The NLRX1-SLC39A7 complex orchestrates mitochondrial dynamics and mitophagy to rejuvenate intervertebral disc by modulating mitochondrial Zn²⁺ trafficking

ABSTRACT

Intervertebral disc degeneration (IDD) is the most critical pathological factor in the development of low back pain. The maintenance of nucleus pulposus (NP) cell and intervertebral disc integrity benefits largely from well-controlled mitochondrial quality, surveilled by mitochondrial dynamics (fission and fusion) and mitophagy, but the outcome is cellular context-dependent that remain to be clarified. Our studies revealed that the loss of NLRX1 is correlated with NP cell senescence and IDD progression, which involve disordered mitochondrial quality. Further using animal and in vitro tissue and cell models, we demonstrated that NLRX1 could facilitate mitochondrial quality by coupling mitochondrial dynamic factors (p-DNM1L, L-OPA1:S-OPA1, OMA1) and mitophagy activity. Conversely, mitochondrial collapse occurred in NLRX1-defective NP cells and switched on the compensatory PINK1-PRKN pathway that led to excessive mitophagy and aggressive NP cell senescence. Mechanistically, NLRX1 was originally shown to interact with zinc transporter SLC39A7 and modulate mitochondrial Zn²⁺ trafficking via the formation of an NLRX1-SLC39A7 complex on the mitochondrial membrane of NP cells, subsequently orchestrating mitochondrial dynamics and mitophagy. The restoration of NLRX1 function by gene overexpression or pharmacological agonist (NX-13) treatment showed great potential for regulating mitochondrial fission with synchronous fusion and mitophagy, thus sustaining mitochondrial homeostasis, ameliorating NP cell senescence and rejuvenating intervertebral discs. Collectively, our findings highlight a working model whereby the NLRX1-SLC39A7 complex coupled mitochondrial dynamics and mitophagy activity to surveil and target damaged mitochondria for degradation, which determines the beneficial function of the mitochondrial surveillance system and ultimately rejuvenates intervertebral discs.

Abbreviations: 3-MA: 3-methyladenine; Baf-A₁: bafilomycin A₁; CDKN1A/p21: cyclin dependent kinase inhibitor 1A; CDKN2A/p16: cyclin dependent kinase inhibitor 2A; DNM1L/DRP1: dynamin 1 like; EdU: 5-Ethynyl-2’-deoxyuridine; HE: hematoxylin-eosin; IDD: intervertebral disc degeneration; IL1B/IL-1β: interleukin 1 beta; IL6: interleukin 6; MAP1LC3B/LC3B: microtubule associated protein 1 light chain 3 beta; MKI67/Ki67: marker of proliferation Ki-67; LBP: low back pain; MMP: mitochondrial membrane potential; MFN1: mitofusin 1; MFN2: mitofusin 2; MFF: mitochondrial fission factor; NP: nucleus pulposus; NLRX1: NLR family member X1; OMA1: OMA1 zinc metallopeptidase; OPA1: OPA1 mitochondrial dynamin like GTPase; PINK1: PTEN induced kinase 1; PRKN: parkin RBR E3 ubiquitin protein ligase; ROS: reactive oxidative species; SASP: senescence-associated secretory phenotype; SA-GLB1/β-gal: senescence-associated galactosidase beta 1; SO: safranin o; TBHP: tert-butyl hydroperoxide; TP53/p53: tumor protein p53; SLC39A7/ZIP7: solute carrier family 39 member 7; TOMM20: translocase of outer mitochondrial membrane 20; TIMM23: translocase of inner mitochondrial membrane 23.

KEYWORDS:

• Intervertebral disc degeneration
• mitochondrial dynamics
• mitophagy
• NLRX1
• nucleus pulposus
• SLC39A7

Introduction

Low back pain (LBP) is the leading health condition that contributes to years lived with disability and need for rehabilitation services [1,2]. In addition to affecting elderly patients, LBP is also commonly observed in children and adolescents [3]. Consistent with the prevalence of LBP in younger populations, studies have reported that the onset of intervertebral disc degeneration (IDD) occurs as early as adolescence and described IDD as a major contributor to the pathogenesis of LBP [4,5]. Decreases in cellular number and function in central gelatinous nucleus pulposus (NP) tissues initiates the disruption of intervertebral disc integrity. However, the pathological and molecular mechanisms are poorly understood, which limits the identification of effective therapeutic targets.

In addition to being the cellular energy source, mitochondria play critical roles in calcium homeostasis, metabolic and redox balance, organelle communication and stress resistance [6]. Disrupted mitochondrial function has been associated with a broad variety of health disorders [7,8], including IDD [9,10]. It has been concluded that NP cell survival, proliferation and extracellular matrix generation rely largely on health mitochondrial quality controlled by mitochondrial surveillance system [9,10], during which mitochondrial dynamics and mitophagy are key aspects. Mitophagy, which mediates the selective removal of defective or superfluous mitochondria, is essential for maintaining mitochondrial quality. Targeting manipulation on mitophagy holds attractive therapeutic potential for various degenerative and inflammatory diseases [7,8,11]. Similarly, the coordinated fusion and fission dynamics also implement widespread benefits on mitochondria by morphologic control, content exchange, equitable inheritance, high-quality mitochondrial DNA maintenance and damaged mitochondria division and degradation [12].

Although enhanced mitophagy has demonstrated therapeutic potential for protecting NP cells against various insults [13,14], excessive mitophagy can actually exacerbate NP cell death and senescence, ultimately promoting disc degeneration [15,16]. Similarly, aberrantly activated mitochondrial fission are also involved in the pathogenesis of IDD [17]. Indeed, it was shown that whether mitochondrial dynamics and mitophagy exert beneficial or detrimental effects in organismal health is determined in real-time by cellular factors and microenvironments that influence the mitochondrial quality surveillance system [18–20]. Briefly, the suitable fission of damaged mitochondria, which promotes mitophagy, would be beneficial for mitochondrial health and tissue integrity [21], while defective or excessive mitochondrial fission could promote widespread mitochondrial disruption and ultimately lead to detrimental mitophagy [22,23]. Thus, well-regulated mitochondrial dynamics and mitophagy are crucial for the beneficial health effects of mitophagy, and the precise regulatory mechanisms that maintain intervertebral disc integrity need to be further elucidated.

NLR family proteins are intracellular sensors that sense pathogen- and danger/damage-associated molecular patterns. NLRX1 is the only member of the NLR family that localizes to the mitochondria, and it has an N-terminal effector domain that contains a mitochondrial targeting sequence, which suggests the fundamental role of NLRX1 in regulating mitochondrial function and cellular physiology. Indeed, NLRX1 can organize mitophagic machinery directly by interacting with MAP1LC3B and driving MAP1LC3B lipidation [24,25] or indirectly by promoting the assembly of the ATG12–ATG5-ATG16L1 autophagy complex and activating FUNDC1‐NIPSNAP1/NIPSNAP2 mitophagic signaling [26,27]. In addition, studies have demonstrated the role of NLRX1 in regulating mitochondrial fission and mitochondria-lysosomal crosstalk [28,29], but its exact roles in organizing the mitochondrial quality surveillance system and controlling intervertebral disc integrity, as well as the underlying molecular mechanisms, are still unclear.

In this study, we emphasized the NLRX1-regulated network that coordinates mitochondrial dynamics and mitophagy for selectively eliminating damaged mitochondria and further ameliorates NP cell senescence during intervertebral disc rejuvenation. Utilizing human NP tissues and a rat model of IDD, we systemically analyzed the potential relationship between NLRX1 expression and IDD progression, and we found that loss of NLRX1 led to global mitochondrial fragmentation and excessive mitophagy via the compensatory PINK-PRKN pathway. Importantly, it was originally shown that NLRX1 interacted with the Zn²⁺ transporter SLC39A7 to modulate mitochondrial Zn²⁺ trafficking, which is a critical determinant of selective mitochondrial fission and beneficial mitophagic activity. The restoration of NLRX1 function by genetic overexpression or selective agonist (NX-13) treatment successfully alleviated IDD progression. Collectively, our findings highlight that the beneficial outcomes of mitophagy are determined by specific molecular settings, at least by the NLRX1-SLC39A7 complex in NP cells. These findings reveal a potential therapeutic target for the efficient intervention of IDD progression.

