1510

INDUCTION OF SYNTHESIS AND RELEASE OF INTERLEUKIN-8 FROM HUMAN ARTICULAR CHONDROCYTES AND CARTILAGE EXPLANTS ANNELIESE D. RECKLIES and ELAINE E. GOLDS Objective. The activation of neutrophils in the joint space may contribute to the destruction of cartilage matrix observed in rheumatoid arthritis. The capacity of articular chondrocytes to synthesize and secrete interleukin-8 (IL-8)and GROa, two potent neutrophil chemoattractant peptides, was investigated to determine whether cartilage itself could serve as a source of these small cytokines. Methods. Induction of IL-8 and GRO protein was studied both at the messenger RNA (mRNA) and the protein level by reverse transcriptase/polymerasechain reaction and metabolic labeling, respectively. Results. Strong induction of IL-8 was observed in primary cultures of articular chondrocytes as well as in cartilage explants stimulated with IL-1p. The increased secretion of the IL-8 protein was accompanied by corresponding increases in mRNA levels. In contrast to other connective tissue cells, a peptide corresponding in molecular size to the GRO proteins was only weakly induced in cartilage explants or primary chondrocyte cultures. However, mRNA for all 3 members of the GRO family was easily detectable in cultured chondrocytes following stimulation with IL-1p. In explanted cartilage, mRNA for only GROy was found to be Presented in part at the 36th Annual Meeting of the Orthopaedic Research Society, New Orleans, LA, February 1990. From the Shriners Hospital for Crippled Children and the Department of Surgery, McGill University, Montreal, Quebec, Canada. Supported by the Shriners of North America and the Arthritis Society of Canada. Anneliese D. Recklies, PhD; Elaine E. Golds, PhD. Address reprint requests to Anneliese D. Recklies, PhD, Joint Diseases Laboratory, Shriners Hospital for Crippled Children, 1529 Cedar Avenue, Montreal, Quebec H3G lA6, Canada. Submitted for publication October 21, 1991; accepted in revised form August 21, 1992. Arthritis and Rheumatism, Vol. 35, No. 12 (December 1992)

induced. Newly synthesized IL-8 was slowly released from cartilage explants over a prolonged time in culture. Conclusion. The results suggest that synthesis and secretion of the diverse members of the IL-8IGRO family is regulated in a tissue-specific or cell-specific manner. The slow release of IL-8 from articular cartilage following induction by IL-lP could establish a chemotactic gradient toward the articular surface and mediate the migration and attachment of neutrophils and lymphocytes to this tissue.

During the pathogenesis of rheumatoid arthritis, a large influx of neutrophils, monocytes, and lymphocytes into the synovium and the joint space occurs at an early stage of the disease, resulting in joint swelling and pain. The migration of immune cells to sites of inflammation is directed by a variety of chemotactic peptides which are released locally in response to primary inflammatory stimuli such as interleukin-1 (IL-I) and tumor necrosis factor a (TNFa). Accumulation of immune cells in the arthritic joint has been associated with the development of autoimmune responses believed to be involved in the chronicity of rheumatoid arthritis (1). Thus, the control of synthesis and secretion of factors involved in the regulation of immune cell tr&c could be an important element in the progress of joint inflammation. It has recently been shown that IL-1 induces the synthesis and release of a number of small cytokines with proinflammatory properties. They show sequence homology with platelet factor 4, and their main function appears to be that of chemoattractants and activators of various immune cells, such as neutrophils, macrophages, and T cells (for review, see ref. 2). The most well characterized of the proinflamma-

