ERK1/2-Mediated Activation of DRP1 Regulates Mitochondrial Dynamics and Apoptosis in Chondrocytes

Abstract

Objective.

To determine the Dynamin-related protein 1 (DRP1) regulation of mitochondrial fission in chondrocytes under pathological conditions, an area which is underexplored in osteoarthritis pathogenesis.

Design.

DRP1 protein expression was determined by IHC or IF staining of cartilage sections. IL-1β-induced DRP1 mRNA expression in chondrocytes was quantified by qPCR and protein expression by immunoblotting. Mitochondrial fragmentation in chondrocytes was visualized by MitoTracker staining or IF staining of mitochondrial marker proteins or by transient expression of mitoDsRed. Mitochondrial ROS levels were determined by MitoSOX staining. Apoptosis was determined by LDH release assay, Caspase 3/7 activity assay, propidium iodide, and TUNEL staining and IF staining of cleaved caspase 3. Cytochrome c release was determined by confocal microscopy. Surgical destabilization of the medial meniscus (DMM) was used to induce OA in mice.

Results.

Expression of DRP1 and mitochondrial damage was high in human OA cartilage and in the joints of mice subjected to DMM surgery which also showed increased chondrocytes apoptosis. IL-1β-induced mitochondrial network fragmentation and chondrocyte apoptosis via modulation of DRP1 expression and activity and induce apoptosis via Bax-mediated release of Cytochrome c. Pharmacological inhibition of DRP1 activity by Mdivi-1 blocked IL-1β-induced mitochondrial damage and apoptosis in chondrocytes. Additionally, IL-1β-induced activation of ERK1/2 is crucial for DRP1 activation and induction of mitochondrial network fragmentation in chondrocytes as these were blocked by inhibiting ERK1/2 activation.

Conclusions.

These findings demonstrate that ERK1/2 is a critical player in DRP1-mediated induction of mitochondrial fission and apoptosis in IL-1β-stimulated chondrocytes.

Introduction

Mitochondrial dysfunction and the associated catabolic effects contribute to several age-related diseases including arthritis. Mitochondrial fission and fusion are fundamental aspects of mitochondrial biology as the balance between fission and fusion regulate the size and distribution of mitochondria in cells. The mitochondrial dynamics is tightly regulated by fusion (mitofusins) and fission proteins (dynamin related protein-1 and fission 1). Dynamin-Related Protein 1 (DRP1, also known as DLP1, DNM1L, or DVLP) is a key regulator of mitochondrial fission which separates the defective mitochondria from the mitochondrial network [1]. DRP1 is a GTPases and functions in the mitochondrial and peroxisomal division and mediates membrane fission through oligomerization into membrane-associated tubular structures that wrap around the scission site to constrict and sever the mitochondrial membrane through a GTP hydrolysis-dependent mechanism. Activation of DRP1 by cellular kinase triggers its translocation to mitochondria through its interaction with mitochondrial outer membrane proteins such as mitochondrial fission protein-1, mitochondrial fission factor, and mitochondrial elongation factor-1 and -2 [2].

Previous studies have shown the importance of mitochondrial dysfunction in osteoarthritis (OA) pathogenesis [3–5]; but the upstream mechanism behind mitochondrial dysfunction remains unknown. Here we show that chondrocytes under pathological conditions in in vivo and in vitro show excessive mitochondrial network fragmentation. We found increased expression and activation of DRP1 in the OA cartilage of human and mouse and in chondrocytes and cartilage explants treated with IL-1β that was associated with excessive mitochondrial network fragmentation, oxidative stress, apoptosis, and cartilage extracellular matrix (ECM) degradation. Importantly, inhibition of DRP1 mediated mitochondrial fission with mitochondrial division inhibitor-1 (Mdivi-1) suppressed the IL-1β-induced mitochondrial damage, Cytochrome c release, and apoptosis in chondrocytes in vitro. Our data also demonstrated that ERK1/2 are the kinases that activate DRP1 under pathological conditions and ERK1/2 inhibition suppressed DRP1 activation, mitochondrial network fragmentation, and apoptosis.

Materials and Methods

Human and mouse chondrocytes preparation.

The use of discarded, deidentified human cartilage was approved by Northeast Ohio Medical University (NEOMED) and SUMMA IRB. Primary OA and normal human chondrocytes were prepared by sequential enzymatic digestion as described in many of our previous studies [6–8]. Briefly, the cartilage was resected from the donor samples and digested with pronase (1 mg/ml) for 1 h followed by Collagenase (1 mg/ml) for overnight. The digested cells were passed through a 70 μm cell strainer to remove any undigested tissue and chondrocytes were suspended in DMEM supplemented with 1% antibiotics and 10% fetal bovine serum at a density of 1×10⁶ cells per well in 6 well plates. Primary human chondrocytes were used without any passage. A total of 21 donor samples [15 OA (5 males, 10 females, mean age 64±7 years) and 6 normal (3 males, 3 females, mean age 41±7 years)] were used in this study. The animal protocol was approved by the Institutional Animal Care and Use Committee (IACUC) of NEOMED. Mice were housed in an environmentally controlled facility and provided food and water ad libitum. Mouse chondrocytes were prepared from 4–5-week-old male and female C57BL/6 mice as described previously [9]. Briefly, mice were sacrificed, and knee joints were collected for harvesting the cartilage. The isolated cartilage tissue was washed with PBS to get rid of any contaminating cells, sliced into small pieces, and digested with Collagenase. The digested cells were passed through a 70 μm cell strainer to remove any undigested tissue and cells were suspended in DMEM supplemented with 1% antibiotics and 10% non-heat inactivated fetal bovine serum at a density of 1×10⁶ cells per well in 6 well plates. Mouse primary chondrocytes were used after two passages.

