Mito-TEMPO

LIFE SCIENCES

Molecular, Cellular and Functional Basic of Theraphy

Parkin and Nrf2 prevent oxidative stress-induced apoptosis in intervertebral endplate chondrocytes via inducing mitophagy and anti-oxidant defenses

Liang Kang, Shiwei Liu, Jingchao Li, Yueyang Tian, Yuan Xue, Xiaozhi Liu

To appear in: Life Sciences

Parkin and Nrf2 prevent oxidative stress-induced apoptosis in intervertebral endplate chondrocytes via inducing mitophagy and anti-oxidant defenses

Liang Kang 1,2,#, Shiwei Liu 1,2,#, Jingchao Li 1,2,3,#, Yueyang Tian 1,2, Yuan Xue 1,2,*, Xiaozhi Liu 4,*

1 Department of Orthopedics, Tianjin Medical University General Hospital, Tianjin 300052, China

2 Tianjin Key Laboratory of Spine and Spinal Cord Injury, Tianjin 300052, China
3 Department of Orthopedics, Tianjin Jinghai District Hospital, Tianjin 301600, China
4 Central Laboratory, The Fifth Central Hospital of Tianjin, Tianjin 300450, China

Correspondence to:
Department of Orthopedics, Tianjin Medical University General Hospital, No. 154 Anshan Road, Heping District, Tianjin 300052, China. E-mail address: [email protected] (Yuan Xue)
Central Laboratory, The Fifth Central Hospital of Tianjin, No. 41 Zhejiang Road, Binhai New District, Tianjin 300450, China. E-mail address: [email protected] (Xiaozhi Liu)# Liang Kang, Shiwei Liu, and Jingchao Li contributed equally to this work.
Abbreviations: IDD, intervertebral disc degeneration; IVD, intervertebral disc; CEP, cartilaginous endplate; ROS, reactive oxygen species; mPTP, mitochondrial permeability transition pore; ΔΨm, mitochondrial membrane potential; Nrf2, nuclear factor E2-related factor 2; SOD-2, superoxide dismutase-2; HO-1, heme oxygenase-1; NQO-1, NAD(P)H quinone oxidoreductase-1; Drp1, dynamin-related protein 1; Mfn2, mitofusin 2; CQ, chloroquine; LAMP1, lysosomal associated membrane protein 1; NAC, N-acetyl-L-cysteine; TUNEL, terminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling.

Graphical Abstract
Abstract
Aims: Endplate chondrocyte apoptosis is an important contributor to the pathogenesis of cartilaginous endplate (CEP) degeneration that leads to the initiation and development of intervertebral disc degeneration (IDD). In this study, we hypothesized that Parkin-mediated mitophagy and nuclear factor E2-related factor 2 (Nrf2)-mediated antioxidant system played an important role in endplate chondrocyte survival under pathological conditions.
Materials and Methods: Human endplate chondrocytes were stimulated with H2O2 to mimic pathological conditions. Western blotting, immunofluorescence staining, and flow cytometry were applied to detect the indicators related to mitochondrial dynamics, mitophagy, Nrf2 signaling, and apoptosis. The puncture-induced rat models were established to evaluate the changes in vivo.
Key findings: Our results showed that H2O2 induced oxidative stress, mitochondrial dysfunction, and apoptosis in endplate chondrocytes. These H2O2-induced detrimental effects were inhibited by pretreatment with the mitochondria-targeted antioxidant Mito-TEMPO. In addition, mitochondrial dynamics, Parkin-mediated elimination of dysfunctional mitochondria, and Nrf2-mediated antioxidant system were promoted by H2O2. Knockdown of Parkin or Nrf2 increased H2O2-induced detrimental effects. Moreover, upregulation of Parkin and Nrf2 by polydatin protected endplate chondrocytes against H2O2-induced mitochondrial dysfunction, oxidative stress, and apoptosis. Finally, puncture-induced rat models showed that polydatin exerted a protective effect on CEP and disc degeneration.
Significance: Targeting Parkin and Nrf2 to improve mitochondrial homeostasis, redox balance and endplate chondrocyte survival may represent a potential therapeutic strategy for preventing IDD.

Keywords: Intervertebral disc degeneration, Endplate chondrocytes, Parkin, Nrf2, Mitochondrial dysfunction, Reactive oxygen species

1. Introduction
Low back pain (LBP) is the leading cause of disability, activity limitation, and lost productivity worldwide, with over 80% of adults suffering from LBP at some point in their lives [1, 2]. Intervertebral disc (IVD) degeneration (IDD) is widely acknowledged to be the major cause of LBP [3]. IVDs are fibro-cartilaginous tissues located between the vertebrae; these tissues function as shock absorbers by distributing the mechanical load along the spine and facilitating trunk mobility [4]. The cartilaginous endplate (CEP) is composed of a layer of hydrated biological tissue between the IVD and the adjacent vertebrae. The CEP is the main route through which nutrients reach IVDs from vertebral capillaries and through which waste products are eliminated from IVDs. CEP degeneration hinders the nutritional supply to IVDs, leading to IVD homeostasis imbalance and initiation of IDD [5]. Many studies have demonstrated that endplate chondrocyte apoptosis plays important roles in CEP degeneration and IDD [6-8]. Therefore, more investigation on endplate chondrocyte apoptosis and the target interventions may not only increase the pathogenetic knowledge of CEP degeneration and IDD, but also provide potential therapeutic strategy.
In some pathological conditions, when the cells cannot sufficiently eliminate
excessive reactive oxygen species (ROS) to maintain ROS at normal levels, oxidative stress may occur and lead to cytotoxicity [9]. High levels of oxidative stress have been detected in degenerated CEPs, suggesting the involvement of oxidative stress in CEP degeneration [10, 11]. Various pathological factors related to IDD, such as mechanical stress and high glucose, can induce oxidative stress and subsequent mitochondrial dysfunction in endplate chondrocytes [12, 13]. Oxidative stress-induced mitochondrial dysfunction in endplate chondrocytes is characterized by mitochondrial membrane depolarization, mitochondrial ROS overproduction, and increased release of pro-apoptotic proteins into the cytoplasm, which are all important contributors to endplate chondrocyte apoptosis under pathological conditions [14, 15]. Furthermore, disruption of mitochondria by ROS-induced damage leads to mitochondrial dysfunction, further enhancing the production of ROS. This

feed-forward vicious cycle between mitochondria and ROS causes sustained oxidative damage and ultimately cell death [11]. Thus, appropriate mitochondrial quality control is essential for endplate chondrocyte survival and normal CEP function.
Autophagy plays critical roles in maintaining cellular homeostasis by degrading and recycling intracellular damaged organelles and proteins in response to increased metabolic requirements of cells or to environmental stressors, such as starvation, hypoxia, and oxidative stress [16]. Recent evidence has shown that activating autophagy protects endplate chondrocytes against mitochondria-dependent apoptosis, suggesting that the selective autophagy process, called mitophagy, may function to clear damaged mitochondria in endplate chondrocytes under pathological conditions [15]. One protein that functions in the removal of damaged mitochondria via mitophagy is the E3 ubiquitin ligase Parkin. Parkin operates in conjunction with PTEN-induced putative kinase 1 and responds to the loss of mitochondrial membrane potential, initiating the clearance of damaged mitochondria [17, 18]. A recent study showed that Parkin-mediated removal of damaged mitochondria is critical for nucleus pulposus cell survival [19]. However, the effects of Parkin-mediated mitophagy on endplate chondrocyte survival remain largely unknown.
Inhibition or inadequate activation of antioxidant systems is related to the
occurrence of oxidative stress induced by excessive production of ROS [20]. Nuclear factor E2-related factor 2 (Nrf2) is a key redox-sensitive transcription factor that confer adaptive protection against oxidative stress by activating the expression of a number of cytoprotective genes [21-23]. In a mouse model of IDD, degenerative changes in IVDs in Nrf2-knockout mice were more severe than those in wild-type mice [24]. Furthermore, some phytochemicals extracted from medicinal plants have been shown to protect IVD cells from mitochondrial dysfunction and apoptosis via activation of Nrf2 signaling [25-27]. In addition, Nrf2 has been found to modulate mitochondrial function and metabolism [28]. However, despite these findings, few studies have explored the roles of Nrf2 signaling in CEP degeneration.
In the present study, we explored the role of Parkin-mediated mitophagy and Nrf2-mediated antioxidant system in endplate chondrocytes under oxidative stress.

We aimed to provide insights into the role of Parkin and Nrf2 pathway in CEP degeneration pathogenesis and to thereby contribute to the development of effective therapeutic strategies for inhibition of IDD progression.