IL-8 IN HUMAN ARTICULAR CARTILAGE

tory peptides is interleukin-8, or neutrophil-activating peptide 1 (3,4), which has been shown to be chemotactic for neutrophils and T lymphocytes both in vitro and in vivo (5). IL-8 also induces neutrophil activation, resulting in the release of glycosidases as well as the proteinase elastase (6-8). Activated macrophages appear to be a major source of chemotactic cytokines, but induction of IL-8 has also been demonstrated in dermal fibroblasts (9,10), endothelial cells (1l), and human synovial cells (10). In addition to being secreted in response to inflammatory stimuli, some members of this family are transiently expressed in fibroblasts in response to mitogenic stimuli (12,13), suggesting a role in tissue remodeling and repair. Another member of the family of chemotactic cytokines is the product of the groa gene, known as GROa or neutrophil-activating peptide 3 (NAP-3). It has properties similar to those of IL-8 and is found to be induced in parallel with IL-8 in human synovial cells (10) and in macrophages (14). While both IL-8 and GROa belong to the same subfamily of peptides, characterized by the conservation of 4 cysteine residues with the first 2 cysteines contained in the N-terminal motif -C-X-C-, there now appear to be at least 3 very closely related peptides in the GRO subgroup (GROa, GROpIMIP-2a, and GROyIMIP-2p) with close to 90% homology at the amino acid level (15,16). All 3 species have been shown to be induced at the messenger RNA (mRNA) level by IL-1 and TNFa (17,18). There appears to be some tissue and cell specificity with respect to the expression of the 3 gro genes, which suggests a differential regulation (18). Whether the protein products of all 3 gro genes have similar functions is currently not known. We have previously reported that IL-8 and GROa are induced by IL-1 in passaged human synovial cells (10). These peptides were identified in the culture medium of stimulated cells by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDSPAGE) gradient gels, where they migrate with a relative molecular size of 7.0 kd (IL-8) and 6 kd (GRO), and by N-terminal sequencing of isolated bands after transfer to polyvinylidene difluoride (PVDF) membranes. Because the resident chondrocytes of articular cartilage can contribute substantially to the degradative environment created in inflammatory conditions, the synthesis and secretion of IL-8 and GRO protein by human chondrocytes in primary or explant culture was investigated. In many respects, the responses of articular chondrocytes to IL-1 are very similar to those

1511

of synovial cells and other connective tissue cells, as has been shown for the induction of procollagenase and prostromelysin. While this work was in progress, secretion of IL-8 from human articular chondrocytes in monolayer cultures was described by van Damme et al (19). However, it is well established that the isolation of articular chondrocytes from their normal matrix environment and their culture change the phenotypic pattern of secreted proteins, in particular, that of the cartilage matrix components (20,21). The responses of chondrocytes to cytokines and growth factors may be modulated in a similar manner by the extracellular environment of the cells. We therefore studied the response of articular chondrocytes in cartilage explants to IL-lp, to determine whether IL-8 and/or GRO were induced and released in a manner similar to that of human synovial cells.

MATERIALS AND METHODS Materials. Recombinant human IL-1p was obtained from Collaborative Research (Bedford, MA). TNFa and the enzyme-linked immunosorbent assay (ELISA) kit used for detection of IL-8 were purchased from R&D Systems (Minneapolis, MN). 35S-cysteineand 35S-methioninewere purchased from Amersham (Montreal, Quebec, Canada). MMLV reverse transcriptase was obtained from Gibco-BRL (Toronto, Ontario, Canada) and Tuq DNA polymerase from Perkin-Elmer Cetus (Emeryville, CA). Oligonucleotide primers were synthesized on an ABI oligonucleotidesynthesizer (AppliedBiosystems, Mississauga, Ontario, Canada) in the Biotechnology Core Facility of the Shriners Hospital (Montreal, Quebec, Canada). Source of cartilage, chondrocytes, and synovial cells.

Human articular cartilage was harvested from grossly normal femoral condyles at autopsy. Explants were established from diced cartilage and maintained in Dulbecco’s modified Eagle’s medium (DMEM), supplemented with 0.2 mg/ml bovine serum albumin (BSA). Chondrocytes were isolated from chopped cartilage by sequential digestion with trypsin and clostridial collagenase (20). For primary cultures, cells were seeded into 6-well plates at 500,000 cells per well and cultured in DMEM supplemented with 10% fetal calf serum (FCS) and antibiotics (penicillin G, 50 unitdml, and streptomycin, 50 pg/ml). Cultures were maintained for 3 days in the presence of FCS before stimulation with cytokines. Human synovial cell lines were established from rheumatoid synovium obtained at surgery from patients undergoingjoint replacement, as described previously (10). Cells were used between passage 4 and 15. Metabolic labeling of secreted proteins. IL-8 contains 4 cysteine residues and no methionine (22). It is easily identified in culture medium when the patterns of secreted peptides, labeled with either 35S-cysteineor 35S-methionine, and analyzed by SDS-PAGE and autoradiography,are com-

