melbourne, Parkville, VIC 3050, Australia 3Department of Medical Oncology, Royal Melbourne Hospital, Parkville, VIC 3050, Australia *Correspondence: visvader@wehi.edu.au DOI 10.1016/j.stem.2012.05.007
2University 1VBCRC

The cancer stem cell (CSC) model has been established as a cellular mechanism that contributes to phenotypic and functional heterogeneity in diverse cancer types. Recent observations, however, have highlighted many complexities and challenges: the CSC phenotype can vary substantially between patients, tumors may harbor multiple phenotypically or genetically distinct CSCs, metastatic CSCs can evolve from primary CSCs, and tumor cells may undergo reversible phenotypic changes. Although the CSC concept will have clinical relevance in specic cases, accumulating evidence suggests that it will be imperative to target all CSC subsets within the tumor to prevent relapse.
Introduction Phenotypic and functional heterogeneity is a dening feature of many leukemias and solid tumors. Several factors contribute to this heterogeneity, including genetic mutations, epigenetic changes, interactions with the microenvironment, and the presence or absence of a cellular hierarchy. Different cellular mechanisms have been postulated to account for intratumoral heterogeneity. The acquisition of genetic (or epigenetic) alterations underpins the clonal evolution theory (Nowell, 1976) in which cells in the dominant clonal population(s) possess similar tumorigenic potential. Conversely, the cancer stem cell (CSC) model postulates a hierarchical organization of cells such that only a small subset is responsible for sustaining tumorigenesis and establishing the cellular heterogeneity inherent in the primary tumor. Although CSCs exhibit the stem cell properties of self-renewal and differentiation, they do not necessarily originate from the transformation of normal tissue stem cells (Figure 1). This model has received wide attention because it provides an explanation for resistance to both radiation and chemotherapy and eventual tumor relapse. In addition, quiescent or slow cycling CSCs may survive therapeutic intervention and result in recurrence. The rst prospective identication of a CSC was made by Dick and colleagues for AML (Bonnet and Dick, 1997; Lapidot et al., 1994). CSCs were subsequently demonstrated to occur in diverse solid tumors (reviewed in Visvader and Lindeman, 2008). Although the existence of CSCs has been well established for specic cancers, it is clear that the CSC model does not account for functional heterogeneity in all tumors. Over the last few years, emphasis in the CSC eld has shifted more toward the use of freshly isolated tumor specimens and early-passage xenografts for transplantation studies rather than the use of cultured tumor cells. Furthermore, there is increased awareness that the nature of the xenotransplantation assay is critical for evaluating the existence of CSCs (Quintana et al., 2008). Here we review recent developments in this rapidly moving eld, including the variable phenotype of CSCs, the presence of multiple CSC pools within individual tumors, the ability of CSCs to undergo genetic evolution, and the potential of nonCSCs to switch to CSC-like cells (Figure 2). These observations highlight the dynamic nature of CSCs and further indicate that the clonal evolution and CSC models can act in concert. They also somewhat dampen the original therapeutic promise of the CSC model, as it seems that all CSC subsets within the tumor will need to be dened and targeted in order to inuence clinical outcome. In the case of solid tumors, which exhibit extraordinary genomic instability, it will probably be necessary to target both CSCs and non-CSCs to achieve durable remission. Despite these complexities, the recent derivation of a stem cell-like or self-renewal gene expression signature that is predictive of patient outcome in human leukemia lends credence to the CSC hypothesis and its clinical relevance (Eppert et al., 2011; Gentles et al., 2010). CSC Markers Are Not Universal for Any Cancer Type CSCs must be dened functionally by well-validated assays such as in vivo transplantation rather than on the basis of immunophenotype alone. Nonetheless, a number of markers have proven useful for the isolation of subsets enriched for CSCs in multiple types of solid tumors, including CD133, CD44, EpCAM, and ALDH activity. In the case of human leukemia, a combination of CD34, CD38, and IL3Ra has enabled the prospective isolation of leukemia stem cells. It should be noted that none of these markers are exclusively expressed by CSCs. With the passage of time, it has become increasingly evident that the CSC phenotype varies between individual patient tumors of the same subtype, raising the question of whether the markedly different clinical outcomes reect differences in their CSC populations. CD133 (prominin) is one example of a cell surface protein that has been widely explored as a CSC marker. CD133 was initially described as a CSC marker for glioblastoma multiforme (Singh et al., 2004). Moreover, direct imaging of matched CSCs and non-CSCs in the same in vivo microenvironment of primary glioblastoma tumors demonstrated that only the CD133+ subset had the ability to maintain tumorigenesis and generate heterogeneity (Lathia et al., 2011a). However, CD133 does not always mark
Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc. 717

Cell Stem Cell

Perspective
Figure 1. Schemata of the Clonal Evolution and Cancer Stem Cell Models
(A) The clonal evolution model is a nonhierarchical model where mutations arising in tumor cells confer a selective growth advantage. Depicted here is a cell (red) that has acquired a series of mutations and produced a dominant clone. Tumor cells (red and orange) arising from this clone have similar tumorigenic capacity. Other derivatives (grey) may lack tumorigenicity due to stochastic events. Tumor heterogeneity results from the diversity of cells present within the tumor. (B) The cancer stem cell model is predicated on a hierarchical organization of cells, where a small subset of cells has the ability to sustain tumorigenesis and generate heterogeneity through differentiation. In the example shown, a mutation(s) in a progenitor cell (depicted as the brown cell) has endowed the tumor cell with stem cell-like properties. These cells have self-renewing capability and give rise to a range of tumor cells (depicted as gray and green cells), thereby accounting for tumor heterogeneity.

CSCs and appears to be modulated by extrinsic factors. The search for more robust markers of CSCs in glioblastoma and other brain tumors has revealed SSEA-1/CD15/Lewis X and a6-integrin. SSEA-1 (stage-specic embryonic antigen) was identied as a CSC marker in both human glioblastoma and syngeneic mouse models of medulloblastoma (Read et al., 2009; Son et al., 2009; Ward et al., 2009). Despite a high proportion of specimens lacking CD133+ cells, SSEA-1 enriched for CSCs by 100-fold in almost every human glioblastoma tumor evaluated (Son et al., 2009). In another approach, Rich and colleagues examined the perivascular microenvironment in which brain CSCs reside and identied a6-integrin as a CSC marker that was required for maintenance of CSCs in vivo (Lathia et al., 2010). Coexpression of CD133 and a6-integrin was observed in some but not all tumors. The limited overlap evident between the phenotypes of CSCs isolated from the same tumor type may reect the presence of multiple CSC pools or technical variation arising from differing enzymatic digestion conditions, the use of cultured versus freshly sorted cells, or extensively passaged versus early xenograft tumors. Another confounding factor is that stringent assays to prove self-renewing activity have not always been applied. The genetic mutation prole may also inuence the nature and phenotype of CSCs, as suggested by studies on different genetic mouse models of lung adenocarcinoma (Curtis et al., 2010), whereas epigenetic changes in regulatory genes could impact marker expression itself. In breast cancer, although CD44 and CD24 have been extensively used to isolate CSCs, they should not be viewed as universal markers. CD44 and CD24 did not selectively enrich for CSCs in ER-negative and triplenegative breast tumors as shown by the fact that CSCs were
718 Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc.

found in both the CD44+CD24 and CD44+CD24+ fractions (Meyer et al., 2010). Furthermore, the ALDHhi and CD44hiCD24lo CSC-enriched subsets in breast cancer bear little overlap within the same tumor (Ginestier et al., 2007). A similar story holds true for colorectal cancer in which the EpCAMhiCD44+ CSC subpopulation shared minor overlap with CD133 (Dalerba et al., 2007), and for pancreatic cancer, where overlap between the CD133+ and CD44+CD24+ populations varied considerably between specimens (Hermann et al., 2007). In ovarian cancers, strikingly little concordance was found between CD133 and reported ovarian CSC markers including CD117, CD44, and ALDH1 activity (Curley et al., 2009; Stewart et al., 2011), most probably explained by many groups relying on cultured cells as opposed to freshly sorted tumors. Finally, in patients with non-small cell lung cancer, although CD133, CD44, and EpCAM proved ineffective for the isolation of CSCs, CD166 emerged as a robust marker in more than 50% of cases (Zhang et al., 2012). Nevertheless, a combination of markers can rene the CSC phenotype. For example, CD44 expression combined with high levels of the tyrosine kinase receptor c-MET provided robust selection of pancreatic CSCs (Li et al., 2011), and high ALDH activity together with CD133 expression resulted in signicant enrichment such that 1 in 11 ovarian tumor cells exhibited CSC properties (Silva et al., 2011). Highly Variable Frequency of CSCs between Tumors The true frequency of CSCs in most human tumors has probably been underestimated because of barriers imposed by xenotransplantation, species-specic differences in growth factors/ receptors, and the level of immune recognition. However, CSCs and tumor-initiating cells in many solid tumors tend to be relatively infrequent, even when measured under more permissive conditions (Ishizawa et al., 2010; Stewart et al., 2011). The term tumor-initiating is generally used by the eld as an operational term to dene cells that initiate tumors upon transplantation but it is not necessarily synonymous with a CSC.

Cell Stem Cell

Perspective
estingly, initiating driver mutations were found to affect the frequencies of LSCs in mouse models of AML (Heuser et al., 2009; Somervaille et al., 2009). In high-grade tumors from patients, an increased pool of CSCs may underlie their aggressive behavior (Boiko et al., 2010; Pece et al., 2010; Zhang et al., 2012). Indeed, the CSC model of heterogeneity may apply more readily to early-stage than to advanced tumors, in which dominant clones probably drive tumor progression. Instability of the CSC Phenotype Substantial differences in the immunophenotype of tumorpropagating cells between primary cancer specimens and their corresponding xenografts have been reported. Whereas most primary serous ovarian cancers contained CD133+ CSCs, the majority of xenografted tumors contained signicant numbers of CD133 tumor-initiating cells that could not be attributed to contamination (Stewart et al., 2011). The marked changes in copy number variation between these primary and xenografted tumors suggest that genetic change is driving tumor progression, although the pre-existence of subclonal diversity cannot be excluded. These data not only reect heterogeneity within the tumor-propagating compartment but also indicate that the CSC phenotype may not be stable upon xenograft passaging. Although the frequency of CSCs for some ovarian (Stewart et al., 2011) and breast cancer (Meyer et al., 2010) xenografts remained constant, the frequency of CSCs has been observed to increase during serial transplantation, thus emphasizing the need to study early-passage tumors (Boiko et al., 2010; Ishizawa et al., 2010). Existence of Multiple CSC Pools within Individual Tumors Cancers can harbor heterogeneous and biologically distinct populations of CSCs. Recent studies have identied molecularly distinct leukemic stem cell populations dened by CD34, CD38, and/or IL3Ra expression. In the majority of AML patients, two hierarchically organized LSCs were shown to coexist. These populations are more closely related to normal progenitor subtypes than hematopoietic stem cells (HSCs) (Goardon et al., 2011), implying that the progenitors have aberrantly acquired stem cell properties. Complementary ndings were made with a large number of AML patient samples, in which both progenitor and more primitive HSC-like fractions contained LSCs and generated the same phenotypic diversity found in the primary samples (Eppert et al., 2011; Sarry et al., 2011). Notably, another study revealed that CD38+ AML cells may have previously escaped detection because of their unexpected clearance by the antibody (Taussig et al., 2008). In a HoxA9-Meis1-driven mouse model of AML, multiple phenotypically distinct LSCs were identied, and each was capable of recapitulating the original disease histopathology (Gibbs et al., 2012). Collectively, these ndings demonstrate heterogeneity within the LSC compartment of individual patient specimens and also indicate that AML often appears as a progenitor disease. In at least some of these AMLs, a hierarchical relationship appears to exist among the different LSC subsets. The observation that different pools can clonally recapitulate the immunophenotype of the primary specimen suggests that LSCs may dedifferentiate or exhibit phenotypic interconversion.
Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc. 719

Figure 2. Schematic Models of Tumor Propagation by CSCs Depicting Variations that Can Contribute to Tumor Heterogeneity
(A) One CSC subset may be present within the tumor. As described in Figure 1, non-CSCs are incapable of generating a tumor. (B) Multiple distinct CSC pools, each independently capable of tumor propagation, may exist with an individual tumor. (C) Long-lived dormant CSCs may produce local and/or distant tumor recurrence after activation (depicted here as a yellow CSC with a red rim) many years after anticancer therapy. (D) As tumor progression occurs, a second distinct CSC may arise as a result of clonal evolution. This may result from the acquisition of an additional mutation or by epigenetic modication. The more aggressive CSC will become dominant and drive tumor formation. (E) The CSC phenotype may be unstable, resulting in phenotypic reversion of cell surface markers and switching of the CSC phenotype. This may occur in response to cell-intrinsic or microenvironmental cues.

Tumors that do not follow a CSC model also contain tumorinitiating cells but these do not exhibit stem cell-like properties. In head and neck, pancreatic, and non-small cell lung cancers, the frequency of tumor-initiating cells varied dramatically but always comprised a very small population (<0.02%). Interestingly, the frequency was not dependent on the immune status of the recipient (Ishizawa et al., 2010). CSCs in serous ovarian cancers were also found to be infrequent (<0.04%) and again varied substantially among patients (Stewart et al., 2011). Analogous to these observations, evaluation of leukemic stem cells (LSCs) under improved xenotransplantation conditions revealed highly variable LSC frequencies in the range of 1 in 103106 cells (Eppert et al., 2011). CSCs may not necessarily constitute a minor component of the tumor. A relatively high proportion of leukemia-propagating cells has been observed in specic syngeneic mouse models of lymphomas and leukemias, whereas lower frequencies of CSCs generally occur in murine models of epithelial and other solid tumors (reviewed in Visvader and Lindeman, 2008). Inter-

Cell Stem Cell

Perspective
Heterogeneous CSC compartments have recently been unmasked in solid tumors. In ovarian, breast, and squamous cell carcinomas, distinct CSC populations that regenerate the phenotypic and functional heterogeneity of the parental tumor have been described (Meyer et al., 2010; Schober and Fuchs, 2011; Stewart et al., 2011). In the case of primary colorectal cancers, three different types of tumor-initiating cells were resolved on the basis of clonal sphere cultures from individual patient tumors: a rare subset of CSCs that maintained tumor growth on serial transplantation, a tumor-initiating cell with limited self-renewal capacity (therefore not dened as a CSC), and a more latent CSC that apparently was activated in secondary or tertiary transplantation assays (Dieter et al., 2011). Because spheres generated from single cells comprised three cell types dened by differences in self-renewal, epigenetic rather than genetic mechanisms may account for the functional differences. Clonal heterogeneity among tumor-initiating cells was also observed in PTEN-decient glioblastoma, in which a series of phenotypically distinct self-renewing cells was observed in both the CD133+ and CD133 fractions (Chen et al., 2010). These cells were arranged in a linear hierarchy and generated tumors with different growth kinetics in serial transplantation experiments. However, both of these studies relied on sphere cultures of cells maintained under specic conditions, and therefore need to be validated with fresh primary tumor samples. Although tumor-propagating ability can reect sphereforming capacity, it is important to note that they do not always equate because the selection of specic cells may occur in vitro (Read et al., 2009). Adding a further layer of complexity, distinct CSC subsets within a tumor have the potential to interconvert. In skin squamous cell carcinomas, two CSC subsets located along the tumor-stroma interface displayed different tumor growth kinetics and could interchange phenotype (Schober and Fuchs, 2011). These may not represent distinct CSC pools but rather stochastic variation within a single CSC population in response to microenvironmental signals. Phenotypic conversion also occurs among nonhierarchically organized tumor cells in melanoma (Quintana et al., 2010). Metastatic CSCs May Be the Same or Distinct from the Primary CSC There is growing evidence for the existence of functionally distinct subsets of tumor cells that impart metastatic activity. CSC subsets within primary tumors may harbor CSC subsets with tumor-propagating and/or metastatic capacity (Figure 3). In breast cancer, noninvasive imaging indicated that primary tumor CSCs characterized by CD44 expression are directly involved in metastasis (Liu et al., 2010). Similarly, in colorectal cancer, metastasis was almost exclusively a property of the CSCs that exhibited long-term self-renewing capacity (Dieter et al., 2011). Multiple disseminating CSCs homed to the bone marrow and generated liver metastases but only single clones were detected in peripheral blood, suggesting that metastatic CSCs enter the circulation transiently. In a related study on colorectal cancer, a subset of CD26+ cells resident within primary and metastatic tumors demonstrated tumor propagation, chemoresistance, and liver metastatic potential after implantation at the orthotopic site (the cecal wall). Signicantly, the presence
720 Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc.

Figure 3. CSC Subsets within Primary Tumors May Harbor Metastatic and Tumor-Propagating Capacity
(A) A CSC may be responsible for both local and disseminated tumor propagation. In the example shown, a CSC (blue) enters the vasculature and metastasizes to a distant organ, where it seeds a heterogeneous tumor deposit exhibiting the hallmark features of the primary tumor. (B) Alternatively, genetic and/or epigenetic mechanisms acting in the primary CSC could lead to the emergence of a self-renewing metastatic CSC (green) expressing distinct markers from the original CSC. This metastatic CSC, through a series of invasive processes, seeds secondary tumors in distant organs.

of CD26+ cells in primary tumors also predicted metastasis in patients (Pang et al., 2010). In these epithelial malignancies, the epithelial mesenchymal transition (EMT) may underlie the metastatic process and give rise to precursors of metastatic CSCs at the invading edge of the tumor (Mani et al., 2008). In other tumors, a unique subset of metastatic CSCs may drive metastasis. In pancreatic cancer, only CD133+CXCR4+ cells (not CD133+CXCR4 cells) demonstrated metastatic activity, despite both subsets having tumor-propagating capacity (Hermann et al., 2007). Moreover, inhibition of CXCR4 signaling profoundly reduced the metastatic potential of pancreatic tumors without altering their tumorigenic potential. This metastatic CSC may have evolved from the primary tumor CSC or, alternatively, from a non-CSC within the tumor. The delineation of functionally distinct pools of CSCs will ultimately require cell tracing studies in vivo, via either mouse models relevant to human disease or minimally manipulated human cells for transplantation. Tracking of tumor cells in the circulation should also provide insight into metastatic CSCs. Not All Cancers Harbor CSCs Not all cancers will be sustained by CSCs. In melanoma, the high proportion of tumorigenic cells (as many as 50%) assayed

Cell Stem Cell

Perspective
under more permissive conditions and with a wide spectrum of markers (e.g., CD271), argues against a CSC model of heterogeneity (Quintana et al., 2010). On the other hand, there are data from two groups indicating that melanoma lesions contain a CSC subset characterized by CD271 expression with nude recipient mice (Boiko et al., 2010; Civenni et al., 2011). However, in more immunocompromised strains such as NOD-SCIDIL2Rg/ mice, both subsets were found to be tumorigenic, although the CD271 fraction did not phenocopy the original tumor histology (Civenni et al., 2011). One factor contributing to these disparities (besides recipient strain) will include the use of trypsin during tissue dissociation (Quintana et al., 2010), given that the CD271 antigen was shown to be sensitive to this enzyme (Civenni et al., 2011). Thus, the inclusion of trypsin in the dissociation procedure will result in contamination of the negative fraction with cells that actually express the antigen. Nevertheless, a large number of markers evaluated by Morrison and colleagues (Quintana et al., 2010) yielded cell populations that were tumorigenic irrespective of marker expression. It seems plausible that other parameters such as implantation conditions and tumor grade also contribute to the discrepant ndings. Another potential issue is that the frequency of CSCs can vary widely from 2.5% to 41% (Boiko et al., 2010), suggesting that the CSC model is not applicable to those tumors containing a high proportion of tumor-forming cells. Melanoma may also use distinct cellular mechanisms from most other solid malignancies, given the highly migratory nature of neural crest cells and their ability to respond to immune-based therapies. Role of the CSC Niche Cancers comprise malignant cells together with inammatory cells, hematopoietic cells, associated stroma, and vasculature. Although some CSCs conceivably do not require a dedicated niche, others will be dependent on a specic set of extrinsic interactions with their microenvironment. The niche effect on tumor cells may be inductive or selective but will inevitably differ for every tumor subtype. The perivascular niche of CSCs in brain cancers is the best characterized to date (reviewed in Gilbertson and Rich, 2007). In at least some glioblastomas, the relationship between the CSC and local environment appears to be bidirectional: the niche can alter the cellular fate of cancer cells and, conversely, CSCs can modify their microenvironment (Heddleston et al., 2009; Hjelmeland et al., 2011; Ricci-Vitiani et al., 2010; Wang et al., 2010b). Indeed, CSCs in glioblastoma have been demonstrated to secrete VEGF that directly supports the development of the local vasculature (Gilbertson and Rich, 2007). In the reverse direction, endothelial cells secrete nitric oxide that induces Notch signaling in glioma cells (Charles et al., 2010). It is relevant that CSCs but not non-CSCs in gliomas were shown to be dependent on nitric oxide synthase-2 (Eyler et al., 2011). Intriguingly, CSCs in glioblastomas can directly contribute to the microvasculature through their transdifferentiation into vascular cells (Ricci-Vitiani et al., 2010; Wang et al., 2010b), underscoring the close relationship between brain CSCs and their niche. The perivascular niche also serves a crucial role in the case of cutaneous squamous cell carcinomas. CSCs in this vascular niche establish an autocrine loop in which VEGF promotes CSC activity by governing both the microenvironment and intrinsic self-renewal pathways in CSCs (Beck et al., 2011). Even in nonsolid tumors, microenvironmental cues from cytokines, growth factors, or the immune-decient strain play an instructive role in determining the lineage fate of LSCs in a human model of leukemia (Wei et al., 2008). Hence, there has been considerable interest in targeting the putative CSC niche. Cells within the tumor-associated stroma, such as myobroblasts, are likely to have a prominent role in controlling CSC homeostasis in many tumor types. In colorectal cancer, myobroblasts secrete HGF that maintains CSC function by activating the Wnt pathway. Interestingly, tumor cells with an active Wnt canonical pathway were preferentially located adjacent to stromal myobroblasts (Vermeulen et al., 2010). Moreover, HGF-mediated activation of the Wnt pathway could induce CSC features and tumorigenic capacity in differentiated cancer cells that otherwise had limited tumorigenic capacity. Although these studies used spheroid cultures of primary colorectal cancers rather than fresh patient specimens, they suggest that the microenvironment can govern tumor cell stemness. Selective targeting of myobroblasts or the HGF/c-MET pathway would be predicted to interfere with the maintenance of CSCs and to potentially prevent the generation of CSCs from the non-CSC compartment. Notably, HGF is a potent inducer of the EMT, which plays a role in mediating invasion and metastasis. Using a mouse model of mammary tumorigenesis, another stromal factor, periostin, was shown to be essential for metastatic colonization by governing interactions between CSCs and their metastatic niche (Malanchi et al., 2012). Finally, the perturbation of other stromal mesenchymal cells such as osteoprogenitors can disrupt homeostasis, resulting in myelodysplasia and secondary leukemia (Raaijmakers et al., 2010). These ndings support the notion of niche-induced transformation and suggest that selective targeting of the tumor microenvironment may represent an alternative or adjunct to targeting the CSC. Even though it will be extraordinarily difcult to delineate the niche for human tumor CSCs and recapitulate the immune system of cancer patients, it is crucial to use an orthotopic transplantation assay to mimic the tumor environment as closely as possible. The coinoculation of human stromal cells to create a more appropriate environment for tumor development is also a relevant parameter to consider. The site of injection has been shown to directly inuence the frequency of tumor-initiating cells, underscoring the relevance of context. For example, the frequency of tumor-initiating cells in ovarian tumors was highest and most reliably read-out by the mammary fat pad assay rather than the ovarian bursa (Stewart et al., 2011). However, the questions arise as to why these microenvironments differentially inuence cell tumorigenicity and whether the tumor-initiating cells measured in the fat pad are in fact different from those assayed in the ovarian bursa. Pathways Regulating CSC Function Elucidation of the pathways that regulate the maintenance and survival of CSCs is important for the development of novel therapies. Not surprisingly, many CSC subsets and normal tissue stem cells seem to share core regulatory genes and developmental pathways such as c-myc, Bmi-1, and the Hedgehog (Hh), Notch, and Wnt pathways. Indeed, there is substantial evidence that restricted progenitors can generate LSCs by the reactivation of distinct self-renewal programs (Krivtsov et al.,
Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc. 721

Cell Stem Cell

Perspective
2006; Somervaille and Cleary, 2006; Somervaille et al., 2009). In chronic myeloid leukemia (CML), Hh signaling is essential for the maintenance and function of LSCs, and loss of Hh activity via disruption of Smoothened led to depletion of LSCs in vivo and prolonged animal survival (Dierks et al., 2008; Zhao et al., 2009). In the transition to the blast cell crisis phase, the LSC appears to originate from the granulocyte-macrophage progenitor cell through the acquired activation of the Wnt pathway (Jamieson et al., 2004). The Wnt pathway also plays a prominent role in the generation and self-renewal of LSCs in AML (Wang et al., 2010c). Signicantly, b-catenin activation endowed progenitor cells with self-renewing capability but was not essential for the renewal of normal adult HSCs. Parallel ndings were made for cutaneous CSCs versus normal skin stem cells in mouse models of skin cancer (Malanchi et al., 2008). Hence, the genetic programs governing self-renewal may be differentially active in normal and malignant stem cells, thereby opening therapeutic avenues. Cell polarity and metabolic pathways have recently been implicated in governing the function of CSCs. TAZ, a transcriptional effector in the Hippo pathway, was found to be frequently overexpressed in high-grade breast cancers and to maintain the selfrenewing capacity of tumorigenic cells isolated from established cell lines (Cordenonsi et al., 2011). A key link was established between the Hippo pathway and the cell polarity gene Scribbled, suggesting that cell polarity pathways may impact CSC function. It will be important to extend these studies to fresh tumor specimens because cancer cell lines do not reect in vivo tumor cell behavior. In a metabolic context, glycine decarboxylase was demonstrated to regulate the activity of tumor-propagating cells in non-small cell lung cancer (Zhang et al., 2012). Aberrant expression of glycine decarboxylase occurs in multiple cancer types and leads to changes in glycine/serine metabolism. The observation that CSC activity was dependent on glycine decarboxylase function provides a direct link between glycine metabolism and tumorigenesis. Although it is presumed that metabolic processes play a crucial role in all tumor cells, it is intriguing that they can selectively inuence CSC function. CSCs and Stemness Signatures In spite of the heterogeneity exhibited by CSCs, recent gene expression proling studies have provided important insights into the prognostic signicance of CSCs. The molecular analyses of functionally dened LSC populations from AML patients led to the generation of a LSC signature that largely reects a self-renewal or stemness signature (Eppert et al., 2011). This signature was found to be a strong predictor of poor prognosis, with the implication that it may be possible to identify patients at highest risk and to inform both the type and duration of their therapy. In a murine model of AML, a conserved signal transduction network was unveiled among different LSCs (Gibbs et al., 2012). Other recently derived stem cell signatures also exhibit prognostic value. In colorectal cancer, a gene signature derived for adult intestinal stem cells predicted relapse in patients rez and identied EphB2-positive CSCs in tumors (Merlos-Sua et al., 2011). Moreover, an embryonic stem cell (ESC)- and metastatic cell-based stem cell signature was found to increase with tumor grade and mortality in multiple tumor types (Shats et al., 2011), while an ESC-like transcriptional program evident in
722 Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc.

diverse epithelial cancers predicted poor prognosis (Wong et al., 2008). Preliminary data suggest that activation of this transcriptional program in adult cells may lead to the generation of CSCs. Another important player in ESCs, STAT3, was implicated in maintaining the stemness of glioma CSCs (Guryanova et al., 2011). Overall, these ndings suggest that targeting self-renewal pathways may represent one of the most effective strategies for eradicating CSCs (see below). Therapeutic Strategies to Target CSCs From a clinical perspective, it is important to decipher mechanisms of chemo- and radioresistance that operate in CSCs. Quiescent CSCs are thought to be more resistant to therapies while most CSCs seem to evade cytotoxic or radiotherapy through active mechanisms. There is clinical evidence for a subpopulation of chemotherapy-resistant CSCs in a number of solid tumors including breast cancer (Li et al., 2008; Yu et al., 2007). Furthermore, the analysis of breast tissue taken from patients pre- and postendocrine therapy or chemotherapy for gene expression changes revealed that residual breast cancers may be enriched for tumor cells with CSC-like and mesenchymal characteristics (Creighton et al., 2009). In patients with del(5q) myelodysplastic syndrome, rare stem cells were found to be refractory to therapeutic targeting in individuals in remission, which probably accounts for relapse (Tehranchi et al., 2010). Different aspects of CSCs have been explored in recent targeting strategies including quiescence, self-renewal pathways, radioresistance, and CSC-specic cell surface molecules. Several reports, predominantly for hematopoietic malignancies, indicate that CSCs can be selectively targeted without ablating normal stem cell function. Stem cell maintenance pathways are emerging as prime targets to eradicate CSCs. This approach, however, will be applicable only if the genetic programs controlling self-renewal are differentially active in malignant versus normal stem cells. It will be imperative to carefully evaluate the toxicity of anti-CSC agents on normal stem cell function in preclinical models. There are little data on the use of differentiation therapy in the context of CSCs but BMPs may be effective in inducing glial differentiation in glioblastomas and attenuating tumor growth (Piccirillo et al., 2006). In all likelihood, given the large number of mutations incurred by solid tumors such as breast (Wood et al., 2007), it will be essential to target multiple pathways that have been activated in CSCs in a given tumor. Quiescence or dormancy is a property of at least some CSCs such as those in leukemia. This feature has recently been exploited to provide a window for therapeutic intervention. Cytokines such as G-CSF efciently induced quiescent LSCs in AML to enter the cell cycle, thus sensitizing them to different chemotherapeutic agents (Saito et al., 2010). Indeed, combined G-CSF with chemotherapy elicited profound apoptosis and eradication of human AML stem cells in vivo. Inhibition of DNA repair mechanisms may also be harnessed for eradication of slow cycling LSCs (Viale et al., 2009). Although LSCs may reside in a more quiescent state, emerging evidence suggests that solid tumor CSCs follow a different pattern. In glioblastoma, CSCs are actively self-renewing and cellular diversity is most probably generated through symmetric cell division (Lathia et al., 2011b). Despite the observation that mouse mammary CSCs appear to

Cell Stem Cell

Perspective

Figure 4. Possible Mechanisms of Metastatic Relapse after Anticancer Therapy and Evaluation of Anti-CSC Treatments
(A) Late relapse can be accounted for by CSC dormancy. Here a dormant CSC (pink) that is resistant to both chemotherapy and targeted therapy has seeded to distant organs. After a considerable latency period, reactivation of a CSC will result in tumor growth and clinical emergence of metastases. Intriguingly, the clinical appearance of metastases is often synchronous in breast cancer. (B) One of the challenges facing the eld is the clinical translation of anti-CSC therapies. Strategies will need to be deployed in clinical trials that enable reproducible assays of CSC activity. In the scenario depicted here, biopsy material from a newly diagnosed breast cancer is subjected to a variety of assays to measure CSCs, including functional assays and gene expression proling, in parallel with the collection of blood samples and tumor imaging. The latter could include nanoparticle-labeling of anti-CSC markers coupled with in vivo imaging. These assays can be repeated after neoadjuvant therapy to determine whether the therapy has elicited a response against putative CSCs.

undergo more frequent symmetrical division than normal mammary stem cells in mammosphere cultures (Cicalese et al., 2009), it is tempting to speculate that the long period between primary tumor detection and relapse in patients with ER-positive breast cancer (up to 20 years) may reect a dormant stem cell subset (Figure 4). Of the key developmental pathways frequently deregulated in CSCs, considerable progress has been made in the case of targeted therapies against Notch and Hh, but the development of Wnt inhibitors has proven difcult. Signicantly, pharmacologic inhibition of the Hh pathway in human and mouse leukemias inhibited the expansion of imatinib-resistant CML (Dierks et al., 2008; Zhao et al., 2009). These ndings have profound implications because they suggest that treatment of imatinibresistant recurrence in CML patients may be achievable via targeting the Hh pathway. However, because Hh pathway activity is required for maintenance of normal HSCs, it will be crucial to determine the effects of these anticancer agents on all aspects of normal HSC function. Pharmacologic or siRNA-

mediated inhibition of Hh signaling in CSCs in glioblastoma, medulloblastoma, breast, pancreatic adenocarcinoma, and multiple myeloma has resulted in markedly reduced tumorigenic potential and, in some cases, ameliorated metastasis (reviewed in Merchant and Matsui, 2010). Hh ligands may play a dual role in the maintenance of CSCs and their niche, given the high stromal expression of these ligands. In terms of Notch signaling, CSCs in brain cancer were rendered more sensitive to radiation by blockade of this pathway (Wang et al., 2010a). Notably, Notch pathway inhibition via a neutralizing antibody against the DLL4 ligand was effective in reducing CSC numbers in diverse solid tumor xenografts (Hoey et al., 2009), whereas inhibition of Notch-4 expressed within the CSC subset largely ablated breast tumor growth (Harrison et al., 2010). A combination of Notch and Hh signaling may drive the self-renewal of CSCs in certain tumors such as undifferentiated pleomorphic sarcomas (Wang et al., 2012). Other self-renewal programs such as those regulated by Nodal and Activin, factors important for ESC maintenance, are also candidate targets. Pharmacologic inhibition of
Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc. 723

