Mitochondrial dysfunction increases inflammatory responsiveness to cytokines in normal human chondrocytes

Authors


Abstract

Objective

Alterations in mitochondria play a key role in the pathogenesis of osteoarthritis (OA). The role of inflammation in the progression of OA has also acquired important new dimensions. This study was undertaken to evaluate the potential role of mitochondrial dysfunction in increasing the inflammatory response of normal human chondrocytes to cytokines.

Methods

Mitochondrial dysfunction was induced by commonly used inhibitors. Interleukin-1β (IL-1β) and tumor necrosis factor α (TNFα) were used as inflammatory mediators. IL-8 and cyclooxygenase 2 (COX-2) protein and messenger RNA (mRNA) expression and prostaglandin E2 (PGE2) levels were assessed. The chemotactic activity of neutrophils was assayed. Additionally, inhibitors of reactive oxygen species (ROS) and NF-κB were used to identify possible inflammatory response pathways induced by mitochondrial dysfunction, and the effects of the natural antioxidant resveratrol were tested.

Results

Pretreatment with antimycin A or oligomycin (inhibitors of mitochondrial respiratory chain complexes III and V, respectively) triggered a strong potentiation of IL-1β–induced IL-8 mRNA and protein expression (mean ± SEM at 18 hours 5,932 ± 1,995 pg/50,000 cells for IL-1β alone versus 16,241 ± 5,843 pg/50,000 cells for antimycin A plus IL-1β and 20,087 ± 5,407 pg/50,000 cells for oligomycin plus IL-1β; P < 0.05). Similar results were observed with TNFα or when expression of the inflammatory mediator COX-2 or PGE2 production was assessed. Mitochondrial dysfunction increased the chemotactic activity induced by cytokines, and ROS and NF-κB inhibitors decreased the production of IL-8. Resveratrol significantly reduced the inflammatory response.

Conclusion

Our findings indicate that mitochondrial dysfunction could amplify the responsiveness to cytokine-induced chondrocyte inflammation through ROS production and NF-κB activation. This pathway might lead to the impairment of cartilage and joint function in OA.

Osteoarthritis (OA), the most common age-related cartilage and joint disorder, is a degenerative disease characterized by degradation of the matrix and cell death resulting in the gradual loss of articular cartilage integrity (1, 2). The chondrocyte, which is the only cell type present in mature cartilage, is responsible for repairing damaged tissue. Although the primary etiology of the disease is undetermined, OA is now believed to involve a disruption of cartilage homeostasis in which proinflammatory stimuli induce an increase in chondrocyte catabolic processes. Consequently, the role of inflammation in the progression of OA has acquired important new dimensions (3–5).

Accumulating evidence indicates that mitochondrial damage may play a significant role in the pathogenesis of OA (6–8). Recent ex vivo studies have demonstrated mitochondrial dysfunction in human OA chondrocytes, and analyses of mitochondrial electron transport chain activity in these cells show decreased activity of mitochondrial respiratory chain complexes I, II, and III compared to normal chondrocytes (9). This mitochondrial dysfunction may affect several pathways that have been implicated in cartilage degradation, including defective chondrocyte biosynthesis and growth response, cartilage matrix calcification, and increased chondrocyte apoptosis, as well as augmentation of the inflammatory and matrix catabolism responses. Most of these processes are potentially related to the production of both reactive oxygen species (ROS) and reactive nitrogen species (RNS) intermediaries (10–15). In this regard, a correlation has been found between the OA disease stage and the presence of oxidative stress (2, 16–18).

As both the predominant site for ROS production and the prime target of these molecules, mitochondria play a key role in oxidative stress. High levels of oxidative stress may also underlie mitochondrial respiratory chain inhibition, ATP decrease, and mitochondrial DNA (mtDNA) mutation, all of which are related to the severity of the inflammatory process (14, 19, 20). In fact, an augmentation of oxidative mtDNA damage has been observed in OA cartilage (14, 21).

Proinflammatory mediators may also alter mitochondrial function. Accordingly, we previously described the modulation of mitochondrial activity by interleukin-1β (IL-1β), tumor necrosis factor α (TNFα), and nitric oxide in normal human chondrocytes (19, 20). In this scenario, mitochondrial damage produces ROS and RNS that in turn reduce mitochondrial bioenergetics, favoring cell damage and death (14, 18, 22). In this sense, increasing experimental findings support a connection between mitochondrial dysfunction and inflammation (22–24). However, the role of mitochondrial dysfunction in OA-related inflammation is not yet fully understood.

