JOINT STIFFNESS AND ARTICULAR GELLING: INHIBITION OF THE FUSION OF
ARTICULAR SURFACES BY SURFACTANT B. A. HILLS and K. THOMAS Paediatric Respiratory Research Centre, Mater Childrens Hospital, Brisbane, Australia SUMMARY It was proposed some years ago that, in osteoarthritis, one source of joint stiness arises from articular gelling, but, if so, why does this not occur in the normal joint? In a preliminary experiment using agar gels, it is shown how such fusion of gel surfaces can be inhibited by surface-active phospholipid (SAPL)both synthetic and humanas quantied by the shear stress needed to cause cleavage between samples after prolonged contact. On the other hand, normal bovine articular cartilage (BAC) does not fuse to itself, but can be made to do so if rinsed with a powerful lipid solvent known to remove the outermost layer of adsorbed SAPL along with the hydrophobicity so characteristic of the normal waxy surface it imparts. It is then shown how the inhibition of gel fusion can be restored by treating both bovine and human articular surfaces with exogenous SAPL derived from human AC and with synthetic SAPL. Samples of human articular cartilage excised from osteoarthritic hips and knees during total joint replacement showed a 55% greater tendency to fuse together than normal BAC. This was exacerbated by solvent rinsing and can be attributed to a deciency in the outermost lining of SAPL previously studied as a load-bearing boundary lubricant capable of reducing friction and wear to the remarkably low levels observed physiologically. Hence, joint stiness can be attributed, in part, to a deciency in the lubricating layer of SAPL lining the normal articular surface where it can inhibit articular gelling/gel fusion, possibly imparting other desirable physiological functions. The possibility of clinical replenishment of SAPL in the osteoarthritic joint is discussed. Krx vons: Articular gelling, Joint stiness, Osteoarthritis, Synovial surfactant. JoiN1 stiness can arise from several sources, and is Problems of unwanted adhesion are usually resolved often associated as much with structures around the in the physical sciences by applying a very thin coating joint as within the joint itself. In rheumatoid arthritis, of a release (anti-stick) agent to the surface. These a neurogenic component has also been identied [1], release agents are usually surfactants in which molec- while in osteoarthritis, which is more a disease of the ules bind to the surface by their polar ends, thus cartilage itself [2], the stiness tends to be more orientating their non-polar moieties outwards to localized [3] and is essentially mechanical in nature impart a hydrophobic surface much less conducive to [4]. Focusing on cartilage, Wright [3, 4] made a most adhesion [6]. The layer may be no more than a interesting observation when attributing stiness to monolayer, while, often, the same surfactant lining is articular gelling, but, apparently, never pursued this also a good boundary lubricant, as witnessed in every- phenomenon in detail. day life in many paper coatings [7]. Although the contribution of this component may It could, therefore, be particularly signicant that a not be as great as Wright originally envisaged, it lining of surface-active phospholipid (SAPL) has been emphasizes how the diarthrodial joint violates one of identied on the normal articular surface and shown the basic principles of lubrication engineering. It is the to possess the capability to reduce friction to the very golden rule, when designing moving systems, never to low levels recorded in the normal joint, and to do so have two surfaces of the same material sliding against under high load [8]. Moreover, this lining has been each other for fear of their welding or binding to shown to be decient in osteoarthritis [9]. As the each other when they come into contact at start-up or outermost layer of boundary lubricant deposited from shut-down as the uid lm is squeezed out. This synovial uid (SF), this coating of SAPL could also aspect is very pertinent in the joint where the articular explain the extreme hydrophobicity of the normal surface has been described in the light of more recent articular surface [10, 11] and many other features studies as a proteoglycan-rich hydrated gel supported reviewed elsewhere [ 12]. on a collagen matrix [5]. Gels such as these, whose The question which this line of deduction poses is rheology is