UCHL3 drives ferroptosis in nucleus pulposus cells by deubiquitinating HMGB1 to activate NCOA4-mediated ferritinophagy

ABSTRACT

Ferroptosis contributes to intervertebral disc degeneration (IVDD), yet its upstream regulatory mechanisms in nucleus pulposus cells (NPCs) remain unclear. Here, we identify a novel UCHL3-HMGB1-NCOA4 signaling axis that drives NPC ferroptosis. Transcriptomic and clinical analyses reveal that UCHL3 is upregulated in degenerative discs, with its expression correlating positively with both IVDD severity and ferroptosis markers. Mechanistically, UCHL3 deubiquitinates and stabilizes HMGB1 by cleaving Lys48-linked polyubiquitin chains, thereby promoting the accumulation of cytoplasmic HMGB1. This stabilized HMGB1 activates NCOA4-mediated ferritinophagy, initiating ferroptosis. Silencing UCHL3 or NCOA4 mitigates these effects. Furthermore, a polydopamine-based nanoplatform (PDA@si-UCHL3) achieves dual therapeutic action, silencing UCHL3 while scavenging ROS and chelating iron, which effectively inhibits ferroptosis and alleviates IVDD in rats. These findings establish UCHL3 as a pivotal regulator of NPC ferroptosis and highlight PDA@si-UCHL3 as a promising therapeutic strategy for IVDD.

Abbreviations: AAV: adeno-associated virus; CCK-8: cell counting kit-8; Co-IP: co-immunoprecipitation; DUBs: deubiquitinating enzymes; ECM: extracellular matrix; GPX4: glutathione peroxidase 4; GSH: glutathione; HE: hematoxylin and eosin; HMGB1: high mobility group box 1; IF: immunofluorescence; IHC: immunohistochemistry; IP/MS: immunoprecipitation mass spectrometry; IVDD: intervertebral disc degeneration; IVDs: intervertebral discs; MRI: magnetic resonance imaging; NCOA4: nuclear receptor coactivator 4; NP: nucleus pulposus; NPCs: nucleus pulposus cells; PDA: polydopamine; PG: Pfirrmann grading; ROS: reactive oxygen species; siRNA: small interfering RNAs; SO-FG: Safranin O-Fast Green; TBHP: tert-butyl hydroperoxide; TEM: transmission electron microscope; UCHL3: ubiquitin C-terminal hydrolase L3; WT: wild-type

Introduction

Intervertebral disc degeneration (IVDD) is a leading cause of chronic low back pain and physical disability worldwide [1], imposing significant socioeconomic burdens on healthcare systems and societies [2,3]. The pathological progression of IVDD is marked by the irreversible loss and dysfunction of nucleus pulposus cells (NPCs) [4], excessive degradation of the extracellular matrix (ECM) [5], and disruption of the homeostatic balance within the intervertebral disc (IVD) microenvironment [6]. While symptomatic interventions, including physical rehabilitation and surgical discectomy, have made significant strides, disease-modifying therapies targeting the underlying molecular mechanisms of IVDD remain lacking [7]. This gap is primarily due to the incomplete understanding of the regulatory networks that govern NPC survival and function under pathological conditions [8].

Growing evidence has implicated various regulated cell death pathways – including apoptosis, necroptosis, and pyroptosis – in the pathogenesis of IVDD [9]. In recent years, ferroptosis, an iron-dependent form of regulated cell death driven by uncontrolled lipid peroxidation and glutathione (GSH) depletion, has emerged as a critical mediator of NPC loss during IVDD progression [10,11]. Unlike other forms of cell death, ferroptosis is characterized by the inactivation of GPX4 (glutathione peroxidase 4), excessive accumulation of lipid reactive oxygen species (ROS), and dysregulated iron metabolism [12,13]. The avascular and hypoxic environment of the IVD renders NPCs particularly susceptible to oxidative stress-induced ferroptosis [14]. Preclinical studies have confirmed that pharmacological or genetic inhibition of ferroptosis alleviates NPC dysfunction and slows IVDD progression [15,16]. Although these observations emphasize ferroptosis as a potential therapeutic target for IVDD, the upstream molecular mechanisms regulating ferroptosis signaling in NPCs remain inadequately characterized.