1512

pared. Both radiolabeled amino acids are incorporated into newly synthesized GRO proteins, as predicted from the amino acid sequence for the putative mature, secreted peptides (12,18). Isolated primary chondrocyte cultures were maintained in serum-containing medium for 3 days to allow attachment of cells. The medium was replaced with serumfree and methionine- or cysteine-free medium containing 1 unit/ml IL-1p (0.1 ng/ml) and 50 pg/ml BSA. IL-1p concentrations of 1-2 units/ml were found to stimulate stromelysin and collagenase production in cultured chondrocytes maximally, with no further increases observed at higher concentrations (Golds EE: unpublished observations), and thus the same concentrations were used in these experiments. '%cysteine or methionine (30 pCi/ml), diluted in regular DMEM was added simultaneously, except where indicated otherwise, to give a final cysteine or methionine concentration of 20 pM. Passaged chondrocytes and human synovial cells were handled in an identical manner, except that the cells were allowed to reach confluence before stimulation with IL-1p. Cartilage explant cultures were established, containing 0.5 gm of tissue in 5 ml of culture medium, and medium was changed every 48 hours, unless indicated otherwise. This ratio of tissue to medium has been established to support the culture of human cartilage without any excessive nutrient depletion (23). Explants were maintained in DMEM for 48-72 hours before stimulation and metabolic labeling. Cartilage explants were equilibrated in cysteine- or methionine-free medium for 2 hours to deplete tissue pools of the amino acids prior to addition of cytokines and radiolabeled amino acid precursors in fresh medium. IL-1p was added at 5 units/ml(O.5 ng/ml); 35S-cysteine or methionine concentrations were the same as above. The IL-1p concentration of 5 units/ml used in cartilage explants was chosen to allow for penetration of the matrix and nonspecific adsorption; however, similar results were obtained with 2 units/ml of IL-1p. Culture media from monolayer or explant cultures were collected after 24 hours, and newly synthesized, secreted proteins were analyzed by SDS-PAGE and autoradiography, as described previously (24). Cartilage explants were extracted for 48 hours with O.05M sodium acetate buffer, pH 6.5, containing 1M sodium chloride, 10 mM EDTA, 1 mM diisopropylfluorophosphate, 1 mM iodoacetamide, and 1 pg/ml pepstatin A. Extracts were analyzed as described above. In some cases, explants were further extracted with 4M guanidine hydrochloride. These extracts were dialyzed directly against SDS sample buffer and concentrated 10-fold by ultrafiltration before analysis by SDSPAGE. Release of newly synthesized IL-8 from cartilage explants. Cartilage explants were established as above and stimulated with IL-lp in the presence of 35S-cysteine or "S-methionine for 24 hours. The medium was removed and explants were incubated with 2 changes of complete DMEM for 20 minutes each time to deplete the label and IL-1p. Stimulation of control explants with IL-lp at this point resulted in undetectable levels of radiolabeled IL-8 in the culture medium, indicating that the radioactive precursor pool had been depleted adequately. Fresh DMEM was then added and replaced every 24 hours for an additional 4 days.