2. Materials and methods
2.1. Tissue specimens and cell culture
This study was approved by the Ethics Committee of Tianjin Medical University General Hospital, and informed consent was obtained from each donor. CEP specimens were collected from patients who received discectomy and spinal fusion surgery for IDD at Tianjin Medical University General Hospital. Detailed information for each patient is listed in Table 1. Endplate chondrocytes isolation and culture were carried out as previously described [29]. Briefly, CEP tissues were minced into small pieces and washed three times with phosphate-buffered saline (PBS, Gibco, Grand Island, NY, USA) before digestion with 0.25% type II collagenase (Invitrogen) for 4 h at 37 °C. After washing in PBS and resuspension, human endplate chondrocytes were cultured in Dulbecco’s modified Eagle’s medium (DMEM)/F12 (Gibco) supplemented with 15% fetal bovine serum (FBS, Gibco) and 1% penicillin-streptomycin (Invitrogen) in 5% CO2 at 37 °C. The culture medium was changed every other day. The second passage of endplate chondrocytes was used in our study.

2.2. Short interfering (si) RNA transfection
SiRNAs against Parkin (si-Parkin), Beclin-1 (si-Beclin-1), and Nrf2 (si-Nrf2), and scrambled siRNA (si-Control) were designed and synthesized by RiboBio (Guangzhou, China). Endplate chondrocytes were transfected for 48 h with 100 nM of each siRNA using Lipofectamine 2000 (Invitrogen) according to the manufacturer’s instructions. After transfection, endplate chondrocytes were further treated as described in the Figure legends, and used for subsequent analysis.

2.3. Flow cytometry
Apoptosis levels, mitochondrial membrane potential (ΔΨm) changes, intracellular

total ROS production, and mitochondrial ROS production in endplate chondrocytes from each treatment group were assessed using Annexin V-APC/7-AAD Apoptosis Detection Kit (KeyGen Biotech, Nanjing, China), JC-1 Assay Kit (Beyotime, Shanghai, China), DCFH-DA (Beyotime), and MitoSOX Red (Invitrogen), respectively, as previously described [30]. After labeling, samples were examined using a FACSCalibur flow cytometer (BD Biosciences, USA).

2.4. Western blotting
Protein samples were separated using SDS-PAGE and transferred to PVDF membranes (Millipore, USA). The primary antibodies were diluted from 1:500-1:1000. Primary antibodies against the following proteins were used: Cytochrome c, Parkin, p62, Nrf2, Dynamin-related protein 1 (Drp1) (Abcam), Cleaved caspase-3, LC3, Beclin-1, Mitofusin 2 (Mfn2), GAPDH (Cell Signaling Technology), Superoxide dismutase-2 (SOD-2), Heme oxygenase-1 (HO-1), NAD(P)H quinone oxidoreductase-1 (NQO-1), COX IV, and Lamin B (Proteintech).

2.5. Immunofluorescence
Immunofluorescence staining of endplate chondrocytes was conducted as described previously [31]. Briefly, samples were blocked with 3% bovine serum albumin at room temperature for 30 min followed by incubation with primary antibodies against Tom20 (mitochondrial antibody), Drp1, Parkin, LC3, LAMP1 (lysosomal associated membrane protein 1, lysosomal antibody), and Nrf2 at 4°C overnight. Then, samples were washed with PBS and incubated with appropriate secondary antibodies at room temperature for 1 h. Nuclei were stained with DAPI (ZSGB-BIO). Fluorescence images were acquired using a laser scanning confocal microscope (FV1000, Olympus, Japan). The mitochondrial and lysosomal antibodies were used to mark mitochondria and lysosomes, respectively.

2.6. Measurement of ATP content and mitochondrial permeability transition pore (mPTP) opening

The ATP production and mPTP opening in endplate chondrocytes were measured using ATP Assay Kit (Beyotime) and mPTP Assay Kit (Genmed, Shanghai, China), respectively, as per manufacturer instructions.

2.7. mRFP-GFP-LC3 assay
Human endplate chondrocytes were infected with mRFP-GFP-LC3 adenoviral vectors (HanBio Technology, Shanghai, China) to evaluate the effect of H2O2 on autophagic flux. The principle of the assay is based on difference in pH stability of the green and red fluorescent proteins as previously described [32]. The fluorescence signal of GFP is quenched under the acidic condition inside the lysosome, but the mRFP fluorescence signal does not change markedly under the acidic condition, so the autophagosomes appear as yellow puncta (RFP+GFP+) and the autolysosomes appear as red puncta (RFP+GFP−). The RFP and GFP puncta in each treatment group were detected using a laser scanning confocal microscope (FV1000, Olympus, Japan).

2.8. Animal model establishment and assessments
The animal experiments were approved by the Animal Care and Use Committee of Tianjin Medical University. Sprague-Dawley rats (3 months old) were randomly divided into three groups: Control group, IDD group, and Polydatin (PD) group. The rats in the IDD and PD groups underwent puncture-induced IDD surgery, which was performed using a needle (27G) on the experimental level rat tail disc (Co7/8) according to the surgical methods reported by Jiang et al [33]. After surgery, the PD group was treated with 50 mg/kg PD by intragastric administration once per day, and the IDD group was administered an equal amount of saline. Daily administration started on the day of operation, and continued until the rats were sacrificed at 4 weeks after operation. Daily monitoring of the rats was carried out to ensure their well-being, and all animals were allowed free unrestricted weight bearing and activity.
Magnetic resonance imaging (MRI) was performed on all rats by using a BioSpec MRI (Bruker, 7.0 T/20 cm) at 4 weeks after operation. The parameters of T2-weighted imaging were those referred to in a previous study [30]. All MR images were

analyzed using the IDD classification by Pfirrmann et al [34].
After MRI examination, disc tissues were harvested for further analyses. The specimens were fixed in formaldehyde, decalcified, dehydrated, embedded in paraffin, and sectioned at a thickness of 5 μm. Sections were stained with hematoxylin-eosin (HE) and safranin O-fast green (SO), and the histological scores were calculated based on the method previously described [35] to quantify the histological results. Immunohistochemistry was performed as previously described [33]. For immunohistochemical analysis, sections were incubated with primary antibodies against Parkin and Nrf2 at 4 °C overnight, after which the sections were incubated with appropriate HRP-conjugated secondary antibodies and counterstained with hematoxylin. Images were captured under a light microscope (Olympus). For terminal deoxynucleotidyl transferase (TdT)-mediated dUTP nick-end labeling (TUNEL) staining assay, the sections were handled using TUNEL Apoptosis Assay Kit (Beyotime), and samples were imaged with a fluorescence microscope (Olympus IX71).
Lastly, the Malondialdehyde (MDA) level and Hydrogen peroxide (H2O2) content
in the CEP tissues were measured using assay kits (Nanjing Jiancheng Bioengineering Institute) for MDA and H2O2, respectively, as per manufacturer’s instructions.

2.9. Statistical analysis
Data are presented as means ± standard deviation (SD) of at least three independent experiments and were analyzed using SPSS v.18.0 software (SPSS, Chicago, IL, USA). Differences between groups were evaluated using Student’s t-test or one-way analysis of variance (ANOVA) followed by Tukey’s test. P< 0.05 was considered statistically significant.

3. Results
3.1. H2O2 promoted ROS generation, mitochondrial dysfunction, and apoptosis in endplate chondrocytes
H2O2 was used to mimic oxidative stress in vitro. The results showed that
compared with control cells, intracellular and mitochondrial ROS production levels were significantly increased in H2O2-treated endplate chondrocytes (Fig. 1A-D). Next, the results showed that less calcein was retained in mitochondria from H2O2-treated endplate chondrocytes than in those from control cells, indicating increased mPTP opening (Fig. 1E). Additionally, flow cytometric analysis indicated that H2O2 treatment induced ΔΨm loss in endplate chondrocytes, as indicated by the decreased ratio of red to green (Fig. 1F-G). We found that H2O2 treatment reduced ATP production in endplate chondrocytes (Fig. 1H). Moreover, the translocation of mitochondrial cytochrome c to the cytoplasm and protein levels of cleaved caspase 3 were increased in endplate chondrocytes treated with H2O2, as indicated by western blotting (Fig. 1I-L). Annexin V-APC/7-AAD staining results showed that the percentage of apoptotic endplate chondrocytes was higher in the H2O2-treated group than in the control group (Fig. 1M-N). Taken together, these results demonstrated that endplate chondrocytes stimulated with H2O2 exhibited increased mitochondrial dysfunction, elevated ROS production, and enhanced apoptosis rates.
Subsequently, we specifically scavenged mitochondrial ROS with a
mitochondria-targeted antioxidant, Mito-TEMPO. As shown in Fig. 2A-D, H2O2-induced increases in mitochondrial and intracellular ROS levels were ameliorated by Mito-TEMPO pretreatment, indicating that mitochondrial ROS contributed to oxidative stress in H2O2-treated endplate chondrocytes. In addition, Mito-TEMPO pretreatment alleviated H2O2-induced prolonged mPTP opening and decreased ΔΨm and ATP production, indicating that H2O2-induced mitochondrial dysfunction was alleviated by Mito-TEMPO (Fig. 2E-H). Moreover, western blotting showed that Mito-TEMPO pretreatment suppressed the translocation of mitochondrial cytochrome c to the cytoplasm and blocked the upregulation of cleaved caspase 3 induced by H2O2 (Fig. 2I-L). Finally, flow cytometric analysis revealed that Mito-TEMPO pretreatment prevented the increase in apoptosis induced by H2O2 in endplate chondrocytes (Fig. 2M-N). These results showed that mitochondrial dysfunction played an important role in H2O2-induced endplate chondrocyte apoptosis.