Results

Degenerated human NP tissues and rat discs exhibit downregulated NLRX1 expression and exacerbated senescent phenotypes

Mitochondrial damage is a critical factor in NP cell senescence and IDD. NLRX1, which is the only Nod-like receptor that localizes to mitochondria, plays various roles in regulating mitochondrial function. Here, we explored the potential relationship between NLRX1 levels and IDD pathogenesis in human NP tissue specimens and a rat model of IDD. As shown, NP tissues with different grades of degeneration, which were determined according to the Pfirrmann grading system, were collected from patients with spinal fusion surgery (Figure 1A). Alcian blue staining was also used to reveal the degenerative patterns in NP tissue specimens, defined as decreased blue staining for proteoglycan and collagen contents with increased degenerative levels (Figure 1B). We next assessed the levels of cell senescence markers in the corresponding NP tissues. Immunohistochemical analysis demonstrated that the protein expression levels of CDKN2A/p16 were significantly increased in degenerated NP tissues, while MKI67/Ki67 protein levels were decreased (Figure 1C). Furthermore, we examined the protein expression levels of NLRX1 and observed decreased positive staining in degenerated NP tissues (Figure 1C). Importantly, linear regression analysis revealed that the NLRX1 levels, as estimated by densitometric analysis of immunostaining, were negatively correlated with the CDKN2A levels but positively correlated with the MKI67 levels (Figure 1D,E). These data suggested that NLRX1 is aberrantly downregulated in degenerated human NP tissues and its expression is closely correlated with NP cell senescence and exacerbated disc degeneration.

Figure 1. Downregulation of NLRX1 correlates with aggravated human NP cell senescence and IDD progression. Human NP tissue specimens with different degenerative grades were collected for histological analysis. (A) Representative MRI images at T2 weight sequence were evaluated by Pfirrmann grading system. II: grade II, III: grade III, IV: grade IV. (B) histological analysis of human NP samples by alcian blue staining, scale bar: 100 μm. (C) immunohistochemical staining of CDKN2A, MKI67 and NLRX1 in different degenerative NP tissues, scale bar: 100 μm. (D and E) linear regression analyses of the tissue staining intensity of CDKN2A and that of NLRX1 (D), or the intensity of MKI67 and that of NLRX1 (E). AOD, average optical density. (F-H) protein expressions of senescence indicators (TP53, CDKN1A, CDKN2A), SASP factors (IL1B, IL6) and NLRX1 in primary human NP cells isolated from different degenerative NP tissues with the treatment of TBHP (100 μM), as determined by western blotting. (I-L) cell senescence (SA-GLB1/β-gal staining), cell proliferation (EdU incorporation) and NLRX1 expression (immunofluorescent staining) in primary human NP cells isolated from different degenerative NP tissues with the treatment of TBHP (100 μM), scale bar: 100 μm. (M and N) MRI examination, hematoxylin and eosin (HE) and safranin-O (SO) staining in sham or operation-induced degenerated disc of rat, scale bar: 500 μm (left panel), 50 μm (right panel). (O and P) immunohistochemical staining of aggrecan, collagen type II, NLRX1 and CDKN2A in sham or operation-induced degenerated disc of rat, scale bar: 500 μm (left panel), 50 μm (right panel). Data are represented as mean ± SD. *p < 0.05, **p < 0.01.

To further verify this finding, we isolated primary NP cells from human NP tissues with different Pfirrmann grades and then induced oxidative stress (100 μM TBHP), which is an important factor in IDD progression. Consistent with the findings in the tissue specimens, more senescent cells were observed among NP cells from degenerated discs, as shown by the increased levels of senescence markers (TP53/p53, CDKN1A/p21, and CDKN2A/p16) and proteins related to the senescence-associated secretory phenotype (SASP; IL1B/IL-1β and IL6) by western blotting (Figure 1F–H). Similarly, increased numbers of senescent cells were also observed among NP cells from degenerated discs, as shown by the increased number of SA-GLB1/β-gal-positive cells and the decreased number of EdU-positive cells; these results indicated that NP cells are prone to senescence when exposed to stress (Figure 1I–K). Notably, NLRX1 expression was lower in NP cells from degenerated discs, according to western blotting and immunofluorescence staining (Figure 1F,I,L). Moreover, we established an IVD degenerative model in rats using needle puncture. As expected, MRI imaging analysis and HE/SO staining confirmed the presence of degenerative changes in the punctured groups (Figure 1M,N). The degenerative and senescent phenotypes were also observed by the decreased expression of the anabolic matrix proteins aggrecan and collagen type II, the increased expression of CDKN2A, and the decreased expression of NLRX1 (Figure 1O,P). Collectively, these findings suggest that NLRX1 loss and NP cell senescence are critical features in IDD.

Restoration of NLRX1 attenuates NP cell senescence and rejuvenates intervertebral disc

To further characterize the function of NLRX1 in NP cell senescence and IDD progression, we knocked down the expression of NLRX1 in human NP cells with siRNA, the efficiency of which was determined by western blotting and RT-qPCR (Figure 2A). Compared with wild-type NP cells, NLRX1 deficiency exacerbated the predisposition to cell senescence under oxidative stress conditions. As shown, NLRX1 knockdown significantly increased SA-GLB1/β-gal activity in NP cells and correspondingly decreased the proportion of EdU-positive cells, indicating reduced proliferation (Figure 2B–D). The western blotting results also confirmed these findings by showing increased expression of senescence indicators (TP53, CDKN1A, and CDKN2A) and SASP factors (IL1B) (Figure 2E,F). Furthermore, overexpression of NLRX1 abrogated the senescence of NP cells isolated from degenerated human NP tissues, as shown by decreased SA-GLB1/β-gal-positive staining and decreased expression of senescence indicators (TP53, CDKN1A, and CDKN2A) and SASP factors (IL1B and IL6) (Figure 2G–J). Similarly, increased EdU-positive staining and increased MKI67 expression also demonstrated the ability of NLRX1 to rescue cell proliferation in degenerated discs (Figure 2G–J). In the rat model of IDD, we observed that the restoration of NLRX1 function by lentiviral plasmid injection significantly alleviated the degenerative patterns of intervertebral discs and reconstituted disc morphology, as reflected by MRI imaging and histological analysis (Figure 2K). These data suggested that NLRX1 is critical for maintaining NP cell proliferation and the therapeutic restoration of its function has great potential for reversing NP cell senescence and disc degeneration.

Figure 2. NLRX1 promotes NP cell proliferation, attenuates cell senescence and senescence-associated secretory phenotypes (SASPs) and rescues disc degeneration. Primary human NP cells isolated from NP tissues with different Pfirrmann grades were prepared. (A) knockdown efficiency of NLRX1 in NP cells by siRNA, analyzed with western blotting and RT-qPCR. (B-D) cell proliferation (EdU incorporation) and cell senescence (SA-GLB1/β-gal staining) in primary human NP cells isolated from health NP tissues treated by PBS or TBHP with NLRX1 knockdown or not, scale bar: 100 μm. (E and F) protein expressions of senescence indicators (TP53, CDKN1A, CDKN2A), SASP factors (IL1B) in primary human NP cells isolated from health NP tissues treated by PBS or TBHP with NLRX1 knockdown or not, as determined by western blotting. (G and H) protein expressions of senescence indicators (TP53, CDKN1A, CDKN2A), SASP factors (IL1B) and cell proliferation (MKI67) in primary human NP cells isolated from degenerated NP tissues with NLRX1 overexpression or not, as determined by western blotting. (I and J) cell proliferation (EdU incorporation) and cell senescence (SA-GLB1/β-gal staining) in primary human NP cells isolated from degenerated NP tissues with NLRX1 overexpression or not, scale bar: 100 μm. (K) rat disc degenerative models treated with NLRX1 overexpression and histologic analysis, upper panel: HE and so staining, lower panel: immunohistochemical staining of collagen type II, CDKN2A and MKI67, scale bar: 500 μm (left panel), 50 μm (right panel). Data are represented as mean ± SD. *p < 0.05, **p < 0.01.

NLRX1 is essential for the beneficial effects of mitophagy in NP cells to alleviate mitochondrial dysfunction and cell senescence

Intervertebral disc integrity requires good regulation of mitochondrial function via the mitochondrial quality surveillance system. Mitophagy is a crucial step for eliminating damaged mitochondria, which requires cooperation between mitophagy and mitochondrial fission. To further elucidate the mechanisms by which NLRX1 functions in NP cell senescence, we explored its involvement in mitophagy and mitochondrial dynamics. Indeed, we revealed that NLRX1 could function as an autophagic receptor for cargo recognition in human NP cells by interacting with MAP1LC3B, which was enhanced by oxidative stress (100 μM TBHP) (Figure 3A,B). Interestingly, it seems that NLRX1 is dispensable for the induction of mitophagy. In degenerated NP cells, it was observed that in both the wild-type and NLRX1-overexpression groups, TBHP treatment could induce the increased colocalization of MAP1LC3B and TOMM20 (Figure 3C), indicating enhanced mitophagic activity. The western blotting results also showed that TBHP treatment induced a significant increase in MAP1LC3B-II, which is an autophagosome marker, and this was accompanied by a decrease in TOMM20 and TIMM23, which are two mitochondrial proteins, in both groups (Figure 3D,E). Moreover, increased mitophagy was also observed in wild-type or NLRX1-knockdown NP cells from healthy discs, as shown by confocal colocalization analysis of MAP1LC3B and TOMM20 and western blotting analysis of mitophagy marker expression (MAP1LC3B-II, TOMM20, TIMM23) (Fig. S1A, B).