Ferritinophagy, the selective autophagic degradation of the iron-storage protein ferritin, plays a critical role in regulating cellular iron homeostasis and acts as a key inducer of ferroptosis [17]. By mediating ferritin degradation, ferritinophagy releases labile iron into the cytoplasm, which then catalyzes the Fenton reaction, producing excessive ROS and triggering lipid peroxidation [18,19]. NCOA4 (nuclear receptor coactivator 4) serves as the specific cargo receptor for ferritinophagy [20,21], and NCOA4-dependent ferritin degradation is essential for intracellular iron release and the subsequent initiation of ferroptosis [22]. Ferritinophagy activation contributes to NPC ferroptosis and IVDD progression, and targeting NCOA4 can effectively inhibit ferritinophagy, offering a protective effect against IVDD [23]. However, the upstream signaling pathways that regulate NCOA4-mediated ferritinophagy in NPCs under oxidative stress conditions remain to be fully understood.

Ubiquitination and deubiquitination are evolutionarily conserved post-translational modifications that regulate protein stability, subcellular localization, and biological activity, playing key roles in cell death and autophagy [24–26]. Deubiquitinating enzymes (DUBs) remove ubiquitin moieties from substrate proteins, serving as essential regulators of this reversible modification, and are implicated in the pathogenesis of various degenerative diseases [27,28]. UCHL3 (ubiquitin C-terminal hydrolase L3), a member of the UCH subfamily of DUBs [29], is involved in proteasomal degradation [30], DNA repair [31], and tumorigenesis [32] by stabilizing oncoproteins through deubiquitination. While its role in oncology is well-established, the function of UCHL3 in degenerative musculoskeletal disorders, particularly in regulating ferroptosis and ferritinophagy in NPCs, remains unexplored.

HMGB1 (high mobility group box 1) is a highly conserved non-histone chromatin-associated protein with dual biological roles in the nucleus and cytoplasm [33]. Under normal conditions, HMGB1 is predominantly localized to the nucleus, where it regulates DNA repair, transcription, and chromatin remodeling [33–36]. In response to cellular stress, HMGB1 undergoes post-translational modifications and translocates to the cytoplasm, where it functions as a damage-associated molecular pattern (DAMP) to modulate autophagic processes and cell death signaling [37,38]. Recent studies have demonstrated that cytoplasmic HMGB1 can activate NCOA4-mediated ferritinophagy and promote ferroptosis in hepatic and neuronal cells [39,40], suggesting the existence of a conserved HMGB1-NCOA4-ferritinophagy axis in ferroptosis regulation. However, the involvement of this axis in NPCs and the mechanisms regulating HMGB1 stability and subcellular localization within the unique IVD microenvironment remain unclear.

This study identified an uncharacterized UCHL3-HMGB1-NCOA4 signaling cascade that drives ferroptosis in NPCs and accelerates IVDD progression. Our findings demonstrate that oxidative stress induces UCHL3 upregulation in NPCs, where UCHL3 binds directly to HMGB1 and mediates its deubiquitination, thereby stabilizing HMGB1 and promoting its cytoplasmic translocation. The stabilized cytoplasmic HMGB1 subsequently activates NCOA4-mediated ferritinophagy, resulting in ferritin degradation, the release of labile iron, and lipid peroxidation, all of which contribute to ferroptosis in NPCs. Furthermore, a polydopamine (PDA)-based nanoparticle delivery system (PDA@si-UCHL3) was developed to target UCHL3-specific siRNA delivery. This nanoplatform exerts a dual protective effect by silencing UCHL3 expression and scavenging intracellular ROS and excess iron, effectively inhibiting ferroptosis and mitigating IVDD progression in vivo. These findings define a novel molecular pathway regulating ferroptosis in NPCs, position UCHL3 as a potential therapeutic target for IVDD, and offer a mechanistic basis for creating disease-modifying therapies for chronic low back pain associated with IVDD.

