J Mol Cell Cardiol 33, 149–160 (2001)

doi:10.1006/jmcc.2000.1288, available online at http://www.idealibrary.com on

Sarcolemmal Blebs and Osmotic Fragility

as Correlates of Irreversible Ischemic
Injury in Preconditioned Isolated Rabbit
Cardiomyocytes
Stephen C. Armstrong, L. Christine Shivell and Charles E. Ganote
Veterans Affairs Medical Center and Department of Pathology, James H. Quillen College of Medicine,
East Tennessee State University, 37614, USA
(Received 14 April 2000, accepted in revised form 17 October 2000, published electronically 4 December 2000)

S. C. A, L. C. S  C. E. G. Sarcolemmal Blebs and Osmotic Fragility as Correlates of
Irreversible Ischemic Injury in Preconditioned Isolated Rabbit Cardiomyocytes. Journal of Molecular and Cellular
Cardiology (2001) 33, 149–160. The hypothesis that irreversible ischemic injury is related to sub-sarcolemmal
blebbing and an inherent osmotic fragility of the blebs was tested by subjecting isolated control and ischemically
preconditioned (IPC) or calyculin A (CalA)-pretreated (protected) rabbit cardiomyocytes to ischemic pelleting
followed by resuspension in 340, 170 or 85 mosmol medium containing trypan blue. At time points from
0–240 min, osmotic fragility was assessed by the percentage of trypan blue permeable cells. Membrane blebs
were visualized with India ink preparations. Bleb formation, following acute hypo-osmotic swelling, developed
by 75 min and increased with longer periods of ischemia. Osmotic fragility developed only after 75 min. Cells
resuspended in 340 mosmol media did not form blebs and largely retained the ability to exclude trypan blue,
even after 240 min ischemia. Although the latent tendency for osmotic blebbing preceded the development of
osmotic fragility, most osmotically fragile cells became permeable without evident sarcolemmal bleb formation.
The onset of osmotic fragility was delayed in protected cells, but protection did not reduce the bleb formation. It
is concluded that blebbing and osmotic fragility are independent manifestations of ischemic injury. The principal
locus of irreversible ischemic injury and the protection provided by IPC may lie within the sarcolemma rather
than at sarcolemmal attachments to underlying adherens junctions.  2000 Academic Press
K W: Cytoskeleton; Isolated cardiomyocytes; Ischemic preconditioning; Protein phosphatases; Ir-
reversible injury; Sarcolemmal blebs.

Introduction onset of cell death during ischemia in the absence of

The potential to salvage ischemic myocardium is
linked to the ability to reperfuse the ischemic tissue.
Reperfusion of irreversibly injured myocardium pro-
duces acute osmotic and mechanical stresses that
manifiest latent sarcolemmal lesions formed during
ischemia1,2 and results in acute contraction band
necrosis associated with formation of large sub-
sarcolemmal blebs and rupture of the sarcolemmal
membrane.3–7 These events contrast with the slow
reperfusion. In myocardium, hypo-osmotic swelling
results in the formation of sub-sarcolemmal blebs
in ischemic cells but not in oxygenated control
cells.2,8 The onset of osmotic fragility correlates
closely with the onset of irreversible injury in
perfused rat hearts and in isolated rat car-
diomyocytes.9,10
It has been postulated that osmotic fragility and
irreversible injury are associated with a latent isch-
emic lesion in the costamere junctions, allowing

Please address all correspondence to: Dr Stephen C. Armstrong, Dept. of Pathology, PO Box 70568, James H. Quillen College of
Medicine, East Tennessee State University, 37614, USA. Tel: 423-439-6210; Fax: 423-439-8060; E-mail: armstron@access.etsu.edu

0022–2828/01/010149+12 $35.00/0  2000 Academic Press

150 S. C. Armstrong et al.
the sarcolemma to detach from the underlying
cytoskeleton to form sub-sarcolemmal blebs.6 The
sarcolemmal membranes overlying blebs are then
deprived of cytoskeletal support and were presumed
inherently fragile and susceptible to rupture fol-
lowing imposition of mechanical or osmotic
stresses11 or during ischemic cellular swelling.12 An
experimentally testable corollary of this hypothesis
is that myocardial cells protected by ischemic pre-
conditioning should demonstrate a concurrent
delay in osmotic fragility and bleb formation during
osmotic swelling.
An in vitro rabbit cardiomyocyte model of isch-
emia and cardioprotection by ischemic pre-
conditioning (IPC) and protein phosphatase
inhibition (CalA) was used to quantitatively define
the relationships between sarcolemmal blebbing
and osmotic fragility. The tendency of cells to bleb
preceded development of osmotic fragility and IPC/
CalA delayed the onset of osmotic fragility but
not bleb formation. The results suggest that bleb
formation during osmotic swelling and osmotic
fragility are independent phenomena. Irreversible
injury may not be related to lesions within at-
tachments of the sarcolemma to the underlying
cytoskeleton, as had been previously hypothesized,
but rather to intrinsic changes of the sarcolemmal
membrane.
Materials and Methods
Cell isolation
Isolated, calcium tolerant, adult rabbit car-
diomyocytes were prepared by collagenase per-
fusion as previously described13 according to the
methods of Hohl et al.,14 with the specific modi-
fication that adenosine was excluded from all media.