The key mediators of inflammation, IL-8 and prostaglandin E2 (PGE2), are up-regulated in inflamed joint tissue (25–27). Both are spontaneously released by cartilage specimens from patients with OA at significantly higher levels than those released by normal cartilage specimens (4, 27). Specifically, IL-8–mediated inflammation can promote cartilage degradation through matrix metalloproteinase 3 (MMP-3) synthesis and altered chondrocyte differentiation and calcification in OA (28, 29). PGE2 contributes to hyperalgesia and the erosion of cartilage and juxta-articular bone (30). Overproduction of both mediators is likely induced by proinflammatory cytokines such as IL-1β and TNFα. Because it controls the transcription of a number of proinflammatory genes, the redox-sensitive NF-κB pathway is considered a key regulator of tissue inflammation, including IL-8 and cyclooxygenase 2 (COX-2) expression, in several cell types, notably OA chondrocytes (31).

Although current evidence indicates that mitochondrial dysfunction and inflammation are important players in the pathogenesis of OA, data are lacking to support the hypothesis that the connection between these processes amplifies and accelerates the mechanisms contributing to the impairment of cartilage and joint function. Recently, we demonstrated that mitochondrial dysfunction produced a slight increase in COX-2 expression and PGE2 production in normal human chondrocytes in vitro (12). In this study, we investigated the effects of mitochondrial dysfunction on exacerbating the inflammatory response induced by cytokines in normal human chondrocytes, which can play a crucial role in governing the onset and progression of OA. Furthermore, the effect of antioxidant treatment on the above processes was examined.

MATERIALS AND METHODS

Chondrocyte culture.

Normal human chondrocytes were obtained as previously described from the knee joints of 30 adult autopsy donors (mean ± SD age 63 ± 11 years; n = 16 men and 14 women) with no history of joint disease (20). Subcultures of chondrocytes isolated from cartilage were performed with trypsin-EDTA (Gibco Life Technologies), and first-passage cells were used for experiments. Chondrocytes were seeded into 6-well plates (Corning Costar) for RNA or flow cytometric analysis, 12-well plates (Costar) for chemotaxis assays, 96-well plates (Costar) for enzyme-linked immunosorbent assay (ELISA) of IL-8 and PGE2, or 8-well chamber slides (Becton Dickinson) for immunocytochemistry studies. When cells reached confluence, they were made quiescent by 48-hour incubation in Dulbecco's modified Eagle's medium (Gibco Life Technologies) containing 0.5% fetal calf serum (FCS; Gibco). After washing, the experiments were performed without FCS. All studies were performed strictly in accordance with current local ethics regulations.

Reagents and cell treatments.

Antimycin A and oligomycin (both from Sigma-Aldrich) were used as inhibitors of mitochondrial respiratory chain complexes III and V, respectively (11, 12, 32, 33). IL-1β or TNFα (both from Sigma-Aldrich) was used to induce an inflammatory response. Chondrocytes were preincubated with antimycin A or oligomycin for 1 hour before adding IL-1β (0.5, 1.5, or 5 ng/ml) or TNFα (5 ng/ml) for the necessary time period for RNA or flow cytometry analysis, ELISA for the determination of IL-8 and PGE2 levels, chemotaxis assays, or immunocytochemistry studies. N-acetylcysteine (NAC; 40 mM) was used as an ROS scavenger, and BAY 11-7085 (5 μM; Calbiochem) was used to prevent NF-κB activation. The natural antioxidant resveratrol (50 or 250 μM; Sigma-Aldrich) was tested for its effect as an inflammatory response modulator. BD GolgiStop (Becton Dickinson) containing monensin (0.7 μl/ml) was used to prevent cytokine release in flow cytometry and immunohistochemistry studies.

Analysis of IL-8 and COX-2 messenger RNA (mRNA) expression.

To isolate mRNA, TRIzol reagent was used according to the recommendations of the manufacturer (Invitrogen). Isolated mRNA from 5 × 105 cells was treated with DNase I (Invitrogen) and reverse transcribed with a Transcriptor First Strand cDNA Synthesis Kit, according to the recommendations of the manufacturer (Roche Diagnostics). Real-time polymerase chain reaction analysis for IL-8, COX-2, and the housekeeping gene porphobilinogen deaminase was performed using a LightCycler 4800 SYBR Green I Master kit and a real-time LightCycler (Roche Diagnostics).