Results

UCHL3 is elevated in NPCs ferroptosis

NPCs were treated with varying concentrations of TBHP for 3 h, with the ferroptosis inducer erastin (10 μM, 12 h) serving as a positive control. TBHP exposure led to a significant reduction in cell viability (Figure S1A), and calcein-AM/PI staining revealed a notable decrease in viable cells (Figure S1B, C). Similar to erastin treatment, TBHP treatment resulted in a substantial increase in ROS levels (Figure S1D, E). Given that GSH depletion, lipid peroxidation, and iron accumulation are hallmark features of ferroptosis, these parameters were assessed after TBHP treatment. As expected, TBHP robustly induced ferroptosis (Figure S1F-K). Additionally, both TBHP and erastin treatments increased the expression of ACSL4, PTGS2, and 4HNE, while decreasing GPX4 levels (Figure S1L, M). Further, Liproxstatin-1, a potent and specific ferroptosis inhibitor, was used to treat NPCs. Liproxstatin-1 effectively inhibited TBHP-induced ferroptosis (Figure S2D-M). In both CCK-8 assays and calcein-AM/PI staining, Liproxstatin-1 reduced cell death (Figure S2A-C). These results confirm that TBHP-induced damage leads to ferroptosis in NPCs.

To explore the molecular mechanisms driving ferroptosis in NPCs, transcriptomic sequencing was performed, identifying UCHL3 as a significantly upregulated gene following ferroptosis induction (Figure S3A). This upregulation was validated at the mRNA level by qPCR (Figure S3B, E), and western blot analysis confirmed a marked increase in UCHL3 protein expression after treatment with either TBHP or erastin (Figure S3C, D, F, G). These results suggest that UCHL3 is significantly upregulated during oxidative stress-induced ferroptosis and may play a pivotal role in regulating this cell death pathway in NPCs.

The expression of UCHL3 increases, and ferroptosis occurs after IVDD

To investigate changes in UCHL3 expression, human NP tissues were harvested from patients with IVDD, classified according to the PG system, and subjected to HE staining (Figure 1(A,B)). The expression of UCHL3, ferroptosis-related markers (4HNE and ACSL4), and IVDD-related markers (COL2/collagen II and MMP3) was assessed in NP tissues (Figure 1(C)). Correlation analyses were performed to evaluate relationships between the expression levels of these markers and PG or histological grading scores (Figure 1(D,E)). The percentage of UCHL3, 4HNE, ACSL4, and MMP3 immunopositive NP cells increased progressively with PG, correlating positively with both PG (UCHL3: R² = 0.438, p < 0.001; 4HNE: R² = 0.439, p < 0.001; ACSL4: R² = 0.326, p < 0.001; MMP3: R² = 0.451, p < 0.001) and histological scores (UCHL3: R² = 0.533, p < 0.001; 4HNE: R² = 0.421, p < 0.001; ACSL4: R² = 0.407, p < 0.001; MMP3: R² = 0.505, p < 0.001) (Figure 1(C-E)). Conversely, immunopositivity for COL2 decreased with PG, and the percentage of immunopositive cells inversely correlated with PG (COL2: R² = 0.469, p < 0.001) and histological scores (COL2: R² = 0.357, p < 0.001) (Figure 1(C-E)). Western blot analysis of tissue lysates confirmed that, as IVDD progressed, levels of UCHL3, ferroptosis markers (ACSL4, PTGS2, and 4HNE) and ECM catabolic markers (MMP3 and MMP13) significantly increased, while GPX4 and ECM anabolic markers (COL2 and ACAN [aggrecan]) expression decreased (Figure 1(F-H) and S4A-C). IF analyses yielded similar results (Figures 1(I) and S4D). Overall, these results suggest that increased UCHL3 expression in degenerative NPCs may contribute to cellular ferroptosis and the progression of IVDD.

Figure 1.

Ferroptosis is induced after IVDD, with increased expression of UCHL3. (A) MRI of IVDD categorized according to the PG system. (B) Representative HE staining of human NP tissues. (C) Representative IHC images showing expression of UCHL3, 4HNE, ACSL4, COL2, and MMP3. (D-E) Linear regression analysis correlating positive cell rates with PG (d) and histological scores (e). (F-H) Representative western blots showing expression levels of UCHL3, GPX4, ACSL4, PTGS2, 4HNE, COL2, ACAN, MMP3, and MMP13 in human NP tissues (n = 3). (I) Immunofluorescence analysis confirming upregulated UCHL3 expression in human NPCs in the context of IVDD (n = 3). Note that n denotes the number of independent biological replicates.