The investigation conforms with the Guide for the
care and use of laboratory animals published by the
US National Institutes of Health (NIH publication
No 85-23, revised 1985). Briefly, hearts were can-
nulated, attached to a recirculating Langendorff
apparatus and perfused at 37°C with a nominally
calcium-free buffer containing in m, NaCl (125),
KCl (4.75), KH2PO4 (1.2), MgCl2 (1.2), HEPES (30)
bovine serum albumin (BSA, Pentax fraction V)
(0.1%), glucose (11), taurine (58.5) creatine (24.9)
including a complete amino acid mixture (GIBCO)
basal medium, Eagle (BME) and modified Eagle’s
medium non-essential amino acid solution diluted
1:100 with Krebs–Henseleit buffer. Collagenase,
(Worthington, Type II) was added to the buffer to
make a concentration of 200 U/ml and perfusion
continued until the hearts were flaccid
(15–20 min). Ventricles were minced and cells dis-
persed with a large bore pipette, followed by fil-
tration through nylon mesh. Cells were washed by
a brief 20×g centrifugation and resuspension in
Krebs–Henseleit buffer, containing in m, NaCl
(125), KCl (4.75), KH2PO4 (1.2), MgCl2 (1.2), HEPES
(30), glucose (11), and 2 percent bovine serum
albumin (BSA, Pentex Fraction V) supplemented
with creatine, taurine and amino acids. Cells were
then incubated for 30 min at 30°C to allow re-
establishment of normal electrolyte gradients, fol-
lowed by calcium addition to a final concentration
of 1.2 m. Calcium tolerant cells were harvested
by two brief centrifugations, discarding the super-
natants. After purification cells were resuspended in
wash solution, buffered to pH 7.4. Isolates averaged
85–90 percent viability and contained sufficient
cells for preparation of 4–6 ischemic pellets. A
separate isolate was used for each experiment, each
series consisted of five to seven experiments.
Ischemia model
The initial cell suspension was equally divided into
control and experimental groups. Cells were pre-
conditioned (IPC) by a 10 min period of in vitro
ischemic pelleting, followed by resuspension of cells
in oxygenated buffer or preincubated for 10 min
with 1
m calyculin A (CalA), obtained from Re-
search Biochemicals International (Natick, MA,
USA), and dissolved in DMSO as a stock solution
of 100
. A single wash was followed by an
oxygenated 15-min post-incubation period. Paired
controls for this group consisted of cells from the
same isolate incubated in oxygenated buffer for
10 min but otherwise subjected to the same wash
and post-incubation protocol, prior to final ischemic
pelleting. An aliquot of each cell suspension was
then placed in a 1.8-ml microcentrifuge tube (Fisher
Scientific, Pittsburgh, PA, USA) and centrifuged into
a pellet. Each cell pellet occupied a volume of about
0.25 ml, and measured 0.8–1 cm in thickness. Ex-
cess supernatant was removed to leave a fluid layer
above the pelleted cells of about one third the
volume of the pellet. After layering with mineral
oil, the cell pellets were incubated without agitation
at 37°C. A 25
l sample of the final cell pellet was
removed through the oil layer, with care to prevent
introduction of air, at the appropriate time points
and resuspended for 3–5 min in the appopriate
osmolality buffer. The initial incubation medium
was formulated to be nearly isotonic at 290–300
 Sarcolemmal Blebs and Osmotic Fragility
mosmol. The medium into which cells were re-
suspended contained a mitochondrial inhibitor,
3 m amytal, to prevent reoxygenation rounding
of the cells, with a final osmolality of 340 mosmol.
The hypotonic swelling media were prepared by
dilution to 170 or 85 mosmol with distilled water.
A 10
l cell sample was mixed with equal volumes
of counting media (0.5% glutaraldehyde in 85
mOsM modified Tyrodes solution, with reduced
NaCl, containing 1% trypan blue) mixed with a
small volume of a concentrated solution of India
ink particles reuspended in PBS. Microscopic ex-
amination at 100× magnification determined the
morphology (rod, round or square), the per-
meability of the cells to trypan blue and the
formation of sub-sarcolemmal macro-blebs. Photo-
microscopy was performed on a Nikon Diaphot
microscope. The term square cell was coined to
depict the shape of isolated rat cardiomyocytes in
ischemic contracture.15 The terminology is retained
although rabbit cardiomyocytes are initially longer
than rat cardiomyocytes and do not attain the 1/
1 length/width ratio typical of rat cardiomyocytes
in full contracture.
Electron microscopy
Oxygenated control cells, 75 and 180 min ischemic
cells were exposed to 85 mosmol media for 3–5 min,
briefly centrifuged and resuspended for 2 h at room
temperature in 1% glutaraldehyde in a 0.05 
cacodylate buffer (pH 7.4) containing 0.05  lysine
and 5 m EGTA, which is a hypotonic fixative that
mordates proteins to render them more electron
dense.3 Cells were washed three times in 0.05 
cacodylate buffer (in the absence of lysine/EGTA)
and postfixed in 1% osmium tetroxide in cacodylate
buffer for 60 min at 4°C, then washed twice with
distilled water. The fixed cells were processed by
serial ethanol dehydration, followed by propylene
oxide and 50/50 propylene oxide/epon prior to
epon embedding by routine tissue techniques. Sec-
tions cut with a diamond knife were placed on
copper grids, double stained with uranyl acetate/
lead citrate and examined in a Philips EM 201
microscope.
Horseradish peroxidase assay
An ultrastructural assessment of sarcolemmal per-
meability by horseradish peroxidase uptake was
conducted by a modification of a published tech-
nique.16 Oxygenated (0 min) and ischemic (75 and
180 min) were suspended in 1 ml of 85 mosmol
151
buffer containing 0.1% horseradish peroxidase type
II (Sigma Chemical Co, St Louis, MO, USA) for
5 min at room temperature. The cells were pelleted,
washed in PBS and resuspended in 4% para-
formaldehyde for 10 min at room temperature. The
cells were washed in 0.1  cacodylate buffer and
reacted for 5 min at room temperature in 0.3 ml of
3,3′diaminobenzidine tetrahydrochloride in a H2O2/
imidazole buffer (DAKO, Carpinteria, CA, USA).