RECKLIES AND GOLDS The cartilage was extracted at the end of the chase period as above, and media and extracts were analyzed for released radioactive components. To quantitate released IL-8, autoradiographs were scanned in 2 dimensions using an LKB UltroScan laser densitometer (LKB Instruments, Piscataway, NJ), and the total densities of the areas corresponding to IL-8 were estimated using the LKB GSXL software in the Biotechnology Care Facility of the Shriners Hospital. N-terminal amino acid sequence analysis. Concentrated culture medium collected from cartilage explants exposed for 3 days to IL-lp (5 unitdml) was separated by SDS-PAGE on 7.5-15% gradients and electroblotted onto PVDF membranes as described by Matsudaira (25). The band corresponding to IL-8 was identified by staining with Coomassie brilliant blue, and sequenced on an ABI 473 pulsed liquid sequencer (Applied Biosystems) in the Biotechnology Core Facility of the Shriners Hospital. Analysis of mRNA for IL-8 and members of the GRO family. RNA was isolated from cultured chondrocytes or cartilage explants according to the method of Chomczynski and Sacchi (26), with minor modifications, as described by Hakala et al (27). RNA preparations were analyzed for the presence of mRNA for IL-8, as well as members of the GRO family, using the reverse transcriptase/polymerase chain reaction technique (RTPCR). Based on the sequence published by Matsushima et a1 (28), the oligodeoxynucleotides AACATGACTTCCAAGCTGG and AAACTTCTCCACAACCCTCTGC were used as antisense and sense primers, respectively. Analysis of RNA prepared from chondrocytes stimulated for 24 hours with 1 unitlml of IL-1p by RTPCR, using these primers, gave rise to a fragment of approximately 280 basepairs (the expected size is 279 bp). This fragment was further characterized by digestion with the restriction endonuclease Hind 111, which gave rise to 2 fragments of sizes similar to those expected (208 bp and 71 bp, respectively). No PCR product was obtained from RNA prepared from control cultures. Glyceraldehyde-3-phosphate dehydrogenase (G-3-PDH) mRNA was used as a standard for the RNA preparations, and PCR products of approximately equal intensity were obtained from both RNA preparations. The primers used have been described previously (27). Primers for GROa, GROP, and GROy were designed based on the sequences published by Haskill et a1 (18) and by Anisowicz et al (29). A common sense ( 5 ' ) primer was used for all 3 members of the GRO family (CCCAAACCGAAGTCATAGC) and the following antisense (3') primers were synthesized for GROa, GROP, and GROy, respectively: GCATGAAGCAGGCTCCTC (expected fragment size 488 bp), TGTGATATGTCATCACGAAG (expected fragment size 490 bp), and CAAGGTGGCTGACACATTATGG (expected fragment size 454 bp). Fragments amplified from chondrocyte RNA preparations using these primers were further characterized by determining the presence of unique restriction enzyme sites ( H i m I1 for GROa, Hha I for GROP, and Dru I for GROy). Each PCR fragment was found to be cut only by the appropriate enzyme, yielding fragments of the appropriate size. This rules out the possibility of cross hybridization

IL-8 IN HUMAN ARTICULAR CARTILAGE

Figure 1. Comparison of induction of secreted proteins in primary articular chondrocytes and synovial cells by interleukin-lp (IL-lp). Culture medium from synovial cells (a) and primary human chondrocytes (b) labeled with 3sS-cysteine (C) or 3sS-methionine(M) for 24 hours was analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis on 7 5 1 5 % gradient gels and autoradiography. Lanes 1 and 5, Low and high molecular weight markers, respectively. Cells were cultured in the absence (a, lanes 2 and 3; b, lanes 1 and 2) or presence (a, lanes 4 and 5; b, lanes 3 and 4) of 1 unitlml of IL- 1p. Double arrowheads show the expected migration positions of prostromelysin (55 kd), IL-6 (23 kd), &-microglobulin (8 kd), IL-8 (7 kd), and GRO protein (6 kd), respectively. IL-1p induced both IL-8 and GRO protein in synovial cells (a, lane 5 ) , while the predominant peptide in this molecular size region secreted by chondrocytes was IL-8 (b, lane 4).

1513

sequence determined for the first 5 residues was A-V-L-P-R, identical to that reported elsewhere for IL-8 secreted from human synovial cells (10). Primary chondrocyte preparations of cartilage specimens from donors of various ages, ranging from newborn to 72 years, were found to respond similarly with respect to induction of IL-8 synthesis. In contrast to synovial cells, only weak induction of the 6-kd peptide identified as GROa in synovial cell culture medium was observed in cultured chondrocytes. The bands on autoradiographs corresponding to IL-8 and GRO in culture medium from human synovial cells and articular chondrocytes stimulated with IL-1 and labeled with 35S-cysteine were quantitated densitometrically . Because both secreted peptides contain 4 cysteine residues, their relative ratios within the same sample can be determined. While medium from synovial cells contained about equal amounts of the 2 peptides, primary chondrocytes secreted 8-10 times as much IL-8 as GRO protein following stimulation with IL-1p. Thus, while IL-1 induces synthesis of IL-8 in both cell types equally well, induction of the gro genes appears to require additional factors in chondrocytes. To determine whether IL-8 synthesis required the continued presence of an inducing stimulus, primary chondrocytes were exposed for 24 hours to IL-1p at 2 unitsiml. Secreted proteins were labeled

between the different downstream primers and GRO species other than the target ones. RNA preparations were amplified following reverse transcription with the 3'-primer for 1 hour at 42"C, followed by 3 0 4 0 cycles of PCR. PCR products were analyzed on 1.2% agarose gels in the presence of ethidium bromide for visualization of bands.