Cell Stem Cell

Perspective
the Nodal/Activin pathway sensitized CSCs to gemcitabine in a human xenograft model and signicantly prolonged survival when combined with a stroma-targeting Hh inhibitor to improve drug delivery (Lonardo et al., 2011). There has been considerable interest in the development of monoclonal antibodies to target CSCs. Markers differentially expressed between normal stem cells and LSCs have been utilized to specically target LSCs in human AML, including CD44 (Jin et al., 2006), IL3R (Jin et al., 2009), and the immunoglobulin mucin TIM-3 (Kikushige et al., 2010). In each case, treatment with antibodies against these cell surface molecules dramatically decreased leukemogenicity and eradicated LSCs as assessed by AML reconstitution in mice. Furthermore, the targeting of CD44 provides a paradigm for targeting CSC-niche interactions. Blocking antibodies against CD47, which serves as a dont eat me signal to tumor macrophages, may also be effective in eliminating LSCs in ALL that express higher levels of this antigen than their normal counterpart (Chao et al., 2011). Several studies have highlighted the radioresistance of CSCs in solid tumors, particularly in brain cancer. CSCs in fresh glioblastoma specimens or glioma xenografts are more resistant to ionizing irradiation (IR) in vivo than non-CSCs because of enhanced DNA repair pathways operating in CSCs (reviewed in Gilbertson and Rich, 2007). In medulloblastoma, targeting of cells in the perivascular region with Akt inhibitors enhanced responsiveness to radiation (Hambardzumyan et al., 2008), indicating that the CSC niche itself may serve as a therapeutic target. Interestingly, the DNA damage checkpoint response and radioresistance of CSCs in glioma is regulated in part by the adhesion molecule L1CAM through the activation of the ATM kinase pathway (Cheng et al., 2011). Similarly, radioresistance has been implicated in breast CSC-like populations that are thought to repair DNA damage more efciently. Inhibition of the Akt pathway led to the selective targeting of CSCs by blocking canonical WNT signaling and repair of DNA damage in these cells, thus sensitizing them to ionizing radiation (Zhang et al., 2010). In some breast tumors, lower ROS levels were found in certain CSC subsets compared with their nontumorigenic counterparts, perhaps conferring resistance to ionizing radiation (Diehn et al., 2009). Other therapeutic targets currently being pursued in the context of CSCs include growth factor receptor signaling networks. In pancreatic cancer, inhibition of the c-MET tyrosine kinase receptor diminished the CSC population and prevented metastasis, either alone or in combination with gemcitabine (Li et al., 2011). These inhibitors may also prove efcacious in colorectal cancer (Vermeulen et al., 2010). Recent studies on squamous carcinoma revealed that selective inhibition of VEGF signaling reduced CSC activity and led to tumor regression (Beck et al., 2011), implying that it may be necessary to target both CSCs and the stroma in which they reside. Cytokine pathways such as IL-8/CXCR1 are also emerging as important modulators of CSC activity. Repertaxin-mediated inhibition of this pathway reduced breast tumorigenesis and metastasis but only in combination with chemotherapy (Ginestier et al., 2010). In the search for novel drug discovery platforms, high throughput screens using small molecule, miRNA, or siRNA libraries have become an area of increasing focus. The application of a high throughput screen to target breast CSCs revealed
724 Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc.

a class of compounds that had previously not been implicated as cancer drugs: salinomycin-reduced tumor growth and lung metastases, possibly via direct targeting of breast CSCs (Gupta et al., 2009). An analogous screen to identify small molecule inhibitors of LSCs in AML led to the recent discovery of kinetic riboside (McDermott et al., 2012). MicroRNA-based therapies are emerging as novel modes of therapeutic intervention. Notably, systemic delivery of miR-34a, which is expressed at low levels in prostate CSCs, inhibited metastasis of prostate cancer cells and prolonged survival of mice (Liu et al., 2011). It is relevant that miR-34a targets CD44, a cell surface marker used to enrich prostate CSCs. Further largescale screens should be enabled by the development of improved surrogate in vitro culture assays that maintain the integrity of primary tumor-derived CSCs, rather than the use of cell lines. Ultimately, the testing of all putative anti-CSC agents requires preclinical mouse models containing early passage xenografts (or leukemic cells) to obviate any changes that occur upon prolonged passage. Evolution of CSCs and Tumor Cell Plasticity It is important to note that the CSC and clonal evolution concepts are not mutually exclusive. Two recent papers have highlighted a high degree of convergence between these models in leukemia. LSCs in acute lymphoblastic leukemia harboring the ETV6-Runx1 translocation were shown to be genetically diverse, exhibiting different degrees of self-renewing and leukemogenic activity in vivo (Anderson et al., 2011). This study provides evidence that CSCs within individual cancer patients can be genetically heterogeneous, presumably accounting for their variable biological properties such as self-renewal potential. Moreover, in BCR-ABL acute lymphoblastic leukemia patients, the leukemia-initiating population displayed profound genetic diversity, with multiple genetically distinct tumor-initiating subclones at diagnosis (Notta et al., 2011). Although it has not yet been determined whether these tumor-initiating cells follow the CSC model, it seems likely. In parallel with the other study, a nonlinear, branching model of tumor evolution was identied. Taken together, these studies illustrate the importance of complementing functional assays of cellular heterogeneity with genetic ngerprinting of the different subsets. In addition to genetic variegation, tumor cell plasticity may contribute to phenotypic and functional heterogeneity. Many cell surface markers on melanoma cells are reversibly expressed, such that phenotypically diverse melanoma cells can recapitulate tumor heterogeneity of the parent tumor, irrespective of whether they arose from marker-positive or markernegative cells (Quintana et al., 2010). Regulatory genes may also be transiently or stochastically expressed. JARID1Bmediated histone demethylation was demonstrated to be reversibly expressed in melanoma cell lines and to be essential for the maintenance of tumorigenic activity (Roesch et al., 2010). Slow cycling JARID1B-expressing cells could arise from a negative population even when initiated from a single cell, suggesting that a nontumorigenic cell may reacquire stem cell-like properties. In addition, cells within breast cancer cell lines were found to transition stochastically between states to establish a stable phenotypic equilibrium (Gupta et al., 2011), and CSC-like cells could arise de novo from transformed breast epithelial cells

Cell Stem Cell

Perspective
(Chaffer et al., 2011). These data suggest that cellular interconversion might occur in a stochastic manner. CSC function may be induced by specic microenvironmental cues from growth factors or in stress-related contexts. For example, HGF has been implicated in the reprogramming of non-CSCs toward a CSC-like phenotype in colon cancer (Vermeulen et al., 2010). Furthermore, induction of an EMT in immortalized human breast epithelial cells can endow them with stem cell-like properties and potentially promote the generation of CSCs from tumor cells (Mani et al., 2008). HIF2a, induced as a cellular response to a hypoxic microenvironment, has been implicated in the maintenance of CSCs in glioblastoma and may promote the interconversion of nonstem (CD133) to CSC-like cells (Heddleston et al., 2009; Li et al., 2009). The reversible state, in its various guises discussed above, has profound implications for the treatment and management of patients. Moreover, therapeutic resistance itself may potentially reect a reversible state (Sharma et al., 2010). Finally, it should be noted that many of the studies described in this section used cell lines or cultured primary cells, and that at present there is no evidence for dedifferentiation occurring in primary tumors in vivo. However, it is noteworthy that dedifferentiation can occur at a low frequency in normal tissues such as the testis (Barroca et al., 2009). Because denitive cell surface markers are lacking for most stem cells and their descendants, at least in solid organs, it is difcult to study cellular plasticity. Further elucidation of differentiation hierarchies may eventually enable plasticity and interconversion of cells between nontumorigenic and tumorigenic states to be formally tested by clonal cell tracking analysis in vivo. Conclusions An emerging consensus in the eld is that cellular state rather than phenotype is important when dening a CSC. The uniformity between LSC signatures that has emerged across diverse samples, despite interpatient variation in CSC markers, conrms that current phenotypic markers are not a reliable measure of CSCs. One corollary of high interpatient variation is that an extensive range of markers will need to be validated in a large number of patient samples, perhaps even necessitating functional assessment for each patient. Indeed, the purication of CSCs using a robust set of markers, even from a given tumor subtype, remains a major challenge for the eld. If CSCs exist in a dynamic state in certain tumors, then this will inevitably confound their prospective isolation. The existence of multiple CSC pools or evolving intratumoral clones in individual tumors demands the monitoring of these populations in pre- and posttreatment samples by multicolor ow cytometry or highresolution molecular imaging to identify residual cells that might drive relapse (Figure 4). The derivation of robust signatures that distinguish CSCs from normal stem cells may also facilitate the evaluation of clinically relevant residual cells. The clinical applicability of the CSC concept to predicting patient response remains a fundamental question. Most putative anti-CSC therapies to date have attenuated rather than eradicated solid tumors in preclinical models, and efcacious response often required concomitant chemotherapy. In order to improve the evaluation of efcacy of anti-CSC agents in clinical trials, there is a pressing need to optimize assays for CSC targeting and measurement of tumor response. In standard clinical trials, tumor response criteria depend on measurements of tumor size, which largely reects tumor response in the nonCSC tumor bulk. Specic response criteria that provide a readout of response to anti-CSC agents in clinical trials remain elusive (Figure 4). Tumor sphere-forming assays and measurement of CSC marker expression are unlikely to provide robust surrogate markers in a clinical setting. The incorporation of other measures such as self-renewal activity into therapeutic strategies will almost certainly be required. We speculate that for most tumor types it will still prove necessary to test novel antiCSC therapies in combination with tumor debulking (non-CSC) therapy, such as conventional chemotherapy.
ACKNOWLEDGMENTS We apologize to those authors whose papers could not be cited due to space constraints. We thank P. Maltezos for expert help with the gures. This work was supported by the Australian National Health and Medical Research Council (NHMRC), NHMRC IRIISS, and the Victorian State Government through Victorian Cancer Agency funding of the Victorian Breast Cancer Research Consortium and OIS. REFERENCES Anderson, K., Lutz, C., van Delft, F.W., Bateman, C.M., Guo, Y., Colman, S.M., Kempski, H., Moorman, A.V., Titley, I., Swansbury, J., et al. (2011). Genetic variegation of clonal architecture and propagating cells in leukaemia. Nature 469, 356361. Barroca, V., Lassalle, B., Coureuil, M., Louis, J.P., Le Page, F., Testart, J., Allemand, I., Riou, L., and Fouchet, P. (2009). Mouse differentiating spermatogonia can generate germinal stem cells in vivo. Nat. Cell Biol. 11, 190196. Beck, B., Driessens, G., Goossens, S., Youssef, K.K., Kuchnio, A., Caauwe, A., Sotiropoulou, P.A., Loges, S., Lapouge, G., Candi, A., et al. (2011). A vascular niche and a VEGF-Nrp1 loop regulate the initiation and stemness of skin tumours. Nature 478, 399403. Boiko, A.D., Razorenova, O.V., van de Rijn, M., Swetter, S.M., Johnson, D.L., Ly, D.P., Butler, P.D., Yang, G.P., Joshua, B., Kaplan, M.J., et al. (2010). Human melanoma-initiating cells express neural crest nerve growth factor receptor CD271. Nature 466, 133137. Bonnet, D., and Dick, J.E. (1997). Human acute myeloid leukemia is organized as a hierarchy that originates from a primitive hematopoietic cell. Nat. Med. 3, 730737. Chaffer, C.L., Brueckmann, I., Scheel, C., Kaestli, A.J., Wiggins, P.A., Rodrigues, L.O., Brooks, M., Reinhardt, F., Su, Y., Polyak, K., et al. (2011). Normal and neoplastic nonstem cells can spontaneously convert to a stemlike state. Proc. Natl. Acad. Sci. USA 108, 79507955. Chao, M.P., Alizadeh, A.A., Tang, C., Jan, M., Weissman-Tsukamoto, R., Zhao, F., Park, C.Y., Weissman, I.L., and Majeti, R. (2011). Therapeutic antibody targeting of CD47 eliminates human acute lymphoblastic leukemia. Cancer Res. 71, 13741384. Charles, N., Ozawa, T., Squatrito, M., Bleau, A.M., Brennan, C.W., Hambardzumyan, D., and Holland, E.C. (2010). Perivascular nitric oxide activates notch signaling and promotes stem-like character in PDGF-induced glioma cells. Cell Stem Cell 6, 141152. Chen, R., Nishimura, M.C., Bumbaca, S.M., Kharbanda, S., Forrest, W.F., Kasman, I.M., Greve, J.M., Soriano, R.H., Gilmour, L.L., Rivers, C.S., et al. (2010). A hierarchy of self-renewing tumor-initiating cell types in glioblastoma. Cancer Cell 17, 362375. Cheng, L., Wu, Q., Huang, Z., Guryanova, O.A., Huang, Q., Shou, W., Rich, J.N., and Bao, S. (2011). L1CAM regulates DNA damage checkpoint response of glioblastoma stem cells through NBS1. EMBO J. 30, 800813. Cicalese, A., Bonizzi, G., Pasi, C.E., Faretta, M., Ronzoni, S., Giulini, B., Brisken, C., Minucci, S., Di Fiore, P.P., and Pelicci, P.G. (2009). The tumor

Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc. 725

Cell Stem Cell

Perspective
suppressor p53 regulates polarity of self-renewing divisions in mammary stem cells. Cell 138, 10831095. Civenni, G., Walter, A., Kobert, N., Mihic-Probst, D., Zipser, M., Belloni, B., Seifert, B., Moch, H., Dummer, R., van den Broek, M., and Sommer, L. (2011). Human CD271-positive melanoma stem cells associated with metastasis establish tumor heterogeneity and long-term growth. Cancer Res. 71, 3098 3109. Cordenonsi, M., Zanconato, F., Azzolin, L., Forcato, M., Rosato, A., Frasson, C., Inui, M., Montagner, M., Parenti, A.R., Poletti, A., et al. (2011). The Hippo transducer TAZ confers cancer stem cell-related traits on breast cancer cells. Cell 147, 759772. Creighton, C.J., Li, X., Landis, M., Dixon, J.M., Neumeister, V.M., Sjolund, A., Rimm, D.L., Wong, H., Rodriguez, A., Herschkowitz, J.I., et al. (2009). Residual breast cancers after conventional therapy display mesenchymal as well as tumor-initiating features. Proc. Natl. Acad. Sci. USA 106, 1382013825. Curley, M.D., Therrien, V.A., Cummings, C.L., Sergent, P.A., Koulouris, C.R., Friel, A.M., Roberts, D.J., Seiden, M.V., Scadden, D.T., Rueda, B.R., and Foster, R. (2009). CD133 expression denes a tumor initiating cell population in primary human ovarian cancer. Stem Cells 27, 28752883. Curtis, S.J., Sinkevicius, K.W., Li, D., Lau, A.N., Roach, R.R., Zamponi, R., Woolfenden, A.E., Kirsch, D.G., Wong, K.K., and Kim, C.F. (2010). Primary tumor genotype is an important determinant in identication of lung cancer propagating cells. Cell Stem Cell 7, 127133. Dalerba, P., Dylla, S.J., Park, I.K., Liu, R., Wang, X., Cho, R.W., Hoey, T., Gurney, A., Huang, E.H., Simeone, D.M., et al. (2007). Phenotypic characterization of human colorectal cancer stem cells. Proc. Natl. Acad. Sci. USA 104, 1015810163. Diehn, M., Cho, R.W., Lobo, N.A., Kalisky, T., Dorie, M.J., Kulp, A.N., Qian, D., Lam, J.S., Ailles, L.E., Wong, M., et al. (2009). Association of reactive oxygen species levels and radioresistance in cancer stem cells. Nature 458, 780783. Dierks, C., Beigi, R., Guo, G.R., Zirlik, K., Stegert, M.R., Manley, P., Trussell, C., Schmitt-Graeff, A., Landwerlin, K., Veelken, H., and Warmuth, M. (2008). Expansion of Bcr-Abl-positive leukemic stem cells is dependent on Hedgehog pathway activation. Cancer Cell 14, 238249. Dieter, S.M., Ball, C.R., Hoffmann, C.M., Nowrouzi, A., Herbst, F., Zavidij, O., Abel, U., Arens, A., Weichert, W., Brand, K., et al. (2011). Distinct types of tumor-initiating cells form human colon cancer tumors and metastases. Cell Stem Cell 9, 357365. Eppert, K., Takenaka, K., Lechman, E.R., Waldron, L., Nilsson, B., van Galen, P., Metzeler, K.H., Poeppl, A., Ling, V., Beyene, J., et al. (2011). Stem cell gene expression programs inuence clinical outcome in human leukemia. Nat. Med. 17, 10861093. Eyler, C.E., Wu, Q., Yan, K., MacSwords, J.M., Chandler-Militello, D., Misuraca, K.L., Lathia, J.D., Forrester, M.T., Lee, J., Stamler, J.S., et al. (2011). Glioma stem cell proliferation and tumor growth are promoted by nitric oxide synthase-2. Cell 146, 5366. Gentles, A.J., Plevritis, S.K., Majeti, R., and Alizadeh, A.A. (2010). Association of a leukemic stem cell gene expression signature with clinical outcomes in acute myeloid leukemia. JAMA 304, 27062715. Gibbs, K.D., Jr., Jager, A., Crespo, O., Goltsev, Y., Trejo, A., Richard, C.E., and Nolan, G.P. (2012). Decoupling of tumor-initiating activity from stable immunophenotype in HoxA9-Meis1-driven AML. Cell Stem Cell 10, 210217. Gilbertson, R.J., and Rich, J.N. (2007). Making a tumours bed: glioblastoma stem cells and the vascular niche. Nat. Rev. Cancer 7, 733736. Ginestier, C., Hur, M.H., Charafe-Jauffret, E., Monville, F., Dutcher, J., Brown, M., Jacquemier, J., Viens, P., Kleer, C.G., Liu, S., et al. (2007). ALDH1 is a marker of normal and malignant human mammary stem cells and a predictor of poor clinical outcome. Cell Stem Cell 1, 555567. Ginestier, C., Liu, S., Diebel, M.E., Korkaya, H., Luo, M., Brown, M., Wicinski, J., Cabaud, O., Charafe-Jauffret, E., Birnbaum, D., et al. (2010). CXCR1 blockade selectively targets human breast cancer stem cells in vitro and in xenografts. J. Clin. Invest. 120, 485497. Goardon, N., Marchi, E., Atzberger, A., Quek, L., Schuh, A., Soneji, S., Woll, P., Mead, A., Alford, K.A., Rout, R., et al. (2011). Coexistence of LMPP-like and GMP-like leukemia stem cells in acute myeloid leukemia. Cancer Cell 19, 138152. Gupta, P.B., Onder, T.T., Jiang, G., Tao, K., Kuperwasser, C., Weinberg, R.A., and Lander, E.S. (2009). Identication of selective inhibitors of cancer stem cells by high-throughput screening. Cell 138, 645659. Gupta, P.B., Fillmore, C.M., Jiang, G., Shapira, S.D., Tao, K., Kuperwasser, C., and Lander, E.S. (2011). Stochastic state transitions give rise to phenotypic equilibrium in populations of cancer cells. Cell 146, 633644. Guryanova, O.A., Wu, Q., Cheng, L., Lathia, J.D., Huang, Z., Yang, J., MacSwords, J., Eyler, C.E., McLendon, R.E., Heddleston, J.M., et al. (2011). Nonreceptor tyrosine kinase BMX maintains self-renewal and tumorigenic potential of glioblastoma stem cells by activating STAT3. Cancer Cell 19, 498511. Hambardzumyan, D., Becher, O.J., Rosenblum, M.K., Pandol, P.P., ManovaTodorova, K., and Holland, E.C. (2008). PI3K pathway regulates survival of cancer stem cells residing in the perivascular niche following radiation in medulloblastoma in vivo. Genes Dev. 22, 436448. Harrison, H., Farnie, G., Howell, S.J., Rock, R.E., Stylianou, S., Brennan, K.R., Bundred, N.J., and Clarke, R.B. (2010). Regulation of breast cancer stem cell activity by signaling through the Notch4 receptor. Cancer Res. 70, 709718. Heddleston, J.M., Li, Z., McLendon, R.E., Hjelmeland, A.B., and Rich, J.N. (2009). The hypoxic microenvironment maintains glioblastoma stem cells and promotes reprogramming towards a cancer stem cell phenotype. Cell Cycle 8, 32743284. Hermann, P.C., Huber, S.L., Herrler, T., Aicher, A., Ellwart, J.W., Guba, M., Bruns, C.J., and Heeschen, C. (2007). Distinct populations of cancer stem cells determine tumor growth and metastatic activity in human pancreatic cancer. Cell Stem Cell 1, 313323. Heuser, M., Sly, L.M., Argiropoulos, B., Kuchenbauer, F., Lai, C., Weng, A., Leung, M., Lin, G., Brookes, C., Fung, S., et al. (2009). Modeling the functional heterogeneity of leukemia stem cells: role of STAT5 in leukemia stem cell selfrenewal. Blood 114, 39833993. Hjelmeland, A.B., Wu, Q., Heddleston, J.M., Choudhary, G.S., MacSwords, J., Lathia, J.D., McLendon, R., Lindner, D., Sloan, A., and Rich, J.N. (2011). Acidic stress promotes a glioma stem cell phenotype. Cell Death Differ. 18, 829840. Hoey, T., Yen, W.C., Axelrod, F., Basi, J., Donigian, L., Dylla, S., Fitch-Bruhns, M., Lazetic, S., Park, I.K., Sato, A., et al. (2009). DLL4 blockade inhibits tumor growth and reduces tumor-initiating cell frequency. Cell Stem Cell 5, 168177. Ishizawa, K., Rasheed, Z.A., Karisch, R., Wang, Q., Kowalski, J., Susky, E., Pereira, K., Karamboulas, C., Moghal, N., Rajeshkumar, N.V., et al. (2010). Tumor-initiating cells are rare in many human tumors. Cell Stem Cell 7, 279282. Jamieson, C.H., Ailles, L.E., Dylla, S.J., Muijtjens, M., Jones, C., Zehnder, J.L., Gotlib, J., Li, K., Manz, M.G., Keating, A., et al. (2004). Granulocyte-macrophage progenitors as candidate leukemic stem cells in blast-crisis CML. N. Engl. J. Med. 351, 657667. Jin, L., Hope, K.J., Zhai, Q., Smadja-Joffe, F., and Dick, J.E. (2006). Targeting of CD44 eradicates human acute myeloid leukemic stem cells. Nat. Med. 12, 11671174. Jin, L., Lee, E.M., Ramshaw, H.S., Buseld, S.J., Peoppl, A.G., Wilkinson, L., Guthridge, M.A., Thomas, D., Barry, E.F., Boyd, A., et al. (2009). Monoclonal antibody-mediated targeting of CD123, IL-3 receptor alpha chain, eliminates human acute myeloid leukemic stem cells. Cell Stem Cell 5, 3142. Kikushige, Y., Shima, T., Takayanagi, S., Urata, S., Miyamoto, T., Iwasaki, H., Takenaka, K., Teshima, T., Tanaka, T., Inagaki, Y., and Akashi, K. (2010). TIM-3 is a promising target to selectively kill acute myeloid leukemia stem cells. Cell Stem Cell 7, 708717. Krivtsov, A.V., Twomey, D., Feng, Z., Stubbs, M.C., Wang, Y., Faber, J., Levine, J.E., Wang, J., Hahn, W.C., Gilliland, D.G., et al. (2006). Transformation from committed progenitor to leukaemia stem cell initiated by MLL-AF9. Nature 442, 818822. Lapidot, T., Sirard, C., Vormoor, J., Murdoch, B., Hoang, T., Caceres-Cortes, J., Minden, M., Paterson, B., Caligiuri, M.A., and Dick, J.E. (1994). A cell initiating human acute myeloid leukaemia after transplantation into SCID mice. Nature 367, 645648. Lathia, J.D., Gallagher, J., Heddleston, J.M., Wang, J., Eyler, C.E., Macswords, J., Wu, Q., Vasanji, A., McLendon, R.E., Hjelmeland, A.B., and

726 Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc.

Cell Stem Cell

Perspective
Rich, J.N. (2010). Integrin alpha 6 regulates glioblastoma stem cells. Cell Stem Cell 6, 421432. Lathia, J.D., Gallagher, J., Myers, J.T., Li, M., Vasanji, A., McLendon, R.E., Hjelmeland, A.B., Huang, A.Y., and Rich, J.N. (2011a). Direct in vivo evidence for tumor propagation by glioblastoma cancer stem cells. PLoS ONE 6, e24807. Lathia, J.D., Hitomi, M., Gallagher, J., Gadani, S.P., Adkins, J., Vasanji, A., Liu, L., Eyler, C.E., Heddleston, J.M., Wu, Q., et al. (2011b). Distribution of CD133 reveals glioma stem cells self-renew through symmetric and asymmetric cell divisions. Cell Death Dis 2, e200. Li, X., Lewis, M.T., Huang, J., Gutierrez, C., Osborne, C.K., Wu, M.F., Hilsenbeck, S.G., Pavlick, A., Zhang, X., Chamness, G.C., et al. (2008). Intrinsic resistance of tumorigenic breast cancer cells to chemotherapy. J. Natl. Cancer Inst. 100, 672679. Li, Z., Bao, S., Wu, Q., Wang, H., Eyler, C., Sathornsumetee, S., Shi, Q., Cao, Y., Lathia, J., McLendon, R.E., et al. (2009). Hypoxia-inducible factors regulate tumorigenic capacity of glioma stem cells. Cancer Cell 15, 501513. Li, C., Wu, J.J., Hynes, M., Dosch, J., Sarkar, B., Welling, T.H., Pasca di Magliano, M., and Simeone, D.M. (2011). c-Met is a marker of pancreatic cancer stem cells and therapeutic target. Gastroenterology 141, 22182227, e5. Liu, H., Patel, M.R., Prescher, J.A., Patsialou, A., Qian, D., Lin, J., Wen, S., Chang, Y.F., Bachmann, M.H., Shimono, Y., et al. (2010). Cancer stem cells from human breast tumors are involved in spontaneous metastases in orthotopic mouse models. Proc. Natl. Acad. Sci. USA 107, 1811518120. Liu, C., Kelnar, K., Liu, B., Chen, X., Calhoun-Davis, T., Li, H., Patrawala, L., Yan, H., Jeter, C., Honorio, S., et al. (2011). The microRNA miR-34a inhibits prostate cancer stem cells and metastasis by directly repressing CD44. Nat. Med. 17, 211215. Lonardo, E., Hermann, P.C., Mueller, M.T., Huber, S., Balic, A., MirandaLorenzo, I., Zagorac, S., Alcala, S., Rodriguez-Arabaolaza, I., Ramirez, J.C., et al. (2011). Nodal/Activin signaling drives self-renewal and tumorigenicity of pancreatic cancer stem cells and provides a target for combined drug therapy. Cell Stem Cell 9, 433446. Malanchi, I., Peinado, H., Kassen, D., Hussenet, T., Metzger, D., Chambon, P., Huber, M., Hohl, D., Cano, A., Birchmeier, W., and Huelsken, J. (2008). Cutaneous cancer stem cell maintenance is dependent on beta-catenin signalling. Nature 452, 650653. nez, A., Susanto, E., Peng, H., Lehr, H.A., DelaMalanchi, I., Santamaria-Mart loye, J.F., and Huelsken, J. (2012). Interactions between cancer stem cells and their niche govern metastatic colonization. Nature 481, 8589. Mani, S.A., Guo, W., Liao, M.J., Eaton, E.N., Ayyanan, A., Zhou, A.Y., Brooks, M., Reinhard, F., Zhang, C.C., Shipitsin, M., et al. (2008). The epithelialmesenchymal transition generates cells with properties of stem cells. Cell 133, 704715. McDermott, S.P., Eppert, K., Notta, F., Isaac, M., Datti, A., Al-Awar, R., Wrana, J., Minden, M.D., and Dick, J.E. (2012). A small molecule screening strategy with validation on human leukemia stem cells uncovers the therapeutic efcacy of kinetin riboside. Blood 119, 12001207. Merchant, A.A., and Matsui, W. (2010). Targeting Hedgehoga cancer stem cell pathway. Clin. Cancer Res. 16, 31303140. rez, A., Barriga, F.M., Jung, P., Iglesias, M., Ce spedes, M.V., RosMerlos-Sua oz, P., sell, D., Sevillano, M., Hernando-Momblona, X., da Silva-Diz, V., Mun et al. (2011). The intestinal stem cell signature identies colorectal cancer stem cells and predicts disease relapse. Cell Stem Cell 8, 511524. Meyer, M.J., Fleming, J.M., Lin, A.F., Hussnain, S.A., Ginsburg, E., and Vonderhaar, B.K. (2010). CD44posCD49fhiCD133/2hi denes xenograftinitiating cells in estrogen receptor-negative breast cancer. Cancer Res. 70, 46244633. Notta, F., Mullighan, C.G., Wang, J.C., Poeppl, A., Doulatov, S., Phillips, L.A., Ma, J., Minden, M.D., Downing, J.R., and Dick, J.E. (2011). Evolution of human BCR-ABL1 lymphoblastic leukaemia-initiating cells. Nature 469, 362367. Nowell, P.C. (1976). The clonal evolution of tumor cell populations. Science 194, 2328. Pang, R., Law, W.L., Chu, A.C., Poon, J.T., Lam, C.S., Chow, A.K., Ng, L., Cheung, L.W., Lan, X.R., Lan, H.Y., et al. (2010). A subpopulation of CD26+ cancer stem cells with metastatic capacity in human colorectal cancer. Cell Stem Cell 6, 603615. Pece, S., Tosoni, D., Confalonieri, S., Mazzarol, G., Vecchi, M., Ronzoni, S., Bernard, L., Viale, G., Pelicci, P.G., and Di Fiore, P.P. (2010). Biological and molecular heterogeneity of breast cancers correlates with their cancer stem cell content. Cell 140, 6273. Piccirillo, S.G., Reynolds, B.A., Zanetti, N., Lamorte, G., Binda, E., Broggi, G., Brem, H., Olivi, A., Dimeco, F., and Vescovi, A.L. (2006). Bone morphogenetic proteins inhibit the tumorigenic potential of human brain tumour-initiating cells. Nature 444, 761765. Quintana, E., Shackleton, M., Sabel, M.S., Fullen, D.R., Johnson, T.M., and Morrison, S.J. (2008). Efcient tumour formation by single human melanoma cells. Nature 456, 593598. Quintana, E., Shackleton, M., Foster, H.R., Fullen, D.R., Sabel, M.S., Johnson, T.M., and Morrison, S.J. (2010). Phenotypic heterogeneity among tumorigenic melanoma cells from patients that is reversible and not hierarchically organized. Cancer Cell 18, 510523. Raaijmakers, M.H., Mukherjee, S., Guo, S., Zhang, S., Kobayashi, T., Schoonmaker, J.A., Ebert, B.L., Al-Shahrour, F., Hasserjian, R.P., Scadden, E.O., et al. (2010). Bone progenitor dysfunction induces myelodysplasia and secondary leukaemia. Nature 464, 852857. Read, T.A., Fogarty, M.P., Markant, S.L., McLendon, R.E., Wei, Z., Ellison, D.W., Febbo, P.G., and Wechsler-Reya, R.J. (2009). Identication of CD15 as a marker for tumor-propagating cells in a mouse model of medulloblastoma. Cancer Cell 15, 135147. Ricci-Vitiani, L., Pallini, R., Biffoni, M., Todaro, M., Invernici, G., Cenci, T., Maira, G., Parati, E.A., Stassi, G., Larocca, L.M., and De Maria, R. (2010). Tumour vascularization via endothelial differentiation of glioblastoma stemlike cells. Nature 468, 824828. Roesch, A., Fukunaga-Kalabis, M., Schmidt, E.C., Zabierowski, S.E., Brafford, P.A., Vultur, A., Basu, D., Gimotty, P., Vogt, T., and Herlyn, M. (2010). A temporarily distinct subpopulation of slow-cycling melanoma cells is required for continuous tumor growth. Cell 141, 583594. Saito, Y., Uchida, N., Tanaka, S., Suzuki, N., Tomizawa-Murasawa, M., Sone, A., Najima, Y., Takagi, S., Aoki, Y., Wake, A., et al. (2010). Induction of cell cycle entry eliminates human leukemia stem cells in a mouse model of AML. Nat. Biotechnol. 28, 275280. Sarry, J.E., Murphy, K., Perry, R., Sanchez, P.V., Secreto, A., Keefer, C., cher, C., et al. (2011). Human Swider, C.R., Strzelecki, A.C., Cavelier, C., Re acute myelogenous leukemia stem cells are rare and heterogeneous when assayed in NOD/SCID/IL2Rgc-decient mice. J. Clin. Invest. 121, 384395. Schober, M., and Fuchs, E. (2011). Tumor-initiating stem cells of squamous cell carcinomas and their control by TGF-b and integrin/focal adhesion kinase (FAK) signaling. Proc. Natl. Acad. Sci. USA 108, 1054410549. Sharma, S.V., Lee, D.Y., Li, B., Quinlan, M.P., Takahashi, F., Maheswaran, S., McDermott, U., Azizian, N., Zou, L., Fischbach, M.A., et al. (2010). A chromatin-mediated reversible drug-tolerant state in cancer cell subpopulations. Cell 141, 6980. Shats, I., Gatza, M.L., Chang, J.T., Mori, S., Wang, J., Rich, J., and Nevins, J.R. (2011). Using a stem cell-based signature to guide therapeutic selection in cancer. Cancer Res. 71, 17721780. Silva, I.A., Bai, S., McLean, K., Yang, K., Grifth, K., Thomas, D., Ginestier, C., Johnston, C., Kueck, A., Reynolds, R.K., et al. (2011). Aldehyde dehydrogenase in combination with CD133 denes angiogenic ovarian cancer stem cells that portend poor patient survival. Cancer Res. 71, 39914001. Singh, S.K., Hawkins, C., Clarke, I.D., Squire, J.A., Bayani, J., Hide, T., Henkelman, R.M., Cusimano, M.D., and Dirks, P.B. (2004). Identication of human brain tumour initiating cells. Nature 432, 396401. Somervaille, T.C., and Cleary, M.L. (2006). Identication and characterization of leukemia stem cells in murine MLL-AF9 acute myeloid leukemia. Cancer Cell 10, 257268. Somervaille, T.C., Matheny, C.J., Spencer, G.J., Iwasaki, M., Rinn, J.L., Witten, D.M., Chang, H.Y., Shurtleff, S.A., Downing, J.R., and Cleary, M.L. (2009). Hierarchical maintenance of MLL myeloid leukemia stem cells employs a transcriptional program shared with embryonic rather than adult stem cells. Cell Stem Cell 4, 129140.

Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc. 727

Cell Stem Cell

Perspective
Son, M.J., Woolard, K., Nam, D.H., Lee, J., and Fine, H.A. (2009). SSEA-1 is an enrichment marker for tumor-initiating cells in human glioblastoma. Cell Stem Cell 4, 440452. Stewart, J.M., Shaw, P.A., Gedye, C., Bernardini, M.Q., Neel, B.G., and Ailles, L.E. (2011). Phenotypic heterogeneity and instability of human ovarian tumorinitiating cells. Proc. Natl. Acad. Sci. USA 108, 64686473. Taussig, D.C., Miraki-Moud, F., Anjos-Afonso, F., Pearce, D.J., Allen, K., Ridler, C., Lillington, D., Oakervee, H., Cavenagh, J., Agrawal, S.G., et al. (2008). Anti-CD38 antibody-mediated clearance of human repopulating cells masks the heterogeneity of leukemia-initiating cells. Blood 112, 568575. Tehranchi, R., Woll, P.S., Anderson, K., Buza-Vidas, N., Mizukami, T., Mead, m, I., Stro mbeck, B., Horvat, A., Ferry, H., et al. A.J., Astrand-Grundstro (2010). Persistent malignant stem cells in del(5q) myelodysplasia in remission. N. Engl. J. Med. 363, 10251037. Vermeulen, L., De Sousa E Melo, F., van der Heijden, M., Cameron, K., de Jong, J.H., Borovski, T., Tuynman, J.B., Todaro, M., Merz, C., Rodermond, H., et al. (2010). Wnt activity denes colon cancer stem cells and is regulated by the microenvironment. Nat. Cell Biol. 12, 468476. Viale, A., De Franco, F., Orleth, A., Cambiaghi, V., Giuliani, V., Bossi, D., Ronchini, C., Ronzoni, S., Muradore, I., Monestiroli, S., et al. (2009). Cell-cycle restriction limits DNA damage and maintains self-renewal of leukaemia stem cells. Nature 457, 5156. Visvader, J.E., and Lindeman, G.J. (2008). Cancer stem cells in solid tumours: accumulating evidence and unresolved questions. Nat. Rev. Cancer 8, 755768. Wang, J., Wakeman, T.P., Lathia, J.D., Hjelmeland, A.B., Wang, X.F., White, R.R., Rich, J.N., and Sullenger, B.A. (2010a). Notch promotes radioresistance of glioma stem cells. Stem Cells 28, 1728. Wang, R., Chadalavada, K., Wilshire, J., Kowalik, U., Hovinga, K.E., Geber, A., Fligelman, B., Leversha, M., Brennan, C., and Tabar, V. (2010b). Glioblastoma stem-like cells give rise to tumour endothelium. Nature 468, 829833. Wang, Y., Krivtsov, A.V., Sinha, A.U., North, T.E., Goessling, W., Feng, Z., Zon, L.I., and Armstrong, S.A. (2010c). The Wnt/beta-catenin pathway is required for the development of leukemia stem cells in AML. Science 327, 16501653. Wang, C.Y., Wei, Q., Han, I., Sato, S., Ghanbari-Azarnier, R., Whetstone, H., Poon, R., Hu, J., Zheng, F., Zhang, P., et al. (2012). Hedgehog and Notch signaling regulate self-renewal of undifferentiated pleomorphic sarcomas. Cancer Res. 72, 10131022. Ward, R.J., Lee, L., Graham, K., Satkunendran, T., Yoshikawa, K., Ling, E., Harper, L., Austin, R., Nieuwenhuis, E., Clarke, I.D., et al. (2009). Multipotent CD15+ cancer stem cells in patched-1-decient mouse medulloblastoma. Cancer Res. 69, 46824690. Wei, J., Wunderlich, M., Fox, C., Alvarez, S., Cigudosa, J.C., Wilhelm, J.S., Zheng, Y., Cancelas, J.A., Gu, Y., Jansen, M., et al. (2008). Microenvironment determines lineage fate in a human model of MLL-AF9 leukemia. Cancer Cell 13, 483495. Wong, D.J., Liu, H., Ridky, T.W., Cassarino, D., Segal, E., and Chang, H.Y. (2008). Module map of stem cell genes guides creation of epithelial cancer stem cells. Cell Stem Cell 2, 333344. blom, T., Leary, R.J., Shen, Wood, L.D., Parsons, D.W., Jones, S., Lin, J., Sjo D., Boca, S.M., Barber, T., Ptak, J., et al. (2007). The genomic landscapes of human breast and colorectal cancers. Science 318, 11081113. Yu, F., Yao, H., Zhu, P., Zhang, X., Pan, Q., Gong, C., Huang, Y., Hu, X., Su, F., Lieberman, J., and Song, E. (2007). let-7 regulates self renewal and tumorigenicity of breast cancer cells. Cell 131, 11091123. Zhang, M., Atkinson, R.L., and Rosen, J.M. (2010). Selective targeting of radiation-resistant tumor-initiating cells. Proc. Natl. Acad. Sci. USA 107, 3522 3527. Zhang, W.C., Shyh-Chang, N., Yang, H., Rai, A., Umashankar, S., Ma, S., Soh, B.S., Sun, L.L., Tai, B.C., Nga, M.E., et al. (2012). Glycine decarboxylase activity drives non-small cell lung cancer tumor-initiating cells and tumorigenesis. Cell 148, 259272. Zhao, C., Chen, A., Jamieson, C.H., Fereshteh, M., Abrahamsson, A., Blum, J., Kwon, H.Y., Kim, J., Chute, J.P., Rizzieri, D., et al. (2009). Hedgehog signalling is essential for maintenance of cancer stem cells in myeloid leukaemia. Nature 458, 776779.

728 Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc.

Cell Stem Cell

Perspective
Germline Stem Cells: Origin and Destiny
Ruth Lehmann1,2,*
Hughes Medical Institute The Helen L. and Martin S. Kimmel Center for Stem Cell Biology, Department of Cell Biology, New York University School of Medicine, 540 First Avenue, New York, NY 10016, USA *Correspondence: lehmann@saturn.med.nyu.edu DOI 10.1016/j.stem.2012.05.016
2Skirball Institute, 1Howard

Germline stem cells are key to genome transmission to future generations. Over recent years, there have been numerous insights into the regulatory mechanisms that govern both germ cell specication and the maintenance of the germline in adults. Complex regulatory interactions with both the niche and the environment modulate germline stem cell function. This perspective highlights some examples of this regulation to illustrate the diversity and complexity of the mechanisms involved.
Germ cells have a unique function in the body. They are not needed for survival or immediate physiological function of the individual, but rather, are endowed with the ability to contribute to the next generation. However, while they are ultimately able to generate all cells in the body, germ cells are unipotent and their normal fate is restricted to sperm or egg. This review will focus on recent developments in our understanding of how primordial germ cells (PGCs) are specied during embryogenesis, and how different strategies, including germline stem cell (GSC) renewal, soma-germline interactions, and physiological signals, are used to maintain a pool of differentiating, meiotic germ cells in adults. Germ Cell Specication In most organisms, PGCs are set aside early during embryogenesis. Two distinct mechanisms have been identied in model organisms that specify germ cells. In ies, worms, sh, and frogs, maternally synthesized germ plasm is deposited into the egg during oogenesis. This specialized cytoplasm harbors germlinespecic, electron-dense RNA-protein particles required for multiple aspects of germ cell fate. Embryonic nuclei inheriting germ plasm are destined to become PGCs. In contrast, in mammals and many other species traversing all animal phyla, germ cells are specied among the cells of the embryo independent of preexisting maternal information. Specication occurs in some species early during development, when germ cell fate is induced in a subset of otherwise pluripotent progenitors cells such as epiblast cells of the early mouse embryo, and in other species at later embryonic and postembryonic stages, when germ cells can originate from multipotent progenitors within the mesoderm. Irrespective of these apparent differences in how the germ cell lineage becomes distinct from somatic cells, conserved molecular principles have emerged (Cinalli et al., 2008). Among these, global epigenetic regulation of germline gene expression and germ-cell-specic posttranscriptional regulation are essential both for specifying, maintaining, and protecting the germline during its life cycle and for ensuring transgenerational success. PGC Specication Requires Suppression of the Somatic Program Transcriptional repression seems to be a necessary component of germline specication (Nakamura and Seydoux, 2008). Depending on the organism this repression is achieved either by globally repressing RNA Polymerase activity or by more specically blocking the somatic program. In C. elegans and Drosophila, PGCs are the rst cells separated from the somatic lineage. In these species, PGCs inherit maternally synthesized germ plasm that is rich in mRNA and RNA binding proteins (RBPs). The newly formed PGCs rely entirely on these maternally provided factors, since transcription is completely blocked, thereby preventing the expression of the zygotic genome. In Drosophila, repression of transcription is achieved by germcell-specic translation of a 71 amino acid peptide encoded by the polar granule component (pgc) gene (Hanyu-Nakamura et al., 2008). Pgc protein inhibits the recruitment of the Positive Transcription Elongation factor-b (P-TEFb) kinase complex to transcription sites. P-TEFb phosphorylates Ser-2-containing motifs in the carboxyl-terminal domain (CTD) repeats of RNA Polymerase II (RNA PolII) and is responsible for transcriptional elongation. Pgc protein is only present in germ cells for a short time and transcriptional activity is observed as soon as germ cells initiate their migration toward the somatic part of the embryo. Robust, high levels of transcription are achieved once germ cells reach the somatic gonad. In C. elegans, transcription in the early germline is also repressed via interference with RNA PolII activity, but by slightly different mechanisms. During the rst two embryonic divisions, the Zn-nger proteins Oma1 and Oma2 sequester TATA-binding protein associated factor 4 (TAF-4), a subunit of TFIID required for assembly of the RNA PolII preinitiation complex, in the cytoplasm, thereby preventing transcription in the entire early embryo (Guven-Ozkan et al., 2008). Subsequently, Pie-1, a CCCH Zn-nger protein, is specically translated in the germ cell precursors and inhibits transcriptional initiation and elongation by interfering with Ser 5 and Ser 2 phosphorylation, respectively, in RNA PolII CTD motifs (Ghosh and Seydoux, 2008). In the mouse, germ cell specication occurs in the proximal epiblast just prior to gastrulation. A BMP4/8 signal from the extraembryonic ectoderm initiates the germ-cell-specic program by activating the repressor protein Blimp1 (also known as Prdm1) (Lawson et al., 1999; Ohinata et al., 2005; Ying et al., 2001). Not all cells that receive the BMP signal become PGCs. Expression is restricted to the posterior epiblast by antagonizing signals from the anterior visceral endoderm and is enhanced by Lin28, which counteracts let-7 miRNA, a Blimp1 suppressor (Ohinata et al., 2009; West et al., 2009). Blimp1 is responsible
Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc. 729

Cell Stem Cell

Perspective
for the repression of almost all of the genes that are downregulated in PGCs, but is only required for the activation of some of the genes that are upregulated in germ cells (Kurimoto et al., 2008). How expression of Stella, Kit, Dnd1/Ter, and Nanos3, which play an important part in the specication, migration, and survival of early germ cells, is activated in germ cells is not fully understood. Beyond repression of the somatic program, Blimp1, together with another PR domain protein, Prdm14, also promotes the epigenetic reprogramming within early germ cells to regain pluripotency (Yamaji et al., 2008). Indeed, PGCs reactivate the pluripotency genes Nanog, Sox2, and Oct4. This nding together with the observation that early PGCs can easily give rise to pluripotent embryonic stem cells (called EG cells) led to the hypothesis that PGCs resemble earlier epiblast cells (Matsui et al., 1992). Indeed, recent evidence using activation of a Blimp1 reporter argues that ESC cultures may originate from PGCs (Chu et al., 2011). However, because Blimp1 activity is not required for ESC derivation, the general implication of these ndings remains uncertain. Once PGCs complete their migration to the genital ridge and associate with the somatic part of the gonad, they lose the ability to easily revert to the pluripotent state and reach a germ-cell-specic epigenetic program (Hayashi and Surani, 2009). In an effort to reveal the principles of germ cell specication, Hayashi et al. (2011) recently developed a protocol to derive germ cells from ESCs and iPSCs. In a stepwise process, pluripotent stem cells were rst cultured into cells resembling pregastrulating epiblast cells, from which in turn PGC-like cells were derived. Cells with PGC-like characteristics were able to produce functional sperm after being injected into the seminiferous tubules of germ-cell-deprived (kit mutant) mice. The efciency was similar to that of transplantation of in-vivo-derived PGCs. These new culture conditions will be critical for developing a more complete understanding of germ cell development and have a clear application for reproductive medicine. A Special Chromatin Program Is Established in Germ Cells Early embryonic germ cells in the mouse are transcriptionally active and express the CTD phospho-epitopes characteristic of active RNA PolII transcription. These marks are no longer detected during the migration of germ cells to the somatic gonad (between E8 and E9) (Seki et al., 2007). In the gonad, epigenetic reprogramming takes place, including global DNA demethylation, exchange of histone variants, large-scale chromatin remodeling of retrotransposon-linked and imprinted genes, and reactivation of both X chromosomes (Hajkova et al., 2008). This remodeling is critical for resetting imprint marks and has also been proposed to play an important role in the activation of genes required for early embryonic development in the next generation (Hayashi and Surani, 2009). For a long time it was unclear whether DNA demethylation was achieved passively as a consequence of replication and lack of maintenance methylation, or whether there were enzymes that actively removed methylation marks from DNA. Recent ndings show that the Tet (ten eleven translocation) family of proteins can catalyze the conversion of the 5-methylcytosine (5mC) base (the methylation mark mostly associated with inactive genes) to 5-hydroxymethylcytosine (5hmC), suggesting that this
730 Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc.

conversion could be the rst step toward demethylation (He et al., 2011; Ito et al., 2010, 2011; Tahiliani et al., 2009). Subsequent studies have shown that 5hmC marks are enriched in ESCs and are found in the promoter regions of totipotency genes that are also active in germ cells (Ficz et al., 2011). The next steps toward complete demethylation are still unclear, and may involve Tet-mediated oxidation, base excision repair, and/or passive, replication-dependent demethylation (Hackett et al., 2012). DNA modications and their regulation play a less prominent role in y and worm germ cells, but modication of chromatinassociated histones has important functions in maintaining the germ cell fate in later stages of germline development. In particular, several recent results argue that global chromatin regulation has a direct impact on the germline-soma fate decision. For example, mutations in the synthetic multivulva (synMuv) genes, the worm homologs of the mammalian proteins retinoblastoma (RB), E2F, HP1, and NuRD complex, lead to transformation of intestinal cells into germline-like cells (Petrella et al., 2011). This complex interacts with l(3)mbt, a gene that causes malignant brain tumors when mutated in the y (Lewis et al., 2004). Interestingly, and in contrast to mutations in other brain tumor genes, l(3)mbt mutant tumor cells express genes normally restricted to the germline (Janic et al., 2010). Chromatin association studies indicate that l(3) MBT protein binds preferentially to insulator regions and suggest a role for l(3)MBT in global repression of germline genes in specic somatic tissues (Richter et al., 2011). Chromatin modiers can also coax the germline into expressing somatic programs. Expression of specic neural transcription factors converts C. elegans germline cells into specic types of terminally differentiated neurons. This conversion requires the removal of the conserved chromatin remodeling protein LIN-53, another component of the synMuv family (Tursun et al., 2011). Curiously, germline-to-soma and soma-to-germline transformations involve mutations in the same class of genes normally associated with a repressive chromatin state. However, transformation of germline into soma requires tissue-specic transcription factors, while such factors have not been identied for soma-to-germline transformations or for normal germline development. Together these ndings point to the important role of global transcriptional and epigenetic regulation in the germline. Indeed, evidence is mounting that epigenetic memory in the germline contributes not only to germline development but also to patterns of inheritance for generations to come (Greer et al., 2011; Katz et al., 2009). Conserved RNA Regulators Are Associated with Germ Cells throughout the Life Cycle While so far no specic, conserved transcriptional regulator for germ cell fate has been identied, RBPs play important roles throughout germ cell development. Germline RBPs prevent somatic differentiation and control many germ cell functions such as specication of germ cell fate and sexual identity, proliferation, survival, migration, and regulation of transposable element activity. RBP families such as Vasa, Nanos, Pumilio, Dnd1, DAZL, PIWI, and Tudor are broadly conserved from planarians to humans and are generally found during multiple stages of the germ cell life cycle (Juliano et al., 2010). These

Cell Stem Cell

Perspective
Figure 1. Drosophila Ovarian Niche
(A) Drawing of the ovarian GSC compartment. Each ovary is subdivided into 1216 ovarioles that generate chains of maturing egg chambers. The niche located at the tip of each ovariole contains three different somatic cell types (terminal lament cells, cap cells, and escort or inner germarial sheath [IGS] cells) and two or three GSCs. Asymmetric division of a GSC perpendicular to the niche (arrow) generates a new GSC and a differentiating daughter that exits the niche. The differentiating daughter, called the cystoblast, subsequently undergoes four rounds of division. Each division is incomplete, generating two, four, eight, and sixteen interconnected germline cysts. One cell within the 16-cell cyst differentiates into the oocyte and undergoes meiosis, while the other cells become polyploid nurse cells that feed into the growing oocyte and eventually die. The spectrosome and fusome, germline organelles without membrane that contain alphaspectrin and adducin, play a role in the orientation of the spindle in the GSC and the regulation of cyst division and oocyte determination in the female. The spectrosome has a round appearance and is found in GSCs and cystoblasts. The fusome extends between dividing cysts. (B) Summary of signaling pathways. Beige: terminal lament. Red: cells of the niche, cap cells, and escort/ IGS cells. Blue: GSCs. Green: differentiating germ cells. Yellow: site of adhesion between GSC and niche. TL, regulation by translation; TS, regulation by transcription; S, regulation of stability. Arrow depicts cell-to-cell signaling. In addition to transcriptional regulation via the DPP/GBB receptors, RNA regulators such as Nanos and Pumilio as well as the miRNA machinery act within germ cells and are required for the maintenance of GSCs, likely by repressing RNA translation and stability of germline differentiation genes (Gilboa and Lehmann, 2004; Park et al., 2007; Wang and Lin, 2004). Recent ndings propose that in GSCs, Mei-P26 interacts with Nanos and the miRNA machinery to repress RNAs, such as brat, required for differentiation, while in the cystoblast Bam/Bgcn inhibit nanos RNA, thereby favoring Pumilio/Brat interactions, which repress RNAs required for self-renewal and proliferation, such as the BMP-mediated transcription factor Mad and the proliferation factor dMyc (Harris et al., 2011); (Li et al., 2012). Mei-P26 is also required for differentiation and its translation may depend on Vasa (Liu et al., 2009). JAK/STAT signaling is required for GSC maintenance; the ligand Unpaired (UPD) expressed in terminal pez-Onieva et al., 2008). lament cells activates the Domeless receptor in cap cells, where it activates transcription of dpp (Lo

proteins associate in large RNA-protein complexes referred to as P or polar granules, nuage, mitochondrial cloud, and Balbani body, which have long been recognized as the morphological hallmark of germ cells (Voronina et al., 2011). Binding of these RBPs to their target RNAs regulates RNA stability and translation and possibly transcription (Richter and Lasko, 2011). It is particularly intriguing that several of these RBPs are present and employed at multiple stages during the germline life cycle. For example, Vasa protein is a highly conserved, ATP-dependent helicase with multiple germline functions including germ plasm assembly, germ cell proliferation and differentiation, and smallRNA-mediated transposable element silencing (Lasko and Ashburner, 1990; Liu et al., 2009; Pek and Kai, 2011; Tomancak et al., 1998; Vagin et al., 2004). Another example is the conserved Zn-nger protein Nanos, with a well-dened role in RNA translational repression (Kadyrova et al., 2007; Suzuki et al., 2010). In ies, worm, and mouse, Nanos plays essential roles in specifying embryonic germ cells and in the maintenance of adult GSCs (Forbes and Lehmann, 1998; Kobayashi et al., 1996; Kraemer et al., 1999; Sada et al., 2009; Saga, 2010; Subramaniam and Seydoux, 1999). Combinatorial interactions between RBPs likely regulate the afnity and selectivity for cognate RNAs during germline development. This is best illustrated by the interplay between Nanos, the sequence-specic RBP Pumilio, the TRIM-NHL domain proteins Brat and Mei-P26, and the miRNA machinery, whose changing associations trigger stage-specic translational repression of specic RNA targets during embryogenesis and GSC differentiation in Drosophila (for more detail, please refer to Figure 1) (Harris et al., 2011; Li et al., 2012; Sonoda and Wharton, 2001).

Several RPBs have recently been implicated in the control of small RNA pathways. Piwi, Tudor, and Argonaute family proteins are part of the piwi interacting (pi) small RNA pathway and control transposable element abundance specically in the germline via the production of small RNAs that target transposable elements for destruction (Khurana and Theurkauf, 2010; Malone and Hannon, 2009). Dazl and DND1 in zebrash counteract the effects of miRNA, thereby preventing the degradation of target RNAs in the germline (Kedde et al., 2007). The interplay between RBPs and small RNA pathways provides a mechanism to ne-tune gene expression largely independently from instructive gene transcription. This function may be particularly relevant in the germline to prevent somatic differentiation and to protect the germline genome throughout the life cycle. GSCs Analysis of GSC development has greatly inuenced the study of stem cell biology in general and has informed our knowledge of human GSC behavior. Because of the early recognition that adult gonads contain a self-renewing stem cell population, much is known about the physical nature of the stem cell compartments and the regulatory networks that keep the balance between self-renewal and differentiation. Here, I will summarize the general features of GSC microenvironments in the gonads of some of the best-studied model organisms, and I will focus on recent results in building regulator GSC networks and observing their response to specic physiological conditions, such as nutritional state, environment, and aging.
Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc. 731

Cell Stem Cell

Perspective
Figure 2. Drosophila Testis Niche
(A) Drawing of the testis GSC compartment. Six to twelve GSCs form a rosette around a single cluster of somatic support cells called the hub. Division is oriented (arrow) by the centrosomes (stars). The mother centrosome remains in the GSC (black star). The hub also maintains somatic cyst cell stem cells (CySCs). Division of the CySCs and GSCs is coordinated such that two cyst cells surround each differentiating germ cell. The differentiating germ cell, called a gonioblast, undergoes four rounds of division, before each spermatocyte of the interconnected 16-cell cyst undergoes meiosis. GSCs predominately divide asymmetrically but can also divide symmetrically to replenish the stem cell pool (de Cuevas and Matunis, 2011). The round spectrosome is found in GSC and gonioblast, while the fusome extends equally between dividing cysts. (B) Summary of signaling pathways. Beige: hub cells. Red: cyst stem cells and cyst cells. Blue: GSCs. Green: differentiating germ cells. Yellow: site of adhesion between GSC and niche. TL, regulation by translation; TS, regulation by transcription; pStat, phosphorylated active Stat. The HOW RNA binding protein binds to bam RNA and represses its translation in stem cells. Ectopic HOW delays BAM protein accumulation and leads to additional spermatogonial divisions (Monk et al., 2010). Once a critical level of BAM is reached spermatogonia cease proliferation and differentiate (Insco et al., 2009).

GSCs Rely on a Specialized Somatic Microenvironment Germ cells are only part of the gonad. Throughout development and during differentiation into gametes, germ cells rely on interactions with specialized somatic cells. For GSCs these interactions take place in specialized somatic gonadal compartment referred to as the stem cell niche. GSCs have been identied in both sexes in C. elegans and Drosophila and in the mouse testis. While some studies suggest that a dormant GSC population exists in the female mammalian germline, lineage-tracing experiments have not yet been performed in vivo and it remains possible that these cells represent remnants of an earlier PGC or gonocyte stage (White et al., 2012). Drosophila Thanks to comparatively simple morphology and ease of genetic manipulation, the Drosophila ovary and testis have become model tissues for the study of stem cell behavior. GSCs are found in both male and female gonads where they maintain egg and sperm production throughout adult life (for detailed description of the morphology of the Drosophila ovary and testis, please consult Figure 1A and Figure 2A). In the female, each ovary contains multiple niches. Each niche is composed of at least three different somatic cell types (terminal lament cells, cap cells, and inner germarial sheath or escort cells), which interact with two or three GSCs and their progeny. Asymmetric division of the GSC generates a new GSC and a differentiating daughter that exits the niches inuence. The differentiating daughter, also called the cystoblast, subsequently undergoes four rounds of division. Each division is incomplete, generating a 16-celled interconnected germline cyst, which will generate the oocyte. In the Drosophila male, each testis contains 612 GSCs that form a rosette around a single cluster of somatic support cells called the hub. The hub also maintains somatic cyst cell stem cells (CySCs). Division of the CySC and GSC is coordinated such that two cyst cells surround each differentiating germ cell. The differentiating germ cell, called a gonioblast, undergoes
732 Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc.

four rounds of division before each spermatocyte of the interconnected 16-cell cyst undergoes meiosis, generating 64 sperm. Asymmetric division of the GSC can be directly linked to spindle orientation, with one centrosome being anchored to the niche via astral microtubules that interact with the adherens junction complex connecting the GSC and the hub. Failure to properly align the mitotic spindle causes a delay in GSC division, suggesting the existence of a checkpoint that monitors centrosome orientation prior to mitosis (Cheng et al., 2008). Furthermore, individual labeling of dividing centrosomes revealed a conservative mode of centrosome division in males, such that the maternal centrosome remains with the stem cell (Yamashita et al., 2007). This mode of inheritance is not followed by the dividing chromosome strands, which, with the possible exception of the Y chromosome, seem to randomly segregate into the daughter cell (Yadlapalli et al., 2011). In both males and females, GSCs predominately divide asymmetrically but can also divide symmetrically to replenish the stem cell pool (de Cuevas and Matunis, 2011). C. elegans Depending on the nutritional state, about 200 mitotically active, undifferentiated germ cells reside in each of the distal ends of the two gonad arms of the hermaphrodite C. elegans. At the end of each arm, one somatic cell, the distal tip cell (DTC), envelops the gonad tip and serves as a niche (Figure 3A). As cells move away from the DTC, they cease division and enter meiosis. Asymmetric divisions have not been observed, but cells closest to the DTC form a pool of undifferentiated GSCs, while cells further away initiate the transition to meiosis. Long extensions of the DTC encompass the mitotically active stem cell compartment (Kimble and Crittenden, 2007). Mammalian Testis In contrast to Drosophila and C. elegans, where specialized somatic cells form a structured niche surrounding the stem cell compartment, the entire process, starting from self-renewing

Cell Stem Cell

Perspective
Figure 3. Distal Tip Cell and Stem Cell Compartment of C. elegans Hermaphrodite Gonad
(A) Drawing of GSC compartment of one gonad arm. Mitotically active, undifferentiated germ cells reside in each of the distal ends of the two gonad arms. The distal tip cell (DTC) envelops the gonad tip and serves as a niche. As cells move away from the DTC they cease division and enter meiosis. (B) Summary signaling pathways. Beige: ASI neuron. Red: distal tip cell. Blue: GSCs. Green: differentiating germ cells. TL, regulation by translation. Arrow depicts cell-tocell signaling. The double pointed arrow indicates the random direction of GSC division.

spermatogonia to release of elongated sperm into the lumen of the seminiferous tubule, occurs within the environment of large, somatic Sertoli cells (Figure 4A). As the cell bodies of these cells extend from the stem cell to the spermatocyst compartment, it is unclear how they could provide the type of microenvironment that denes a stem cell niche. However, tight junctions separate the basal compartment, which contains early spermiogenic stages, from the adluminal compartment lled with later stages and could thereby provide a structural separation of function (Oatley and Brinster, 2008). The existence of stem cells in the mouse testis is well documented by cell-lineage and transplantation reconstitution assays (Brinster and Zimmermann, 1994). Only 0.02%0.03% of the germ cells in a total testis have reconstitution potential upon transplantation into a testis devoid of germ cells. This percentage can be signicantly increased if the germ cells are enriched for undifferentiated spermatogonia (Oatley and Brinster, 2008). Traditionally it was thought that among the undifferentiated spermatogonia, the single-celled (As) spermatogonia constitute the self-renewing population, and that differentiation proceeds through a hierarchy of four mitotic divisions producing sequentially paired (Apr) and Aaligned4-Aal16 spermatogonia connected by intercellular bridges. In this scenario, after the fourth division the Aal16 spermatogonia would mature into A1 spermatogonia, which enter the differentiation path leading to meiosis. In recent years, this hierarchical description of early spermatogenesis has been revised. Lineage marking and live imaging revealed that Apr-Aal16 spermatogonia can resolve their connecting bridges and thereby contribute to the self-renewing stem cell population. Furthermore, these experiments suggested that As-A8 spermatogonia are also capable of directly entering differentiation without completing all rounds of division (Nakagawa et al., 2007, 2010). Taken together, these experiments revealed an apparent heterogeneity among the undifferentiated spermatogonia independent of the mitotic hierarchy and suggest that As to Aal16 spermatogonia contribute to the spermatogonial progenitor cell (SGP) pool. It remains unclear

how the self-renewal versus differentiation decision is achieved for an individual As-A16 spermatogonium. SGPs are found along the basement membrane of the seminiferous tubule and are preferentially associated with vasculature enriched for Leydig and interstitial cells (Yoshida et al., 2007). As spermatogonia undergo differentiation they leave the basal region, suggesting that a vascular niche protects SGPs from differentiation. SGPs are highly motile, so one possibility is that uctuations in the interactions between the individual spermatogonia and their microenvironment create heterogeneity. Regulatory Networks Control the Balance between Self-Renewal, Differentiation, and Regeneration After the initial discovery of specic factors that promote either stem cell proliferation and maintenance or differentiation, it has become apparent that there are many levels of regulation. In most stem cell systems, one major GSC signaling pathway has been identied that relies on a signal provided by the niche and received by the germ cells. Mosaic analysis, tissue-specic gene expression assays, and transplantation experiments have determined the tissue dependence of the respective factors. Finally, knockout and overexpression or ectopic expression experiments have been used to examine whether a particular signaling pathway plays an instructive role. The major signaling pathways identied include BMP, JAK/STAT, Notch, and GDNF. More detailed analysis of these and additional signaling pathways has provided insight into intricate regulatory networks. As an increasing amount of information emerges, it is becoming clearer how a balance between self-renewal and differentiation is achieved at the molecular level. Because of the sheer complexity of the present networks and at times incomplete connections, I will focus on specic examples highlighting recent progress in our understanding. For a more detailed discussion of specic germline systems, please refer to recent reviews (Kimble, 2011; Oatley and Brinster, 2008; Spradling et al., 2011). Multiple Levels of Control Restrict the Availability of the Self-Renewal Signal and Limit Tissue Response in the Drosophila Ovary In the Drosophila ovary, recent studies have led to a detailed understanding of how the inuence of the GSC signal is restricted within the somatic niche to prevent ectopic production of GSCs and how receptor activity in the GSCs is regulated to allow differentiation (Figure 1B). GSCs in the ovary depend on
Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc. 733

Cell Stem Cell

Perspective
Figure 4. Mouse Spermatogonial Niche
(A) Drawing of GSC compartment as a quadrant of cross sections of a seminiferous tubule. The entire spermiogenesis process, starting from self-renewing spermatogonia to release of elongated sperm into the lumen of the seminiferous tubule, occurs within the environment of large, somatic Sertoli cells. Tight junctions separate the basal compartment, which contains early spermiogenic stages, from the adluminal compartment lled with later stages. (B) Summary of signaling pathways. Beige: blood vessel and Leydig cells, which may contribute additional factors regulating stem cell maintenance, including Colony stimulating factor 1 (CSF-1) and FGF. Red: Sertoli cells. Note that tight junctions form the blood-testis barrier between the Basal (BC) and Adluminal (AL) compartment, which separates differentiating spermatocytes from undifferentiating SGPs. Blue: SGPs. Green: differentiating germ cells. All depicted interactions are transcriptional. RA: retinoic acid activates KIT receptor during the transition to differentiation. SGPs express high levels of the GDNF receptor GFRa1, the translational repressor Nanos2, and the transcriptional repressor PLZF. Increased levels of the bHLH transcription factors Ngn3 and Nanos3 characterize SGPs that are more likely to differentiate. Etv5 expressed in Sertoli cells leads to production of CCL9 ligand, which is recognized by CCR1 receptor in germ cells and attracts SGPs to Sertoli cells (double headed arrow) (Simon et al., 2010).

