Abstract
Foci of chondrocyte hypertrophy that commonly develop in osteoarthritic (OA) cartilage can promote dysregulated matrix repair and pathologic calcification in OA. The closely related chemokines IL-8/CXCL8 and growth-related oncogene α (GROα)/CXCL1 and their receptors are up-regulated in OA cartilage chondrocytes. Because these chemokines regulate leukocyte activation through p38 mitogen-activated protein kinase signaling, a pathway implicated in chondrocyte hypertrophic differentiation, we tested whether IL-8 and GROα promote chondrocyte hypertrophy. We observed that normal human and bovine primary articular chondrocytes expressed both IL-8Rs (CXCR1, CXCR2). IL-8 and the selective CXCR2 ligand GROα (10 ng/ml) induced tissue inhibitor of metalloproteinase-3 expression, markers of hypertrophy (type X collagen and MMP-13 expression, alkaline phosphatase activity), as well as matrix calcification. IL-8 and the selective CXCR2 ligand GROα also induced increased transamidation activity of chondrocyte transglutaminases (TGs), enzymes up-regulated in chondrocyte hypertrophy that have the potential to modulate differentiation and calcification. Under these conditions, p38 mitogen-activated protein kinase pathway signaling mediated induction of both type X collagen and TG activity. Studies using mouse knee chondrocytes lacking one of the two known articular chondrocyte-expressed TG isoenzymes (TG2) demonstrated that TG2 was essential for murine GROα homologue KC-induced TG activity and critically mediated induction by KC of type X collagen, matrix metalloproteinase-13, alkaline phosphatase, and calcification. In conclusion, IL-8 and GROα induce articular chondrocyte hypertrophy and calcification through p38 and TG2. Our results suggest a novel linkage between inflammation and altered differentiation of articular chondrocytes. Furthermore, CXCR2 and TG2 may be sites for intervention in the pathogenesis of OA.
Low-grade inflammatory alterations within synovium and cartilage modulate pathogenesis in osteoarthritis (OA), 3 with effects of IL-1 on cartilage a prime example (1, 2, 3, 4). IL-1 induces expression of several matrix-degrading proteases, including matrix metalloproteinases (MMPs), 1, 3, and 13, and a disintegrin and metalloproteinase ts4/5 (1, 2, 3, 4, 5). IL-1 also induces inducible NO synthase expression and up-regulated NO production, which suppresses PG synthesis and promotes MMP activation (6, 7, 8, 9). Furthermore, IL-1 induces matrix calcification by chondrocytes (10, 11), a common pathologic event associated with OA in situ (12). However, the level at which IL-1 is expressed within OA cartilage is variable, and chondrocyte responsiveness to IL-1 in OA can be diminished by soluble IL-1R antagonist and type II IL-1β decoy receptor (1, 2, 4). Therefore, inflammatory cytokines other than IL-1 that can promote OA are of interest.
Chondrocytes in OA cartilages frequently develop increased expression of chemokines, including the CC chemokine subfamily member RANTES and the CXC chemokine subfamily members IL-8 (CXCL8) and GROα (CXCL1) (13, 14, 15). Significantly, the highly selective IL-8R CXCR1 and the relatively promiscuous CXC chemokine receptor CXCR2 are both expressed in situ in normal cartilage, and CXCR1 expression may moderately increase in OA cartilage (16, 17).
Chemokine functions in leukocyte adhesion and migration and organization of inflammatory reactions are recognized (18). But RANTES induces chondrocyte expression of inducible NO synthase, MMP-1, and IL-6, and stimulates PG depletion in cartilage explants (13), which has focused attention on potential pro-OA effects of chemokines exerted directly on chondrocytes. In this regard, high concentrations of growth-related oncogene α (GROα) (at ≥1 μg/ml) induce at least two matrix-degrading enzymes, MMP-3 and lysosomal N-acetyl-β-d-glucosaminidase, in cartilage explants or cultured chondrocytes, respectively (16). GROα (at ≥1 μg/ml) also stimulates chondrocyte apoptosis (19). Hence, GROα and IL-8 effects on chondrocytes may be pertinent in OA.