Destabilization of medial meniscus (DMM) surgery and induction of experimental OA in mice.

C57BL/6 male mice (12-weeks, n=7 per group) were assigned to DMM and Sham surgery group as described previously [9,10]. Mice were sacrificed 8-week post DMM and knee joints were collected for histology and immunostaining. Histological evaluation of disease severity was determined using the OARSI scoring system [10].

Determination of mitochondrial network fragmentation in chondrocytes.

Mitochondrial network fragmentation was determined as described previously [10]. Briefly, primary human or mouse chondrocytes or TC28 cells (0.2×106) were seeded on glass coverslips in a 24-well plate for 2–3 days and stained with MitoTracker Deep Red (100 nM) for 10 minutes followed by treatment with IL-1β (5 ng/ml) for 30 minutes or indicated time. Cells were washed with 1×PBS and fixed with 10% neutral buffered formalin for 10 minutes followed by permeabilization with 0.1% Triton X-100 in 1×PBS for 10 minutes. The nuclei were counterstained with 4’,6-Diamidino-2-phenylindole (DAPI). The coverslips were mounted with an antifade mounting media and images were captured by confocal microscope (FV1000, Olympus). For live-cell imaging of mitochondrial network, cells were seeded in a glass-bottom 35-mm dish and stained with MitoTracker as above. Cells were focused under the microscope and then treated with IL-1β. Images were captured at a 1-minute interval.

Statistical analyses.

The sample size for the in vitro experiments was calculated for a two-fold difference between groups with a standard deviation of 0.5 and 80% power. The sample size for the in vivo experiment was calculated using the following parameters: type-1 error probability =5%, difference in mean OA score between groups =33%, within group standard deviation of the OA score =1, ratio of control to experimental mice =1, and power =80%. All the experiments were performed with cartilage or chondrocytes from n≥5 independent donor samples in duplicates. “n” represents the number of donors in human and mouse experiments and number of repeats in case of cell line experiments. All the statistical analyses were performed with GraphPad Prism software (version 9.2). Statistical significance between the two groups was determined by Student’s T-test for parametric data and the Mann-Whitney test for non-parametric data. Statistical significance between more than two groups was determined by one-way analysis of variance followed by Dunnett’s post hoc analysis. The result is plotted as mean ± 95% CI. A p value of ≤0.05 was considered significant.

Results

DRP1 is highly expressed in OA cartilage which correlated with excessive mitochondrial fragmentation and apoptosis in chondrocytes.

DRP1 is a mitochondrial fission protein that is primarily found in cytoplasm, but during stress DRP1 translocates to the dysfunctional mitochondria and causes the fission/fragmentation of the mitochondrial network. To study the role of DRP1 in the regulation of mitochondrial dynamics in chondrocytes, we analyzed the expression of DRP1 in human [n=6 normal (no known history of any rheumatic disease), n=9 low-grade OA cartilage (Manikin score 1–2) and n=6 high-grade OA cartilage (Makin score 3 and above)] and mouse (n=7 per group for immunofluorescence staining and n=3 per group for qPCR) OA cartilage. Compared to normal cartilage, expression of DRP1 was increased in human OA cartilage at protein level (Figure 1a, 1b and supplementary figure S1) and mRNA level Figure 1c) and mouse OA cartilage (Figure 1d, 1e, and 1f). There was no significant difference in the expression of DRP1 in low-grade and high-grade OA cartilage. We also determined the levels of mitofusin 1 (Mfn1) mRNA in the DMM joints. Interestingly, the expression of Mfn1 was significantly reduced in DMM knee joints (Figure 1f).

Figure 1. DRP1 expression is high in OA cartilage and is associated with excessive mitochondrial network fragmentation and apoptosis in OA chondrocytes.

(a) DRP1 protein expression in normal (n=6) and low-grade OA (n=9) and high-grade OA (n=6) human cartilage was determined by IHC staining. Scale bar 20 μm. (b) Quantification of DRP1 IHC staining. (c) DRP1 mRNA expression in normal (n=6) and OA (n=9) human cartilage was determined by qPCR. (d) Drp1 expression in control and DMM mouse knee joint (n=7 per group) was determined by IF staining. Scale bar 100 μm. (e) Quantification of Drp1 staining in control and DMM mouse knee joint. (f) Drp1 and Mfn1 mRNA expression levels in control and DMM knee joint cartilage (n=3) was determined by qPCR. (g) Mitochondrial network in normal and OA human chondrocytes (n=5 per group) was analyzed by MitoTracker staining and immunostaining with (h) SDHA or (i) SDHB. Scale bar 20 μm. (j) Quantification of chondrocytes with fragmented mitochondrial network. (k) Aspect ratio of mitochondria (ratio of major to minor axis). (l) TUNEL staining in mouse control and DMM knee sections (n=7 per group). Scale bar 50 μm. (m) Quantification of TUNEL in mouse knee. (n) TUNEL staining in human normal and OA cartilage (n=5). *p<0.05, **p<0.005.