3.2. Parkin-mediated mitophagy was activated in endplate chondrocytes treated with H2O2
Next, we exposed endplate chondrocytes to 200 μM H2O2 to detect Parkin expression under oxidative stress conditions. We found that Parkin protein levels were increased in H2O2-treated endplate chondrocytes compared with that in control cells (Fig. 3A-B). Because H2O2 exposure results in mitochondrial damage/depolarization in endplate chondrocytes, we investigated whether Parkin was translocated to damaged/depolarized mitochondria in H2O2-treated endplate chondrocytes. Western blotting revealed that H2O2 increased mitochondrial Parkin levels (Fig. 3C-D). Additionally, elevated colocalization of Parkin and Tom20 was observed in H2O2-treated cells compared with that in control cells (Fig. 3E-F).
Next, we evaluated engulfment of damaged mitochondria by LC3-decorated autophagosomes using immunofluorescence and western blotting. The immunofluorescence results showed that H2O2 exposure promoted colocalization of LC3 and Tom20 in endplate chondrocytes (Fig. 3G-H). Additionally, western blot analysis indicated that mitochondrial LC3-II protein levels were increased in H2O2-treated cells (Fig. 3I-J). Furthermore, colocalization of Tom20 and LAMP1 was increased in H2O2-treated cells compared with that in control cells, suggesting that dysfunctional mitochondria engulfed in mitophagosomes enter lysosomes in H2O2-treated endplate chondrocytes (Fig. 3K-L).
Finally, we investigated whether Parkin was essential for the activation of mitophagy in endplate chondrocytes upon exposure to H2O2. As shown in Fig. 3M-N, si-Parkin transfection effectively inhibited Parkin protein expression. Immunofluorescence analysis showed that knockdown of Parkin reduced the colocalization of Tom20 and LAMP1 in H2O2-treated endplate chondrocytes (Fig. 3K-L). Taken together, these data suggested that Parkin protein expression levels and mitochondrial translocation were increased in endplate chondrocytes under H2O2 treatment, resulting in promotion of mitophagy.
After mitochondria engulfment by autophagosomes, mitophagy merges into

autophagic flux. Undamaged autophagic flux is critical for selective removal of dysfunctional mitochondria. We measured the conversion of cytosolic, soluble LC3-I to autophagosomal membrane-attached, lipidated LC3-II and the protein expression of Beclin-1, a key autophagy inducer, in endplate chondrocytes under oxidative stress. Western blot results indicated that H2O2 treatment induced an increase in the LC3-II/I ratio and Beclin-1 expression level in endplate chondrocytes (Fig. 4A-D). To verify whether autophagic flux was enhanced upon H2O2 treatment, LC3-II levels were examined in endplate chondrocytes treated with the autophagic flux inhibitor chloroquine (CQ). Western blotting revealed that LC3-II levels were increased by CQ treatment and that this effect was further enhanced by H2O2 exposure, suggesting that increased LC3-II levels following H2O2 exposure were caused by enhanced autophagic flux (Fig. 4E-F). Furthermore, we transfected a tandem mRFP-GFP-LC3 construct into endplate chondrocytes to directly monitor changes in autophagic flux. Compared with the control group, the H2O2 treatment group showed increased yellow puncta and red puncta (Fig. 4G). These results indicated that H2O2 exposure promoted autophagic flux in endplate chondrocytes.

3.3. H2O2 promoted mitochondrial dynamics in endplate chondrocytes
Mitochondria are highly dynamic organelles that continuously undergo fission and fusion, known as mitochondrial dynamics. Mitophagy is preceded by mitochondrial fission, which generates individual mitochondrial fragments of manageable size for encapsulation [36]. Therefore, we assessed the expression of Drp1 and Mfn2, which are necessary for mitochondrial fission and fusion, respectively. Western blotting revealed that H2O2 increased the protein levels of Drp1 and Mfn2 (Fig. 5A-C). Additionally, recruitment of Drp1 to mitochondria is confirmed to be a prerequisite for mitochondrial fission. We found that H2O2 increased mitochondrial Drp1 levels and decreased cytosolic Drp1 levels (Fig. 5D-F). The immunofluorescence results showed an elevated colocalization of Drp1 and Tom20 in H2O2-treated endplate chondrocytes (Fig. 5G-H).

3.4. Knockdown of Parkin promoted H2O2-induced damage to endplate chondrocytes
Subsequently, we explored the biological roles of mitophagy in H2O2-induced endplate chondrocyte mitochondrial dysfunction and apoptosis. The results showed that knockdown of Parkin augmented the production of intracellular and mitochondrial ROS (Fig. 6A-D), the loss of ΔΨm (Fig. 6E-F), and apoptosis (Fig. 6G-H) induced by H2O2 in endplate chondrocytes, as determined by flow cytometry.
Because autophagic flux was enhanced in endplate chondrocytes treated with H2O2, we next investigated whether Parkin-mediated removal of dysfunctional mitochondria required increased autophagic flux in endplate chondrocytes. Western blotting showed that transfection with si-Beclin-1 significantly decreased Beclin-1 protein levels (Fig. 7A-B). Flow cytometric analysis indicated that H2O2-induced increases in mitochondrial ROS production, ΔΨm loss, and apoptosis were enhanced by si-Beclin-1 (Fig. 7C-H). Next, endplate chondrocytes were pretreated with CQ followed by H2O2 stimulation. As shown in Fig. 7I-N, endplate chondrocytes showed increased mitochondrial ROS production, decreased ΔΨm, and enhanced apoptosis rates after combined H2O2 and CQ treatment compared with H2O2 treatment alone. Taken together, these results indicated that Parkin-mediated removal of damaged mitochondria was essential for preventing mitochondrial dysfunction, oxidative stress, and apoptosis in H2O2-treated endplate chondrocytes and that this protective effect depended largely on effective autophagic flux.

3.5. Nrf2 signaling was activated in H2O2-treated endplate chondrocytes
We then investigated whether Nrf2 signaling was activated under pathological conditions. Western blotting showed that the expression levels of Nrf2 and its targets SOD-2, HO-1, and NQO-1 were increased in endplate chondrocytes after H2O2 treatment compared with those in the control group (Fig. 8A-E). Additionally, we found that H2O2 exposure induced an increase in protein expression of nuclear Nrf2 (Fig. 8F-G). Nrf2 immunofluorescence also showed that H2O2 facilitated the translocation of Nrf2 into the nucleus (Fig. 8H). These results suggested that H2O2

activated Nrf2 signaling in endplate chondrocytes.
3.6. Nrf2 deficiency enhanced H2O2-induced damage to endplate chondrocytes
Next, we specifically knocked down Nrf2 in endplate chondrocytes and then treated the cells with H2O2. Western blotting showed that si-Nrf2 significantly suppressed the expression of Nrf2 and its targets (Fig. 9A-E), and flow cytometric analysis indicated that H2O2 exposure resulted in increased intracellular and mitochondrial ROS production, enhanced ΔΨm loss, and elevated apoptosis rates, which were further enhanced by si-Nrf2 (Fig. 9F-M). Together, these findings indicated that increased Nrf2 expression and activity in endplate chondrocytes partially prevented H2O2-induced oxidative stress, mitochondrial impairment, and apoptosis.

3.7. Increased ROS production was involved in the activation of Parkin and Nrf2 signaling in H2O2-treated endplate chondrocytes
To improve our understanding of the mechanisms of Parkin and Nrf2 signaling activation, endplate chondrocytes were pretreated with the ROS scavenger N-acetyl-L-cysteine (NAC), followed by H2O2 exposure. The results showed that pretreatment with NAC significantly inhibited H2O2-induced increases in intracellular ROS levels in endplate chondrocytes (Fig. 10A-B). Moreover, protein expression of Nrf2 and Parkin and the ratio of LC3-II/I were lower in the H2O2 and NAC combination treatment group than in the H2O2 alone treatment group (Fig. 10C-G).
Additionally, the relationship between Parkin and Nrf2 signaling in endplate chondrocytes exposed to oxidative stress was assessed. We found that H2O2-induced Nrf2 expression was further enhanced by si-Parkin transfection (Fig. 10H-I). Moreover, si-Nrf2 transfection augmented the H2O2-induced expression of Parkin and conversion of LC3-I to LC3-II (Fig. 10J-L).