Light photomicroscopy of the cells was performed
prior to post-fixation in 1% buffered osmium tetr-
oxide for 60 min at 4°C. The cells were then pro-
cessed for electron microscopy as described above.
Data was analysed by ANOVA (SupraANOVA II,
Abacus Concepts Inc., Berkeley, CA, USA) followed
by Scheffe’s or Fisher’s PSDL post hoc tests. A
P<0.05 was considered significant.
Results
Isolated rabbit cardiomyocytes exclude trypan blue
and retain an elongated rod-shaped configuration
over a 4-h period of oxygenated incubation. Our in
vitro model of myocardial ischemia mimics tissue
ischemia in that centrifugation of cardiomyocytes
into pellets excludes oxygen and restricts extra-
cellular fluid, providing the hypoxic and metabolite
accumulation components of ischemia during in-
cubation by consumption of available oxygen and
accumulation of lactate as a result of anaerobic
glycolysis. Following cessation of glycolysis, ATP
depletion17 is accompanied by ischemic contracture
of cardiomyocytes. A stable rod to square con-
version occurs by 75 min of ischemia (Figs 1A,
1B and 2). Oxygenated and 75 min ischemic cells
resisted development of trypan blue permeability
when exposed to 85 mosmol swelling medium (Figs
1B, 1C and 3). A significant rise in osmotic fragility
occured only after 75 min of ischemia. Prolonged
periods of ischemia resulted in an increasing pro-
portion of osmotically fragile cells (Figs 1D, 1E and
3). Between 75 and 180 min of ischemia there was
a small increase (8.37±1.18% to 25.2±3.43%)
in the trypan blue permeability of cells exposed to
340 mosmol media. There was a graduated increase
in membrane permeability (76.3±3.6% and
64.9±2.2%) in cells exposed to 85 and 170 mosmol
swelling, respectively, after 240 min of ischemia.
The rates of cell squaring in the IPC/CalA groups
were similar to that of controls (Fig. 2). Protected
cells demonstrated a significant delay in the onset
of osmotic fragility (Fig. 3), with an approximate
50% reduction of osmotically fragile square cells
152 S. C. Armstrong et al.

Figure 1 Light micrographs of trypan blue and India ink preparations of control and ischemic isolated cardiomyocytes,
resuspended in 85 mosmol hypotonic swelling media. (A) At 0 min of ischemia, the cell suspensions consisted of
elongated, rod-shaped cells with a smaller proportion of viable and trypan blue permeable rounds (×185). (B) After
75 min of ischemia, the rods contracted into shortened, square-shaped cells. The majority of squares excluded trypan
blue in the presence of prominent dome-shaped macro-blebs (×185). (C) A higher magnification of a blebbed viable
square cell after 75 min of ischemia demonstrate a distinct margin of the blebs on the sarcolemmal membrane (×360).
(D) At 180 min of ischemia, the majority of square cells are trypan blue permeable and lack evident macro-blebs,
although the viable squares in the preparation retain macro-blebs (×240). (E) A higher magnification at 180 min of
ischemia. The cells excluding trypan blue have large blebs but the trypan blue-stained cells have no visible blebs
(×400).
 Sarcolemmal Blebs and Osmotic Fragility
Figure 2 Graph of the percentage of cells retaining an
elongated, rod-shaped conformation. The initial round
cells contaminating the preparation were excluded from
the cell counts. By 75 min of ischemia, 80–90% of cells
had undergone ischemic contracture. There was no dif-
ference in the rates of ischemic contracture between the
control (open symbols and solid lines) and the IPC or
CalA-pretreated (filled symbols and dashed lines) groups.
Figure 3 Graph showing the percent of trypan blue
permeable squared cells, relative to the total population
of squared cells, as a function of duration of ischemic
incubation and osmolality of the resuspension buffer. The
loss of viability of control cells (empty symbols) was
increased following hypotonic swelling. The cells re-
suspended in 340 mosmol media (boxes) died at a rel-
atively slow rate. There was a progressive increase in
the percentage of trypan blue permeable cells, following
resuspenson in 170 (triangles) and 85 mosmol (circles)
media. Hypo-osmotic, protected cells (filled symbols) had
a significant (P<0.05) reduction in the percentage of
trypan blue permeable cells as compared to the control
cells, indicated by the asterisks.
153
Figure 4 Graph showing the percentage of viable rod
(0 min) or square (75, 180, 240 min) cells possessing
blebs in India ink preparations. Rods resisted blebbing.
By 75 min of ischemia, the majority of cells were viable
squares and 50% possessed blebs after resuspension in
85 mosmol media (circles). A smaller percentage of
squares had blebs after resuspension in 170 mosmol
media (triangles) and few blebs were observed after 340
mosmol resuspension (boxes). Bleb formation in viable
squares was slightly increased at 180 min of ischemia.
The bleb data obtained at 240 min of ischemia may be
inaccurate due to the reduced population of viable cells.
No reduction in hypo-osmotic bleb formation was ob-
served in IPC or CalA cells (solid symbols).
as compared to 180 min ischemic controls, after
exposure to either 170 or 85 mosmol swelling
media.