the production of the BMP-type ligands DPP (Decapentaplegic) and GBB (Glass bottom boat), which are secreted from niche cells (Chen et al., 2011; Xie and Spradling, 1998). The DPP/ GBB signal is received in germ cells by the Thickveins (TKV) and Saxophone (SAX) receptors. Receptor activation leads to transcriptional repression of the bag of marbles (bam) gene in the GSCs. As a consequence, bam expression is restricted to the GSC daughter, the cystoblast, and its progeny (Song et al., 2004). Bam is an instructive factor for germ cell differentiation: loss of bam causes accumulation of undifferentiated germ cells and ectopic expression of bam causes GSC differentiation and germline depletion (Ohlstein and McKearin, 1997). Multiple levels of control restrict the availability of the BMP signal and the activity of its receptor to the stem cell compartments (Chen et al., 2011). Niche cells produce the BMP ligand DPP. Its diffusion is limited by the heparin sulfate proteoglycan DALLY, ensuring that a high level of BMP remains close to the cap cells that are located directly adjacent to the GSCs (Guo and Wang, 2009). dally expression itself is negatively regulated by the activity of the Drosophila EGF receptor (EGFR) in the somatic escort cells, which intermingle and send ne extensions between the dividing, differentiating germ cells (Liu et al., 2010). Differentiating germ cells express and process the ligand for the EGFR, thereby providing a negative feedback loop to limit the GSC signal. In addition to Dally, Collagen IV directly binds to and antagonizes DPP, which further restricts its reach (Wang et al., 2008). Germ-cell-intrinsic mechanisms control the activity of the receptor by regulating its expression within GSCs and promoting its degradation by a complex of the E3 ubiquitin ligase Smurf and the serine-threonine kinase Fused (Xia et al., 2010). This regulation, together with posttranscriptional repression of the downstream transcription factor Mad by the RNA regulator complex Pumilio/Brat in cystoblasts, ensures that the GSC signal is only transmitted to a small group of GSCs (Harris
734 Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc.

et al., 2011). (For additional examples of regulation within the stem cell compartment, please refer to Figure 1B.) Experimental Models Distinguish between Germ-Cell-Autonomous Functions and Intercellular Competition in the Drosophila Testis In the male gonad, genetic analysis had pointed to a role for the JAK/STAT pathway in GSC self-renewal (Kiger et al., 2001; Tulina and Matunis, 2001). Recent studies revealed, however, that this function is dispensable as long as STAT levels are kept uniform among GSCs. However, in genetic mosaics of wild-type and stat mutant GSCs, cells with higher STAT levels will outcompete mutant cells and take over the niche (Leatherman and Dinardo, 2010). Uniform loss of JAK/STAT signaling in GSCs causes a weak GSC maintenance phenotype due to reduced adhesion between the GSCs and the hub. In contrast to this apparent minor role in GSCs, the JAK/STAT pathway plays an indispensable and instructive role in maintenance of the CySCs (Issigonis et al., 2009). It does so through activation of two downstream transcription factors, ZFHI and CHINMO in CySCs, which are required for CySC self-renewal. Conversely, overexpression of either ZFHI or CHINMO, or constitutive activation of JAK, leads to a CySC expansion. Activation of the JAK/ STAT pathway in the somatic cells controls GSC maintenance, and overexpression of ZFHI and CHINMO causes accumulation of undifferentiated germ cells (Flaherty et al., 2010; Leatherman and Dinardo, 2008). How the somatic cells signal back to the GSCs remains unclear, but could be directly by controlling the expression, production, or secretion of the BMP ligands GBB and DPP, which have been implicated in GSC maintenance in the testis, or indirectly by preventing GSC differentiation (Shivdasani and Ingham, 2003). These new results point to an interesting experimental dilemma in the analysis of stem cell behavior: most signaling pathways are used at multiple times during development and in multiple tissues. Thus, analysis of

Cell Stem Cell

Perspective
mutant phenotypes caused by the loss of a particular factor is often limited to conditional knockout approaches. These approaches in general produce mosaics of wild-type and mutant cells within the same tissue. Such mosaics do not allow one to distinguish between a bona de, tissue-specic requirement for a gene product and a more indirect effect due to the competition between cells with different levels of the respective factor. Transplantation assays as used in the mammalian testis; tissuespecic complementation of the gene in question in the soma, but not the germline; or germline-specic RNAi knockout can be utilized to create genetically uniform cell populations. An RNA Regulatory Network Maintains GSCs and Promotes the Mitosis-to-Meiosis Decision in C. elegans The mitosis-to-meiosis decision in C. elegans is dependent on signals produced from the DTC (Figure 3A). The DTC produces the Notch ligand, and activation of the Notch receptor, GLP-1, in the germline is sufcient to maintain a proliferating stem cell population, while loss of GLP-1 activity leads to entry into meiosis (Cinquin et al., 2010). The regulatory network downstream of Notch/Glp-1 is one of the best-understood models of how the balance between stem cell self-renewal and proliferation versus differentiation and meiosis is established and maintained (Figure 3B). Central to this network is the crosstalk between several RNA regulatory factors. In the stem cell compartment, Notch transcriptionally activates fbf-1 and fbf-2 (Lamont et al., 2004). These homologs of the conserved RBP Pumilio repress translation of another set of RNA regulators, GLD1, GLD2, and GLD3, in the stem cell compartment. GLD1, a member of the STAR family of RBPs, in turn inhibits stem cell maintenance and proliferation by repressing Notch/Glp-1 and Cyclin E translation, respectively (Biedermann et al., 2009; Marin and Evans, 2003). GLD2 and GLD3 are part of a cytoplasmic polyA-polymerase translational activator complex that promotes translation of gld-1 RNA, thereby reinforcing the meiosis decision (Suh et al., 2006). The regulatory pathway downstream of Notch demonstrates the important role that RNA regulatory networks play in germ cell development. Indeed, reporter studies have revealed that independent of a specic transcriptional input, 30 UTRs are the primary determinants regulating spatial gene expression in the C. elegans germline (Merritt et al., 2008). Opposing Functions of Transcription Factors Dene the Germline Progenitor Pool in the Mammalian Testis The molecular mechanisms underlying self-renewal and differentiation of SGPs are beginning to be understood due to the establishment of culture conditions to grow, select, and increase the SGP population, as well as to tissue-specic knockout strategies. A critical factor in the maintenance of a self-renewing SGP population is the TGFb-related growth factor GDNF, which is produced by Sertoli cells (Figure 4B). GDNF supports maintenance of a stem cell population in vitro and in vivo, and overexpression of GDNF in Sertoli cells leads to an increase in As and Apr spermatogonia (Meng et al., 2000). Two GDNF receptors, GFRa1 and c-RET, are required in spermatogonia for stem cell maintenance (Naughton et al., 2006; Oatley et al., 2007). Downstream effectors expressed in SGPs in response to GDNF include a number of transcription factors, such as Oct4, Bcl6b, and Taf4b, but also the RNA regulator NANOS-2, which, upon forced expression, can partially rescue stem cell loss in the absence of GDNF (Sada et al., 2012). When specically removed from the germline, these genes affect SGP proliferation, self-renewal, and survival. One critical transcription factor that is regulated apparently independently of GDNF in SGPs is PLZF. This Zn-nger transcriptional repressor forms a complex with the Spalt-like protein SALL4, and the stoichiometry of SALL4 to PLZF in germ cells inuences the self-renewal-to-differentiation decision (Hobbs et al., 2012). Consistent with the opposing effects of these two genes, SALL4 deletion affects spermatogonial differentiation, while PLZF deletion causes loss of stem cell self-renewal. A high PLZF-to-SALL4 ratio determines self-renewal, at least in part, via expression of SAL1, while a high SALL4-to-PLFZ ratio favors differentiation possibly via kit expression (Filipponi et al., 2007). SALL4 is not only active in SPGs but also in embryonic germ cells, where PLZF levels are low. It has been proposed that SALL4 preserves the undifferentiated state of the germ cell, possibly by acting through the KIT receptor, which is required for PGC survival and proliferation. Later in the adult SALL4 again regulates kit expression, but this time, to promote differentiation (Hobbs et al., 2012). Adhesion between GSCs and Niche Control Asymmetry of GSC Division and Tissue Regenerative Potential Application of live imaging and differential labeling techniques are beginning to reveal intriguing aspects of GSC regulation (Cheng et al., 2008; Morris and Spradling, 2011). One critical aspect is the role of adhesion between GSCs and the niche. In the Drosophila ovary, GSCs are anchored to the cap cells of the niche via gap junctions and E-cadherin-rich adherens junctions. E-cadherin downregulation favors GSC differentiation, while loss of gap junction communication causes the death of differentiating germ cells (Jin et al., 2008; Tazuke et al., 2002). In the male, adherens junctions between the hub and GSCs orient the mitotic spindle and may directly contribute to keeping the mother centrosome with the GSC (Inaba et al., 2010). As in the female, E-cadherin mutants rapidly lose GSCs. Interestingly, a recent study suggests that BMP receptor activation in the male is localized to the adherens junctions. In this model trafcking of the BMP ligand together with adherens junction components provides a short-range source for BMP receptor activation (Michel et al., 2011). Rather than E-cadherin, spermatogonial stem cells in the mouse require beta-1-integrin to home and adhere to their niche, and beta-1-integrin expression in germ cells is necessary for normal spermatogenesis (Kanatsu-Shinohara et al., 2008). Taken together, these studies indicate that adhesion molecules are required to anchor stem cells to their niche and play a role in orienting stem cell divisions. Furthermore, regulation of the strength of adhesion by environmental factors or during aging can inuence the balance between selfrenewal and differentiation and thereby affect the regenerative ability of the tissue (Pan et al., 2007; Yamashita, 2010). Environmental and Systemic Factors Regulate Gonad Homeostasis Cooperation between GSCs and their niches maintains homeostasis by setting the balance between self-renewal and differentiation. This balance has to be exible to adapt to changes in environmental conditions like food availability or to systemic changes caused by hormonal uctuation and aging. During
Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc. 735

Cell Stem Cell

Perspective
Drosophila larval growth, levels of the hormone ecdysone coordinate the development and differentiation of the niche and GSCs in the ovaries. Ecdysone prevents precocious maturation of the niche and PGC differentiation during early larval development, but promotes niche morphogenesis and germ cell differentiation during later larval stages (Gancz et al., 2011). In the adult, niche signals, GSC-to-niche adhesion, GSC proliferation, and apoptosis are affected by both external and internal conditions. Aging is associated with delay or complete failure in GSC self-renewal in both males and females. The effects of aging can be overcome by the expression of niche factors, such as the ligands of the BMP and JAK/STAT pathways, the adhesion molecule E-cadherin, or the cell cycle regulator STRING, a Cdc25 homolog (Inaba et al., 2011; Pan et al., 2007). Interestingly, with aging the number of GSCs decreases less than expected, likely due to replacement of lost stem cells by symmetrical division, leading to clonal expansion of a subset of GSCs (Cheng et al., 2008; Wallenfang et al., 2006). If a similar process occurs in mammals, such a clonal expansion could contribute to the accumulation of genomic defects in the progeny of aged parents. Nutrition has a dramatic effect on egg production in Drosophila. Flies that are fed a protein-rich diet produce up to 15 times more progeny than those on a protein-poor diet (Drummond-Barbosa and Spradling, 2001). In addition to decreased cell death during oogenesis, GSC proliferation rate increases with a high-protein diet. These effects are in part mediated by Insulin and Tor, which regulate niche size, GSC-niche adhesion, and germline cyst differentiation (Hsu and Drummond-Barbosa, 2011; Sun et al., 2010). Environmental factors such as food abundance and population density also inuence the rate of proliferation and the mitosis-to-meiosis decision in C. elegans. Limited food availability restricts the size of the stem cell pool during larval growth, and starvation leads to a dramatic resorption of the germline. Even under these conditions, about 30 stem cells remain, which, if conditions improve, can regenerate the entire germline (Angelo and Van Gilst, 2009). The Insulin/ IGF and Tor/S6K pathways regulate germline proliferation during larval stages (Korta et al., 2012; Michaelson et al., 2010). Favorable conditions are also sensed by a group of neurons located close to the pharynx, triggering activation of the TGFb receptor in the DTC and promoting GSC proliferation (Dalfo et al., 2012). In addition to proliferation, Tor/S6K and TGFb also affect the mitosis-to- meiosis decision. Insulin-, Tor-, and TGFb-pathway mutants exacerbate the glp-1 phenotype, suggesting that these pathways exert their function at least in part independently of the Notch regulatory network. How these sensors and mediators of environmental conditions are integrated at the level of germ cell self-renewal and differentiation is not understood. Environmental stimuli may also regulate the SGP pool in the mouse testis. Recent studies show that PLZF and mTORC1, one of two Tor complexes that mediate cell growth in response to nutrients, growth factors, and cellular stress in mammals, are engaged in a negative feedback loop that can shift the selfrenewal versus differentiation balance depending on environmental conditions. PLZF represses mTORC1 via activation of Redd1, an mTORC1 inhibitor; mTORC1 activation in turn leads to downregulation of the GDNF receptors (Hobbs et al., 2010). Thus, PLZF expression in SPGs may render these cells less
736 Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc.

sensitive to environmental inuence. Furthermore, the ability of mTORC1 to downregulate GDNF receptors could entice SGPs to leave their niche and initiate the differentiation program. The balanced interactions of PLZF with SALL4 and mTORC1 nicely exemplify how small, local changes in the levels of regulators of cell growth and differentiation can contribute to the heterogeneity and potential of the SGP population. Summary and Outlook Systematic analysis of germ cell development in the embryo and stem cell homeostasis in the adult gonad has brought insight into the molecular mechanisms that establish and regulate niche-stem cell interactions. In contrast to all other stem cell systems in the body, homeostasis of the gonadal system is directly related to reproductive success. Thus specic mechanisms such as global epigenetic regulation of the germ cell program and the need for a complete reprogramming of the germ cell genome during development are features of essential importance for germ cells to give rise to a new generation. How transgenerational inheritance is transmitted at the gene and genome level is still largely unclear. The regulatory networks controlling GSC self-renewal and differentiation are beginning to be unraveled. These include germ-cell-specic RNA regulators that are conserved between species. The targets of these regulators are still largely unknown, but in some cases they control strikingly similar processes, such as the role of Nanos family proteins in PGC specication and GSC maintenance in worms, ies, and mice. These RNA regulators, together with their targets, which include small regulatory RNAs, are packaged into intracellular particles that change in composition and cellular location during the germ cell life cycle. Compartmentalization may allow the same RBP to entertain multiple regulatory relationships during different stages of germline development. Niche-stem cell interactions play a crucial role for GSC homeostasis. These interactions protect GSCs from differentiation, regulate proliferation, and, in some cases, orient the GSCdaughter cell division. In Drosophila, the path from GSC to differentiated germline cyst was considered to be rather hierarchical until the discovery that male and female GSCs can regenerate from advanced, interconnected germline cysts (Brawley and Matunis, 2004; Kai and Spradling, 2004; Wong and Jones, 2012). In the mouse testis, heterogeneity among SPGs seems to be the rule rather than the exception, since more advanced cysts routinely contribute to the precursor pool. The functional and hereditary consequences of this plasticity are not understood. Germ cells, through their potential to differentiate into sperm and egg, have the ability to create a new organism. Analysis of the regulatory networks that control germ cell specication, self-renewal, and differentiation may ultimately lead to a better understanding of the control mechanisms that balance the need for genomic delity with the opportunity for evolutionary change.
ACKNOWLEDGMENTS I would like to thank members of my lab and my colleagues in the Developmental Genetics Program for discussion, and specically Drs. Daria Siekhaus

Cell Stem Cell

Perspective
and Ryan Cinalli for critically reading the manuscript. I apologize to those colleagues whose work I was unable to properly cite due to space restrictions. Research on germ cell development in my lab is supported by the NIH (R01HD041900), the HHMI, and the Kimmel Center for Stem Cell Biology. REFERENCES Angelo, G., and Van Gilst, M.R. (2009). Starvation protects germline stem cells and extends reproductive longevity in C. elegans. Science 326, 954958. Biedermann, B., Wright, J., Senften, M., Kalchhauser, I., Sarathy, G., Lee, M.H., and Ciosk, R. (2009). Translational repression of cyclin E prevents precocious mitosis and embryonic gene activation during C. elegans meiosis. Dev. Cell 17, 355364. Brawley, C., and Matunis, E. (2004). Regeneration of male germline stem cells by spermatogonial dedifferentiation in vivo. Science 304, 13311334. Brinster, R.L., and Zimmermann, J.W. (1994). Spermatogenesis following male germ-cell transplantation. Proc. Natl. Acad. Sci. USA 91, 1129811302. Chen, S., Wang, S., and Xie, T. (2011). Restricting self-renewal signals within the stem cell niche: multiple levels of control. Curr. Opin. Genet. Dev. 21, 684689. rkel, N., Hemati, N., Fuller, M.T., Hunt, A.J., and Yamashita, Y.M. Cheng, J., Tu (2008). Centrosome misorientation reduces stem cell division during ageing. Nature 456, 599604. Chu, L.F., Surani, M.A., Jaenisch, R., and Zwaka, T.P. (2011). Blimp1 expression predicts embryonic stem cell development in vitro. Curr. Biol. 21, 1759 1765. Cinalli, R.M., Rangan, P., and Lehmann, R. (2008). Germ cells are forever. Cell 132, 559562. Cinquin, O., Crittenden, S.L., Morgan, D.E., and Kimble, J. (2010). Progression from a stem cell-like state to early differentiation in the C. elegans germ line. Proc. Natl. Acad. Sci. USA 107, 20482053. Dalfo, D., Michaelson, D., and Hubbard, E.J. (2012). Sensory Regulation of the C. elegans Germline through TGF-beta-Dependent Signaling in the Niche. Curr. Biol. 22, 712719. de Cuevas, M., and Matunis, E.L. (2011). The stem cell niche: lessons from the Drosophila testis. Development 138, 28612869. Drummond-Barbosa, D., and Spradling, A.C. (2001). Stem cells and their progeny respond to nutritional changes during Drosophila oogenesis. Dev. Biol. 231, 265278. Ficz, G., Branco, M.R., Seisenberger, S., Santos, F., Krueger, F., Hore, T.A., Marques, C.J., Andrews, S., and Reik, W. (2011). Dynamic regulation of 5-hydroxymethylcytosine in mouse ES cells and during differentiation. Nature 473, 398402. Filipponi, D., Hobbs, R.M., Ottolenghi, S., Rossi, P., Jannini, E.A., Pandol, P.P., and Dolci, S. (2007). Repression of kit expression by Plzf in germ cells. Mol. Cell. Biol. 27, 67706781. Flaherty, M.S., Salis, P., Evans, C.J., Ekas, L.A., Marouf, A., Zavadil, J., Banerjee, U., and Bach, E.A. (2010). chinmo is a functional effector of the JAK/STAT pathway that regulates eye development, tumor formation, and stem cell selfrenewal in Drosophila. Dev. Cell 18, 556568. Forbes, A., and Lehmann, R. (1998). Nanos and Pumilio have critical roles in the development and function of Drosophila germline stem cells. Development 125, 679690. Gancz, D., Lengil, T., and Gilboa, L. (2011). Coordinated regulation of niche and stem cell precursors by hormonal signaling. PLoS Biol. 9, e1001202. Ghosh, D., and Seydoux, G. (2008). Inhibition of transcription by the Caenorhabditis elegans germline protein PIE-1: genetic evidence for distinct mechanisms targeting initiation and elongation. Genetics 178, 235243. Gilboa, L., and Lehmann, R. (2004). Repression of primordial germ cell differentiation parallels germ line stem cell maintenance. Curr. Biol. 14, 981986. Greer, E.L., Maures, T.J., Ucar, D., Hauswirth, A.G., Mancini, E., Lim, J.P., Benayoun, B.A., Shi, Y., and Brunet, A. (2011). Transgenerational epigenetic inheritance of longevity in Caenorhabditis elegans. Nature 479, 365371. Guo, Z., and Wang, Z. (2009). The glypican Dally is required in the niche for the maintenance of germline stem cells and short-range BMP signaling in the Drosophila ovary. Development 136, 36273635. Guven-Ozkan, T., Nishi, Y., Robertson, S.M., and Lin, R. (2008). Global transcriptional repression in C. elegans germline precursors by regulated sequestration of TAF-4. Cell 135, 149160. Hackett, J.A., Zylicz, J.J., and Surani, M.A. (2012). Parallel mechanisms of epigenetic reprogramming in the germline. Trends Genet. 28, 164174. Hajkova, P., Ancelin, K., Waldmann, T., Lacoste, N., Lange, U.C., Cesari, F., Lee, C., Almouzni, G., Schneider, R., and Surani, M.A. (2008). Chromatin dynamics during epigenetic reprogramming in the mouse germ line. Nature 452, 877881. Hanyu-Nakamura, K., Sonobe-Nojima, H., Tanigawa, A., Lasko, P., and Nakamura, A. (2008). Drosophila Pgc protein inhibits P-TEFb recruitment to chromatin in primordial germ cells. Nature 451, 730733. Harris, R.E., Pargett, M., Sutcliffe, C., Umulis, D., and Ashe, H.L. (2011). Brat promotes stem cell differentiation via control of a bistable switch that restricts BMP signaling. Dev. Cell 20, 7283. Hayashi, K., and Surani, M.A. (2009). Resetting the epigenome beyond pluripotency in the germline. Cell Stem Cell 4, 493498. Hayashi, K., Ohta, H., Kurimoto, K., Aramaki, S., and Saitou, M. (2011). Reconstitution of the mouse germ cell specication pathway in culture by pluripotent stem cells. Cell 146, 519532. He, Y.F., Li, B.Z., Li, Z., Liu, P., Wang, Y., Tang, Q., Ding, J., Jia, Y., Chen, Z., Li, L., et al. (2011). Tet-mediated formation of 5-carboxylcytosine and its excision by TDG in mammalian DNA. Science 333, 13031307. Hobbs, R.M., Seandel, M., Falciatori, I., Rai, S., and Pandol, P.P. (2010). Plzf regulates germline progenitor self-renewal by opposing mTORC1. Cell 142, 468479. Hobbs, R.M., Fagoonee, S., Papa, A., Webster, K., Altruda, F., Nishinakamura, R., Chai, L., and Pandol, P.P. (2012). Functional antagonism between Sall4 and Plzf denes germline progenitors. Cell Stem Cell 10, 284298. Hsu, H.J., and Drummond-Barbosa, D. (2011). Insulin signals control the competence of the Drosophila female germline stem cell niche to respond to Notch ligands. Dev. Biol. 350, 290300. Inaba, M., Yuan, H., Salzmann, V., Fuller, M.T., and Yamashita, Y.M. (2010). E-cadherin is required for centrosome and spindle orientation in Drosophila male germline stem cells. PLoS ONE 5, e12473. Inaba, M., Yuan, H., and Yamashita, Y.M. (2011). String (Cdc25) regulates stem cell maintenance, proliferation and aging in Drosophila testis. Development 138, 50795086. Insco, M.L., Leon, A., Tam, C.H., McKearin, D.M., and Fuller, M.T. (2009). Accumulation of a differentiation regulator species transit amplifying division number in an adult stem cell lineage. Proc. Natl. Acad. Sci. USA 106, 22311 22316. Issigonis, M., Tulina, N., de Cuevas, M., Brawley, C., Sandler, L., and Matunis, E. (2009). JAK-STAT signal inhibition regulates competition in the Drosophila testis stem cell niche. Science 326, 153156. Ito, S., DAlessio, A.C., Taranova, O.V., Hong, K., Sowers, L.C., and Zhang, Y. (2010). Role of Tet proteins in 5mC to 5hmC conversion, ES-cell self-renewal and inner cell mass specication. Nature 466, 11291133. Ito, S., Shen, L., Dai, Q., Wu, S.C., Collins, L.B., Swenberg, J.A., He, C., and Zhang, Y. (2011). Tet proteins can convert 5-methylcytosine to 5-formylcytosine and 5-carboxylcytosine. Science 333, 13001303. Janic, A., Mendizabal, L., Llamazares, S., Rossell, D., and Gonzalez, C. (2010). Ectopic expression of germline genes drives malignant brain tumor growth in Drosophila. Science 330, 18241827. Jin, Z., Kirilly, D., Weng, C., Kawase, E., Song, X., Smith, S., Schwartz, J., and Xie, T. (2008). Differentiation-defective stem cells outcompete normal stem cells for niche occupancy in the Drosophila ovary. Cell Stem Cell 2, 3949. Juliano, C.E., Swartz, S.Z., and Wessel, G.M. (2010). A conserved germline multipotency program. Development 137, 41134126.

Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc. 737

Cell Stem Cell

Perspective
Kadyrova, L.Y., Habara, Y., Lee, T.H., and Wharton, R.P. (2007). Translational control of maternal Cyclin B mRNA by Nanos in the Drosophila germline. Development 134, 15191527. Kai, T., and Spradling, A. (2004). Differentiating germ cells can revert into functional stem cells in Drosophila melanogaster ovaries. Nature 428, 564569. Kanatsu-Shinohara, M., Takehashi, M., Takashima, S., Lee, J., Morimoto, H., ssler, R., and Shinohara, T. (2008). Chuma, S., Raducanu, A., Nakatsuji, N., Fa Homing of mouse spermatogonial stem cells to germline niche depends on beta1-integrin. Cell Stem Cell 3, 533542. Katz, D.J., Edwards, T.M., Reinke, V., and Kelly, W.G. (2009). A C. elegans LSD1 demethylase contributes to germline immortality by reprogramming epigenetic memory. Cell 137, 308320. Kedde, M., Strasser, M.J., Boldajipour, B., Oude Vrielink, J.A., Slanchev, K., le Sage, C., Nagel, R., Voorhoeve, P.M., van Duijse, J., rom, U.A., et al. (2007). RNA-binding protein Dnd1 inhibits microRNA access to target mRNA. Cell 131, 12731286. Khurana, J.S., and Theurkauf, W. (2010). piRNAs, transposon silencing, and Drosophila germline development. J. Cell Biol. 191, 905913. Kiger, A.A., Jones, D.L., Schulz, C., Rogers, M.B., and Fuller, M.T. (2001). Stem cell self-renewal specied by JAK-STAT activation in response to a support cell cue. Science 294, 25422545. Kimble, J. (2011). Molecular regulation of the mitosis/meiosis decision in multicellular organisms. Cold Spring Harb Perspect Biol 3, a002683. Kimble, J., and Crittenden, S.L. (2007). Controls of germline stem cells, entry into meiosis, and the sperm/oocyte decision in Caenorhabditis elegans. Annu. Rev. Cell Dev. Biol. 23, 405433. Kobayashi, S., Yamada, M., Asaoka, M., and Kitamura, T. (1996). Essential role of the posterior morphogen nanos for germline development in Drosophila. Nature 380, 708711. Korta, D.Z., Tuck, S., and Hubbard, E.J. (2012). S6K links cell fate, cell cycle and nutrient response in C. elegans germline stem/progenitor cells. Development 139, 859870. Kraemer, B., Crittenden, S., Gallegos, M., Moulder, G., Barstead, R., Kimble, J., and Wickens, M. (1999). NANOS-3 and FBF proteins physically interact to control the sperm-oocyte switch in Caenorhabditis elegans. Curr. Biol. 9, 10091018. Kurimoto, K., Yabuta, Y., Ohinata, Y., Shigeta, M., Yamanaka, K., and Saitou, M. (2008). Complex genome-wide transcription dynamics orchestrated by Blimp1 for the specication of the germ cell lineage in mice. Genes Dev. 22, 16171635. Lamont, L.B., Crittenden, S.L., Bernstein, D., Wickens, M., and Kimble, J. (2004). FBF-1 and FBF-2 regulate the size of the mitotic region in the C. elegans germline. Dev. Cell 7, 697707. Lasko, P.F., and Ashburner, M. (1990). Posterior localization of vasa protein correlates with, but is not sufcient for, pole cell development. Genes Dev. 4, 905921. Lawson, K.A., Dunn, N.R., Roelen, B.A., Zeinstra, L.M., Davis, A.M., Wright, C.V., Korving, J.P., and Hogan, B.L. (1999). Bmp4 is required for the generation of primordial germ cells in the mouse embryo. Genes Dev. 13, 424436. Leatherman, J.L., and Dinardo, S. (2008). Zfh-1 controls somatic stem cell selfrenewal in the Drosophila testis and nonautonomously inuences germline stem cell self-renewal. Cell Stem Cell 3, 4454. Leatherman, J.L., and Dinardo, S. (2010). Germline self-renewal requires cyst stem cells and stat regulates niche adhesion in Drosophila testes. Nat. Cell Biol. 12, 806811. Lewis, P.W., Beall, E.L., Fleischer, T.C., Georlette, D., Link, A.J., and Botchan, M.R. (2004). Identication of a Drosophila Myb-E2F2/RBF transcriptional repressor complex. Genes Dev. 18, 29292940. Li, Y., Maines, J.Z., Tastan, O.Y., McKearin, D.M., and Buszczak, M. (2012). Mei-P26 regulates the maintenance of ovarian germline stem cells by promoting BMP signaling. Development 139, 15471556. Liu, N., Han, H., and Lasko, P. (2009). Vasa promotes Drosophila germline stem cell differentiation by activating mei-P26 translation by directly interacting with a (U)-rich motif in its 30 UTR. Genes Dev. 23, 27422752. Liu, M., Lim, T.M., and Cai, Y. (2010). The Drosophila female germline stem cell lineage acts to spatially restrict DPP function within the niche. Sci. Signal. 3, ra57. pez-Onieva, L., Ferna ndez-Min n, A., and Gonza lez-Reyes, A. (2008). Jak/ a Lo Stat signalling in niche support cells regulates dpp transcription to control germline stem cell maintenance in the Drosophila ovary. Development 135, 533540. Malone, C.D., and Hannon, G.J. (2009). Small RNAs as guardians of the genome. Cell 136, 656668. Marin, V.A., and Evans, T.C. (2003). Translational repression of a C. elegans Notch mRNA by the STAR/KH domain protein GLD-1. Development 130, 26232632. Matsui, Y., Zsebo, K., and Hogan, B.L. (1992). Derivation of pluripotential embryonic stem cells from murine primordial germ cells in culture. Cell 70, 841847. nen, M.E., Parvinen, M., de Rooij, D.G., Hess, Meng, X., Lindahl, M., Hyvo M.W., Raatikainen-Ahokas, A., Sainio, K., Rauvala, H., Lakso, M., et al. (2000). Regulation of cell fate decision of undifferentiated spermatogonia by GDNF. Science 287, 14891493. Merritt, C., Rasoloson, D., Ko, D., and Seydoux, G. (2008). 30 UTRs are the primary regulators of gene expression in the C. elegans germline. Curr. Biol. 18, 14761482. Michaelson, D., Korta, D.Z., Capua, Y., and Hubbard, E.J. (2010). Insulin signaling promotes germline proliferation in C. elegans. Development 137, 671680. Michel, M., Raabe, I., Kupinski, A.P., Perez-Palencia, R., and Bokel, C. (2011). Local BMP receptor activation at adherens junctions in the Drosophila germline stem cell niche. Nat. Comm. 2, 415. Monk, A.C., Siddall, N.A., Volk, T., Fraser, B., Quinn, L.M., McLaughlin, E.A., and Hime, G.R. (2010). HOW is required for stem cell maintenance in the Drosophila testis and for the onset of transit-amplifying divisions. Cell Stem Cell 6, 348360. Morris, L.X., and Spradling, A.C. (2011). Long-term live imaging provides new insight into stem cell regulation and germline-soma coordination in the Drosophila ovary. Development 138, 22072215. Nakagawa, T., Nabeshima, Y., and Yoshida, S. (2007). Functional identication of the actual and potential stem cell compartments in mouse spermatogenesis. Dev. Cell 12, 195206. Nakagawa, T., Sharma, M., Nabeshima, Y., Braun, R.E., and Yoshida, S. (2010). Functional hierarchy and reversibility within the murine spermatogenic stem cell compartment. Science 328, 6267. Nakamura, A., and Seydoux, G. (2008). Less is more: specication of the germline by transcriptional repression. Development 135, 38173827. Naughton, C.K., Jain, S., Strickland, A.M., Gupta, A., and Milbrandt, J. (2006). Glial cell-line derived neurotrophic factor-mediated RET signaling regulates spermatogonial stem cell fate. Biol. Reprod. 74, 314321. Oatley, J.M., and Brinster, R.L. (2008). Regulation of spermatogonial stem cell self-renewal in mammals. Annu. Rev. Cell Dev. Biol. 24, 263286. Oatley, J.M., Avarbock, M.R., and Brinster, R.L. (2007). Glial cell line-derived neurotrophic factor regulation of genes essential for self-renewal of mouse spermatogonial stem cells is dependent on Src family kinase signaling. J. Biol. Chem. 282, 2584225851. Ohinata, Y., Payer, B., OCarroll, D., Ancelin, K., Ono, Y., Sano, M., Barton, S.C., Obukhanych, T., Nussenzweig, M., Tarakhovsky, A., et al. (2005). Blimp1 is a critical determinant of the germ cell lineage in mice. Nature 436, 207213. Ohinata, Y., Ohta, H., Shigeta, M., Yamanaka, K., Wakayama, T., and Saitou, M. (2009). A signaling principle for the specication of the germ cell lineage in mice. Cell 137, 571584.

738 Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc.