3.8. Polydatin prevented H2O2-induced mitochondrial dysfunction and apoptosis through Parkin and Nrf2 in endplate chondrocytes

Polydatin (PD) is a glucoside of resveratrol that is extracted from the rhizomes of Polygonum cuspidatum [37]. Previous studies have demonstrated that PD is able to activate Parkin and Nrf2 pathway in different disease models [38-41]. Thus, we investigated its effect in H2O2-treated endplate chondrocytes. As shown in Fig. 11A-C, PD increased the protein levels of total and mitochondrial Parkin in H2O2-treated endplate chondrocytes. In addition, the protein expression of Nrf2 and its nuclear translocation were upregulated in the H2O2 and PD combination treatment group compared to the H2O2 alone treatment group (Fig. 11D-F). Moreover, flow cytometric analysis revealed that PD pretreatment alleviated H2O2-induced increased intracellular and mitochondrial ROS production, enhanced ΔΨm loss, and elevated apoptosis rates in endplate chondrocytes (Fig. 11G-N). Knockdown of Parkin or Nrf2 partially abolished these protective effects of PD (Fig. 11G-N). Together, these findings indicated that PD-activated Parkin and Nrf2 pathway could partially prevent the mitochondrial dysfunction, ROS overproduction, and apoptosis induced by H2O2 in endplate chondrocytes.

3.9. PD ameliorated CEP and disc degeneration in vivo
Finally, the potential therapeutic effect of PD on CEP and disc degeneration in vivo was evaluated. A rat IDD model was established through disc puncture surgery. MRI was performed at 4 weeks after the puncture to assess the level of disc degeneration in rats. MR image showed a decrease of T2-weighted signal intensity in IDD group, while the T2-weighted signal intensity in PD group was higher than in IDD group (Fig. 12A). In addition, the Pfirrmann MRI grade scores, which indicates the degree of disc degeneration, were lower in the PD group than in the IDD group (Fig. 12B). The histomorphological changes in the IVD in the rat model were assessed by using HE and SO staining. HE staining results showed that the structure of CEP and nucleus pulposus disappeared, and the fibrous ring was markedly irregular in IDD group, while in PD group CEP as well as nucleus pulposus were better preserved and regular fibrous ring were still found within the discs (Fig. 12D). The SO staining demonstrated that loss of CEP and nucleus pulposus was alleviated by PD

compared to the IDD group (Fig. 12D). The histological score also indicated the PD protection against CEP and disc degeneration (Fig. 12C). Next, the oxidative stress and apoptosis in CEP were assessed. MDA and H2O2 are key markers of oxidative stress. As shown in Fig. 12E-F, compared with the control group, there was a significant increase in MDA and H2O2 content in CEP of the IDD group. The results of TUNEL staining revealed that the number of TUNEL-positive cells (green) in CEP was increased in the IDD group compared with the control group (Fig. 12G-H). But PD administration reversed these pathological phenomenon in vivo (Fig. 12E-H). Based on the results of in vitro studies, the activation of Parkin and Nrf2 by PD in vivo was further verified. Consistent with in vitro findings, immunohistological staining and its corresponding quantification showed that PD promoted the Parkin and Nrf2 expression in CEP (Fig. 12I-J). Together, these findings suggested that PD ameliorated CEP and disc degeneration in puncture-induced rat model; meanwhile, the PD protection on CEP was carried out likely by upregulating Parkin and Nrf2.

4. Discussion
CEP degeneration is widely recognized as an important contributor to the onset and development of IDD because it can hinder the transport of nutrients to IVDs [5]. Some studies have shown that oxidative stress, mitochondrial dysfunction, and consequent endplate chondrocyte apoptosis play important roles in the pathogenesis of CEP degeneration [7, 14]. Therefore, we explored the roles of Parkin and Nrf2 signaling in regulating the redox balance, mitochondrial homeostasis, and endplate chondrocyte survival under H2O2 treatment. To the best of our knowledge, this is the first report to establish the roles of Parkin and Nrf2 in CEP degeneration.
In this study, H2O2 was selected as a stimulant for oxidative stress. Recent studies have demonstrated that exposure of endplate chondrocytes to H2O2 leads to mitochondrial dysfunction, ROS overproduction, and endplate chondrocyte apoptosis [14]. Here, we obtained the same results following stimulation with H2O2. Mito-TEMPO is a mitochondria-targeted antioxidant that effectively alleviates mitochondrial ROS-originated oxidative stress and maintains mitochondrial

homeostasis [30]. Using Mito-TEMPO, we demonstrated that mitochondrial homeostasis disorders played important roles in H2O2-induced endplate chondrocyte apoptosis.
Mitophagy has attracted substantial interest in recent years owing to its critical roles in the clearance of damaged mitochondria and maintenance of mitochondrial homeostasis [42]. Parkin provides a mechanistic link between mitochondrial damage and mitophagic clearance [18]. First, Parkin is recruited to the mitochondrial outer membrane of damaged mitochondria to transmit autophagy signals. Then, Parkin-labeled damaged mitochondria are engulfed by LC3-labeled autophagosomes to form mitophagosomes, which are fused with lysosomes to form mitolysosomes, where damaged mitochondria are degraded by hydrolases. Alterations in the expression of Parkin affect the clearance of damaged mitochondria, which can lead to the pathogenesis of several diseases [36, 43]. For example, mutations in the gene encoding Parkin cause early-onset familial Parkinson’s disease through defects in mitophagy [44]. In this study, we tested whether Parkin could function to eliminate damaged mitochondria and reduce oxidative stress in endplate chondrocytes. We found that Parkin protein levels and mitochondrial Parkin translocation were increased in endplate chondrocytes upon H2O2 stimulation, which was accompanied by increased colocalization of Tom20 with LC3 and LAMP1, indicating mitophagy activation. Moreover, knockdown of Parkin reduced the colocalization of Tom20 and LAMP1, indicating that Parkin was necessary for lysosomal degradation of damaged mitochondria. Our results also showed that endplate chondrocytes with depleted Parkin expression exhibited ROS overproduction, mitochondrial dysfunction, and apoptosis, indicating the important roles of Parkin in the clearance of dysfunctional mitochondria and its impact on oxidative stress and survival in endplate chondrocytes. Autophagy, a highly conserved lysosomal-degradation mechanism responsible for the quality control of cellular organelles and proteins, is thought to be essential for the maintenance of cellular homeostasis. Autophagy dysfunction is closely related to several diseases, including IDD [45, 46]. The efficiency of the autophagy degradation pathway depends on the patency of autophagic flux. When all the stages of autophagic

flux remain unobstructed, degradation of the substrate can be achieved. In contrast, when downstream autophagic flux is blocked, the substrate cannot be efficiently degraded. A recent study reported that tumor necrosis factor-α blocks autophagic flux via generating ROS in nucleus pulposus cells and further inhibiting the effects of Parkin-mediated clearance of damaged mitochondria, thereby contributing to apoptosis [19]. In this study, we found that the LC3-II/I ratio and lysosomal quenching of GFP fluorescence were increased in endplate chondrocytes treated with H2O2, indicating that H2O2 enhanced autophagic flux in endplate chondrocytes. One possible reason for this discrepancy was the differences in reactions of various cell types following induction of oxidative stress. Furthermore, we found that autophagy inhibition sensitized endplate chondrocytes to H2O2-induced ROS overproduction, mitochondrial dysfunction, and apoptosis. Taken together, our results showed that Parkin-mediated removal of damaged mitochondria prevented ROS overproduction, mitochondrial impairment, and apoptosis in endplate chondrocytes under oxidative stress.
Mitochondrial fusion and fission play an important role in maintaining the quality
of mitochondria [47]. Mitochondrial fusion facilitates the exchange of mitochondrial components, such as DNA, proteins, and metabolites. Mitochondrial fission promotes the separation of dysfunctional mitochondrial regions for their degradation through mitophagy [36]. In this study, we found that H2O2 not only increased the expression of the mitochondrial fusion protein Mfn1, but also increased the expression of Drp1 and mitochondrial translocation of Drp1 in endplate chondrocytes, suggesting that H2O2 promotes mitochondrial dynamics. Mitochondrial dynamics and mitophagy serve as coordinated quality-control mechanisms for maintaining a healthy mitochondrial population [36]. In endplate chondrocytes exposed to oxidative stress, activated Parkin-mediated mitophagy clears nonfunctional and undesirable mitochondrial fragmentation from mitochondrial fission, which contributes to the maintenance of mitochondrial homeostasis and cell survival.
The transcription factor Nrf2, a master regulator of cellular redox homeostasis, orchestrates the expression of approximately 250 genes involved in various cellular