RESULTS Induction and synthesis of IL-8 and GRO protein in isolated articular chondrocytes. IL-1p caused a strong induction of IL-8 both in synovial cells (Figure la) and articular chondrocytes in primary culture (Figure lb). The 7.0-kd IL-8 band, seen when the cells were stimulated by IL-lfi in the presence of 35Scysteine, was completely 'absept when 35S-methionine was used as precursor (Figures 'la and b, lanes 4 and 3, respectively), consistent with the lack of methionine in mature IL-8 (12). N-terminal sequence analysis of the band corresponding to IL-8 confirmed its identity. The

Figure 2. Synthesis and secretion of interleukin-8 (IL-8) during and after removal of IL-lp. Human chondrocytes were Labeled in the presence of 2 unitslml IL-lp (a), during the 24 hours following a 24-hour stimulation with 2 unitslml IL-lp (b), or during the last 24 hours of a 48-hour stimulation period with IL-lp (c). 3sS-cysteine(C) or 3sS-methionine(M) were the labeled amino acids used. Lanes 1 and 2, Controls of cells cultured in Dulbecco's modified Eagle's medium only; lanes 3 and 4, cells exposed to IL-lp as outlined above. The expected migration positions for prostromelysin, IL-8, and GRO protein are shown to the right.

1514

either concurrently or following the removal of IL-lp and analyzed by SDS-PAGE (Figure 2). A strong band corresponding to IL-8 can be seen when cells are labeled in the presence of IL-1p (Figures 2a and c). No further synthesis was detected upon removal of the cytokine (Figure 2b), suggesting that IL-8 synthesis is shut down very rapidly. Increased secretion of components previously identified as prostromelysin (55 kd) and procollagenase (52 kd) (23) was maintained after removal of IL-1p. In the continued presence of IL-lp, IL-8 synthesis was maintained at a high level (Figure 2c, lanes 3 and 4). Synthesis of GRO did not increase during the 48-hour exposure of the cells to IL-1. Thus, the induction of IL-8 synthesis is not transient in the presence of IL-lp, but requires the presence of this cytokine for continuous synthesis. Induction of synthesis and secretion of IL-8 in cartilage explants. Since isolation and primary culture of articular chondrocytes radically alters their normal tissue environment and can influence their responses to extracellular agents, we investigated the responses of cultured cartilage explants. The response of cartilage explants to stimulation with IL-1p is shown in Figure 3. A cysteine-rich, methionine-negative band of 7.0 kd was observed in the culture medium. The peptide in this band was also identified by N-terminal sequence analysis as IL-8. A major portion of the newly synthesized material appeared to be retained in the matrix, since 1M NaCl extracts of the explants after exposure to IL-1 for 24 hours showed the presence of the 7.0-kd IL-8 band. Induction of GRO protein appeared to be variable, but the corresponding band around 6 kd was always much less intense than the band corresponding to IL-8. Densitometric analysis of the 2 bands suggested that the amounts of IL-8 were approximately 8-10-fold higher than those of GRO protein, similar to the ratio observed in cultured chondrocytes. Neither IL-8 nor GRO protein was detected in 4M guanidine HC1 extracts following the salt-extraction, indicating that all the newly synthesized material was recovered in the 1M NaCl extract. IL-8 synthesis in cartilage explants was dependent on the presence of IL-Ip, as was observed with cultured chondrocytes. Cartilage explants were labeled either during a 24-hour induction period with IL-lp, during a 24-hour period following the removal of IL-lp, or from 24-48 hours in the continuous presence of IL-1p (Figure 4). While IL-8 was detectable both in culture media (Figure 4a) and cartilage extracts (Figure 4b) after IL-1p was removed, its level of synthesis was much reduced compared with that

RECKLIES AND GOLDS

Figure 3. Synthesis and release of interleukin-8 (IL-8) from human articular cartilage. Explants from a 64-year-old donor (a) and a 72-year-old donor (b) were cultured in the absence (-) or presence (+) of 5 unitslml of human recombinant IL-lp. Culture media and 1M NaCl extracts of the cultured cartilage explants were concentrated and analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. The lines in the center indicate the migration positions of prostromelysin and IL-8. a, Explants were labeled with 3SS-cysteine(C) or 35S-methionine(M); b, explants were labeled with '%-cysteine only.

