Induction of increased cAMP levels in articular chondrocytes blocks matrix metalloproteinase–mediated cartilage degradation, but not aggrecanase-mediated cartilage degradation

Authors


Abstract

Objective

Calcitonin has been suggested to have chondroprotective effects. One signaling pathway of calcitonin is via the second messenger cAMP. We undertook this study to investigate whether increased cAMP levels in chondrocytes would be chondroprotective.

Methods

Cartilage degradation was induced in bovine articular cartilage explants by 10 ng/ml oncostatin M (OSM) and 20 ng/ml tumor necrosis factor (TNF). In these cultures, cAMP levels were augmented by treatment with either forskolin (4, 16, or 64 μM) or 3-isobutyl-1-methyl xanthine (IBMX; 4, 16, or 64 μM). Cartilage degradation was assessed by 1) quantification of C-terminal crosslinking telopeptide of type II collagen fragments (CTX-II), 2) matrix metalloproteinase (MMP)–mediated aggrecan degradation by 342FFGV- G2 assay, 3) aggrecanase-mediated degradation by 374ARGS-G2 assay, 4) release of sulfated glycosaminoglycans (sGAG) into culture medium, 5) immunohistochemistry with a monoclonal antibody recognizing the CTX-II epitope, and 6) toluidine blue staining of proteoglycans. MMP expression and activity were assessed by gelatin zymography.

Results

OSM and TNF induced an 8,000% increase in CTX-II compared with control (P < 0.001). Both forskolin and IBMX dose-dependently inhibited release of CTX-II (P < 0.001). OSM and TNF induced a 6-fold increase in 342FFGV-G2, which was abrogated by forskolin and IBMX (by >80%). OSM and TNF stimulated MMP expression as visualized by zymography, and MMP expression was dose-dependently inhibited by forskolin and IBMX. The highest concentration of IBMX lowered cytokine-induced release of sGAG by 72%.

Conclusion

Levels of cAMP in chondrocytes play a key role in controlling catabolic activity. Increased cAMP levels in chondrocytes inhibited MMP expression and activity and consequently strongly inhibited cartilage degradation. Specific cAMP modulators in chondrocytes may be potential treatments for cartilage degenerative diseases.

Osteoarthritis (OA) is the most common form of destructive joint diseases affecting both cartilage and bone tissue (1, 2). Experimental and clinical observations suggest links between these 2 compartments, since the structural integrity of articular cartilage is dependent not only on intact chondrocyte function, but also on normal subchondral bone turnover (1, 3). Accordingly, drugs that inhibit both the resorptive activity of osteoclasts and the degradation of cartilage by chondrocytes could constitute an ideal medication for the prevention/treatment of OA.

Calcitonin is a physiologic modulator of osteoclast function and a well-established antiresorptive medication that has long been used for the treatment of osteoporosis (4). A range of studies have elucidated that binding of calcitonin to its receptor activates the cAMP–protein kinase A (PKA) and the Ca2+-PKC signaling pathways (5–7). We recently demonstrated that calcitonin receptors are expressed not only on osteoclasts, but also on chondrocytes (8). Calcitonin was shown to directly attenuate cartilage degradation in a manner involving both chondrocyte-mediated increased cAMP levels and a general inhibition of matrix metalloproteinase (MMP) activity (8). Since calcitonin attenuates cartilage degradation partly through a cAMP-dependent mechanism, this raises the possibility that other cAMP-dependent mechanisms might have positive effects on cartilage health. Increased levels of cAMP in chondrocytes may be induced by stimulation of intracellular cAMP production through activation of the adenylate cyclase by forskolin, or by inhibition of cAMP degradation by inhibition of the phosphodiesterases (PDEs) by, for example, 3-isobutyl-1-methyl xanthine (IBMX), a nonspecific PDE inhibitor (9). Thus, whether cAMP may be a common descriptor for cartilage health could potentially be investigated under appropriate experimental conditions using these molecular tools.

