The synthesis of hyaluronic acid by human synovial fibroblasts is influenced by the nature of the hyaluronate in the extracellular environment

Transcription

1 Rheumatol Int (1987) 7: Rheumataln tm'~ne~'r~at. OY Clinical and Experimental Investigations Springer-Verlag 1987 The synthesis of hyaluronic acid by human synovial fibroblasts is influenced by the nature of the hyaluronate in the extracellular environment M. M. Smith and P. Ghosh Raymond Purves Research Laboratories, University of Sydney, Royal North Shore Hospital, St. Leonards, N.S.W. 2065, Australia Received December 1, 1986/Accepted March 16, 1987 Summary. Various cell lines of human synovial fibroblasts derived from synovium obtained at the time of biopsy or total joint-replacement surgery have been established. The synthesis of 3H-labelled hyaluronic acid (HA) in these cells has been determined, and the effects of adding HA of varying molecular size to the cultured cells examined. The results obtained clearly show that the in vitro synthesis of HA by these cells is influenced by the concentration and molecular weight (MW) of the HA in their extracellular environment. Synovial fibroblasts derived from an osteoarthritic joint demonstrated the most marked response on exposure to exogenous HA, showing a stimulation of HA synthesis with preparations of weight-average molecular weight (Mw) > in a concentration dependent manner. HA preparations with Mw < showed little or no effect except at high concentrations where a suppression of biosynthesis was observed. A model to explain these findings is proposed. Key words: Synovial fibroblasts - Hyaluronate synthesis - Hyaluronate molecular size Introduction The synovium is a layer of loose, highly vascularised connective tissue lining the innermost cavity of diarthrodial (synovial) joints. The cells of this tissue, particularly the fibroblasts (type-b cells), are responsible for the synthesis of hyaluronic acid (HA), an important component of synovial fluid (SF) [1]. The remarkable rheological properties of SF are dependent on the HA concentration as well as its molecular weight (MW) [2-4]. In arthritic joints the concentration and molecular size of HA usually declines [5-71. The inflamed synovial membrane is more permeable to plasma proteins, and consequently the synovial volume increases [2]. A recent paper by Dahl et al. [8] showed, using non-commercial gels, that the MW of HA within the joint falls, on average, from 7.0x 106 in normal SF to about 4.8x 106 in SF from patients with rheumatoid arthritis (RA), although at least 30% of the HA present in the latter was still of high MW. These workers suggested that while depolymerisation of normal HA occurs within the pathological joint, dilution of SF by plasma dialysate is of major importance. It has been suggested that depolymerisation of HA may be due to lysosomal hydrolases such as N-acetylglucosaminidase, fl-glucuronidase and a hyaluronidase [9, 10]; however, these enzymes are optimally active at acid ph, and it is difficult to rationalise their action extracellularly. A more likely mechanism involves the mediation of free radicals, particularly those derived from singlet oxygen (ODFR), which are generated in significant quantities by activated polymorphonuclear leukocytes (PMNs) and macrophages that accumulate within the inflamed joint [11]. Free radicals have been shown to rapidly depolymerise HA in vitro [12-16], and evidence is accumulating which implicates their role in the degradation of HA in vivo [17]. Furthermore, these highenergy species are toxic to flbroblasts [18] and inhibit HA and proteoglycan synthesis by chondrocytes [19-21]. While degradation of some synovial-fluid HA within the inflamed joint is of undoubted significance, the possibility also exists that HA synthesis by synovial fibroblasts is disturbed in the pathological state. HA synthesised in vitro by cultures of normal fibroblasts appears to be similar to that found in the synovial fluid of the joint from which the cells were isolated. However, synovial cells derived from pathological joints demonstrate an abnormal pattern of synthesis, which is maintained through multiple passages [5, 6, 22-24]. These cells can thus provide a useful means for comparative investigation of HA synthesis by fibroblasts from normal and abnormal tissues. Numerous studies have examined the characteristics of HA synthesised by fibroblasts in vitro [5-7, 22, 23, 25-29] and in vivo [2, 8, , while other studies have described the effects of such diverse substances as corticosteroids [5, 34-37], Arteparon [38-40], sulphated dextrans [40] and interleukin-1 [41, 42] on HA biosynthesis. Within recent years, intra-articular administration of HA has been proposed as a treatment for certain arthropathies, and clinical trials have been conducted in humans [43-46] and in horses [47-52].