Oxygenated rod-shaped cells, immediately prior
to ischemic pelleting, were remarkably resistant to
blebbing (Figs 1A and 4). Only 17.5±8.2% of rods
had blebs when exposed to 85 mosmol. No blebs
were observed after 170 mosmol swelling or re-
suspension in 340 mosmol buffer. Bleb formation
at 75 min of ischemia was indirectly related to the
osmolality of the resuspension buffer, with
51.5±7.6%, 31.0±6.7% and 2.3±2.3% of viable
squares possessing blebs at 85 mosmol, 170 mos-
mol, and 340 mosmol, respectively (Figs 1B, 1C
and 4). Cell viability was largely maintained at
75 min of ischemia after resuspension in 85, 170
and 340 mosmol buffer with 70, 85 and 92%
respectively of squares excluding trypan blue. Os-
motic fragility increased at 180 min of ischemia
but viable squares still accounted for a significant
fraction of the cell population. Bleb formation in
viable squares, after resuspension in 85 mosmol
buffer, increased only slightly from 51.5±7.6% to
62.9±6.7% at 180 min of ischemia. There was a
larger incremental increase from 31.0±6.7% to
154 S. C. Armstrong et al.
Figure 5 Graph showing the percentage of non-viable,
trypan blue permeable squares that possessed blebs visible
in India ink preparations. There were no blue squares at
the 0 min time point and only a small percentage at the
75 min time point, and thus these early time counts are
unreliable due to small sample size. A significant number
of trypan blue permeable squares were present at 180
and 240 min of ischemia. Following resuspension in 85
(circles) or 170 (triangles) mosmol buffer only a small
percentage of trypan blue permeable squares possessed
blebs. No blebbed, trypan blue permeable cells were
observed in the group resuspended in 340 mosmol buffer
(squares). Protected cell groups (filled symbols) were not
significantly different from the control groups.
48.9±6.8% after 170 mosmol swelling, while only
12.8±9.7% of cells had blebs in 340 mosmol buffer.
These results indicate that bleb formation increases
with hypo-osmotic stress and duration of ischemia.
The percent of blue squares with blebs was small
relative to the viable squares. Following 85 mosmol
swelling only 5.45±3.69% of the blue squares
possessed blebs at 240 min of ischemia (Figs 1D,
1E and 5). The delay in the onset of osmotic fragility
by cell protection (IPC/CalA) (Fig. 3) was not ac-
companied by a reduction in bleb formation at any
osmolality of resuspension buffer (Figs 4 and 5).
Transmission electron microscopy (TEM) of oxy-
genated control preparations, following 85 mosmol
swelling, showed cells with relaxed sarcomeres,
mitochondria of moderate density and an intact
sarcolemma (Fig. 6A). Cell swelling was evident
as a diffuse separation of cellular organelles, but
membrane blebs were rarely observed. Higher mag-
nifications of the sarcolemma (Fig. 6B) showed an
absence of the basal lamina, a uniform and dense
extracellular glycocalyx and salloping at the points
in which the sarcolemma was bound to the myo-
fibrils by intact costamere junctions. At 75 min of
ischemia, the majority of cells contained contracted
sarcomeres and two types of sub-sarcolemmal blebs
could be distinguished. Type II blebs (Fig. 6C) were
most common in cells in which mitochondria had
prominent or multiple amorphous matrix densities.
Type II blebs were devoid of background flocculent
material. Breaks in the continuity of the sar-
colemmal membrane covering type II blebs were
frequently observed. It was assumed, but not con-
firmed by serial sectioning, that the occasional
type II bleb with apparently intact membranes had
membrane breaks outside of the plane of section
being examined. Occasionally both blebs types
occured on the same cell. Only type I blebs were
seen on oxygenated cells. At 75 and 180 min of
ischemia, cells were observed with either type of
bleb. However, intact type I blebs were frequent at
75 min and the ruptured type II blebs were most
common after 180 min of ischemia.
An ultrastructural assessment of sarcolemmal
permeability in ischemic cardiomyocytes was per-
formed by horseradish peroxidase uptake and elec-
tron microscopy. Light microscopy demonstrated
that at 75 min of ischemia most of the contracted
cardiomyocytes (swollen in 85 mosmol media con-
taining HRP) were unstained, although many of
the round cells were HRP positive. Blebs were not
observed in the absence of India ink (Fig. 7A). At
180 min of ischemia, the majority of the contracted
cardiomyocytes were HRP-positive although some
remained HRP negative (Fig. 7B). The proportion
of HRP negative and positive cells was similar to
trypan blue stained preparations. Electron micro-
scopy demonstrated type I blebs on 75 min ischemic
cells (swollen in hypotonic media containing HRP).
Reaction product was seen adhering to the outer
surface of the intact sarcolemma overlying the bleb
and small amounts were associated with T-tubules,
but no reaction product was observed within the
bleb space. The mitochondria were swollen but did
not contain amorphous matrix densities (Fig. 7C).
Electron microscopy of 180 min ischemic car-
diomyocytes showed the presence of reaction prod-
uct within cells that had a ruptured sarcolemma.
HRP reaction product penetrated into the cells out-
lining the outer mitochondrial membranes and was
diffusely bound to the myofibrils. The mitochondria
of these cells were swollen and contained numerous
large amorphous matrix densities (Fig. 7D).
Discussion
This study tested and rejected the hypothesis that
sarcolemmal bleb formation is a primary ischemic
 Sarcolemmal Blebs and Osmotic Fragility 155

Figure 6 (A) Electron micrograph of an oxygenated control cardiomyocyte following swelling in 85 mosmol buffer.
Cell swelling is evident by the diffuse separation of cellular organelles and focal lucent areas within the cell. The cell
has relaxed sarcomeres, moderately swollen mitochondria and is surrounded by an intact sarcolemma (×4500). (B)
Electron micrograph of an oxygenated control cardiomyocyte with relaxed sarcomeres. The sarcolemma is comprised
of a lipid bilayer, which is obscured by an apposed dense internal coating of electron dense material, and an outer,
diffuse glycocalyx. The basal lamina (basement membrane) is absent due to collagenase digestion during cell isolation.
Despite exposure to 85 mosmol media prior to fixation, the sarcolemmal membrane is closely apposed to the myofibrils
at lateral costamere junctions (×37 000). (C) Type II bleb on a cell after 180 min of ischemia. Type II blebs generally
occurred in cells in which numerous mitochondria contained amorphous matrix densities. The bleb space is devoid of
dispersed lightly stained flocculent material. Sarcolemmal membranes covering type II blebs are ruptured (arrow)
(×4000).

