Oxygen-Sensing Nox4 Generates Genotoxic ROS to Induce Premature Senescence of Nucleus Pulposus Cells through MAPK and NF-κB Pathways

Senescence is a crucial driver of intervertebral disc degeneration (IDD). Disc cells are exposed to high oxygen tension due to neovascularization in degenerative discs. However, the effect of oxygen tension on disc cell senescence was unknown. Herein, rat nucleus pulposus (NP) cells were cultured under 20% O2 or 1% O2. Consequently, ROS induced by 20% O2 caused DNA damage and then activated p53-p21-Rb and p16-Rb pathways via ERK signaling to induce NP cell senescence. It also induced catabolic and proinflammatory phenotype of NP cells via MAPK and NF-κB pathways. Furthermore, 20% O2 was found to upregulate Nox4 in NP cells. Small interfering RNA against Nox4 reduced ROS production induced by 20% O2 and consequently suppressed premature senescence of NP cells. On the contrary, NP cells overexpressing Nox4 produced more ROS and rapidly developed senescent signs. In consistent with the in vitro studies, the expression of Nox4, p21, and Rb was upregulated in rat degenerative discs. This study, for the first time, demonstrates that Nox4 is an oxygen-sensing enzyme and a main ROS source in NP cells. Nox4-dependent ROS are genotoxic and a potent trigger of NP cell senescence. Nox4 is a potential therapeutic target for disc cell senescence and IDD.


Introduction
The accumulation of senescent intervertebral disc (IVD) cells is a hallmark of IVD degeneration (IDD) [1,2]. Senescent disc cells are irreversibly cell cycle arrested, causing a decrease in the number of functional and viable disc cells gradually due to apoptosis or necrosis. On the other hand, an aberrant secretion of extracellular matrix (ECM) catabolic proteases and proinflammatory mediators is strongly associated with the senescence-associated secretory phenotype (SASP) of disc cells [3,4]. The SASP causes ECM turnover and inflammation in the microenvironment of IVDs and also exerts negative effects on neighboring normal disc cells [5,6]. Together, disc cell senescence is crucial to the establishment and progression of IDD [7,8]. Investigating the molecular mechanism underlying disc cell senescence contributes to understanding the pathogenesis of IDD further in depth.
Preventing or delaying disc cell senescence is a potential therapeutic approach for IDD.
The microenvironment of degenerative discs is characterized by acidic PH, low nutrition, high osmotic pressure, and inflammation [9]. Oxygen tension in avascular IVDs is low (1% O 2 in nucleus pulposus (NP)). NP cells are subjected to a hypoxic but not entirely anaerobic microenvironment [10,11]. NP cells adapt to hypoxic microenvironment well and still have a certain level of oxidative metabolism [12,13]. Nevertheless, degenerative discs with neovascularization are characterized by increased oxygen tension. Therefore, the resident NP cells are exposed to higher oxygen tension [14,15]. High oxygen tension (20% O 2 ) has been reported to enhance the production of mitochondrialderived reactive oxygen species (ROS) in NP cells to perturb ECM homeostasis of IVDs [16]. On the contrary, lower oxygen tension (2% O 2 ) was shown to decrease ROS production in NP cells and consequently to suppress the excessive autophagy of NP cells [17]. The findings suggest that oxygen tension affects the function and viability of NP cells through regulating ROS metabolism. However, the effect of oxygen tension on disc cell senescence remains unknown.
Numerous lines of evidence indicate that ROS are involved in the signal transduction, ECM metabolism, programed cell death, senescence, and phenotypic shift of disc cells [4,18,19]. With respect to the source of ROS in disc cells, mitochondrion is a major site of ROS generation in disc cells [16,20]. However, nonmitochondrial ROS production through nicotinamide adenine dinucleotide phosphate (NADPH) oxidase (Nox) in disc cells is unclear. The Nox family consists of seven isoforms (Nox1-5 and dual oxidase 1-2). They are professional ROS-generating enzymes and have been documented in various diseases, including cardiovascular diseases, lung injury, and central nervous system diseases [21][22][23]. Uniquely, Nox4 is an oxygen-sensing and constitutively active enzyme [24,25]. It is abundant in kidneys, vascular system, and many other tissues [26,27], suggesting a more general function of Nox4 than that of some other Nox members. Nox4-depedent ROS have been identified as a critical mediator in cell senescence induced by resveratrol, oncogene, and cytokine [28][29][30]. However, the expression of Nox4 in disc cells has not been determined. The role of Nox4 in disc cell senescence is unclear.
In this study, we investigated the effect of oxygen tension in 20% O 2 or 1% O 2 on disc cell senescence using rat NP cells. And then, the involvement of Nox4-dependent ROS in mediating the regulatory effect of oxygen tension on NP cell senescence was analyzed. At the same time, the association between Nox4 and IDD was investigated using the rat caudal IDD model produced by disc needle puncture. Furthermore, the role of MAPK and NF-κB pathways on the downstream of ROS in regulating NP cell senescence was also evaluated. Our study gives a novel insight into the mechanism of disc cell senescence and the pathogenesis of IDD. The abbreviations used in this study are listed in the supplementary materials (Supplementary Material, Table 1