Cartilage is predominantly composed of type II collagen (60–70% of dry weight) and proteoglycans (10% of dry weight), of which aggrecan is the most abundant (10). The key mediators of cartilage degradation include the MMPs and the closely related aggrecanases, which are members of the ADAMTS family (11, 12). Aggrecan is degraded by both MMPs and aggrecanases, whereas type II collagen is degraded by MMPs (13). The action of these proteases results in the release of collagen and aggrecan fragments that can be measured both in vitro and in vivo (14). With these protease activities in mind, it is logical to target the action of these key players to stop the progression of cartilage loss in joint diseases such as OA. Accordingly, many investigators have focused their attention on single, selected proteases as targets for potential therapeutics. However, other opportunities may be favorable for modulating chondrocyte phenotype and inhibiting protease activities, in preference to monofocused drugs. For example, potential treatments that are chondrocyte specific but that panspecifically affect protease activity may be preferable, but these have yet to be demonstrated.

Articular cartilage explants exposed to catabolic cytokines such as oncostatin M (OSM) and tumor necrosis factor (TNF) provide a useful ex vivo model of cartilage degradation with a strong likeness to in vivo conditions, since the extracellular matrix is intact and possesses both the regulators and structural components of articular cartilage (13, 15). We have investigated whether pharmacologic interventions that increase intracellular cAMP levels in chondrocytes can inhibit cartilage degradation and thus open new possibilities for the management of OA. We investigated the influence of stimulators of the adenylate cyclase enzyme and of inhibitors of PDEs on degradation of both type II collagen and aggrecan.

MATERIALS AND METHODS

Reagents.

All reagents used were of analytic grade. The culture medium comprised Dulbecco's modified Eagle's medium (DMEM) containing penicillin and streptomycin (Life Technologies, Gaithersburg, MD). Forskolin, IBMX, and human recombinant OSM were from Sigma-Aldrich (Poole, UK). Human recombinant TNF was from R&D Systems (Abingdon, UK).

Cell number.

For the quantification cell viability, we used the alamarBlue assay (Trek Diagnostic Systems, Westlake, OH) as described by the manufacturer.

Experiments on cartilage explants.

To obtain articular cartilage explants, isolated heifer stifle joints from animals ages <1.5 years were used. Samples were cut from the superficial cartilage layer using a surgical scalpel to obtain cartilage explants without adherent calcified cartilage. The dimensions of the articular cartilage explants were ∼1–1.5 × 3 × 3 mm with a homogenous appearance, and the individual explants were weighed (mean ± SD 12 ± 2 mg). The explants were placed in 96-well plates and cultured under the following serum-free conditions: 1) in DMEM only; 2) in DMEM containing OSM (10 ng/ml) and TNF (20 ng/ml); 3) in DMEM containing OSM and TNF with doses of forskolin (a potent stimulator of cAMP) at 4, 16, or 64 μM; or 4) in DMEM containing OSM and TNF with doses of IBMX (a nonspecific inhibitor of cAMP PDEs) at 4, 16, or 64 μM. The conditioned medium (200 μl/well) was fully replaced 3 times a week, and the explants were cultured for 19 days. Additionally, cartilage explants were freeze-thawed 3 times in N2 to kill the chondrocytes and were included as controls for non–chondrocyte-mediated release. These were referred to as metabolically inactive.

Isolation of chondrocytes and cAMP measurements.

Articular chondrocytes were isolated by enzymatic digestion with 0.5% trypsin (Sigma-Aldrich) for 30 minutes and 0.5% collagenase (Wako, Osaka, Japan) for 5 hours at 37°C. Chondrocytes were obtained after filtration and centrifuged at 2,200g for 10 minutes. The pellet was resuspended in DMEM containing 10% fetal bovine serum (Sigma-Aldrich) and subcultured for 2 days at 37°C in 5% CO2. The cells were lifted, centrifuged, cultured under serum-free conditions for 24 hours, and subsequently stimulated with either forskolin or IBMX for 1 hour for the concentration-dependent experiments. To investigate time-dependent effects of the cAMP modulators, chondrocytes were stimulated for 5 or 20 minutes or 1, 8, or 24 hours. After the first passage, the chondrocytes retain their spherical appearance and do not dedifferentiate. Quantification of intracellular cAMP was performed using the cAMP kit (Amersham, London, UK).

Detection of C-terminal crosslinking telopeptide of type II collagen fragments (CTX-II).