2 114 Although the results of these trials have been variable, some benefits have been claimed, particularly with the equine arthropathies. Unfortunately, the scientific basis for such intra-articular therapy has not been clearly defined and has led to a proliferation of commercial HA preparations of variable molecular size, all of which are claimed by the manufacturers to be effective. In the present study we have examined the effects of a mimber of commercial HA preparations, as well as a series of HA standards of well-defined weight-average molecular weight (19Iw), on the synthesis of HA by human synovial fibroblasts isolated from both normal and pathological joints. The results indicate that the in vitro biosynthesis of HA by these cells is sensitive to the concentration and molecular size of the HA to which they are exposed. Materials Dulbecco's modified Eagle's medium (DME) and trypsin (from bovine pancreas; EC ) were obtained from GIBCO (New York, USA); collagenase (from Clostridium histolytieum; EC ) from Sigma Chemical Co. (St. Louis, USA); foetal calf serum (FCS) from Commonwealth Serum Laboratories (Melbourne, Vic., Australia); hyaluronidase (Strel)tomyces hyalurolyticus nov. sp.; EC ) from Calbiochem-Behring (Aust) Pry. Ltd. (Kingsgrove, NSW, Australia) and sodium [3i.i] acetate (specific activity= 2-5mCi/mmol) from Amersham Australia Pty. Ltd., (Surry Hills, NSW, Australia). Gentamicin (as Garamycin) was obtained from Essex Laboratories Pty. Ltd. (Baulkham Hills, NSW, Australia). Sepharose CL-2B, pure preparations of HA of defined 1VIw as described in Table 1 and Hylartil were supplied by Pharmacia (Southseas) Pty. Ltd., Sydney, Australia. Hyalovet was a product of Fidia S.p.A. (Abano Terme, Italy) and Remobilase was obtained from Arnolds Veterinary Products Ltd. (Boronia, Vic., Australia). All HA preparations were derived from rooster combs. Methods Tissue sources. Biopsies of synovium from patients undergoing investigative arthroscopies were classifed as "normal" if there was no gross evidence of cartilage degeneration or synovitis. Synovium from osteoarthritic joints was obtained at the time of surgery from patients undergoing prosthetic knee-joint replacements. All biopsy material was kindly provided by Dr. David Sonnabend of this hospital. Rheumatoid synovial fibroblasts, also derived from specimens obtained at the time of joint-replacement surgery, were generously provided by Dr. Victor Danis, also of this hospital. Culture methods. Fibroblasts were prepared from tissue samples by a modification of the method of Dayer et al. [53] as described previously [40]. Briefly, diced samples of the synovial membrane were washed twice with Dulbecco's calcium and magnesium-free phosphate-buffered saline (CMF-PBS) and then placed in 5 ml serum-free DME containing 40 ~tg gentamicin/ml and 4 mg/ml collagenase for 3 or 4 h at 37 C. The digest was mixed regularly. After the incubation, an equal volume of 0.05% trypsin and 0.02% ethylene diamine tetraacetic acid (EDTA) in CMF-PBS was added, and a further 1 h incubation under the same conditions was performed. The resulting cell suspension was centrifuged for 3 rain at 400 g; the cell pellet was washed three times in DME containing 10% inactivated FCS and then plated in a 25 cm 2 culture flask. The media were changed every 2-3 days, and the cells subcultured on reaching confluency (usually after 7-10 days depending on the cell line). Cells were maintained at 37 C in a moist atmosphere of 95% air and 5% CO2. For experiments, cells were plated in 10 cm ~ wells at a density of cells/well and used 4-7 days later. Experiments involved incubating cells at 70% confluency with 50 gciph]-acetate/well in 2.0 ml DME+ 10% FCS for 24 h at 37 C in the presence and absence of test substances. Each experiment was conducted on cells from the same line and passaged at the same stage of confluency with appropriate controls to minimise variability due to these parameters. The medium was then removed and the cells rinsed twice with PBS, the washings being added to the media samples. These samples were then assayed in triplicate for radioactive HA. Assay for [3H]HA. Samples were kept on ice unless otherwise stated. 1.0 ml aliquots of media were placed separately in 10 ml plastic tubes sealed with dialysis membrane. These tubes were then inverted in 0.1 M sodium acetate buffer, ph 5.6, which was changed twice a day for 2 days at 4 C. The samples were then made up to 2.0 ml and 2x0.8 ml aliquots placed in clean tubes. To one of these aliquots was added 5 turbidity-reducing units (TRU, as defined by Calbiochem-Behring) of Streptomyces hyaluronidase; the samples were mixed and tightly sealed, and both aliquots (with and without enzyme) incubated at 60 C for 3 h. Dialysis membranes were again placed over the tubes and the samples were redialysed as before in the same buffer. The dialysed sample volumes were then made up to 2.0 ml and a 0.5 ml aliquot counted for radioactivity. The radioactivity due to newly synthesised HA was calculated as disintegrations per minute (dpm) of the blank sample minus the dpm of the hyaluronidase-treated sample corrected for cell number or DNA content. DNA was measured after lysing cells with 0.1% trypsin containing 0.02% EDTA and solubilising the DNA in perchloric acid as described by Burton [54]. Table 1. Exogenous HA preparations studied HA Weight-average Source Protein preparation molecular weight content (Mw) % (w/w) Hylartil Pharmacia 0.22 Hyalovet Fidia 0.10 Remobilase Sterivet 1.48 H Pharmacia 0.10 H Pharmacia 0.14 H Pharmacia 0.20 H Pharmacia 0.21 Gel-exclusion chromatography. Radioactive HA samples were chromatographed on a Sepharose CL-2B column ( cm) equilibrated with 0.5 M sodium acetate, ph 6.8, at 4 C, eluting at 12ml/h. The column was calibrated with Hylartil (hyaluronic acid, 1VIw=3.8 l0 s) and [3H]acetate to mark the void volume (V o) and total volume (Vt) respectively. Fractions were monitored for radioactivity and/or hexuronic acid [55]. Protein determination. The protein content of the HA preparations was measured using the method of Peterson [56] after the samples and standard protein solutions had been incubated with 5 TRU Streptomyces hyaluronidase for 3 h at 60 C.