Materials and Methods
2.1. Ethics Statement. This research was approved by the Ethical Committee of Xinqiao Hospital. All procedures described in this study were in accordance with the standards set forth in the eighth edition of Guide for the Care and Use of Laboratory Animals published by the National Academy of Sciences (Washington, DC, USA).

Rat NP Cell
Culture at 1% O 2 and High Oxygen Tension (20% O 2 ) Treatment. Rat NP cells were isolated as previously mentioned [31]. Isolated NP cells were cultured at 37°C and 1% O 2 . We cultured rat NP cells in low oxygen tension until they reached to the second or third passages. The period the cells remained in low oxygen tension was 4-5 weeks. To increase oxygen tension, NP cells were treated with 20% O 2 for 72 h. In order to scavenge ROS, NP cells were preincubated with GSH (1 mg/ml) and NAC (100 μg/ml) for 30 min before high oxygen treatment. Then, DMEM/F12 medium containing GSH or NAC was updated every 24 h. The cells used in this study were at the second or third passages.

Nox4
Overexpression. NP cells at 1% O 2 or 20% O 2 were transfected with pRP[EXP]-EGFP-EF1A>rNox4 vector (Cyagen, Santa Clara, CA, USA) using lipofectamine 2000. At 48 h after transfection, the expression of green fluorescent protein (GFP) in the transfected cells was observed under a microscope (Olympus, Tokyo, Japan). The expression of Nox4 was measured using RT-qPCR and immunoblot analysis.
2.6. Treatment of MAPK and NF-κB Inhibitors. The phosphorylation of ERK, JNK, p38, and p65 was investigated in NP cells treated with 20% O 2 for 10, 30, 45, 60, and 120 min using immunoblot analysis. In order to determine the role of Nox4-dependent ROS in the activation of MAPK and NF-κB, NP cells were pretreated with GSH, NAC, and siNox4 as mentioned above followed by high oxygen treatment. Furthermore, NP cells were pretreated with the p38 inhibitor (SB202190, 1 μg/ml), the JNK inhibitor (SP600125, 1 μg/ml), the ERK inhibitor (U0126, 1 μg/ml), and the NF-κB inhibitor (PDTC, 1 μg/ml) for 30 min followed by high oxygen treatment to determine the roles of MAPK and NF-κB pathways in regulating NP cell senescence.
2.7. SA-β-gal Staining. The SA-β-gal staining of NP cells was investigated using SA-β-gal staining kit (Cell Signaling Technology). Fixative NP cells seeded in 12-well culture plates were incubated with the staining solution at 37°C for 12 h. Nine random fields per well were imaged using a phasecontrast microscope (200x magnification, Olympus). The mean percentage of SA-β-gal-positive cells was calculated.
2.8. ROS Measurement. ROS levels in NP cells were analyzed using the Amplex Red Hydrogen Peroxide/Peroxidase Assay kit (Thermo Scientific, Waltham, MA, USA). 20 μl of NP cell suspension (1-1.5 × 10 4 ) was incubated with 100 μl of Amplex Red reaction mixture in 96-well microplate at 37°C for 10 min. The absorbance was measured at 560 nm using a spectrophotometer (Varioskan Flash, Thermo Scientific, Waltham, MA, USA). The concentration of ROS was calculated based on the absorbance according to the H 2 O 2 standard curve.
2.9. BrdU Incorporation. For BrdU incorporation, NP cells seeded in coverslips were incubated with DMEM/F12 medium containing BrdU (1 μg/ml, BD Biosciences, San Jose, CA, USA) for 24 h. After fixation with 4% paraformaldehyde, NP cells were incubated with 1 ml HCl (2 mol/l) under room temperature for 30 min. Then, the cells were used to perform the immunofluorescence assay.