Overexpression of UCHL3 increases the oxidative stress-induced NPC ferroptosis

To investigate whether upregulated UCHL3 directly contributes to ferroptosis induction, UCHL3 expression was modulated using siRNA (si-UCHL3) for targeted knockdown and an overexpression plasmid (OE-UCHL3) for enforced upregulation. Western blot analysis confirmed that si-UCHL3 significantly reduced, while OE-UCHL3 significantly increased, UCHL3 protein levels (Figure 2(A-C)). Further investigation revealed that OE-UCHL3 aggravated ferroptosis (Figures 2(D-I) and S5E, F), whereas UCHL3 knockdown reversed these effects (Figures 2(D-I) and S5E, F). Additionally, UCHL3 knockdown alleviated the partial inhibition of cell viability, while UCHL3 overexpression further reduced cell viability (Figure S5B-D). Western blot analysis demonstrated that UCHL3 overexpression exacerbated oxidative stress-induced ferroptosis in NPCs, as evidenced by decreased GPX4 levels and increased ACSL4, PTGS2, and 4HNE expression (Figure 2(J,K)). Moreover, UCHL3 overexpression promoted the expression of ECM-degrading enzymes MMP3 and MMP13 while suppressing the synthesis of key matrix components, COL2 and ACAN (Figure S5G-L), highlighting its dual role in both ferroptosis and ECM catabolism. Given the established association between ferroptosis and mitochondrial dysfunction, this relationship was further explored by assessing TBHP-induced alterations in mitochondrial morphology and function following UCHL3 knockdown or overexpression in NPCs. MitoSOX Red fluorescence assays showed that si-UCHL3 significantly reduced TBHP-induced mitochondrial superoxide accumulation, while OE-UCHL3 exacerbated mitochondrial oxidative stress (Figure 2(L,M)). JC-1 staining indicated that UCHL3 knockdown effectively restored mitochondrial membrane potential, whereas UCHL3 overexpression caused significant depolarization (Figure 2(N,O)). IF imaging of mitochondrial architecture revealed that TBHP treatment severely disrupted cristae integrity, which was partially rescued by UCHL3 knockdown but further aggravated by UCHL3 overexpression (Figure 2(P)). TEM analysis confirmed that TBHP induced profound ultrastructural abnormalities, including cristae fragmentation and outer membrane deformation, which were ameliorated with UCHL3 knockdown but intensified with OE-UCHL3 (Figure 2(Q)). These results demonstrate that UCHL3 plays a critical role in maintaining mitochondrial homeostasis, where its upregulation disrupts both the structural and functional integrity of mitochondria, thereby enhancing oxidative stress-driven ferroptosis in vitro.

Figure 2.

In vitro analysis demonstrates that UCHL3 promotes ferroptosis in NPCs. (A) Experimental setup for this part of the study. (B-C) Confirmation of UCHL3 transfection efficiency via western blot analysis (n = 3). (D-G) Measurement of ROS, GSH, and MDA levels (n = 3). (H-I) Detection of intracellular Fe²⁺ using FerroOrange and corresponding quantitative analysis of relative fluorescence intensity (n = 3). (J-K) Western blot analysis of ferroptosis-related protein expression across experimental groups, with quantitative data (n = 3). (L-M) Mitochondrial ROS accumulation detected by MitoSOX immunofluorescence and corresponding statistical analysis (n = 3). (N-O) Mitochondrial membrane potential assessed by JC-1 staining, with quantification of JC-1 aggregates and monomers (n = 3). (P) Mitochondrial function in NPCs evaluated by MitoTracker staining. (Q) Representative TEM images showing mitochondrial ultrastructure in NPCs. *p < 0.05; **p < 0.01; ***p < 0.001. Note that n denotes the number of independent biological replicates.