3 115 Results Measurement of HA synthesis. The method described for the determination of newly synthesised HA by synovial fibroblasts and released into the media proved to be sensitive, reproducible and cost-efficient, and allowed as many as a hundred samples to be processed at a time. Tritiated acetate, rather than the more specific [14C]glucosamine or [SH]glucosamine was used since this was readily incorporated into HA (J. R.E. Fraser, personal communication) and was considerably cheaper than the radiolabelled aminosugars. While acetate was also incorporated into proteins, the use of a highly purified Streptomyces hyaluronidase preparation, which was specific for hyaluronic acid [57] eliminated this potential problem. An additional papain digestion step to degrade proteins was found to be unnecessary, and in fact was detrimental, as the cysteine used to activate this enzyme also depolymerised the newly synthesised HA. Sepharose CL-2B chromatography was used to confirm that the Streptomyces hyaluronidase specifically degraded all of the newly synthesised HA. Figure 1A shows the media radioactivity profile before and after incubation with normal fibroblasts. Figure 1 B shows that in the absence of hyaluronidase no breakdown of included material occurs on incubating at 60 C for 3 h, and by comparison with Fig. 1A, no included material is lost on dialysis. Figure 1 C demonstrates that most of the included material is degraded by Streptomyces hyaluronidase and efficiently removed by redialysis. From 4 to 6 h after addition of [3H]acetate to the cultures, radio-labelled HA was released into the media surrounding the cells, the level of which reached a plateau at approximately 48 h (Fig. 2). From these data, we selected 24 h as a suitable incubation time for routine radio-labelling studies, as the system was not at equilibrium, and any differences in the rate of HA synthesis could be readily detected. Using the experimental conditions described, HA synthesis was found to be proportional to both cell number and DNA content. Cells were used during the growth phase (at approximately 70% confluency) rather than when confluent, as the rate of HA synthesis in vitro was found to be greater at this stage, as has been described previously [39]. The amount of labelled HA still associated with the cell layer was also measured in a number of experiments and found not to vary from the control value (approx. 5% of the amount of [3H]HA released into the media). This suggests that the size of the pericellular coat remains constant regardless of changes in the rate of HA synthesis. Thus, the measured changes in the amount of [3H]HA released into the media was an accurate estimation of the rate of HA synthesis by these cell lines. HA synthesis by "normal" and pathological cell lines. Radio-labelled media from a number of cell lines were fractionated on Sepharose CL-2B. Although the tritiated acetate was not specific for the polysaccharide, HA was the E 2~ IX. lgt u "O rr ~ Distribution Coefficient Fig. I A-C. Sepharose CL-2B chromatograms ~f culture media obtained from 24 h incubation of normal synovial fibrohlasts With [3H]acetate. A Media+[~H]acetate before (... ) and after ( ) cell incubation. B Dialysed media before ( ) and after (... ) incubation at 60 C for 3 h with no hyaluronidase. C Dialysed media treated with Streptomyces hyaluronidase for 3 h at 60 C before ( ) and after (... ) re-dialysis Q_ '2 Incubation Time (hours) Fig. 2. Kinetics of synthesis of 3H-labelled hyaluronic acid by normal synovial fibroblasts. Results show the mean + SEM (n = 3) major macromolecular product of these cells, and profiles of radioactivity provided a good indication of the hydrodynamic size of the newly synthesised HA during the 24 h experimental period. The media were examined directly without modification or manipulation, therefore it was unlikely that artifactual depolymerisation of the HA had occurred. A representative sample of the gel-filtration profiles obtained for the different cell lines (Fig. 3) showed wide variation in the hydrodynamic size of HA newly 1 i vt

4 i IOO0 50O 0 i Oi , 0.0 0,5 1.0 vo I! --, J o.o o'.5 1;o DlstrlbuUon Coefficient D Fig, 3. Sepharose CL-2B chromatograms of culture media obtained from 24 h incubation With: A and B "normal", C and D rheumatoid, E and F osteoarthritic human synovial fibroblasts with [3H]acetate. Note the small amount of high-mw HA synthesised by osteoarthritic (OA) and rheumatoid (RA) cell lines, as indicated by exclusion from the column synthesised by fibroblasts derived from joints affected by the various diseases. Generally, "normal" fibroblasts appeared to synthesise a high proportion of void-volume material, i.e. HA with a large proportion of molecules of MW> Rheumatoid synovial fibroblasts synthesised a higher proportion of HA which was included in the column, i.e. HA of reduced MW and high polydispersity. Similar results have been previously reported by Vuorio and others [6, 7, 39]. The cell lines derived from joints of patients with osteoarthritis (OA) synthesised less total HA than normal and RA fibroblasts, and the polysaccharide was also highly polydisperse (Fig. 3). Contrary to the findings of other workers [5, 6], we found that the rate of synthesis of HA in the RA-derived cell lines tended to be the same or lower than in cells from "normal" joints (Table 2). Since only a limited number of cell lines were examined in the present study, this