functions and contributes to cytoprotection against environmental stress and oxidative damage [48]. The Nrf2-activating drug Dimethyl Fumarate has been approved by the US Food and Drug Administration for the treatment of multiple sclerosis, in part based on its anti-oxidative and anti-apoptotic effects [49]. Furthermore, Nrf2 plays an important role in maintaining mitochondrial homeostasis by affecting ΔΨm, respiration, oxidative phosphorylation, ATP synthesis, mitochondrial biogenesis, and mitochondrial integrity [28, 50]. In this study, we found that H2O2 treatment increased the protein levels of Nrf2 and its downstream targets HO-1, SOD-2, and NQO-1 and enhanced the nuclear translocation of Nrf2 in endplate chondrocytes. Thus, activation of Nrf2 signaling may represent an adaptive mechanism through which endplate chondrocytes survive in the disease microenvironment characterized by increased levels of oxidative stress. Our findings provided evidence to support this hypothesis, and our data showed that antioxidant treatment significantly suppressed H2O2-induced Nrf2 expression in endplate chondrocytes, suggesting that Nrf2 activation in endplate chondrocytes was a response to oxidative stress. Moreover, we found that siRNA-mediated depletion of Nrf2 resulted in further elevation of H2O2-induced ROS production, mitochondrial dysfunction, and endplate chondrocyte apoptosis. Collectively, these results suggested that H2O2 increased ROS levels to activate Nrf2 signaling, which in turn alleviated oxidative stress, mitochondrial dysfunction, and apoptosis in endplate chondrocytes. However, the intrinsic activity of Nrf2 in H2O2-treated endplate chondrocytes was not sufficient to fully counteract the pathogenic events induced by H2O2. Therefore, pharmacological intervention to enhance Nrf2 expression and activity in endplate chondrocytes may be an effective therapeutic strategy for CEP degeneration.
Because both Parkin and Nrf2 signaling are required to manage oxidative stress
and mitochondrial dysfunction and promote endplate chondrocyte survival, we hypothesize that there may be interactions between these pathways. Our results indicated that inhibition of Nrf2 increased Parkin expression and the LC3-II/I ratio induced by H2O2 and that inhibition of Parkin also increased H2O2-induced Nrf2 expression. Accordingly, there could be a negative interaction between Parkin and

Nrf2 signaling when endplate chondrocytes were stimulated with high levels of ROS. Our findings were inconsistent with a previous study showing that autophagy activity was enhanced in wild-type nucleus pulposus cells treated with H2O2 compared with that in Nrf2-knockout nucleus pulposus cells [24]. This discrepancy could be explained by differences in cell types and H2O2 concentrations. In our study, Parkin and Nrf2 signaling were suggested to be adaptive mechanisms to cope with inhibition of Parkin or Nrf2 expression-induced increases in oxidative stress damage in H2O2-treated endplate chondrocytes. Further studies are needed to fully elucidate the mechanisms regulating the interactions between Parkin and Nrf2 signaling.
Subsequently, we used a pharmacological method to regulate the Parkin and Nrf2 pathway in order to further study their effects. Previous studies have reported that PD exhibits beneficial effects in various diseases, including nonalcoholic fatty liver disease [39], spinal cord injury [40], and diabetic vascular complication [51]. It is worth noting that there is evidence indicating that the protective effect of PD is performed by activating the Parkin and Nrf2 pathway [38, 41]. In this study, we explored the potential application of PD in endplate chondrocytes under oxidative stress and the results showed that PD could protect endplate chondrocytes against H2O2-induced mitochondrial dysfunction, oxidative stress, and apoptosis, in which activating Parkin and Nrf2 pathway were deeply involved. In addition to the in vitro study, puncture-induced rat models were used to evaluate the effect of PD in vivo. As a result, PD administration could ameliorate CEP degeneration in rat models, along with the upregulation of Parkin and Nrf2 level in CEP. These results suggested that PD-mediated upregulation of Parkin and Nrf2 plays a protective role in CEP degeneration. Moreover, aside from CEP, the nucleus pulposus was also better preserved in the PD group than in the IDD group in the in vivo study, suggesting that PD may alleviate the disc degeneration also by targeting nucleus pulposus.
Meanwhile, we cannot neglect the limitations of this study due to the use of only
degenerated endplate chondrocytes from patients with IDD. In order to enhance the physiological relevance of our results, it is necessary to compare the normal human endplate chondrocytes from normal disc tissues with the degenerated human endplate

chondrocytes to determine if in vivo processes lead to similar signaling pathway activation as seen in vitro with H2O2 exposure. Moreover, the response of degenerated human endplate chondrocytes to H2O2 stimulation in vitro may be different from that of normal endplate chondrocytes. Due to limited access to normal disc tissues, these issues were not investigated in the current study.

5. Conclusion
In conclusion, our findings demonstrated, for the first time, that induction of Parkin and Nrf2 alleviated redox imbalance and mitochondrial dysfunction and eventually improved cell survival in endplate chondrocytes under oxidative stress. Our findings suggested that Parkin and Nrf2 signaling in endplate chondrocytes may be new and effective therapeutic targets for the management of IDD.

Conflict of interest
These authors have no conflict of interest to declare.

Author contributions
YX, XZL, and LK conceived and designed the experiments. LK, SWL, JCL, YYT, YX, and XZL performed the experiments and analyzed the data. LK and YX wrote the paper. YX reviewed and revised the manuscript. All authors have read and approved the final version of the manuscript.

Acknowledgements
This work was supported by the National Natural Science Foundation of China (No. 81871124).

References
[1] J. Hartvigsen, M.J. Hancock, A. Kongsted, Q. Louw, M.L. Ferreira, S. Genevay, D. Hoy, J. Karppinen, G. Pransky, J. Sieper, R.J. Smeets, M. Underwood, What low back pain is and why we need to pay attention, Lancet

(London, England) 391 (2018) 2356-2367.
[2] S. Clark, R. Horton, Low back pain: a major global challenge, Lancet (London, England) 391 (2018) 2302.
[3] K. Luoma, H. Riihimaki, R. Luukkonen, R. Raininko, E. Viikari-Juntura, A. Lamminen, Low back pain in relation to lumbar disc degeneration, Spine 25 (2000) 487-492.
[4] N. Henry, J. Clouet, J. Le Bideau, C. Le Visage, J. Guicheux, Innovative strategies for intervertebral disc regenerative medicine: From cell therapies to multiscale delivery systems, Biotechnology Advances 36 (2018) 281-294.
[5] J. Wong, S.L. Sampson, H. Bell-Briones, A. Ouyang, A.A. Lazar, J.C. Lotz,
A.J. Fields, Nutrient supply and nucleus pulposus cell function: effects of the transport properties of the cartilage endplate and potential implications for intradiscal biologic therapy, Osteoarthritis and Cartilage 27 (2019) 956-964.
[6] K. Ariga, S. Miyamoto, T. Nakase, S. Okuda, W. Meng, K. Yonenobu, H. Yoshikawa, The relationship between apoptosis of endplate chondrocytes and aging and degeneration of the intervertebral disc, Spine 26 (2001) 2414-2420.
[7] K. Chen, X. Lv, W. Li, F. Yu, J. Lin, J. Ma, D. Xiao, Autophagy Is a Protective Response to the Oxidative Damage to Endplate Chondrocytes in Intervertebral Disc: Implications for the Treatment of Degenerative Lumbar Disc, Oxidative Medicine and Cellular Longevity,2017,(2017-02-22) 2017 (2017) 4041768.
[8] C.Q. Zhao, Y.H. Zhang, S.D. Jiang, L.S. Jiang, L.Y. Dai, Both endoplasmic reticulum and mitochondria are involved in disc cell apoptosis and intervertebral disc degeneration in rats, Age (Dordrecht, Netherlands) 32 (2010) 161-177.
[9] A. Belenguer-Varea, F.J. Tarazona-Santabalbina, J.A. Avellana-Zaragoza, M. Ingles, C. Mas-Bargues, M. Martinez-Reig, Oxidative stress and exceptional human longevity: Systematic review, Free radical biology & medicine (2019).
[10] E. Belge Kurutas, M. Senoglu, K.Z. Yuksel, V. Unsal, I. Altun, Oxidative/Nitrosative Stress in Patients With Modic Changes: Preliminary Controlled Study, Spine 40 (2015) 1101-1107.