Analysis of IL-8 and COX-2 protein levels by flow cytometry.

IL-8 and COX-2 protein expression measurements were obtained by flow cytometric analysis using a FACSCalibur cytometer (Becton Dickinson). Four hours prior to harvesting, BD GolgiStop (0.7 μl/ml) was added to retain IL-8 within the cells. Stimulated cells (5 × 105 cells per well) were collected, washed with phosphate buffered saline (PBS), and fixed and permeabilized with 0.2% saponin (Sigma-Aldrich) in 4% paraformaldehyde (Panreac Química). Nonspecific binding was blocked by washing twice with PBS–1% bovine serum albumin (Sigma-Aldrich), and cells were incubated overnight with fluorescein isothiocyanate–labeled anti–IL-8 (Abcam) or phycoerythrin-labeled anti–COX-2 (Becton Dickinson).

IL-8 and PGE2 assays.

The levels of IL-8 and PGE2 in culture supernatants from chondrocytes (5 × 104) were determined using commercially available ELISA kits for IL-8 (ImmunoTools) and PGE2 (Amersham) according to the recommendations of the manufacturers. Data are expressed as picograms released per 50,000 cells. The working range was between 1.0 and 240 pg/ml for IL-8 and between 2.5 and 230 pg/well for PGE2.

Chemotaxis assay.

Polymorphonuclear leukocytes were isolated from the anticoagulated blood of healthy volunteers by Histopaque (Sigma-Aldrich) density-gradient centrifugation, gelatin sedimentation, and hypotonic lysis of red cells. The cells were suspended in 20 mM HEPES (Sigma-Aldrich), pH 7.4, at a concentration of 1 × 107 cells/ml. The chemotactic activity of culture supernatants was evaluated in 24-well Transwell chemotaxis chambers (3-μm pore polycarbonate membrane; Corning Costar). Neutrophils (100 μl HEPES; 1 × 106 cells) were placed in the upper chamber, and the lower well was loaded with 600 μl of supernatants from cultured chondrocytes. The plate was incubated at 37°C in 5% CO2 for 1 hour, and the migrating cells in the lower compartment were then counted by flow cytometry. Random migration was defined as the response to medium alone. In some experiments, supernatants were preincubated with 500 ng/ml IL-8 neutralizing antibody for 1 hour at 37°C before chemotactic activity was tested. Specific chemotaxis was represented by the mean number of migrating cells minus the mean number of randomly migrating cells. Controls included oligomycin in the lower compartment.

Immunocytochemistry.

Chondrocytes (2 × 104 cells per well in 8-well chamber slides) were fixed with ice-cold acetone for 10 minutes. BD GolgiStop (0.7 μl/ml) was added 4 hours prior to fixing the cells to retain IL-8 within the cells. After pretreatment with 1% hydrogen peroxidase–methanol to inactivate endogenous peroxidase, the cells were washed and incubated with rabbit anti-human IL-8 (Becton Dickinson) for 1 hour. The chambers were then washed with PBS, and peroxidase-labeled goat anti-rabbit secondary antibody (Dako) was added and incubated for 30 minutes. The cells were then stained with Gill's hematoxylin no. 3 (Merck), washed, and examined using a microscope (Olympus BX61; Olympus Biosystems).

Statistical analysis.

Data are expressed as the mean ± SEM or representative results are shown, as indicated. Individual donor samples were studied in duplicate or triplicate; cells from different donors were not pooled for any procedure. A GraphPad Prism version 5 statistical software package was used to perform one-way analysis of variance and Friedman's test. Wilcoxon's paired comparison test was used to assess paired groups. P values less than 0.05 were considered significant.

RESULTS

Mitochondrial inhibitors exacerbate the IL-8 and COX-2 expression induced by cytokines in normal human chondrocytes.