UCHL3 interacts with HMGB1

To investigate the mechanistic role of UCHL3 accumulation in ferroptosis, IP/MS and molecular docking analyses were performed to identify molecules interacting with UCHL3. The results revealed a strong interaction between HMGB1 and UCHL3 (Figure 3(A,B)). Co-IP experiments confirmed this finding, showing that UCHL3 precipitated HMGB1 but not IgG in NPCs (Figure 3(C)). Reverse co-IP further validated that HMGB1 effectively binds to and precipitates UCHL3 in NPCs (Figure 3(D)). To confirm this interaction, Flag-tagged UCHL3 and MYC-tagged HMGB1 were co-transfected into HEK293T cells, and co-IP was conducted using antibodies against the corresponding tags. As expected, Flag-tagged UCHL3 and MYC-tagged HMGB1 strongly interacted in HEK293T cells (Figure 3(E,F)). Moreover, UCHL3 expression and its binding to HMGB1 were upregulated in both erastin- and TBHP-treated NPCs (Figure 3(G,H)). IF staining further demonstrated the colocalization of UCHL3 and HMGB1 in NPCs (Figure 3(I)). To pinpoint the binding region, a series of truncated Flag-tagged UCHL3 and MYC-tagged HMGB1 fragments were used to assess their interaction in HEK293T cells (Figure 3(J)). This analysis identified that fragments containing amino acids 81–230 within the C terminus of UCHL3 bind to HMGB1 (Figure 3(K)). Further co-IP assays confirmed that UCHL3 specifically interacts with the domain of HMGB1 spanning residues 80–163 (Figure 3(L)).

Figure 3.

UCHL3 interacts with HMGB1 in NPCs. (A) Secondary spectrum of HMGB1 from IP/MS results. (B) Molecular docking analysis showing the interaction between UCHL3 and HMGB1. (C) UCHL3 precipitates HMGB1 in NPCs, confirming the interaction. (D) Reverse co-IP demonstrating that UCHL3 is precipitated by HMGB1 in NPCs. (E-F) Exogenously Flag-tagged UCHL3 and MYC-tagged HMGB1 can effectively co-precipitate in HEK293T cells. (G) Co-IP of UCHL3 and HMGB1 in NPCs treated with or without erastin (n = 3). (H) Co-IP of UCHL3 and HMGB1 in NPCs treated with or without TBHP (n = 3). (I) IF colocalization analysis confirming the interaction between UCHL3 and HMGB1 in NPCs (n = 3). (J) Diagram showing the protein domains of UCHL3 and HMGB1. (K) Co-IP showing that the fragment containing the 81–230 amino acids of UCHL3 binds to HMGB1. (L) Reverse co-IP demonstrating that the fragment containing the 80–163 amino acids of HMGB1 binds to UCHL3. Note that n denotes the number of independent biological replicates.

UCHL3 inhibits HMGB1 ubiquitination and degradation

The effect of UCHL3 silencing on HMGB1 expression in NPCs was examined. Compared to the si-NC, UCHL3 knockdown significantly reduced HMGB1 protein levels without affecting its mRNA abundance, regardless of treatment with erastin (Figure S6A-C) or TBHP (Figure S6D-F). These results suggest that UCHL3 regulates HMGB1 post-transcriptionally, likely by modulating its protein stability rather than transcription. Treatment with the proteasome inhibitor MG132 reversed the reduction in HMGB1 protein levels induced by UCHL3 depletion (Figure 4(A,B)). To investigate whether UCHL3 stabilizes HMGB1, wild-type (WT) UCHL3 overexpression was shown to significantly increase HMGB1 protein levels, whereas overexpression of the catalytically inactive C95A mutant failed to produce a similar effect (Figure 4(C)). CHX chase assays revealed that UCHL3 knockdown accelerated HMGB1 degradation, while UCHL3 overexpression markedly delayed its turnover (Figure 4(D-G)), confirming that UCHL3 stabilizes HMGB1 in a proteasome-dependent manner. IF staining demonstrated clear colocalization between WT UCHL3 and HMGB1, whereas no significant colocalization was observed between the catalytically inactive UCHL3 C95A mutant and HMGB1 (Figure 4(H)). Further analysis was conducted to determine whether UCHL3 regulates HMGB1 ubiquitination and subsequent degradation. Compared to the si-NC, UCHL3 knockdown significantly increased HMGB1 ubiquitination while concurrently reducing its protein levels (Figure 4(I,J)). To confirm UCHL3’s role in regulating HMGB1 ubiquitination, HEK293T cells were co-transfected with MYC-HMGB1, Flag-UCHL3 (WT or C95A mutant), and HA-Ub. Overexpression of WT UCHL3 significantly suppressed HMGB1 ubiquitination, an effect that was abolished by the C95A mutant (Figure 4(K,L)). Additionally, HMGB1 ubiquitination and degradation were significantly upregulated in both the sham and IVDD groups compared to WT rats injected with AAV-UCHL3 (Figure 4(M,N)). The impact of UCHL3 on the polyubiquitination of HMGB1 was also investigated. UCHL3 specifically cleaves Lys48-linked polyubiquitin chains without significantly affecting Lys63-linked chains (Figure 4(O)). To confirm the essential role of Lys48-linked polyubiquitination in UCHL3-mediated HMGB1 degradation, a Lys48-resistant ubiquitin mutant (Lys48R) was expressed in UCHL3-knockdown HEK293T cells. The expression of Lys48R ubiquitin completely prevented the reduction in HMGB1 protein levels induced by UCHL3 knockdown (Figure 4(P)), highlighting the critical role of Lys48-linked ubiquitination in this regulatory axis. In conclusion, these results demonstrate that UCHL3 regulates the stability of HMGB1 in NPCs through modulation of its ubiquitination and proteasomal degradation.