[11] C. Feng, M. Yang, M. Lan, C. Liu, Y. Zhang, B. Huang, H. Liu, Y. Zhou, ROS: Crucial Intermediators in the Pathogenesis of Intervertebral Disc Degeneration, Oxidative Medicine and Cellular Longevity,2017,(2017-3-14) 2017 (2017) 1-12.
[12] Z. Jiang, W. Lu, Q. Zeng, D. Li, L. Ding, J. Wu, High glucose-induced excessive reactive oxygen species promote apoptosis through mitochondrial damage in rat cartilage endplate cells, Journal of orthopaedic research : official publication of the Orthopaedic Research Society 36 (2018) 2476-2483.
[13] D. Kong, T. Zheng, M. Zhang, D. Wang, S. Du, X. Li, J. Fang, X. Cao, Static mechanical stress induces apoptosis in rat endplate chondrocytes through MAPK and mitochondria-dependent caspase activation signaling pathways, PLoS One 8 (2013) e69403-e69403.
[14] Y. Han, X. Li, M. Yan, M. Yang, S. Wang, J. Pan, L. Li, J. Tan, Oxidative damage induces apoptosis and promotes calcification in disc cartilage endplate cell through ROS/MAPK/NF-kappaB pathway: Implications for disc degeneration, Biochemical and biophysical research communications 516 (2019) 1026-1032.
[15] Z. Zhang, J. Lin, N. Tian, Y. Wu, Y. Zhou, C. Wang, Q. Wang, H. Jin, T. Chen,
M. Nisar, Melatonin protects vertebral endplate chondrocytes against apoptosis and calcification via the Sirt1‐autophagy pathway, Journal of cellular and molecular medicine (2019).
[16] G.P. Kaushal, K. Chandrashekar, L.A. Juncos, Molecular Interactions Between Reactive Oxygen Species and Autophagy in Kidney Disease, International journal of molecular sciences 20 (2019) 3791.
[17] C. Zhao, Z. Chen, X. Xu, X. An, S. Duan, Z. Huang, C. Zhang, L. Wu, B. Zhang, A. Zhang, C. Xing, Y. Yuan, Pink1/Parkin-mediated mitophagy play a protective role in cisplatin induced renal tubular epithelial cells injury, Experimental cell research 350 (2017) 390-397.
[18] A. Pilsl, K.F. Winklhofer, Parkin, PINK1 and mitochondrial integrity: emerging concepts of mitochondrial dysfunction in Parkinson’s disease, Acta

neuropathologica 123 (2012) 173-188.
[19] Z. Zhang, T. Xu, J. Chen, Z. Shao, K. Wang, Y. Yan, C. Wu, J. Lin, H. Wang,
W. Gao, X. Zhang, X. Wang, Parkin-mediated mitophagy as a potential therapeutic target for intervertebral disc degeneration, Cell death & disease 9 (2018) 980.
[20] Q. Ma, Role of nrf2 in oxidative stress and toxicity, Annual review of pharmacology and toxicology 53 (2013) 401-426.
[21] F. Sivandzade, A. Bhalerao, L. Cucullo, Cerebrovascular and neurological disorders: protective role of NRF2, International journal of molecular sciences 20 (2019) 3433.
[22] Z.H. Wan, T.Y. Jiang, Y.Y. Shi, Y.F. Pan, Y.K. Lin, Y.H. Ma, C. Yang, X.F.
Feng, L.F. Huang, X.N. Kong, RPB5 ‐ mediating protein promotes cholangiocarcinoma tumorigenesis and drug resistance by competing with NRF2 for KEAP1 binding, Hepatology (2019).
[23] M. Rojo de la Vega, E. Chapman, D.D. Zhang, NRF2 and the Hallmarks of Cancer, Cancer cell 34 (2018) 21-43.
[24] Z. Tang, B. Hu, F. Zang, J. Wang, X. Zhang, H. Chen, Nrf2 drives oxidative stress-induced autophagy in nucleus pulposus cells via a Keap1/Nrf2/p62 feedback loop to protect intervertebral disc from degeneration, Cell death & disease 10 (2019) 510.
[25] W. Fang, X. Zhou, J. Wang, L. Xu, L. Zhou, W. Yu, Y. Tao, J. Zhu, B. Hu, C. Liang, Wogonin mitigates intervertebral disc degeneration through the Nrf2/ARE and MAPK signaling pathways, International immunopharmacology 65 (2018) 539-549.
[26] W. Jianle, H. Chongan, L. Zongze, P. Xiangxiang, C. Jiaoxiang, Z. Gang, T. Naifeng, Y. Yingzhao, Z. Zengjie, H. Jianing, Polydatin suppresses nucleus pulposus cell senescence, promotes matrix homeostasis and attenuates intervertebral disc degeneration in rats, Journal of cellular and molecular medicine.
[27] Y. Wang, R. Zuo, Z. Wang, L. Luo, J. Wu, C. Zhang, M. Liu, C. Shi, Y. Zhou,

Kinsenoside ameliorates intervertebral disc degeneration through the activation of AKT-ERK1/2-Nrf2 signaling pathway, Aging 11 (2019).
[28] A.T. Dinkova-Kostova, A.Y. Abramov, The emerging role of Nrf2 in mitochondrial function, Free Radical Biology and Medicine 88 (2015) 179-188.
[29] N. Zhou, X. Lin, W. Dong, W. Huang, W. Jiang, L. Lin, Q. Qiu, X. Zhang, J. Shen, Z. Song, SIRT1 alleviates senescence of degenerative human intervertebral disc cartilage endo-plate cells via the p53/p21 pathway, Sci Rep 6 (2016) 22628.
[30] Y. Song, S. Li, W. Geng, R. Luo, W. Liu, J. Tu, K. Wang, L. Kang, H. Yin, X. Wu, Y. Gao, Y. Zhang, C. Yang, Sirtuin 3-dependent mitochondrial redox homeostasis protects against AGEs-induced intervertebral disc degeneration, Redox biology 19 (2018) 339-353.
[31] L. Kang, J. Hu, Y. Weng, J. Jia, Y. Zhang, Sirtuin 6 prevents matrix degradation through inhibition of the NF-kappaB pathway in intervertebral disc degeneration, Experimental cell research 352 (2017) 322-332.
[32] T. Yu, F. Guo, Y. Yu, T. Sun, D. Ma, J. Han, Y. Qian, I. Kryczek, D. Sun, N. Nagarsheth, Y. Chen, H. Chen, J. Hong, W. Zou, J.Y. Fang, Fusobacterium nucleatum Promotes Chemoresistance to Colorectal Cancer by Modulating Autophagy, Cell 170 (2017) 548-563.e516.
[33] C. Jiang, Q. Guo, Y. Jin, J.-J. Xu, Z.-M. Sun, D.-C. Zhu, J.-H. Lin, N.-F. Tian,
L.-J. Sun, X.-L. Zhang, Y.-S. Wu, Inhibition of EZH2 ameliorates cartilage endplate degeneration and attenuates the progression of intervertebral disc degeneration via demethylation of Sox-9, EBioMedicine 48 (2019) 619-629.
[34] C.W. Pfirrmann, A. Metzdorf, M. Zanetti, J. Hodler, N. Boos, Magnetic resonance classification of lumbar intervertebral disc degeneration, Spine 26 (2001) 1873-1878.
[35] H.J. Mao, Q.X. Chen, B. Han, F.C. Li, J. Feng, Z.L. Shi, M. Lin, J. Wang, The effect of injection volume on disc degeneration in a rat tail model, Spine 36 (2011) E1062-1069.

[36] S. Rovira-Llopis, C. Bañuls, N. Diaz-Morales, A. Hernandez-Mijares, M. Rocha, V.M. Victor, Mitochondrial dynamics in type 2 diabetes: Pathophysiological implications, Redox biology 11 (2017) 637-645.
[37] H.-L. Wang, J.-P. Gao, Y.-L. Han, X. Xu, R. Wu, Y. Gao, X.-H. Cui,
Comparative studies of polydatin and resveratrol on mutual transformation and antioxidative effect in vivo, Phytomedicine 22 (2015) 553-559.
[38] T. Li, Y. Liu, W. Xu, X. Dai, R. Liu, Y. Gao, Z. Chen, Y. Li, Polydatin mediates Parkin-dependent mitophagy and protects against mitochondria-dependent apoptosis in acute respiratory distress syndrome, Laboratory Investigation 99 (2019) 819.
[39] X.-J. Zhao, H.-W. Yu, Y.-Z. Yang, W.-Y. Wu, T.-Y. Chen, K.-K. Jia, L.-L. Kang,
R.-Q. Jiao, L.-D. Kong, Polydatin prevents fructose-induced liver inflammation and lipid deposition through increasing miR-200a to regulate Keap1/Nrf2 pathway, Redox biology 18 (2018) 124-137.
[40] R. Lv, L. Du, L. Zhang, Z. Zhang, Polydatin attenuates spinal cord injury in rats by inhibiting oxidative stress and microglia apoptosis via Nrf2/HO-1 pathway, Life Sciences 217 (2019) 119-127.
[41] H. Zeng, Y. Wang, Y. Gu, J. Wang, H. Zhang, H. Gao, Q. Jin, L. Zhao, Polydatin attenuates reactive oxygen species-induced airway remodeling by promoting Nrf2-mediated antioxidant signaling in asthma mouse model, Life sciences 218 (2019) 25-30.
[42] J.M. Bravo-San Pedro, G. Kroemer, L. Galluzzi, Autophagy and Mitophagy in Cardiovascular Disease, Circulation research 120 (2017) 1812-1824.
[43] S. Pickles, P. Vigié, R.J. Youle, Mitophagy and Quality Control Mechanisms in Mitochondrial Maintenance, Current Biology 28 (2018) R170-R185.
[44] J.-Y. Lee, Y. Nagano, J.P. Taylor, K.L. Lim, T.-P. Yao, Disease-causing mutations in Parkin impair mitochondrial ubiquitination, aggregation, and HDAC6-dependent mitophagy, The Journal of Cell Biology 189 (2010) 671.
[45] G. Zheng, Z. Pan, Y. Zhan, Q. Tang, F. Zheng, Y. Zhou, Y. Wu, Y. Zhou, D. Chen, J. Chen, X. Wang, W. Gao, H. Xu, N. Tian, X. Zhang, TFEB protects