Based on the above data, we hypothesized that high expression of DRP1 induces excessive mitochondrial network fragmentation in chondrocytes in diseased chondrocytes disrupting its homeostasis. To test our hypothesis, we analyzed the mitochondrial network integrity in chondrocytes prepared from normal and OA cartilage (n=5 per group). The mitochondrial network was intact in normal chondrocytes with a tubular appearance as determined by MitoTracker staining (Figure 1g), succinate dehydrogenase A (SDHA) (Figure 1h), and SDHB (Figure 1i) immunostaining but severely fragmented in OA chondrocytes (Figure 1g, 1h, 1i, and 1j) indicating dysregulated mitochondrial dynamics in diseased conditions. The mitochondrial aspect ratio (ratio of the major to the minor axis) showed a significant decrease in OA chondrocytes (1.5±0.1) compared to normal chondrocytes (2.15±0.2) suggestive of mitochondrial fragmentation (Figure 1k). These results indicate that increased DRP1 expression in OA cartilage is involved in excessive mitochondrial network fragmentation in chondrocytes in the disease condition. Mitochondrial damage is associated with increased apoptosis in chondrocytes and other cell types [11,12] and indeed we found increased mitochondrial damage and apoptosis in the joints of mice subjected to DMM surgery (Figure 1l and 1m) and human OA cartilage (Figure 1n) compared to normal cartilage.

Inflammatory cytokines induce DRP1 expression and excessive mitochondrial network fragmentation in chondrocytes.

Inflammation in OA joints plays a critical role in chondrocyte apoptosis and disease progression [13,14]. We treated primary human OA chondrocytes and mouse chondrocytes with IL-1β (5 ng/ml) to determine the role of IL-1β, a proinflammatory cytokine believed to play an important role in OA pathogenesis [18], in inducing mitochondrial network fission in chondrocytes. The DRP1 expression increased both at mRNA and protein levels in primary human (n=6) OA chondrocytes (Figure 2a and 2b) and mouse (n=3 for mRNA and n=6 for protein) chondrocytes (Figure 2c and 2d) upon IL-1β treatment. We determined the mitochondrial network fission by MitoTracker dye staining in primary human normal and OA chondrocytes (n=5 in each group) treated with IL-1β. IL-1β-induced mitochondrial fission in both normal (Figure 2e and 2f) and OA chondrocytes (Figure 2g and 2h). Live cell imaging showed mitochondrial fission at a single cell level in IL-1β treated primary human OA chondrocytes (Figure 2i and Supplementary video 1) in real time. Mouse chondrocytes treated with IL-1β and TNFα also showed increased mitochondrial network fission as determined by MitoTracker staining (Figure 2j and 2k) and Sdha staining (Figure 2l and 2m). Similarly, TC28 cells showed increased mitochondrial network fragmentation upon treatment with IL-1β and TNFα (10 ng/ml) as determined by MitoTracker staining (Figure 2n and 2o) and live cell imaging (Figure 2p and supplementary video 2). We also observed the formation of donut mitochondria (Figure 2p, green arrows). The mitochondria showed abnormal distribution in perinuclear areas in IL-1β treated normal, OA, and TC28 chondrocytes, compared to even distribution in untreated chondrocytes (Figure 2n and 2o). To avoid any bias due to the dye (MitoTracker) based assay, we transfected mouse chondrocytes and TC28 cells with mitoDsRed to visualize mitochondria and treated them with IL-1β and TNFα. As observed with MitoTracker staining, mitoDsRed also revealed fragmentation of mitochondrial network in mouse chondrocytes (Figure 2q and 2r) and TC28 cells (Figure 2s and 2t). We generated a stable HTB-94 cell line expressing mitoDsRed (HTB-mtDsRed cells) and determined mitochondrial network fragmentation upon IL-1β stimulation. HTB-mtDsRed cells showed mitochondrial network fragmentation and swelling upon treatment with IL-1β after 20 minutes (Figure 2u). DRP1 is a cytosolic protein that translocates to mitochondria and triggers mitochondrial network fragmentation. IL-1β and TNFα-induced the translocation of Drp1 in mouse chondrocytes as determined by colocalization of Drp1 with MitoTracker (Figure 3a) and Sdha (Figure 3b). The translocation of DRP1 to mitochondria and mitochondrial fragmentation upon IL-1β stimulation was also confirmed in HTB94-mtDsRed cells by IF staining (Figure 3c). These results suggest that inflammatory cytokines promote DRP1 expression which then translocates to mitochondria and induce mitochondrial network fragmentation in chondrocytes.

Figure 2. IL-1β and TNFα induced DRP1 expression and excessive mitochondrial network fragmentation in chondrocytes.