CTX-II was measured in the preclinical CartiLaps enzyme-linked immunosorbent assay (ELISA) (Nordic Bioscience Diagnostics, Herlev, Denmark), which is based on a mouse monoclonal antibody (F46) recognizing the 6–amino acid epitope EKGPDP at the C-terminal telopeptide of type II collagen. The assay can be used for measuring levels of CTX-II in conditioned media from explant cultures. The detection limit of the assay is 0.75 ng/ml, and the intra- and interassay coefficients of variation are <6%.

Detection of aggrecan fragments 342FFGV-G2.

Monoclonal antibody AF-28 recognizing the N-terminal neoepitope generated by MMP cleavage of the amino acid sequence IPEN341-342FFGV localized in the interglobular domain of aggrecan has previously been described (16). The 342FFGV-G2 assay combines 2 monoclonal antibodies in a sandwich ELISA; the other antibody, F78, recognizes epitopes in the G1 and G2 globular domains of aggrecan (17).

Detection of aggrecan fragment 374ARGS.

The ELISA detecting the aggrecanase-derived fragments of the N-terminal 374ARGS combines 2 monoclonal antibodies in a sandwich ELISA system. Maxisorp plates were coated with 100 μl rabbit anti-mouse antibody (10 μg/ml) in 1M Na2CO3 buffer (pH 9.6) overnight at 4°C. The next day, plates were washed 5 times and incubated for 1 hour at 20°C with shaking (at 300 revolutions per minute) with 100 μl (500 ng/ml with 1.5% mouse serum) monoclonal BC-3 (Abcam, Cambridge, UK), which recognizes the aggrecanase-generated neoepitope 374ARGS of aggrecan. After washing, 100 μl standards (ADAMTS-4–digested bovine aggrecan [0–10 μg/ml]) or diluted conditioned media from articular cartilage explants were added and incubated for 1 hour at 20°C with shaking. The plates were washed, and 100 μl peroxidase-labeled monoclonal antibody F-78 recognizing the G1 and G2 domains of aggrecan (17), diluted to 2 μg/ml plus 5% normal mouse serum (Calbiochem, Abingdon, UK), was added, and the plates were incubated for another 1 hour at 20°C with shaking. Plates were washed, and 3,3′,5,5′-tetramethylbenzidine solution substrate was added and incubated in the dark for 15 minutes at 20°C with shaking. The reaction was stopped by 0.18M H2SO4, and the absorbance was measured at 450 nm, with 650 nm as the reference wavelength, on a microtiter plate reader (Molecular Devices, Sunnyvale, CA).

Detection of sulfated glycosaminoglycans (sGAG)

The sGAG levels were measured in the conditioned medium using the Alcian blue–binding assay (Euro-Diagnostica, Malmö, Sweden) according to the manufacturer's instructions.

Histology and immunohistology.

On the last day of culture, the articular cartilage explants were fixed in formaldehyde buffer (pH 7.0), washed, paraffin-embedded, and subsequently sectioned into 5-μm sections. Toluidine blue (Bie and Berntsen, Rødovre, Denmark) staining was used to stain the proteoglycans. CTX-II was immunolocalized using the monoclonal F46 antibody diluted 1:5,000 in Tris buffered saline with 5% casein (Sigma-Aldrich). For a control, the CTX-II peptide (Pepceuticals, Leicester, UK) was used at a 10-molar excess to evaluate the specificity of the CTX-II staining (data not shown). Peroxidase-labeled mouse Envision was used as secondary antibody (DakoCytomation, Glostrup, Denmark). Finally, the immunoreactivity was visualized with a liquid 3,3′-diaminobenzidine chromogen solution (Sigma-Aldrich), after which the slides were rinsed in tap water. The nuclei were counterstained using Ehrlich's hematoxylin, and the slides were then dehydrated, cleared, and mounted with glass. Digital images were taken using a BX60 microscope and a C5050-zoom camera (both from Olympus, Lake Success, NY).

Statistical analysis.

Results are shown as the mean ± SEM. All in vitro experiments were repeated at least 4 times with 4–8 replicates of cartilage explants for each treatment. Differences between mean values were compared by Student's t-test for unpaired observations using GraphPad Prism software (GraphPad Software, San Diego, CA), assuming normal distribution where 4 replicates were used. P values less than 0.05 were considered significant.