determined largely by hydrogen bonding whether the lining of SAPL demonstrated on the [5], are highly likely to fuse together, reviving the normal articular surface can act as a release agent and, concept of articular gelling. At least, it poses the vital in particular, whether it can prevent gel fusion, includ- question of why this does not occur in the normal ing articular gels. SAPL has been shown to be a good joint and what agent could be present to prevent gel release agent, maintaining patency in the Eustachian fusion from occurring. tube [13], but this need not apply to the aqueous environment of the articular surface. This study has been designed to address these questions by undertak- Submitted 18 July 1997; revised version accepted 10 November ing four basic experiments designed to test the concept 1997. of articular gelling as an example of gel fusion and its Correspondence to: B. A. Hills, Paediatric Respiratory Research Centre, South Brisbane, Queensland, 4101, Australia. inhibition by SAPL by: (i) demonstrating that when 1998 British Society for Rheumatology 532 HILLS AND THOMAS: SURFACTANT INHIBITS ARTICULAR FUSION 533 articular surfaces (human or bovine) are rinsed with a Preliminary experiment A very simple experiment was performed in which lipid solvent to remove all lipid in Zone I cartilage, two agar blocks, as used in culture dishes, were placed those surfaces will fuse together if left in contact; together face to face after removing any excess moisture (ii) demonstrating that when the lipid is left intact by gently blotting those faces. The blocks were then on normal bovine cartilage, fusion is inhibited; (iii ) laid horizontally on a watchglass and placed in a demonstrating that when articular surfaces from refrigerator at 4C overnight after covering them with osteoarthritic hips and knees are placed in contact, aluminium foil. The following day, they were found to there is some degree of fusion which can be eliminated be fused together and, if forced apart, the plane of by applying additional human SAPL; (iv) a very cleavage did not follow the original boundary. Clearly, preliminary experiment is also included to demonstrate this was an example of gel fusion as depicted in Fig. 1. how SAPL can also inhibit the fusion of gels not The experiment was then repeated with the exception specic to the joints, such as agar, which is a classical that 200 ml of a solution of DPPC in PG were spread example of a structure held together by hydrogen across the surface (without owing over the sides) bonding [14]. before the agar gel surfaces were placed together. The following day, the two blocks separated very easily MATERIALS AND METHODS along their original plane of contact. In these studies, the degree of release is quantied as the shearing force needed to separate the surfaces Quantication of gel fusion after a standardized period of contact, shearing being In repeating the above experiment, the dierences considered more relevant to the joint than a perpendic- were not always as dramatic and it became clear that ular pull as employed in standard adhesion/release the degree of fusion needed to be quantied. The tests. In the physical sciences, release is considered a simplest index, and one considered most relevant clinic- dierent mechanism to lubrication, arising from over- ally, was measurement of the shearing force needed to coming fusion/binding as opposed to friction [7]; break the bonding caused by gel fusion, shearing also although the same agents may be used to reduce both. being the relevant mode of initiating cleavage in a sti joint. This was achieved by attaching each agar block to Materials a balsa wood strip using superglue and clamping these Nutrient agar gel was prepared from Nutrient Agar stripsone to the moveable carriage and the other to Powder (Sigma #N0394) as a solution of 23 g/l in the base plate of the machine normally employed to water and allowed to set as a slab 3 mm in thickness. measure the coecient of friction. This device is essen- This slab was then cut into blocks 2 2 cm. tially a modication of the apparatus devised by Radin Bovine articular cartilage was obtained from the and co-workers [16] for measuring kinetic friction, as patellae of steers killed at the local abattoir and placed described in detail elsewhere [8]. The basic congura- in saline at 0C within 15 min of death. Strips of tion is shown in Fig. 2. Essentially, the carriage