2.10. Immunofluorescence Assay. The NP cells with BrdU incorporation were used to perform BrdU staining. For γ-H2A.X staining, NP cells seeded in coverslips were fixed with 4% paraformaldehyde. After permeabilization and antigen retrieval, NP cells were incubated with the primary antibodies against BrdU (1 : 500 dilution) and γ-H2A.X (1 : 400 dilution) overnight at 4°C. After rinsing, NP cells were incubated with the Alexa Fluor 647 dye-conjugated secondary antibodies (1 : 400 dilution) in the dark and then stained with DAPI (0.1 mg/ml, Sigma). The cells without primary antibody incubation served as the negative control. Three random fields per coverslip were imaged using a confocal microscope (200x magnification, Leica, Wetzlar, Germany). The mean percentages of BrdU-positive cells and γ-H2A.X-positive cells were calculated by using Image J software.
2.11. RT-qPCR. Total RNA was isolated from NP cells using TRIzol reagent (Takara Bio, Shiga, Japan) and quantitated using a NanoDrop ND-2000 spectrophotometer (Thermo Scientific, Waltham, MA, USA). 1 μg RNA was reverse transcribed into cDNA using a Prime Script RT reagent kit (Takara Bio). Real-time quantitative PCR was carried out in triplicate using a ViiA™7 Real-Time PCR system (Thermo Scientific). The SYBR® Premix Ex Taq™ II kit (Takara Bio) was used. The reaction condition was 95°C for 30 seconds, followed by 40 cycles of 95°C for 5 seconds and 60°C for 30 seconds. The relative expression of genes was calculated using the 2 −ΔΔCt method. GAPDH served as the internal reference gene. The primers used in the current study are listed in Table 1. The expression of matrix metalloproteinase 2 (MMP2), MMP3, MMP13, a disintegrin and metalloproteinase with thrombospondin motifs 4 (ADAMTS4), ADAMTS5, interleukin 6 (IL6), C-C motif ligand 2 (CCL2), CCL5, and C-X-C motif ligand 10 (CXCL10) in NP cells was analyzed to investigate the SASP of rat NP cells.

Rat Caudal IDD Model.
A total of ten eight-week-old male SD rats were used. The rats were randomized into two groups. Under anesthesia, caudal discs (C6-C9) were exposed through posterior incision. Disc puncture was performed using a 23-gauge needle on C7-C8 to cause IDD, as mentioned previously [19,32,33]. Untreated C6-C7 and C8-C9 served as the internal control. For the sham-operated control, caudal discs were exposed without puncture. After surgery, the rats were fed normally and monitored. MRI was performed 1 month after surgery. Caudal discs were harvested for H&E staining and immunohistochemistry analysis.
2.14. Immunohistochemistry Analysis. Deparaffinized IVD sections were incubated with primary antibodies against Nox4 (1 : 10 dilution), p21 (1 : 50 dilution), and Rb (1 : 50 dilution) overnight at 4°C followed by the incubation with HRP-conjugated secondary antibodies at room temperature. Nuclei were counterstained with hematoxylin. All sections were viewed under a microscope (Olympus). The integrated optical density (IOP) and area of sections were measured using Image-Pro Plus 6.0. The mean optical density (MOD, MOD = IOP/area) of sections was calculated to determine the expression intensity of targeted proteins.