Figure 4.

UCHL3 interacts with HMGB1 and mediates the deubiquitination of HMGB1. (A-B) Analysis of HMGB1 and UCHL3 in si-UCHL3-transfected NPCs with or without MG132 proteasome inhibitor treatment (n = 3). (C) Transfection of Flag-tagged UCHL3 (wt or C95A mutant) into HEK293T cells, followed by immunoblotting with anti-HMGB1 and anti-Flag antibodies (n = 3). (D-E) Assessment of HMGB1 protein stability after UCHL3 knockdown in NPCs (n = 3). (F-G) Evaluation of HMGB1 protein stability following UCHL3 overexpression in NPCs (n = 3). (H) IF staining showing that Flag-tagged UCHL3 significantly colocalizes with HMGB1, unlike the C95A mutant UCHL3 tagged with Flag. (I-J) Lysates from si-NC or si-UCHL3-transfected NPCs treated with MG132 before collection, subjected to immunoprecipitation and analyzed by the indicated antibodies (n = 3). (K-L) HEK293T cells transfected with HA-Ub and MYC-HMGB1, along with Flag-tagged UCHL3 (wt) or C95A mutant, were immunoprecipitated and analyzed by anti-MYC, anti-HA, and anti-Flag immunoblotting (n = 3). (M-N) NP tissue lysates from rats injected with AAV-Con or AAV-UCHL3 were immunoprecipitated with anti-HMGB1 (n = 3). Ubiquitinated HMGB1 was detected by western blot using anti-Ub and anti-HMGB1 antibodies. (O) HEK293T cells co-transfected with MYC-HMGB1, Flag-UCHL3, and HA-Ub Lys0, Lys48 alone, or Lys63 alone plasmids, followed by HMGB1 ubiquitination analysis (n = 3). (P) in the presence of si-NC or si-UCHL3, HEK293T cells transfected with either Ub wt or Ub Lys48R were cultured for 72 h, and cell lysates were analyzed by immunoblotting with anti-HMGB1 and anti-UCHL3 (n = 3). *p < 0.05; **p < 0.01; ***p < 0.001. Note that n denotes the number of independent biological replicates.

HMGB1 promotes NCOA4-mediated ferritinophagy

HMGB1 is a highly conserved nuclear protein involved in critical processes such as DNA replication, repair, transcription, and chromatin remodeling [41]. Emerging evidence suggests that under cellular stress, HMGB1 translocates from the nucleus to the cytoplasm [42]. Once in the cytosol, HMGB1 plays a key role in initiating both selective and nonselective autophagy by directly interacting with the core autophagy protein Beclin 1 [37,43]. Immunofluorescence staining revealed a significant increase in cytoplasmic HMGB1 in NPCs following treatment with either TBHP or erastin, indicating that these ferroptosis inducers facilitate the translocation of HMGB1 from the nucleus to the cytoplasm (Figure 5(A,B)). To explore the mechanistic role of HMGB1 in regulating ferroptosis in NPCs, HMGB1 expression was modulated using siRNA (si-HMGB1) or an overexpression plasmid (OE-HMGB1) in the context of TBHP-induced oxidative stress. NPCs were subsequently transduced with an mRFP-GFP-LC3 adenovirus to assess autophagic flux. si-HMGB1 treatment significantly reduced the number of autophagosomes (yellow puncta) and autolysosomes (red-only puncta), whereas OE-HMGB1 notably enhanced their formation (Figure 5(C,D)). Western blot analysis showed that HMGB1 overexpression upregulated NCOA4 and LC3-II, while downregulating ferritin heavy chain 1 (FTH1) and SQSTM1/p62, which are consistent with the activation of ferritinophagy (Figure 5(E,F)). Additionally, IF co-staining for LC3 (red) and NCOA4 (green) revealed a marked increase in colocalization in the OE-HMGB1 group, indicating a direct spatial association between HMGB1 and the autophagosomal machinery (Figure 5(G,H)). These results suggest that HMGB1 activates ferritinophagy by upregulating NCOA4, facilitating ferritin degradation, and promoting ferroptosis in NPCs.