is cartilage were excised from these bones, each measur- moved at 3 mm/min, pulled by a cord attached to a ing 2.5 cm1 cm23 mm. Similar strips of arthritic human cartilage were excised from discarded hips and knees replaced by the orthopaedic surgeons during routine total joint replacement. In the knees, tissue was excised from the lateral chondyles since those surfaces did not display the deep brillation found on the medial chondyles. In hips, care was taken to avoid any worn surfaces and any marginal areas of hyaline cartilage distinguishable as opaque white tissue [15]. Thus, for human tissue, samples were standardized as 1 1 cm. SAPL was provided in the form of i-a dipalmitoyl phosphatidylcholine (DPPC) from Sigma (#P9671). This was dispensed in two forms: either as a solution in a non-volatile solvent (propylene glycol; PG) or in a standard volatile solvent of hexane:ethanol (9:1) (HexEth) widely used as a vehicle for depositing SAPL Fic. 1.Depicting how two blocks of an aqueous gel (a), if placed onto various liquid and solid surfaces [6]. These solu- in contact (b), will tend to fuse together (c) and, depending upon tions ranged in concentration from 10 to 200 mg/ml contact time, will require an appreciable shearing force (F) to for both solvents. separate them; although cleavage (d) may not occur in the original Human SAPL was obtained by rinsing the articular plane of contact (b). This could be relevant to adhesive wear [22]. surfaces of 12 hips and 32 knees replaced at surgery On the other hand, cleavage can occur spontaneously, or at a much with Folch reagent (chloroform:methanol 2:1) and lower shearing force ( f ), if the surfaces are coated with a monolayer of a release agent such as human SAPL prior to contact (e). evaporating the elutriant to dryness under N 2 . BRITISH JOURNAL OF RHEUMATOLOGY VOL. 37 NO. 5 534 curved surfaces, the clamp was also needed to keep them at, a task made easier by reducing the sample size to 1 1 cm. The rinsing procedure followed that described in detail elsewhere [11] for the recovery of adsorbed SAPL from articular surfaces for analysis and evaluation of lubrication capability. The weight (W) needed to keep the surfaces at was kept constant, gel fusion being conveniently quantied as a dimen- sionless ratio: Cleavage index =F/W (2) Fic. 2.Demonstrating the principle of the machine used for meas- where F is the shearing force needed to initiate uring the shearing force and, hence, shear stress needed to cause cleavage. cleavage between the two agar blocks depicted in Fig. 1 or between In an ancillary experiment, we attempted to restore two articular surfaces from the same joint. In the case of articular the release/anti-fusion property of the original articular cartilage, the weight (W) clamping the surfaces together and surfaces by applying SAPL in two forms: either syn- keeping them at increases the shearing force (F or f ), requiring the force of cleavage to be quantied as a cleavage index dened by thetic DPPC or lipid harvested from rinsings of the equation (2). This simplied diagram accentuates the thicknesses of articular surfaces of human hips and knees, as the layers for demonstration purposes, while the shearing force was described in detail elsewhere [11]. applied in the plane of fusion. Human articular cartilage The above experiment was next performed upon force transducer in the plane of the fusing surfaces. This system records the maximum force (F) which is human articular surfaces excised from arthritic knees replaced at surgery as described above. Unlike normal generated just before shearing occurs between the two gel samples. The experiment is then repeated using the bovine articular cartilage, there was some degree of fusion and this was again quantied as the cleavage same gel, but interposing a very thin SAPL layer between the samples, as described above, and the new index dened in equation (2). This experiment was repeated upon the same pairs of surfaces after rinsing shearing force ( f) is recorded. The ability of the surfactant to eect release, i.e. to inhibit gel fusion, is with Folch reagent to remove any adsorbed SAPL. It is important to remember that these values of the calculated as: cleavage index, involving the rupture of gel fusion, will Release factor =100(Ff )/F% (1) be much higher than coecients of friction (kinetic or static) which involves sliding only. In this experiment, no weight (W) was used as