In leukocytes, IL-8 and GROα regulate adhesion and migration by defined signal transduction mechanisms, including the p38 mitogen-activated protein kinase (MAPK) pathway (20, 21). In chondrocytes, p38 MAPK pathway activation promotes both hypertrophic differentiation and apoptosis (22, 23). Furthermore, resting articular chondrocytes in OA cartilage can undergo transition to hypertrophic and apoptotic cells, partially recapitulating chondrocyte differentiation changes in endochondral mineralization (2, 11). Chondrocyte hypertrophy and apoptosis in OA can contribute to dysregulation of matrix repair via alterations in collagen subtype and MMP expression, and in cell viability, respectively (24, 25, 26). Chondrocytes that have undergone hypertrophic differentiation or apoptosis in OA cartilage most likely promote pathologic calcification, analogous to the major contribution of chondrocyte hypertrophy to physiologic growth plate mineralization (11, 27). Pathologic calcification in OA cartilage can stimulate crystal-induced inflammation that promotes further cartilage degradation (11, 12).
Markers of chondrocyte hypertrophic differentiation in the growth plate include up-regulated expression of the transglutaminase (TG) isoenzymes TG2 (also termed tissue TG and Gh) and factor XIIIA (FXIIIA) (10, 11). TGs are mediators of tissue repair, and catalyze calcium-dependent transamidation, resulting in covalent cross-linking of available substrate glutamine residues to a primary amino group (EC 2.3.2.13) (10, 11, 28) that can thereby stabilize matrix proteins, including several collagen subtypes and fibronectin (10, 11, 28). Furthermore, TG transamidation activity, which is up-regulated in part by posttranslational modifications of TGs (28), increases in both an OA severity-dependent and age-dependent manner in joint cartilages (10).
Selective gain-of-function of either chondrocyte TG2 or FXIIIA TG activity via transfection stimulates cultured chondrocytes to calcify their matrix (10). But IL-1β induces TG transamidation activity and stimulates calcification in cultured chondrocytes in a TG2-dependent manner (10, 11). In addition, TG2 is required for retinoic acid-induced articular chondrocyte hypertrophic differentiation in vitro (11).
In this study, we demonstrate that the CXCR2 ligands IL-8 and GROα induce chondrocyte hypertrophy mediated by p38 signaling and TG2. Our results suggest that chemokine-mediated inflammation can promote altered chondrocyte differentiation and calcification in OA.
Materials and Methods
Reagents
Bovine and human chondrocyte isolation and culture
Normal human knee articular chondrocytes were obtained from M. Lotz (The Scripps Research Institute, La Jolla, CA), as described (29). Primary bovine chondrocytes were obtained from articular cartilage slices removed from the medial and lateral condyles, the patellar grove, and the tibial plateau of normal mature bovine knee joints (Animal Technologies, Tyler, TX) (11). Chondrocytes were isolated using collagenase digestion and placed in monolayer culture in DMEM/high glucose supplemented with 10% FCS, 1% l-glutamine, 100 U/ml penicillin, and 50 μg/ml streptomycin at 37°C with 5% CO2, as described (11). For treatment of chondrocytes with cytokines under nonadherent conditions, aliquots of 0.1 × 106 cells were plated on polyhydroxylethylmethacrylate (polyHEME)-coated 96-well round-bottom plates in DMEM high glucose supplemented with only 1% FCS, 100 U/ml penicillin, and 50 μg/ml streptomycin (11). For studies of calcification with bovine and human cells, the medium was additionally supplemented with 2.5 mM sodium phosphate and 25 μg/ml of l-ascorbic acid 2-phosphate (11).
Mouse knee chondrocyte isolation and culture
We established a breeding colony of TG2 knockout mice (30), as well as congenic wild-type control mice on a C57BL6/129SVJ background, as described (11). Primary mouse articular chondrocytes were isolated by dissection of the tibial plateaus and femoral condyles of mice at 2 mo of age, as described (11). In brief, the articular cartilage was carefully peeled off with a scalpel to avoid subchondral bone disruption, and chondrocytes extracted by collagenase digestion were plated in monolayer culture in DMEM high glucose supplemented with 10% FCS, 1% glutamine, 100 U/ml penicillin, and 50 μg/ml streptomycin at 37°C in 5% CO2 for 7 days before initiation of each experiment. Subconfluent chondrocytes were >95% type II collagen expression. Approximately 2500 primary chondrocytes were obtained initially from a pair of knee joints from each mouse. Mouse chondrocytes were allowed to proliferate for 5 days, yielding ∼10,000 cells per pair of knees per mouse. For the described experiments, knees from 30 mice of each genotype were harvested, and the chondrocytes of each genotype were pooled upon isolation and plated at ∼70% confluency in DMEM high glucose, with 1% FCS, 1 mM sodium phosphate, 50 μg/ml of ascorbic acid, and glutamine and antibiotics, as above.