RESULTS

Inhibition of MMP-mediated cartilage degradation by cAMP modulators.

To test the effect of cAMP modulators on articular cartilage degradation, we initially exposed isolated articular chondrocytes to either forskolin or IBMX, which resulted in a dose-dependent increase in intracellular cAMP levels (Figures 1A and B). At the highest concentration (64 μM), exposure of chondrocytes to forskolin resulted in a significant induction of cAMP levels (>50-fold) (P < 0.001), whereas exposure to IBMX resulted in a smaller, but still significant, induction (3.3-fold) (P < 0.05). In addition, we investigated the time-dependent profile of cAMP induction by forskolin and IBMX in isolated articular chondrocytes. We found that both forskolin and IBMX already induced significantly increased cAMP levels after 5 minutes, and these levels were sustained after 24 hours, as shown in Figures 1C and D. The induction of cAMP levels was significantly higher (P < 0.001) after forskolin stimulation than after IBMX stimulation, as seen in Figures 1A and B.

Figure 1.

Time- and concentration-dependent effects of forskolin and 3-isobutyl-1-methyl xanthine (IBMX) on cAMP levels in isolated articular chondrocytes. Articular chondrocytes were isolated from the extracellular matrix and stimulated with either forskolin or IBMX at concentrations ranging from 0.062 μM to 64 μM for 1 hour or with 64 μM forskolin or IBMX for 5 or 20 minutes or for 1, 8, or 24 hours. A and B, Effect of various concentrations of forskolin (A) or IBMX (B) on cAMP levels. C and D, Time-dependent effect of stimulation with forskolin (C) or IBMX (D). Values are the mean and SEM. ∗ = P < 0.05; ∗∗ = P < 0.01; ∗∗∗ = P < 0.001. Control = medium without cAMP modulators.

Thereafter, to investigate whether modulation of intracellular cAMP in chondrocytes could be linked to changes in MMP-mediated type II collagen degradation or aggrecan degradation, we investigated the release of CTX-II fragments and aggrecan fragments (measured by the novel 342FFGV-G2 ELISA) from articular cartilage explants exposed to OSM and TNF in the presence or absence of forskolin or IBMX. The CTX-II release was measured in the conditioned medium and accumulated throughout the culture period until day 19. OSM and TNF induced a >85-fold increase in released CTX-II (Figure 2A) compared with unstimulated and metabolically inactive control explants (P < 0.001). Forskolin and IBMX significantly inhibited the CTX-II release in a dose-dependent manner. The highest concentrations of both forskolin and IBMX almost abrogated type II collagen degradation (P < 0.001). As shown in Figure 2B, OSM and TNF induced a >600% increase in 342FFGV-G2 compared with unstimulated and metabolically inactive control explants. With regard to MMP-mediated aggrecan degradation, both forskolin and IBMX significantly inhibited the release of 342FFGV-G2 (P < 0.001) at all doses used for these experiments, with a >80% reduction.

Figure 2.

Effect of cAMP modulators on matrix metalloproteinase (MMP)–mediated cartilage degradation. Articular cartilage explants (control) and metabolically inactive explants were cultured without stimulation or were stimulated with 10 ng/ml oncostatin M (OSM) and 20 ng/ml tumor necrosis factor (TNF) with either forskolin or 3-isobutyl-1-methyl xanthine (IBMX). A, Release of neoepitopes from type II collagen degradation in the conditioned medium (C-terminal crosslinking telopeptide of type II collagen fragments [CTX-II]). B, Release of MMP-derived aggrecan degradation products (342-G2) measured by the 342FFGV-G2 enzyme-linked immunosorbent assay. Each bar represents the mean and SEM level of accumulated release from 19 days of culture from 4 individual wells. The values are adjusted for the amount of cultured cartilage. ∗ = P < 0.05; ∗∗ = P < 0.01; ∗∗∗ = P < 0.001, versus TNF + OSM treatment alone.

Lack of effect of increased cAMP levels in articular chondrocytes on aggrecanase-mediated aggrecan degradation.