depicted in Fig. 2 since the gel surfaces were perfectly at. RESULTS Agar gel Surfactant runs The procedure described above was repeated by It was found that SAPL greatly reduced the shearing force needed to cause cleavage of agar gel blocks held depositing the 200 ml of SAPL solution over one of the surfaces prior to contact with its counterface at concen- together overnight in the refrigerator. The results are summarized in Table I and depicted graphically in trations of 10, 50, 100, 150 and 200 mg DPPC/ml PG. A further control was performed with PG alone. Fig. 3. It can be seen that gel fusion was inhibited by SAPL whether this agent was dispensed in PG or In a second series of runs, 200 ml of DPPC solution in hexane:ethanol were touched o onto a horizontal hexane:ethanol. It can be seen that the solvent PG promotes fusion, while hexane:ethanol alone is not gel surface over which it spread spontaneously as though on water. The solvent then evaporated, leaving signicantly dierent from the untreated control. It is also noted that the shearing force is independ- the DPPC as a very thin coating. This procedure was repeated at concentrations of 10, 50, 100, 150 and ent of DPPC concentration when deposited from hexane:ethanol, but somewhat dose dependent when 200 mg/ml, all runs being performed 10 times. the solvent is PG. Otherwise, the release properties are quite similar and range from 61 to 84%. Bovine articular surfaces A preliminary experiment was performed in which On two further agar blocks, we used SAPL of human origin in hexane:ethanol when the release factors were two samples of normal bovine articular cartilage were clamped together with their articular surfaces in con- 70 and 74%, indicating the universal nature of the natural material as a release agent. tact and placed in the refrigerator as described above. The following day, the two samples fell apart upon In the preliminary experiment using the agar gels, application of Students t-test to the data in Table I removing the clamp. When this simple experiment was repeated, but both showed a highly signicant dierence (P<0.01) between controls and surfaces treated with DPPC. surfaces were now rinsed with a powerful lipid solvent ( Folch reagent) prior to contact, partial fusion was There was no signicant dierence between shearing forces for dierent concentrations of adsorbed DPPC, found to have occurred upon removing the clamp the following day. Since the cartilage was excised from a nding consistent with the basic theory of release HILLS AND THOMAS: SURFACTANT INHIBITS ARTICULAR FUSION 535 TABLE I Shearing force needed to cleave agar gels with and without an SAPL coating Maximum shearing force Surface ( f in mN) s.r.x. n Release factor* treatment F=196 8 10 Nonecontrol PG alone (control ) f =366 18 10 87% 10 mg DPPC/ml PG 76 9 10 61% 50 mg DPPC/ml PG 44 7 10 78% 100 mg DPPC/ml PG 57 8 10 71% 150 mg DPPC/ml PG 56 8 10 71% 200 mg DPPC/ml PG 31 3 10 84% Hexane:ethanol control 207 9 10 1% 10 mg DPPC/ml HexEth 77 6 10 61% 50 mg DPPC/ml HexEth 86 1 10 56% 100 mg DPPC/ml HexEth 61 6 10 69% 150 mg DPPC/ml HexEth 71 4 10 64% 200 mg DPPC/ml HexEth 65 4 10 67% *As dened by equation (1). ponding to a cleavage index of 0.12 0.017 (n =5). These results are depicted in Fig. 4. The results for bovine articular cartilage, as depicted in Fig. 4, show a highly signicant increase in shearing force needed to eect cleavage (P<0.001) when the surface is rinsed with either solvent for SAPL. However, any dierence from normal becomes insigni- cant when SAPL is reapplied either as a human extract or as synthetic DPPC. This must add strong support to the concept that these treatments are replen- ishing SAPL extracted by solvent rinsing. Human articular cartilage Human articular cartilage from osteoarthritic hips and knees tended to fuse more readily than bovine articular cartilage, as witnessed by the results given in Fic. 3.Depicting the results quoted in Table I in which it can be seen that DPPC is an eective release agent inhibiting the fusion of agar blocks whether applied to the surfaces as a solution in propylene glycol (PG) or using hexane:ethanol (9:1) as a vehicle which is allowed to evaporate. Each result is the average of 10 runs s.r.x. agents in that a monolayer should suce [6] and any additional surfactant is superuous. Bovine articular cartilage When the shearing force needed to initiate motion was measured as a cleavage index as