CXCR1 and CXCR2 assessment by flow cytometry and IL-8 ELISA
Primary human and bovine articular chondrocytes were permeabilized using the BD Cytofix/Cytoperm kit (BD PharMingen, San Diego, CA) and incubated for 30 min at 4°C with murine monoclonal anti-CXCR1 (Biosource, Camarillo, CA) and anti-CXCR2 (BD PharMingen), or murine IgG isotype control. Washed cells were incubated with FITC-conjugated goat F(ab′)2 anti-mouse Ig (Biosource) for 30 min at 4°C. Fluorescence was detected using a FACSCaliber apparatus (BD Biosciences, San Jose, CA) with data analyzed using CellQuest software (Purdue University, West Lafayette, IN). Expression of IL-8 was determined using the human IL-8 ELISA kit (Biosource), according to the manufacturer protocol.
SDS-PAGE/Western blotting and RT-PCR analyses
Preparations of cell lysates, protein concentration determinations, and 10% SDS-PAGE/Western blots were performed, as described (11). Primary and secondary Ab dilutions were 1/2000. Commercial polyclonal Abs to type X collagen, TG2, FXIIIA (from Calbiochem), MMP-13 (Chemicon, Temecula, CA), tissue inhibitor of metalloproteinases (TIMP-3) (Chemicon), and tubulin (Sigma-Aldrich) were used. To assess MMP-13 protein, aliquots of 0.2 ml conditioned medium were precipitated using 15% trichloroacetic acid and centrifuged, and the resulting pellet was resuspended in 0.05% NaOH. Aliquots of 10 μg of precipitated protein were studied by SDS-PAGE/Western blotting.
For RT-PCR analyses, total RNA was prepared and reversed transcribed, as described (29). Primers and RT-PCR conditions for the ribosomal housekeeping gene L30 were described earlier (31). Type X collagen primers were sense 5′-TAGGAGCTAAAGGAGTGCCTGGAC-3′ and antisense 5′-GCATACCTGTTACCCCGTGGTTAG-3′, which amplified a 369-bp product in human and bovine chondrocytes confirmed by sequencing to be type X collagen.
NO production, and MMP-13, alkaline phosphatase (AP), TG, and kinase assays
NO production was measured using the Greiss reaction, as previously described (10). To assay MMP-13 activity, aliquots of 50 μl conditioned medium were added to individual wells in a 96-well plate containing 25 μM of MMP-13 fluorogenic substrate (7-methoxycoumarin-4-yl)-acetyl-Pro-Cha-Gly-Nva-His-Ala-Dpa-NH2 (Cha = l-cyclohexylalanine; Dpa = 3-(2,4-dinitrophenyl)-l-2,3-diaminopropionyl; Nva = l-norvaline) (Calbiochem) (32) in 50 μl of 200 mM NaCl, 50 mM Tris-HCl, 5 mM CaCl2, 20 μm ZnSO4, and 0.05% BRIJ 35, pH 7.5, for 18 h at 37°C. Fluorescence was read at excitation 325 nm, emission 393 nm. We determined AP sp. act., as previously described, with units of AP designated as moles of substrate hydrolyzed per hour (per gram protein in each sample) (11).
To assay TG transamidation activity, we modified a spectrophotometric protocol for 5-biotinamidopentylamine binding to dimethylcasein (33), as previously described (11). One unit of TG was designated as 1 μM substrate catalyzed per hour.
Kinase assays for activities of extracellular signal-regulated kinase 1/2, JNK, and p38 mitogen-activated protein kinases were conducted using substrates and protocols from Cell Signaling Technology (Beverly, MA), with aliquots of 2 × 106 bovine chondrocytes studied in individual wells of a polyHEME-coated 24-well culture plate.
Matrix calcification
To quantify matrix calcification, we used a previously described spectrophotometric assay system for Alizarin Red S binding to precipitated calcium (10, 11).
Statistical analyses
Where indicated, error bars represent SD. Statistical analyses were performed using the Student’s t test (paired two-sample testing for means).