To investigate whether increased levels of intracellular cAMP in chondrocytes could be linked to changes in aggrecan degradation mediated by aggrecanases, which are the main in situ mediators of aggrecan degradation (11, 12, 18, 19), we investigated aggrecan release from cartilage explants using a novel 374ARGS-G2 ELISA. Aggrecanase cleavage at TEGE373-374ARGS is required for reactivity in the assay, since the 374ARGS epitope is generated by aggrecanases, and intact aggrecan does not react. Thus, the assay is specific for aggrecanase-mediated aggrecan degradation. Articular cartilage explants were stimulated with OSM and TNF in the presence or absence of forskolin or IBMX. The 374ARGS-G2 release was measured in the conditioned medium and accumulated over the 19-day culture period. OSM and TNF induced a significant increase in 374ARGS-G2 (>300%) as compared with unstimulated and metabolically inactive control explants, as shown in Figure 3. Neither forskolin nor IBMX affected the cytokine-stimulated release of 374ARGS-G2 fragments.

Figure 3.

Aggrecanase-mediated aggrecan degradation not blocked by increased cAMP levels in chondrocytes. Articular cartilage explants (control) and metabolically inactive explants were cultured without stimulation or were stimulated with 10 ng/ml OSM and 20 ng/ml TNF with either forskolin or IBMX. The release of aggrecanase-mediated aggrecan fragments (374-G2) into the conditioned medium was measured by the 374ARGS-G2 enzyme-linked immunosorbent assay. Each bar represents the mean and SEM level of accumulated release from 19 days of culture from 4 individual wells. ∗∗ = P < 0.01; ∗∗∗ = P < 0.001, versus TNF + OSM treatment alone. See Figure 2 for definitions.

Inhibition of sGAG release from cartilage explants by IBMX, but not by forskolin.

Sulfated GAG residues are linked to the aggrecan core protein and are predominantly chondroitin sulfate, but also include some keratan sulfate chains. The high negative charge on GAGs confers the water-retaining properties of aggrecan, which is critical for its ability to bear weight and resist compression. Aggrecan fragments released from cartilage explants into conditioned medium contain variable amounts of substituted sGAG chains; therefore, the assay of sGAG in medium is traditionally used to measure total aggrecan release. To further investigate the effect of cAMP modulators on aggrecan turnover, articular cartilage explants were cultured with OSM and TNF in the presence or absence of forskolin or IBMX. The conditioned medium from days 3, 5, and 7 was evaluated in the sGAG assay. As shown in Figure 4, exposure of articular cartilage to OSM and TNF resulted in a significant increase in sGAG release (45%) (P < 0.05). IBMX, but not forskolin, significantly attenuated sGAG release by 72% (P < 0.001), although only at the highest dose.

Figure 4.

Inhibition of sulfated glycosaminoglycan (sGAG) release from articular cartilage by IBMX, but not by forskolin. Articular cartilage explants (control) and metabolically inactive explants were cultured without stimulation or were stimulated with 10 ng/ml OSM and 20 ng/ml TNF with either forskolin or IBMX. The release of sGAG was measured in conditioned medium from days 3, 5, and 7 and adjusted for the amount of cultured cartilage, and the values were accumulated. Bars represent the mean and SEM of 4 individual wells. ∗ = P < 0.05; ∗∗∗ = P < 0.001, versus TNF + OSM treatment alone. See Figure 2 for other definitions.

Protection by cAMP modulators against cartilage degradation, evaluated immunohistochemically and histologically.

To correlate the in situ cartilage composition with the biochemical markers measured above, we evaluated cultured articular explants using immunohistochemical and histologic staining methods. Articular cartilage explants were cultured in the combined presence of OSM and TNF with either forskolin or IBMX at 64 μM. As shown in Figure 5, we assessed proteoglycan content using toluidine blue staining and CTX-II by examining the localization of the CTX-II neoepitope in the articular cartilage matrix. In the absence of cytokine stimulation, aggrecan was readily detected by staining with toluidine blue. In contrast, aggrecan staining was almost completely lost following stimulation with OSM and TNF, indicating significant depletion by these cytokines. In contrast to the control, the cytokines induced vast CTX-II staining in complete alignment with the biochemical markers for CTX-II release (see Figure 2A).

Figure 5.