Cell Stem Cell

Perspective
Ohlstein, B., and McKearin, D. (1997). Ectopic expression of the Drosophila Bam protein eliminates oogenic germline stem cells. Development 124, 36513662. Pan, L., Chen, S., Weng, C., Call, G., Zhu, D., Tang, H., Zhang, N., and Xie, T. (2007). Stem cell aging is controlled both intrinsically and extrinsically in the Drosophila ovary. Cell Stem Cell 1, 458469. Park, J.K., Liu, X., Strauss, T.J., McKearin, D.M., and Liu, Q. (2007). The miRNA pathway intrinsically controls self-renewal of Drosophila germline stem cells. Curr. Biol. 17, 533538. Pek, J.W., and Kai, T. (2011). A role for vasa in regulating mitotic chromosome condensation in Drosophila. Curr. Biol. 21, 3944. Petrella, L.N., Wang, W., Spike, C.A., Rechtsteiner, A., Reinke, V., and Strome, S. (2011). synMuv B proteins antagonize germline fate in the intestine and ensure C. elegans survival. Development 138, 10691079. Richter, J.D., and Lasko, P. (2011). Translational control in oocyte development. Cold Spring Harb Perspect Biol 3, a002758. ller, J., and Knoblich, J.A. (2011). The Richter, C., Oktaba, K., Steinmann, J., Mu tumour suppressor L(3)mbt inhibits neuroepithelial proliferation and acts on insulator elements. Nat. Cell Biol. 13, 10291039. Sada, A., Suzuki, A., Suzuki, H., and Saga, Y. (2009). The RNA-binding protein NANOS2 is required to maintain murine spermatogonial stem cells. Science 325, 13941398. Sada, A., Hasegawa, K., Pin, P.H., and Saga, Y. (2012). NANOS2 acts downstream of glial cell line-derived neurotrophic factor signaling to suppress differentiation of spermatogonial stem cells. Stem Cells 30, 280291. Saga, Y. (2010). Function of Nanos2 in the male germ cell lineage in mice. Cell. Mol. Life Sci. 67, 38153822. Seki, Y., Yamaji, M., Yabuta, Y., Sano, M., Shigeta, M., Matsui, Y., Saga, Y., Tachibana, M., Shinkai, Y., and Saitou, M. (2007). Cellular dynamics associated with the genome-wide epigenetic reprogramming in migrating primordial germ cells in mice. Development 134, 26272638. Shivdasani, A.A., and Ingham, P.W. (2003). Regulation of stem cell maintenance and transit amplifying cell proliferation by tgf-beta signaling in Drosophila spermatogenesis. Curr. Biol. 13, 20652072. Simon, L., Ekman, G.C., Garcia, T., Carnes, K., Zhang, Z., Murphy, T., Murphy, K.M., Hess, R.A., Cooke, P.S., and Hofmann, M.C. (2010). ETV5 regulates sertoli cell chemokines involved in mouse stem/progenitor spermatogonia maintenance. Stem Cells 28, 18821892. Song, X., Wong, M.D., Kawase, E., Xi, R., Ding, B.C., McCarthy, J.J., and Xie, T. (2004). Bmp signals from niche cells directly repress transcription of a differentiation-promoting gene, bag of marbles, in germline stem cells in the Drosophila ovary. Development 131, 13531364. Sonoda, J., and Wharton, R.P. (2001). Drosophila Brain Tumor is a translational repressor. Genes Dev. 15, 762773. Spradling, A., Fuller, M.T., Braun, R.E., and Yoshida, S. (2011). Germline stem cells. Cold Spring Harb Perspect Biol 3, a002642. Subramaniam, K., and Seydoux, G. (1999). nos-1 and nos-2, two genes related to Drosophila nanos, regulate primordial germ cell development and survival in Caenorhabditis elegans. Development 126, 48614871. Suh, N., Jedamzik, B., Eckmann, C.R., Wickens, M., and Kimble, J. (2006). The GLD-2 poly(A) polymerase activates gld-1 mRNA in the Caenorhabditis elegans germ line. Proc. Natl. Acad. Sci. USA 103, 1510815112. Sun, P., Quan, Z., Zhang, B., Wu, T., and Xi, R. (2010). TSC1/2 tumour suppressor complex maintains Drosophila germline stem cells by preventing differentiation. Development 137, 24612469. Suzuki, A., Igarashi, K., Aisaki, K., Kanno, J., and Saga, Y. (2010). NANOS2 interacts with the CCR4-NOT deadenylation complex and leads to suppression of specic RNAs. Proc. Natl. Acad. Sci. USA 107, 35943599. Tahiliani, M., Koh, K.P., Shen, Y., Pastor, W.A., Bandukwala, H., Brudno, Y., Agarwal, S., Iyer, L.M., Liu, D.R., Aravind, L., and Rao, A. (2009). Conversion of 5-methylcytosine to 5-hydroxymethylcytosine in mammalian DNA by MLL partner TET1. Science 324, 930935. Tazuke, S.I., Schulz, C., Gilboa, L., Fogarty, M., Mahowald, A.P., Guichet, A., Ephrussi, A., Wood, C.G., Lehmann, R., and Fuller, M.T. (2002). A germlinespecic gap junction protein required for survival of differentiating early germ cells. Development 129, 25292539. Tomancak, P., Guichet, A., Zavorszky, P., and Ephrussi, A. (1998). Oocyte polarity depends on regulation of gurken by Vasa. Development 125, 1723 1732. Tulina, N., and Matunis, E. (2001). Control of stem cell self-renewal in Drosophila spermatogenesis by JAK-STAT signaling. Science 294, 2546 2549. Tursun, B., Patel, T., Kratsios, P., and Hobert, O. (2011). Direct conversion of C. elegans germ cells into specic neuron types. Science 331, 304308. Vagin, V.V., Klenov, M.S., Kalmykova, A.I., Stolyarenko, A.D., Kotelnikov, R.N., and Gvozdev, V.A. (2004). The RNA interference proteins and vasa locus are involved in the silencing of retrotransposons in the female germline of Drosophila melanogaster. RNA Biol. 1, 5458. Voronina, E., Seydoux, G., Sassone-Corsi, P., and Nagamori, I. (2011). RNA granules in germ cells. Cold Spring Harb Perspect Biol 3, a002774. Wallenfang, M.R., Nayak, R., and DiNardo, S. (2006). Dynamics of the male germline stem cell population during aging of Drosophila melanogaster. Aging Cell 5, 297304. Wang, Z., and Lin, H. (2004). Nanos maintains germline stem cell self-renewal by preventing differentiation. Science 303, 20162019. Wang, X., Harris, R.E., Bayston, L.J., and Ashe, H.L. (2008). Type IV collagens regulate BMP signalling in Drosophila. Nature 455, 7277. West, J.A., Viswanathan, S.R., Yabuuchi, A., Cunniff, K., Takeuchi, A., Park, I.H., Sero, J.E., Zhu, H., Perez-Atayde, A., Frazier, A.L., et al. (2009). A role for Lin28 in primordial germ-cell development and germ-cell malignancy. Nature 460, 909913. White, Y.A., Woods, D.C., Takai, Y., Ishihara, O., Seki, H., and Tilly, J.L. (2012). Oocyte formation by mitotically active germ cells puried from ovaries of reproductive-age women. Nat. Med. 18, 413421. Wong, C., and Jones, D.L. (2012). Efciency of spermatogonial dedifferentiation during aging. PLoS ONE 7, e33635. Xia, L., Jia, S., Huang, S., Wang, H., Zhu, Y., Mu, Y., Kan, L., Zheng, W., Wu, D., Li, X., et al. (2010). The Fused/Smurf complex controls the fate of Drosophila germline stem cells by generating a gradient BMP response. Cell 143, 978990. Xie, T., and Spradling, A.C. (1998). decapentaplegic is essential for the maintenance and division of germline stem cells in the Drosophila ovary. Cell 94, 251260. Yadlapalli, S., Cheng, J., and Yamashita, Y.M. (2011). Drosophila male germline stem cells do not asymmetrically segregate chromosome strands. J. Cell Sci. 124, 933939. Yamaji, M., Seki, Y., Kurimoto, K., Yabuta, Y., Yuasa, M., Shigeta, M., Yamanaka, K., Ohinata, Y., and Saitou, M. (2008). Critical function of Prdm14 for the establishment of the germ cell lineage in mice. Nat. Genet. 40, 10161022. Yamashita, Y.M. (2010). Cell adhesion in regulation of asymmetric stem cell division. Curr. Opin. Cell Biol. 22, 605610. Yamashita, Y.M., Mahowald, A.P., Perlin, J.R., and Fuller, M.T. (2007). Asymmetric inheritance of mother versus daughter centrosome in stem cell division. Science 315, 518521. Ying, Y., Qi, X., and Zhao, G.Q. (2001). Induction of primordial germ cells from murine epiblasts by synergistic action of BMP4 and BMP8B signaling pathways. Proc. Natl. Acad. Sci. USA 98, 78587862. Yoshida, S., Sukeno, M., and Nabeshima, Y. (2007). A vasculature-associated niche for undifferentiated spermatogonia in the mouse testis. Science 317, 17221726.

Cell Stem Cell 10, June 14, 2012 2012 Elsevier Inc. 739

Cell Stem Cell

Brief Report
Il2rg Gene-Targeted Severe Combined Immunodeciency Pigs
Shunichi Suzuki,1 Masaki Iwamoto,3 Yoriko Saito,4 Daiichiro Fuchimoto,1 Shoichiro Sembon,1 Misae Suzuki,1 Satoshi Mikawa,2 Michiko Hashimoto,3 Yuki Aoki,4 Yuho Najima,4 Shinsuke Takagi,4 Nahoko Suzuki,4 Emi Suzuki,5 Masanori Kubo,6 Jun Mimuro,7 Yuji Kashiwakura,7 Seiji Madoiwa,7 Yoichi Sakata,7 Anthony C.F. Perry,8 Fumihiko Ishikawa,4,* and Akira Onishi1,*
Animal Research Center Genome Research Unit National Institute of Agrobiological Sciences, Tsukuba, Ibaraki 305-0901, Japan 3Prime Tech Ltd., Tsuchiura, Ibaraki 300-0841, Japan 4Research Unit for Human Disease Model, RIKEN Research Center for Allergy and Immunology, Yokohama, Kanagawa 230-0045, Japan 5Laboratory of Mammalian Molecular Embryology, RIKEN Research Center for Developmental Biology, Kobe 650-0047, Japan 6Center for Animal Disease Control and Prevention, National Institute of Animal Health, Tsukuba, Ibaraki 305-0856, Japan 7Division of Cell and Molecular Medicine, Center for Molecular Medicine, Jichi Medical University, Tochigi-ken 329-0498, Japan 8Laboratory of Mammalian Molecular Embryology, Department of Biology and Biochemistry, University of Bath, Bath BA2 7AY, UK *Correspondence: f_ishika@rcai.riken.jp (F.I.), onishi@affrc.go.jp (A.O.) DOI 10.1016/j.stem.2012.04.021
2Animal 1Transgenic

SUMMARY

A porcine model of severe combined immunodeciency (SCID) promises to facilitate human cancer studies, the humanization of tissue for xenotransplantation, and the evaluation of stem cells for clinical therapy, but SCID pigs have not been described. We report here the generation and preliminary evaluation of a porcine SCID model. Fibroblasts containing a targeted disruption of the X-linked interleukin2 receptor gamma chain gene, Il2rg, were used as donors to generate cloned pigs by serial nuclear transfer. Germline transmission of the Il2rg deletion produced healthy Il2rg+/ females, while Il2rg/Y males were athymic and exhibited markedly impaired immunoglobulin and T and NK cell production, robustly recapitulating human SCID. Following allogeneic bone marrow transplantation, donor cells stably integrated in Il2rg/Y heterozygotes and reconstituted the Il2rg/Y lymphoid lineage. The SCID pigs described here represent a step toward the comprehensive evaluation of preclinical cellular regenerative strategies.
The common gamma chain, IL2RG, is an IL-2 receptor subunit (Takeshita et al., 1992) shared by IL-4, IL-7, IL-9, IL-15, and IL-21 receptors (Kondo et al., 1993; Noguchi et al., 1993a; Russell et al., 1993; Giri et al., 1994; Kimura et al., 1995; Asao et al., 2001). IL2RG is present in T, NK, NKT, and dendritic cells (Ishii et al., 1994) and plays an essential role in lymphoid development by activating, through its cytoplasmic domain, Janus kinase 3 (Nakamura et al., 1994; Nelson et al., 1994, 1997). Mammalian IL2RG orthologs are typically located on the X chromosome; in humans, IL2RG mutations result in X-linked severe combined immunodeciency (XSCID) in which T and

NK cells are absent or profoundly reduced in number, while B cells are numerically normal (or increased) but functionally impaired (Noguchi et al., 1993b; Leonard, 1996; Fischer et al., 1997). Gene-targeted mice lacking Il2rg also exhibit immunological defects (Cao et al., 1995; Ohbo et al., 1996) including the ablation of NK cell activity. NOD/SCID/Il2rgnull, NOG, and Rag2null/Il2rgnull mice permit the functional reconstitution of human hematopoietic and immune systems following the injection of puried human hematopoietic stem cells (Traggiai et al., 2004; Ishikawa et al., 2005; Shultz et al., 2005). Unfortunately, phenotypic differences exist between XSCID humans and Il2rg null mice, including a pronounced numerical B cell reduction in the latter. Two dog breeds develop SCID caused by Il2rg mutations (Felsburg et al., 1999; Perryman, 2004), but dogs are poorly characterized research models. In contrast, the pig more closely resembles humans regarding anatomy, hematology, physiology, size, and longevity. By enabling long-term follow-up, pig models will permit the evaluation of human cancer and stem cell transplantation over clinically relevant time frames. We here describe disruption of the porcine Il2rg gene to generate SCID pigs, their phenotypic characterization, and proof-of-principle transplantation studies. A conventional positive-negative selection Il2rg gene targeting vector (TV) (Figure S1A) enabled the functional inactivation of porcine Il2rg by removing exon 6 (Kanai et al., 1999). Following fetal broblast transfection, selection, and PCR screening, 1 of 3 TV-targeted cell lines was expanded for nuclear transfer (Table S1). Even after screening, PCR-positive colonies often contain a substantial proportion of nontargeted cells (not shown) that would result in a corresponding proportion of nontargeted cloned pigs following nuclear transfer. To ensure that all clones harbored a targeted Il2rg locus, we adopted a serial cloning strategy. Nine embryonic day 35 (E35) or E39 nuclear transfer embryos were collected and screened by PCR and Southern blotting. PCR (not shown) and Southern blotting (Figure S1B) revealed that six contained the genomic conguration predicted for a single-targeted Il2rg allele. Fibroblasts were cultured from one targeted embryo and used in secondary nuclear transfer to
Cell Stem Cell 10, 753758, June 14, 2012 2012 Elsevier Inc. 753

Cell Stem Cell


Il2rg-Targeted Nuclear Transfer SCID Pigs

754 Cell Stem Cell 10, 753758, June 14, 2012 2012 Elsevier Inc.

Cell Stem Cell


Il2rg-Targeted Nuclear Transfer SCID Pigs

produce 31 cloned F0 piglets; all were females heterozygously targeted at their Il2rg genes as judged by PCR (not shown) and Southern blotting (Figure S1C). Fourteen of the 31 were stillborn and 3 of the 17 live-born died neonatally of unknown cause(s) (Table S1). Ten survivors died from pneumonia and severe arthritis (ve were euthanized) between postnatal day 7 (P7) and P70. The remaining four (#5, 9, 15 and 20) survived for >1 year. Stillborn and neonatal fatalities often had spleens with hypoplastic lymphoid aggregations (Figure S1D). Most (24/31, 77%) had undetectable or severely hypoplastic thymi (Figure S1E and Table S2). F0 clones that died within 10 weeks also lacked detectable thymi and had few, if any, T cells either in their spleens or circulating (Figures S1F and S1G). In contrast, levels of CD4+ and CD8+ T cells in four long-lived F0 lines were comparable to those of WT controls. Analysis of peripheral blood (PB) mononuclear cell (PBMC) RNA corroborated this: Il2rg, CD4, and CD8 transcript levels were reduced in athymic Il2rg+/ clones that perished relative to respective levels in long-lived clones and WT controls (Figure S1H). Thus, most Il2rg+/ clones exhibited SCID-like phenotypes, albeit that they had one WT allele. We attributed this high proportion to aberrant X-inactivation, a previously observed epigenetic cloning phenotype (Senda et al., 2004; Nolen et al., 2005; Jiang et al., 2008). However, epigenetic cloning phenotypes are corrected by germline transmission (Shimozawa et al., 2002). To conrm this and isolate the Il2rg+/ phenotype, we analyzed progeny derived by fertilization from Il2rg+/ cloned female #9. Female Il2rg+/ #9 inseminated with WT sperm produced 19 F1 (12 m, 7 f) offspring, and of these F1 offspring, two Il2rg+/ females produced 21 F2 (13 m, 8 f) when inseminated with WT sperm. Autopsies of representative F1 and F2 progeny revealed that, as expected, all Il2rg/Y males had undetectable thymi, whereas Il2rg+/ females had thymi of normal size (Figure 1A and Table S3). Hematological parameters in PB exhibited a signicantly (p = 0.0041) reduced white blood cell (WBC) count in F1 Il2rg/Y males (6.4 1.6 3 103/ml, n = 5) compared to WT controls (17.8 2.3 3 103/ml, n = 6), while hemoglobin levels and platelet counts were unaffected (Figure 1B). F1 Il2rg+/ females and WT littermates yielded comparable PB T, NK, and B cell numbers, indicative of intact acquired and innate immunity (Figures 1C, 1E, and 1F). In contrast, Il2rg/Y males harbored signicantly reduced PB T cells (Il2rg/Y males, 1.5% 1.0%; Il2rg+/Y males, 57.3% 4.3%; n = 4 each, p < 0.0001) and NK cells (Il2rg/Y males, 0.1% 0.1%; Il2rg+/Y males, 3.6% 1.1%; n = 4 each, p = 0.0162) (Figures 1D, 1E, and 1F). In proportion to the PB reductions, Il2rg/Y spleens exhibited signicant numerical

reductions of T cells (Il2rg/Y males, 2.3% 1.1%; Il2rg+/Y males, 13.0% 1.4%; n = 4 each, p = 0.0011) and NK cells (Il2rg/Y males, 0.1% 0.0%; Il2rg+/Y males, 0.8% 0.2%; n = 4 each, p = 0.0162) (Figures 1E and 1F). B cells and myeloid cells accounted for the majority of CD45+ leukocytes in Il2rg/Y males, indicating that their immune deciency was limited to T and NK cell lineages. The presence in Il2rg/Y males of CD33+ myeloid cells with both mononuclear and polynuclear properties suggests the differentiation of both granulocyte lineages and antigen-presenting cells, including monocytes and dendritic cells (Figures 1C and 1D). We next evaluated humoral immune status in Il2rg-targeted pigs (Figure 1G). Serum IgG and IgA levels were high at 1 week (P7) and decreased gradually from 3 to 5 weeks in Il2rg/Y (males), WT littermates, and Il2rg+/ female controls. After 7 weeks, IgG and IgA levels re-elevated in WT controls, while levels of both remained low in Il2rg/Y males. Serum IgM was low at 1 week and increased gradually in controls but remained low in Il2rg/Y males. Because both IgG and IgA at 1 week of age are entirely transferred via the colostrum in pigs, these results indicate that there had been no de novo Ig production in Il2rg/Y males after weaning at 4 weeks. Impaired antibody production by Il2rg/Y B cells is likely due to the absence of critical CD4+ T helper cells. Consistent with their impaired immunity, all F1 Il2rg/Y males became systemically ill in the conventional housing conditions used, while F1 Il2rg+/ females appeared healthy. Collectively, this shows that when produced by conventional breeding, Il2rg/Y F1 males, but not Il2rg+/ females, present SCID phenotypes. SCID-like phenotypes observed in many Il2rg+/ female clones are attributable to aberrant, nonrandom X-inactivation during somatic cell cloning. Following germline transmission, Il2rg-targeted phenotypes resembled those of X-linked SCID in other species, with greatly reduced T and NK cell development and function (Cao et al., 1995; Puck et al., 1987). Il2rg-targeted pigs harbor B cells and thereby recapitulate human XSCID more closely than do Il2rg-targeted mice. We next performed proof-of-principle allogeneic transplantation experiments using Il2rg-targeted pigs as recipients. Preliminary conditioning with orally administered udarabine and busulfan produced signicantly decreased WBC and platelet counts (not shown), which might promote the engraftment of transplanted cells. However, 2 of 6 Il2rg/Y males died within 2 weeks postadministration, suggesting that the regimen was lethal for some piglets. Bone marrow (BM) cells from WT siblings were intravenously transplanted to four P11-12 Il2rg/Y males with (#113, #115) or without (#605, #610) conditioning. Ubiquitously GFP-expressing (#184) BM cells (Watanabe et al., 2005)

(A) Thymic phenotype in an Il2rg+/ female at 10 weeks and an Il2rg/Y male at 9 weeks. (B) Peripheral blood (PB) white blood cell (WBC), hemoglobin (Hgb), and platelet count (Plt) at 2 months of age in WT controls, Il2rg+/+ female littermates, Il2rg+/ female littermates, and IL2rg/Y males. (C) Identication of acquired and innate immune subsets in an Il2rg+/+ female by surface phenotype of CD45RA+CD3 B cells, CD3+CD4+ T cells, CD3+CD8+ T cells, and CD3CD16+ NK cells among the nonmyeloid fraction and myeloid cells. (D) Analysis as for (C), but of an IL2rg/Y male. (E) Proportion of CD3+ T cells in the spleen (spl), PB, and thymus (thy) in control Il2rg+/+ females and male Il2rg+/Y littermates and nonlittermates (WT). (F) Analysis as for (E), except showing the proportion of CD3CD16+ NK cells. (G) Changes with time postpartum in serum IgG (left), IgA (middle), and IgM levels. All error bars indicate SEM. See also Figure S1 and Tables S1S3.

Figure 1. Phenotypes of F1 and F2 Progeny Derived from Il2rg+/ Clone #9 by Germline Transmission

Cell Stem Cell 10, 753758, June 14, 2012 2012 Elsevier Inc. 755

Cell Stem Cell


Il2rg-Targeted Nuclear Transfer SCID Pigs

Figure 2. Allogeneic Bone Marrow Transfer into Il2rg/Y Males


(AH) Flow cytometric quantication of each subtype of leukocytes in Il2rg/Y males (red lines) at different times after BM transfer. Controls are WT littermates (blue lines) and Il2rg/Y males (black lines) without BM transfer. (A)(D) correspond to BMT without conditioning, the case for #605 and #610. (E)(H) correspond to BMT with conditioning, the case for #105, #113, and #122. (A) and (E) show the proportion of CD3+ T cells in PB at different times after BM transfer. (B) and (F) show proportion of CD45RA+CD3 B cells in PB at different times after BM transfer. (C) and (G) show proportion of granulocytes in PB at different times after BM transfer. (D) and (H) show proportion of monocytes in PB at different times after BM transfer.

756 Cell Stem Cell 10, 753758, June 14, 2012 2012 Elsevier Inc.

Cell Stem Cell


Il2rg-Targeted Nuclear Transfer SCID Pigs

were transferred to a Il2rg/Y male (#122) with conditioning. Recipient #115 died 15 days after BM transfer (BMT), possibly due to conditioning toxicity, and recipient #605 died of presumptive pneumonia (not shown) 140 days posttransplantation; nevertheless, it survived longer than Il2rg/Y controls without BMT, which died within 54 days. The remaining three recipients have survived >516 days (#610) and >321 days (#113 and #122) posttransplantation, with PB T cell populations exhibiting similar dynamics (Figures 2A and 2E); T cell counts increased $6 weeks posttransplantation and remained high. Following early variability, PB B cell counts equilibrated at detectably low levels. The exception was #113, in which B cell counts were clearly higher than those of WT controls until 16 weeks posttransplantation, gradually decreasing thereafter to a level comparable with other recipients (Figures 2B and 2F). Compared to WT controls, recipient #610 PB contained similar T cell counts and a diminished but substantial number of B and NK cells 42 weeks posttransplantation (Figures 2A, 2B, and 2I). Similar proles were observed in #113 and 122 (Figures 2E and 2F; not shown for NK). Surviving recipients and WT littermates had comparable granulocyte and monocyte numbers (Figures 2C, 2D, 2G, and 2H). The provenance of immune cells in surviving recipients was examined by microsatellite marker analysis of genomic DNA from the ear, whole blood, and sorted T, B, and NK cells. GFP uorescence of immune cells was also detected in the case of #122 (Figures 2E2H and 2J). Lymphoid lineages were totally of donor origin in all recipients. Myeloid lineages were mainly of donor origin in #113 and of mixed donor/host origin in #122 and #610. Thus, our conditioning regimen enabled donorderived myeloid lineage reconstitution in #113, but not #122. All surviving recipient PB contained IgG, IgA, and IgM (Figure 2K), albeit at varying levels, strongly suggesting that humoral immunity had been reconstituted and that donor-derived recipient B cells produced antibodies. Thus, allogeneic BM transplantation to Il2rg/Y SCID pigs reproducibly resulted in enduring functional donor cell engraftment and reconstituted acquired immunity. Assuming that the allogeneic model reported here reects the behavior of cells transplanted from different species, SCID pigs promise to become a valuable tool in xenogeneic transplantation studies of human stem cells, such as hematopoietic, embryonic, and induced pluripotent stem cells (Takahashi et al., 2007). In particular, they promise to serve as platforms for the evaluation of therapeutic outcomes over several years, possibly after further genetic manipulation such as disruption of recombination activating genes 1 or 2 (Rag1, Rag2), which play critical roles in both cellular and humoral immunity (Shinkai et al., 1992) and may facilitate efcient human stem cell engraftment. Preliminary xenotransplantation of human BM cells to porcine Il2rg/Y recipients in the absence of preconditioning permitted limited engraftment, underscoring the importance of further genetic manipulation, and optimized

preconditioning (not shown). The porcine SCID model described here therefore represents an essential step toward the translational evaluation of human stem cells for long-term clinical applications.
SUPPLEMENTAL INFORMATION Supplemental Information includes one gure, three tables, and Supplemental Experimental Procedures and can be found with this article online at doi:10. 1016/j.stem.2012.04.021. ACKNOWLEDGMENTS This work was supported in part by a Grant-in-Aid from the Ministry of Agriculture, Forestry, and Fisheries of Japan. We thank staff in the Pig Management Section of the National Institute of Livestock and Grassland Science for assistance with animal management and Dr. T. Yagi for providing PKJ2 and MC1DTA-p(A). Received: November 17, 2011 Revised: March 13, 2012 Accepted: April 18, 2012 Published: June 13, 2012 REFERENCES Asao, H., Okuyama, C., Kumaki, S., Ishii, N., Tsuchiya, S., Foster, D., and Sugamura, K. (2001). Cutting edge: the common gamma-chain is an indispensable subunit of the IL-21 receptor complex. J. Immunol. 167, 15. Cao, X., Shores, E.W., Hu-Li, J., Anver, M.R., Kelsall, B.L., Russell, S.M., Drago, J., Noguchi, M., Grinberg, A., Bloom, E.T., et al. (1995). Defective lymphoid development in mice lacking expression of the common cytokine receptor gamma chain. Immunity 2, 223238. Felsburg, P.J., Hartnett, B.J., Henthorn, P.S., Moore, P.F., Krakowka, S., and Ochs, H.D. (1999). Canine X-linked severe combined immunodeciency. Vet. Immunol. Immunopathol. 69, 127135. Fischer, A., Cavazzana-Calvo, M., De Saint Basile, G., DeVillartay, J.P., Di Santo, J.P., Hivroz, C., Rieux-Laucat, F., and Le Deist, F. (1997). Naturally occurring primary deciencies of the immune system. Annu. Rev. Immunol. 15, 93124. Giri, J.G., Ahdieh, M., Eisenman, J., Shanebeck, K., Grabstein, K., Kumaki, S., Namen, A., Park, L.S., Cosman, D., and Anderson, D. (1994). Utilization of the beta and gamma chains of the IL-2 receptor by the novel cytokine IL-15. EMBO J. 13, 28222830. Ishii, N., Takeshita, T., Kimura, Y., Tada, K., Kondo, M., Nakamura, M., and Sugamura, K. (1994). Expression of the IL-2 receptor gamma chain on various populations in human peripheral blood. Int. Immunol. 6, 12731277. Ishikawa, F., Yasukawa, M., Lyons, B., Yoshida, S., Miyamoto, T., Yoshimoto, G., Watanabe, T., Akashi, K., Shultz, L.D., and Harada, M. (2005). Development of functional human blood and immune systems in NOD/SCID/ IL2 receptor g chain(null) mice. Blood 106, 15651573. Jiang, L., Lai, L., Samuel, M., Prather, R.S., Yang, X., and Tian, X.C. (2008). Expression of X-linked genes in deceased neonates and surviving cloned female piglets. Mol. Reprod. Dev. 75, 265273. Kanai, N., Yanai, F., Hirose, S., Nibu, K., Izuhara, K., Tani, T., Kubota, T., and Mitsudome, A. (1999). A G to A transition at the last nucleotide of exon 6 of the gamma c gene (868G/A) may result in either a splice or missense mutation in

(I) Identication of immune subsets in recipient Il2rg/Y male, #610, 12 weeks posttransfer as CD45RA+CD3 B cells, CD3+CD45RA+ naive T cells, CD3+CD45RA memory T cells, and CD3CD16+ NK cells among nonmyeloid and myeloid cells. (J) Representative microsatellite PCR analyses to distinguish between donor and recipient DNA illustrated by the discriminatory marker (SW1263 for #610, SWR1367 for # 113, and SW24 for #122). (K) Changes of IgG (left), IgA (middle), and IgM levels in serum from recipients (red lines) and wild-type controls (blue lines).

Cell Stem Cell 10, 753758, June 14, 2012 2012 Elsevier Inc. 757

Cell Stem Cell


Il2rg-Targeted Nuclear Transfer SCID Pigs

patients with X-linked severe combined immunodeciency. Hum. Genet. 104, 3642. Kimura, Y., Takeshita, T., Kondo, M., Ishii, N., Nakamura, M., Van Snick, J., and Sugamura, K. (1995). Sharing of the IL-2 receptor gamma chain with the functional IL-9 receptor complex. Int. Immunol. 7, 115120. Kondo, M., Takeshita, T., Ishii, N., Nakamura, M., Watanabe, S., Arai, K., and Sugamura, K. (1993). Sharing of the interleukin-2 (IL-2) receptor gamma chain between receptors for IL-2 and IL-4. Science 262, 18741877. Leonard, W.J. (1996). The molecular basis of X-linked severe combined immunodeciency: defective cytokine receptor signaling. Annu. Rev. Med. 47, 229239. Nakamura, Y., Russell, S.M., Mess, S.A., Friedmann, M., Erdos, M., Francois, C., Jacques, Y., Adelstein, S., and Leonard, W.J. (1994). Heterodimerization of the IL-2 receptor beta- and gamma-chain cytoplasmic domains is required for signalling. Nature 369, 330333. Nelson, B.H., Lord, J.D., and Greenberg, P.D. (1994). Cytoplasmic domains of the interleukin-2 receptor beta and gamma chains mediate the signal for T-cell proliferation. Nature 369, 333336. Nelson, B.H., McIntosh, B.C., Rosencrans, L.L., and Greenberg, P.D. (1997). Requirement for an initial signal from the membrane-proximal region of the interleukin 2 receptor gamma(c) chain for Janus kinase activation leading to T cell proliferation. Proc. Natl. Acad. Sci. USA 94, 18781883. Noguchi, M., Nakamura, Y., Russell, S.M., Ziegler, S.F., Tsang, M., Cao, X., and Leonard, W.J. (1993a). Interleukin-2 receptor gamma chain: a functional component of the interleukin-7 receptor. Science 262, 18771880. Noguchi, M., Yi, H., Rosenblatt, H.M., Filipovich, A.H., Adelstein, S., Modi, W.S., McBride, O.W., and Leonard, W.J. (1993b). Interleukin-2 receptor gamma chain mutation results in X-linked severe combined immunodeciency in humans. Cell 73, 147157. Nolen, L.D., Gao, S., Han, Z., Mann, M.R., Gie Chung, Y., Otte, A.P., Bartolomei, M.S., and Latham, K.E. (2005). X chromosome reactivation and regulation in cloned embryos. Dev. Biol. 279, 525540. Ohbo, K., Suda, T., Hashiyama, M., Mantani, A., Ikebe, M., Miyakawa, K., Moriyama, M., Nakamura, M., Katsuki, M., Takahashi, K., et al. (1996). Modulation of hematopoiesis in mice with a truncated mutant of the interleukin-2 receptor gamma chain. Blood 87, 956967.

Perryman, L.E. (2004). Molecular pathology of severe combined immunodeciency in mice, horses, and dogs. Vet. Pathol. 41, 95100. Puck, J.M., Nussbaum, R.L., and Conley, M.E. (1987). Carrier detection in X-linked severe combined immunodeciency based on patterns of X chromosome inactivation. J. Clin. Invest. 79, 13951400. Russell, S.M., Keegan, A.D., Harada, N., Nakamura, Y., Noguchi, M., Leland, P., Friedmann, M.C., Miyajima, A., Puri, R.K., Paul, W.E., et al. (1993). Interleukin-2 receptor gamma chain: a functional component of the interleukin-4 receptor. Science 262, 18801883. Senda, S., Wakayama, T., Yamazaki, Y., Ohgane, J., Hattori, N., Tanaka, S., Yanagimachi, R., and Shiota, K. (2004). Skewed X-inactivation in cloned mice. Biochem. Biophys. Res. Commun. 321, 3844. Shimozawa, N., Ono, Y., Kimoto, S., Hioki, K., Araki, Y., Shinkai, Y., Kono, T., and Ito, M. (2002). Abnormalities in cloned mice are not transmitted to the progeny. Genesis 34, 203207. Shinkai, Y., Rathbun, G., Lam, K.P., Oltz, E.M., Stewart, V., Mendelsohn, M., Charron, J., Datta, M., Young, F., Stall, A.M., et al. (1992). RAG-2-decient mice lack mature lymphocytes owing to inability to initiate V(D)J rearrangement. Cell 68, 855867. Shultz, L.D., Lyons, B.L., Burzenski, L.M., Gott, B., Chen, X., Chaleff, S., Kotb, M., Gillies, S.D., King, M., Mangada, J., et al. (2005). Human lymphoid and myeloid cell development in NOD/LtSz-scid IL2R gamma null mice engrafted with mobilized human hemopoietic stem cells. J. Immunol. 174, 64776489. Takahashi, K., Tanabe, K., Ohnuki, M., Narita, M., Ichisaka, T., Tomoda, K., and Yamanaka, S. (2007). Induction of pluripotent stem cells from adult human broblasts by dened factors. Cell 131, 861872. Takeshita, T., Asao, H., Ohtani, K., Ishii, N., Kumaki, S., Tanaka, N., Munakata, H., Nakamura, M., and Sugamura, K. (1992). Cloning of the gamma chain of the human IL-2 receptor. Science 257, 379382. Traggiai, E., Chicha, L., Mazzucchelli, L., Bronz, L., Piffaretti, J.C., Lanzavecchia, A., and Manz, M.G. (2004). Development of a human adaptive immune system in cord blood cell-transplanted mice. Science 304, 104107. Watanabe, S., Iwamoto, M., Suzuki, S., Fuchimoto, D., Honma, D., Nagai, T., Hashimoto, M., Yazaki, S., Sato, M., and Onishi, A. (2005). A novel method for the production of transgenic cloned pigs: electroporation-mediated gene transfer to non-cultured cells and subsequent selection with puromycin. Biol. Reprod. 72, 309315.

