Mitochondrial biogenesis is impaired in osteoarthritic chondrocytes but reversible via peroxisome proliferator-activated receptor-γ coactivator 1α

Reagents

All chemical reagents were obtained from Sigma-Aldrich (St Louis, MO), unless otherwise indicated. Selective AMPK pharmacologic activator A-769662 was from LC laboratories (Woburn, MA). SRT-1720 (a selective SIRT1 activator) and EX527 (a selective SIRT1 inhibitor) were from Cayman Chemical (Ann Arbor, MI). Antibodies to phospho-AMPKα (Thr172), total AMPKα, AMPKα1, and NRF1 were from Cell Signaling Technology, Inc. (Danvers, MA). Antibodies to SIRT1 (C-terminal) that recognize full-length SIRT1, PGC-1α, NRF2 were from Abcam (Cambridge, MA). SIRT1 (N-terminal) antibody that recognize both full-length and truncated from of SIRT1 (75 KDa) was from EMD Millipore (Billerica, MA). Antibodies for immunoprecipitation of SIRT1 and PGC-1α were from Abcam (Cambridge, MA) and Santa Cruz Biotechnology (Santa Cruz, CA), respectively. Anti-human total OXPHOS complex kit was from Life Technologies (Carlsbad, CA), as were human SIRT1 siRNA, PGC-1α siRNA, TFAM siRNA and control siRNA.

Studies of human and mouse articular chondrocytes

Studies were performed in compliance with institutional IRB reviewed and approved human subjects protocols at the Scripps Research Institute and VA Medical Center. Human knee chondrocytes were isolated from autopsy donors that were graded macroscopically according to a modified Outerbridge scale (32). Grade I represents intact cartilage surface (normal); grade II represents minimal fibrillation (OA); grade III and IV represents overt fibrillation (OA). Human chondrocytes were cultured in Dulbecco's modified Eagle's (DMEM) high glucose medium with 10% fetal calf serum (FCS), 100 μg/ml streptomycin, and 100 IU/ml penicillin at 37°C, and no later than first passage chondrocytes were used for all experiments. Mouse chondrocytes were isolated from femoral head cartilage of 6-8 weeks old of mice. Unless otherwise indicated, chondrocytes were plated at 2.5 × 10⁵ cells per well in 12 well plates or 4 × 10⁵ cells per well in 6-well plates.

SDS-PAGE/Western Blot

Cells were lysed in RIPA buffer with 2 mM sodium vanadate and protease inhibitor cocktails (Roche, Mannheim, Germany). Cell lysates (10-15 μg) were separated by gradient 4-20% SDS-PAGE and transferred onto nitrocellulose membranes (Bio-Rad, Hercules, CA), probed with antibodies, exposed to SuperSignal West Pico Chemiluminescent Substrate (Thermo Scientific, Waltham, MA), and visualized by autoradiography.

Immunohistochemistry (IHC)

Slides of human knee cartilage sections from both normal and advanced OA donors, as well as slides of mouse knee sections from 12 and 24 month-old donors were firstly treated with 3% H₂O₂ for 10 min, then blocked with 10% goat serum for 2 hours at room temperature. After washing with TBS, rabbit antibodies to TFAM and UQCRC1 (ubiquinol-cytochrome c reductase core protein I), and the negative control rabbit IgG (1 μg/ml) were applied to the sections and incubated overnight at 4°C. Next, the sections were washed with TBS, incubated with biotinylated goat anti-rabbit IgG secondary antibody for 30 min, and then incubated for 30 minutes using the Histostain Plus kit (Invitrogen, Carlsbad, CA). Finally, the sections were washed and incubated with 3,3′-diaminobenzidine (DAB) substrate for 2-5 min.

Measurement of NAD⁺/NADH, oxygen consumption, and ATP and nitric oxide

NAD⁺/NADH ratio, oxygen consumption as a gauge for OXPHOS (presented as oxygen consumption rate calculated as a function of time (picomoles per minute)) and intracellular ATP were quantified from chondrocytes, using a NAD/NADH quantification colorimetric kit (BioVision), the XF extracellular flux analyzer (Seahorse Bioscience), and an ATP colorimetric assay kit (BioVision), respectively, according to manufacturer instructions. Nitric oxide (NO) release was measured as the concentration of nitrite in conditioned media by the Griess reaction, using NaNO₂ as standard.