To test the hypothesis that mitochondrial dysfunction modulates the inflammatory response induced by cytokines in normal human chondrocytes, we evaluated whether mitochondrial dysfunction amplifies the IL-8 expression induced by IL-1β. First, the effect of mitochondrial inhibitors (antimycin A and oligomycin) on chemokine production was tested. Both antimycin A and oligomycin induced a faint but significant dose-dependent increase in IL-8 mRNA expression (P ≤ 0.05; n = 6 experiments) (Figure 1A). After 4 hours of incubation, the maximum increase in IL-8 mRNA expression was observed with a concentration of 20 μg/ml antimycin A or 20 μg/ml oligomycin (mean ± SEM increase of 38.01 ± 21.5 and 174.84 ± 95.83, respectively, compared with a basal level set at 1). The positive control, IL-1β (5 ng/ml), induced a 3.9 × 104-fold increase in IL-8 mRNA expression (Figure 1A). When the concentration of IL-8 protein was evaluated by ELISA, the results obtained were consistent with those obtained for IL-8 mRNA expression (Figure 1B), with the highest levels observed at 18 hours of incubation with 20 μg/ml antimycin A or 20 μg/ml oligomycin (P ≤ 0.05; n = 6 experiments).

Figure 1.

Antimycin A and oligomycin induce interleukin-8 (IL-8) mRNA and protein expression in cultured normal human chondrocytes. A, IL-8 mRNA expression, quantified by real-time reverse-transcription polymerase chain reaction, in chondrocytes incubated under basal conditions, in the presence of the positive control IL-1β (5 ng/ml), or with increasing doses of an inhibitor of mitochondrial respiratory chain complex III, antimycin A (10, 20, or 40 μg/ml) or an inhibitor of mitochondrial respiratory chain complex V, oligomycin (5, 10, or 20 μg/ml) for 4 hours. B, Levels of IL-8 released into the culture medium, determined by enzyme-linked immunosorbent assay, from chondrocytes stimulated with IL-1β (5 ng/ml), antimycin A (10 or 20 μg/ml), or oligomycin (10 or 20 μg/ml) for 9 or 18 hours. Values are the mean ± SEM (n = 6 independent experiments performed in duplicate). ∗ = P ≤ 0.05; ∗∗ = P ≤ 0.01; ∗∗∗ = P ≤ 0.001 versus basal level.

Second, we demonstrated that pretreatment of chondrocytes with antimycin A or oligomycin intensified the IL-8 mRNA expression induced by IL-1β, significantly lowering the threshold concentration of IL-1β required for the expression of IL-8 (P ≤ 0.05; n = 6 experiments) (Figure 2A). Similar results were observed when IL-8 protein expression induced by different doses of IL-1β was evaluated by flow cytometry (Figure 2B). Antimycin A or oligomycin pretreatment produced modulation similar to that seen with a greater concentration of IL-1β alone. These results were confirmed by immunocytochemistry (Figure 2C). Parallel results were obtained when IL-8 protein expression was quantified using ELISA (Figure 2D). As shown in Figure 2D, pretreatment with antimycin A and oligomycin resulted in increases up to 279% and 339%, respectively, over that induced by IL-1β (0.5 ng/ml) alone (mean ± SEM at 18 hours 5,932 ± 1,995 pg/50,000 cells for IL-1β alone versus 16,241 ± 5,843 pg/50,000 cells for antimycin A plus IL-1β and 20,087 ± 5,407 pg/50,000 cells for oligomycin plus IL-1β).

Figure 2.

Mitochondrial dysfunction modulates the interleukin-8 (IL-8) expression induced by IL-1β in normal human chondrocytes. A, IL-8 mRNA expression in chondrocytes incubated with 10 μg/ml of either antimycin A (AA) or oligomycin (OLI) for 1 hour and then treated with 1.5 ng/ml IL-1β for 4 hours. IL-1β (5 ng/ml) was used as a positive control. Values are the mean ± SEM (n = 6 independent experiments performed in duplicate). B, IL-8 protein expression, quantified by flow cytometry, in chondrocytes incubated under basal conditions or with increasing doses of IL-1β (0.5, 1.5, or 5 ng/ml) for 8 hours, with or without pretreatment with antimycin A or oligomycin. Values are the mean ± SEM median fluorescence intensity (n = 6 independent experiments). C, Immunochemical analysis of IL-8 expression. Results are representative of 3 independent experiments. Original magnification × 200. D, IL-8 concentration, determined by enzyme-linked immunosorbent assay, in chondrocytes stimulated with medium alone or with the indicated stimuli for 18 hours. Values are the mean ± SEM (n = 6 independent experiments performed in duplicate). E, Neutrophil chemotactic activity of supernatants collected from chondrocytes preincubated with oligomycin (5 μg/ml) for 1 hour before stimulation with 0.5 ng/ml IL-1β for 18 hours. Values are the mean ± SEM (n = 6 independent experiments). ∗ = P ≤ 0.05; ∗∗ = P ≤ 0.01 versus IL-1β alone.