Figure 3. NLRX1 is essential for the beneficial action of mitophagy in alleviating mitochondrial dysfunction and NP cell senescence. Primary human NP cells isolated from NP tissues with different Pfirrmann grades were prepared. (A) confocal analysis of NLRX1 and MAP1LC3B with if staining in NP cells isolated from health NP tissues treated by PBS or TBHP, scale bar: 10 μm. (B) proteins immunoprecipitated (IP) from NP cells isolated from health NP tissues treated by PBS or TBHP followed by western blotting, left panel: IP with anti-MAP1LC3B antibody, right panel: IP with anti-NLRX1 antibody. (C) confocal analysis of TOMM20 and MAP1LC3B with if staining in NP cells isolated from degenerated NP tissues treated by PBS or TBHP with NLRX1 overexpression or not, scale bar: 10 μm. (D and E) protein expressions of mitophagy indicators (MAP1LC3B-II, TOMM20, TIMM23) in primary human NP cells isolated from degenerated NP tissues treated by PBS or TBHP with NLRX1 overexpression or not, as determined by western blotting. (F-I) JC-1 incubation for detecting mitochondrial membrane potential (MMP; F, H) and DCFH incubation for detecting reactive oxidative species (ROS; G, I) in NP cells isolated from degenerated NP tissues treated by PBS, TBHP or 3-MA with NLRX1 overexpression or not, as determined by flow cytometry. (J and K) protein expressions of senescence indicators (TP53, CDKN2A) and SASP factors (IL1B, IL6) in NP cells isolated from degenerated NP tissues treated by PBS, TBHP or 3-MA with NLRX1 overexpression or not, as determined by western blotting. (L and M) cell proliferation (MKI67 immunofluorescent staining, EdU incorporation) and cell senescence (SA-GLB1/β-gal staining) in primary human NP cells isolated from degenerated NP tissues treated by PBS, TBHP or 3-MA with NLRX1 overexpression or not, scale bar: 50 μm (IF images), 100 μm (white light images). Data are represented as mean ± SD. *p < 0.05, **p < 0.01.

To further elucidate the relationships of NLRX1 and mitophagy activation in the regulation of human NP cell senescence, we administered 3-methyladenine (3-MA) to wild-type and NLRX1-overexpressing NP cells from degenerated discs to block autophagosome formation. As shown, 3-MA treatment significantly decreased the ratio of MAP1LC3B-II to MAP1LC3B-I and blocked the degradation of the TOMM20 and TIMM23 proteins (Fig. S1C, D). The TBHP-induced loss of mitochondrial membrane potential (MMP) was also significantly alleviated by 3-MA treatment in wild-type NP cells (Figure 3F,H). Interestingly, the loss of MMP in NLRX1-overexpressing NP cells was not obvious, but it was exacerbated by 3-MA treatment (Figure 3F,H). The results of reactive oxidative species (ROS) assessment also demonstrated the beneficial effect of 3-MA on mitochondrial stress in wild-type NP cells, but it exerted a detrimental effect in NLRX1-overexpressing NP cells (Figure 3G,J). Consistently, senescent events were exacerbated in NLRX1-overexpressing NP cells from degenerated discs after autophagy was blocked by 3-MA, but senescence was mitigated in wild-type NP cells from degenerated discs, as shown by western blotting (TP53, CDKN2A, IL1B, and IL6) and cell staining (MKI67, EdU, and SA-GLB1/β-gal) (Figure 3J–M). In addition, we further established wild-type or NLRX1-deficient NP cells after their isolation from healthy discs. As shown, the inhibition of autophagy by 3-MA exacerbated MMP loss and ROS generation in wild-type NP cells and mitigated MMP loss and ROS generation in NLRX1-knockdown NP cells (Fig. S1E, F). Similar results were observed with MKI67 staining and western blotting analyses (Fig. S1G, H). Taken together, these data indicate that the induction of mitophagy could be induced in NP cells with or without NLRX1 function, but NLRX1 function is essential for promoting the beneficial effects of mitophagy for alleviating mitochondrial dysfunction and cell senescence.

NLRX1 sustains adaptive mitochondrial morphology by inducing selective mitochondrial fission and mitophagy

Intact autophagic flux is the main determinant of the biological consequence of the induction of autophagy as well as mitophagy [30]. To further elucidate the molecular mechanism underlying NLRX1-mediated beneficial mitophagy, we first examined the status of the autophagic flux using NP cells stably expressing tandem mCherry-GFP-MAP1LC3B, which were used to analyze autophagic activation or autophagic deposition. As shown, TBHP treatment (100 μM) induced punctate MAP1LC3B distribution in both wild-type and NLRX1-overexpressing NP cells (Figure 4A). Treatment with bafilomycin A₁ (Baf-A₁), which is a known autophagic flux inhibitor, induced the formation of yellow puncta, and the puncta structures of the MAP1LC3B protein in TBHP-treated NP cells were predominantly red; these results indicated the initiation of autophagosome formation and normal autophagic flux (Figure 4A). In parallel, the western blotting results demonstrated increased MAP1LC3B-II:MAP1LC3B-I ratios after TBHP treatment in both wild-type and NLRX1-overexpressing NP cells, and this ratio was further enhanced by Baf-A₁ treatment. Moreover, the increased degradation of the TOMM20 and TIMM23 proteins after TBHP treatment was blocked by Baf-A₁ (Figure 4B,C). In addition, normal autophagic flux was also observed in both wild-type and NLRX1-knockdown NP cells from healthy discs, as shown by fluorescence analysis of mCherry-GFP-MAP1LC3B and western blotting (Fig. S2A-C). Thus, the autophagic flux was almost unchanged regardless of NLRX1 expression levels in human NP cells.

Figure 4. NLRX1 induces selective mitochondrial fission and mitophagy to maintain adaptive mitochondrial morphology. Primary human NP cells isolated from degenerated NP tissues were prepared. (A) Co-localization analysis of mCherry and GFP in live stable mCherry-GFP-MAP1LC3B-expressing NP cells following the treatments of PBS, TBHP or Baf-A₁ with NLRX1 overexpression or not, scale bar: 200 μm. (B and C) protein expressions of mitophagy indicators (MAP1LC3B-II, TOMM20, TIMM23) in primary human NP cells isolated from degenerated NP tissues following the treatments of PBS, TBHP or Baf-A₁ with NLRX1 overexpression or not, as determined by western blotting. (D-F) mitochondrial morphology analysis by fluorescence microscope with MitoTracker Red CMXRos label (D and E) and transmission electron microscopy (TEM) (F) in primary human NP cells isolated from degenerated NP tissues following the treatments of PBS or TBHP with NLRX1 overexpression or not, fluorescent scale bar: 5 μm, TEM scale bar: 2 μm (upper panel), 500 nm (lower panel). (G and H) protein expressions of mitochondrial dynamics indicators (p-DNM1L, DNM1L, MFF, MFN1, MFN2, OPA1, OMA1) in primary human NP cells isolated from degenerated NP tissues following the treatments of PBS or TBHP with NLRX1 overexpression or not. (I and J) protein expressions of PINK1 and PRKN in primary human NP cells isolated from degenerated NP tissues following the treatments of PBS or TBHP with NLRX1 overexpression or not. (K) confocal analysis of MitoTracker labeling and PRKN protein with if staining in primary human NP cells isolated from degenerated NP tissues following the treatments of PBS or TBHP with NLRX1 overexpression or not, scale bar: 10 μm. Data are represented as mean ± SD. *p < 0.05, **p < 0.01.