dened in equation (2), the result for 10 samples was a value of 0.096 0.006 (means.r.x.). When the surfaces were rinsed with either Folch reagent (n =4) or hexane:ethanol (n =5), the force needed to produce Fic. 4.Depicting the relative shearing forces needed to cleave two cleavage (F) went o scale, even when W was reduced normal bovine articular surfaces clamped together and how they from 120 to 70 N (12 to 7 kg). However, when the can be made to fuse together (cleavage index o scale) if these SAPL was replenished on the surface by depositing a surfaces are rst rinsed with two powerful lipid solvents ( Folch human extract (100 mg) in hexane:ethanol as described reagent and hexane:ethanol 9:1). Results are plotted as the cleavage above, the shearing force was greatly reduced, corres- index dened by equation (2), the clamping load (W) being the ponding to a cleavage index of 0.140 0.009 (n =10). same in all runs. It is also shown how the application of human or When this experiment was repeated using DPPC synthetic SAPL, i.e. DPPC, almost restores the original release (100 mg) to replenish the SAPL elutriated by solvent properties. The shearing force is quantied as the cleavage index as dened in equation (2). rinsing, the shearing force was reduced further, corres- BRITISH JOURNAL OF RHEUMATOLOGY VOL. 37 NO. 5 536 TABLE II Shear stress needed to cleave human articular surfaces before and after solvent rinsing and replenishment of surface SAPL Surface Shear stress Cleavage index Release factor treatment (N/cm2) [equation (2)] n* [equation (1)] None f =4.36 0.62 0.141 0.020 11 55% Rinsed with Folch reagent (control ) F=9.60 1.09 0.310 0.035 12 Control treated with 500 mg human SAPL f =3.36 0.37 0.108 0.012 12 65% Control treated with 500 mg DPPC f =1.16 0.38 0.037 0.012 12 88% *n refers to samples from dierent joints. result, even including replenishment with human SAPL. DISCUSSION The simple experiment with the agar gel demon- strates the principle of gel fusion and how eectively it can be inhibited by SAPL. At the molecular level, gelling has always been associated with hydrogen bond- ing [14] and so its inhibition by SAPL could be related to the nding that even an adsorbed monolayer of DPPC can impose a barrier to hydrogen ions, decreas- ing permeability by an order of magnitude [17]. Hydrogen bonding has also been emphasized as a major factor in the structure of the very hydrophilic Fic. 5.Depicting how arthritic human cartilage derived from proteoglycan gel [5] comprising the thick surface layer replaced hips and knees displays a greater degree of fusion than of articular cartilage supported on the tough collagen normal bovine as reproduced from Fig. 4. Also demonstrated is how matrix providing structural integrity. Hence, it is not replenishment of SAPL can restore release and reduce articular surprising that articular surfaces rinsed free of the fusion almost to that of normal bovine articular cartilage. This implies a deciency of SAPL on arthritic human cartilage as outermost lipid coating are prone to gelling to them- conrmed in other studies [20]. Note how synthetic DPPC reduces selves, whether bovine ( Fig. 4) or human ( Fig. 5). This articular fusion even further. Shearing forces are compared as the outermost lipoidal layer is responsible for the very cleavage index dened in equation (2) where the clamping load (W) hydrophobic nature of the normal articular surface is the same in all runs. [10, 11], and is also an excellent boundary lubricant capable of reducing wear [ 18] and friction to the physiological range, which has been described as mira- Table II and summarized in Fig. 5. It can be seen that culously low by engineering criteria [19]. Moreover, the average cleavage index of 0.141 0.020 compares it imparts such remarkable lubrication under high load with 0.096 0.006 for normal bovine articular cartil- of >130 N/cm2 (>13 kg/cm2). This same release mech- age. When rinsed with solvent, this value was greatly anism might still apply even if other mechanisms, e.g. increased, indicating a loss of release agent. When the related to synovial cells, were contributing to the SAPL was replenished with the human extract of mechanical component of joint stiness. SAPL, the shear stress needed to cause cleavage was The observation that articular surfaces from arthritic reduced by 65%, and by 85% when replenished by joints demonstrated a greater tendency to fuse together