(a) Primary human OA chondrocytes (hCCs) (n=6) were treated with IL-1β (5 ng/ml) for 24 h and DRP1 mRNA expression was determined by qPCR and (b) immunoblotting. (c) The expression of Drp1 in mouse primary chondrocytes (mCCs) (n=5) treated with IL-1β (5 ng/ml) for 24 h was determined by qPCR and (d) immunoblotting (n=6). (e) and (f) Normal human chondrocytes (n=5) were treated with IL-1β for 30 minutes and the mitochondria were visualized by MitoTracker staining. Scale bar 20 μm. (g) and (h) Mitochondrial network fragmentation in OA human chondrocytes (n=5) treated with IL-1β was determined by MitoTracker staining. Mitochondrial donuts are marked by green arrow. Scale bar 10 μm. (i) Human OA chondrocytes were stained with MitoTracker and treated with IL-1β for real time live cell imaging. Scale bar 20 μm. (j) and (k) Primary mouse chondrocytes (n=5) were treated with IL-1β (5 ng/ml) or TNFα (10 ng/ml) for 30 minutes and mitochondria were visualized by MitoTracker staining. Scale bar 10 μm. (l) and (m) Mitochondrial fragmentation in primary mouse chondrocytes (n=5) was determined by Sdha IF staining. Scale bar 10 μm. (n) and (o) Mitochondrial network fragmentation in TC28 cells treated with IL-1β or TNFα was determined by MitoTracker staining. Scale bar 20 μm. (p) TC28 cells were stained with MitoTracker and treated with IL-1β for live cell imaging. Donut mitochondria are marked by green arrow. Scale bar 10 μm. (q) and (r) Mouse primary chondrocytes (n=5) were transfected with mitoDsRed2 plasmid to visualize mitochondria and treated with IL-1β or TNFα. Scale bar 10 μm. (s) and (t) TC28 cells were transfected with mitoDsRed2 plasmid to visualize mitochondria and treated with IL-1β or TNFα. Scale bar 20 μm. (u) HTB94 cells stably expressing mitoDsred2 (HTB-mtDsRed) were treated with IL-1β and mitochondrial network fragmentation was determined by real time live cell imaging. Scale bar 20 μm. *p<0.05, **p<0.005.

Figure 3. DRP1 is translocated to mitochondria under pathological conditions.

(a) Mitochondrial translocation of Drp1 in mouse chondrocytes treated with IL-1β and TNFα was determined by IF staining. Mitochondria were visualized by MitoTracker (MT) staining. Scale bar 20 μm. (b) Mitochondrial translocation of Drp1 in primary mouse chondrocytes treated with IL-1β and TNFα was determined by IF staining. Mitochondria were visualized by SDHA staining. Scale bar 20 μm. (c) Translocation of DRP1 in HTB-mtDsRed cells upon IL-1β stimulation was determined by IF staining of DRP1. Scale bar 20 μm.

IL-1β and TNFα-induced mitochondrial network fragmentation is associated with mitochondrial depolarization, oxidative stress, and apoptosis in chondrocytes.

Excessive mitochondrial network fragmentation and alteration in mitochondrial dynamics may trigger mitochondrial dysfunction and apoptosis. IL-1β or TNFα increased the ROS levels in TC28 chondrocytes (n=5 in each group) as determined by DCFDA (Figure 4a), DHR123 (Figure 4b), and MitoSOX staining (Figure 4c). IL-1β and TNFα increased ROS levels also in mouse chondrocytes as determined by DCFDA (Figure 4d), DHR123 (Figure 4e), and MitoSOX staining followed by fluorimetry (Figure 4f) and confocal microscopy (Figure 4g). IL-1β and TNFα also increased mitochondrial ROS production in primary human OA chondrocytes as determined by MitoSOX staining (Figure 4h). IL-1β stimulation also decreased the levels of reduced glutathione (GSH), an indicator of oxidative stress (Figure 4i) in mouse chondrocytes. As shown previously in primary human OA chondrocytes [3,10,11], IL-1β and TNFα-induced mitochondrial membrane potential/ΔΨ loss in mouse chondrocytes (Figure 4j and 4k) and in TC28 cells (Supplementary Figure 2) as determined by JC-1 staining.

Figure 4. Inflammation induced mitochondrial dysfunction and oxidative stress in chondrocytes.

IL-1β and TNFα induced ROS production in TC28 cells (n=5 per group) was determined by (a) DCFDA, (b) DHR123 or (c) MitoSOX staining followed by fluorimetry. IL-1β and TNFα induced ROS production in mouse chondrocytes (n=5) was determined by (d) DCFDA, (e) DHR123 or (f) MitoSOX staining followed by fluorimetry or (g) fluorescence microscopy. Scale bar 50 μm. (h) Primary human OA chondrocytes (n=5) were treated with IL-1β and TNFα and stained with MitoSOX to determine mitochondrial superoxide production. Scale bar 50 μm. (i) GSH levels in mouse chondrocytes (n=5) treated with IL-1β. (j) MMP/ΔΨ in mouse chondrocytes (n=5) treated with IL-1β or TNFα was determined by JC-1 staining followed by fluorimetry or (k) fluorescence microscopy. Scale bar 5 μm. (l) LDH assay in culture supernatant of primary human OA chondrocytes (n=5) treated with IL-1β. (m) LDH assay in culture supernatant of primary human normal chondrocytes (n=5) treated with IL-1β. (n) LDH assay in primary mouse chondrocytes (n=5) treated with IL-1β and TNFα. (o) PI uptake assay in primary mouse chondrocytes (mCCs) (n=5) treated with IL-1β and TNFα as determined by flow cytometry and (p) fluorescence microscopy. Inset image shows DAPI alone. Scale bar 10 μm. (q) Mouse chondrocytes were treated with IL-1β and chondrocytes apoptosis was determined by Caspase 3/7 luciferase activity assay. (r) IL-1β induced translocation of Bax to mitochondria and (s) release of Cytochrome c from mitochondria in HTB-mtDsRed cells. Scale bar 20 μm.