Statistical Analysis.
All experiments were performed independently in three replicates at least. The results are presented as mean ± SEM. For comparisons between two independent groups, a two-tailed Student's t-test was used. Comparisons between three or more groups were tested by one-way ANOVA and LSD multiple comparisons. The results of RT-qPCR analysis and immunoblot analysis were statistically tested using Kruskal-Wallis nonparametric analysis and Mann-Whitney U post hoc tests. All results were analyzed using GraphPad Prism 6 and SPSS version 22.0 software programs. P < 0 05 was considered statistically significant.

High Oxygen Tension Caused Excessive ROS Production
to Induce NP Cell Senescence. 20% O 2 significantly enhanced ROS production in NP cells (Figure 1(a)). At the same time, the expression of methionine sulfoxide reductase A (MsrA), MsrB1, and MsrB2 in NP cells was upregulated by 20% O 2 (Figure 1(b)). Msr catalyzes the reduction of methionine residues in proteins to repair the oxidative damage of proteins. It has been recognized as an oxidative stress marker of disc cells [34]. High oxygen tension also increased the percentage of γ-H2A.X-positive cells, indicating an enhanced DNA damage induced by high oxygen tension  Figure S1). In addition, the catabolic and proinflammatory signaling in NP cells also was activated by high oxygen tension. The expression of MMP2, MMP3, MMP13, ADAMTS4, ADAMTS5, IL6, CCL2, CCL5, and CXCL10 in NP cells was markedly upregulated (Figures 2(f) and 2(g)). In summary, high oxygen tension induced premature senescence of NP cells.
In order to elucidate the role of ROS in high oxygen tension-induced premature senescence of NP cells, GSH and NAC were used. As a result, both antioxidants suppressed ROS production and expression of MsrA, MsrB1, and MsrB2 in NP cells treated with high oxygen tension (Figures 1(a) and 1(b)). The percentage of γ-H2A.X-positive cells was significantly reduced by the antioxidants (Figures 1(c) and 1(d)). Subsequently, NP cell senescence was retarded. Moreover, the inductive effect of high oxygen tension on the catabolic and proinflammatory phenotype of NP cells was antagonized by both antioxidants (Figure 2, Supplementary Material, Figure S1, S7A).
p38, and p65 showed an enhanced phosphorylation after high oxygen tension treatment for 15 min while ERK showed an enhanced phosphorylation at 45 min (Figure 3(a), Supplementary Material, Figure S8A). GSH and NAC suppressed the phosphorylation of p38, ERK, JNK, and p65 in NP cells subjected to high oxygen tension (Figure 3(b), Supplementary Material, Figure S8B). On the other hand, U0126 dramatically downregulated the expression of p53, p21, p16, and Rb in NP cells more potently than SP600125, SB202190, and PDTC ( Figure 3(c), Supplementary Material, Figure S7E). Consequently, the percentage of SA-β-gel-positive cells decreased, and the BrdU incorporation of NP cells increased   Figure S2). Noticeably, all signaling inhibitors suppressed the high oxygen tension-induced expression of matrix proteases and cytokines in NP cells (Figures 3(f) and 3(g)).