758 Cell Stem Cell 10, 753758, June 14, 2012 2012 Elsevier Inc.

Cell Stem Cell

Article
Protein Kinase A Determines Timing of Early Differentiation through Epigenetic Regulation with G9a
Kohei Yamamizu,1,2 Mayako Fujihara,3 Makoto Tachibana,4 Shiori Katayama,1,2 Akiko Takahashi,5 Eiji Hara,5 Hiroshi Imai,3 Yoichi Shinkai,4 and Jun K. Yamashita1,2,*
1Department of Stem Cell Differentiation, Institute for Frontier Medical Sciences, Kyoto University, Kyoto 606-8507, Japan 2Department of Cell Growth and Differentiation, Center for iPS Cell Research and Application (CiRA), Kyoto University, Kyoto 606-8507, Japan 3Laboratory of Reproductive Biology, Graduate School of Agriculture, Kyoto University, Kyoto 606-8502, Japan 4Experimental Research Center for Infectious Disease, Institute for Virus Research, Kyoto University, Kyoto 606-8507, Japan 5Cancer Institute, Japanese Foundation for Cancer Research, Tokyo 135-8550, Japan *Correspondence: juny@frontier.kyoto-u.ac.jp DOI 10.1016/j.stem.2012.02.022

SUMMARY

Timing of cell differentiation is strictly controlled and is crucial for normal development and stem cell differentiation. However, underlying mechanisms regulating differentiation timing are fully unknown. Here, we show a molecular mechanism determining differentiation timing from mouse embryonic stem cells (ESCs). Activation of protein kinase A (PKA) modulates differentiation timing to accelerate the appearance of mesoderm and other germ layer cells, reciprocally correlated with the earlier disappearance of pluripotent markers after ESC differentiation. PKA activation increases protein expression of G9a, an H3K9 methyltransferase, along with earlier H3K9 dimethylation and DNA methylation in Oct3/4 and Nanog gene promoters. Deletion of G9a completely abolishes PKA-elicited acceleration of differentiation and epigenetic modication. Furthermore, G9a knockout mice show prolonged expressions of Oct3/4 and Nanog at embryonic day 7.5 and delayed development. In this study, we demonstrate molecular machinery that regulates timing of multilineage differentiation by linking signaling with epigenetics.
INTRODUCTION Embryonic development is a highly orchestrated process, and the timing of cell differentiation from stem cells during development is strictly controlled. Nevertheless, the underlying molecular mechanism regulating differentiation timing remains unknown. Embryonic stem cells (ESCs), which are derived from the inner cell mass (ICM) of early blastocysts and are able to differentiate from a pluripotent state into all three germ layer populations in vitro, are one of the most potent tools for investigating cell-differentiation processes (Evans and Kaufman, 1981; Martin, 1981; Thomson et al., 1998). Even though numerous methods have been reported to induce ESC differentiation

in vitro, currently the timing of differentiated cell emergence from ESCs cannot be intentionally controlled. Cell differentiation is often compared to a ball rolling down an epigenetic landscape from the pluripotent state to lineage-committed states (Waddington, 1957; Yamanaka, 2009). Though the destination and route of the ball (i.e., nerve, cardiomyocyte, liver, etc.) has been intensively investigated, the timing of differentiationthat is, arrival time to the destinationhas not been highlighted and remains an open question. Histone methylation contributes to chromatin remodeling as well as transcriptional activity and operates the epigenetic landscape. Methylation of lysine 9 of histone H3 (H3K9) is a wellconserved epigenetic mark for heterochromatin formation and transcriptional silencing (Jenuwein, 2006). G9a (Ehmt2) is an H3K9 methyltransferase (mainly mono- and dimethylation, me1 and me2) of euchromatin (Tachibana et al., 2002; Tachibana et al., 2005). Recent studies on the mouse Oct3/4 gene, a pluripotent gene, have demonstrated that the postimplantation inactivation process of Oct3/4 is carried out in a multilayer manner that involves direct inhibition of transcription, heterochromatinization, through the methylation of H3K9 by G9a and subsequent DNA methylation (Feldman et al., 2006). Furthermore, G9a itself is capable of causing de novo DNA methylation to be independent of its histone methyltransferase activity by recruiting DNA methyltransferases Dnmt3a and Dnmt3b. These events should cause irreversible silencing of Oct3/4 in differentiated cell lineages (Epsztejn-Litman et al., 2008). G9a-decient mice displayed severe growth retardation and early lethality (Tachibana et al., 2002), indicating that cell differentiation would be affected by the inhibition of pluripotent factors through epigenetic modication in early differentiation. Nevertheless, the regulatory mechanisms of most epigenetic modiers, including G9a, and the involvement of them in differentiation kinetics are yet unknown. Previously, we developed an ESC differentiation system that recapitulates early cardiovascular development in vitro (Nishikawa et al., 1998; Yamashita et al., 2000; Yamashita et al., 2005). Flk1 (also known as vascular endothelial growth factor [VEGF] receptor-2) is one of the earliest differentiation markers for endothelial cells (ECs) and blood cells, and it is a marker of lateral plate mesoderm (Nishikawa et al., 1998; Yamashita et al., 2000). We induced Flk1+ cells from ESCs, and they were
Cell Stem Cell 10, 759770, June 14, 2012 2012 Elsevier Inc. 759

Cell Stem Cell


PKA-G9a for Differentiation Timing

then puried by uorescence-activated cell sorting (FACS) and recultured. We succeeded in inducing the major cardiovascular cell types from the common Flk1+ progenitor cells: vascular ECs, mural cells (pericytes and vascular smooth muscle cells) (Yamashita et al., 2000), and cardiomyocytes (Yamashita et al., 2005). We recently reported that the activation of cyclic adenosine monophosphate (cAMP)/protein kinase A (PKA) signaling in Flk1+ cells markedly enhanced EC differentiation (Yamamizu et al., 2009). cAMP/PKA signaling plays an important role beyond vasculature in virtually every known physiological action, such as metabolism, gene expression, cell division and growth, and cell differentiation. Adenylate cyclase generates cAMP from ATP in essentially all tissues of the body. This enzyme is activated by transmembrane receptors that are coupled to trimeric G-proteins (Howe, 2004). The effects of cAMP are mediated mainly by PKA. Mice decient in PKA catalytic subunits (both Ca and Cb) displayed early embryonic lethality around gastrulation (Huang et al., 2002), and in Drosophila, genetically induced mutations in the PKA catalytic subunits resulted in severe developmental defects (Lane and Kalderon, 1993; Lane and Kalderon, 1995), indicating that the PKA pathway is crucial in early development. However, the roles of the cAMP/PKA pathway in early development and stem cell differentiation are not fully elucidated. In this study, we attempt to elucidate roles of PKA signaling in early-stage cell differentiation. We demonstrate that activation of PKA regulates the timing of differentiation in ESCs and early embryos through epigenetic inhibition of pluripotent gene expression by G9a. RESULTS PKA Accelerates Differentiation Timing of the Three Germ Layers from ESCs To investigate the roles of the cAMP/PKA pathway in early differentiation from ESCs, we rst measured intracellular cAMP concentration during ESC differentiation. Interestingly, intracellular cAMP concentration signicantly increased from differentiation day-2.5 (D2.5) preceding the emergence of the three germ layers (around D3.5) (Figure 1A). We then tried to manipulate PKA signaling during ESC differentiation. We generated an ESC line carrying a constitutive active form of the PKA (CA-PKA) gene, which can be induced with the depletion of tetracycline (TetOff system) (Yamamizu et al., 2009), and activated the PKA pathway from the beginning of ESC differentiation (Figure 1B). In the control condition (Dox+), the appearance of Flk1+ mesoderm cells occurs from day 3.5 (D3.5) in our system. To our surprise, when PKA was activated (Dox) from D0, Flk1+ cells were able to be observed from D1.5 and became more prominent in the following 2 days (Figures 1C and 1D). A similar amount of Flk1+ cells were detected approximately two times faster with PKA activation compared to the control condition. In contrast, SSEA1+ cells, an undifferentiated ESC marker, were decreased signicantly from D1.5 by CA-PKA expression (Dox) (Figures 1C and 1E). Another mesoderm marker, platelet-derived growth factor receptor-a (PDGFRa), was also increased signicantly earlier with PKA activation (Figures 1F and 1G). Double immunostaining for Flk1 and SSEA1 similarly revealed an early decrease in SSEA1 and a reciprocal early appearance of Flk1.
760 Cell Stem Cell 10, 759770, June 14, 2012 2012 Elsevier Inc.

Flk1 expression was rst detected from D1.5 in the Dox condition, compared to D3.5 in the control condition. In contrast, SSEA1 expression completely disappeared at D3.5 in the Dox condition when it was still predominant in the control condition (Dox+) (Figure 1H). Immunostaining of brachyury, a mesendoderm marker, also showed an apparent earlier appearance with PKA activation than in the control condition (Figure 1I). RT-PCR and quantitative RT-PCR (qPCR) showed that emergence and peak expression of brachyury, Flk1, and PDGFRa mRNA were accelerated by PKA activation. Inversely, when endogenous PKA was blocked by treatment with a PKA inhibitor, PKI, the mRNA expression of these markers was delayed in comparison to the control condition (Figures 1J and 1K). These results indicate that PKA activity in early differentiating ESCs regulates differentiation timing from ESCs to mesoderm cells. Next, we examined endoderm and ectoderm induction from ESCs with PKA activation. Induction of Foxa2+ endoderm cells appeared from D3.5 in the control condition. In contrast, PKA activation signicantly increased an earlier appearance of Foxa2+ cells from D2.5 (Figures 2A and 2B). Nuclear expression of denitive endoderm markers, Foxa2 and Sox17, was observed in distinct populations at D1.5 and increased in the following 2 days under PKA activation (Dox), whereas these expressions were still sparse at D3.5 in the control condition (Figures 2C and 2D). RT-PCR and qPCR revealed an earlier emergence and peak expression of Foxa2 and Sox17 mRNA by PKA activation and an inverse delay in their expressions with PKI treatment (Figures 2E and 2F). Appearance of an ectoderm marker protein, Nestin, was also induced signicantly earlier than in the control condition by PKA activation (Figures 2G2I). mRNA expression kinetics of ectoderm markers, Nestin and Fgf5, was similarly modied in accordance with PKA activation or inhibition (Figures 2J and 2K). Thus, differentiation timing of ESCs to all three germ layer cells was positively regulated by PKA activity. As proliferation and differentiation are considered to be closely related to each other, we examined cell proliferation during ESC differentiation under PKA activation (Figure S1). Total cell counts were not different between the Dox+ condition and the Dox condition. In particular, the cell number was almost completely identical until D2.5, when early appearance of germ layer cells has already been achieved. EdU (an analog of BrdU) incorporation also showed that overall cell proliferation was not affected by PKA activation during ESC differentiation. Taken together, these results indicate that PKA should accelerate differentiation timing of the three germ layers from mouse ESCs with independent mechanisms from cell proliferation. PKA Activation Accelerates Disappearance of Pluripotent Markers after Differentiation Next, we examined the undifferentiated status of ESCs after differentiation induction under PKA activation. In the control condition (Dox+), alkaline phosphatase (AP)-negative differentiated colonies appeared from D3.5. In contrast, PKA activation (Dox) induced the appearance of AP-negative colonies from D1.5. They showed various cellular morphologies, such as endoderm- or mesoderm-like cells, in the following 2 days (Figures 3A

Cell Stem Cell


PKA-G9a for Differentiation Timing

Figure 1. PKA Accelerates Differentiation Timing of Mesodermal Cells from ESCs


(A) Intracellular cAMP concentration during ESC differentiation (n = 3, **p < 0.01 versus undifferentiated cells). (B) Experimental system for PKA activation. ESC lines expressing a constitutive active form of PKA (CA-PKA) via a tetracyclin-inducible expression system (Tet-Off) were established. (C) Time course of Flk1+ or SSEA1+ cell appearance evaluated by FACS. PKA was activated from the beginning of ESC differentiation. Upper panels: Dox treatment (1 mg/ml; Dox+; control). Lower panels: Dox free (Dox; PKA activation). Percentages of Flk1+ or SSEA+ cells in total cells are indicated. (D and E) Quantitative evaluation of effects of PKA activation on Flk1+ or SSEA1+ cell appearance during ESC differentiation. Percentages of the Flk1+ (D) or SSEA1+ (E) population (n = 3; *p < 0.05, **p < 0.01 versus Dox+) are indicated. (F) FACS analysis for PDGFRa+ cell appearance. Percentages of PDGFRa+ in total cells are indicated. (G) Quantitative evaluation of PDGFRa+ cell percentages (n = 3; **p < 0.01 versus Dox+). (H) Double immunostaining for Flk1 (green) and SSEA1 (red) during ESC differentiation. Upper panels: Dox+. Lower panels: Dox. Scale bar represents 200 mm. (I) Immunostaining for Brachyury (green). Scale bar represents 200 mm. (J and K) RT-PCR (J) and qPCR (K) showing mRNA expression of Brachyury, Flk1, and PDGFRa during ESC differentiation. Blue lines, Dox+ (control); red lines, Dox (PKA activation); green lines, Dox+ with PKA inhibitor, PKI, treatment (10 mM).

Cell Stem Cell 10, 759770, June 14, 2012 2012 Elsevier Inc. 761

Cell Stem Cell


PKA-G9a for Differentiation Timing

and 3B; Movies S1 and S2). Immunostaining for pluripotent ESC markers, Oct3/4 and Nanog, also revealed an earlier disappearance of these proteins with PKA activation (Figures 3C and 3D). mRNA expression of Oct3/4, Nanog, and Sox2 was decreased earlier with PKA activation than in the control condition (Figures 3E and 3F). On the contrary, inhibition of endogenous PKA with PKI treatment delayed the disappearance of these gene expressions. Next, we examined the effect of PKA activation on the maintenance of undifferentiated status in the presence of leukemia inhibitory factor (LIF). ESC growth was not affected by PKA activation. In addition, these cells maintained pluripotent gene expressions until 4 weeks after activation (Figures S2AS2C), indicating that PKA itself is not able to initiate ESC differentiation when the maintenance machinery is working. Taken together, these results indicate that PKA accelerates the disappearance of pluripotent genes after the exit from undifferentiated status but cannot initiate ESC differentiation. PKA Induces H3K9me2 and DNA Methylation through G9a in Early Differentiation from ESCs Our data show that PKA is broadly involved in many gene expressions during ESC differentiation. Thus, we hypothesized that the PKA pathway could be responsible for the regulation of epigenetic modications. We examined the overall status of several histone H3 methylations during ESC differentiation. Among those that we tested, H3K9me2 showed a remarkable increase from D1.5 with PKA activation (Figure 4A). We then examined the expression of an H3K9 dimethyltransferase, G9a, during ESC differentiation, and we found that G9a protein was signicantly increased in the very early stages of differentiation with PKA activation, from D0.5 to D2.5, which precedes H3K9me2 (Figures 4B and 4C). Early downregulation of pluripotent gene expression (i.e., Oct3/4, Nanog, and Sox2) and increase in an inhibitory epigenetic marking (H3K9me2) suggest epigenetic silencing of these pluripotent gene expressions via PKA. We then conrmed the epigenetic markings in the promoter regions of Oct3/4 and Nanog. A chromatin precipitation assay (ChIP) revealed that H3K9me2 was signicantly increased in the promoter regions of Oct3/4 and Nanog from D1.5 after CA-PKA expression (Figure 4D). Bisulfate sequence analysis showed that PKA activation apparently increased DNA methylation in the promoter regions of Oct3/4 and Nanog at D5.5 (Figures 4E and 4F). These results suggest that PKA increases G9a protein level in the early stages of differentiation, thereby enhancing H3K9me2 and subsequent DNA methylation in the promoter regions of Oct3/4 and Nanog.

Figure 2. PKA Accelerates Differentiation Timing of Endodermal and Ectodermal Cells from ESCs
(A and B) Quantitative evaluation of effects of PKA activation on Foxa2+ cell appearance during ESC differentiation. (A) FACS analysis for Foxa2+ cell appearance. Percentages of Foxa2+ in total cells are indicated. (B) Quantitative evaluation of Foxa2+ cell percentages (n = 3; *p < 0.05, **p < 0.01 versus Dox+). (C and D) Immunostaining for Foxa2 (green, C) and Sox17 (green, D) during ESC differentiation. Upper panels: Dox+. Lower panels: Dox. Scale bar represents 200 mm.

(E and F) RT-PCR (E) and qPCR (F) showing mRNA expression of Foxa2 and Sox17 during ESC differentiation. Blue lines, Dox+; red lines, Dox; green lines, Dox+ with PKI treatment. (G) FACS analysis for Nestin+ cell appearance. Percentages of Nestin+ in total cells are indicated. (H) Quantitative evaluation of Nestin+ cell percentages (n = 3; *p < 0.05, **p < 0.01 versus Dox+). (I) Immunostaining for Nestin (red). Scale bar represents 200 mm. (J and K) RT-PCR (J) and qPCR (K) showing mRNA expression of Nestin and Fgf5.

762 Cell Stem Cell 10, 759770, June 14, 2012 2012 Elsevier Inc.

Cell Stem Cell


PKA-G9a for Differentiation Timing

Figure 3. PKA Regulates Undifferentiated Status during Differentiation from ESCs


(A) Alkaline phosphatase (AP) staining (purple) during ESC differentiation. Upper panels: Dox+. Lower panel: Dox. Scale bar represents 100 mm. Arrow indicates the early appearance of APnegative cells. Arrowhead indicates cells with endodermal-like morphology. (B) Morphological evaluation of ESC colonies after PKA activation. Black bars, undifferentiated colonies; gray bars, partially differentiated colonies; white bars, totally differentiated colonies (n = 3). (C and D) Immunostaining for Oct3/4 (red, C) and Nanog (green, D) during ESC differentiation. Upper panels: Dox+. Lower panels: Dox. Scale bar represents 200 mm. (E and F) RT-PCR (E) and qPCR (F) showing mRNA expression of Oct3/4, Nanog, and Sox2 during ESC differentiation. Blue lines, Dox+; red lines, Dox; green lines, Dox+ with PKI treatment.

Effects of PKA on Early Differentiation Are Dependent on G9a To investigate whether regulations of differentiation timing by PKA are G9a-mediated effects, we established an ESC line carrying a dual inducible system combining tetracycline-regulatable CA-PKA expression and a tamoxifen (OHT)-inducible G9a conditional knockout (cKO) system by using a fusion protein (MerCreMer) of Cre protein and mutant estrogen receptor, which has a single copy of the G9a gene anked by loxP recombination sites (/ox) (Yokochi et al., 2009) (Figures 5A and S3A). The addition of an estrogen receptor ligand, 4-hydroxytamoxifen (OHT), induced the rapid deletion of G9a and subsequent H3k9me2 reduction (Figure 5B). Earlier PDGFRa+ mesoderm appearance induced by PKA activation (Dox) was almost completely abolished by the deletion of G9a (OHT+) (Figure 5C). Similarly, PKA-elicited early brachyury (Figure 5D) and Flk1 (data

not shown) appearances were abolished in the absence of G9a. qPCR performed on the mRNAs of these genes demonstrated that under the G9a-deleted condition (OHT+), mRNA expression showed delayed patterns similar to those shown under treatment with PKA inhibitor (Figure 1K), and more importantly, effects of PKA activation (Dox) were completely abolished (Figure 5E). G9a deletion also completely abolished PKA-elicited acceleration of differentiation to endoderm (foxa2 and sox17) and ectoderm (nestin and fgf5), examined with immunostaining and qPCR (Figures 5F5I). On the contrary, early disappearance of pluripotent genes induced by PKA activation was cancelled with G9a deletion (Figures 5J and 5K). Moreover, bisulfate sequence analysis revealed that G9a deletion abolished PKA-elicited DNA methylation on the promoter regions of Oct3/4 and Nanog (Figures 5L and 5M). These results indicate that the acceleration of differentiation timing induced by PKA is mediated by G9a. PKA and G9a Regulate Differentiation Timing In Vivo Finally, we examined the functional relevance of PKA and G9a in differentiation timing in vivo. First, we conrmed PKA expression in the early embryo. Immunostaining of mouse embryos revealed that the PKA catalytic subunit (PKAc) was highly expressed in metaphase II-(MII) oocytes and in the 1-cell (1C) stage of the zygote. PKAc once decreased from the 2-cell (2C) stage to the compact morula stage, and it intensied at blastocyst stages in both inner cell mass (ICM) and trophectoderm (TE) (Figure 6A). qPCR for PKAc further revealed that PKAc was highly expressed at the MII and 1C stages, once decreased at the 2C stage, and was reupregulated at blastocyst stages (Figure 6B), indicating
Cell Stem Cell 10, 759770, June 14, 2012 2012 Elsevier Inc. 763

Cell Stem Cell


PKA-G9a for Differentiation Timing

Figure 4. PKA Regulates G9a Expression Accompanied by H3K9me2 and DNA Methylation in Early Differentiation from ESCs
(A) Western blot for various histone methylations during ESC differentiation. Upper panels: Dox treatment (Dox+). Lower panel: Dox-free (Dox). (B) Western blot for G9a during ESC differentiation. (C) Quantitative evaluation of G9a protein expression. Relative expression normalized with b-actin expression is shown. Blue line, Dox+; red line, Dox (n = 3; **p < 0.01 versus Dox+). (D) ChIP assays for H3K9me2 on Oct3/4 (left panels) or Nanog (right panels) promoter regions. Lower panels show quantitative evaluation of H3K9me2 (n = 3; **p < 0.01 versus Dox+). (E and F) DNA methylation in the promoter regions of Oct3/4 (E) and Nanog (F) with bisulte genomic sequencing. Open circles, nonmethylated CpGs; closed circles, methylated CpGs.

that PKA should be activated in the early stages of development. We conrmed the functional signicance of PKA and G9a in the early differentiation of the mouse embryos with an early blastocyst culture assay (from embryonic day 3.25 [E3.25] to E4.75) in the presence of a PKA inhibitor, H89, or a G9a inhibitor, BIX 01294. The E3.25 embryos that were collected were at approximately the 32-cell stage and did not express Gata4, a primitive endoderm marker (data not shown). After 1.5 days of treatment
764 Cell Stem Cell 10, 759770, June 14, 2012 2012 Elsevier Inc.

in E3.25 embryos, immunostaining revealed that the addition of H89 (3 mM) or BIX01294 (0.5 mM) signicantly increased Nanog+ cells in compared to control embryos (Figures 6C and 6D). Whereas a clear array of Gata4+ primitive endoderm cells was observed in control embryos, H89 or BIX01294 treatment remarkably inhibited Gata4+ cell appearance (Figures 6E and 6F). Oct3/4+ cell number was not affected with the inhibitor treatments, which could be due to the fact that Oct3/4 was expressed in both ICM and primitive endoderm at that stage. The total cell number after BIX01294 treatment was decreased (Figures 6E and 6F). These results suggested that physiological PKA and G9a critically control the timing of early differentiation from ICM. We further conrmed the in vivo functional signicance of G9a in differentiation timing with G9a KO mice. We previously generated G9a KO mice that displayed severe growth retardation and died at E8.5E12.5 (Tachibana et al., 2002). Thus, we hypothesized that the loss of G9a induced prolongation of Oct3/4 and Nanog expression and perturbed differentiation in early embryos. We performed in situ hybridization for Oct3/4 and Flk1 in early embryos. At E6.5, Oct3/4 was highly expressed in the distal side of the embryos (Figure 6G). Wild-type (WT) E7.5 embryos showed only a trace of Oct3/4 expression (Figure 6H, arrow), but Flk1 was expressed in wedge-shaped areas on the proximal lateral sides (Figure 6H). G9a KO embryos at E7.5 showed growth-retarded embryonic morphology and sustained expressions of Oct3/4 in the whole embryo proper (Figure 6H, arrowheads) together with decreased Flk1 expression (Figure 6H) (detailed results are shown in Figure S4). qPCR using whole WT and G9a KO embryos at E7.5 further demonstrated the remarkable increase in Oct3/4 and Nanog mRNA in the E7.5

Cell Stem Cell


PKA-G9a for Differentiation Timing

G9a KO embryo and the signicantly reduced Flk1 mRNA expression (Figure 6I). Both Oct3/4 and Nanog mRNA expression were decreased from E8.5 to E9.5 in G9a KO embryos (data not shown). Taken together, these results indicate that PKA and G9a should be regulating the differentiation timing from early pluripotent cells in the embryo. DISCUSSION In this study, we demonstrated a molecular mechanism regulating differentiation timing. PKA signaling accelerates the differentiation timing of three germ layer cells through an H3K9 methyltransferase, G9a expression. G9a modied H3K9me2 in the promoter region of Oct3/4 and Nanog, thereby inhibiting Oct3/4 and Nanog transcription and in turn inducing earlier differentiation (Figure 7). Elucidation of these functional linkages from signaling (PKA) to epigenetic modication (G9a) to gene expression (Oct3/4 and Nanog) and nally to cell behavior (differentiation) would provide a valuable insight regarding the underlying machinery in the nely regulated stem cell differentiation and orchestrated embryonic development. The striking observations are the broad control of the differentiation of the three germ layers by PKA signaling. Although various growth factors such as BMPs, Wnts, and Fgfs play important functions in cell-fate determination, PKA is here identied as a signal that commonly regulates differentiation timing toward various cell fates. Currently, upstream signals involved in the activation of the cAMP/PKA pathway in this process are still unknown. We previously revealed that adrenomedullin (AM) is one of the regulatory factors activating cAMP/PKA signaling and enhances endothelial cell differentiation from Flk1+ vascular progenitors (Yurugi-Kobayashi et al., 2006). We examined whether AM could be an endogenous ligand that regulates differentiation timing. mRNAs from AM and its receptors, RAMP3 (receptor activity modifying protein 3) and CRLR (calcitonin receptor-like receptor), were detected in the early ESC differentiation stage. Treatment of ESCs with AM weakly but positively enhanced an early appearance of three germ layers and a reciprocal disappearance of pluripotent markers after ESC differentiation (Figures S5AS5L). Li et al. reported that preimplanting blastocysts at E3.5 expressed AM in ICM and trophectoderm cells (Li et al., 2006). We conrmed that AM was highly induced from the 8-cell stage during embryogenesis (Figure S5M), suggesting that an AM ligand-receptor system should exist and act in the early stages of development. Taken together, AM could be at least one of the endogenous ligands that activate cAMP/ PKA signaling for the regulation of differentiation timing in the early developmental stage. Though G9a was demonstrated to be downstream of PKA for differentiation timing, it is still unclear how G9a expression is regulated by PKA. We further investigated the relationship between PKA and G9a expression by examining the role of a major PKA downstream transcription factor, cAMP-response element binding protein (CREB). We generated an ESC line with an OHT-inducible dominant-negative form CREB (DNCREB) activation system (Figure S6A). Inhibition of CREB with OHT treatment did not affect the protein level of G9a (Figure S6B). Furthermore, qPCR showed that PKA activation did not affect the G9a mRNA level during ESC differentiation (Fig-

ure S6C), indicating that PKA does not regulate G9a expression through transcriptional regulation with CREB. APC/CCdh1 ubiquitin ligase was recently demonstrated to cause proteasomal degradation of G9a (Takahashi et al., 2012). To investigate whether APC/CCdh1 ubiquitin ligase is involved in the degradation of G9a during ESC differentiation, we performed overexpression of Cdh1, an activator of APC/C, in ESCs. Cdh1 expression decreased G9a protein expression at D1.5 in control (Dox+). When PKA was activated (Dox), G9a protein expression was increased and interestingly, the level of G9a protein was not decreased even with induction of Cdh1 expression (Figure S6D). Some reports have shown that the function of APC/CCdh1 ubiquitin ligase is inhibited by the cAMP/PKA pathway (Yamashita et al., 1996; Yamada et al., 1997), suggesting that PKA should inhibit the function of APC/CCdh1 ubiquitin ligase thereby decreasing degradation of G9a. These evidences suggest that PKA should increase G9a protein during ESC differentiation not through transcriptional regulation with CREB, but posttranslational regulation with inhibition of APC/CCdh1 ubiquitin ligase. Mouse ESCs are able to maintain pluripotency in the presence of LIF by the transcriptional regulatory circuit with genes such as Oct3/4, Nanog, and Sox2 (Nishiyama et al., 2009; Takahashi and Yamanaka, 2006; Walker et al., 2009; Arnold and Robertson, 2009; Ng and Surani, 2011). Given that PKA activation as well as G9a deletion did not affect the undifferentiated status of ESCs (Figure S2) in the maintenance condition with LIF, PKA and G9a cannot interfere or break down the pluripotent transcriptional circuit in the presence of LIF. Some protective mechanisms from epigenetic silencing of pluripotent genes should be working during the undifferentiated state (Ng and Surani, 2011). For example, H3K9 demethylases, Jmid1a and Jmid2c, were reported to prevent the accumulation of repressive methylation at the promoters of genes that maintain pluripotency of ESCs (Loh et al., 2007). After exit from such protective machinery in the undifferentiated state, G9a should be able to silence pluripotent genes through PKA activation. Though Oct3/4 but not Nanog was reported to be silenced by G9a (Feldman et al., 2006), PKA activation induced epigenetic silencing of Oct3/4 and Nanog, both of which were abolished by G9a deletion (Figures 4D4F, 5L, and 5M). Moreover, Nanog expression was increased by G9a inhibition or deletion in vivo (Figures 6C, 6D, and 6I). Taken together, these results indicate that G9a should be a key factor in the silencing of pluripotent genes at early differentiation stages. Here, we report that the machinery for differentiation timing consists of a direct molecular link between PKA signaling and epigenetic control and was independently regulated from that for cell-fate determination and cell division. Meshorer et al. reported that although undifferentiated ESCs have diffuse heterochromatin structure in the nuclei, differentiation processes cause reshaping of the global genomic architecture, particularly the reorganization of compact heterochromatin (Meshorer et al., 2006; Meshorer and Misteli, 2006). We speculate that PKA critically regulates the kinetics of compact heterochromatin formation through G9a expression (see Graphical Abstract). Of course, the regulatory mechanisms for differentiation timing would be variable depending on cell types and differentiation stages. Nevertheless, elucidation of this critical molecular linkage for
Cell Stem Cell 10, 759770, June 14, 2012 2012 Elsevier Inc. 765

Cell Stem Cell


PKA-G9a for Differentiation Timing

Figure 5. Effects of PKA on Early Differentiation Are Dependent on G9a


(A) Experimental system for PKA activation and loss of G9a. An ESC line expressing CA-PKA via a tetracyclin-inducible expression system (Tet-Off) together with a tamoxifen (OHT)-inducible G9a conditional knockout (cKO) system using a fusion protein (MerCreMer) of Cre protein and a mutant estrogen receptor, which has a single copy of the G9a gene anked by loxP recombination sites (/ox) (tetCA-PKA/G9a cKO cells). (B) Western blot for G9a and H3K9me2 during ESC differentiation (Dox+). Upper panels: control (OHT). Lower panels: OHT (150 ng/mL) treatment (OHT+; G9adeleted condition). (C) FACS analysis at D2.5 for PDGFRa+ cell appearance. Left panels: Dox treatment (Dox+). Right panels: Dox-free (Dox; PKA activation). Percentages of PDGFRa+ in total cells are indicated.