Another factor that determines the consequence of autophagy is the initiation of autophagic cargo selection, which is referred to as mitochondrial dynamics in mitophagy. In wild-type NP cells from healthy discs, confocal microscopy analysis revealed normal elongated-tubular mitochondrial structures, which gradually shifted toward significantly punctuated structures under conditions of progressive oxidative stress (Fig. S2D, E). Correspondingly, the expression of mitochondrial fission- and fusion-related proteins was evaluated. The results showed that the p-DNM1L levels were significantly increased in TBHP-treated NP cells (Fig. S2F, G). Moreover, the mitochondrial protease OMA1 and its substrate OPA1, which regulate membrane fusion events, were expressed at increased levels, and more cleaved isoforms were present (Fig. S2F, G); these results indicated the transformation of selected fragments and global collapse in mitochondria as oxidative stress worsened. Regarding the role of NLRX1 in mitochondrial dynamics, we observed that low-level oxidative stress (100 μM TBHP) induced the transformation of mitochondrial morphology to punctuated structures in wild-type NP cells from degenerated discs, and this effect could be rescued by NLRX1 overexpression (Figure 4D,E). Consistently, transmission electron microscopy revealed that NLRX1 overexpression restored the elongated tubular mitochondrial network with condensed cristae compared to the NLRX1-deficient groups, which had swollen and round mitochondria with low crista integrity (Figure 4F). The western blotting results also demonstrated higher OMA1 expression and cleaved OPA1 isoforms in NP cells from degenerated discs after TBHP treatment (100 μM), and these effects were attenuated by NLRX1 overexpression, even though both groups showed increased p-DNM1L protein levels (Figure 4G,H). Increased unconnected and punctuated mitochondria with greater OMA1 and cleaved OPA1 expression were also observed after NLRX1 silencing (Fig. S2H-J), which further indicated the essential role of NLRX1 in maintaining mitochondrial morphology in NP cells. The classical nonreceptor-mediated pathway (PINK1-PRKN) showed crosstalk with and compensation for receptor-mediated mitophagy, especially under conditions of disrupted mitochondrial fusion and membrane potential depolarization [13,31]. Interestingly, we observed a significant increase in the protein levels of PINK1 and PRKN in wild-type NP cells from degenerated discs (Figure 4I–K), which had low NLRX1 levels (Figure 1F,I). In contrast, these changes were not obvious in the NLRX1 overexpression group (Figure 4I–K). A similar overall profile was observed after NLRX1 silencing in NP cells from healthy discs, as shown by significantly increased PINK1 and PRKN protein levels (Fig. S2K, L). Moreover, in wild-type NP cells from healthy discs, we observed that significant retention of the PINK1 and PRKN proteins was not induced until the cells were exposed to high-level oxidative stress (400 μM TBHP) (Fig. S2M), while mitophagy and the NLRX1-MAP1LC3B interaction were induced after low-level oxidative stress (TBHP 50–100 μM) (Figure 4B,C, Fig. S2K). These results suggested that NLRX1 maintained an adaptive mitochondrial morphology by inducing selective mitochondrial fission and mitophagy, and disrupted mitochondria with global fragmentation were subsequently degraded via the compensatory PINK1-PRKN-mediated mitophagy pathway.

The zinc transporter SLC39A7 interacts with NLRX1 in NP cells

To elucidate the mechanism by which NLRX1 regulates mitochondrial fission and mitophagy to promote NP cell function, mass spectrometry analysis was performed to identify proteins that were immunoprecipitated by an NLRX1 antibody from extracts of human NP cells treated with TBHP (100 μM) or the blank control (Figure 5A). The top 20 up- and downregulated proteins were analyzed and are listed; the zinc transporter SLC39A7 was the most highly ranked among these proteins (Figure 5A). The transportation of Zn²⁺, which is a divalent cation, through the mitochondrial membrane can regulate mitochondrial function, including membrane potential and mitochondrial dynamics; thus, SLC39A7 was considered to be a suitable candidate. Next, we sought to validate the physical interaction between NLRX1 and SLC39A7 in cells. In human NP cells, endogenous NLRX1 could be successfully coimmunoprecipitated by an anti-SLC39A7 antibody, and vice versa (Figure 5B). In addition, exogenous GFP-NLRX1 or HA-SLC39A7 proteins could be reciprocally precipitated from the 293T cell line (Figure 5C). Moreover, their actions in response to oxidative stress (100 μM TBHP) in human NP cells were also examined. As shown, the endogenous interaction of the NLRX1 and SLC39A7 proteins were decreased according to coimmunoprecipitation analysis (Figure 5D). Similarly, confocal microscopy analysis after immunofluorescence staining revealed fewer SLC39A7 puncta that colocalized with NLRX1 after TBHP treatment (Figure 5E). Thus, SLC39A7 May function as a novel NLRX1-interacting protein that regulates NP cell function under oxidative stress.

Figure 5. Zinc transporter SLC39A7 interacts with NLRX1 in NP cells. (A) the proteins that potentially interact with NLRX1 in human NP cells were immunoprecipitated with anti-NLRX1 antibody and analyzed by mass spectrometry. (B) endogenous protein immunoprecipitated (IP) from NP cells followed by western blotting, left panel: IP with anti-SLC39A7 antibody, right panel: IP with anti-NLRX1 antibody. (C) exogenous protein immunoprecipitated (IP) from 293T cells with plasmid transfection (mock GFP or HA, GFP-NLRX1, HA-SLC39A7) followed by western blotting, left panel: IP with anti-GFP antibody, right panel: IP with anti-HA antibody. (D) endogenous protein immunoprecipitated (IP) from NP cells isolated from health NP tissues following the treatments of PBS or TBHP, followed by western blotting, left panel: IP with anti-SLC39A7 antibody, right panel: IP with anti-NLRX1 antibody. (E) confocal analysis of NLRX1 and SLC39A7 with if staining in NP cells isolated from health NP tissues treated by PBS or TBHP, scale bar: 10 μm. Data are represented as mean ± SD. *p < 0.05, **p < 0.01.

SLC39A7 maintains mitochondrial Zn²⁺ homeostasis to alleviate mitochondrial damage in NP cells

To elucidate the potential roles of the protein complex, we first defined the function of SLC39A7 in NP cells. SLC39A7 is a membrane protein, and it localizes and functions in multiple organelles, including mitochondria. In human NP cells, immunofluorescence images that were obtained using confocal microscopy revealed the punctiform localization of SLC39A7 in the mitochondria by MitoTracker Red (Figure 6A). Subcellular fractionation and western blotting demonstrated that both endogenous SLC39A7 and exogenous HA-SLC39A7 could be detected in mitochondria (Figure 6B,C). Furthermore, small amounts of FluoZin3-AM-labeled Zn²⁺ were observed in elongated-tubular mitochondria, but SLC39A7 knockdown resulted in punctuated mitochondrial structures at baseline, exacerbated oxidative stress (100 μM TBHP)-induced mitochondrial fragmentation, and increased Zn²⁺ accumulation in punctuated mitochondria (Figure 6D); these results indicated a role for SLC39A7 in Zn²⁺ export from mitochondria. The mitochondrial accumulation of Zn²⁺, which is a divalent cation, can directly depolarize the mitochondrial membrane potential by diluting the proton electrochemical gradient. As expected, the decrease in the JC-1 aggregation:monomer ratio under oxidative stress (100 μM TBHP) was significantly exacerbated by SLC39A7 knockdown (Figure 6E), and similar results were observed for ROS generation (Figure 6F). Subsequently, SLC39A7 knockdown significantly enhanced mitophagy, which was partially dependent on the PINK1-PRKN pathway (Figure 6G,H). To further confirm the involvement of Zn²⁺ in mitophagy and mitochondrial function, we exposed NP cells to Zn²⁺ supplementation (ZnCl₂) and chelator (TPEN) pretreatment. As shown, ZnCl₂ alone induced mitochondrial damage (MMP loss, ROS generation) or exacerbated mitochondrial damage in human NP cells under oxidative stress conditions (100 μM TBHP) (Figure 6I,J). Additionally, PINK1-PRKN-mediated mitophagy was activated by ZnCl₂ supplementation (Figure 6K). In addition, TPEN pretreatment alleviated MMP loss and mitophagy in SLC39A7-knockdown NP cells under physiological conditions, but it exerted minimal effects on ROS production and the PINK1-PRKN pathway (Figure 6I–K). In contrast, the substantial mitochondrial damage that was observed under oxidative stress conditions (100 μM TBHP) was significantly inhibited by TPEN pretreatment, and PINK1-PRKN-mediated mitophagy was inhibited (Figure 6I–K). Therefore, SLC39A7 plays a crucial role in maintaining mitochondrial Zn²⁺ homeostasis, which facilitates the resistance of NP cells to oxidative stress.

Figure 6. SLC39A7 maintains mitochondrial zinc homeostasis to resist mitochondrial damage in NP cells. (A) confocal analysis of MitoTracker labeling and SLC39A7 protein with if staining in primary human NP cells isolated from health NP tissues, scale bar: 10 μm. (B) endogenous protein lysates extracted specially in mitochondria and cytoplasm of human NP cells isolated from health NP tissues, followed by western blotting. (C) exogenous protein lysates extracted specially in mitochondria and cytoplasm of human NP cells with plasmid transfection (mock HA, HA-SLC39A7), followed by western blotting. (D) zinc distribution in primary human NP cells isolated from health NP tissues following the treatments of PBS or TBHP with SLC39A7 knockdown or not, detected by FluoZin™-3-AM labelling and confocal fluorescence microscope, scale bar: 10 μm. (E and F) JC-1 incubation for detecting mitochondrial membrane potential (MMP; E) and DCFH incubation for detecting reactive oxidative species (ROS; F) in NP cells isolated from health NP tissues following the treatments of PBS or TBHP with SLC39A7 knockdown or not, as determined by flow cytometry. (G and H) protein expressions of mitophagy indicators (MAP1LC3B-II, TOMM20, TIMM23), PINK1 and PRKN in primary human NP cells isolated from health NP tissues following the treatments of PBS or TBHP with SLC39A7 knockdown or not. (I-N) primary human NP cells isolated from health NP tissues following the treatments of PBS, TBHP, ZnCl₂ and TPEN with SLC39A7 knockdown or not, followed by flow cytometry analysis of JC-1 (I, L) and DCFH (J, M) labelling and western blotting analysis of mitophagy indicators (MAP1LC3B-II, TOMM20, TIMM23), PINK1 and PRKN (K, N). Data are represented as mean ± SD. *p < 0.05, **p < 0.01.