Results
Effects of IL-8 on chondrocyte differentiation
Both normal human and bovine knee articular chondrocytes in first passage culture expressed both IL-8Rs (CXCR1 and CXCR2) (Fig. 1⇓). Next, we compared responses to IL-1 and IL-8 in chondrocytes using polyHEME-coated plates to establish nonadherent culture conditions. We confirmed (10, 11) that IL-1 (10 ng/ml) stimulated matrix calcification, and IL-1 also was confirmed (6, 8) to markedly increase NO release (i.e., by well over 10-fold) in chondrocytes (Fig. 2⇓). IL-8 (10 ng/ml) stimulated calcification comparably to IL-1, but induced NO release by only 3- to 6-fold in human and bovine chondrocytes (Fig. 2⇓). Having observed results comparable in human and bovine knee chondrocytes for IL-8R expression and responses to IL-1 and IL-8 (Figs. 1⇓ and 2⇓), we further examined IL-8 effects on differentiation using predominantly bovine knee chondrocytes, which were more readily available. Where appropriate (e.g., for experiments using an ELISA specific for human protein), we used normal human knee chondrocytes.
Flow cytometric analysis of CXCR1 and CXCR2 expression in human and bovine articular chondrocytes. Primary articular chondrocytes of the indicated species were permeabilized, and flow cytometric analysis was performed using anti-CXCR1, anti-CXCR2, and IgG isotype control, as described in Materials and Methods. These data are representative of three independent experiments.
Altered matrix calcification and NO release in response to IL-8. Primary human (A) or bovine (B) articular chondrocytes (1 × 105 cells/well in a 96-well plate) were stimulated with 10 ng/ml IL-1 or IL-8 in the presence of medium supplemented with 2.5 mM sodium phosphate and 25 μg/ml ascorbic acid. Fresh medium and stimuli were added every third day. Bound Alizarin Red S was analyzed at 14 days. The data represent the average of 8 replicates from 10 different human and 8 different bovine donors. Primary human (C) and bovine (D) articular chondrocytes were stimulated, as in A. NO release was measured via the Greiss reaction at 24 h after stimulation. The data are pooled from 5 different human and 4 different bovine donors, studied in replicates of 8. ∗, p < 0.05.
IL-8 (10 ng/ml) induced the stereotypic marker of chondrocyte hypertrophy type X collagen at the mRNA and protein expression levels in bovine articular chondrocytes (Fig. 3⇓, A and B). Under the same conditions, an absence of induction of type X collagen by IL-1 (10 ng/ml) was confirmed (11) (Fig. 3⇓, A and B). Because IL-1 can induce IL-8 expression in cultured chondrocytes (34, 35, 36), we assessed for the levels of basal vs IL-1-stimulated IL-8 expression by chondrocytes in the nonadherent culture system used in this study. To do so, we exclusively used human cells because the IL-8 ELISA used did not detect bovine IL-8. We observed that IL-1 (10 ng/ml) significantly induced IL-8 release, but that the IL-1-induced IL-8 release achieved only modest levels (<100 pg/ml in the conditioned medium) (Fig. 3⇓C). Furthermore, IL-8-induced chondrocyte type X collagen expression, which was robust in cells treated with 10 ng/ml IL-8, was modest in cells treated with only 0.1 and 1 ng/ml IL-8 (Fig. 3⇓D).
Differential induction of type X collagen in response to IL-1 and IL-8. Primary human articular chondrocytes (1 × 105 cells/well in a 96-well plate) were stimulated with 10 ng/ml rIL-1 or rIL-8. A, Total RNA was harvested on days 0, 1, and 3. RT-PCR analyses for type X collagen and L30 were conducted, as described. B, SDS-PAGE and Western blotting analysis for type X collagen was performed on cell lysates at 5 days in culture. Tubulin was used as a loading control. C, IL-8 ELISA was performed on conditioned medium from cells stimulated with 10 ng/ml IL-1 at 5 days in culture. D, SDS-PAGE and Western blotting analysis for type X collagen was performed on cell lysates at 5 days in culture after stimulation with 0.1, 1, or 10 ng/ml IL-8, as indicated. Tubulin was used as the internal loading control. Results illustrated are all from human cells (representative of four independent experiments). ∗, p < 0.05.
We next tested for effects of IL-8 on production of MMP-13, which parallels type X collagen during maturation of growth plate chondrocytes and mediates mineralization (37, 38). IL-8 and IL-1 both induced MMP-13 secretion from normal bovine articular chondrocytes (Fig. 4⇓A). In contrast, IL-8, but not IL-1, induced the MMP inhibitor TIMP-3 (Fig. 4⇓B). Concordantly, increased MMP-13 enzyme activity in conditioned medium was induced by IL-1, but not IL-8 (Fig. 4⇓C).