conclusion is still open to question. It should be noted, however, that the fibroblast lines investigated retained the same pattern of HA synthesis through multiple passages, as has been reported by others [6]. All cell lines were discarded after the tenth passage, even though no visible changes in the morphology of the fibroblasts were evident. Influence on HA biosynthesis of exogenous HA added to culture media. The sources, protein contents and weightaverage molecular weights (IVlw) of the HA preparations used in these experiments are listed in Table 1. Except for Remobilase, which has 1.48% (w/w) protein, the protein content of all the HA preparations was determined to be less than 0.25% (w/w). The IVlw data were calculated from information provided by the supplier, but were compatible with the partition co-efficients determined by gel filtration and shown in Fig. 4. The effects of these pure HA preparations, at the concentrations indicated, after incubation with the fibroblast cultures for the 24 h are summarised in Table 3 and represented graphically in Figs. 5 Table 2. Radioactive acetate incorporation into hyaluronic acid (HA) by various cell lines. RA= rheumatoid arthritis; OA=osteoarthritis; n = no. of experiments; 2 = mean value; SEM = standard error of the mean, dpm = disintegrations per minute Cell Age/sex Pathology HA synthesis line of donor ofsynovium (dpm/24 h/1000 cells) 2±SEM 1 61/male "Normal" /male "Normal" 754_ /female RA 560_ /female RA 551_ /female RA 754_ /female RA 756_ /female RA 572_ /male OA 318_ /female OA 413_ and 6. Hylartil (1Vlw= 3.8 x 10 ~ ) stimulated HA synthesis in a dose-dependent manner in all cell lines tested. However, the stimulation was particularly strong in a cell line derived from osteoarthritic synovium. The stimulation in this case reached a maximum of 170% above the control value at 400 ~tg Hylartil/ml, which was the highest concentration used. HA synthesis was also stimulated by Hylartil in the rheumatoid and "normal" cell lines, but to a lesser extent. Hyalovet (Mw=6.2x 10 s) caused a slight stimulation at 400 ~tg/ml in the rheumatoid and normal lines, but produced this effect at lower concentrations in one osteoarthritic cell line. Overall, however, the stimulation of synthesis was much less significant than for Hylartil. Remobilase (l~iw = ) had little stimulatory effect and in some cell lines inhibited HA synthesis, in one case by as much as 60% (results not shown). The Pharmacia HA standards (Table 1 ) were examined in osteoarthritic and rheumatoid cell lines to determine n

5 O0O,o-/. vt HA 34O,00O /! O.2.. t 0.1 " -"'" ' ' i... 0 HA,54o,ooo 0,1 0 "' ' '" " T~ ' 0 0,2 a ~a 0.1 o 2...!! j 6 "~-" -' , 'o l" l i" l HA 4.70O.00O ---! Fig. 4. Sepharose CL-2B chromatograms of culture media obtained from 24 h incubations ofsynovial fibroblasts from osteoarthritic joints with [~H]acetate ( ) and exogenous HA (200 ix/ml) of various lvlws, which are shown in each panel. Exogenous HA profiles (... ) were determined by hexuronic-acid analysis [55] Table 3. Effects of exogenous HA on HA synthesis by a number of human synovial fibroblast lines. Results (2_+ SEM, n= 3) are expressed as dpm/24 h/gg DNA HA preparation added Concentration ofhapreparation added (gg/ml) Osteoarthritic cell #he 9 (Figs. 5 A and 6 A) Hylartil Hyalovet Remobilase H _+ 208 H9 H5 H3 Rheumatoid cell line 5 (Fig. 5 B) Hylartil 5 650_+ 480 Hyalovet Remobilase Normal cell line 2 (Fig. 5 C) Hyartil 5 270_+ 100 Hyalovet Remobitase Rheumatoid cell line 6 (Fig. 6 B) H H9 H5 H _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ t _ _+ 60

6 t00 "N 0 ~, 1oo "6 d whether the correlation between stimulation of HA synthesis and IVlw of exogenous HA displayed by the commercial preparations could be confirmed. This was found to be the case (Fig. 6); however, with the standard of l(4w=4.7x 106, the stimulation exhibited in the RA and OA cell lines was less than with Hylartil and declined with increasing concentration. Although the HA standard with lvlw = 8.8 x 105 stimulated biosynthesis in an osteoarthritic line, it was not effective in a rheumatoid line and the two smaller HA standards (1Vlw = and ) showed no stimulatory effect in either cell line. :~ 0 N Hylartil Hyalovet Remobilase 10ug/ml 25ug/mi 50ug/ml 100ug/ml 200ug/ml 400ug/ml [] [] [] [] i Fig. 5. Histographical representation of the effects of three commercial HA preparations over the concentration range ~tg/ml on the synthesis of [3H]HA by synovial fibroblasts derived from osteoarthritic A, rheumatoid B and "normal" C knee joints. Results shown are mean + SEM (n = 3) A C Effect of exogenous HA on the hydrodynamic size of newly synthesised HA. The synovial fibroblasts obtained from osteoarthritic joints appeared to show the greatest stimulation of HA synthesis on incubation with the higher 1Viw preparations. Media samples from these cells incubated in the presence and absence of the various exogenous HA preparations at a concentration of 200 ~tg/ml were fractionated on Sepharose CL-2B (Fig. 4) and assayed both for radioactivity (to give a measure of newly synthesised HA) and hexuronic acid (to detect the presence of the added HA). It was found that the hydrodynamic size-distribution of HA synthesised by this synovial fibroblast line was not substantially influenced by the MW of the exogenous HA preparations to which it was exposed during the culture period (Fig. 4). (See note added in proof.) Discussion < -I , ,000 