NLRX1 promotes the mitochondrial localization of SLC39A7 that is essential for selective mitochondrial fission and mitophagy

To determine whether SLC39A7 participates in IVDD progression and NLRX1-mediated protection, we next examined the changes in SLC39A7 expression that occur with IVDD progression. In human NP specimens and isolated primary NP cells, we observed few changes in the SLC39A7 protein levels in samples with different Pfirrmann grades, and similar results were observed in the rat model of disc degeneration (Figure 7A,B, Fig. S3A, B). To further confirm the potential effect of NLRX1 loss, we measured the expression of SLC39A7 in NLRX1-deficient NP cells. Western blotting and immunofluorescence staining demonstrated no significant changes in NLRX1-deficient NP cells compared to wild-type cells (Fig. S3C). Due to the essential roles of mitochondrial localization in SLC39A7 function, we further determined whether NLRX1 could regulate the localization of SLC39A7 to mitochondria. As shown in Figure 6A, the punctiform SLC39A7 protein colocalized well with mitochondria in NP cells from healthy discs, and this colocalization was significantly reduced by NLRX1 knockdown (Figure 7C,D). In contrast, NLRX1 overexpression restored the localization of SLC39A7 to mitochondria in NP cells from degenerated discs (Figure 7C,E). Thus, NLRX1 is important for maintaining the localization of SLC39A7 in mitochondria. Similarly, the accumulated Zn²⁺ signal (FluoZin3) was also observed within punctuated mitochondrial structures in NLRX1-deficient NP cell (NLRX1-knockdown or degenerated NP cells), which could be recovered by NLRX1 overexpression (Fig. S3D). Due to the decreased SLC39A7-NLRX1 interaction after TBHP treatment, we introduced a novel NLRX1-targeting agonist, NX-13 [32], to further determine whether NLRX1 activation occurs along with SLC39A7 disaggregation. As shown, the amount of SLC39A7 protein precipitated by NLRX1 was significantly decreased after NX-13 administration (Figure 7F). Correspondingly, western blotting of mitochondrial and cytoplasmic lysates revealed that the SLC39A7 protein translocated from the mitochondria to the cytoplasm after NX-13 administration (Figure 7G). Moreover, immunofluorescence double staining of SLC39A7 and mitochondria also showed that their disassociation mainly occurred at the site of elongated-tubular structure fission (Figure 7H). Thus, these data indicated that the trafficking function of SLC39A7, which is regulated by NLRX1 activation, may be involved in mitochondrial quality surveillance.

Figure 7. NLRX1 participates in modulating SLC39A7 trafficking at mitochondria that is essential for selective mitochondrial fission and mitophagy. (A and B) the expression of SLC39A7 in human different degenerative NP tissues (A) and rat disc degenerative models (B) examined by IHC and IF, human tissue scale bar: 25 μm (upper panel), 50 μm (lower panel), rat scale bar: 500 μm (upper panel), 500 μm (lower panel). (C) confocal analysis of MitoTracker labeling and SLC39A7 protein with if staining in primary human NP cells isolated from health NP tissues with NLRX1 knockdown or not (upper two panel), or from degenerated NP tissues with NLRX1 overexpression or not (lower two panel), scale bar: 10 μm. (D and E) protein lysates extracted specially in mitochondria and cytoplasm of human NP cells isolated from health NP tissues with NLRX1 knockdown or not (D), or from degenerated NP tissues with NLRX1 overexpression or not (E), followed by western blotting. (F) protein immunoprecipitated (IP) using anti-NLRX1 antibody from NP cells treated by NX-13 or not, followed by western blotting. (G) protein lysates extracted specially in mitochondria and cytoplasm of human NP cells treated by TBHP, NX-13 or not, followed by western blotting. (H) confocal analysis of MitoTracker labeling and SLC39A7 protein with if staining in primary human NP cells treated by TBHP, NX-13 or not, scale bar: 10 μm. (I and J) mitochondrial morphology analysis by fluorescence microscope with MitoTracker Red CMXRos label and transmission electron microscopy (TEM) in primary human NP cells isolated from degenerated NP tissues following the treatments of PBS or TBHP with NLRX1 overexpression, SLC39A7 knockdown or not, fluorescent scale bar: 5 μm, TEM scale bar: 2 μm (left panel), 500 nm (right panel). (K) protein expressions of mitochondrial dynamics indicators (p-DNM1L, DNM1L, MFF, MFN1, MFN2, OPA1, OMA1) in primary human NP cells isolated from degenerated NP tissues following the treatments of PBS or TBHP with NLRX1 overexpression, SLC39A7 knockdown or not. Data are represented as mean ± SD. *p < 0.05, **p < 0.01.

To further elucidate the involvement of SLC39A7 in NLRX1-mediated protection, we then silenced SLC39A7 in NLRX1-overexpressing NP cells. We found that the elongated tubular structures and condensed cristae that had been restored by NLRX1 overexpression in degenerated NP cells were significantly disrupted by SLC39A7 knockdown (Figure 7I,J). In addition, higher levels of cleaved OPA1 isoforms and OMA1 were observed after SLC39A7 silencing (Figure 7K). Furthermore, SLC39A7 silencing inhibited the protective effects of NLRX1 overexpression on mitochondrial damage (MMP loss, and ROS generation) and NP cell senescence under oxidative stress conditions (100 μM TBHP) (Fig. S3E-H); these results indicated the essential roles of SLC39A7 in the NLRX1-mediated protection of mitochondrial quality and NP cell function.

Pharmacological activation of NLRX1 attenuates disc degeneration in an experimental rat model and human disc tissue culture

Given the importance of NLRX1 activation in modulating mitochondrial SLC39A7 and Zn²⁺ trafficking in NP cells, we further explored the therapeutic outcome of the pharmacological administration of NX-13, which targets the NLRX1 pathway. First, the in vitro results showed that the mitophagy levels in NP cells from degenerated discs were increased after NX-13 treatment, as shown by western blotting and immunofluorescence staining (Figure 8A,C). Additionally, NX-13 displayed remarkable therapeutic properties in alleviating NP cell senescence and SASP generation (Figure 8B,D). To further verify the effect of NX-13 in vivo, 200 ng/ml NX-13 was administered by intradisc injection every two weeks beginning at the establishment of the degenerative animal model. Treatment with NX-13 markedly preserved disc integrity (Figure 8E,F), sustained NP cell proliferation and collagen type II synthesis (Figure 8E,G), and reduced cell senescence (Figure 8E,G). Moreover, tissue culture of human NP cells was also used to evaluate the therapeutic potential of NX-13. We first used DCFH-DA loading to assess the oxidative conditions. The results revealed partial mitochondrial damage; accumulated DCF signals were observed in degenerated NP tissues, and this effect was reduced after NX-13 treatment (Figure 8H). Similarly, histochemical analysis demonstrated that NX-13 treatment significantly rescued NP cell proliferation (MKI67-positive staining) (Figure 8J,K), maintained functional extracellular matrix (Alcian blue staining) (Figure 8J,L), and decreased senescence phenotypes (CDKN2A-positive staining) (Figure 8J,M). Taken together, these data show that the NLRX1 pharmacological agonist NX-13 has potential for treating NP cell senescence and senescence-associated inflammation, and targeting NLRX1 provides an early intervention strategy for IDD therapy.