Since we observed that mitochondrial dysfunction increased cytokine-induced chemokine expression, we used Transwell migration assays to investigate whether up-regulated IL-8 was functional and also resulted in an enhancement of polymorphonuclear cell chemotaxis. As shown in Figure 2E, pretreatment of chondrocytes with oligomycin resulted in increased chemotactic activity of neutrophils, showing a mean ± SEM 218 ± 41% increase in cell migration compared with that induced by IL-1β (0.5 ng/ml) alone. Incubation of chondrocyte supernatant with anti–IL-8 antibody resulted in a marked reduction of neutrophil migration (46.65% inhibition), indicating that chemotaxis activity was due at least in part to the production of IL-8. In these experiments, oligomycin failed to demonstrate intrinsic chemotactic activity at the concentration used (data not shown).

COX-2 is also a major player in the inflammatory process. We further demonstrated that mitochondrial dysfunction in normal human chondrocytes significantly intensified COX-2 gene and protein expression and PGE2 production in cells exposed to IL-1β (P ≤ 0.05; n = 5 experiments) (Figure 3). Additionally, TNFα is, together with IL-1β, one of the principal cytokines that drive cartilage destruction. We established that pretreatment of chondrocytes with oligomycin amplified the IL-8 expression induced by TNFα, at both the gene level (Figure 4A) and the protein (Figure 4B) level (P ≤ 0.05; n = 7 experiments). Similar results were obtained for PGE2 production (data not shown).

Figure 3.

Mitochondrial dysfunction modulates the cyclooxygenase 2 (COX-2) mRNA and protein expression and prostaglandin E2 (PGE2) production induced by interleukin-1β (IL-1β) in normal human chondrocytes. A, COX-2 mRNA expression in chondrocytes incubated with 10 μg/ml oligomycin (OLI) for 1 hour before treatment with 1.5 ng/ml IL-1β for 4 hours. Values are the mean ± SEM (n = 5 independent experiments performed in duplicate). B, COX-2 protein expression, quantified by flow cytometry, in chondrocytes incubated under basal conditions or with the indicated stimuli for 6 hours. Values are the mean ± SEM median fluorescence intensity (n = 4 independent experiments). C, PGE2 concentration, determined by enzyme-linked immunosorbent assay, in chondrocytes incubated under basal conditions or with the indicated stimuli for 18 hours. Values are the mean ± SEM (n = 10 independent experiments performed in duplicate). ∗ = P ≤ 0.05 versus IL-1β alone.

Figure 4.

Mitochondrial dysfunction modulates the interleukin-8 (IL-8) mRNA and protein expression induced by tumor necrosis factor α (TNFα) in normal human chondrocytes. A, IL-8 mRNA expression in chondrocytes preincubated with 5 μg/ml oligomycin (OLI) for 1 hour before stimulation with 5 ng/ml TNFα for 4 hours. B, IL-8 concentration, determined by enzyme-linked immunosorbent assay, in chondrocytes incubated under basal conditions or with the indicated stimuli for 18 hours. Values are the mean ± SEM (n = 7 independent experiments performed in duplicate). ∗ = P ≤ 0.05 versus TNFα alone.

ROS and NF-κB involvement in exacerbation of the inflammatory response to cytokines produced by mitochondrial dysfunction.

The mitochondrion is the principal source and target of ROS. We found that preincubation with NAC, a ROS scavenger, for 1 hour significantly decreased the expression of IL-8 protein in cells pretreated with antimycin A or oligomycin before stimulation with IL-1β (P ≤ 0.05; n = 4 experiments) (Figure 5A). Because increased oxidative stress may lead to the up-regulation of redox-sensitive transcription factors such as NF-κB, and because NF-κB participates in the induction of IL-8 gene expression in several cell types, we preincubated chondrocytes with BAY 11-7085, an NF-κB inhibitor. Preincubation with BAY 11-7085 for 1 hour significantly decreased the expression of IL-8 protein induced by pretreatment with mitochondrial respiratory chain inhibitors in IL-1β–stimulated chondrocytes (P ≤ 0.05; n = 4 experiments) (Figure 5B).

Figure 5.