Knockdown of SIRT1 and TFAM and overexpression of SIRT1 and PGC-1α in human knee articular chondrocytes

Cultured human knee articular chondrocytes (passage 1) were transfected with SIRT1 siRNA, TFAM siRNA, and the non-target control using the X-tremeGene siRNA transfection reagent (Roche Applied Science), or plasmids of pcDNA4-PGC-1α (Addgene, #10974) and pcDNA4 vector controls, respectively, using the Lipofectamine 3000 transfection reagent (Life Technology), according to manufacturer protocols. Levels of expression of SIRT1, PGC-1α and TFAM were examined by SDS-PAGE/Western blot analysis.

SIRT1 activity assay

SIRT1 was immuoprecipitated from protein lysates of chondrocytes under the conditions indicated with SIRT1 antibody (Abcam) using the Pierce Classic IP kit (Thermo Scientific). The SIRT1 immunoprecipitates were then subjected to SIRT1 enzymatic activity assay using a fluorescence-based SIRT1 activity kit (Cayman Chemical) according to manufacturer protocol.

Examination of acetylation of PGC-1α

PGC-1α was immuoprecipitated from cell lysates using the Pierce Classic IP kit (Thermo Scientific). The PGC-1α immunoprecipitates were then subjected to SDS-PAGE/Western blot analysis using an antibody to acetyl-lysine (Cell Signaling).

Assessment of chondrocyte mitochondrial biogenesis

Statistical analyses

All data were expressed as mean ± SD. Statistical analyses were performed by either student's t-test or 2-way ANOVA with Bonferroni post-hoc test using GraphPad PRISM 6. P values less than 0.05 were considered significant.

AMPK regulated PGC-1α expression via SIRT1 in articular chondrocytes

The selective AMPK pharmacological activator A-769662 induced elevated NAD⁺/NADH ratio, associated with increased expression of NAMPT, an enzyme in the NAD+ salvage pathway, and increase in both expression and activity of SIRT1 (Figure 1A) in human knee chondrocytes. In contrast, SIRT1 expression was almost diminished in AMPKα1 KO mouse chondrocytes (Figure 1B). Although AMPKα1 is the predominant AMPKα isoform expressed in chondrocytes, AMPKα2 KO chondrocytes exhibited impaired SIRT1 expression as well (data not shown). Expression of PGC-1α, but not phosphorylation and expression of AMPKα, was reduced in the SIRT1 knockdown chondrocytes (Figure 1C). Since PGC-1α expression was decreased in both AMPKα1 KO and AMPKα2 KO mouse chondrocytes (21), AMPK appeared to regulate expression of PGC-1α via SIRT1.

Stimulation of human chondrocytes with a selective SIRT1 activator SRT-1720 (5 μM) or a selective SIRT1 inhibitor EX527 (3 μM) resulted in significantly increased or decreased SIRT1 activity, respectively (Figure 1D), despite that SIRT1 expression was not affected (Figure 1E). In addition, either SIRT1 activator or inhibitor had any effect on phosphorylation and expression of AMPKα (Figure 1 E), which further confirmed that SIRT1 is downstream of AMPK. Moreover, SRT-1720 almost completely inhibited acetylation of PGC-1α and strongly enhanced expression of PGC-1α (Figure 1E). Opposite results were observed with EX527 (Figure 1E). Taken together, AMPK regulated PGC-1α acetylation and expression through promoting SIRT1 activity in chondrocytes.