Figure 8. Pharmacological activation of NLRX1 by NX-13 attenuates disc degeneration in rat degenerative disc model and tissue culture of human disc. (A and B) protein expressions of mitophagy indicators (MAP1LC3B-II, TOMM20, TIMM23), senescence indicators (TP53, CDKN2A) and SASP factors (IL1B, IL6) in primary human NP cells isolated from degenerated NP tissues following the treatments of PBS or NX-13. (C) confocal analysis of TOMM20 and MAP1LC3B protein with if staining in primary human NP cells isolated from degenerated NP tissues following the treatments of PBS or NX-13, scale bar: 10 μm. (D) cell proliferation (MKI67 immunofluorescent staining) and cell senescence (SA-GLB1/β-gal staining) in primary human NP cells isolated from degenerated NP tissues following the treatments of PBS or NX-13, scale bar: 50 μm (IF images), 100 μm (white light images). (E-G) rat disc degenerative models treated with NX-13 and histologic analysis, upper panel: HE and so staining, lower panel: immunohistochemical staining of collagen type II, CDKN2A and MKI67, scale bar: 500 μm (left panel), 50 μm (right panel). (H and I) ROS detection by DCFH labelling in frozen section of human NP tissues in-vitro cultured by NX-13 or not, scale bar: 500 μm. (J-M) histologic analysis of human NP tissues in-vitro cultured by NX-13 or not with alcian blue staining or IHC staining (MKI67, CDKN2A), scale bar: 20 μm. Data are represented as mean ± SD. *p < 0.05, **p < 0.01.

Discussion

Cellular senescence and SASP-related inflammation are important pathological characteristics of NP tissue degeneration that disrupt intervertebral disc integrity. Mitochondrial dysfunction has been identified as a major contributor to the variability of NP cell senescent phenotypes in the context of different stress inducers and tissues [33,34]. NLRX1 is the only Nod-like receptor that localizes to mitochondria, and it plays an important role in sensing mitochondrial damage and regulating mitochondrial function [24,25,35]. Nonetheless, the intrinsic molecular mechanisms by which NLRX1 regulates the mitochondrial quality surveillance system remain largely unknown, as does its function in modulating intervertebral disc integrity. In the present study, we elucidated the essential function of NLRX1 in regulating mitochondrial homeostasis and alleviating NP cell senescence. Mechanistically, the zinc transporter SLC39A7, which is a novel NLRX1-interacting protein, was identified and demonstrated to govern mitochondrial dynamics to promote synchronous beneficial mitophagy. Pharmacologically targeting the NLRX1-SLC39A7 pathway showed great potential for intervertebral disc rejuvenation.

Mitophagy, which is a specialized form of selective autophagy that targets impaired mitochondria for degradation, is a pivotal mechanism for maintaining mitochondrial homeostasis [36]. Mitophagic deficiency is mostly observed in organismic disorders that correlate with a series of pathological processes, including cell senescence and aberrant inflammation, during which the restoration or induction of mitophagy by genetic or pharmacological techniques can be very beneficial [37,38]. During IDD progression, it has been revealed that various risk factors, such as high-magnitude loading and oxidative stress, can induce mitochondrial damage by suppressing mitophagy, which ultimately results in NP cell senescence. Thus, enhancing mitophagy is thought to be a potential therapeutic approach for alleviating IDD [39,40]. Nevertheless, this is not always the case. It was reported that an extended duration of mechanical loading induced excessive mitochondria removal by mitophagy, which exacerbated NP cell senescence, and the inhibition of mitophagy by cyclosporine A or PINK1-shRNA exerted positive effects [16]. Similarly, enhanced mitophagy was also observed in NP cells that were isolated from pathological tissue specimens after exposure to oxidative stress, and this enhanced mitophagy subsequently exacerbated NP cell dysfunction [15]. It has been concluded that whether mitophagy is beneficial or detrimental to health depends on cellular and microenvironmental factors [18]. Our present study showed that after exposure to oxidative stress, different biological outcomes were observed in NLRX1-overexpressing and NLRX1-deficient NP cells, even though mitophagy was consistently induced. This result indicated that NLRX1 was a critical determinant of the function of mitophagy in regulating intervertebral disc homeostasis. Therefore, our findings suggested that enhancing mitophagy is not always a good approach for treating IDD, and treatment decisions should be based on a clear understanding of the molecular context, partially relying on the presence of NLRX1.

Indeed, autophagy is a highly conserved and multistep process that sequesters cytoplasmic cargo in autophagosomes for subsequent degradation in acidic lysosomes; this process is called the autophagy flux, and any disorders in this process can disrupt the function of autophagy. Previous studies have demonstrated that in IDD, impaired mitochondrial autophagosome-lysosome fusion or lysosomal dysfunction that blocks the autophagy flux can induce detrimental autophagic consequences in the regulation of NP cell function [41,42]. Nevertheless, our experiment revealed normal autophagic flux in both NLRX1-overexpressing and NLRX1-deficient NP cells, indicating sustained ongoing autophagy. Regarding the role of autophagy initiation, it has been suggested that a strategy that precisely separates damaged parts and reestablishes healthy portions of mitochondria would provide a great foundation for facilitating the beneficial effects of mitophagy [43]. Indeed, this process is sustained by well-regulated mitochondrial fusion and fission machinery. Mitochondrial fission is performed by the mitochondrial divisome, which consists of the dynamin-like GTPase DNM1L and its membrane-associated adaptors [44]. Although DNM1L activation is essential for selective mitophagy and mitochondrial integrity [20,45], it must be considered that loss of DNM1L or excessive DNM1L activation exacerbates widespread mitochondrial damage or extensive mitochondrial fragmentation, which induce nonselective and excessive mitochondrial clearance and the failure of mitochondrial rejuvenation [23,46]. Moreover, the mitochondrial fusion machinery is equally well regulated and modulates mitochondrial morphology. The distribution of long and short OPA1 proteins, which is regulated by the i-AAA protease OMA1, plays an equally important role in regulating mitochondrial inner membrane fusion and mitochondrial cristae morphogenesis [47,48]. Modest OPA1 cleavage and the short isoform S-OPA1 participate in L-OPA1-mediated membrane fusion, which can be disrupted by aggressive OPA1 cleavage [49–51]. Our present study found that NLRX1 is crucial for the maintenance of mitochondrial integrity and morphology, even after exposure to oxidative stress. Regarding the in-depth mechanism, we observed higher expression of the mitochondrial protease OMA1 and its substrate OPA1with increased cleaved isoforms (S-OPA1), in NLRX1-deficient NP cells, even though p-DNM1L was increased in both NLRX1-overexpressing and NLRX1-deficient NP cells. Thus, published reports and our present findings suggest the presence of a novel network in which NLRX1 regulates mitochondrial fission and mitophagy to promote mitochondrial rejuvenation via collaborative mitochondrial fusion and fission machinery.

The functions of NLRX1 have been increasingly defined and extended due to its roles in mitochondria-associated inflammatory signaling [26], mitochondrial respiratory function [52], mitochondrial RNA processing [53], mitochondria-lysosomal crosstalk [29] and mitochondrial autophagy [24]. Specifically, NLRX1 can regulate mitophagic machinery directly by interacting with MAP1LC3B and driving MAP1LC3B lipidation [24,25] or indirectly by promoting the assembly of the ATG12–ATG5-ATG16L1 autophagy complex and inhibiting FUNDC1‐NIPSNAP1/NIPSNAP2 mitophagy signaling [26,27]. Even so, certain functions of NLRX1 itself or effects of NLRX1-regulated mitophagy are controversial [54–56]; such controversy is attributed to the variable and complex underlying molecular mechanisms. In human NP cells, we identified a novel interaction between the zinc transporter SLC39A7 and NLRX1, which modulated Zn²⁺ distribution in mitochondria and regulated NLRX1 function.

Indeed, increasing numbers of zinc transporters (SLC30A/ZnTs, SLC39A/ZIPs, and TRPMs) have been shown to function in mitochondria and regulate mitochondrial Zn²⁺ concentrations, thus playing important roles in maintaining mitochondrial homeostasis [57–60]. Zinc transporter SLC39A7 was initially shown to localize to the early secretory pathway, including the ER and Golgi apparatus, and control Zn²⁺ mobilization between the cytosol and intracellular vesicles. Recently, it has been suggested that SLC39A7 also localizes to mitochondria or is redistributed from other structures [61]. Pathological changes in SLC39A7 can exacerbate mitochondrial oxidative stress due to subsequent Zn²⁺ accumulation in mitochondria and defective mitophagy [62]. In the present study, we confirmed the important role of SLC39A7 in modulating the mitochondrial Zn²⁺ distribution and maintaining mitochondrial membrane polarization in human NP cells. Under oxidative stress conditions, the interaction between SLC39A7 and NLRX1 is lost, and we speculated that SLC39A7 loss occurs in degenerative discs. Surprisingly, the protein levels of SLC39A7 were not significantly changed in degenerated human NP tissues and the rat model of disc degeneration, which suggested that the loss of the SLC39A7-NLRX1 interaction may represent one response to stress. Indeed, we confirmed that NLRX1 is indispensable for maintaining SLC39A7 in the mitochondria of human NP cells. NLRX1 converts into active state to promote mitophagy, so we further used NX-13 to positively stimulate NLRX1 activation; we found that both the interaction of SLC39A7 with NLRX1 and its localization to mitochondria were decreased. Thus, mitochondrial SLC39A7, which is sustained by NLRX1, is crucial for maintaining the mitochondrial membrane potential, while the loss of this interaction was observed at sites of mitochondrial damage and may induce local mitochondrial depolarization without affecting the overall membrane potential. This phenomenon has been observed and defined as the fluctuation or flickering of the mitochondrial membrane potential to allow for the precise adaptation of mitochondria [46,47]. Moreover, the local fluctuation in the mitochondrial membrane potential largely contributes to regulating the function and interconnection of mitochondrial fission/fusion dynamics. In DNM1L-deficient cells, transient flickering of the membrane potential could activate the OMA1 protease and result in the proteolytic cleavage of OPA1, which induces compensatory fission and prevents the formation of a deleterious hyperfused mitochondrial network [47]. In physiological processes, a threshold of membrane potential loss is thought to balance fission/fusion dynamics. In brief, membrane potential above this threshold facilitates the presence of fusion-active L-OPA1 and maintains the mitochondrial fission/fusion balance, while membrane potential below this threshold activates both DNM1L-mediated fission and OMA1-mediated OPA1 cleavage and thus disrupts the mitochondrial dynamic balance [63]. Therefore, one model in which the NLRX1-SLC39A7 interaction participates in regulating the flickering of the mitochondrial membrane potential and subsequently initiating mitochondrial fission/fusion dynamic homeostasis is proposed.