nucleus pulposus cells against apoptosis and senescence via restoring autophagic flux, Osteoarthritis Cartilage 27 (2019) 347-357.
[46] S.Y. Bae, S. Byun, S.H. Bae, D.S. Min, H.A. Woo, K. Lee, TPT1 (tumor protein, translationally-controlled 1) negatively regulates autophagy through the BECN1 interactome and an MTORC1-mediated pathway, Autophagy 13 (2017) 820-833.
[47] C. Vásquez‐Trincado, I. García‐Carvajal, C. Pennanen, V. Parra, J.A. Hill,
B.A. Rothermel, S. Lavandero, Mitochondrial dynamics, mitophagy and cardiovascular disease, The Journal of physiology 594 (2016) 509-525.
[48] J.M. Lee, J.A. Johnson, An important role of Nrf2-ARE pathway in the cellular defense mechanism, Journal of biochemistry and molecular biology 37 (2004) 139-143.
[49] R. Gold, L. Kappos, D.L. Arnold, A. Bar-Or, G. Giovannoni, K. Selmaj, C. Tornatore, M.T. Sweetser, M. Yang, S.I. Sheikh, K.T. Dawson, Placebo-controlled phase 3 study of oral BG-12 for relapsing multiple sclerosis, The New England journal of medicine 367 (2012) 1098-1107.
[50] K. Itoh, P. Ye, T. Matsumiya, K. Tanji, T. Ozaki, Emerging functional cross-talk between the Keap1-Nrf2 system and mitochondria, Journal of clinical biochemistry and nutrition 56 (2015) 91-97.
[51] N. Pang, T. Chen, X. Deng, N. Chen, R. Li, M. Ren, Y. Li, M. Luo, H. Hao, J. Wu, Polydatin prevents methylglyoxal-induced apoptosis through reducing oxidative stress and improving mitochondrial function in human umbilical vein endothelial cells, Oxidative medicine and cellular longevity 2017 (2017).

Figure legends
Fig. 1. Effects of H2O2 on ROS production, mitochondrial homeostasis and apoptosis of endplate chondrocytes. Endplate chondrocytes were treated with H2O2 (0, 100 and 200 μM) for 2 h. (A-B) The intracellular ROS levels in the endplate chondrocytes were detected using the DCFH-DA and measured by flow cytometry. (C-D) The mitochondrial ROS levels were detected using the MitoSOX Red and measured by flow cytometry. (E) Calcein fluorescence intensity in mitochondria was measured to reflect the opening level of mPTP. (F-G) ΔΨm was detected by JC-1 staining and measured by flow cytometry. (H) Intracellular ATP levels in the endplate chondrocytes. (I-L) The protein levels of mitochondrial cytochrome c (mito-cyt c), cytoplasmic cytochrome c (cyto-cyt c), and cleaved caspase-3 (c-caspase 3) in endplate chondrocytes were measured by western blotting. (M-N) Annexin V-APC/7-AAD staining results showing the rate of apoptosis in endplate chondrocytes. The results in figure B, D, E, H, J, K and L were normalized to control. Data are represented as the mean ± SD. ***P < 0.001, **P < 0.01 and *P < 0.05 versus control group.

Fig. 2. Effects of Mito-TEMPO on H2O2-induced ROS accumulation, mitochondrial dysfunction, and apoptosis in endplate chondrocytes. Endplate chondrocytes were pretreated with Mito-TEMPO (20 μM) for 1 h and then treated with H2O2 (200 μM) for 2 h. (A-B) The intracellular ROS levels in the endplate chondrocytes were detected using the DCFH-DA and measured by flow cytometry. (C-D) The mitochondrial ROS levels were detected using the MitoSOX Red and measured by flow cytometry. (E) Calcein fluorescence intensity in mitochondria was measured to reflect the opening level of mPTP. (F-G) ΔΨm was detected by JC-1 staining and measured by flow cytometry. (H) Intracellular ATP levels in the endplate chondrocytes. (I-L) The protein levels of mito-cyt c, cyto-cyt c, and c-caspase 3 in endplate chondrocytes were measured by western blotting. (M-N) Annexin V-APC/7-AAD staining results showing the rate of apoptosis in endplate chondrocytes. The results in figure B, D, E, H, J, K and L were normalized to control.

Data are represented as the mean ± SD. ***P < 0.001 versus control group. ###P <
0.001 and ##P < 0.01 versus H2O2 treatment only group.
Fig. 3. Effects of H2O2 on Parkin-mediated mitophagy in endplate chondrocytes. (A-J) Endplate chondrocytes were treated with H2O2 (200 μM) for 2 h. (A-B) The protein level of Parkin in the endplate chondrocytes was measured by western blotting. (C-D) The protein level of mito-Parkin in the endplate chondrocytes was measured by western blotting. (E-F) The co-localization of Parkin and Tom20 was examined by confocal microscopy. Scale bar: 10 μm. (G-H) The co-localization of LC3 and Tom20 was examined by confocal microscopy. Scale bar: 10 μm. (I-J) The protein level of mitochondrial LC3 in the endplate chondrocytes was measured by western blotting. (K-L) Endplate chondrocytes were transfected with si-Parkin (100 nM) or si-Control (100 nM) for 48 h followed by the administration of H2O2 (200 μM, 2 h). The co-localization of LAMP1 and Tom20 was examined by confocal microscopy. Scale bar: 10 μm. (M-N) The protein level of Parkin in the endplate chondrocytes was measured by western blotting. The results in figure B, D, F, H, J, L and N were normalized to control. Data are represented as the mean ± SD. ***P < 0.001 and **P
< 0.01 versus control group. ###P < 0.001 versus H2O2 + si-Control treatment group.
Fig. 4. Effects of H2O2 on autophagic flux in endplate chondrocytes. (A-D) Endplate chondrocytes were treated with H2O2 (200 μM) for 2 h. The protein levels of LC3 and Beclin-1 in the endplate chondrocytes were measured by western blotting. (E-F) Endplate chondrocytes were either untreated or treated with H2O2 (200 μM) for 2 h in the presence or absence of CQ (10 μM). The protein level of LC3-II in the endplate chondrocytes was measured by western blotting. (G) Representative images of the endplate chondrocytes expressing mRFP-GFP-LC3 were obtained by confocal microscopy. Scale bar: 10 μm. The results in figure B, D and F were normalized to control. Data are represented as the mean ± SD. ***P < 0.001 and **P < 0.01 versus control group. ##P < 0.01 versus CQ treatment only group.

Fig. 5. H2O2 promoted mitochondrial dynamics in endplate chondrocytes. Endplate chondrocytes were treated with H2O2 (200 μM) for 2 h. (A-C) The protein levels of Drp1 and Mfn2 in the endplate chondrocytes were measured by western blotting. (D-F) The protein levels of mito-Drp1 and cyto-Drp1 in the endplate chondrocytes were measured by western blotting. (G-H) The co-localization of Drp1 and Tom20 was examined by confocal microscopy. Scale bar: 10 μm. The results in figure B, C, E, F and H were normalized to control. Data are represented as the mean
± SD. **P < 0.01 and *P < 0.05 versus control group.
Fig. 6. Knockdown of Parkin promotes H2O2-induced damage to endplate chondrocytes. Endplate chondrocytes were transfected with si-Parkin (100 nM) or si-Control (100 nM) for 48 h followed by the administration of H2O2 (200 μM, 2 h). (A-B) The intracellular ROS levels in the endplate chondrocytes were detected using the DCFH-DA and measured by flow cytometry. (C-D) The mitochondrial ROS levels were detected using the MitoSOX Red and measured by flow cytometry. (E-F) ΔΨm was detected by JC-1 staining and measured by flow cytometry. (G-H) Annexin V-APC/7-AAD staining results showing the rate of apoptosis in endplate chondrocytes. The results in figure B and D were normalized to control. Data are represented as the mean ± SD. ***P < 0.001 versus si-Control treatment only group. ##P < 0.01 and #P < 0.05 versus H2O2 + si-Control treatment group.