MWof Exogenous HA Standards Fig. 6. Histograph_ical representation of the effects of HA standards of known Mw over the concentration range ~tg/ml on the synthesis of [3H]HA by synovial fibroblasts from osteoarthritic (A) and rheumatoid (B) knee joints A B These in vitro studies suggest that synovial fibroblasts obtained from knee joints of patients with well-defined chronic OA synthesise HA at a lower rate than cells derived from "normal" and RA synovia. As we did not have access to synovial tissues from healthy individuals, our so-called normal cell lines were obtained from biopsies taken at the time of investigative arthroscopy and classified by the surgeon as normal. However, since arthroscopy is generally not undertaken without clinical evidence of joint dysfunction or pain, the classification of our tissues as normal may be questioned. In contrast, the tissues obtained at the time of total joint replacement were pathologically abnormal and the classification provided can be considered valid. Our finding that the synthesis of HA by synovial fibroblasts derived from pathological joints was influenced by the concentration and molecular size of HA present in their extracellular environment appears to be novel. These cells responded to the presence of high molecular-weight HA (MW>5x l0 s) in culture media by increasing the biosynthesis of HA. This response was concentrationdependent (Figs. 5 and 6). The most pronounced effect was observed using synovial cells derived from an OA joint. In the light of the small percentage of protein present in the HA preparations, it is possible (but unlikely) that

7 119 It II M'wHA < 500, ,000 < Mw HA < 4,000,000 Mw HA > 4,000,000 A g C Fig. 7A-C. Proposed model of HA binding to receptors on the surface of synovial fibroblasts. A With HA preparations of Mw < weak binding occurs and the biosynthesis of HA is not stimulated. B With exogenous HA of < IVlw < , it is proposed that strong binding occurs and that, because of the number of receptors stimulated (simultaneously?), HA biosynthesis is enhanced. C While maximal receptor binding occurs with high-mw HA (1Vlw > ), the large domains of these molecules limit the number of sites that can be occupied on the cell surface. We propose that HA biosynthesis is not strongly stimulated by such a configuration. At high concentrations of the high-mw species, binding by smaller HA species may be reduced due the preferential binding of the former. This may decrease the stimulation of HA biosynthesis the inhibition of HA synthesis observed at high exogenous HA concentrations could be due to a biologically active inhibitory protein present in these preparations. Previous studies by Balazs and Darzynkiewicz [58] had shown that low concentrations (0.1 mg/ml) of HA (molecular size not defined) caused a several-fold increase in the incorporation of [3H]-glucosamine into HA synthesised in vitro by fibroblasts obtained from monkey subcutaneous tissue. At higher concentrations (2 mg/ml) incorporation was depressed. From these experiments, it was concluded that HA present in the extracellular environment of these cells induced a specific effect on the metabolism of fibroblast surface glycoproteins, which in turn modified the biosynthesis of HA. The effects of extracellular HA on the biosynthesis of proteoglycans (PG) by chick and porcine articular chondrocytes and human fibroblast lines have been investigated by Wiebkin and Muir [59-62] and others [63-65]. Low concentrations (< 1.0 btg/ml) of HA were found to inhibit PG synthesis by the cartilage chondrocytes, but not by chick corneal epithelium [66], muscle fibroblasts [67], mature human skin or synovial fibroblasts [59]. In the latter experiments, only low concentrations of exogenous HA were examined ( b~g/ml) and the molecular size of the HA preparation was not defined. These previous studies are not altogether consistent with the present observations using "normal" fibroblasts, since we find that these cells respond by enhanced HA synthesis in the presence of high-mw HA at 10 btg/ml (Fig. 5). While it is possible that the HA preparations used by previous workers had MWs somewhere between that of Hylartil and Hyalovet (Fig. 5), which in our system would provide a borderline situation, a more likely explanation for the discrepancy relates to the different fibroblast cell lines used and perhaps the classification of our cells as "normal". Although an early experimental study [68] failed to demonstrate a beneficial effect of HA administered intraarticularly to rabbits in which a joint arthropathy had been induced by immobilisation, subsequent experimental and clinical studies in horses [47-50, 52] suggested that such therapy was useful in some instances. The arthrotides affecting racehorses are generally of traumatic etiology and are accompanied by synovial-membrane inflammation, HA depolymerisation and infiltration of PMN leukocytes into the joint [69]. Apart from the immediate restoration of synovial-fluid viscosity and rheological properties, the high l~lw HA preparations were considered, on the basis of Balazs' earlier studies [3, 58], to restrict the entry of PMNs into the joint fluid from capillaries of the inflamed synovium. Although such an explanation wourd be reasonable for the short-term relief of symptoms, it fails to account for the long-term benefits observed, particularly as the HA so administered has been shown to be cleared from the joint synovial fluid within a few days of administration (G. Lindblad, personal communication). The results obtained in the present study offer one possible solution to this problem, for if the synovial cells of the inflamed joint respond as indicated by our in vitro experiments, their contact with HA of a certain molecular size could result in stimulation of HA synthesis by these cells. Once initiated, such a process might be self-sustaining, provided the rate of HA biosynthesis exceeds the rate of its metabolism. Such a view is consistent with the clinical observations of Peyron and Balazs [43], who found that the limiting viscosity number of synovial fluid aspirated from osteoarthritic joints was higher 1-3 weeks