Mitochondria are metabolically highly active organelle in the cell and home for several proapoptotic proteins and during stress may promote cell death. As shown previously [11,15], IL-1β treatment increased chondrocyte death in human OA (Figure 4l) and normal chondrocytes (Figure 4m). Mouse chondrocytes treated with IL-1β and TNFα also showed increased cell death (Figure 4n, 4o, and 4p). The IL-1β-induced cell death in chondrocytes through the induction of apoptosis was confirmed by caspase 3/7 luciferase reporter assay (Figure 4q). IL-1β-induced the translocation of Bax to mitochondria (Figure 4r) and caused Cytochrome c release (Figure 4s). These results indicate that increased levels of inflammatory cytokines promote mitochondrial dysfunction and apoptosis in chondrocytes in the diseased joint.

IL-1β-induces mitochondrial network fragmentation in chondrocytes in mouse femoral head cartilage explants.

To avoid any bias introduced due to monolayer culture of chondrocytes or due to lack of cartilage ECM or other changes introduced due to cell culture conditions, we used mouse femoral head cartilage to confirm the above results. Mouse femoral head cartilage (n=8) showed increased Drp1 expression upon IL-1β treatment (Figure 5a and 5b). We also stained the mouse femoral head (n=8) with MitoTracker and treated it with IL-1β. MitoTracker was able to penetrate the femoral head and stain the chondrocyte mitochondria which was visualized by confocal microscopy at different magnifications. Imaging of the femoral head was done with a confocal microscope at every 2 μm interval for up to 250–300 μm and the 3D structure was reconstructed (Figure 5c and Supplementary video 3). Mitochondria appeared as tubular structure in chondrocytes in cartilage as in chondrocytes monolayer culture in control samples and exposure with IL-1β triggered mitochondrial network fragmentation (Figure 5c and 5d) and decreased the mitochondrial aspect ratio (Figure 5e). The translocation of Drp1 to mitochondria in chondrocytes upon IL-1β stimulation was also confirmed in mouse femoral head cartilage sections (n=8) by immunostaining of Drp1 and Sdha (Figure 5f). Treatment of mouse femoral head cartilage (n=7) increased the green fluorescence of JC-1, an indicator of MMP/ΔΨ loss (Figure 5g). IL-1β also induced ROS levels in mouse femoral head cartilage (n=7) as determined by DCFDA staining (Figure 5h and 5i). IL-1β-induced oxidative damage in mouse femoral head cartilage (n=8) was determined by 3-nitrotyrosine antibody staining (Figure 5j and 5k) and DNA/RNA oxidative damage marker antibody (Figure 5l and 5m). We also determined IL-1β-induced cell death in mouse femoral head cartilage (n=8) by LDH release assay (Figure 5n) and cleaved caspase-3 immunostaining (Figure 5o and 5p). The apoptosis in femoral cartilage (n=8) was also confirmed by TUNEL staining (Figure 5q). These results show that inflammatory cytokines induce chondrocyte apoptosis in vitro as observed in vivo in the diseased condition in human and mouse OA cartilage (Figure 1).

Figure 5. IL-1β induced mitochondrial network fragmentation in chondrocytes in their native environment.

(a) The expression of Drp1 in mouse femoral head cartilage (n=8 per group) treated with IL-1β for 72 h was determined by IF staining. Scale bar 100 μm (b) Quantification of Drp1 expression in mouse femoral head cartilage treated with IL-1β. (c) and (d) Mouse femoral head cartilage (n=8 per group) was treated with IL-1β and stained with MitoTracker. Scale bar 100 μm in whole femoral head and 2 μm in 60x image. (e) Quantification of the mitochondrial aspect ratio. (f) The translocation of Drp1 to mitochondria was also determined in mouse femoral head cartilage (n=8 per group) sections by IF staining of Drp1 and SDHA. (g) JC-1 staining in mouse femoral head cartilage (n=7 per group) treated with IL-1β to determine MMP/ΔΨ. (h) and (i) DCFDA staining in mouse femoral head cartilage (n=7 per group) treated with IL-1β followed by confocal microscopy. Scale bar 100 μm. (j) and (k) 3-Nitrotyrosine IF staining in mouse femoral head cartilage (n=8 per group) section. Scale bar 100 μm. (l) and (m) DNA/RNA oxidative stress damage IF staining [nucleic acid (NA) ox dam] in mouse femoral head cartilage (n=8 per group) section. Scale bar 100 μm. *p<0.05, **p<0.005. (n) LDH assay in mouse femoral head explants (n=8 per group) treated with IL-1β. (o) Cleaved Caspase-3 immunostaining in femoral head cartilage sections (n=8 per group) treated with IL-1β. Scale bar 100 μm. (p) Quantification of cleaved Caspase-3 IF staining in femoral head sections (n=8 per group). (q) TUNEL staining in mouse femoral head cartilage section (n=8 per group). *p<0.05, **p<0.005. Scale bar 100 μm.

Pharmacological Inhibition of DRP1 suppressed IL-1β-induced mitochondrial network fragmentation and apoptosis in chondrocytes.