In addition, the profile of mitochondrial membrane polarization and potential also determined the effects of mitophagy on mitochondrial function and cell fate [18,58]. As known, the interconnected mitochondria with electrochemical gradients can be well-organized by rapid diffusion of ions. In our present study, NLRX1-SLC39A7-mediated mitochondrial Zn²⁺ trafficking could control the distribution of mitochondrial membrane potential that facilitates NLRX1-mediated mitophagy and sustain intact mitochondrial morphology and oxidative homeostasis. This may be attributed to the efficient division of damaged mitochondria and mitophagic removal to prevent the accumulation of damage-associated molecular patterns [7]. In contrast, global loss of membrane potential and mitochondrial depolarization would undergo progressive structural collapse and extensive mitophagic degradation. It not only disturbs mitochondrial energy generation, but also weaken the ability of mitochondrial stress resistance, which ultimately resulted in cell senescence and lifespan shortening [18,19]. As known, mitochondrial depolarization was one key inducer of PINK-PRKN activation and subsequent degradation of damaged mitochondria [31,64]. In NLRX1- and SLC39A7-deficient NP cells, we observed that accumulated mitochondrial Zn²⁺ and aggravated disruption of mitochondrial membrane occurred and activated PINK-PRKN-mediated mitophagy, along with increased ROS levels and NP cell senescence. Thus, this partially revealed the complementary or crosstalk mechanisms between PINK-PRKN-dependent and -independent mitophagy pathways [25,65].

The identification of novel molecular contexts and subsequent biotherapeutics is one important route for studying the clinical management of IDD. Our findings provide key mechanistic insights into the NLRX1-SLC39A7 complex, which regulates mitochondrial dynamics and mitophagy to promote intervertebral disc rejuvenation (Figure 9). In addition, our findings suggest the significant clinical relevance of NLRX1 activation for IDD treatment. More studies on bioactive substances that target NLRX1 activation and the NLRX1-SLC39A7 complex may provide novel promising therapeutic approaches for IDD.

Figure 9. Schematic depicting the molecular mechanism through which NLRX1-SLC39A7 facilitates intervertebral disc rejuvenation via orchestrating mitochondrial dynamics and mitophagy. NLRX1 recruits zinc transporter SLC39A7 to mitochondrial membrane in order to maintain mitochondrial zinc homeostasis, that is essential for selective segregation of damaged mitochondria during mitochondrial dynamics and subsequent degradation by mitophagy. The coordinated network of mitochondrial dynamics and mitophagy is essential for mitochondrial homeostasis and disc rejuvenation. Conversely, IDD progression is characterized by decreased expression of NLRX1 in NP cells, which results in SLC39A7 loss and aberrant zinc accumulation in mitochondria. Consequently, excessive mitochondrial fission and mitophagy aggravate NP cell senescence and IDD progression.

Materials and methods

Human nucleus pulposus (NP) tissue samples

Clinical samples of human NP tissue were obtained from patients who underwent spinal fusion due to lumbar spinal stenosis, lumbar disc herniation or vertebral fracture. Magnetic resonance images (MRI) were collected from patient medical records and used to evaluate the degree of disc degeneration according to the MRI-based Pfirrmann grading system. Correspondingly, the patients were divided into three groups: patients with Pfirrmann Grade II (n = 6, mean age = 35.63 ± 12.48 years), patients with Pfirrmann Grade III (n = 6, mean age = 46.17 ± 16.39 years), and patients with Pfirrmann Grade IV (n = 6, mean age = 51.82 ± 9.43 years). Ethical approval was obtained from the Ethics Committee of Tongji Medical College, Huazhong University of Science and Technology (NO. S214). Informed consent was obtained from all the participants in this study.

Animal experiments

Sprague-Dawley rats (three-months old) were purchased from the Laboratory Animal Center of Huazhong University of Science and Technology (Wuhan, China). The surgical procedure was performed as previously described [66] and approved by the Animal Experimentation Committee of Huazhong University of Science and Technology. After anesthetization with 2% (w:v) pentobarbital (40 mg/kg), the coccygeal vertebrae (Co6/7, 7/8, and 8/9) of the rats were punctured with 29-G needles, which was confirmed by trial radiography. This surgery was performed twice a month to induce disc degeneration. In contrast, the discs of normal rats were analyzed as the “Sham” group. To evaluate the therapeutic effects of Nlrx1 gene overexpression or treatment with the pharmacological agent NX-13 (MedChemExpress, HY-141521) on disc rejuvenation, intradisc injection was performed along with disc puncture. Similarly, an equal volume of PBS (Boster Biological Technology, PYG0021) was injected into the untreated group of rats with disc degeneration. All the animals were housed under SPF conditions and allowed free, unrestricted weight bearing and activity. The rats were sacrificed 4 weeks after the surgery (n = 3 per group).

Human NP cell isolation and culture

NP tissues from patients in different Pfirrmann groups were collected and stored in Hank’s balanced salt solution for transportation. The NP tissues were cut into pieces and used to seed and expand NP cells via culture in Dulbecco’s modified Eagle’s medium supplemented with the F12 nutrient mixture and 15% fetal bovine serum (Gibco, 10099141C) at 37°C in 5% CO₂. Cells from the second passage were used in further experiments. NP cells from Pfirrmann Grade II tissues were considered to be healthy cells and were used to establish a degenerative cell model, while cells from Pfirrmann Grade IV tissues were considered to be severely degenerated cells and were used to examine therapeutic outcomes. In in vitro experiments, the degenerative cell model was established by exposing NP cells to tert-butyl hydroperoxide (TBHP, 0–400 μM; Sigma-Aldrich, A13926) for 4 h followed by culture under normal conditions. For the treatment of NP cells with some agents, NP cells were pretreated with 3-methyladenine (3-MA, 5 mM; MedChemExpress, HY-19312) or bafilomycin A₁ (Baf-A₁, 5 μM; MedChemExpress, HY-100558) or treated with ZnCl₂ (80 μM; Aladdin, 7646-85-7), TPEN (0.5 μM; MedChemExpress, HY-100202), or NX-13 (0.5 μM; MedChemExpress, HY-141521). To knock down NLRX1 or SLC39A7 expression, cells were transfected for 48 h with 100 nM NLRX1- or SLC39A7-targeting small-interfering RNA (siRNA) or scrambled siRNA (GENERAL BIOL, A130878/9, A130870/1) using Lipofectamine 2000 (Invitrogen 11,668,019), and then, the cells were treated with certain agents. In addition, exogenous protein overexpression was achieved by a lentiviral plasmid carrying the NLRX1 sequence (Lenti-NLRX1 with blank or GFP-tag), a lentiviral plasmid carrying the SLC39A7 sequence (Lenti-SLC39A7 with blank or HA-tag), or a flanking sequence control (Lenti-vector).