Concurrent MMP-13 and TIMP-3 induction by IL-8. Primary bovine articular chondrocytes (1 × 105 cells/well in a 96-well plate) were stimulated with 10 ng/ml IL-1 or IL-8. A, SDS-PAGE and Western blot analysis for MMP-13 was performed on the precipitated conditioned medium at day 3 of stimulation, as described in Materials and Methods, with 50- and 57-kDa forms detected. B, SDS-PAGE and Western blot analysis for TIMP-3 was performed on cell lysates at 3 days of stimulation. Two alternatively glycosylated forms are visible at 24 and 30 kDa. C, MMP-13 activity was measured fluorometrically in triplicate (excitation 325 nm, emission 393 nm) at days 2, 3, 5, and 7 and expressed in picograms of substrate cleaved per 105 cells, as described in Materials and Methods. Data representative of three independent experiments. ∗, p < 0.05.
Next, we observed that IL-8 induced two additional markers of hypertrophic chondrocyte differentiation, increased TG and AP activity (10, 11), in normal bovine articular chondrocytes (Fig. 5⇓). IL-1 was confirmed (10, 11) to induce TG activity (Fig. 5⇓A), but IL-1 significantly suppressed the level of AP activity in normal articular chondrocytes (Fig. 5⇓B). Hence, there was substantial dissociation between IL-1 and IL-8 effects on chondrocyte differentiation (Figs. 2–5⇑⇑⇑⇓).
TG activity and alkaline phosphatase (AP) activity in response to IL-8. Primary bovine articular chondrocytes (1 × 105 cells/well in a 96-well plate) were stimulated with 10 ng/ml IL-1 or IL-8 in medium supplemented wth 2.5 mM sodium phosphate and 25 μg/ml ascorbic acid, as described in Materials and Methods. A, TG activity was assayed at 48 h, as described in Materials and Methods, with these data pooled from five different donors in replicates of three. B, AP activity was analyzed at 72 h, as described in Materials and Methods, with these data pooled from seven donors studied in triplicate. ∗, p < 0.05.
GROα induces chondrocyte hypertrophic differentiation comparable to IL-8
IL-8 can differentially activate signal transduction and cell functions through CXCR1 and CXCR2 (39, 40). Because GROα is an activating ligand of CXCR2, but not CXCR1 (18), we next assessed whether GROα shared the capacity of IL-8 to stimulate hypertrophic differentiation in normal bovine articular chondrocytes (Fig. 6⇓). We observed that comparable to IL-8, GROα induced matrix calcification, type X collagen, and MMP-13 expression, and TG activity consistent with hypertrophic differentiation. GROα also induced TIMP-3 expression and did not induce increased MMP-13 activity (Fig. 6⇓), comparable to the effects of IL-8 (Fig. 4⇑). Thus, CXCR2-mediated signaling appeared sufficient to promote articular chondrocyte hypertrophy.
Induction of chondrocyte hypertrophy by GROα. Primary bovine chondrocytes (1 × 105 cells/well in a 96-well plate) were stimulated with 10 ng/ml GROα. Using the methods described above, we performed analyses of matrix calcification (A), type X collagen expression by RT-PCR (B), and SDS-PAGE and Western blotting (C), TG activity (D), MMP-13 protein released in conditioned medium (E), MMP-13 activity (E), TIMP-3 protein (relative to tubulin as a control) in cell lysates, with results spliced from the same blot (F), and MMP-13 activity (G). Results of B are from human cells, and the rest are from bovine cells. ∗, p < 0.05.
Role of p38 pathway signaling in IL-8-induced chondrocyte hypertrophy
Because IL-8-induced p38 pathway signaling mediates certain functional responses in leukocytes (21), we next examined the roles of signaling through p38 and the related MAPKs p44/42 and JNK in IL-8-induced chondrocyte hypertrophy. IL-8 rapidly induced kinase activities of p44/42, JNK, and p38 in bovine chondrocytes (Fig. 7⇓A). Selective pharmacologic inhibition of the p38 pathway (using SB203580, 25 μM), but not selective PD98059 and JNK inhibitor II inhibition of the p44/42 or JNK pathways, respectively, using inhibitor doses and conditions validated to give selective suppression of each MAPK pathway (not shown), blocked type X collagen expression in response to IL-8 (Fig. 7⇓B). Similarly, selective pharmacologic p38 inhibition, but not p44/42 or JNK inhibition, blocked the capacity of not only IL-8, but also GROα to induce TG activity in chondrocytes (Fig. 7⇓C).