Abrogation of CTX-II protein generation in forskolin- and IBMX-treated articular cartilage explants. Bovine articular cartilage explants were cultured without stimulation (control), with 10 ng/ml OSM and 20 ng/ml TNF, or with 10 ng/ml OSM and 20 ng/ml TNF and either 64 μM forskolin or 64 μM IBMX. On day 19, the explants were formaldehyde-fixed, paraffin-embedded, sectioned, and immunostained for neoepitopes of type II collagen degradation (CTX-II) (brown staining). Toluidine blue was used to detect aggrecan (blue staining). Bar = 10 μm. (Original magnification × 60.) See Figure 2 for definitions.

When articular explants were cultured in the presence of either forskolin or IBMX, immunohistochemistry revealed that type II collagen degradation and immunoreactivity with the CTX-II antibody were strongly inhibited. This is consistent with the effects of forskolin and IBMX on CTX-II release shown in Figure 2A, as well as with the dose-dependent decrease in MMP activity caused by treatment with forskolin and IBMX (see below). IBMX, but not forskolin, blocked aggrecan release from cytokine-treated explants, consistent with the effect of these agents on the release of sGAG shown in Figure 4.

Possible association of increased cAMP levels with decreased apoptosis.

To eliminate the possibility that the reduced MMP activity seen with cAMP modulators was due to cell toxicity, we measured cell viability with a cell-diffusible dye, alamarBlue, the fluorescence of which correlates with cell numbers (20, 21). The results showed that none of the tested compounds had inhibitory effects on cell viability at the concentrations used in these experiments (Figure 6). In contrast, IBMX showed a positive effect on cell numbers in the presence of OSM and TNF, suggesting less apoptosis. A similar (although nonsignificant) trend was observed for forskolin treatment.

Figure 6.

Effect of increased cAMP levels on chondrocyte viability in articular explants. Articular cartilage explants (control) and metabolically inactive explants were cultured without stimulation or were stimulated with 10 ng/ml OSM and 20 ng/ml TNF with either forskolin or IBMX. Cell viability was evaluated on day 19 with the metabolic dye alamarBlue. Each bar represents the mean and SEM value from 4 individual wells. ∗ = P < 0.05 versus TNF + OSM treatment alone. See Figure 2 for definitions.

Dose-dependent inhibition of MMP activity and expression by elevation of cAMP levels.

OSM and TNF have previously been shown to induce high levels of MMP expression and activity in articular explants (13, 15). To further investigate the mode of action by which stimulators of endogenous levels of cAMP attenuate cartilage degradation, we investigated changes in MMP activity, focusing on MMP-2 and MMP-9 activity in conditioned medium on day 19 using gelatin zymography. OSM and TNF substantially induced gelatinase expression and activity, detected as bands corresponding to the molecular weights of MMP-2 and MMP-9 (Figure 7). Only modest gelatinase activity was observed in the absence of cytokine stimulation. Forskolin and IBMX dose-dependently inhibited MMP-2 and MMP-9 activity and expression (Figure 7).

Figure 7.

Dose-dependent inhibition of MMP activity and expression by elevation of cAMP levels. Articular cartilage explants (control) and metabolically inactive (MI) explants were cultured without stimulation (WO) or were stimulated with 10 ng/ml OSM and 20 ng/ml TNF with either forskolin or IBMX. The MMP activity in the conditioned medium from day 19 is identified on the zymography gel by the standards for MMP-2 and MMP-9 (for proMMP-9 and active MMP-9, 92 kd and 86 kd, respectively; for proMMP-2 and active MMP-2, 72 kd and 66 kd, respectively). See Figure 2 for definitions.

DISCUSSION

Calcitonin has been shown to induce cAMP in chondrocytes and to be chondroprotective (22, 23). The current data clearly demonstrate for the first time that cAMP elevation in articular chondrocytes results in a dose-dependent inhibition of cartilage degradation via inhibition of MMP expression and activity. Taken together, these findings are the first to suggest that chondrocyte-specific cAMP modulators could provide a useful approach for the treatment of progressive and destructive joint diseases.

We investigated the effect of increased intracellular cAMP levels in an ex vivo model of bovine articular cartilage explants subjected to the synergistic catabolic effects of the cytokines OSM and TNF (13). We monitored MMP-generated cartilage degradation by assessment of type II collagen degradation (CTX-II) and MMP-mediated aggrecan degradation by a novel 342FFGV-G2 ELISA. We demonstrated that both forskolin and IBMX dose-dependently inhibited type II collagen degradation and MMP-mediated aggrecan degradation.