8 120 after intra-articular administration of high-molecularweight HA (Healon). An important caveat to this explanation is that the molecular weight of the HA introduced into the joint should not be less than 5 x HA preparations with molecular weights within this range were found to be only marginally effective, while those of lower hydrodynamic size were inactive or suppressed biosynthesis. An intriguing finding in the present study was that the very high 1VIw HA fractions (IVIw= ) were less effective in stimulating HA synthesis by the RA and OA cell lines than Hylartil (IV[w = 3.8 x 106 ). We interpret this result to indicate that synovial fibroblasts do not increase the biosynthesis of HA when the HA in their environment is of a molecular weight within a range which could be functionally acceptable. It is noteworthy in this regard that the MW distribution of HA present in synovial fluid from normal individuals is significantly higher than that obtained from synovial fluid of arthritic joints, i.e. the stimulus for increased biosynthesis of HA may only become operational when the hydrodynamic-size distribution of extrinsic HA falls within a particular mean range. The manner in which the synovial fibroblasts sense their extracellular HA environment and translate such information into biosynthetic mechanisms is presently unknown. However, a recent study by Prehm [70] using a hybrid B6-cell line has shown that HA is synthesised at the inner surface of the plasma membrane and extruded from the cell surface as a continuous filament. It is not unreasonable to expect that such a synthetic process would necessitate a means of monitoring the amount of HA released into the pericellular region: specific cell-surface receptors for this macromolecule would seem to be a necessary requirement for homeostasis. These receptors might also be activated by exogenous HA. Such a view has long been promulgated by Underhill and Toole [71-73] who showed, using transformed 3T3 cells, that as the MW of exogenous HA increases, so does the affinity of the HA for cell-surface receptors. However, the number of molecules bound at saturation decreased, and these workers proposed that high-mw HA binds to multiple receptors. Laurent et al. [74], using liver endothelial cells confirmed the relationship between binding of HA to cell receptors and the length of the HA chain, but also proposed that high-affinity binding occurs with larger HA molecules by virtue of the increased probability that the repetitive sequence along the HA chain will interact fruitfully with the cell receptors. This theory discounts the hypothesis of co-operative multiple-site attachment suggested by Underhill and Toole [71]. The Laurent et al. study would indicate that when the hydrodynamic size of the exogenous HA is very great, as for example with our 1~Iw=4.7 x 10-6 preparation, the number of molecules that could occupy the cell-surface receptors available would be reduced, due to size exclusion. If a specific number of receptor sites need to be activated simultaneously to promote biosynthesis of HA by the cell, then only HA molecules of a particular molecular domain could produce a maximal response. According to this model, HA molecules of very small size and low binding affinity would fail to trigger a successful response. A simplified representation of these concepts is shown in Fig. 7. Acknowledgements. We gratefully acknowledge financial support from the National Health and Medical Research Council and the competent technical assistance of Ms. Carole Grey. We also thank Ms. Janet MacKenney for typing this manuscript. References 1. Fraser JRE, Clarris BJ, Baxter E (1979) Patterns of induced variation in the morphology, hyaluronic acid secretion and lysosomal enzyme activity of cultured synovial cells. Ann Rheum Dis 38: Balazs EA, Watson D, Duff IF, Roseman S (1967) Hyaluronic acid in synovial fluid. I. Molecular parameters of hyaluronic acid in normal and arthritic human fluids. Arthritis Rheum 10: Balazs EA (1974) The physical properties of synovial fluid and the special role of hyaluronic acid. In: Helfet A (ed) Disorders of the knee. Lippincott, Philadelphia, pp Cullis-Hill D, Ghosh P (in press) The role of hyaluronic acid in joint stability - a hypothesis for hip dysplasia and allied disorders. Med Hypotheses 5. Castor CW, Dorstewitz EL (1966) Abnormalities of connective tissue cells cultured from patients with rheumatoid arthritis. I. Relative unresponsiveness of rheumatoid synovial cells to hydrocortisone. J Lab Clin Med 68: Vuorio E (1977) Rheumatoid disease in cultured human synovial cells - a biochemical study on glycosaminoglycans, proteins and plasma membranes of synovial fibroblasts in culture. Scand J Clin Lab Invest 