observed in the presence of IL-1p. The continued presence of IL-1p resulted in maintenance of a high level of IL-8 synthesis, but again, no further induction of GRO protein was apparent. No IL-8 synthesis was detectable when explants were labeled from 2448 hours after IL-1p had been removed (data not shown). Induction of mRNA for IL-8 and members of the GRO family. To investigate the induction of IL-8 and the 3 GRO species at the transcriptional level, RNA was prepared from cultured chondrocytes and cartilage explants, unstimulated or stimulated for 24 hours with IL-lp, and the presence of mRNA for IL-8, as well as for GROa, -p, and -7, was analyzed by RT/PCR. As shown in Figure 5a for cultured chondrocytes and in Figure 5b for cartilage explant cultures, mRNA for IL-8 can be clearly detected following stimulation with IL-1p. Interestingly, mRNA for all GRO species was detectable in cultured chondrocytes stimulated with IL-lp, but only mRNA for GROy was detectable in cartilage explants. No PCR products for GROa or GROp could be detected even after 50 cycles of amplification, indicating that in intact cartilage only the groy gene is induced by IL-1p. No message for any of these cytokines was detectable in unstimulated chondrocytes or cartilage explants or in RNA preparations from uncultured cartilage. Identical results were obtained for RNA preparations from 2 different adult donors. Cul-

IL-8 IN HUMAN ARTICULAR CARTILAGE

1515

Figure 4. Synthesis of interleukin-8 (IL-8) in articular cartilage explants depends on the presence of IL-I. Explants from a 54-year-old donor were exposed to 5 units/ml of IL-Ip and labeled either

simultaneously or after removal of IL-lp, as indicated across the bottom. a, Autoradiograph of proteins secreted into the culture medium; b, newly synthesized proteins extractable from the cartilage matrix with 1M NaCl at the end of the labeling period. Explants were labeled with either 35S-cysteine (C) or 3sS-methionine (M).

ture medium from the above cell and explant cultures was collected and analyzed for IL-8 protein, using a sandwich ELISA specific for this cytokine (Table 1). While no IL-8 could be detected in medium from control cultures, IL-8 was clearly elevated in media from cultures stimulated with IL-1p. These results are consistent with those obtained by metabolic labeling of proteins induced by IL-1p. Release of newly synthesized IL-8 from cartilage matrix. Since a fairly large proportion of the newly synthesized IL-8 was found to be present in the

cartilage matrix, a pulse-chase analysis was conducted to determine whether this material would be slowly released or whether it would be retained in the matrix. Cartilage explants were labeled with 35Scysteine or 35S-methioninein the presence of IL-lp for 24 hours, at which point the medium was replaced with a chase medium consisting of complete DMEM. Medium was collected every 24 hours for 4 more days and analyzed for the presence of radioactive components. As is shown in Figure 6, IL-8 could be detected even after 5 days. Panels a-e show culture media from

Figure 5. Induction of messenger RNA (mRNA) for interleukin-8 (IL-8) and the GRO family in human chondrocytes by IL-1p. Total cellular RNA was prepared from primary chondrocytes cultured as monolayers (a) or from cartilage explants (b). The presence of mRNA for glyceraldehyde-3-phosphate dehydrogenase (G3PDH), IL-8, as well as members of the gro gene family, was analyzed by reverse transcriptase/polymerasechain reaction. Lane 1 , RNA prepared from control cultures for each mRNA species investigated; lane 2, RNA from cultures stimulated with IL-IP (1 u n i t h l for cell cultures or 5 unitshl for explant cultures); lane 3, control, omitting the RNA.

RECKLIES AND GOLDS

1516

Table 1. Interleukin-8 chondrocytes*

Control IL-IP

(IL-8)

release

by

human

articular

Monolayer cultures (ng/lOs cells)

Explants (ng/100 mg wet weight)

ND 57.8 f 7.4

ND 38.1 f 12.4

chondrocytes and in articular cartilage. The major difference in the response of articular chondrocytes compared with that of human synovial cells was the

* Monolayer cultures were established in T25 flasks and grown to confluence before stimulation with IL-Ip (1 unit/ml). The cells were at passage level 2. Explants were established from autopsy specimens and stimulated with 5 unitshl IL-lp after 24 hours of preculture. IL-8 was determined after 24 hours of stimulation, using a sandwich enzyme-linked immunosorbent assay. The data are the average of 2 different specimens and 3 explant cultures for each specimen or of 2 culture flasks for each chondrocyte preparation. ND = not detectable (