Gelatin zymography detects both MMP-2 and MMP-9 and, to a lesser extent, MMP-13 (13, 24). MMP-13 is readily recognized by casein zymography, and only to a lesser extent by gelatin zymography (25). Both forskolin and IBMX dose-dependently inhibited and virtually abrogated MMP-2 and MMP-9 activity at the highest dose. Both the active and latent forms of MMP-9 and MMP-2 were strongly attenuated, suggesting that elevated cAMP levels decreased both the expression and activation of these enzymes. The zymography results are consistent with the results for the biochemical markers of MMP-mediated cartilage degradation of type II collagen (CTX-II) and aggrecan (342FFGV-G2), since these parameters were also inhibited in a dose-dependent manner by forskolin and IBMX. Overall, these 3 very different measurements of the MMP activity of chondrocytes correlated well, and this suggests that in the presence of elevated cAMP levels, MMP activity of chondrocytes is attenuated.

In addition to MMP activity, aggrecanase activity is even more important for aggrecan degradation (11, 12, 18, 19). In explant cultures in vitro, aggrecanase activity is found initially, whereas MMPs are expressed at later stages (26). Using an aggrecanase-specific cleavage ELISA, we investigated whether increased cAMP levels would counter the aggrecanase-mediated aggrecan degradation. We found that neither IBMX nor forskolin inhibited aggrecanase-mediated aggrecan degradation. We observed a nonsignificant trend toward increased aggrecanase activity in response to higher levels of IBMX. We have previously investigated protease profiles in both chondrocytes (13) and osteoclasts (27). Those studies demonstrated that in response to panfamily-specific protease inhibition, compensatory mechanisms take place, leading to altered protein degradation profiles. These compensatory mechanisms might in part explain the observed increase in aggrecanase activity in the face of strongly attenuated MMP activity.

We investigated total aggrecan release by measuring loss of sGAG into culture medium. We focused on the initial period of culture, since other investigators have shown that aggrecan release occurs in the early stage in these cultures (26). We found that IBMX, but not forskolin, attenuated aggrecan loss, but only at the highest dose. This suggests that even though both molecules increased cAMP levels, there are differences between their modes of action that require further investigation.

Using immunohistochemistry, we found that exposure of cartilage explants to OSM and TNF resulted in increased CTX-II immunostaining as well as aggrecan depletion. Both forskolin and IBMX completely inhibited the generation of MMP-derived type II collagen fragments, consistent with the abrogation of MMP expression and activity. IBMX, but not forskolin, inhibited aggrecan loss, as determined by both toluidine blue staining of cultured explants and assay of total aggrecan release in the dye-binding assay. The 6-fold increase in 342FFGV-G2 neoepitope is consistent with the ability of cAMP to decrease MMP-2 and MMP-9 expression and activation, since both these MMPs cleave aggrecan at IPEN341-342FFGV (28). Thus, agents such as forskolin and IBMX that inhibit MMP-2 and MMP-9 activation will also block aggrecan cleavage at IPEN341-342FFGV.

On the other hand, forskolin and IBMX were generally unable to block generation of the aggrecanase neoepitope 374ARGSV; only the highest concentration of IBMX was able to inhibit aggrecan release measured in the dye-binding assay. Since aggrecan loss is predominantly driven by aggrecanases, the inability of cAMP modulators to block aggrecanase activity is consistent with their generalized lack of effect on total aggrecan release. However, at 64 μM IBMX, total aggrecan loss (but not the 374ARGSV neoepitope) was significantly inhibited, which presumably does not reflect an effect on aggrecanase activity, but rather, on the proportion of MMP-mediated aggrecan loss that is prevented by the decrease in MMP activity and expression induced by high IBMX concentrations, since MMPs in previous studies have been shown to be directly linked to aggrecan degradation (29–31).