Reactive oxygen species and NF-κB are implicated in the exacerbation of the inflammatory response to cytokines produced by mitochondrial dysfunction. A, Interleukin-8 (IL-8) protein expression, quantified by flow cytometry, in chondrocytes preincubated with the antioxidant N-acetylcysteine (NAC; 40 mM) for 1 hour before treatment with antimycin A (AA; 10 μg/ml) plus IL-1β (1.5 ng/ml) or oligomycin (OLI; 10 μg/ml) plus IL-1β (1.5 ng/ml) for 8 hours. Values are the mean ± SEM median fluorescence intensity (n = 4 independent experiments). ∗ = P < 0.05; ∗∗ = P ≤ 0.01 versus control without antioxidant treatment. B, IL-8 protein expression, quantified by flow cytometry, in chondrocytes preincubated with BAY 11-7085 (BAY; 5 μM) for 1 hour before treatment with the indicated stimuli for 8 hours. Values are the mean ± SEM median fluorescence intensity (n = 4 independent experiments). ∗ = P ≤ 0.05; ∗∗ = P ≤ 0.01 versus control without BAY 11-7085 treatment.

Because the goal of treatment strategies using antioxidants is to attenuate mitochondrial oxidative stress, we tested the ability of the natural antioxidant resveratrol to reduce the inflammatory response induced by the synergy between mitochondrial dysfunction and cytokines. Evaluation of IL-8 protein expression by flow cytometry showed that coincubation with resveratrol greatly reduced the IL-8 levels exacerbated in our model of synergy between mitochondrial dysfunction and IL-1β (P ≤ 0.01; n = 4 experiments) (Figure 6A). This effect of resveratrol was confirmed when IL-8 protein levels were assessed by ELISA (P ≤ 0.05; n = 6 experiments) (Figure 6B). Furthermore, treatment with resveratrol caused a decrease in the chemotactic activity induced by IL-1β in our model of mitochondrial dysfunction (92% inhibition) (P ≤ 0.05; n = 5 experiments) (Figure 6C).

Figure 6.

Resveratrol inhibits the inflammatory response to cytokines up-regulated by mitochondrial dysfunction. A, Interleukin-8 (IL-8) protein expression, quantified by flow cytometry, in normal human chondrocytes coincubated with the natural antioxidant resveratrol (RSV; 250 μM) and treated with antimycin A (AA; 10 μg/ml) plus IL-1β (1.5 ng/ml) or oligomycin (OLI; 10 μg/ml) plus IL-1β (1.5 ng/ml) for 8 hours. Values are the mean ± SEM (n = 4 independent experiments). B, IL-8 concentration, determined by enzyme-linked immunosorbent assay, in supernatants from chondrocytes incubated with oligomycin (5 μg/ml) plus IL-1β (0.5 ng/ml) in the absence or presence of resveratrol (250 or 50 μM) for 18 hours. Values are the mean ± SEM (n = 6 independent experiments performed in duplicate). C, Neutrophil chemotactic activity of supernatants collected from chondrocytes treated with oligomycin (5 μg/ml) plus IL-1β (0.5 ng/ml) in the absence or presence of resveratrol (50 μM) for 18 hours. Values are the mean ± SEM (n = 5 independent experiments). ∗ = P ≤ 0.05; ∗∗ = P ≤ 0.01 versus control without resveratrol treatment.

DISCUSSION

In recent years, accumulating evidence has indicated that mitochondrial damage may play a significant role in the pathogenesis of OA (6–8). In addition, the role of inflammation in the progression of OA has acquired important new dimensions (3, 4). In this sense, increasing experimental data support a connection between inflammation and mitochondrial dysfunction (22–24). Importantly, we previously demonstrated that this mitochondrial dysfunction per se may generate low-grade inflammatory and matrix degradation processes in normal human chondrocytes in vitro (12, 13). However, whether the preexisting mitochondrial dysfunction described in OA chondrocytes (9) intensifies cytokine-induced chondrocyte inflammation remained unknown.

In this study, we focused on the potential of mitochondrial dysfunction to increase the cytokine-induced inflammatory response in normal human chondrocytes. This is the first study to show that mitochondrial dysfunction in normal human chondrocytes increases the production of inflammatory mediators such as IL-8, COX-2, and PGE2 in response to cytokines through ROS generation and activation of the transcription factor NF-κB. In addition, we showed that resveratrol significantly reduced this inflammatory response.