Nox4 Was a Critical Mediator in High Oxygen
Tension-Induced Premature Senescence of NP Cells. High oxygen tension-induced Nox4 expression in NP cells was prominently knockdown by siNox4 (Figures 4(a) and 4(b),     Supplementary Material, Figure S9A). Consequently, ROS production and Msr expression in NP cells were decreased (Figures 4(c) and 4(d)). The percentage of γ-H2A.X-positive cells declined (Figures 4(e) and 4(f)). Moreover, the phosphorylation of p38, ERK, JNK, and p65 in NP cells induced by high oxygen tension was markedly suppressed siNox4. Concomitantly, the expression of p53, p21, p16, and Rb in NP cells was downregulated (Figures 5(a) Figure S4) and showed an increased Nox4 expression (Figures 6(a) and 6(b), Supplementary Material, Figure S9B). Accordingly, an increased ROS production along with an upregulation of MsrB1 and MsrB2 was observed in the transfected cells (Figures 6(a) and 6(c)). As also, the percentage of     γ-H2A.X-positive cells increased (Figures 6(d) and 6(e)). The expression of p53, p21, p16, and Rb in the transfected cells was slightly upregulated at 1% O 2 and was markedly upregulated at 20% O 2 (Figures 7(a) and 7(b), Supplementary Material, Figure S7D), causing an increased SA-β-gel-staining and a decreased BrdU incorporation in NP cells (Figures 7(c), 7(d), and 7(e), Supplementary Material, Figure S5). Furthermore, the upregulation of matrix proteases and proinflammatory mediators in the transfected cells was determined by RT-qPCR (Figures 7(f) and 7(g)).
3.5. The Involvement of Nox4 in IDD. MRI and histological analysis demonstrated the degeneration of rat caudal discs induced by a puncture. The punctured discs showed lower T2 signal intensity, smaller NP area, disorganized AF, and enhanced connective tissue infiltration (Supplementary Material, Figure S6A, B). Notably, the MOD of Nox4 in the punctured NP was significantly higher than that in the sham group (Figures 8(a) and 8(d)). Also, the MOD of p21 and Rb markedly increased in the punctured NP (Figures 8(b)