Role of p38 pathway signaling in IL-8-induced type X collagen expression and TG activity. A, Aliquots of 2 × 106 bovine chondrocytes were stimulated with IL-8 for the times indicated, as described in Materials and Methods. The nonradioactive kinase assays for p38, JNK, and p44/42 activities were performed, as described in Materials and Methods; data representative of three independent experiments. B, Aliquots of 1 × 105 bovine chondrocytes were pretreated for 1 h with the indicated selective pharmacological inhibitor of individual MAPK signaling pathways (PD98059, 50 μM for p44/42; JNK inhibitor II, 25 μM for JNK; SB203580, 25 μM for p38), then stimulated with IL-8 for 5 days. SDS-PAGE and Western blotting for type X collagen and tubulin were performed, as described above; data representative of three independent experiments. C, Aliquots of 1 × 105 bovine chondrocytes were pretreated for 1 h with the selective p44/42, JNK, and p38 pathway inhibitors, as described above, then stimulated with 10 ng/ml IL-8 or GROα for 48 h. Chondrocyte TG activity was analyzed as above; data pooled from four independent experiments in triplicate. ∗, p < 0.05.
Central role of TG2 in chondrocyte hypertrophic differentiation induced by KC/GROα
Results to this point indicated that IL-8 and GROα induced calcification and multiple markers of chondrocyte hypertrophic differentiation, including increased TG activity. We confirmed (10, 11) by SDS-PAGE/Western blot analysis that articular chondrocytes expressed the TG isoenzymes FXIIIA and TG2 under the culture conditions used in this study (data not shown). Because TG2 can modulate cell differentiation (11), we tested the role of TG2 in CXCR2 ligand-induced TG activity, hypertrophic differentiation, and calcification. To do so, we evaluated knee articular chondrocytes from the TG2 null mouse (11, 30). Because mice do not express homologues of either IL-8 or CXCR1 (41), we tested functional responses to the murine GROα homologue KC (Fig. 8⇓).
Altered effects of KC/GROα on murine TG2−/− chondrocytes. To determine KC/GROα TG activity (A), we isolated primary knee articular chondrocytes from 2-mo-old TG2+/+ and TG2−/− mice and plated the cells in 24-well plates (5 × 104 cells/well), as described in Materials and Methods. Adherent cells were stimulated with 10 ng/ml KC/GROα, where indicated, for 48 h, as described in Materials and Methods, and TG activity was determined as above, with cells studied in triplicate for each condition. For studies in B–E, aliquots of adherent primary mouse articular chondrocytes (1 × 104 cells/well in a 96-well plate) were grown for up to 14 days, as described in Materials and Methods (supplemented with 1% FCS, 1 mM sodium phosphate, and 50 μg/ml of ascorbic acid), in the presence of 10 ng/ml KC/GROα, where indicated. In B, SDS-PAGE and Western blot analysis for type X collagen in cell lysates was performed at 10 days. In C, SDS-PAGE and Western blotting was performed for MMP-13 using aliquots of 10 μg of protein precipitated from conditioned medium after 3 days. In D, AP sp. act. was measured at 7 days, as described above. In E, the fresh medium and agonist were replaced every 3 days, and matrix calcification was assayed at 14 days. Data all are representative of three experiments, with each result pooled from chondrocytes of 30 mice of TG2+/+ and TG2−/− genotypes. ∗, p < 0.05.
TG2 and FXIIIA are the only TG isoenzymes detectable in normal articular chondrocytes, and TG2−/− mouse chondrocytes retain FXIIIA expression (11). In this study, we first confirmed (11) that TG transamidation activity was not absent, but reduced by ∼50%, in resting TG2−/− chondrocytes compared with congenic TG2+/+ controls (Fig. 8⇑A). KC/GROα induced a significant increase of TG activity in wild-type, but not TG2 null chondrocytes (Fig. 8⇑A). Concurrently, the capacity of KC/GROα to induce type X collagen and MMP-13 was markedly diminished, and induction of AP activity attenuated, in TG2−/− chondrocytes (Fig. 8⇑, B–D). Last, the capacity of KC/GROα to induce calcification was attenuated in TG2−/− mouse chondrocytes (Fig. 8⇑E).