lesion that leads to osmotic fragility and plasma
membrane rupture. An artificially high, hypo-os-
motic stress was imposed upon cells, only at the
termination of each ischemic time point, to deter-
mine the presence of a latent pre-existing lesion.
Sub-sarcolemmal blebs were not actually present on
cells during ischemic incubation. Cardiomyocytes,
within an ischemic pellet, develop an osmolar load
due to an accumulation of lactate and other meta-
bolic products of anaerobic glycolysis.1,17 This in-
ternal osmotic load was presumed to be balanced
by the relative hyper-osmolality of the ‘‘isotonic’’
(340 mosmol) resuspension medium. In the absence
of an imposed osmotic stress few cells developed
blebs and development of trypan blue permeability
progressed slowly. Resuspension of ischemic cells
in hypotonic (85 and 170 mosmol) buffer revealed
that the latent tendency for bleb formation coincides
with rigor contracture and precedes osmotic fra-
gility. Under our experimental conditions blebbing
was a function of both duration of ischemia and
degree of osmotic stress, allowing quantitative cor-
relation of these variables. The large sub-sar-
colemmal blebs described in this study must be
differentiated from microblebs, previously described
in anoxic cardiomyocytes18 as focal protrusions
of the sarcolemma enclosing single mitochondria.
Microblebs may form as a result of microtubular
disassociation19 or mechanical deformation during
schemic contracture. Microblebs were observed in
our studies in 85 mosmol swollen control cells and
in late ischemic cells resuspended in 340 mosmol
buffer but were not quantitated in the absence of
a direct correlation with ischemic injury. The large
macroblebs described in this study are comparable
to those described in irreversibly injured anoxic rat
hearts6 or ischemic–reperfused4 or autolytic3 canine
myocardium.
The original hypothesis, upon which this study
was based,20 assumed that osmotic swelling of isch-
emic cells manifested a latent ischemic lesion formed
by a detachment of the sarcolemma from adherens-
type lateral costamere junctions that link the cell
membrane to the underlying myofibrils. Increasing
wall tensions in the unattached plasma membrane
covering enlarging blebs were then postulated to
cause membrane rupture. The corollaries of this
hypothesis would be that: 1) few viable cells would
have blebs while most trypan blue permeable cells
would possess blebs; 2) the onset of bleb formation
would coincide with the onset of irreversible injury;
and 3) cell protection (IPC/CalA) would result in
reduced bleb formation. Morphological observations
by light microscopy did not validate these pre-
156 S. C. Armstrong et al.

Figure 7 (A) Light micrograph of 75 min ischemic cells swollen in 85 mosmol media containing horseradish peroxidase
(HRP). The ischemic, square cardiomyocytes are unstained (arrow), although the round cells (arrow head) are HRP
positive. Blebs are not observed in the absence of India ink (×240). (B) Light micrograph of 180 min ischemic cells
swollen in 85 mosmol media containing HRP. The majority of the contracted cardiomyocytes were HRP-positive (arrow)
although some remained HRP negative (arrow head). HRP negative and positive round cells are also shown. The
proportion of HRP negative and positive cells was similar to trypan blue stained preparations. Blebs are not observed
in the absence of India ink (×240). (C) Electron micrograph of a type I bleb on a 75 min ischemic cell incubated in
hypotonic media containing HRP. Reaction product is seen adhering to the outer surface of the intact sarcolemma
overlying the bleb and small amounts were associated with T-tubules, but no reaction product is observed within the
bleb space. The mitochondria are swollen but do not contain amorphous matrix densities (×10 700). (D) Electron
micrograph of a 180 min ischemic cardiomyocyte swollen in 85 mosmol media containing HRP. The sarcolemma
overlying the large bleb has been ruptured. HRP reaction product has penetrated into the cells outlining the outer
mitochondrial membranes and is diffusely bound to the myofibrils. The mitochondria are swollen and contain numerous
large amorphous matrix densities (×9000).

dictions, but a critical analysis of the absence of
these blebs in osmotically fragile cells is required to
reject or modify the original hypothesis. An initial
increase and subsequent decline of bleb formation
during prolonged ischemia is unlikely. The majority
of acutely swollen trypan blue cells did not have
evident blebs. However, bleb formation may have
preceded membrane rupture with subsequent bleb
collapse or indetectability due to the presence of
India ink in the bleb space. Alternatively, the thresh-
old of osmotic pressure required for rupture of
damaged membranes may be lower than that for
bleb formation. Thus, early membrane rupture
would abolish the osmotic gradient, impeding the
development of detectable macro-blebs.
The cell protection results reinforce the con-
clusion derived from the morphological data, in that
osmotic fragility occurs independent of sarcolemmal
blebbing. IPC and CalA activate similar signal trans-
duction pathways and each may afford protection
by similar mechanisms.21 Cell protection did not
reduce bleb formation in ischemic cells, but did
significantly delay the onset of osmotic fragility. If
the assumption is made that a progressive sar-
colemmal lesion developed, that was expressed ini-
tially as the related events of sarcolemmal
 Sarcolemmal Blebs and Osmotic Fragility
Figure 8 Illustration of: A) the original hypothesis upon
which this study’s experimental design was based, and:
B) a modified hypothesis based on the results of this
study. The original hypothesis assumed that a basic latent
injury responsible for the transition of ischemic cells
to an irreversible state was weakening of sarcolemmal
attachments to the underlying cytoskeleton. Imposition
of osmotic swelling was postulated to induce formation
of subsarcolemmal blebs, which were inherently fragile
and prone to rupture. The results of this study required
a revision of the original hypothesis A, to the modified
hypothesis B. An early lesion occuring coincident with
ATP-depletion contracture is a reversible weakening of
sarcolemmal attachments which can be exposed as mem-
brane blebbing during a severe osmotic stress. More
prolonged ischemia results in development of osmotic
fragility which is a irreversible event manifest upon
osmotic swelling by rupture of the sarcolemmal mem-
brane without an initial formation of subsarcolemmal
blebs. During prolonged isotonic ischemic incubation in
the absence of imposed osmotic or mechanical stresses,
a third event renders sarcolemmal membranes permeable.