Discussion
In the current study, we demonstrated that exposure of NP cells to high oxygen tension causes oxidative stress and DNA damage in NP cells. Consequently, high oxygen tension induced the cell cycle arrest of NP cells via the p53-p21-Rb and p16-Rb pathways. Also, it showed a catabolic and proinflammatory effect on the SASP of NP cells. These findings suggest that high oxygen tension is a potent trigger of disc cell senescence.
Noticeably, GSH and NAC significantly abrogated excessive ROS production in NP cells subjected to 20% O 2 . As a result, the DNA damage foci of NP cells were decreased, indicating the genotoxicity of ROS. However, the correlation between ROS and DNA damage is sophisticated. The involvement of DNA damage in ROS production of NP cells should be investigated in further studies. More importantly, the prosenescent of high oxygen tension on NP cells was prominently abolished by antioxidants, suggesting that ROS are critical mediators in high oxygen tension-induced premature senescence of NP cells. Interestingly, recent studies have been mentioned that oxidative stress essentially contributes to the establishment and progression of IDD [4,19,35]. Particularly, in aged or degenerative IVDs with neovascularization, disc cells inevitably were exposed to higher oxygen tension, leading to excessive ROS production that exerts detrimental effects on the viability and function of disc cells [14][15][16]. Undoubtedly, our study provides a strong support to this idea. High oxygen tension causes oxidative stress to suppress the proliferation of disc cells and to induce the expression of matrix protease and proinflammatory mediators by disc cells, which promotes the matrix remodeling and inflammation of degenerative discs. Thus, oxidative stress is a promising therapeutic target for IDD. Antioxidation is a potential approach for preventing disc cell senescence and disc degeneration. MAPK and NF-κB pathways are able to be activated by various stimuli in disc cells, including cytokines, growth factors, and high osmotic pressure. Consequently, physiological functions of disc cells are regulated to adapt to the stressful microenvironment [36,37]. As mentioned above, NP cells developed senescent signs in response to high oxygen tension. Therefore, we were interested in revealing the roles of MAPK and NF-κB pathways in high oxygen tensioninduced premature senescence of NP cells. Herein, we found that the activation of both pathways induced by high oxygen tension was suppressed by antioxidants, indicating that the pathways are the downstream signaling of ROS in NP cells. More importantly, the different roles of MAPK and NF-κB pathways in high oxygen tension-induced premature senescence of NP cells were determined using signaling inhibitors. On the one hand, the ERK inhibitor suppressed the expression of p53, p21, p16, and Rb in NP cells subjected to high oxygen tension, indicating that ERK activates the p53-p21-Rb and p16-Rb pathways to induce the cell cycle arrest of NP cells. On the other hand, the inhibitors of ERK, JNK, p38, and NF-κB were shown to suppress the catabolic and proinflammatory effect of high oxygen tension on NP cells, suggesting that both MAPK and NF-κB pathways are activated by ROS to regulate the catabolic and proinflammatory phenotype of NP cells. The involvement of MAPK and NF-κB pathways in regulating disc cell viability and function in response to stimuli is required to be investigated in detail.
ROS originate from oxygen-using metabolic processes. Mitochondrion is a major site of ROS production in disc cells [16,20]. A previous study has reported that increased oxygen tension (20% O 2 ) significantly boosts mitochondrial-derived ROS production in NP cells. ROS disturbed matrix homeostasis of disc cells [16]. In consistent with the findings, we also found that 20% O 2 prominently augmented ROS generation by NP cells. However, herein, we investigated nonmitochondrial ROS production through Nox4 in NP cells. Nox4 also is a major site of ROS production. Nox4-dependent ROS play crucial roles in regulating cell functions [26,27]. Nevertheless, there were no studies that elucidate the roles of Nox in the ROS production and senescence of disc cells. In this study, high oxygen tension was found to markedly upregulate the expression of Nox4 in NP cells, suggesting that Nox4 is an oxygen-sensing enzyme and is regulated by oxygen tension at transcriptional level in NP cells. Furthermore, Nox4 knockdown abrogated the high oxygen tensioninduced oxidative stress and DNA damage of NP cells. The activation of MAPK and NF-κB pathways induced by high oxygen tension was also suppressed by siNox4. As a result, NP cell senescence induced by high oxygen tension was retarded, suggesting an essential contribution of Nox4 to high oxygen tension-induced senescence of NP cells. To determine the involvement of Nox4 in NP cell senescence further in depth, NP cells were transfected with Nox4 vectors for Nox4 overexpression. The cells overexpressing Nox4 produced excessive ROS and rapidly developed senescent signs, although at 1% O 2 , indicating that Nox4 is a major source of ROS in NP cells. The upregulation of Nox4 boots oxygen utilization to generate more genotoxic ROS to potently induce premature senescence of disc cells.
The rat caudal IDD model caused by needle puncture has been widely used [32,33]. Therefore, we investigated the expression of Nox4 in degenerative discs in vivo using this model. As a result, the increased expression of Nox4 in degenerative NP tissues suggests the involvement of Nox4 in the pathogenesis of IDD. Furthermore, in consistent with Nox4, the expression of p21 and Rb in degenerative NP tissues also increased, providing a further in vivo evidence to support the association between Nox4 and NP cell senescence.
There are several limitations associated with our study. One limitation is that we investigated the effects of high oxygen tension and Nox4 on disc cell senescence based on rat disc cells. More studies based on human disc cells and IVD specimens should be performed to support our idea further. Secondly, although we discussed the involvement of Nox4dependent ROS in NP cell senescence, the contribution of mitochondria-mediated ROS to disc cell senescence remains to be elucidated in the future. Besides, except for MAPK and NF-κB pathways, more signaling pathways involved in regulating disc cell senescence should be discussed.

Conclusion
In summary, high oxygen tension markedly upregulates the expression of Nox4 to produce excessive ROS in NP cells. Oxygen-sensing Nox4-dependent ROS are genotoxic and activate the p53-p21-Rb and p16-Rb pathways via ERK signaling to induce cell cycle arrest of NP cells. Also, they induce the expression of matrix catabolic and proinflammatory genes in NP cells through MAPK and NF-κB pathways. Moreover, with IDD progression, the expression of Nox4 is increased along with the upregulation of p21 and Rb in NP cells. Our study gives a novel insight into the causes and molecular mechanism of disc cell senescence, contributing to our understating of the pathogenesis of IDD. Ameliorating