Figure 5.

HMGB1 activates NCOA4-mediated ferritinophagy. (A-B) IF analysis revealed partial translocation of HMGB1 from the nucleus to the cytoplasm in NPCs following treatment with TBHP or erastin (n = 3). (C-D) Confocal microscopy was used to detect the number of autophagosomes and autophagolysosomes after transfection of NPCs with mRFP-GFP-LC3 adenovirus (n = 3). (E-F) Western blot analysis determined the expression of NCOA4, FTH1, SQSTM1, and LC3-I/II proteins (n = 3). (G-H) the expression of LC3 and NCOA4 after HMGB1 knockdown or overexpression was analyzed by immunofluorescence (n = 3). *p < 0.05; ***p < 0.001. Note that n denotes the number of independent biological replicates.

UCHL3 regulates ferroptosis in NPCs by modulating NCOA4–mediated ferritinophagy

To investigate whether UCHL3 regulates ferroptosis in NPCs through NCOA4-mediated ferritinophagy, rescue experiments were performed by co-overexpressing UCHL3 and knocking down NCOA4 under TBHP-induced oxidative stress conditions. Western blot analysis revealed that NCOA4 knockdown significantly reduced UCHL3-induced ferroptosis, as evidenced by the upregulation of GPX4 and the downregulation of ferroptosis markers, including ACSL4, PTGS2, and 4HNE (Figure 6(A,B)). Several hallmark features of ferroptosis were assessed. UCHL3 overexpression significantly elevated intracellular ROS levels, increased MDA and Fe²⁺ concentrations, enhanced lipid peroxidation, and depleted GSH levels in NPCs (Figures 6(C-H) and S7D, E). In contrast, NCOA4 knockdown effectively alleviated these ferroptotic alterations (Figures 6(C-H) and S7D, E). CCK-8 viability assays and calcein-AM/PI dual staining further demonstrated that si-NCOA4 significantly mitigated the UCHL3-induced increase in cell death (Figure S7A-C). Next, the impact of NCOA4 knockdown on the expression of ECM-related proteins in NPCs was evaluated. Western blot and IF analyses revealed that NCOA4 silencing significantly upregulated ECM anabolic markers (COL2 and ACAN), while downregulating the catabolic enzymes MMP3 and MMP13 (Figure S7F-K). These results demonstrate that UCHL3 promotes ferroptosis in NPCs via NCOA4-mediated ferritinophagy, ultimately contributing to IVDD through disruption of ECM homeostasis.

Figure 6.

The UCHL3-HMGB1-NCOA4 axis mediates ferroptosis in NPCs. NPCs were transfected with either OE-UCHL3 or si-NCOA4, then treated with TBHP (100 µM) for 3 h. (A-B) Representative western blot images and statistical quantification of ferroptosis-related proteins (n = 3). (C-D) Intracellular ROS levels assessed by flow cytometry with corresponding quantification (n = 3). (E-F) Quantitative measurement of GSH and MDA levels in NPCs (n = 3 per group). (G-H) Detection of intracellular Fe²⁺ using FerroOrange and quantitative analysis of relative fluorescence density (n = 3). (I-J) Immunofluorescence detection and quantification of mitochondrial ROS using MitoSOX in NPCs (n = 3). (K-L) Assessment of mitochondrial membrane potential by JC-1 staining, with quantification based on the JC-1 aggregate-to-monomer ratio (n = 3). (M) Mitochondrial function detected using the MitoTracker kit. (N) Representative TEM images of NPCs treated with different groups. *p < 0.05; **p < 0.01; ***p < 0.001. Note that n denotes the number of independent biological replicates.