detachment and membrane fragility, then cell pro-
tection should reduce bleb formation as a prelude
to a delay in the onset of osmotic fragility. The delay
in osmotic fragility by IPC/CalA, in the absence of
a reduction in bleb formation, suggests that bleb-
bing and osmotic fragility are separate mani-
festations of injury.
The observation that membrane blebbing cor-
related with ATP-depletion suggests that a po-
tentially reversible early structural lesion,
associated with weakening of membrane-cyto-
skeletal attachments, may have developed. Al-
ternatively, it is possible that bleb formation is
related to a functional disturbance in cell volume
control that potentiates hypo-osmotic cellular swell-
ing in cells with intact plasma membranes. In this
case IPC, which does not reduce the rate of ischemic
ATP depletion, would not attenuate bleb formation
in reversibly injured cells.
157
This scenario is consistent with the observation
that blebbed cells were not inherently osmotically
fragile. A minority of cardiomyocytes hypercontract
into viable, disc-shaped rounded forms during isol-
ation, with extensive blebbing under isotonic con-
ditions. Round cells are mechanically fragile22 but
are not inherently osmotically fragile since viable
rounds are seen after hypotonic cell swelling and
round cells do not exceed square cells in their rate of
development of membrane permeability. A com-
parison between the behaviour of blebbed round cells
and blebbed ischemic square cells may not be entirely
valid since the blebs that form as a result of mech-
anical distortion during rounding may not be ident-
ical with respect to membrane structure as are those
that form in ischemic cells. In any case isolated car-
diomyocytes reoxygenated in the early ATP-de-
pletion stage of injury recover in the presence of an
intact sarcolemmal membrane,23 indicating that the
latent ischemic lesion responsible for blebbing is non-
lethal and reversible. A dichotomy between blebbing
and osmotic fragility is further evidenced by the ob-
servation that the protein phosphatase inhibitors,
fostriecin and CalA, when added in late ischemia,
afford protection in in vitro ischemic24 and perfused
heart models.25 This protection from osmotic fragility
is initiated after the onset of ischemic contracture,
after bleb formation, reinforcing the conclusion that
the latent tendency for cells to bleb is reversible. The
results in cardiomyocytes are similar to those re-
ported in hypoxic liver cells26 or in ATP-depleted
renal cells,27 which conclude that blebbing of non-
muscle cell membranes is a reversible phenomenon.
The results of the present study therefore require a
modification of the original hypothesis illustrated in
Figure 8A to that depicted in Figure 8B.
The ultrastructure of cells following sudden ex-
posure to severely hypotonic media revealed that
oxygenated cells withstand this stress without de-
veloping obvious ultrastructural damage. Notably,
the mitochondria remained only moderately swol-
len and the sarcolemma retained costameric at-
tachments and a thick glycocalyx coat. Ischemic
cells appeared less able to withstand a hypo-osmotic
stress and appeared more severely swollen than
oxygenated cells. This apparent increased swelling
was associated with formation of sub-sarcolemmal
blebs and massive mitochondrial swelling. This dif-
ference in the degree of swelling between ischemic
and oxygenated cells could be due to either: 1) a
metabolic deficiency in volume control; 2) an in-
crease in the osmotic stress due to cellular ac-
cumulation of osmolytes (lactate); or 3) a
weakening of the cytoskeletal scaffolding of the cell
which permits a larger increase of cell volume for
158 S. C. Armstrong et al.
any given degree of osmotic stress. Anoxic perfused
hearts and metabolically inhibited cardiomyocytes,
which do not accumulate lactate or osmolytes,
develop osmotic fragility to a similar degree as
ischemic cardiomyocytes, indicating that the critical
event is not an increase of swelling pressure from
intracellular osmolyte accumulation. The role of
mitochondria in the pathophysiology of increasing
cell volume is undetermined. Swelling was induced
in the absence of reoxygenation and in the presence
of respiratory inhibition by sodium amytal, sug-
gesting that the massive mitochondrial swelling
was a passive response, possibly related to structural
damage. Two bleb types are observed in ischemic
cells. The first type retained a background of
amorphous material within the bleb space and had
an intact sarcolemmal membrane covering the bleb
space. While membrane rupture in another plane
of section could not be ruled out without serial
sectioning of the entire cell, it is likely that the type
I blebs correspond to the blebs seen in trypan
blue impermeable cardiomyocytes. Cells with type
II blebs and obvious ruptures of the overlying sar-
colemma might be analogous to the trypan blue
permeable cells seen by light microscopy. The latter
cells leak cytosolic proteins and sarcolemmal dis-
ruption might allow entry of colloidal carbon part-
icles, accounting for the difficulty in visualization
of blebs in India ink preparations of trypan blue
permeable cardiomyocytes. The cells possessing
both type I and II blebs may be analogous to the
blebbed and trypan blue permeable cardiomyocytes
in India ink preparations.