IL-8 and PGE2 are 2 key players that are up-regulated in inflamed joint tissue (25, 26, 34, 35) and in other age-related inflammatory diseases, including Alzheimer's disease, cancer, and atherosclerosis. In fact, cartilage specimens from patients with OA spontaneously released PGE2 and expressed IL-8 mRNA in ex vivo culture at levels at least 50-fold higher and 15-fold higher, respectively, than those observed in normal cartilage (4, 27). This up-regulation may occur secondary to the activation of inflammatory cytokines (27, 36) but may also be independent of cytokine activation (37). With regard to the latter possibility, we recently demonstrated that the inhibition of mitochondrial respiratory chain activity induces a slight increase in COX-2 expression and PGE2 production in normal human chondrocytes (12).

In the present study, when mitochondrial dysfunction was induced in normal chondrocytes with inhibitors of mitochondrial respiratory chain complexes III and V (antimycin A and oligomycin, respectively), a slight but significant increase in IL-8 mRNA and protein levels was also observed, which was dependent on both dose and incubation time. These findings are consistent with those obtained in human liver slices, where mitochondrial injury induced by pharmaceutical inhibitors of fatty acid oxidation led to a significant increase in IL-8 gene and protein expression (37). Mitochondrial dysfunction also induces PGE2 liberation through 4-hydroxynonenal, a lipid peroxidation end product, which is produced abundantly in OA articular tissue and was recently identified as a potent catabolic factor in OA cartilage (38). In fact, 4-hydroxynonenal induces COX-2 expression and PGE2 release in human OA chondrocytes (38).

To examine the potential role of mitochondrial dysfunction in increasing the vulnerability of cells to a cytokine-induced inflammatory response, we evaluated whether mitochondrial dysfunction in chondrocytes increases the level of IL-8 and PGE2 production induced by cytokines. Numerous in vitro and in vivo studies have shown that IL-1β and TNFα are the primary proinflammatory and catabolic cytokines involved in the initiation and progression of articular cartilage destruction (5, 14, 20). Furthermore, the increased levels of catabolic enzymes, prostaglandins, ROS, nitric oxide, and other markers in OA fluids and tissue appear to be related to elevated levels of IL-1β and TNFα. In the present study, antimycin A and oligomycin treatment of normal human chondrocytes resulted in synergistic amplification of the inflammatory response induced by the cytokines IL-1β or TNFα. Hence, we observed that mitochondrial dysfunction increases the IL-8 expression and chemotactic activity induced by these cytokines. We also found that mitochondrial dysfunction aggravates the COX-2 expression and production of its metabolic end product, PGE2, induced by IL-1β.

The findings of the present study are supported by those obtained in other studies demonstrating that mitochondrial dysfunction increases the inflammatory response to different catabolic stimuli. In lung epithelial cells, preexisting mitochondrial dysfunction induced by antisense oligonucleotides to ubiquinol cytochrome c reductase core protein II in mitochondrial respiratory chain complex III increased mitochondrial ROS generation, resulting in a marked potentiation of ragweed pollen extract–induced accumulation of inflammatory cells in the airways (39). Interestingly, recent data also revealed that PGE2 promotes IL-1 expression in articular chondrocytes, thus amplifying the local inflammatory process (30). In addition, the authors of that previous report showed that combined treatment of IL-1 with PGE2 synergistically accelerated the expression of pain-associated molecules, including nitric oxide synthase and IL-6 (30). These findings confer even more physiologic relevance to our findings regarding the effects of mitochondrial dysfunction.

Our results demonstrate a significant mitochondrial dysfunction–related increase in the chondrocyte inflammatory response. Consistent with our results, OA chondrocytes are more responsive to IL-1β than are normal chondrocytes. The spontaneous production of PGE2 by human OA cartilage is notably greater than that by normal cartilage, and the addition of cytokines augments this effect to ∼30-fold greater production compared to that observed in normal cartilage treated with cytokines (27). Similar effects have been described with proMMP-3 and proMMP-9 (40). Specifically, mitochondria from OA chondrocytes have been reported to be more sensitive to the DNA-damaging effects of the proinflammatory cytokines IL-1β and TNFα than are mitochondria from normal chondrocytes (14). Our results, and those of other investigators, suggest that mitochondrial dysfunction may contribute to the inflammatory phenotype observed in OA through the expression of inflammatory mediators.