766 Cell Stem Cell 10, 759770, June 14, 2012 2012 Elsevier Inc.

Cell Stem Cell


PKA-G9a for Differentiation Timing

differentiation timing would provide a clue for understanding and dissecting how cell differentiation and developmental processes are synchronized and orchestrated. Shedding light on the timing of the cell differentiation and elucidating and manipulating the machinery at the molecular level would largely contribute to basic stem cell and developmental biology as well as, more broadly, to applied regenerative medicine.
EXPERIMENTAL PROCEDURES Antibodies Monoclonal antibody (MoAb) for murine Flk1 (AVAS12) was prepared as described previously (Nishikawa et al., 1998). MoAb for Oct3/4 was purchased from Santa Cruz (Santa Cruz, CA, USA). MoAb for Nestin was purchased from StemCell Technologies (Vancouver, Canada). MoAb for PDGFRa was purchased from eBioscience (San Diego, CA, USA). MoAb for b-actin was purchased from Sigma (St. Louis, MO, USA). MoAb for SSEA-1 and the polyclonal antibodies for Brachyury and Sox17 were purchased from R&D Systems (Minneapolis, MN, USA). The polyclonal antibody for Nanog was purchased from ReproCell (Tokyo, Japan). Polyclonal antibodies for Foxa2, trimethyl-Histone H3 (Lys9), and trimethyl-Histone H3 (Lys27) were purchased from Millipore (Billerica, MA, USA). Polyclonal antibodies for PKA catalytic subunit a, G9a, and trimethyl-Histone H3 (Lys4) were purchased from Cell Signaling (Danvers, MA, USA). The polyclonal antibody for dimethyl-Histone H3 (Lys9) was purchased from Abcam (Cambridge, UK). Polyclonal antibodies for Nestin and Foxa2 for FACS analysis were purchased from Bioss (Woburn, MA, USA). ESC Lines An ESC line carrying a tetracycline-regulatable constitutively active form of the PKA (CA-PKA) gene in EStTA-ROSA cells was generated as described previously (Yamamizu et al., 2009). To generate an ESC line carrying both CA-PKA and a 4-hydroxytamoxifen (OHT)-regulated G9a conditional KO system, we constructed a plasmid carrying the tetracycline transactivator (tTA) and the puromycin-resistant gene and both CA-PKA and red uorescent protein (RFP) under the control of the tetracycline-responsive-element-regulatable cytomegalovirus (BiTRE) promoter (tetCA-PKA) (Figure S3A). We introduced the tetCA-PKA plasmid into G9a conditional KO ESC line (Yokochi et al., 2009), using the mouse ESC Nucleofector Kit (Lonza, Basel, Swizerland) (Figures 5A and S3A). Cells were plated on 10 cm dishes containing 1 mg/ml doxycycline (Dox+). After 1 day, the medium was changed to Dox+ medium containing 1 mg/ml puromycin for selection of puromycin-resistant colonies (tetCA-PKA/G9a cKO cells). The induction of CA-PKA was checked by the expression of RFP. RFP+ cells were puried after transient treatment in the Dox condition for 2 days in the presence of LIF; then, puried cells were maintained again in the Dox+ condition (Figure S3B). RFP+ cells were successfully induced in more than 85% of total cells by Dox depletion during differentiation (Figure S3C). Cell Culture and Differentiation ESC lines were maintained as described elsewhere (Yamashita et al., 2000; Yamamizu et al., 2009; Yamamizu et al., 2010). Differentiation was induced in ESC lines with the use of differentiation medium (DM) [alpha minimal essential medium (MEM; GIBCO Grand Island, NY, USA) supplemented with 10% fetal calf serum (FCS; Japan Bioserum, Tokyo, Japan) and 5 3 105 M 2-mer-

captoethanol (2-ME; GIBCO)] as described previously (Yamashita et al., 2000; Yamamizu et al., 2009; Yamamizu et al., 2010). Treatment of 1 mg/ml Doxycycline (Dox) did not affect differentiation in control ESCs (EStTA-ROSA cells; Yamamizu et al., 2009), whereas the acceleration of three germ layer genes was observed similarly in another independent CA-PKA transgenic ESC clone (clone no. 2) (Figure S7). Differentiated cells were examined by immunostaining and ow cytometric analysis. FACS Analysis Cultured cells were harvested in DM at D1.5, D2.5, and D3.5 after differentiation and stained with combinations of APC-conjugated AVAS12 MoAb, PE-conjugated anti-SSEA MoAb (R&D Systems), PE-conjugated antiPDGFRa MoAb (eBioscince), or Biotin-conjugated anti-PDGFRa MoAb (eBioscience) followed by streptavidin-conjugated APC (BD PharMingen, San Diego, CA, USA), and they were then subjected to analysis with FACS Aria (Becton Dickinson, Franklin Lakes, NJ, USA). For intracellular proteins, cultured cells were xed with 4% paraformaldehyde and washed by PBS with 5% FCS and 0.75% Saponin (Sigma) (Uosaki et al., 2011). Fixed cells were stained with PE-conjugated anti-Foxa2 or Nestin antibody (Bioss) and then subjected to analysis with FACS Aria (Becton Dickinson). Immunohistochemistry and Alkaline Phosphatase Staining Immunostaining for cultured cells was carried out as described previously (Yamashita et al., 2000; Yamamizu et al., 2009; Yamamizu et al., 2010). In brief, 4%-paraformaldehyde-xed cells were blocked by 1% skim milk (BD Biosciences) and incubated overnight with primary antibodies at 4 C. For immunouorescence staining, anti-mouse, -rat, -rabbit, or -goat IgG antibodies conjugated with Alexa488 or Alexa546 (Invitrogen) were used as secondary antibodies. Nuclei were visualized with DAPI (Invitrogen). Stained cells were photographed with inverted uorescent microscopy (Eclipse TE2000-U, Nikon, Tokyo, Japan) and a digital camera system (AxioCam HRc) with the use of AxioVision 4.7.1 Software (Carl Zeiss, Jena, Germany). For oocytes and preimplantation embryo staining, oocytes and embryos at various developmental stages were xed with 4% paraformaldehyde, permeabilized for 30 min with 0.5% Triton X-100 in PBS, blocked for 1 hr in blocking buffer (1.5% BSA in PBS), and subsequently incubated overnight with primary antibodies at 4 C. For immunouorescence staining, anti-mouse or -rabbit IgG antibodies conjugated with Alexa488 or Alexa546 (Invitrogen) were used as secondary antibodies. Nuclei were visualized with Hoechst 33342 (SigmaAldrich). Fluorescence microscopy was used for the observation of uorescence (BZ-9000, Keyence, Osaka, Japan). For alkaline phosphatase staining, differentiated cells were xed with 4% paraformaldehyde, washed once in PBS, and stained with NBT/BCIP solution (Roche, Mannheim, Germany). Undifferentiated ESC colonies, partially differentiated cell colonies, and totally differentiated cell colonies were evaluated by AP staining. Undifferentiated indicates AP-positive cell only, partially differentiated indicates a mixture of AP-positive and -negative cells, and totally differentiated indicates AP-negative cell only. Measurement of Intracellular cAMP Undifferentiated ESCs or differentiated cells were lysed in lysis buffer. We measured protein concentration with the Qubit Protein Assay Kit (Invitrogen). The intracellular cAMP concentration was evaluated in protein (1 ng) with the cAMP-Screen System (Applied Biosystem), according to the manufacturers instructions. Luminescence was measured with a microplate reader (ARVO MX; PerkinElmer, Waltham, MA, USA).

(D) Immunostaining for Brachyury (green) at D2.5. (E) qPCR showing Brachyury, Flk1, or PDGFRa mRNA expression with OHT treatment during ESC differntiation. Blue lines, Dox+; red lines, Dox. (F) Immunostaining for Foxa2 (green) at D2.5. (G) qPCR showing Foxa2 and Sox17 mRNA expression with OHT treatment. (H) Immunostaining for Nestin at D2.5. (I) qPCR showing Nestin and Fgf5 mRNA expression with OHT treatment. (J) Immunostaining for Oct3/4 (green) at D3.5. (K) qPCR showing Oct3/4, Nanog, or Sox2 mRNA expression with OHT treatment. (L and M) DNA methylation in the promoter regions of Oct3/4 (L) and Nanog (M) at D5.5. Scale bars represent 200 mm.

Cell Stem Cell 10, 759770, June 14, 2012 2012 Elsevier Inc. 767

Cell Stem Cell


PKA-G9a for Differentiation Timing

Figure 6. Inhibition of PKA- or G9a-Induced Sustained Pluripotent Gene Expression and Delayed Development In Vivo
(A) Immunostaining for PKA catalytic subunit (PKAc) (green) in oocytes and preimplantation embryos. (B) qPCR showing PKAc mRNA expression during early embryogenesis. (C and D) Effects of H89 (3 mM) or BIX01294 (0.5 mM) on Nanog expression in early ex vivo embryo culture (1.5 days from E3.25 stage). (C) Immunostaining for Nanog (green) and DAPI (blue). Scale bar represents 25 mm. (D) Quantitative evaluation of Nanog+ cell percentages in total cells (control; n = 11, H89; n = 15, BIX01294; n = 13, *p < 0.05 versus control). (E and F) Effects of H89 or BIX01294 on the primitive endoderm development in ex vivo culture. (E) Double immunostaining for Oct3/4 (red) and Gata4 (green). Note that a clear array of GATA4+ primitive endoderm is observed in controls. Oct3/4 is positive for both ICM and primitive endoderm. Scale bar represents 25 mm. (F) Quantitative evaluation of Oct3/4+ or Gata4+ cell percentages in total cells (upper and middle panels) and total cell number in each embryo (lower panel) (control; n = 12, H89; n = 13, BIX01294; n = 12, *p < 0.05, **p < 0.01 versus control). (G) in situ hybridization for Oct3/4 at E6.5 in a WT embryo (purple). Scale bars represent 500 mm (left) or 100 mm (right). (H) In situ hybridization for Oct3/4 or Flk1 (purple) in E7.5 embryos. Upper panels: WT. Lower panels: G9a KO embryos. Arrow indicates a trace Oct3/4 expression in WT. Arrowheads indicate sustained Oct3/4 expression in G9a KO embryo. Scale bars represent 100 mm. (I) qPCR showing G9a, Oct3/4, Nanog, or Flk1 mRNA expression in E7.5 embryos (n = 3 from independent pregnant mice, **p < 0.01 versus WT).

an alkaline phosphate color substrate. The sections were counterstained with Kernechtrot (Muto Pure Chemicals, Tokyo, Japan). Stained cells were photographed with an inverted uorescence microscope (Leica DM2500; Leica, Tokyo, Japan) and a digital camera system (COOLPIX995; Nikon). Western Blotting Western blotting was performed as previously described (Yamamizu et al., 2009; Yamamizu et al., 2010). In brief, undifferentiated and differentiated cells were lysed in lysis buffer, and the samples were run on SDS-PAGE with the use of gradient gel (Atto, Tokyo, Japan), followed by electrophoretic transfer onto nitrocellulose membranes. After the blots were incubated for 1 hr in the blocking agent Blocking One (Nacalai Tesque, Kyoto, Japan), they were incubated overnight with the respective primary antibodies at 4 C. Anti-mouse or -rabbit IgG antibodies conjugated with Horseradish peroxidase (HRP) was used as secondary antibodies (1:10,000). The Can Get Signal Immunoreaction Enhancer Solution Kit (Toyobo, Osaka, Japan) was used for signal enhancement. Immunoreactivity was detected with the enhanced chemiluminescence kit Chemi-Lumi One (Nacalai Tesque). Signal intensity was calculated with Scion Image software (Scion, Frederick, MD, USA). Chromatin Immunoprecipitation Assay Immunostaining for cultured cells was carried out as described previously (Yamamizu et al., 2010). In brief, after 1.5, 2.5, and 3.5 days, differentiatied cells were subjected to crosslinking with 3.7% formaldehyde, followed by ChIP

In Situ Hybridization A DNA fragment corresponding to a nucleotide sequence in mouse Oct3/4 (GenBank accession number NM_013633) or Flk1 gene (GenBank accession number NM_010612) was subcloned into pGEMT-Easy vector (Promega) and was used for generation of sense or antisense RNA probes (Table S1). Parafn-embedded embryo sections (6 mm) of C57BL/6 mouse (Sankyo Labo Service) or G9a KO mice were obtained from Genostaff (Tokyo, Japan). For in situ hybridization, the sections were hybridized with digoxigenin-labeled RNA probes at 60 C for 16 hr. The bound label was detected with NBT-BCIP,

768 Cell Stem Cell 10, 759770, June 14, 2012 2012 Elsevier Inc.

Cell Stem Cell


PKA-G9a for Differentiation Timing

RNA Isolation, RT-PCR, and qPCR Total RNA was isolated from undifferentiated ESCs, differentiated cells, oocytes, fertilized embryos before implantation, and E7.5 embryos with the use of RNeasy (QIAGEN) according to the manufacturers instructions. Reverse transcription was performed with the SuperScript III First-Strand Synthesis System (Invitrogen). qPCR was performed with Power SYBR Green PCR Master Mix (Applied Biosystems, San Diego, CA) and a StepOnePlus system (Applied Biosystems). The amount of target RNA was determined from the appropriate standard curve and normalized relative to the amount of Gapdh mRNA. Primer sequences are shown in Table S3. Collection and Culture of Oocytes and Embryos All animal experiments were performed in accordance with the guidelines for Animal Experiments of Kyoto University, which conform to the Guide for the Care and Use of Laboratory Animals in Japan. Mouse oocytes and preimplantation embryos were collected as previously described (Tsukamoto et al., 2006). In brief, 6- or 7-week-old female Crj; CD-1 (ICR) mice (Shimizu Laboratory Supplies, Kyoto, Japan) were superovulated by intraperitoneal (i.p.) injection of 5 IU eCG (ASKA Pharmaceutical, Tokyo, Japan), followed 46 hr later by 5 IU hCG (Sankyo Zoki, Tokyo, Japan). Unfertilized metaphase II (MII) oocytes were collected 16 hr after hCG from the oviductal ampullae. Fertilized embryos were recovered at 20 hr (1-cell stage) after hCG injection from the oviductal ampullae, or at E3.25 from the uterus, respectively. Collected embryos were cultured in vitro in KSOM media (EmbryoMax, Millipore) to the blastocyst stage for immunohistochemical staining. The E3.25 embryos were cultured in KSOM media with various reagents, PKA inhibitor, H89 and G9a inhibitor, and BIX01294 for 1.5 days. After 1.5 days, cultured embryos were examined by immunohistochemistry. G9a Knockout Mice Generation of the G9a KO mice was carried out as described previously (Tachibana et al., 2002). Mice were allowed to mate naturally at night. E0.5 was considered to be noon on the day a vaginal plug was observed. Embryos at E6.5 and E7.5 were subjected to in situ hybridization and qPCR. Statistical Analysis At least three independent experiments were performed. Statistical analysis of the data was performed with an ANOVA; p < 0.05 was considered signicant. Values are reported as means SEM. SUPPLEMENTAL INFORMATION Supplemental Information includes Supplemental Experimental Procedures, seven gures, three tables, and two movies and can be found with this article online at doi:10.1016/j.stem.2012.02.022. ACKNOWLEDGMENTS We are grateful to Dr. Shinji Kuninaka and Dr. Hideyuki Saya for the Cdh1 plasmid. We are also grateful to Dr. Shin Yonehara for the DN-CREB plasmid. We thank Dr. Meiko Takahashi (Kyoto University, Kyoto, Japan) for a critical reading of the manuscript. This study was supported by grants from the Ministry of Education, Science, Sports and Culture of Japan; the Ministry of Health, Labour and Welfare of Japan; the Project for Realization of Regenerative Medicine; the Japan Society for the Promotion of Science, and a Japan Heart Foundation Young Investigators Research Grant. K.Y. performed all experiments and wrote the manuscript; M.F. and H.I. helped with embryo culture; M.T. and Y.S. helped with studies of G9a-decient mice; S.I. helped with qRT-PCR and editing of gures; A.T., and E.H. supervised studies for G9a and Cdh1. H.I. and Y.S. supervised studies for embryo culture and G9a, respectively. J.K.Y. supervised all experiments and wrote the manuscript. Received: September 2, 2011 Revised: January 20, 2012 Accepted: February 24, 2012 Published: June 14, 2012

Figure 7. Molecular Linkages Regulating Differentiation Timing in Early ESC Differentiation and Embryogenesis
PKA signaling accelerates differentiation timing from mouse ESCs and early embryos through regulation of G9a, H3K9 methyltransferase. AM is one of the endogenous ligands that activates cAMP/PKA signaling for the regulation of ESC differentiation timing. PKA increases G9a protein during ESC differentiation through posttranslational regulation with the inhibition of APC/CCdh1 ubiquitin ligase. G9a modies H3K9me2 and DNA methylation in the promoter region of Oct3/4 and Nanog, thereby inhibiting Oct3/4 and Nanog expression. This direct molecular linkage of signaling molecules (PKA signaling), epigenetic modiers (G9a), gene expressions (Oct3/4 and Nanog), and cell behavior (cell differentiation) forms a machinery regulating cell-differentiation timing in early ESC differentiation and in vivo embryonic development. assay via a ChIP Assay Kit (Millipore). Chromatin was sheared to an average length of 0.41.0 kb. Antibodies to dimethyl-Histone H3 (Lys9) (Abcam) were used for immunoprecipitation. Sets of primers were used to amplify DNA sequences (Table S2). PCR amplication was conducted with a variable number of cycles (94 C for 30 s, 60 C for 30 s, and 72 C for 30 s). Bisulte Genomic Sequencing Total DNA was isolated from undifferentiated ESCs and differentiated cells with the use of the DNeasy Tissue Kit (QIAGEN, Valencia, CA, USA) according to the manufacturers instructions. Bisulte treatment was performed with the CpGenome Modication Kit (Chemicon) according to the manufactuers recommendations. PCR products were gel puried with the QIAquick Gel Extraction Kit (QIAGEN) and cloned into T-Vector (Novagen, Darmstadt, Germany) with Ligation high (Toyobo). Ten random clones were picked and sequenced with T7 or U-19-mer primer. PCR primers are listed in Table S2 (Imamura et al., 2006; Takahashi and Yamanaka, 2006).

Cell Stem Cell 10, 759770, June 14, 2012 2012 Elsevier Inc. 769

Cell Stem Cell


PKA-G9a for Differentiation Timing

REFERENCES Arnold, S.J., and Robertson, E.J. (2009). Making a commitment: cell lineage allocation and axis patterning in the early mouse embryo. Nat. Rev. Mol. Cell Biol. 10, 91103. Epsztejn-Litman, S., Feldman, N., Abu-Remaileh, M., Shufaro, Y., Gerson, A., Ueda, J., Deplus, R., Fuks, F., Shinkai, Y., Cedar, H., and Bergman, Y. (2008). De novo DNA methylation promoted by G9a prevents reprogramming of embryonically silenced genes. Nat. Struct. Mol. Biol. 15, 11761183. Evans, M.J., and Kaufman, M.H. (1981). Establishment in culture of pluripotential cells from mouse embryos. Nature 292, 154156. Feldman, N., Gerson, A., Fang, J., Li, E., Zhang, Y., Shinkai, Y., Cedar, H., and Bergman, Y. (2006). G9a-mediated irreversible epigenetic inactivation of Oct-3/4 during early embryogenesis. Nat. Cell Biol. 8, 188194. Howe, A.K. (2004). Regulation of actin-based cell migration by cAMP/PKA. Biochim. Biophys. Acta 1692, 159174. Huang, Y., Roelink, H., and McKnight, G.S. (2002). Protein kinase A deciency causes axially localized neural tube defects in mice. J. Biol. Chem. 277, 19889 19896. Imamura, M., Miura, K., Iwabuchi, K., Ichisaka, T., Nakagawa, M., Lee, J., Kanatsu-Shinohara, M., Shinohara, T., and Yamanaka, S. (2006). Transcriptional repression and DNA hypermethylation of a small set of ES cell marker genes in male germline stem cells. BMC Dev. Biol. 6, 34. Jenuwein, T. (2006). The epigenetic magic of histone lysine methylation. FEBS J. 273, 31213135. Lane, M.E., and Kalderon, D. (1993). Genetic investigation of cAMP-dependent protein kinase function in Drosophila development. Genes Dev. 7 (7A), 12291243. Lane, M.E., and Kalderon, D. (1995). Localization and functions of protein kinase A during Drosophila oogenesis. Mech. Dev. 49, 191200. Li, M., Yee, D., Magnuson, T.R., Smithies, O., and Caron, K.M. (2006). Reduced maternal expression of adrenomedullin disrupts fertility, placentation, and fetal growth in mice. J. Clin. Invest. 116, 26532662. Loh, Y.H., Zhang, W., Chen, X., George, J., and Ng, H.H. (2007). Jmjd1a and Jmjd2c histone H3 Lys 9 demethylases regulate self-renewal in embryonic stem cells. Genes Dev. 21, 25452557. Martin, G.R. (1981). Isolation of a pluripotent cell line from early mouse embryos cultured in medium conditioned by teratocarcinoma stem cells. Proc. Natl. Acad. Sci. USA 78, 76347638. Meshorer, E., and Misteli, T. (2006). Chromatin in pluripotent embryonic stem cells and differentiation. Nat. Rev. Mol. Cell Biol. 7, 540546. Meshorer, E., Yellajoshula, D., George, E., Scambler, P.J., Brown, D.T., and Misteli, T. (2006). Hyperdynamic plasticity of chromatin proteins in pluripotent embryonic stem cells. Dev. Cell 10, 105116. Ng, H.H., and Surani, M.A. (2011). The transcriptional and signalling networks of pluripotency. Nat. Cell Biol. 13, 490496. Nishikawa, S.I., Nishikawa, S., Hirashima, M., Matsuyoshi, N., and Kodama, H. (1998). Progressive lineage analysis by cell sorting and culture identies FLK1+VE-cadherin+ cells at a diverging point of endothelial and hemopoietic lineages. Development 125, 17471757. Nishiyama, A., Xin, L., Sharov, A.A., Thomas, M., Mowrer, G., Meyers, E., Piao, Y., Mehta, S., Yee, S., Nakatake, Y., et al. (2009). Uncovering early response of gene regulatory networks in ESCs by systematic induction of transcription factors. Cell Stem Cell 5, 420433. Tachibana, M., Sugimoto, K., Nozaki, M., Ueda, J., Ohta, T., Ohki, M., Fukuda, M., Takeda, N., Niida, H., Kato, H., and Shinkai, Y. (2002). G9a histone methyltransferase plays a dominant role in euchromatic histone H3 lysine 9 methylation and is essential for early embryogenesis. Genes Dev. 16, 17791791.

Tachibana, M., Ueda, J., Fukuda, M., Takeda, N., Ohta, T., Iwanari, H., Sakihama, T., Kodama, T., Hamakubo, T., and Shinkai, Y. (2005). Histone methyltransferases G9a and GLP form heteromeric complexes and are both crucial for methylation of euchromatin at H3-K9. Genes Dev. 19, 815826. Takahashi, K., and Yamanaka, S. (2006). Induction of pluripotent stem cells from mouse embryonic and adult broblast cultures by dened factors. Cell 126, 663676. Takahashi, A., Imai, Y., Yamakoshi, K., Kuninaka, S., Ohtani, N., Yoshimoto, S., Hori, S., Tachibana, M., Anderton, E., Takeuchi, T., et al. (2012). DNA damage signaling triggers degradation of histone methyltransferases through APC/C(Cdh1) in senescent cells. Mol. Cell 45, 123131. Thomson, J.A., Itskovitz-Eldor, J., Shapiro, S.S., Waknitz, M.A., Swiergiel, J.J., Marshall, V.S., and Jones, J.M. (1998). Embryonic stem cell lines derived from human blastocysts. Science 282, 11451147. Tsukamoto, S., Ihara, R., Aizawa, A., Kishida, S., Kikuchi, A., Imai, H., and Minami, N. (2006). Oog1, an oocyte-specic protein, interacts with Ras and Ras-signaling proteins during early embryogenesis. Biochem. Biophys. Res. Commun. 343, 11051112. Uosaki, H., Fukushima, H., Takeuchi, A., Matsuoka, S., Nakatsuji, N., Yamanaka, S., and Yamashita, J.K. (2011). Efcient and scalable purication of cardiomyocytes from human embryonic and induced pluripotent stem cells by VCAM1 surface expression. PLoS ONE 6, e23657. Waddington, C.H. (1957). The Strategy of Genes: A Discussion of Some Aspects of Theoretical Biology (London: George Allen & Unwin), pp. 262. Walker, M.R., Patel, K.K., and Stappenbeck, T.S. (2009). The stem cell niche. J. Pathol. 217, 169180. Yamada, H., Kumada, K., and Yanagida, M. (1997). Distinct subunit functions and cell cycle regulated phosphorylation of 20S APC/cyclosome required for anaphase in ssion yeast. J. Cell Sci. 110, 17931804. Yamamizu, K., Kawasaki, K., Katayama, S., Watabe, T., and Yamashita, J.K. (2009). Enhancement of vascular progenitor potential by protein kinase A through dual induction of Flk-1 and Neuropilin-1. Blood 114, 37073716. Yamamizu, K., Matsunaga, T., Uosaki, H., Fukushima, H., Katayama, S., Hiraoka-Kanie, M., Mitani, K., and Yamashita, J.K. (2010). Convergence of Notch and beta-catenin signaling induces arterial fate in vascular progenitors. J. Cell Biol. 189, 325338. Yamanaka, S. (2009). Elite and stochastic models for induced pluripotent stem cell generation. Nature 460, 4952. Yamashita, Y.M., Nakaseko, Y., Samejima, I., Kumada, K., Yamada, H., Michaelson, D., and Yanagida, M. (1996). 20S cyclosome complex formation and proteolytic activity inhibited by the cAMP/PKA pathway. Nature 384, 276279. Yamashita, J., Itoh, H., Hirashima, M., Ogawa, M., Nishikawa, S., Yurugi, T., Naito, M., Nakao, K., and Nishikawa, S. (2000). Flk1-positive cells derived from embryonic stem cells serve as vascular progenitors. Nature 408, 9296. Yamashita, J.K., Takano, M., Hiraoka-Kanie, M., Shimazu, C., Peishi, Y., Yanagi, K., Nakano, A., Inoue, E., Kita, F., and Nishikawa, S. (2005). Prospective identication of cardiac progenitors by a novel single cell-based cardiomyocyte induction. FASEB J. 19, 15341536. Yokochi, T., Poduch, K., Ryba, T., Lu, J., Hiratani, I., Tachibana, M., Shinkai, Y., and Gilbert, D.M. (2009). G9a selectively represses a class of late-replicating genes at the nuclear periphery. Proc. Natl. Acad. Sci. USA 106, 1936319368. Yurugi-Kobayashi, T., Itoh, H., Schroeder, T., Nakano, A., Narazaki, G., Kita, F., Yanagi, K., Hiraoka-Kanie, M., Inoue, E., Ara, T., et al. (2006). Adrenomedullin/cyclic AMP pathway induces Notch activation and differentiation of arterial endothelial cells from vascular progenitors. Arterioscler. Thromb. Vasc. Biol. 26, 19771984.

770 Cell Stem Cell 10, 759770, June 14, 2012 2012 Elsevier Inc.

Cell Stem Cell

Article
Self-Formation of Optic Cups and Storable Stratied Neural Retina from Human ESCs
Tokushige Nakano,1,2,4,5 Satoshi Ando,1,2,4 Nozomu Takata,1 Masako Kawada,1 Keiko Muguruma,1 Kiyotoshi Sekiguchi,6 Koichi Saito,4 Shigenobu Yonemura,3 Mototsugu Eiraku,1,2 and Yoshiki Sasai1,2,5,*
and Neurogenesis Group of Human Stem Cell Technology 3Electron Microscopy Laboratory RIKEN Center for Developmental Biology, Kobe 650-0047, Japan 4Environmental Health Science Laboratory, Sumitomo Chemical Co., Ltd., Osaka 554-8558, Japan 5Department of Medical Embryology, Graduate School of Medicine, Kyoto University, Kyoto 606-8501, Japan 6Laboratory of Extracellular Matrix Biochemistry, Institute for Protein Research, Osaka University, Suita 565-0871, Japan *Correspondence: yoshikisasai@cdb.riken.jp DOI 10.1016/j.stem.2012.05.009
2Division 1Organogenesis

SUMMARY

In this report, we demonstrate that an optic cup structure can form by self-organization in human ESC culture. The human ESC-derived optic cup is much larger than the mouse ESC-derived one, presumably reecting the species differences. The neural retina in human ESC culture is thick and spontaneously curves in an apically convex manner, which is not seen in mouse ESC culture. In addition, human ESC-derived neural retina grows into multilayered tissue containing both rods and cones, whereas cone differentiation is rare in mouse ESC culture. The accumulation of photoreceptors in human ESC culture can be greatly accelerated by Notch inhibition. In addition, we show that an optimized vitrication method enables en bloc cryopreservation of stratied neural retina of human origin. This storage method at an intermediate step during the time-consuming differentiation process provides a versatile solution for quality control in large-scale preparation of clinical-grade retinal tissues.

INTRODUCTION Complex organs such as the eye consist of many different kinds of cells that are integrated in a spatially restricted manner. Much progress has been made over the last decade in understanding the molecular mechanism of cellular differentiation, enabling in vitro generation of retinal cells, including photoreceptors and retinal pigment epithelium, from pluripotent stem cells (Boucherie et al., 2011; Lund et al., 2006; Lu et al., 2009a; Meyer et al., 2011; Osakada et al., 2008; Idelson et al., 2009; Lamba et al., 2006, 2010; Carr et al., 2009; Ikeda et al., 2005; Kawasaki et al., 2002). In contrast, little is known about the control of the complex 3D shape and pattern in retinal development.

We recently reported the formation of a 3D retina from mouse ESC (mESC) aggregates (Eiraku et al., 2011) using a versatile oating culture in serum-free and growth-factor-reduced medium (SFEBq culture, or serum-free culture of embryoid body-like aggregates with quick aggregation; Wataya et al., 2008; Eiraku et al., 2008). In SFEBq culture, mESCs spontaneously form a hollow vesicle of continuous neuroepithelium that differentiate into rostral forebrain tissues that include the retina-forming eld. In particular, a modied SFEBq culture using Matrigel in low-growth-factor medium efciently steers mESCs not only to differentiate into retinal progenitors, but also to form a two-walled cup-like structure mimicking the embryonic optic cup (Eiraku et al., 2011) (Figures S1A and S1B available online). Importantly, the optic cup formation is driven by self-organization; it spontaneously occurs via an intrinsic program of orchestrated local cellular interactions in a simple culture starting with the aggregation of mESCs in homogenous culture medium. During early embryogenesis, the retinal anlage rst appears as the optic vesicle, an epithelial vesicle evaginating laterally from the diencephalon. Subsequently, its distal portion invaginates to form the optic cup. The outer wall develops into retinal pigment epithelium (RPE), while the inner wall forms the neurosensory layers called neural retina (NR) (Graw, 2010). The optic cup generated in the modied SFEBq culture has RPE and NR epithelium with the correct topology, and the shape of the cup develops without the inuence of external structures (Eiraku et al., 2011). Furthermore, the mESC-derived NR epithelium, when isolated and cultured in suspension for additional 2 weeks, grows into a stratied NR tissue with six kinds of NR cells (photoreceptors, bipolar cells, ganglion cells, horizontal cells, amacrine cells, and ller cells) aligned in multiple layers as seen in the postnatal Mu eye (Eiraku et al., 2011). The formation of the stratied NR tissue is another example of self-organization in mESC culture; this multilayered structure spontaneously emerges from monolayered NR epithelium. Thus, this SFEBq culture enables the self-formation of the highly ordered 3D retinal pattern from patternless aggregates of mESCs in vitro (Eiraku et al., 2011; Ali and Sowden, 2011).
Cell Stem Cell 10, 771785, June 14, 2012 2012 Elsevier Inc. 771

Cell Stem Cell


Optic Cup and 3D Retina Self-Forming from hESCs

In the present study, we demonstrate that the optic cup structure and stratied NR also form from hESCs in a self-organizing manner. We also discuss some intriguing differences between mESC- and hESC-derived retinal tissues in the size, epithelial deformation mechanics, rod-versus-cone differentiation, and developmental speed. RESULTS Efcient Generation of Retinal Progenitors from hESCs in 3D Culture In mESC culture, self-organized optic cup formation was most frequently seen when the culture contained a high percentage of retinal progenitors (typically 35%70%). We therefore optimized retinal differentiation using hESC lines with venus cDNA knocked in at the locus of the early retina marker gene Rx (also called RAX; Mathers et al., 1997) (Figures S1C and S1D). SFEBq culture was started by reaggregating dissociated hESCs in a low-cell-adhesion 96-well plate (9,000 cells/well) and was continued in serum-free differentiation medium (Figure S1E). Unlike mESCs, reaggregation of hESCs is slow and often incomplete when they are cultured in serum-free medium using a conventional U-bottomed well; in such conditions, the aggregates tend to vary in size, causing substantial variance in differentiation efciency. This problem was largely solved by using a newly designed V-bottomed 96-well plate (conical-shaped; see Experimental Procedures), in which the aggregation was facilitated by the pointed bottom (Figure S1F). One key for the efcient retinal differentiation in our previous mESC study was to reduce the concentration of knockout serum replacement additive (KSR; usually used at 10%20%) in the medium to 1.5%2%. However, hESCs inefciently formed retinal epithelium at such low KSR concentrations in SFEBq culture (data not shown). A high KSR concentration tends to cause moderate caudalization in ESC-derived neural progenitors (data not known); this is not helpful for retinal differentiation, which occurs in the rostral diencephalon in vivo. Therefore, we counteracted the caudalizing activity of high KSR by treating the culture with a Wnt inhibitor (Lu et al., 2009b), which is known to have a rostralizing effect, during the early culture phase. In addition, to avoid dissociation-induced apoptosis, we added a ROCK inhibitor (Y-27632) to the SFEBq culture of hESCs, as we reported previously (Watanabe et al., 2007; Ohgushi et al., 2010). Thus, we cultured hESC aggregates in 20% KSR-containing medium supplemented with the Wnt inhibitor IWR1-endo (IWR1e), Y-27632, and Matrigel for the rst 12 days (Figure 1A). Then, we examined the effects of various signaling modulators on retinal differentiation by adding them to medium on day 12. One effective enhancer of retinal differentiation turned out to be fetal bovine serum (FBS; 10%) (Figure S1H). In addition, in the presence of FBS, treatment with the Hedgehog agonist smoothened agonist (SAG; 100 nM) augmented retinal differentiation in SFEBq culture of hESCs such that the proportion of Rx::venus+ cells reached >70% of total cells (Figures 1B, 1C, and S1G; the majority of venus tissues were N-cadherin+ and Sox2+ neural progenitors; Figures S1I and S1J).
772 Cell Stem Cell 10, 771785, June 14, 2012 2012 Elsevier Inc.