An electron dense extracellular marker, horse-
radish peroxidase reaction product, was used to
assess sarcolemmal permeability of cells at an ul-
trastructural level, as previously described in isch-
emic myocardium.16 By light microscopy, the uptake
of HRP (as determined by staining with reaction
product) reproduces the results obtained with try-
pan blue permeability in that neither oxygenated
cardiomyocytes (data not shown) nor reversibly
injured cells (75 min ischemia) showed any uptake
of HRP during a hypotonic incubation, although
reaction product was evident in a subset of hy-
percontracted, round cells. Irreversibly injured (os-
motically fragile) cells were detectable by HRP
uptake during a hypotonic incubation of cells sub-
jected to 180 min ischemia. The percentage of ir-
reversibly injured cells was similar when
determined by either HRP or trypan blue uptake.
The reversibly injured cells containing type I blebs
excluded HRP and intracellular reaction product
was only observed within T-tubules, with no prod-
uct detectable in the cytosol or bound to myofibrils.
Irreversibly injured cells with ruptured type II blebs
had HRP reaction product dispersed throughout the
cardiomyocyte cytoplasm. These results confirm, at
an ultrastructural level, that bleb formation occurs
prior to osmotic fragility as defined by sarcolemmal
permeability following a hypotonic challenge of
ischemic cardiomyocytes.
It is presumed that cell death during prolonged
ischemia, in the absence of an imposed osmotic
stress, is due to a progression of the lesion(s) re-
sponsible for osmotic fragility. However, the in-
creased permeability of cells during prolonged
ischemic incubation may occur by a separate mech-
anism. The phospholipase inhibitor mepacrine de-
layed cell death during sustained metabolic
inhibition of isolated rat cardiomyocytes, but did
not delay the onset of osmotic fragility.28 The ac-
cumulated evidence suggests that cell injury may
pass through three etiologically separate stages
during prolonged ischemia (Fig. 8B). The initial
event is coincident with ATP depletion and ischemic
contracture and consists of either a loss of volume
control or a potentially reversible18 weakening of
sarcolemmal attachments to the cell cytoskeleton,
allowing bleb formation. The second event, which
develops within the first hour after ATP-depletion
and ischemic contracture, is the onset of osmotic
fragility. The third event is an irreversible and
delayed loss of sarcolemmal membrane integrity,
possibly related to phospholipase degradation of the
lipid bilayer. This late stage of injury might be
related to necrosis in that the severely injured cells
are destined to die, either by oncosis or apoptosis,
despite any intervention. Temporally heterogenous
injury in a cell population produces an overlapping
of these stages in most experimental models.
Large sub-sarcolemmal blebs, covered by a dis-
continuous plasma membrane, are a prominent fea-
ture of irreversible injury in ischemic hearts,1,3,4 but
blebs are rarely seen on reversibly injured cells. Sub-
sarcolemmal bleb formation, during in vitro osmotic
swelling, does not directly correlate with irreversible
injury. However, bleb formation could be a con-
tributing factor to irreversible injury in intact hearts.
A reduction of cell swelling, during reperfusion, and
cell contracture, during reoxygenation, in-
dependently mediate protection in intact hearts.29
The osmotic load of a severely ischemic cell has been
estimated at about 60 mosmol in intact dog and pig
hearts.1,7 Thus, the osmotic forces at reperfusion of
ischemic myocardium are not as severe as those used
experimentally in this study and a latent tendency
to form blebs may not be fully expressed. The sar-
colemma is attached, via transmembrane and extra-
cellular proteins, to the interstitial collagen matrix
 Sarcolemmal Blebs and Osmotic Fragility
such that traction forces could initiate formation of
blebs by breaking sarcolemmal membrane at-
tachments to the underlying cytoskeleton. Swollen
cells with blebs may be susceptible to membrane rup-
ture as a result of mechanical stresses of contracture,
such that reoxygenation contracture could mech-
anically rupture sub-sarcolemmal blebs. A rupture of
sarcolemmal membranes during reperfusion would
initiate massive calcium influxes, contraction band
formation, enzyme release, and necrosis. Mechanical
stresses of reperfusion hypercontracture are not re-
produced in an isolated cardiomyocyte model23,30,31
and thus in vitro experiments cannot determine the
relationship between bleb formation by hypo-os-
motic swelling and mechanically fragility. Although
blebbing could potentiate irreversible injury in intact
ischemic/reperfused hearts, it is evident that bleb-
bing per se is not the major determinant of cell death.
This is indicated by the significant in vivo and in vitro
protection afforded by IPC, in the absence of an at-
tenuation of bleb formation. Bleb formation during
early ischemia, when cells are reversibly injured, in-
dicates that blebbed cells are not inherently os-
motically fragile. This study suggests that in vitro IPC
is mediated by pathways that either: 1) retard cell
swelling during hypo-osmotic stress; or 2) directly
delay the loss of sarcolemmal membrane integrity as
opposed to the original hypothesis that IPC main-
tains sarcolemmal membrane-cytoskeletal at-
tachment sites at the lateral costamere junctions.
Acknowledgements
The authors wish to thank Judy Whittimore for her
expert assistance in the processing, examination and
photography of transmission electron microscopy
specimens. This research was supported by a grant
from the National Institutes of Health, HL 51859-
04.
References
1. J RB, R KA, S C. Editorial.
Myocardial ischemia revisited. The osmolar load,
membrane damage and reperfusion. J Mol Cell Cardiol
1986; 18: 769–780.
2. S CJ, H ML, J RB. Volume re-
gulation and plasma membrane injury in aerobic, an-
aerobic and ischemic myocardium in vitro : Effect of
osmotic swelling on plasma membrane integrity. Circ
Res 1985; 57: 864–875.
3. S MD, J RB. Cytoskeletal injury and sub-
sarcolemmal bleb formation in dog heart during in
vitro total ischemia. Am J Pathol 1988; 133: 327–337.
4. K RA, G CE, W DA, J RB.
159
Effects of a transient period of ischemia on myocardial
cells. II. Fine structure during the first few minutes of
reflow. Am J Pathol 1974; 74: 399–422.