Even though both forskolin and IBMX result in increased intracellular cAMP levels, this is the product of 2 very different molecular mechanisms. As a result, the pharmacodynamics are very different, which in part may explain the more potent effect of IBMX than of forskolin on proteoglycan degradation. However, further investigation of the role of the amplitude and duration of the cAMP response in chondrocytes is needed to further dissect out this exact mechanism. Similarly, diverse pharmacologic effects are observed with parathyroid hormone (PTH), which also mediates signals in part through cAMP activation. It is well established that transient PTH activation leads to a bone anabolic signal, while sustained actions lead to a catabolic signal (32).

One important question was whether the apparent effect of the cAMP modulators on cartilage degradation represented direct functional effects or was merely due to decreased chondrocyte viability. To exclude this confounder, we carefully assessed this parameter using validated methods (14, 20, 21, 33). We did not identify any significant decreases in cell viability, suggesting that the effect of increased cAMP levels is functional. Furthermore, intracellular modulation of this second messenger might well be an indicator of the metabolic function (anabolic or catabolic) of the chondrocytes.

The present data suggest that specific cAMP modulators in chondrocytes may be chondroprotective. Previous studies by Tenor et al (34) have shown that PDE IV inhibitors, but not PDE III or PDE V inhibitors, blocked nitric oxide synthesis in response to interleukin-1 (IL-1). Thus, the current experiments with a panspecific PDE inhibitor may for the major part be mediated through PDE IV, although others have implicated PDE V in IL-1–mediated events in chondrocytes (35). In addition, aggrecan degradation stimulated by Salmonellatyphosa endotoxin and measured by sGAG release was attenuated by cAMP modulators (36); however, no molecular explanation was provided. This highlights the fact that more focused research on the role of chondrocyte PDEs in response to different stimuli is needed to further understand the role of PDEs in the catabolic processes. These studies may suggest new possibilities for developing specific chondroprotectors.

The current investigations focused on cartilage degradation in ex vivo tissue cultures. Previous investigations have mainly aimed at protein synthesis in isolated chondrocytes. Malemud et al (37–40) investigated the effect of cAMP modulators on cartilage synthesis in isolated chondrocytes. Interestingly, forskolin and IBMX were shown to stimulate proteoglycan synthesis (37) in the absence of OSM and TNF. Taken together, treatment of cartilage with chondrocyte-specific cAMP modulators may result not only in inhibition of cartilage degradation, but also in a shift in the metabolism of the chondrocyte into a cartilage synthesis. In vivo studies in relevant preclinical OA models are needed to further substantiate these preliminary findings and speculations.

Another important limitation of the current experiments is that they have relied on a model system using freshly isolated intact bovine cartilage, which may have a different metabolism from that of human OA cartilage. Thus, further experiments are needed to understand the importance of the current results regarding both human cartilage and, particularly, human OA cartilage.

Cell death and hypertrophy are well-established aspects of the progression of OA (41). Thus, specific inhibitors of chondrocyte hypertrophy and cell death may counteract the progression of OA. In the present experiments, we investigated the effects of OSM and TNF on articular cartilage and found that treatment with IBMX resulted in increased chondrocyte viability. Although the results remain to be demonstrated in hypertrophic articular chondrocytes, they raise the possibility that cAMP modulators may counter hypertrophy and apoptosis and thus improve cartilage health. Further evidence for this hypothesis is found in the growth plate, where inducers of cAMP may down-regulate chondrocyte differentiation and thus delay hypertrophy.

In conclusion, an increasing amount of evidence suggests that cAMP levels in chondrocytes are an important indicator of the metabolic function of the chondrocytes. Low levels reflect cartilage degradation, whereas increased levels suggest attenuation of cartilage degradation and increased cartilage synthesis. PDE inhibitors are frequently used drugs, which illustrates the potential clinical relevance of these findings, although so far, no reports of decreased OA incidence have been presented. However, the present results suggest that it is prudent to investigate the effects of cAMP modulators for their chondroprotective potential in preclinical in vivo models, as well as to reanalyze clinical trials.

AUTHOR CONTRIBUTIONS

Dr. Karsdal had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study design. Karsdal, Christiansen, Sondergaard.

Acquisition of data. Sumer, Wulf, Madsen, Sondergaard.

Analysis and interpretation of data. Karsdal, Wulf, Madsen, Sondergaard.

Manuscript preparation. Karsdal, Fosang, Sondergaard.

Statistical analysis. Karsdal, Wulf, Madsen, Sondergaard.

Ancillary