37 (suppl 149): Vuorio E, Einola S, Hakkarainen S, Penttinen R (1982) Synthesis of underpolymerised hyaluronic acid by fibroblasts cultured from rheumatoid and non-rheumatoid synovitis. Rheumatol Int 2: Dab1 LB, Dahl IMS, Engstrom-Laurent A, Granath K (1985) Concentration and molecular weight of sodium hyaluronate in synovial fluid from patients with rheumatoid arthritis and other arthropathies. Ann Rheum Dis 44: Stephens RW, Sutherland J, Ghosh P, Taylor TKF (1976) Human serum and synovial fluid hyaluronidase - bovine testicular hyaluronidase is not a valid substitute in drug evaluation studies. Biochem Pharmaco125: Kar NC, Cracchiolo A, Mirra J, Pearson CM (1976) Acid, neutral and alkaline hydrolases in arthritic synovium. Am J Clin Patho165: Greenwald RA, Moak SA (1986) Degradation of hyaluronic acid by polymorphonuclear leukocytes. Inflammation 10: Pigman W, Rizvi S, Holley HL (1961) Depolymerisation of hyaluronic acid by the ORD reaction. Arthritis Rheum 4: Matsumura G, Herp A, Pigman W (1966) Depolymerization of hyaluronic acid by autoxidants and radiations. Radiat Res 28: McCord JM (1974) Free radicals and inflammation: Protection of synovial fluid by superoxide dismutase. Science 185: Greenwald RA, Moy WW (1980) Effect of oxygen-derived free radicals on hyaluronic acid. Arthritis Rheum 23: McNeil JD, Wiebkin OW, Betts WH, Cleland, LG (1985) Depolymerisation products of hyaluronic acid after exposure to oxygen-derived free radicals. Ann Rheum Dis 44: Puig-Parellada P, Planas JM (1978) Synovial fluid degradation induced by free radicals. In vitro action of several free radical scavengers and anti-inflammatory drugs. Biochem Pharmacol 27:

9 Simon RH, Scoggin CH, Patterson D (1981) Hydrogen peroxide causes the fatal injury to human fibroblasts exposed to oxygen radicals. J Biol Chem 256: Bates EJ, Lowther DA, Handley CJ (1984) Oxygen freeradicals mediate an inhibition of proteoglycan synthesis in cultured articular cartilage. Ann Rheum Dis 43: Bates EJ, Johnson CC, Lowther DA (1985) Inhibition of proteoglycan synthesis by hydrogen peroxide in cultured bovine articular cartilage. Biochim Biophys Acta 838: Bates EJ, Lowther DA, Johnson CC (1985) Hyaluronic acid synthesis in articular cartilage: An inhibition by hydrogen peroxide. Biochem Biophys Res Commun 132: Smith CA, Hamerman D (1969) Signifcance of persistent differences between normal and rheumatoid synovial membrane cells in culture. Arthritis Rheum 12: Marsh JM, Spencer S, Wiebkin O, Maini RN, Dumonde DC, Muir H (1974) Heterogeneity of human synovial tissue cultures. Ann Rheum Dis 33: Castor CW, Lewis RB (1976) Connective tissue activation X. Current studies of the process and its mediators. Scand J Rheumatol 5 (suppl 12): Smith CA, Hamerman D (1971) Properties of synovial cells in culture. J Exp Med 134: Castor CW, Dorstewitz EL, Rowe K, Ritchie JC (1971) Abnormalities of connective tissue cells cultured from patients with rheumatoid arthritis II. Defective regulation of hyaluronate and collagen formation. J Lab Clin Med 77: Castor CW, Smith SF, Ritchie JC, Dorstewitz EL (1971) Connective tissue activation II. Abnormalities of cultured rheumatoid synovial cells. Arthritis Rheum 14: Hamerman D, Smith C, Keiser HD, Craig R (1982) Gtycosaminoglycans produced by human synovial cell cultures. Coll Rel Res 2: Dahl IMS, Husby G (1985) Hyaluronic acid production in vitro by synoviat lining cells from normal and rheumatoid joints. Ann Rheum Dis 44: Barker SA, Bayyuk SHI, Brimacombe JS, Hawkins CF, Stacey M (1963) Fingerprinting the hyaluronic acid component of normal and pathological synovial fluids. Clin Chim Acta 8: Kofoed JA, Barcelo AC (1978) The synovial fluid hyaluronic acid in rheumatoid arthritis. Experentia 34: Balazs EA, Briller SO, Denlinger JL (1980) Na-hyaluronate molecular size variations in equine and human arthritic synovial fluids and the effect on phagocytic cells. Semin Arthritis Rheum 11 (suppl 1): Bjelle A, Andersson T, Granath K (1982) Molecular weight distribution of hyaluronic acid of human synovial fluid in rheumatic diseases. Scand J Rheumatol 12: Castor CW, Prince RK (1964) Modulation of the intrinsic viscosity of hyaluronic acid formed by human "fibroblasts" in vitro. The effects of hydrocortisone and colchicine. Biochim Biophys Acta 83: Saarni H, Tammi M, Vuorio E (1977) Effects of cortisol on glycosaminoglycans synthesized by normal and rheumatoid synovial fibroblasts in vitro. Scand J Rheumatol 6: Saarni H, Hopsu-Havu VK (1978) The decrease ofhyaluronate synthesis by anti-inflammatory steroids in vitro. Br J Dermato198: Myers SL (1985) Suppression of hyaluronic acid synthesis in synovial organ cultures by corticosteroid suspensions. Arthritis Rheum 28: Kleesiek K, Greiling H (1982) Effect of anti-inflammatory agents on the glycosaminoglycan metabolism in cultured human synovial cells. Rheumatol Int 2: Verbruggen G, Veys EM (1977) Influence ofsulphated glycosaminoglycans upon proteoglycan metabolism of the synovial lining cells. Acta Rheumatol 1: Smith MM, Ghosh P (1986) The effects of some polysulphated polysaccharides on hyaluronate (HA) synthesis by human synovial fibroblasts. Agents and Actions [Suppl] 18: Hamerman D, Wood DD (1984) Interleukin I enhances synovial cell hyaluronate synthesis. Proc Soc Exp Biol Med 177: Heino J (1986) Connective tissue components in synovial fibroblast cultures exposed to interleukin 1 and prostaglandin E2. Virchows Arch [Cell Pathol] 50: Peyron JG, Balazs EA (1974) Preliminary clinical assessment of Na-hyaluronate injection into human arthritic joints. Pathol Bio122: Weiss C, Balazs EA, St Onge R, Denlinger JL (1980) Clinical studies of the intraarticular injection of Healon (sodium hyaluronate) in the treatment of osteoarthritis of human knees. Semin Arthritis Rheum 11 (suppl 1): Namiki O, Toyoshima H, Morisaki N (1982) Therapeutic effect of intra-articular injection of high molecular weight hyaluronic acid on osteoarthritis of the knee. Int J Clin Pharmacol Ther Toxicol 20: Kopp S, Wenneberg B, Haraldson T, Carlsson GE (1985) The short-term effect of intra-articular injections of sodium hyab uronate and corticosteroid on temporomandibular joint pain and dysfunction. J Oral Maxillofac Surg 43: Rydell NW, Butler J, Balazs EA (1970) Hyaluronic acid in synovial fluid. VI. Effect of intra-articular injection of hyaluronic acid on the clinical symptoms of arthritis in track horses. Acta Vet Scand 11: Rydell N, Balazs EA (1971) Effect of intra-articular injection of hyaluronic acid on the clinical symptoms of osteoarthritis and on granulation tissue formation. Clin Orthop 80: Asheim A, Lindblad G (1976) Intra-articular treatment of arthritis in race-horses with sodium hyaluronate. Acta Vet Scand 17: Auer TA, Fackelman GE, Gingerich DA, Fetter AW (i980) Effect of hyaluronic acid in naturally occurring and experimentally induced osteoarthritis. Am J Vet Res 41: Phillips MW (1980) Intra-articular sodium hyaluronate in the horse: A clinical trial. Proc 26th Ann Conv AAEP, pp Gingerich DA, Auer JA, Fackelman GE (1981) Effect of exogenous hyaluronic acid on joint function in experimentally induced equine osteoarthritis: dosage titration studies. Res Vet Sci 30: Dayer J-M, Krane SM, Russell RGG, Robinson DR (1976) Production of collagenase and prostaglandins by isolated adherent rheumatoid synovial cells. Proc Natl Acad Sci USA 73: Burton K (1968) Determination of DNA concentration with diphenylamine. In: Colowick SP, Kaplan NO (eds) Methods in enzymology, part 12/lB. Academic Press, New York London, pp Blumenkrantz N, Asboe-Hansen G (1973) New method for quantitative determination of uronic acids. Anal Biochem 54: Peterson GL (1977) A simplification of the protein assay method of Lowry et al. which is more generally applicable. Anal Biochem 83: Ohya T, Kaneko Y (1970) Novel hyaluronidase from streptomyces. Biochim Biophys Acta 98: Balazs EA, Darzynkiewicz Z (1973) The effect of hyaluronic acid on fibroblasts, mononuclear phagocytes and lymphocytes, In: Kulonen E, Pikkarainen J (eds) Biology of fibroblast. Academic Press, New York London, pp Wiebkin OW, Muir H (1973) The inhibition of sulphate incorporation in isolated adult chondrocytes by hyaluronic acid. FEBS Lett 37: Wiebkin OW, Hardingham TE, Muir H (1975) Hyaluronic acid-proteoglycan interaction and the influence of hyaluronic acid on proteoglycan synthesis by chondrocytes from adult cartilage. In: Slavkin HC, Greulich RC (eds) Extracellular matrix influences on gene expression. Academic Press, New York London, pp Wiebkin OW, Muir H (1975) Influence of the cells of the pericellular environment. The effect of hyaluronic acid on

10 122 proteoglycan synthesis and secretion by chondrocytes of adult cartilage. Philos Trans R Soc Lond 271: Wiebkin OW, Muir H (1977) Synthesis of proteoglycans by suspension and monolayer cultures of adult chondrocytes and de novo cartilage nodules - the effect of hyaluronic acid. J Cell Sci 27: Toole BP (1973) Hyaluronate and hyaluronidase in morphogenesis and differentiation. Am Zool 13: Solursh M, Vaerewyck SA, Reiter RS (1974) Depression by hyaluronic acid of glycosaminoglycan synthesis by cultured chick embryo chondrocytes. Dev Biol 41: Handley C J, Lowther DA (1978) Inhibition of proteoglycan biosynthesis by hyaluronic acid in chondrocytes in cell culture. Biochim Biophys Acta 444: Meier S, Hay ED (1974) Stimulation of extracellular matrix synthesis in the developing cornea by glycosaminoglycans. Proc Natl Acad Sci USA 71 : Angello JC, Hauschka SD (1980) Hyaluronate-cell interaction. Effects of exogenous hyaluronate on muscle fibroblast cell surface composition. Exp Cell Res 125: Wigren A, Wik O (1974) The influence of hyaluronic acid on immobilized knees. An experimental study on adult rabbits. Ups J Med Sci [Suppl] 16: Weiner LB (1985) Applications of sodium hyaluronate in equine practice. Equine Vet Sci 5: Prehm P (1984) Hyaluronate is synthesized at plasma membranes. Biochem J 220: Underhill CB, Toole BP (1979) Binding of hyaluronate to the surface of cultured cells. J Cell Biol 82: Underhill CB, Toole BP (1980) Physical characteristics of hyaluronate binding to the surface of Simian Virus 40-transformed 3T3 cells. J Biol Chem 255: Underhill CB, Toole BP (1981) Receptors for hyaluronate on the surface of parent and virus-transformed cell lines. Binding and aggregation studies. Exp Cell Res 131: Laurent TC, Fraser JRE, Pertoft H, Smedsrod B (1986) Binding of hyaluronate and chondroitin sulfate to liver endothelial cells. Biochem J 234: Note added in proof The hydrodynamic size distribution of newly synthesised HA can only be measured when MW < 106 by Sepharose CL 2B chromatography. It is possible that there are substantial influences of exogenous HA at MW> l0 s which cannot be detected by this method. Further investigations are continuing using non-commercial high porosity gels with greater exclusion limits provided by Pharmacia AB, Sweden (see [8]).