Discussion
In this study, IL-8 and GROα, which are expressed in OA cartilage, induced type X collagen, MMP-13, AP, and TG activity, and calcification in chondrocytes. Our findings were consistent with comparable states of chondrocyte hypertrophic differentiation induced through CXCR2-mediated signaling that involved downstream p38 pathway activation and TG2 activation (Fig. 9⇓). We observed that articular chondrocyte CXCR2 expression was substantially more prominent in bovine than human cells used in these studies. Nevertheless, CXCR2 ligands induced hypertrophic differentiation of both human and bovine articular chondrocytes.
Model for IL-8- and GROα-induced articular chondrocyte hypertrophy. In this model, up-regulated IL-8 and GROα expression in OA cartilage stimulates chondrocyte hypertrophic differentiation in the disease. Based on the data obtained in this study, the chemokine-induced chondrocyte hypertrophic differentiation is transduced through p38 MAPK-mediated stimulation of TG2 activity.
Effects of IL-8 and GROα on chondrocytes diverged from those of IL-1. First, the CXC chemokines (but not IL-1) induced TIMP-3. Second, IL-1 failed to induce type X collagen and AP. Third, IL-1, but not the CXC chemokines induced markedly increased extracellular MMP-13 activity, despite the fact that IL-1 and the chemokines both induced MMP-13. The relatively weak NO release induced by CXCR2 ligands, in addition to induction of TIMP-3 expression, most likely contributed to the negligible MMP-13 activity in conditioned medium of IL-8- and GROα-treated cells.
Chronic inflammatory oxidative stress, and several cytokines expressed in OA joints (e.g., IL-1, TNF-α, and IL-17), have the potential to induce IL-8 and GROα expression in chondrocytes (34, 35, 36, 42, 43). In addition, IL-1 stimulates both p38 pathway signaling (44) and a TG2-dependent increase in TG activity (11) in chondrocytes. But under the conditions we used, the extracellular concentration of IL-8 (<100 pg/ml) in response to 10 ng/ml IL-1 was below the threshold for the robust type X collagen expression in response to IL-8 present at 1 and 10 ng/ml. It is possible that under more sustained culture, or under different culture conditions, IL-1 might induce chondrocyte hypertrophic differentiation mediated by IL-8 and GROα expression. But differences in IL-1 and chemokine signal transduction and in NO generation may have been contributors to the lack of chondrocyte hypertrophy in response to IL-1 under the conditions used in this study. We speculate that differing levels of expression of the broad-spectrum MMP and a disintegrin and metalloproteinase ts4/5 inhibitor TIMP-3 (45), of MMP-13 activity, and of AP activity also were significant factors modulating the distinct differentiation responses to IL-1 and the CXC chemokines seen in this study. In addition, AP generates Pi (46, 47), a direct promoter of expression of type X collagen and certain other genes expressed by hypertrophic chondrocytes (48, 49, 50). Significantly, AP activity and Pi metabolism exert marked regulatory effects on both chondrocyte differentiation and organization in the endochondral growth plate (51). In view of the results in this study, the capacity of inhibition of MMP-13 to prevent prehypertrophic growth plate chondrocytes from progressing to hypertrophy and mineralizing the matrix in vitro appears paradoxical (37).
In this study, we defined relatively robust chondrocyte responses to doses of IL-8 and GROα of 10 ng/ml. IL-8 and GROα concentrations in the range of 5–20 ng/ml are commonly detected in inflammatory joint fluids, whereas OA synovial fluid concentrations of IL-8 have been reported to range between 1 and 5 ng/ml (52, 53), with GROα levels averaging 0.1 ng/ml in one study (54). However, low-level type X collagen induction was detectable in chondrocytes treated with 0.1 ng/ml IL-8. In prior studies of GROα, effects on cultured chondrocytes and cartilage explants, investigators used much higher concentrations of GROα (1–5 μg/ml), which induced MMP-3 expression, N-acetyl-β-D-glucosaminidase release, and apoptosis (16, 19). We did not examine these particular responses in this study. But it should be noted that hypertrophy precedes apoptosis in sequential endochondral chondrocyte differentiation (55). Moreover, hypertrophic and apoptotic cells colocalize near calcifications in association with cartilage degeneration (27, 55). As such, CXCR2 ligands are potentially factors in promoting both hypertrophy and apoptosis in OA cartilage.