Fig. 7. Impairment of autophagic flux promotes H2O2-induced damage to endplate chondrocytes. (A-B) Endplate chondrocytes were transfected with si-Beclin-1 (100 nM) or si-Control (100 nM) for 48 h. The protein level of Beclin-1 in the endplate chondrocytes was measured by western blotting. (C-H) Endplate chondrocytes were transfected with si-Beclin-1 (100 nM) or si-Control (100 nM) for 48 h followed by the administration of H2O2 (200 μM, 2 h). (C-D) The mitochondrial ROS levels were detected using the MitoSOX Red and measured by flow cytometry. (E-F) ΔΨm was detected by JC-1 staining and measured by flow cytometry. (G-H) Annexin V-APC/7-AAD staining results showing the rate of apoptosis in endplate
chondrocytes. ***P < 0.001 and **P < 0.01 versus si-Control treatment only group. ##P < 0.01 and #P < 0.05 versus H2O2 + si-Control treatment group. (I-N) Endplate chondrocytes were pretreated with CQ (10 μM) for 2 h before the administration of H2O2 (200 μM, 2 h). (I-J) The mitochondrial ROS levels were detected using the MitoSOX Red and measured by flow cytometry. (K-L) ΔΨm was detected by JC-1 staining and measured by flow cytometry. (M-N) Annexin V-APC/7-AAD staining results showing the rate of apoptosis in endplate chondrocytes. The results in figure B, D and J were normalized to control. Data are represented as the mean ± SD. ***P <
0.001 versus Control group. #P < 0.05 versus H2O2 treatment only group.
Fig. 8. Nrf2 signaling was activated in H2O2-treated endplate chondrocytes. Endplate chondrocytes were treated with H2O2 (200 μM) for 2 h. (A-E) The protein levels of Nrf2, SOD-2, HO-1, and NQO-1 in the endplate chondrocytes were measured by western blotting. (F-G) The protein expression of nucleus Nrf2 (Nu-Nrf2) was measured by western blotting. (H) The nuclear translocation of Nrf2 was examined by immunofluorescence on confocal microscope. Scale bar: 10 μm. The results in figure B, C, D, E and G were normalized to control. Data are represented as the mean ± SD. ***P < 0.001, **P < 0.01 and *P < 0.05 versus Control group.

Fig. 9. Nrf2 deficiency enhances H2O2-induced damage to endplate chondrocytes. (A-E) Endplate chondrocytes were transfected with si-Nrf2 (100 nM) or si-Control (100 nM) for 48 h. The protein levels of Nrf2, SOD-2, HO-1, and NQO-1 in the endplate chondrocytes were measured by western blotting. (F-M) Endplate chondrocytes were transfected with si-Nrf2 (100 nM) or si-Control (100 nM) for 48 h followed by the administration of H2O2 (200 μM, 2 h). (F-G) The intracellular ROS levels in the endplate chondrocytes were detected using the DCFH-DA and measured by flow cytometry. (H-I) The mitochondrial ROS levels were detected using the MitoSOX Red and measured by flow cytometry. (J-K) ΔΨm was detected by JC-1 staining and measured by flow cytometry. (L-M) Annexin V-APC/7-AAD staining results showing the rate of apoptosis in endplate chondrocytes. The results in figure B,
C, D, E, G and I were normalized to control. Data are represented as the mean ± SD.
P < 0.001, P < 0.01 and *P < 0.05 versus si-Control treatment only group. ##P <
0.01 and #P < 0.05 versus H2O2 + si-Control treatment group.

Fig. 10. ROS production elevation is a critical event for H2O2-induced Parkin and Nrf2 signaling in endplate chondrocytes. (A-G) Endplate chondrocytes were pretreated with NAC (10 mM) for 1 h followed by the administration of H2O2 (200 μM, 2 h). (A-B) The intracellular ROS levels in the endplate chondrocytes were detected using the DCFH-DA and measured by flow cytometry. (C-E) The protein levels of Parkin and LC3 in the endplate chondrocytes were measured by western blotting. (F-G) The protein level of Nrf2 in the endplate chondrocytes was measured by western blotting. ***P < 0.001 and P < 0.01 versus Control group. ##P < 0.01 versus H2O2 treatment only group. (H-L) Endplate chondrocytes were transfected with si-Parkin (100 nM), si-Nrf2 (100 nM) or si-Control (100 nM) for 48 h followed by the administration of H2O2 (200 μM, 2 h). (H-I) The protein level of Nrf2 in the endplate chondrocytes was measured by western blotting. (J-L) The protein levels of Parkin and LC3 in the endplate chondrocytes were measured by western blotting. The results in figure B, D, E, G, I, K and L were normalized to control. Data are represented as the mean ± SD. ***P < 0.001, **P < 0.01 and *P < 0.05 versus si-Control treatment only group. ##P < 0.01 and #P < 0.05 versus H2O2 + si-Control treatment group.

Fig. 11. Polydatin prevented H2O2-induced mitochondrial dysfunction and apoptosis through Parkin and Nrf2 in endplate chondrocytes. (A-F) Endplate chondrocytes were pretreated with Polydatin (PD) (200 μM) for 2 h followed by the administration of H2O2 (200 μM, 2 h). (A-C) The protein levels of Parkin and mito-Parkin in the endplate chondrocytes were measured by western blotting. (D-F) The protein levels of Nrf2 and Nu-Nrf2 in the endplate chondrocytes were measured by western blotting. ***P < 0.001 and **P < 0.01 versus Control group. ##P < 0.01 and #P < 0.05 versus H2O2 treatment only group. (G-N) Endplate chondrocytes were

transfected with si-Parkin (100 nM), si-Nrf2 (100 nM), or si-Control (100 nM) for 48 h followed by the administration of PD and H2O2. (G-H) The intracellular ROS levels in the endplate chondrocytes were detected using the DCFH-DA and measured by flow cytometry. (I-J) The mitochondrial ROS levels were detected using the MitoSOX Red and measured by flow cytometry. (K-L) ΔΨm was detected by JC-1 staining and measured by flow cytometry. (M-N) Annexin V-APC/7-AAD staining results showing the rate of apoptosis in endplate chondrocytes. The results in figure B, C, E, F, H and J were normalized to control. Data are represented as the mean ± SD. ***P < 0.001,
**P < 0.01 and *P < 0.05 versus H2O2 + si-Control treatment group. ##P < 0.01 and #P < 0.05 versus H2O2 + PD + si-Control treatment group.

Fig. 12. PD ameliorated CEP and disc degeneration in vivo. (A) The discs from the tails of rat obtained from different treatment groups were examined using MRI at T2-weighted signal (white arrows). (B) Quantitative analysis of the degree of disc degeneration based on the Pfirrmann MRI grade system. (C) The histological scores of the tail discs according to the histological grading scale. (D) HE and SO staining of CEP and disc in different treatment groups. Scale bar: 200 μm in CEP; Scale bar: 500 μm in general disc. (E) The content of H2O2 in CEP samples of the rat model. (F) The level of MDA in CEP samples of the rat model. (G-H) TUNEL staining and fluorescence microscope analysis were used to evaluated the apoptosis in CEP. Scale bar: 100 μm. (I-J) Immunohistochemical staining showing the expression of Parkin and Nrf2 proteins in CEP. Scale bar: 20 μm. The results in figure E and F were normalized to control. Data are represented as the mean ± SD. P < 0.001, **P <
0.01 and *P < 0.05 versus Control group. P < 0.001, ##P < 0.01 and #P < 0.05
versus IDD group.

Table 1. Patient information enrolled in this study

Case no. Age (years) Sex Disc level Pfirrmann grade
1 52 Male L4-5 IV
2 56 Female L4-5 IV
3 61 Female L5-S1 IV
4 65 Male L4-5 IV
5 62 Female L3-4 IV
6 49 Male L5-S1 V
7 51 Female L4-5 IV
8 47 Male L5-S1 IV
9 65 Male L5-S1 V
10 59 Female L4-5 V
11 57 Male L5-S1 V
12 69 Male L4-5 IV

Highlights

Mitophagy and mitochondrial dynamics are promoted in H2O2-treated endplate chondrocytes
Nrf2 signaling is activated in H2O2-treated endplate chondrocytes

Parkin or Nrf2 deficiency enhances H2O2-induced damage to endplate chondrocytes
Upregulation of Parkin and Nrf2 by polydatin protects Mito-TEMPO endplate chondrocytes against the oxidative damage