Reduced mitochondrial biogenesis capacity and function in human knee OA chondrocytes

OA chondrocytes had decreased basal levels of phosphorylation of AMPKα and expression of SIRT1 and PGC-1α compared to normal donor chondrocytes, and associated with increased acetylation of PGC-1α (Figure 2A). In parallel, OA chondrocytes demonstrated decreased protein expression of the mitochondrial biogenesis mediators NRF1, NRF2, TFAM (Figure 2A), and reduced mitochondrial DNA content and mitochondrial mass, indicated by almost 40% decrease in ratio of cytochrome C oxidase I or II (COXI or COXII, mitochondrial-encoded gene) and 18S rDNA (nuclear gene) and less Mitotracker Green FM staining, respectively (Figure 2B). There was decreased expression of mitochondrial respiratory complexes I to IV and ATP synthase (complex V) in OA chondrocytes (Figure 2C). In function analyses, basal levels of oxygen consumption and intracellular ATP level in OA chondrocytes were decreased by 34% and 21%, respectively, compared to normal chondrocytes (Figure 2D). Therefore, human knee OA chondrocytes had reduced capacity for mitochondrial biogenesis with associated broadly impaired mitochondrial functions.

Because inflammation contributes to OA progression, we tested the effect of IL-1β on expression of key players involved in mitochondrial biogenesis and mitochondrial DNA content in chondrocytes. We confirmed (25) that phosphorylation of AMPKα was decreased by IL-1β in normal human knee chondrocytes (passage 1) (Supplemental Figure 1A). Under these conditions, expression of full-length SIRT1 was reduced, but expression of the 75 KDa truncated SIRT1 (inactive variant) was markedly induced by IL-1β (Supplemental Figure 1A). This was associated with decreased expression of PGC-1α, increased acetylation of PGC-1α, decreased expression of TFAM, NRF1 and NRF2 (Supplemental Figure 1A), and significantly decreased mitochondrial DNA content (Supplemental Figure 1B). These results indicated that IL-1β can impair mitochondrial biogenesis capability in chondrocytes. Interestingly, activation of AMPK by A-769662, which is previously shown to attenuate chondrocyte catabolic responses to IL-1β (25), inhibited expression of the 75 KDa truncated SIRT1 and prevented decreased expression of full-length SIRT1 in response to IL-1β in chondrocytes (Supplemental Figure 1C), suggesting that activation of AMPK could potentially preserve chondrocyte mitochondrial biogenesis capacity under inflammatory condition.

Reduced TFAM expression in both human and mouse knee OA cartilages in situ, and functional implications of cultured chondrocyte TFAM deficiency in vitro

Since TFAM is a key activator of mitochondrial transcription, we examined in situ expression of TFAM, and noted it to be decreased by 47.6% (percentage of cells stained positively with TFAM relative to cells stained with hematoxylin which indicates cellularity) in human knee OA cartilage particularly in the clustered chondrocytes (black arrows), compared to normal control, and associated with 47.4% decrease in expression of UQCRC1, a subunit of mitochondrial respiratory complex III (Figure 3A and B). Similar results were seen in mouse knee OA cartilage as well (Supplemental Figure 2). We also observed significantly reduced in situ expression of TFAM in knee cartilages from AMPKα1 KO mice, compared with congenic WT mice (Supplemental Figure 3).

To specifically examine effects of aging on in situ expression of TFAM and UQCRC1 in mouse knee cartilages, we selected 12 and 24 months old mouse knee (tibial plateau) sections that exhibited no obvious tissue damage and no significant loss of proteoglycan content in cartilage confirmed by safranin-o staining (Figure 3C). Twenty-four months old mouse tibial knee cartilage showed mild OA with an OARSI (Osteoarthritis Research Society International) score of 1, compared with 12 months old. Hematoxylin staining indicated loss of cellularity in 24 months, compared to 12 months old mouse tibial knee cartilages (Figure 3C). However, after normalization with the cells stained with hematoxylin, expression of TFAM and UQCRC1 was still significantly decreased in 24 months (decreased by 32.5% and 41.5%, respectively), compared with 12 months old mouse knee tibial cartilages (Figure 3D). Notably, in the enlarged boxed areas (on the left for 12 months, and on the right for 24 months) that were from the same region of the tibial cartilages, few cells stained positively for either TFAM or UQCRC1 despite good cellularity in the area in 24 months old tibial cartilages (Figure 3C).