To investigate the signaling pathway associated with increased mitochondrial network fission in chondrocytes, we treated mouse chondrocytes (n=5 per group) with IL-1β for indicated times and determined the phosphorylation of Drp1^Ser-592 (the human equivalent of DRP1^Ser616) by immunoblotting. IL-1β stimulation increased Drp1^Ser-592 phosphorylation in mouse chondrocytes (Figure 6a and 6b). Drp1 phosphorylation was also increased in mouse femoral head cartilage (n=8 per group) upon IL-1β stimulation (Figure 6c and 6d). We also found increased Drp1 phosphorylation in mouse OA cartilage (n=7 per group) in vivo (Figure 6e and 6f). To confirm the role of DRP1 in mitochondrial network fragmentation in chondrocytes, we treated HTB-mtDsRed cells with IL-1β in the presence of Mdivi-1 or DMSO alone as a control. DRP1 inhibition blocked IL-1β-induced fragmentation of mitochondrial network in HTB-mtDsRed chondrocytes (Figure 6g and 6h). Inhibition of Drp1 blocked IL-1β-induced chondrocytes death (Figure 6i and 6j). In addition, Mdivi-1 blocked IL-1β induced translocation of Bax to mitochondria (Figure 6k) and release of Cytochrome c from mitochondria (Figure 6l). These results show that Drp1 inhibition blocked IL-1β-induced chondrocytes apoptosis. We also treated mouse femoral head cartilage (n=7 per group) with IL-1β in the presence of Mdivi-1 or DMSO (control) and determined mitochondrial network fragmentation. Mdivi-1 treatment inhibited IL-1β-induced mitochondrial network fragmentation in mouse femoral head chondrocytes [Figure 7a, 7b and supplementary video V4 (control), V5 (IL-1β) and V6 (Mdivi-1+IL-1β)]. Mdivi-1 treatment also suppressed IL-1β-induced cell death in femoral head cartilage (Figure 7c and 7d). We also determined cartilage degradation by measuring the release of soluble glycosaminoglycan (sGAG) in the culture supernatant of mouse femoral cartilage treated with IL-1β in the presence of Mdivi-1. IL-1β-induced release of sGAG was mitigated upon Drp1 inhibition (Figure 7e). Safranin O staining also revealed protection from IL-1β-induced aggrecan degradation upon Drp1 inhibition (Figure 7f). In addition, Mdivi-1 treatment also alleviated IL-1β-induced MMP/ΔΨ loss (Figure 7g) and ROS production (Figure 7h) in mouse chondrocytes (n=10 per group). siRNA-mediated depletion of Drp1 (Figure 7i) in mouse chondrocytes inhibited IL-1β-induced MMP/ΔΨ loss (Figure 7j) and improved the expression of Col2a1 (Figure 7k) and Mmp13 (Figure 7l). Drp1 knockdown also inhibited IL-1β-induced apoptosis in mouse chondrocytes (Figure 7m). These results show that DRP1-mediated excessive mitochondrial network fragmentation was involved in chondrocyte apoptosis and cartilage ECM degradation and inhibition of DRP1 mitigated these pathologies.

Figure 6. Inhibition of DRP1 alleviated IL-1β induced mitochondrial network fragmentation and chondrocytes apoptosis.

(a) Mouse chondrocytes (n=5) were treated with IL-1β for indicated time and the level of P-Drp1 was determined by immunoblotting. (b) Quantification of P-Drp1 blots. (c) and (d) Phosphorylation of Drp1 in mouse femoral head cartilage (n=8 per group) treated with IL-1β was determined by IF staining. Scale bar 200 μm. (e) and (f) IF staining of P-Drp1 in control and DMM mouse knee joints (n=7 per group). Scale bar 100 μm. (g) and (h) Mdivi-1 (10 μM) inhibited IL-1β induced mitochondrial network fragmentation in HTB-mtDsRed cells. Scale bar 20 μm. (i) LDH assay in primary mouse chondrocytes (n=8) treated with IL-1β in the presence of Mdivi-1. (j) PI uptake assay followed by flow cytometry of primary mouse chondrocytes (n=5) treated with IL-1β in the presence of Mdivi-1. (k) Mdivi-1 inhibited IL-1β induced translocation of Bax to mitochondria in HTB-mtDsRed cells. (l) Mdivi-1 inhibited IL-1β induced release of Cytochrome c from mitochondria in HTB-mtDsRed cells. Scale bar 20 μm.

Figure 7. Inhibition of DRP1 alleviated IL-1β induced mitochondrial network fragmentation and chondrocytes apoptosis in cartilage explant.