Self-Formation of Optic Cups in 3D hESC Culture Importantly, most of the retinal epithelium generated under the Matrigel+IWR1e+FBS+SAG (MIFS) conditions was positive for Chx10 (NR progenitor marker; Liu et al., 1994) and Pax6, and nez-Morales et al., negative for the RPE marker Mitf (Mart 2004) (Figures 1D1K). This indicates that the induced retinal epithelium preferentially differentiated into NR epithelia in this culture (NR-selective culture, hereafter). We therefore searched for RPE-generating culture conditions. We rst examined the effect of prolonged treatment with FBS (days 1221), which is known to support the growth of RPE (Kawasaki et al., 2002), but did not observe any substantial RPE-inducing effects (data not shown; instead, prolonged FBS treatment enhanced the formation of continuous NR epithelia; Figures 1L and 1M). In contrast, treatment with the Wnt agonist CHIR99021 (GSK3b inhibitor) during days 1821 efciently induced the formation of large thin epithelia expressing Mitf (Figures 1N and 1O; indeed, further culture of these aggregates developed RPE with massive pigmentation; Figures 1P and 1Q) and inhibited NR differentiation (data not shown). The RPEinducing effect of Wnt signaling (when added after the cells are committed to the retinal fate) is consistent with previous reports of in vivo mouse studies (Fuhrmann, 2008; Westenskow et al., 2009; Liu et al., 2010; Fujimura et al., 2009) and our previous in vitro mESC study. The formation of the optic cup structure needs both NR and RPE tissues in a balanced fashion. For this balance, the temporal window of the Wnt agonist treatment was critical, and the addition of CHIR99021 to medium during days 1518 substantially increased Mitf expression, without suppressing Chx10 expression (Figures 2A, 2B, and S2A). Under these conditions (SFEBq/MIFS+W, hereafter), Rx::venus+ epithelia appeared as patches on day 12 and subsequently underwent characteristic retinal morphogenesis, evaginating to form optic-vesicle-like structures (found in 58%73% of aggregates; four experiments, 96 aggregates per experiment; Figures 2C and S2B and Movie S1). The Chx10+ distal portion of the hESC-derived retinal epithelium at this stage already exhibited a morphological characteristic of early pseudostratied epithelium (Figure 2D). Live imaging showed that the cells in the distal epithelium of the hESC-derived optic vesicle were highly proliferative and underwent interkinetic nuclear migration, typical for pseudostratied epithelium (Norden et al., 2009) (Figures 2E and 2F and Movie S2). During days 1924, the distal portion of the vesicle bent inwards (invagination), forming a double-walled optic cup structure (Figures 3A3C and S3AS3C and Movie S3; found in 21% 24% of SFEBq/MIFS+W aggregates; ve experiments; 96 aggregates per experiment). As in vivo, while the thick inner portion expressed Chx10 (Figure 3D, blue; n = 30), the outer epithelium expressed Mitf (Figure 3D, red) and gradually started moderate pigment accumulation (Figure S3C, arrowheads). As with the mESC culture, there were no external structures pushing the tissue inwards such as lens or surface ectoderm, suggesting that the invagination of hESC-derived NR was self-driven by an intrinsic program within the retinal epithelium. As seen in the embryonic retina (Hilfer and Yang, 1980), the pseudostratied NR epithelium was tall, and the RPE epithelium

Cell Stem Cell


Optic Cup and 3D Retina Self-Forming from hESCs

Figure 1. Optimized Generation of Retinal Epithelia in 3D hESC Culture


(A) Schematic of NR-selective SFEBq culture of hESCs. (B) FACS analysis of Rx::venus+ cell population on day 18 (green). Black, control undifferentiated Rx::venus hESCs. Reproducible results were obtained from three independent studies. (C) Positive effects of FBS and SAG (added during days 1218) on Rx::venus+ retinal epithelial formation. (DI) Immunostaining of day-18 Rx::venus+ epithelia generated from hESCs in NR-selective SFEBq culture. Stains are for GFP (D, F, and H; green), Chx10 (E and H; light blue), Pax6 (G; red), and Mitf (I; red; no substantial expression). Dark blue, nuclear staining with DAPI (H). (J and K) Immunostaining of day-26 Rx::venus+ epithelia. Stains are for GFP (J; green), Chx10 (J; light blue) and Mitf (K; red). Dark blue, nuclear staining with DAPI (J). The RPE maker was not expressed in the retinal epithelia generated under these conditions, even at this late stage. (LO) Additional treatment of day-18 retinal epithelia (generated as in A) with CHIR99021 (N and O) but not with FBS (M and not shown) during days 1821 induced thin epithelia (N; arrowheads) with Mitf expression (O) on day 21. Bright-eld views (LN) and Mitf immunostaining are shown on a cryosection (O; red). Dark blue, nuclear staining with DAPI (O). (P and Q) Example of massive pigmentation was seen after longer culture in the day-56 SFEBq aggregates treated with CHIR99021 (during days 1825). 10% FBS and 50 ng/ml Activin were added to culture medium during days 1856 and days 1825, respectively. When these pigmented tissues were replated on the culture dish and cultured for additional 14 days, cells exhibited polygonal morphology typical for RPE (Q). Scale bars: 500 mm (C, L, and P); 200 mm (D and O); 50 mm (F, H, J, and Q). See also Figure S1.

Cell Stem Cell 10, 771785, June 14, 2012 2012 Elsevier Inc. 773

Cell Stem Cell


Optic Cup and 3D Retina Self-Forming from hESCs

Figure 2. Self-Organized Formation of Optic Vesicles from hESC-Derived Retinal Epithelium


(A) Schematic of optic-cup-forming SFEBq culture of hESCs. The effects of SAG treatment during days 1518 on the retinal formation were similar (not shown). (B) Additional treatment with CHIR99021 during days 1518 promoted Mitf expression in qPCR analysis (n = 3 experiments; each experiment with 12 aggregates). **p < 0.01, Students t test. (C) Live imaging of self-formation of an optic vesicle (Rx::venus+) evaginating from hESC-derived retinal epithelium cultured under the conditions of (A). (D) Immunostaining of Chx10 (right; light blue) and Mitf (right; red) in the day-18 Rx::venus+ optic vesicle (left; the same section). Chx10 expression and Mitf expression were found in the distal and proximal portions of the retinal epithelium. (E and F) Time-lapse multiphoton images showing interkinetic nuclear migration of retinal progenitors in the day-20 optic-vesicle epithelium. Cells were attracted toward the apical surface (red line) and underwent proliferation. The square region in (E) corresponds to the region in (F). Partial labeling of cells was done by mixing the venus-reporter cells with nonlabeled hESCs at reaggregation. **p < 0.01; error bars show SEM. Scale bars: 200 mm (C); 100 mm (D); 50 mm (E); 20 mm (F). See also Figure S2 and Movie S1 and Movie S2.

was shorter (Figures 3D3F). The hinge epithelium, which connects the NR and RPE tissues at an angled joint, formed a wedge-shaped domain with a very small apical area (Figure 3G). The appearance of the wedge-shaped hinge epithelium could be traced back to day 17 (Figure 3H; 100%, n = 15), when the optic vesicle epithelium started to make a visible exure (Figure 2C). The aPKC+ apical surface was inside, and the laminin774 Cell Stem Cell 10, 771785, June 14, 2012 2012 Elsevier Inc.

accumulating basal end was outside (Figure 3I; 100%, n = 20). A notable domain-specic mechanical property found in the mESC-derived optic cup is that its NR has a very low level of myosin activity (as shown by low pMLC2 levels), while the RPE contains high levels of pMLC2, particularly on the apical side (Eiraku et al., 2011, 2012). This makes the NR much less stiff than the RPE. A similar situation of region-specic actomyosin

Cell Stem Cell


Optic Cup and 3D Retina Self-Forming from hESCs

Figure 3. Self-Organized Optic Cups from hESCs

Formation

of

(A and B) Low-magnication views of the day-26 optic cup (yellow bracket) formed in the hESC aggregate. (A), bright-eld view; (B), uorescence image. (C) Optical cross-section of the day-26 optic cup by multiphoton imaging. (D) Immunostaining of the day-26 optic cup for Chx10 (light blue) and Mitf (red). (E) Phalloidin (red) and DAPI (dark blue) staining of the day-26 optic cup. The hinge region was indicated by broken lines. (F) Elongated morphology of progenitors in the pseudostratied NR epithelium on day 30. Multiphoton image. Partial labeling of cells was done by mixing the venus-reporter cells with nonlabeled hESCs at reaggregation. (G) High-magnication image of Rx::venus expression in the hinge region of the day-26 optic cup. (H) Wedge-like shape of the hinge region in the day-17 optic vesicle generated from hESCs. Partial labeling of cells was done by cell mixing as above. (I) Immunostaining of the day-26 optic cup for aPKC (apical; red) and Laminin (basement membrane; light blue). (J and K) Immunostaining of the day-26 optic cup (Rx::venus+; green) for phospho-MLC2 (pMLC2; red) (K is a high-magnication view). In the optic cup, pMLC2 was strongly accumulated on the apical side of the RPE, but not substantially observed in the NR. Scale bars: 100 mm (A, B, C, I, and J); 50 mm (D and E); 40 mm (F and H); 20 mm (G and K). See also Figure S3.

control was seen in the hESC-derived optic cup; a substantial accumulation of pMLC2 was observed in the RPE but not in the NR (Figures 3J and 3K). Taken together, these ndings indicate that hESC-derived retinal epithelium spontaneously forms an optic cup structure by self-organization, as has been reported for mESC culture. Structural and Morphogenetic Differences between hESC and mESC Cultures Although the self-organizing formation of optic cup structures occurs in both mESC and hESC cultures, there are several intriguing differences between them. First, the developmental process took much longer than in mESC culture (e.g., optic cup formation took $24 days and $9 days in hESC and mESC culture, respectively). Second, the early optic cups generated from hESCs were substantially bigger (initially $550 mm in long-axis diameter; Figures 4A and 4B) than those from mESCs (initially $250300 mm; Eiraku et al., 2011; see Figure S1A). The hESC-derived NR was also substantially thicker (120150 mm; Figure 4C) than the mESC-derived NR in the early optic cup (6080 mm; Figure S1A). These observations are in accord with the differences seen in the human and mouse optic cups during early retinogenesis; the early human optic cup at Carnegie stage

ller, 15, or fetal day 35, is $500 mm in diameter (ORahilly and Mu 1999), whereas the mouse optic cup on embryonic day 1010.5 is around 250 mm (Figure S1B). In theory, the infolding of the thick epithelia-like hESC-derived NR should require more mechanical elaboration than that of the thinner one from mouse cells. In this respect, we found another intriguing difference in the deformation mechanics between hESC and mESC cultures. Whereas the early retinal epithelia before invagination were apically concave (Figure 4D; Eiraku et al., 2011), the NR epithelia after invagination became apically convex in both hESC and mESC cultures. This direction of curving was xed at this stage, and maintained even when the NR epithelia were isolated from the main body of the aggregates (Figure S4A). Our previous study demonstrated that the NR eversion (from apically concave to convex) in mESC culture requires the presence of RPE and the hinge (Eiraku et al., 2012) and does not spontaneously occur in NR epithelium isolated before invagination (Figure 4D, top; Eiraku et al., 2011). In contrast, the hESC-derived NR epithelia, when isolated on day 18, gradually underwent spontaneous eversion and became apically convex over the next few days (100%, four experiments; 20 aggregates per experiment; Figures 4D4F and Movie S4), regardless of culture conditions before isolation
Cell Stem Cell 10, 771785, June 14, 2012 2012 Elsevier Inc. 775

Cell Stem Cell


Optic Cup and 3D Retina Self-Forming from hESCs

d26
547m

B
Long-axis diameter (m)

800

d26

***
600 400 200 0 mESC Optic Cup hESC Optic Cup

Rx::Venus

Mitf Chx10 DAPI Rx::Venus

D
mESC Optic vesicle

E
Apically Concave

d18

d30

Rx::Venus
hESC

eversion
Apically Convex

NR-Selective Culture

apical (aPKC) basal (Laminin)

aPKC DAPI

G
basal

d18 d22

+25hr
+Y-27632

+37.5hr
+Anti-IntB1
basal

+50hr d26 (cup)


basal

basal

DAPI aPKC Laminin

DAPI Rx::Venus

d22

+Anti-IntB1

d30

basal basal

Rx::Venus DAPI

apical

d30
basal

d30

basal

DAPI Rx::Venus

Chx10

Brn3 DAPI

Figure 4. Optic Cup Features Specic to hESC Culture


(A and B) Size measurement of the day-26 optic cup. (A) The long and short axes were always perpendicular (yellow line) and parallel to the central axis of the cup (NR center to the center of the cup opening), respectively (n = 20). (B) Long-axis diameters of the day-26 optic cups derived from hESCs (n = 8 cups), compared with those of mESC-derived day-9 optic cups (n = 8). ***p < 0.001, Students t test. (C) Thick NR epithelium (yellow bracket) in the hESC-derived day-26 optic cup. (D) Schematic of the NR eversion, which occurs in the isolated retinal epithelia of hESC culture (excised on day 18; before invagination), but not in those of mESC culture (excised on day 7; before invagination). (E) Apical-basal eversion of hESC-derived NR epithelia excised from SFEBq aggregates generated under the NR-selective conditions. Left and right panels show hESC-derived NR epithelia on day 18 and day 30, respectively, in the isolation culture. The aPKC+ apical surface (red; bottom; shown by arrows) was inside on day 18 but outside on day 30. (F) Live-imaging analysis of the eversion of NR epithelia excised on day 18. Green dots indicate cutting edges. (G) Immunostaining of the day-22 NR (isolation culture from day 18) for aPKC (apical; red) and Laminin (basement membrane; light blue) with DAPI nuclear staining (dark blue). Left, control; middle, treated with the ROCK inhibitor Y-27632 (10 mM) during days 1822; right, treated with anti-b1 integrin-neutralizing antibody

776 Cell Stem Cell 10, 771785, June 14, 2012 2012 Elsevier Inc.

Cell Stem Cell


Optic Cup and 3D Retina Self-Forming from hESCs

(SFEBq/MIFSW). These ndings suggest that the hESCderived NR epithelium has an intrinsic tendency to bend in an apically convex manner, which can contribute to the infolding of this thick tissue during invagination. The spontaneous apical eversion of hESC-derived NR was not sensitive to ROCK inhibitor, which blocks the Rho-myosin pathway. In contrast, the apical eversion was inhibited when integrin signaling was blocked by its neutralizing antibody (blocked in 75%, n = 24; see the reduced phospho-paxillin levels on the basal side; Figure S4C), suggesting an essential role of basement-membrane-mediated signals in this process. There were no substantial changes in proliferation (phosho-H3, Ki67) or apoptosis (active caspase 3) (Figures S4D and S4E). One obvious difference along the apical-basal axis of the NR was the positions of the nuclei (i.e., the location of the cell bodies of the radial-glia-like NR progenitors), which were strongly deviated to the apical side in hESC culture (Figure 4H). This strong deviation in cell position is also found in the human fetal optic cup ller, 1999), but not in the early mouse optic (ORahilly and Mu cup or the mESC-derived optic cup (e.g., Figures S1A and S1B; Eiraku et al., 2011). The apically biased localization of cell bodies can theoretically contribute to the preferential volume expansion on the apical side, potentially leading to the apically convex curvature in the NR. Interestingly, integrin blockade diminished the apical nuclear deviation within the hESC-derived NR (Figures 4I and 4J), along with the inhibition of the eversion, suggesting that both depend on some basement membrane signals. Multilayered NR Containing Photoreceptors, Ganglion Cells, and Interneuron Precursors Early NR tissues obtained from both NR-selective (Figure 1) and optic-cup-forming (Figure 2) cultures grew to form stratied tissues (in this report, we mainly show the results from the NRselective culture below). When retinal epithelia were isolated en bloc (typically 300500 mm in diameter) from day-18 SFEBq aggregates and cultured in suspension, the NR epithelia grew into larger epithelial vesicles and remained Rx::venus+ on day 30 (Figures 4K and 4L; the long-axis diameter was typically $1 mm). The great majority of cells in the day-30 NR were positive for the progenitor marker Chx10 (Figure 4M) and the mitotic marker Ki67 (not shown). In the embryo, NR epithelia generate several kinds of retinal components in a temporally controlled manner, with the rst-born cell lineage being ganglion cell (Brn3+/Pax6+) (Mu and Klein, 2004; Cayouette et al., 2006). As in the embryonic eye, postmitotic ganglion cells began to emerge during early phases of hESC culture (around day 24; data not shown). Their number gradually increased over the next week, and formed a distinct layer at

the basal-most zone of the hESC-derived NR (Figure 4N and data not shown). We next investigated photoreceptor differentiation (Figure 5) using hESC lines with venus cDNA knocked in at the locus of the early photoreceptor marker gene Crx (Furukawa et al., 1997) (Figure S5A). On days 2834, a relatively small number of Crx::venus+ cells were present, most of which were found between the ganglion (Brn3+) and the progenitor (Chx10+) cell layers (hereafter referred to as the intermediate-deep zone) (Figure 5A and not shown). On day 42, in addition to the intermediate-deep layer, the apical layer of the NR also accumulated Crx::venus+ cells (Figure S5B). The photoreceptor-precursor marker Blimp1 (Katoh et al., 2010; Brzezinski et al., 2010) was also expressed in these cells (Figure S5C). At this stage, TuJ1+ axons of ganglion cells already formed bundles in the basalmost zone (Figure S5D). On day 60, NR tissues became $2 mm in long-axis diameter and $200 mm in thickness. Despite this thickness, the NR tissue remained semitransparent (Figure 5B). Crx::venus+ photoreceptors located in both the apical-most and intermediate-deep layers (Figure 5C) expressed the bona de photoreceptor marker Recoverin (Figure S5E) and were negative for Ki67 (mitotic marker; Figure S5F). During days 4260, Chx10+ progenitors (Chx10+, Ki67+) still constituted the largest population in culture and formed a thick layer between the two Crx::venus+ zones (not shown), while Brn3+/Tuj1+ ganglion cells were preferentially localized to the basal side (Figure S5G) as seen on day 42. Later, on day 126 (thickness was >250 mm; Figure 5D), Crx::venus+ photoreceptors in the apical-most layer further increased in number (12%18% of total cells; n = 4; >1,000 cells for each) and were densely packed (Figure 5E). The apically located photoreceptors expressed not only Recoverin (Figure 5F), but also the rod-specic marker Nrl (Mears et al., 2001) or the early cone marker Rxr-gamma (Janssen et al., 1999; also expressed in ganglion cells) (Figures 5G and 5H). The ratio of Nrl+ to Rxr-gamma+ cells in the apical-most layer of the day-126 NR was 4.1 1.6: 1 (n = 4; >200 cells for each). On the apical side, these photoreceptors formed ZO-1+ tight junctions (Figure 5I, red) and substantially accumulated the visual pigments rhodopsin and color opsins (Figures 5J5L, S5H, and S5I). Chx10+ progenitors were still found in the intermediate layer, while Tuj1+ postmitotic neurons were found on the basal side (Figures 5M, 5N, and S5J). Ptf1a+ interneuron precursors, becoming amacrine and horizontal cells later (Fujitani et al., 2006), were present mostly near the Crx::venus+ cells in the intermediate-deep layer (Figure 5O). Calretinin+ cells (marker for amacrine and ganglion cells) and GABA+ neurons (Figures 5P, 5Q, and S5K) were also present in the intermediate-deep

(5 mg/ml) during days 1822. In the controls and the Y-27632-treated groups (n = 20), all NR epithelia underwent apical-basal eversion, while the integrin-antibodytreated group remained apically concave (75%, n = 24; two independent experiments). (HJ) Apically derived nuclear localization (white zone; dark blue, DAPI nuclear staining) in the Rx::venus+ NR epithelium (yellow) of the day-26 optic cup (H) and day-22 NR (I). The anti-integrin antibody caused an even distribution of nuclear positions as seen in mESC-derived NR (J). (K) Semitransparent day-30 NR epithelium in isolation culture. (LN) Cross-sections of day-30 NR epithelia (forming a vesicle). Rx::venus expression was observed throughout the epithelium (L; green; dark blue, DAPI nuclear staining). Chx10+ progenitors (light blue; the same section as L) were located preferentially in the apical (outer) zone (M). (N) Brn3+ ganglion cells were located on the basal side. ***p < 0.001; error bars show SEM. Scale bars: 200 mm (K and L); 100 mm (A, C, E, F, G, and N); 50 mm (H); 20 mm (I and J). See also Figure S4 and Movie S4.

Cell Stem Cell 10, 771785, June 14, 2012 2012 Elsevier Inc. 777

Cell Stem Cell


Optic Cup and 3D Retina Self-Forming from hESCs

Figure 5. Self-Formation of Multilayered NR Tissue from hESCs


(A) A relatively small number of Crx+ photoreceptor precursors were found in the intermediate-deep zone of the day-34 NR epithelium. (B) Semitransparent NR epithelium on day 60. Crx::venus uorescence from the inside of NR could be seen by external observation. (C) Immunostaining of the day-60 NR epithelium for GFP (Crx::venus+ photoreceptor precursors; green), Chx10 (progenitors; red), and Pax6 (strong staining in ganglion cells; light blue) with DAPI nuclear staining (dark blue). (DQ) Immunostaining of day-126 NR epithelia. The day-126 NR epithelium was thick (>250 mm; D). (E) Accumulation of Crx::venus+ photoreceptor precursors on the apical-most zone. (FL) The apical-most zone contained many recoverin+ photoreceptors (F); some expressed the rod marker Nrl (red, G; white, H), and others expressed the cone marker Rxr-gamma (red, H). The apical photoreceptors formed ZO-1+ apical junctions (I). The visual pigments rhodopsin (red, J and K) and S-opsin (red, L) were accumulated on the apical surface. (M and N) Chx10+ NR-progenitors were present in the intermediate zone (M), while TuJ1+ postmitotic neurons were mostly found in the more-basal domain (N). (O) Ptf1a+ interneuron precursors (red) were found in the intermediatedeep and deep zones. (P and Q) Most of the calretinin+ neurons (P) and GABA neurons (Q) were located in the basal zone near the intermediatedeep zone (basal Crx::venus+ photoreceptor precursors). (R) Schematic of multilayered NR generated from hESCs. Scale bars: 200 mm (BD); 100 mm (A, E, F, M, N, and P); 50 mm (G and Q); 20 mm (HJ and L). See also Figure S5.

and deep layers. In addition, a small number of calbindin+ cells with immature features were found; unlike mature calbindin+ horizontal cells (located near the apical photoreceptor layer and extending neurites horizontally), the majority of calbindin+ cells in this culture were located near the basal zone, while some calbindin+ cells in the intermediate zone were vertically long in shape (indicative of immaturity; Figures S5L and S5M). Bipolar cells are late-born neurons in retinogenesis in vivo. Consistent with the idea, few postmitotic bipolar cells (PKC+) were observed even at this stage (n = 20; not shown). Also in accordance with the lack of bipolar cells, PSD95+ plexiform
778 Cell Stem Cell 10, 771785, June 14, 2012 2012 Elsevier Inc.

layers (synaptic zones) were not observed (n = 20; not shown). On day 126, outside of the apical junctions, the photoreceptors had apical protrusions of monociliated, mitochondria-rich cell bodies, reminiscent of inner segments (ellipsoids), which carry connecting cilia (Ishikawa and Yamada, 1969; Besharse and Insinna, 2010) (Figures S5NS5P; some cilia had a membrane vesicle on their tips, but no obvious outer-segment formation was seen; Figure S5P, arrow). Taken together, these ndings demonstrate that hESCderived retinal epithelium can spontaneously generate stratied fetal-like NR tissues containing multiple components in a spatially arranged manner (Figure 5R). Slow Photoreceptor Generation Accelerated by Notch Inhibition Unlike mESC culture, the number of photoreceptors (Crx::venus+ precursors and Recoverin+ cells) increased only gradually in hESC culture (over 1318 weeks). However,

Cell Stem Cell


Optic Cup and 3D Retina Self-Forming from hESCs

Figure 6. Accelerated Generation of Photoreceptors by Notch Inhibition


(AC) Control NR epithelia (carrying Crx::venus), excised on day 19. (A) Crx::venus uorescence on day 41. (B and C) Immunostaining of day-43 NR epithelia for Recoverin and Crx::venus (B). (DF) NR epithelia (carrying Crx::venus), excised on day 19 and treated with DAPT during days 29 43. (D) Crx::venus uorescence on day 41. (E and F) Immunostaining of day-43 NR epithelia for Crx::venus (E) and Recoverin. Bracket, expanded photoreceptor layer (Crx+/Recoverin+). (GI) Immunostaining of control day-43 NR epithelia (carrying Crx::venus, excised on day 19) for GFP (green; G and I), Brn3 (red, G; light blue, I), Ki67 (red, H), pH3 (light blue, H), and active caspase 3 (red, I). (JL) Immunostaining of day-43 NR epithelia treated with 10 mM DAPT during days 2943 for Crx::venus (J and L), Brn3 (J and L), Ki67 (K), pH3 (K), and active caspase 3 (L). (M and N) Immunostaining of control day-37 NR epithelia (carrying Crx::venus; excised on day 18) for acetyltubulin (AcTub; red; M and N) and Chx10 (white, N; the same section as M). (O and P) Immunostaining of day-37 NR epithelia (carrying Crx::venus; excised on day 18) treated with DAPT during days 2937 for AcTub (red; O and P) and Chx10 (white, P; the same section as O). Both Chx10 cells and the apical-basal arrangement of AcTub signals (indicative of the presence of radial glia-like progenitors) disappeared after DAPT treatment. Scale bars: 500 mm (A and D); 100 mm (B, E, G, I, J, and L); 50 mm (M). See also Figure S6.

treatment of day-29 NR epithelia with the Notch inhibitor DAPT (Jadhav et al., 2006; Osakada et al., 2008) for 1214 days dramatically increased the number of Crx::venus+/Recoverin+ photoreceptor cells (40%78% of NR cells on day 43; Figures 6A6F and S6AS6F, and not shown), showing that a large portion of progenitors at this stage had the competence for photoreceptor differentiation (enhanced differentiation was already seen after 4 day DAPT treatment; Figures S6GS6J). The number of Brn3+ ganglion cells did not change dramatically (Figures 6G and 6J; 13%19% of NR cells regardless of DAPT treatment; four experiments). In contrast, the number of mitotic progenitors (Ki67+ cells) substantially decreased after the treatment (Figures 6H and 6K). There was a marginal increase in apoptosis (active caspase3+) on day 43, but its level was only moderate (Figures 6I and 6L). These ndings indicate that photoreceptor differentiation in hESC culture is a slow process that is decelerated by Notch signaling but can be strongly accelerated by treatment with a Notch inhibitor. On day 43 (after 2 weeks of DAPT treatment), the NR epithelia became somewhat irregular in shape (Figures 6E and 6F). We infer that this was presumably tissue-architectural perturbation caused by the rapid loss of Chx10+ retinal progenitors, which are structurally similar to radial glia in the

developing cortex. Along with the accelerated differentiation into photoreceptors, the loss of progenitors occurred during the rst several days of DAPT treatment (Figures 6M6P). Efcient En Bloc Cryopreservation of hESC-Derived NR Epithelium The generation of stratied NR tissues takes much longer in hESC culture than in mESC culture, presumably reecting the difference in gestational period between the two species (>260 days versus 20 days). It is technically challenging, however, to continue large-scale culture for more than several weeks in a row while controlling the product quality in a reproducible manner. To address this problem, we sought to develop an en bloc cryopreservation method that would enable the NR epithelia to be stored partway through the differentiation process. According to our observations, the most critical period for successfully generating stratied NR from hESCs is the rst 3040 days; NR that grew well up to this point tended to continue growing as a healthy-looking epithelium with multiple layers. Therefore, we aimed to develop an efcient cryopreservation method for storing the NR epithelia at this stage. Rx::venus+ NR tissues were isolated from SFEBq-cultured hESC aggregates on day 18 and cultured in suspension until
Cell Stem Cell 10, 771785, June 14, 2012 2012 Elsevier Inc. 779

Cell Stem Cell


Optic Cup and 3D Retina Self-Forming from hESCs

A
d18 excise e eversion d30, d33 or d40

B
100%

Integrity grades of NR epithelia (5 days after thawing)

75%

bad low

Freeze

Thaw
50%

high excellent

Pretreatment on ice hESC aggregate

Vitrification in liquid N2

Cryopreservation

Further suspension culture

25%

0%

ng Vi tr i ca tio n

Sl ow

C
Pretreatment Control (No Freezing) Vitrification (No Pretreatment) DMSO DMSO+EG

Pretreatment

DMSO +EG+Sucrose

Rx::Venus

D
Frozen@d40

d60

Rx::Venus

Pax6 Chx10

Blimp1

G
Frozen@d105

d126 H

d126

Rx::Venus

DAPI Recoverin

Rxr DAPI Nrl

J
Frozen@d33

d49

d49

d49

DAPI Ptf1a +DMSO Crx::Venus (d38-)


Figure 7. Efcient Cryopreservation of hESC-Derived NR Epithelia

+DAPT Brn3 Crx::Venus (d38-)

Brn3 Crx::Venus

(A) Schematic of the cryopreservation process. (B) Semiquantitative comparison of different cryopreservation conditions for NR tissue survival (extent of epithelial integrity) 5 days after thawing (frozen on day 30; 1520 NR epithelia/condition). Integrity of the epithelium of each vesicle: grade 1 (excellent), continuous smooth surface in >90% of the epithelial region; grade 2 (high), 50%90%; grade 3 (low), 25%50%; grade 4 (bad), less than 25%. Frequency of each grade found among the thawed NR epithelia is shown by percentage.

780 Cell Stem Cell 10, 771785, June 14, 2012 2012 Elsevier Inc.

D M

SO +E G SO +E G +S uc

ng

Fr ee zi

Fr ee zi

D M

SO

N o

D M

Cell Stem Cell


Optic Cup and 3D Retina Self-Forming from hESCs

day 30. Then, the NR epithelia were subjected to cryopreservation en bloc under various conditions, and the tissues survival was examined 5 days after thawing (see Figure 7A). We found that a vitrication method with pretreatment in 10% DMSO + 5% ethylene glycol (EG) + 10% sucrose on ice was quite effective for NR cryopreservation not only in terms of tissue survival (Figures 7B and 7C, top row; see the legend for details on epithelial integrity rating), but also for the maintenance of Rx::venus expression (Figure 7C, bottom row and S7A). In contrast, by either criterion, neither the simple one-step vitrication without pretreatment nor the conventional slow-freezing method (with DMSO) was nearly as effective as vitrication with DMSO/EG/ sucrose pretreatment (Figures 7B, 7C, and S7AS7C). The hESC-derived NR epithelia frozen under the optimized vitrication conditions retained their integrity and semitransparency after freezing and thawing, and grew in the prolonged culture. For instance, the NR epithelia frozen on day 40 uniformly expressed a substantial level of Rx::venus on day 60 (20 days after thawing; Figure 7D). These NR epithelia showed layer formation of Chx10+ progenitors and Pax6+ ganglion cells, and contained Blimp1+ photoreceptor precursors (Figure 7E and 7F and not shown), indicating that the stratied NR tissues had been preserved. The integrity, transparency, and photoreceptor-generating ability (both cones and rods) were kept well even when day-105 NR tissues were cryopreserved (Figures 7G7I and S7D). In addition, DAPT treatment of NR epithelia that had been cryopreserved and thawed caused a massive induction of Crx::venus+ photoreceptor precursors (Figure 7J7L and S7E; in these panels, the NR epithelia were frozen on day 33 and treated with the vehicle DMSO or DAPT during days 3849). DISCUSSION Self-Organization of Optic Cup from hESCs In this study, we demonstrated that spontaneous emergence of a highly ordered retinal structure (optic cup) occurs in 3D culture of an unpatterned, homogenous aggregate of hESCs. This nding supports the idea that there is a latent intrinsic order for retinal morphogenesis present in early progenitors across species. In SFEBq culture of hESCs, the combination of early Wnt inhibition, ECM addition, Hedgehog signaling, and FBS treatment strongly promotes the generation of retinal epithelium. Hedgehog signaling may be involved in the adjustment of positional information along the dorsal-ventral axis within hESCderived forebrain tissue. FBS has also been shown to enhance retinal differentiation from mESCs (Ikeda et al., 2005). Lowmolecular-weight factors in FBS do not appear to have a major role, since dialyzed FBS showed a similar activity. The effect of

FBS could be not substituted by IGF, Activin, or FGF2 (growth factors implicated for retinal differentiation; data not shown) or by bronectin or vitronectin (matrix proteins abundantly present in FBS) (Figures S2CS2F and data not shown). After the initial retinal specication in these conditions, Wnt augmentation at a specic time window was critical for the balanced generation of both NR and RPE and for the reproducible self-formation of optic cup structures in the hESC culture. The efcient timed control of the canonical Wnt pathway (switching off and on during days 012 and day 1518, respectively) was made possible by applying the chemical antagonist and agonist. Features Specic to hESC-Derived Optic Cup The hESC-derived optic cup is substantially larger than the mESC-derived cup. The control mechanism for optic-cup size, which appears to be intrinsic to the retinal epithelium, is a very intriguing topic for future investigation. In addition, the isolated hESC-derived NR, which is thick, has a tissue-intrinsic tendency toward apically convex curvature, which was not observed in the isolated mESC-derived NR (Eiraku et al., 2011). Our previous study pointed out three local mechanical rules in optic cup invagination (relaxation-expansion model; Eiraku et al., 2011, 2012), all of which were also seen in the hESC-derived optic cup: myosin inactivation in NR, the wedge shaping of the hinge, and the tangential NR expansion. The addition of a fourth local rule observed here, the intrinsic apical bending property of human NR, may contribute to the strong deformation of this thick epithelial structure. One possible mechanism for interpreting this autonomous curvature is that the apical deviation of cell-body positions, which is obvious during invagination only in the hESC-derived NR, could contribute to the preferential apical expansion of the tissue. An alternative possibility is basal narrowing of the NR epithelium (e.g., by basal microlament contraction), an active role of which has been suggested in previous studies using sh (Martinez-Morales et al., 2009). So far, however, we have failed to observe basally dominant accumulation of actin or pMLC2 in the everting NR (Figure S4F and data not shown). On the other hand, the dependence of the NR eversion on integrin activity suggests an essential role for some basementmembrane-derived polarization signals. In the future, it will be intriguing to investigate whether the integrin-sensitive apical deviation of their cell bodies is caused actively, by nuclear attraction toward the apical direction (e