DPPC. than normal bovine articular cartilage is compatible Statistical analysis of the results for arthritic human with Wrights original concept [4] of articular gelling cartilage ( Table II ) shows a highly signicant increase as the mechanism responsible for stiness in in shearing force upon rinsing with Folch reagent osteoarthritis. This is demonstrated by the higher (P<104). Application of human SAPL to the solv- shearing force, as quantied by the higher cleavage ent-rinsed arthritic cartilage shows a highly signicant index in Fig. 5 than recorded for normal bovine artic- decrease in shearing force (P<0.001), but the mean ular cartilage in Fig. 4. The dierence could be attrib- is lower than the original value to a moderate level of uted to a deciency of SAPL. The origin of such a signicance (P<0.05), indicating that the arthritic deciency is probably related to the change in the lipid human cartilage was probably decient in surfactant prole of synovial uid recorded soon after a non- in the rst instance. This is borne out by the moderately fracture traumatic injury of the type expected to induce signicant dierence (P<0.03) between untreated osteoarthritis in the abused joint [20]. Our recent arthritic human and untreated normal bovine. DPPC studies analysing solvent rinsings of the articular sur- appeared remarkably eective in restoring release, faces of hips and knees replaced at surgery have demonstrating highly signicant (P<0.001) decreases in the cleavage index by comparison with any other indicated deciencies of up to 60% [ 9]. HILLS AND THOMAS: SURFACTANT INHIBITS ARTICULAR FUSION 537 In the preliminary experiment with agar, it was the funding provided by the Australian Research found that when untreated gel surfaces fused together, Council. the plane of subsequent cleavage did not coincide with the original plane of contactan observation RrrrrNcrs consistent with the concept of adhesive wear in 1. Helliwell PS, Howe A, Wright V. Lack of objective osteoarthritis [21]. evidence of stiness in rheumatoid arthritis. Ann Rheum It was interesting that the potentiation of gel fusion Dis 1988;47:7548. introduced by rinsing normal (bovine) articular cartil- 2. Moskowitz RW. Osteoarthritissymptoms and signs. age with a lipid solvent could be largely reversed by In: Moskowitz RW, Howell DS, Goldberg VM, Mankin depositing more SAPL (see Fig. 4), either as the lipid HJ, eds. Osteoarthritisdiagnosis and management. elutriated from human articular cartilage or as syn- Philadelphia: Saunders, 1984:14984. thetic DPPC. This indicates that these treatments are 3. Wright V, Johns RJ. Physical factors concerned with the actually replenishing the SAPL known to be elutriated stiness of normal and diseased joints. Bull Johns Hopkins Hosp 1960;160:21531. by analysis of solvent rinsings [11]. The fact that 4. Wright V. A study of articular gelling. In: Wright V, ed. replenishment reduces the shearing force for cleavage Lubrication and wear in joints. London: Sector, 1969: to values appreciably lower than the original value (see 1349. Fig. 5) indicates that the latter was already decient in 5. Ghosh P. Non-steroidal anti-inammatory drugs and SAPL. This indication is also borne out by the moder- chondroprotection. Drugs 1993;46:83446. ately signicant dierence between arthritic human 6. Hills BA. The biology of surfactant. Cambridge: and normal bovine values. The observation that human Cambridge University Press, 1988. SAPL also inhibited the fusion of agar gel tends to 7. Hern JF. Lubricants. In: Landes CG, Kroll L, eds. Paper conrm that it is acting as a general release agent. coating additives. Atlanta: Tech Assoc Pulp & Paper The signicantly better inhibition of articular Industry, 1978:80105. gelling/gel fusion obtained with synthetic DPPC indi- 8. Hills BA. Oligolamellar lubrication of joints by surface cates a possible clinical application of this agent in active phospholipid. J Rheumatol 1989;16:8291. 9. Hills BA, Monds MK. Deciency of lubricating surfact- relieving joint stiness in osteoarthritis. Many other ant lining the articular surfaces of replaced hips and agents have been administered to the joints as lubric- knees. Br J Rheumatol 1998;37:14347. ants by intra-articular injection, as reviewed elsewhere 10. Chappuis J, Sherman IA, Neumann AW. Surface tension [22], but most proved biologically incompatible with of animal cartilage as it relates to friction in joints. Ann the human joint. This should not apply to DPPC, Biomed Eng 1983;11:43549. which is normally present in the joint, produced as 11. Hills BA. Oligolamellar nature of the articular surface. lamellar bodies identied in Type B synoviocytes [ 23, J Rheumatol 1990;17:34956. 