TFAM siRNA knockdown decreased mitochondrial DNA content in cultured normal human chondrocytes, as indicated by ratio of COXI or COXII to 18S rDNA, and by mitochondrial mass, assessed by Mitotracker Green FM staining (Figure 4A). Despite low basal level of TFAM expression in human knee OA chondrocytes (Figure 2A), similar extent of decrease in mitochondrial DNA content and mitochondrial mass was observed in OA chondrocytes transfected with TFAM siRNA (data not shown). Accordingly, transfection with TFAM siRNA decreased expression of mitochondrial respiratory complexes I to IV and ATP synthase (complex V) in normal chondrocytes, and nearly attenuated their expression in OA chondrocytes (Figure 4B).

Chondrocytes with TFAM-mediated reduction in mitochondrial biogenesis potential mounted increased catabolic responses to IL-1β. Specifically, TFAM expression was lower (Figure 4B), and IL-1β-induced release of NO (Figure 4E) and MMP-3 and MMP-13 (Figure 4E) were higher in OA chondrocytes compared with normal chondrocytes transfected with control siRNA; these effects were greatly amplified in chondrocytes transfected with TFAM siRNA (Figure 4E).

Overexpression of PGC-1α ameliorates mitochondrial biogenesis in human knee OA chondrocytes

Overexpression of the mitochondrial biogenesis master PGC-1α by transfection in human OA chondrocytes led to enhanced expression of NRF1, NRF2 and TFAM (Supplemental Figure 4). In addition, mitochondrial DNA content, indicated by the ratio of COXI/18S rDNA and COXII/18S rDNA, was increased by 2.15 and 1.9 fold, respectively, in PGC-1α overexpressing chondrocytes (Supplemental Figure 4), associated with increased mitochondrial mass showed by amplified intensity of Mitotracker Green FM staining (Supplemental Figure 4), and augmented expression of mitochondrial respiratory complexes (particularly I, II, and III) (Supplemental Figure 4). Induction of PGC-1α expression by the SIRT1 activator STR-1720 (Figure 1E) also led to increased expression of NRF1, NRF2 and TFAM (Supplemental Figure 5A), and increased mitochondrial DNA content (Supplemental Figure 5B). Opposite results were seen with the SIRT1 inhibitor EX527 (Supplemental Figure 5).

AMPK promoted both improved mitochondrial biogenesis and function in OA chondrocytes and critically acts via PGC-1α

We tested if pharmacologic AMPK activation improves the established impairments in mitochondrial biogenesis and function in OA chondrocytes. As shown in Figure 5A, A-769662 increased phosphorylation of AMPKα and expression of SIRT1 (full-length) and PGC-1α, inhibited acetylation of PGC-1α, and increased expression of NRF1, NRF2, and TFAM. In addition, A-769662 increased mitochondrial DNA content by roughly 1.5 fold for both COXI/18S rDNA and COXII /18S rDNA, and augmented mitochondrial mass indicated by increased Mitotracker Green FM staining (Figure 5B), and enhanced expression of mitochondrial respiratory complexes (Figure 5C). Furthermore, A-769662 significantly increased oxygen consumption and intracellular ATP, reflecting improved functional OXPHOS (Figure 5D). Notably, A-769662 at 0.125 mM appeared sufficient to increase mitochondrial biogenesis potential in OA chondrocytes.

To further determine if A-769662 improved chondrocyte mitochondrial biogenesis and function through PGC-1α, we knocked down PGC-1α expression in chondrocytes via siRNA transfection, an approach we first validated to reduce expression of NRF1, NRF2 and TFAM (Figure 6A, lane 2), mitochondrial DNA content (Figure 6B), mitochondrial mass (Figure 6C), expression of mitochondrial respiratory complex I, II and III (Figure 6D), and intracellular ATP level (Figure 6E). AMPK activation using A-769662 had decreased capacity to increase each of these same mitochondrial readouts (Figure 6A-E) in PGC-1α knockdown chondrocytes. Our collective results on AMPK, SIRT1, and PGC-1α pathway in promoting chondrocyte mitochondrial biogenesis and function are summarized and schematically presented in Figure 6E.