(a) and (b) IL-1β induced mitochondrial network fragmentation was blocked by Mdivi-1 in mouse femoral head cartilage (n=7 per group). Scale bar 100 μm. (c) LDH assay in the culture supernatant of mouse femoral head cartilage (n=7 per group) treated with IL-1β in the presence of Mdivi-1. (d) TUNEL assay in mouse femoral head cartilage (n=7 per group) treated with IL-1β in the presence of Mdivi-1. Scale bar 100 μm. (e) sGAG levels in the culture supernatant of mouse femoral head cartilage (n=7 per group) treated with IL-1β in the presence of Mdivi-1. (f) Safranin O staining of femoral head (n=7 per group) sections treated with IL-1β in the presence of Mdivi-1. Scale bar 100 μm. (g) MMP/ΔΨ in primary mouse chondrocytes (n=10 per group) treated with IL-1β in the presence of Mdivi-1. (h) ROS levels as determined by DCFDA staining in primary mouse chondrocytes (n=10 per group) treated with IL-1β in the presence of Mdivi-1. (i) siRNA-mediated depletion of Drp1 in mouse chondrocytes (n=3). MMP/ΔΨ (j), Col2a1 mRNA expression (k), Mmp13 mRNA expression (l) and TUNEL staining (m) in primary mouse chondrocytes depleted of Drp1 levels.

ERK1/2-mediated phosphorylation of DRP1 induces mitochondrial network fragmentation in chondrocytes.

Previous studies in other cell types have shown ERK1/2-mediated phosphorylation of DRP1 [16,17] leads to its activation and mitochondrial network fragmentation. To determine if ERK1/2 is involved in DRP1 activation in chondrocytes, we first determined the activation of ERK1/2 in IL-1β treated mouse chondrocytes (n=5 per group). ERK1/2 phosphorylation was increased upon IL-1β treatment (Figure 8a and 8b). The phosphorylation of ERK1/2 was also found increased in mouse (Figure 8c and 8d) and human OA cartilage (Figure 8e and 8f). Inhibition of ERK1/2 with PD98059 suppressed the Drp1^Ser-592 phosphorylation (Figure 8g) and mitochondrial fission (Figure 8h and 8i) and death in mouse chondrocytes (Figure 8j and 8k). HTBmtDsRed cells treated with two different ERK inhibitors, PD98059 and SCH772984 showed similar results (Figure 8l and 8m). Furthermore, this effect was not restricted to chondrocytes in monolayer as inhibition of ERK1/2 in femoral head cartilage explants by SCH772984 also blocked the IL-1β-induced mitochondrial network fragmentation (n=7 per group) (Figure 8n and 8o). ERK1/2 inhibition also blocked the translocation of Drp1 to the mitochondria (Figure 8p).

Figure 8. ERK1/2-mediated phosphorylation of Drp1 triggers mitochondrial fission.

(a) Mouse chondrocytes (n=5 per group) were treated with IL-1β for indicated time and the level of phospho-Erk1/2 was determined by immunoblotting. (b) Quantification of P-Erk1/2 blot. (c) and (d) IF staining of P-Erk1/2 in control and DMM mouse knee joints (n=7 per group). Scale bar 100 μm. (e) IHC staining of P-ERK1/2 in normal and OA human cartilage (n=5 in each group). Scale bar 20 μm. (f) Quantification of P-ERK1/2 IHC staining. (g) P-Drp1 level in mouse chondrocytes was determined in the presence of ERK inhibitor (PD98059). (h) and (i) Mitochondrial network fragmentation in primary mouse chondrocytes (n=5) treated with IL-1β in the presence of PD98059. Scale bar 10 μm. (j) LDH assay in primary mouse chondrocytes (n=5) treated with IL-1β in the presence of PD98059. (k) PI uptake assay in primary mouse chondrocytes treated with IL-1β in the presence of PD98059. (l) and (m) Mitochondrial network fragmentation in HTB-mtDsRed cells treated with IL-1β in the presence of PD98059 (PD98) and SCH772984 (SCH). Scale bar 20 μm. (n) and (o) Mitochondrial network fragmentation in mouse femoral head cartilage (n=7 per group) treated with IL-1β in the presence of SCH772984 (SCH). Scale bar 200 μm. (p) Translocation of Drp1 to mitochondria in primary mouse chondrocytes (n=5) treated with IL-1β in the presence of PD98059 (PD98). Scale bar 10 μm.

Discussion

Mitochondrial dysfunction is a central player in OA pathogenesis; however, the underlying mechanism remains elusive. Altered mitochondrial dynamics is a pathological feature of several human diseases including cancer, neurodegenerative, and bone diseases [16–20], but the role of mitochondrial dynamics in OA remains unknown. DRP1 is a GTPase primarily located in the cytosol, which translocates to mitochondria in response to stress and regulates mitochondrial dynamics in the cell [2,21]. In this study, we observed excessive mitochondrial network fragmentation in chondrocytes under pathological conditions which was associated with increased DRP1 expression, reduced Mfn1 expression, mitochondrial dysfunction, oxidative stress, apoptosis and cartilage ECM degradation. IL-1β and TNFα stimulation increased DRP1 expression, excessive mitochondrial network fragmentation, apoptosis and cartilage ECM degradation in chondrocytes and cartilage explants. Pharmacological inhibition and siRNA-mediated knockdown of DRP1 inhibited IL-1β-induced mitochondrial network fragmentation, mitochondrial depolarization, oxidative stress, and apoptosis. We found that ERK1/2-mediated phosphorylation of DRP1 is a key step in mitochondrial network fragmentation and apoptosis and that ERK1/2 inhibition by PD98059 and SCH772984 inhibited this. The maintenance of the articular cartilage ECM depends solely on the chondrocytes, thus protecting chondrocyte function and survival is crucial for overall joint health. Here we showed that inhibition of DRP1 improved mitochondrial function and suppressed chondrocytes apoptosis and cartilage ECM degradation.