Using TG2−/− knee chondrocytes, we discovered that TG2 was centrally involved in not only GROα-induced increases in TG activity, but also in hypertrophic differentiation and calcification. We have not observed induction of TG2 or FXIIIA expression by GROα or IL-8 in chondrocytes (D. Merz et al., unpublished observations). Hence, TG2 activation by the chemokines was most likely exerted primarily at the level of posttranslational modifications (11, 28, 56, 57). FXIIIA and TG2 are the only TG isoenzymes we have detected in articular chondrocytes (10, 11). We confirmed (11) in this study that TG activity was only reduced by ∼50% in resting TG2−/− mouse knee chondrocytes, consistent with a major contribution of FXIIIA to basal TG activity. Physiologically, the partial functional redundancy of TG2 and FXIIIA in cartilages (10) appears to contribute to the grossly normal skeletal development of TG2−/− mice (11, 30). But in response to certain pathologic stressors, including GROα and IL-8, TG2 in articular chondrocytes may function to modulate cartilage repair and stability.
Cytosolic and externalized TG2 both could regulate chondrocyte differentiation in response to CXCR2 ligands (28, 56, 57). Significantly, TG2-induced matrix cross-linking consistent with externalization of active TG2 has been detected in OA cartilage (58). Cell surface TG2 cannot only promote fibronectin cross-linking via transamidation, but also directly modulate cell attachment to fibronectin via the TG2 N-terminal fibronectin binding domain in TG2 (59). Interestingly, TG2 regulates adhesion and migration in mononuclear phagocytes and fibroblasts (60, 61). By modulating cell adhesion, fibronectin binding, and matrix assembly (28, 62), TG2 has the potential to regulate matrix signals provided to chondrocytes, thereby modulating chondrocyte differentiation and growth (46, 63, 64, 65). Significantly, TG2 catalyzes formation of intramolecular cross-links that increase phospholipase A2 (PLA2) activity (66). TG2-modulated PLA2 activation can modulate inflammatory responses (66). It will be of interest to define whether PLA2 activation mediates effects of CXCR2 ligands on chondrocytes.
Although IL-8 and GROα induced TG2-dependent transamidation activity in chondrocytes, the relative contribution of transamidation events catalyzed by TG2 to chemokine-induced chondrocyte hypertrophy is not known. TG2 is a dual function enzyme (TG and GTPase/ATPase), a unique property of TG2 among human TGs (11, 28, 57). It is possible that the capacities of TG2 to regulate signal transduction and Pi production through GTPase and ATPase activities (11) modulate CXCR2 ligand-induced chondrocyte hypertrophy.
Our results add to a growing body of evidence that chemokine ligands of CXCR2, such as IL-8 and GROα, function beyond recruitment and activation of leukocytes in sites of tissue injury and inflammation (67, 68, 69). For example, such chemokines exert direct effects on MMP expression, fibroblast growth, endothelial cell growth, and angiogenesis, and on tumorigenesis (67, 68, 69). Timing of CXCR2 ligand exposure to chondrocytes, levels of CXCR2 expression, and potential combinatorial effects of CXCR2 ligands with those of other cytokines all have the potential to modulate chondrocyte activation in vivo. Nevertheless, our identification of the capacity of CXCR2 ligands to stimulate hypertrophy and calcification in cultured articular chondrocytes, and the major contribution of TG2 to these events, reveals a novel mechanism by which chronic inflammation can alter the differentiation and function of resident cells of mesenchymal origin in arthritis. CXCR2 and TG2 may be novel sites for therapeutic intervention in OA.
Acknowledgments
We gratefully acknowledge the assistance of Deborah van Etten in managing the TG2−/− mouse-breeding colony.
Footnotes
-
↵1 Supported by grants to R.T. from Department of Veterans Affairs and National Institutes of Health (P01AGO7996) and VC BioSTAR program, and a National Institutes of Health R03 Award and an award from the University of California at San Diego Stein Institute for Aging to R.L.
-
↵2 Address correspondence and reprint requests to Dr. R. Terkeltaub, VA Medical Center, 3350 La Jolla Village Drive, San Diego, CA 92161. E-mail address: rterkeltaub{at}ucsd.edu
-
↵3 Abbreviations used in this paper: OA, osteoarthritis; AP, alkaline phosphatase; FXIIIA, factor XIIIA; GRO, growth-related oncogene; JNK, c-Jun N-terminal kinase; MAPK, mitogen-activated protein kinase; MMP, matrix metalloproteinase; PLA2, phospholipase A2; polyHEME; polyhydroxyethylmethacrylate; TG, transglutaminase; TIMP, tissue inhibitor of metalloproteinases.
- Received March 13, 2003.
- Accepted August 7, 2003.
- Copyright © 2003 by The American Association of Immunologists
References
In this issue