5. G CE. Contraction band necrosis and ir-
reversible myocardial injury. J Mol Cell Cardiol 1983;
15: 67–73.
6. G CE, V H RS. Cytoskeletal lesions in
anoxic myocardial injury: A conventional and high
voltage electron microscopic and immuno-
fluorescence study. Amer J Pathol 1987; 129: 327–
344.
7. T-J J, J MJ, F JWT, K JG,
D’A CN, D D. Tissue osmolality, cell
swelling, and reperfusion in acute regional myo-
cardial ischemia in the isolated porcine heart. Circ Res
1981; 49: 364–381.
8. J RB, M C, R KA. Myocardial effects
of brief periods of ischemia followed by reperfusion.
Silent Myocardial Ischemia: A Critical Apprasial. Kel-
lermann, JJ and Braunwald, E (eds). Adv Cardiol. Ba-
sel, Karger. 1990; 37: 7–31.
9. G CE, V H RS. Irreversible injury of
isolated adult rat myocytes: Osmotic fragility during
metabolic inhibition. Am J Path 1988; 132: 212–222.
10. V H RS, R D, H CM, G CE. An
in vitro model of myocardial ischemia utilizing isolated
adult rat myocytes. J Mol Cell Cardiol 1990; 22: 165–
181.
11. G CE. Cell to cell interactions contributing to the
‘‘oxygen paradox’’. Basic Res Cardiol 1985; 80 [Suppl
2]: 141–146.
12. S CJ, H ML, J RB. Cytoskeletal
damage during myocardial ischemia: Changes in vin-
culin immunofluorescence staining during total in
vitro ischemia in canine heart. Circ Res 1987; 60:
478–486.
13. V H RS, A RA, L KG, G
CE. Modification of caffeine-induced injury in Ca2+-
free perfused rat hearts: Relationship to the calcium
paradox. Am J Pathol 1986; 123: 351–364.
14. H CM, A A, A RA, B GP. Con-
tracture of isolated rat heart cells on anaerobic to
aerobic transition. Amer J Physiol 1982; 242: H1022–
H1030.
15. A SC, G CE. Effects of 2,3-butanedione
monoxime (BDM) on contracture and injury of isol-
ated rat myocytes following metabolic inhibition and
ischemia. J Mol Cell Cardiol 1991; 23: 1001–1014.
16. T E, A M. Pathologic assessment of myo-
cardial cell necrosis and apoptosis after ischemia and
reperfusion with molecular and morphologic
markers. J Mol Cell Cardiol 2000; 32: 209–224.
17. A SC, K R, G W, et al. Comparison of
in vitro preconditioning responses of isolated pig and
rabbit cardiomyocytes: effects of a protein phos-
phatase inhibitor, fostrecin. J Mol Cell Cardiol 1997;
29: 3009–3024.
18. P HM, S P, S R, H JF, S-
 PG. Anoxic injury of adult cardiac myocytes.
Basic Res in Cardiol 1987; 80 (Suppl 1): 37–42.
19. I K, H M, K A, et al. Disruption of
microtubules as an early sign of irreversible ischemic
injury. Immunohistochemical study of in situ canine
hearts. Circ Res 1990; 67: 694–706.
20. G CE, A SC. Ischemia and the myocyte
160 S. C. Armstrong et al.
cytoskeleton: Review and speculation. Cardiovas Res
1993; 27: 1387–1403.
21. A SC, D M, G CE. Phos-
phorylation state of hsp27 and p38 MAPK during
preconditioning and protein phosphatase inhibitor
protection of rabbit cardiomyocytes. J Mol Cell Cardiol
1999; 31: 555–567.
22. A RA, G L, A A, H C, K FA,
B GP. Calcium tolerance of isolated rat heart
cells. J Mol Cell Cardiol 1980; 12: 1383–1395.
23. S B, K A, K T, S P, P HM.
Sarcolemmal integrity and metabolic competence of
cardiomyocytes under anoxia-reoxygenation. Amer J
Physiol 1990; 258 (Heart Circ Physiol 27): H285–
H291.
24. A SC, G W, L JR, G CE. Protein
phosphatase inhibitors calyculin A and fostriecin pro-
tect rabbit cardiomyocytes in late ischemia. J Mol Cell
Cardiol 1998; 30: 61–74.
25. W C, B BS, L G-S, et al. Fostriecin
an inhibitor of protein phosphatase 2A, limits infarct
size even when administered after onset of ischemia.
Circulation 1998; 98: 899–905.
26. L JJ, S CJ, S JL, T
RG. Cell surface changes and enzyme release during
hypoxia and reoxygenation in the isolated, perfused
rat liver. J Cell Biol 1983; 97: 778–786.
27. D RB, Z DV, M LJ. Loss of plasma
membrane structural support in ATP-depleted renal
epithelia. Amer J Physiol 1997; 272 (Cell Physiol 42):
C439–C449.
28. A SC, G CE. Effects of the phos-
pholipase inhibitor mepacrine on irreversible injury
in ischemic and metabolically inhibited adult isolated
myocytes. Amer J Pathol 1991; 138: 545–555.
29. G CE, H SM. Effects of anoxic or oxy-
genated reperfusion in globally ischemic, isovolumic,
perfused rat hearts. Am J Pathol 1985; 120: 129–145.
30. S B, K T, S P, P HM. Tem-
porary contractile blockade prevents hyper-
contracture in anoxic-reoxygenated cardiomyocytes.
Am J Physiol 1991; 260 (Heart Circ Physiol 29): H426–
H435.
31. S̈ K-D, S P, S B, P HM.
Prevention of the oxygen paradox in the hypoxic-
reoxygenated heart (Abst). FASEB J 1991; 5: A1103.