Mitochondria are highly dynamic organelles that undergo continuous fission and fusion events to maintain mitochondrial quality. Understanding the factors involved in the regulation of mitochondrial fission and fusion is crucial as perturbations in these factors may result in a pathological state. Excessive mitochondrial fission has been implicated in many neurological diseases such as Parkinson’s and Alzheimer’s disease [18,22]. We found that the mitochondrial network remains intact in normal chondrocytes, however, large number of chondrocytes from OA cartilage display excessive mitochondrial network fragmentation. Excessive mitochondrial network fragmentation caused MMP/ΔΨ loss and increased mitochondrial ROS levels. We also observed increased mitochondrial network fragmentation in chondrocytes upon IL-1β or TNFα stimulation in monolayer culture and in mouse femoral head cartilage. In addition, we found mitochondrial swelling and altered distribution of mitochondria in OA chondrocytes and in chondrocytes treated with IL-1β and TNFα. These findings show that the mitochondrial dynamics and quality control is dysregulated in chondrocytes under pathological conditions. The formation of donut mitochondria has been reported in certain cells under stress [23]. We also observed the formation of donut mitochondria in chondrocytes treated with IL-1β, however, we do not yet know the implication of donut mitochondria in chondrocytes biology.

DRP1 is a key regulator of mitochondrial fission [1]. Here we show that increased expression of DRP1 in chondrocytes under diseased conditions is associated with excessive mitochondrial network fragmentation, oxidative stress, and apoptosis. We showed here that IL-1β increased the expression and mitochondrial translocation of DRP1 and promoted Bax-mediated Cytochrome c release. DRP1 has been shown to cause Bax oligomerization and Cytochrome c release from mitochondria and loss of DRP1 expression leads to reduced Cytochrome c release [24]. DRP1 was shown to translocate to mitochondria in COS7 cells and overexpression of dominant-negative DRP1-K38A mutant inhibited mitochondrial network fragmentation and apoptosis by blocking the release of Cytochrome c [25]. Mdivi-1 is a small molecule inhibitor of mitochondrial division by targeting DRP1 [26] and has been shown to inhibit apoptosis in the ischemic retina [27]. DRP1 inhibition by Mdivi-1 blocked IL-1β-induced mitochondrial network fragmentation, MMP/ΔΨ loss, and oxidative stress. Also, DRP1 inhibition blocked mitochondrial translocation of Bax and diminished Cytochrome c release and chondrocytes apoptosis. DRP1 inhibition also inhibited cartilage ECM degradation suggesting that DRP1-mediated mitochondrial dysfunction is associated with cartilage ECM degradation. Similarly, siRNA-mediated depletion of Drp1 inhibited IL-1β-induced MMP/ΔΨ loss, enhanced Col2a1 expression and reduced Mmp13 expression and chondrocyte apoptosis. These results show that excessive mitochondrial network fragmentation driven by DRP1 is a pathological feature in OA and is linked to oxidative stress, cartilage ECM degradation and chondrocytes apoptosis. These results also strengthen the idea that mitochondrial dysfunction is a key aspect of OA and that improving mitochondrial function and quality control could help reduce OA pathogenesis. The limitation of our study is the lack of Drp1 knockout mice to investigate the in vivo effect of Drp1 deletion on OA pathogenesis and DRP1 overexpression studies.

Phosphorylation-induced activation of DRP1 is required for its translocation to mitochondria and mitochondrial network fragmentation. Different studies have identified ERK [16,28,29] and other kinases [30] as an activator of DRP1. Here, we found increased activation of DRP1 was mediated by ERK1/2. We also showed here that inhibition of ERK1/2 using PD98059 and SCH772984 suppressed IL-1β-induced mitochondrial network fragmentation and chondrocytes apoptosis. Similarly, inhibition of ERK1/2 by PD98059 suppressed mitochondrial network fragmentation in SHSY-5Y neuronal cells [31]. ERK1/2 pathway promotes mitochondrial fission in neuronal cells and mouse embryonic fibroblasts [32]. Inhibition of ERK1/2 suppressed mitochondrial fission and ROS production in rat liver cells and cardiac myoblast cells [33]. Others have shown that mitochondrial fission is upstream to ROS production and its inhibition suppressed ROS and apoptosis [34,35]. ERK1/2 pathway is activated in OA cartilage compared to normal cartilage [36–38]. Oral administration of ERK inhibitor suppressed OA severity in a dog [39] and a rabbit model of OA [40]. Previous studies by others and us have shown the activation of ERK1/2 in primary human chondrocytes upon IL-1β stimulation [3,41].

In conclusion, we showed that chondrocytes in diseased conditions show excessive fragmentation of mitochondrial network. Altered mitochondrial fission leads to impaired mitochondrial quality control and mitochondrial dysfunction in OA chondrocytes. We showed that DRP1-mediated mitochondrial network fragmentation is linked to the increase in chondrocyte apoptosis and cartilage ECM degradation observed in the diseased cartilage. We also show that ERK1/2 is activated in diseased chondrocytes in vivo and in vitro in response to IL-1β treatment and that it is essential for DRP1 activation and translocation to mitochondria. These findings suggest that ERK1/2-mediated DRP1 activation is a critical step in mitochondrial network fragmentation and chondrocytes apoptosis, and it could be targeted for OA therapy.