24] just as alveolar Type II cells produce DPPC as 12. Hills BA. Synovial surfactant and the hydrophobic artic- surfactant in the lung [25]. Current clinical trials ular surface. J Rheumatol 1996;23:12335. administering exogenous SAPL (formulated in this 13. Hills BA. Analysis of Eustachian surfactant and its laboratory) to osteoarthritic knees, ostensibly as a function as a release agent. Arch Otol 1984;110:39. lubricant [26], have indicated a reduction in joint 14. Glasstone S. Textbook of physical chemistry. London: Macmillan, 1953:1196217. stiness lasting for 2 months or more. 15. Ghadially FN. Fine structure of synovial joints. London: This study demonstrates that the adsorption of Butterworths, 1983:23. SAPL tends to prevent the fusion of adjacent gel 16. Radin EL, Paul IL, Swann DA, Schottstaedt ES. surfaces, including articular surfaces, thus adding Lubrication of synovial membrane. Ann Rheum Dis another highly desirable physiological function for the 1971;30:3225. outermost lipoidal layer of SAPL, originally studied 17. Hills BA, Kirwood CA. Gastric mucosal barrier: barrier for its lubricating capabilities [27]. Other possible to hydrogen ions imparted by gastric surfactant in vitro. functions include chondroprotection and control of Gut 1992;33:103941. the water content of articular cartilage by means of 18. Hills BA. Remarkable anti-wear properties of joint sur- the hydrophobicity imparted by adsorbed SAPL, factant. Ann Biomed Eng 1995;23:1125. hydration being a common nding in osteoarthritic 19. McCutchen CW. The frictional properties of animal cartilage [28]. In recent pulmonary studies [29], it has joints. Wear 1962;5:117. 20. Rabinowitz JL, Gregg JR, Nixon JE. Lipid composition been shown that SAPL is an eective scavenger for of the tissues of human knee joints. II Synovial uid in oxygen free radicals which have been implicated in the trauma. Clin Orthop 1983;190:2928. death of chondrocytes [ 30]. Whether such speculative 21. Meachim G, Brooke G. The pathology of osteoarthritis. properties are ultimately conrmed or not, the lubricat- In: Moskowitz RW, Howell DS, Goldberg VM, Mankin ing and release properties alone would make it highly HJ, eds. Osteoarthritisdiagnosis and management. desirable to maintain the integrity of that hydrophobic Philadelphia: Saunders, 1984:2942. outermost lining of SAPL in combating osteoarthritis. 22. Fife RP, Brandt KD. Experimental modes of therapy in osteoarthritis. In: Moskowitz RW, Howell DS, Goldberg AcxNovirncrxrN1s DS, Mankin HJ, eds. Osteoarthritisdiagnosis and We wish to thank Dr Peter McMeniman for his management. Philadelphia: Saunders, 1984:54959. assistance in providing us with the hips and knees 23. Dobbie JW, Hind C, Meijers P, Bodart C, Tasiaux N, Perret J et al. Lamellar body secretion: ultrastructural which he had replaced at surgery. We also appreciate BRITISH JOURNAL OF RHEUMATOLOGY VOL. 37 NO. 5 538 analysis of an unexplored function of synoviocytes. Br 27. Hills BA, Butler BD. Surfactants identied in synovial uid and their ability to act as boundary lubricants. Ann J Rheumatol 1995;34:1323. 24. Schwarz IM, Hills BA. Synovial surfactant: Lamellar Rheum Dis 1984;48:517. 28. Bollet AJ, Nance JL. Biochemical ndings in normal bodies in Type B synoviocytes and proteolipid in synovial uid and the articular lining. Br J Rheumatol 1996; and osteoarthritic articular cartilage: II. Chondroitin sulfate concentration and chain length, water and ash 35:8217. 25. Stratton CJ. Morphology of surfactant producing cells content. J Clin Invest 1966;45:11707. 29. Ghio AJ, Fracica PJ, Young SL, Piantadosi CA. and of the alveolar lining layer. In: Robertson B, Van Golde LMG, eds. Pulmonary surfactant. Amsterdam: Synthetic surfactant scavenges oxidants and protects against hyperoxic lung injury. J Appl Physiol 1994;77: Elsevier, 1984:68118. 26. Hills BA. OA: Controversies in pathogenesis and treat- 121723. 30. Burkhardt H, Schwingel M, Menninger H, Macartney ment: Surfactant. Proceedings of the Annual Scientic Meeting of the Australian Rheumatology Association, HW, Tschesche H. Oxygen radicals as eectors of cartil- age destruction. Arthritis Rheum 1986;29:37987. Brisbane, Sydney: ARA, 1997.