In a previous study, we demonstrated that chondrocytes stimulated with antimycin A or oligomycin produced ROS, triggering an inflammatory response that was significantly reduced by pretreatment with the ROS scavenger NAC (12). In this sense, the mitochondrial respiratory chain is one of the most important sites for ROS generation. The cumulative oxidative stress caused by ROS has been implicated as a key factor in OA. The present study demonstrated that when chondrocytes were treated with the ROS scavenger NAC, IL-8 expression induced by the synergy between mitochondrial dysfunction and cytokines was significantly reduced, confirming that ROS production is a key step in this inflammatory pathway. However, because NAC did not completely block the inflammatory response, other mitochondrial damage–associated molecules such as ATP, mtDNA, or RNS may also be involved in this inflammatory process (11, 22, 24). In other cell types, mitochondrial dysfunction also increased the generation of ROS, resulting in potentiation of cytotoxicity or inflammatory cell accumulation (39, 41). TNF receptor type I mutant cells exhibited altered mitochondrial function with enhanced ROS generation, and pharmacologic blockade of mitochondrial ROS (i.e., with NAC) reduced inflammatory cytokine production induced by lipopolysaccharide (42). Additionally, antioxidant treatments can improve disease progression in animal models of OA (43).

Increased oxidative stress may lead to the up-regulation of redox-sensitive transcription factors, such as NF-κB, that contribute to the proinflammatory phenotypic alterations in OA tissue, including the induction of IL-8 and COX-2 expression. In our study, the pharmacologic modulation of NF-κB with BAY 11-7085, an inhibitor of NF-κB activation, significantly prevented up-regulation of IL-8 expression, suggesting that NF-κB may modulate the cytokine-induced IL-8 expression associated with mitochondrial dysfunction. Furthermore, we also found that oligomycin potentiated NF-κB activation induced by IL-1β (Vaamonde-García C, et al: unpublished observations). Other groups have found similar results, reporting enhanced sensitivity to activation of NF-κB in the setting of mitochondrial dysfunction (41, 44). However, other redox-sensitive transcription factors or other means of posttranscriptional regulation, such as modulation of mRNA degradation, could also be involved (45, 46).

Because of the critical roles of oxidative stress and inflammation in the pathology of inflammatory diseases such as OA, a debate over various therapeutic strategies using antioxidant compounds is well under way. Resveratrol, a natural compound found in high concentrations in grape skin and red wine, shows great promise as an antioxidant and antiinflammatory agent. In animal models, resveratrol protected against age-related diseases in mice and improved mitochondrial function (47, 48). In vitro, this molecule has proven to have a number of beneficial effects in several cell types, i.e., by reducing NF-κB activation, PGE2 production, and free radical formation or by inducing mitochondrial biogenesis and protecting against chondrocyte apoptosis (49). In this study, we demonstrated that supplementation with this natural antioxidant significantly reduced the IL-8 levels and the chemotaxis induced by the synergy between mitochondrial dysfunction and cytokines. Similar findings have been reported in human monocytic U937 cells, in which resveratrol inhibited phorbol 12-myristate 13-acetate–induced IL-8 production at the protein and mRNA levels (50). Our results suggest that resveratrol may be a useful treatment strategy in OA.

On the whole, we have provided additional support to the hypothesis that a decline in mitochondrial function participates in the chondrocyte inflammatory phenotype observed in OA pathology. The results of the present study verify that mitochondrial dysfunction alone may generate low-grade inflammation in chondrocytes and demonstrates, for the first time, that a decline in mitochondrial function increases chondrocyte inflammatory responsiveness to cytokines, accelerating the mechanisms that may contribute to the impairment of cartilage and joint function in OA and aging. Future studies should be undertaken to evaluate whether resveratrol is a possible treatment of OA.

AUTHOR CONTRIBUTIONS

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. López-Armada 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 conception and design. Vaamonde-García, Riveiro-Naveira, Valcárcel-Ares, Blanco, López-Armada.

Acquisition of data. Vaamonde-García, Riveiro-Naveira, Valcárcel-Ares, Hermida-Carballo, López-Armada.

Analysis and interpretation of data. Vaamonde-García, Riveiro-Naveira, Valcárcel-Ares, López-Armada.

Acknowledgements

We are grateful to Ms Pilar Cal Purriños for expert secretarial assistance. We express appreciation to Lourdes Sanjurjo and Maria Dolores Velo from the Orthopedic Department of the Complexo Hospitalario Universitario A Coruña (CHUAC) and to Francisco Jose Pérez Llarena from the Microbiology Department of CHUAC for mycoplasma analysis.

Ancillary