Histological and immunohistological analysis

The NP tissues from patient specimens or intact discs from model rats were immediately fixed using 4% paraformaldehyde overnight, washed with PBS, dehydrated in gradient ethanol solutions, and cleared with xylene before being embedded in paraffin. Tissue slices were generated from paraffin blocks at a thickness of 4 µm, and the histological features were examined by hematoxylin and eosin (HE), safranin-O (SO) and alcian blue staining. To identify certain proteins using immunological analysis, tissue sections were probed with the corresponding primary antibodies, including antibodies targeting CDKN2A (1:50; Abcam, ab51243), MKI67 (1:100; Affinity, AF0198), NLRX1 (1:100; Invitrogen, PA5–21018), ACAN/aggrecan (1:300; Proteintech 13,880–1-AP), COL2A1/collagen type II (1:1000; Proteintech 28,459–1-AP), and SLC39A7 (1:50; Proteintech 19,429–1-AP), at 4°C overnight. Then, the samples were washed thoroughly and labeled for 60 min at room temperature with the appropriate secondary antibody (diluted 1:200; Proteintech, PR30011 for rabbit and PR30012 for mouse). For immunofluorescence analysis, tissue sections were prepared in the same way as described above. After blocking, the sections were probed with primary antibodies against SLC39A7 (1:100; Affinity, DF4635) at 4°C overnight and incubated with an Alexa Fluor 568-conjugated anti-rabbit secondary antibody at room temperature for 1 h.

For cell slide staining, NP cells at 60% confluence were exposed to certain treatments and then fixed with 4% paraformaldehyde at room temperature (RT) for 30 min. After permeabilization and blocking, the cell slides were incubated with primary antibodies against NLRX1 (1:100; Affinity, DF12124), SLC39A7 (1:100; Affinity, DF4635), TOMM20 (1:100; Proteintech 11,802–1-AP), MAP1LC3B (1:100; Affinity, AF4650), and MKI67 (1:100; Affinity, AF0198) and incubated with Alexa Fluor 488-conjguated or Alexa Fluor 568-conjugated anti-rabbit secondary antibodies (Invitrogen, A-11008 for 488 and A-11011 for 568) at room temperature for 1 h. After washing with PBS, the nuclei were stained with DAPI (Beyotime, P0131-5 ml). The slides were observed using a laser scanning confocal microscope (Zeiss LSM800) or (Olympus, BX53).

Co-immunoprecipitation followed by mass spectrometry (MS)

Proteins that interact with NLRX1 were identified from among endogenous proteins that were harvested from human NP cells. NP cell lysates were coimmunoprecipitated with an anti-NLRX1 antibody or isotype antibody with protein A/G beads (MedChemExpres, HY-K0202) at 4°C overnight on a rotator. After washing in precooled NP-40 lysis buffer five times, the eluted solution was collected and analyzed by liquid chromatography – tandem mass spectrometry (LC‒MS/MS) on a Q Exactive HF-X mass spectrometer (Thermo Fisher Scientific). The LC‒MS/MS data were used to search the UniProt protein database with using MaxQuant (V1.6.2.10).

Western blotting analysis

Protein extracts from NP cells were prepared and collected with RIPA lysis buffer (Beyotime, P0013B) supplemented with protease and phosphatase inhibitors. Proteins were separated by sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis and transferred to polyvinylidene difluoride (PVDF) membranes. The transferred proteins were incubated with primary antibodies against NLRX1 (1:100; Invitrogen, PA5–21018), SLC39A7 (1:500; Affinity, DF4635), TOMM20 (1:3000; Proteintech 11,802–1-AP), TIMM23 (1:2000; Proteintech 11,123–1-AP), MAP1LC3B (1:1000; Affinity, AF4650), PRKN (1:1000; Proteintech 14,060–1-AP), PINK (1:1000; Proteintech 23,274–1-AP), MKI67 (1:1000; Affinity, AF0198), TP53/p53 (1:5000; Proteintech 60,283–2-Ig), CDKN1A/p21 (1:1000; Affinity, DF6423), IL6 (1:1000; Affinity, DF6087), IL1B (1:1000; Affinity, AF5103), p-DNM1L (1:2000; Affinity, DF2980), DNM1L (1:4000; Proteintech 12,957–1-AP), MFF (1:2000; Proteintech 17,090–1-AP), MFN1 (1:1000; Proteintech 13,798–1-AP), MFN2 (1:3000; Proteintech 12,186–1-AP), OPA1 (1:1000; BD Pharmingen 612,606), OMA1 (1:1000; Affinity, DF12435), and ACTB/actin (1:20000; Proteintech 66,009–1-Ig) overnight at 4°C, followed by incubation with diluted horseradish peroxidase (HPR)-conjugated Affinipure Goat Anti-Rabbit/Mouse IgG (Proteintech, SA00001–2/1). Finally, enhanced chemiluminescence reagents (Affinity, KF001) were used to visualize the protein expression using the Chemi-DocMP Imaging System (Bio-Rad, 12003154Hercules). Semiquantification of protein expression was performed with ImageJ.

EdU incorporation assay and GLB1/β-galactosidase staining

The proliferation of NP cells was examined by EdU labeling (25 × 10⁻⁶ M of 5-ethynyl-2′-deoxyuridine, EdU; RiboBio, C10338) for 12 h at 37°C, followed by fixation with 4% paraformaldehyde. After washing and permeabilization, NP cells were incubated with Apollo488 for 30 min, and subsequently, the nuclei were labeled with Hoechst 33,342 for 30 min. The cells were visualized and images were captured using a microscope (Olympus, BX53). At least 500 cells in random visual fields were counted for every group. For β-galactosidase staining, NP cells were fixed with the designated fixative for 15 min at room temperature. After washing, the cells were incubated with fresh staining solution at 37°C without CO₂ for 12 h. Images were captured with a microscope (Olympus, BX53). At least 500 cells in random visual fields were counted for every group.

MitoTracker staining and protein immunofluorescence

Following the indicated treatments, NP cells were treated with 200 nM MitoTracker Red (30 min, 37°C; Invitrogen, M7512) in complete medium. After washing with PBS, the NP cells were fixed in 4% paraformaldehyde at room temperature and then washed with PBS three times. The cell slides were imaged after DAPI labeling. Then, mitochondrial morphology (branch length) was analyzed using Fiji software through binary and skeleton according to previous literature [67]. For concomitant staining for certain proteins, the cell slides were permeabilized, blocked and incubated with primary antibodies (PRKN, 1:10, Proteintech 14,060–1-AP; SLC39A7, 1:100, Affinity, DF4635). After incubation with an Alexa Fluor 488-conjguated anti-rabbit secondary antibody at room temperature for 1 h, the cell slides were labeled with DAPI and observed using a laser scanning confocal microscope (Zeiss LSM800).

Imaging of mitochondria and intracellular Zn²⁺ distribution

Intracellular distribution of Zn²⁺, mainly in mitochondria and cytoplasm, was assessed in human NP cells by the labels of FluoZin™-3-AM (1 μM; Invitrogen, F24195) for Zn²⁺ and MitoTracker Red CMXRos for mitochondria. Human NP cell slides were washed with DMEM and incubated in the dark in DMEM containing 1 mM FluoZin™-3-AM for 4 hours at 37°C. The slides were then washed twice with DMEM and incubated with MitoTracker Red CMXRos (200 nM) diluted in DMEM for 30 min at 37°C. After washing with DMEM, cells were imaged using a laser scanning confocal microscope (Zeiss LSM800).

Mitochondrial membrane potential and reactive oxygen species analysis

Mitochondrial membrane potential (MMP) was measured by JC-1 staining (Beyotime, C2006) and analyzed by a FACSCalibur flow cytometer (BD Biosciences). The ratio of fluorescence intensity (red/aggregates:green/monomers) indicated the degree of mitochondrial membrane depolarization; a decreased ratio of red to green indicated the loss of MMP. Total reactive oxygen species (ROS) levels in NP cells were measured by staining with 2’,7’-dichlorofluorescin diacetate (DCFH-DA; Beyotime, S0033), which can be rapidly oxidized to generate highly fluorescent compounds. After labeling, the fluorescence was measured with a FACSCalibur flow cytometer (BD Biosciences). For ROS staining in tissues, labeled slides were restained with DAPI and imaged under a microscope (Olympus, BX53).

Transmission electron microscopy

NP cell samples were fixed in 2.5% glutaraldehyde (Sigma‒Aldrich, USA) for 1 h followed by 2% osmium tetraoxide for 2 h. After washing with water, the cell samples were stained with 0.5% uranyl acetate for 12 h, followed by dehydration and polymerization. Ultrathin sections were cut at thicknesses of 70–90 nm with an ultramicrotome (EM UC7, Leica) and imaged with a Tecnai G2 TWIN transmission electron microscope (FEI, USA).

Statistical analysis

All the data are presented as the mean ± SD of at least three independent experiments. Two-tailed Student’s t test and one-way or two-way ANOVA were used to assess the statistical significance of the differences. Linear regression analysis was used to test the possible relationships between two parameters. Statistical analysis was performed with GraphPad Prism 8.0 software, and P < 0.05 was considered statistically significant.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

All data supporting the findings of this study are available from the corresponding author on reasonable request.

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/15548627.2023.2274205.

Additional information

Funding

The work was supported by the National Natural Science Foundation of China [82130072, 82072505]; National Natural Science Foundation of China [81902259]; National Natural Science Foundation of China [81902260].