Mitochondrial function was then assessed. Pre-treatment with OE-UCHL3 significantly elevated mitochondrial superoxide levels in NPCs, an effect markedly attenuated by NCOA4 knockdown (Figure 6(I,J)). Consistent with this, JC-1 staining revealed that si-NCOA4 partially restored the OE-UCHL3-induced loss of mitochondrial membrane potential (Figure 6(K,L)). Mitochondrial morphology, assessed by MitoTracker staining, further confirmed that NCOA4 deficiency alleviated mitochondrial dysfunction (Figure 6(M)). Ultrastructural analysis via TEM demonstrated that both TBHP treatment and OE-UCHL3 overexpression induced severe mitochondrial damage, characterized by cristae disintegration, matrix condensation, and outer membrane rupture, phenotypes that were substantially mitigated upon NCOA4 silencing (Figure 6(N)). In conclusion, these in vitro results demonstrate that NCOA4 knockdown preserves mitochondrial integrity and alleviates ferroptosis-induced damage in NPCs under oxidative stress.

Knockdown of UCHL3 significantly alleviates the progression of IVDD in vivo

Delaying the progression of IVDD remains a key therapeutic goal. Based on our previous findings that UCHL3 drives ferroptosis and contributes to NPC degeneration, this study explored the therapeutic potential of targeting UCHL3 in vivo. To this end, a rat tail-needle puncture model of IVDD was established, alongside sham-operated controls, with AAV-shUCHL3 delivered via intradiscal injection (Figure 7(A)). At the end of this study, IVD samples were collected and subjected to comprehensive histopathological and imaging analyses. IVDD severity was evaluated using X-ray, MRI, HE, SO-FG, Alcian blue, and IHC staining. X-ray imaging revealed that UCHL3 knockdown mitigated the loss of disc height caused by IVDD surgery (Figure 7(B,C)). Similarly, MRI demonstrated that UCHL3 silencing alleviated the reduction in disc hydration post-surgery (Figure 7(D,E)). Histological analysis via HE, SO-FG, and Alcian blue staining revealed significant pathological changes in the IVDs following needle puncture, including a marked reduction in NPC numbers and proteoglycan content, disorganization of annulus fibrosus (AF) lamellae, and loss of the demarcation between the NP and AF in the IVDD + AAV-shCtrl groups (Figure 7(F)). Notably, treatment with AAV-shUCHL3 effectively mitigated these degenerative changes, preserving the structural integrity of the NP, AF, and endplate tissues, as well as the boundaries between these compartments. Additionally, the histological scoring system revealed that AAV-shUCHL3 played a pivotal protective role in the progression of IVDD (Figure 7(G)). IHC and western blot analyses further confirmed that AAV-shUCHL3 treatment significantly downregulated the expression of MMP3 and MMP13, while concurrently upregulating COL2 and ACAN (Figures 7(H-M) and S8A, B), effectively inhibiting IVDD progression.

Figure 7.

Knockdown of UCHL3 ameliorates rat IVDD. (A) Schematic illustration of the experimental design (n = 6). (B-C) Representative X-ray images of rat caudal IVDs treated with intradiscal injection of AAV-shCtrl or AAV-shUCHL3 (n = 6). (D-E) Representative MRI images of rat caudal IVDs treated with intradiscal injection of AAV-shCtrl or AAV-shUCHL3 (n = 6). (F-G) HE, SO-FG, and Alcian blue staining of rat caudal IVDs (n = 6). (H-M) IHC staining of UCHL3, COL2, and MMP3 in rat coccygeal IVDs, along with quantitative analysis of UCHL3, COL2, and MMP3 immunostaining (n = 6). ***p < 0.001. Note that n denotes the number of independent biological replicates.

Ferroptosis was next evaluated in vivo. Needle puncture significantly increased MDA levels and Fe²⁺ accumulation, while markedly reducing GSH content, collectively indicating robust activation of ferroptosis in the IVDD model (Figure S8C-E). These alterations were substantially reversed by AAV-shUCHL3 delivery (Figure S8C-E). Further IF analysis revealed that UCHL3 knockdown significantly suppressed the expression of ferroptosis-associated markers (4HNE and ACSL4) compared to the AAV-shCtrl group (Figure S8F-I). In summary, these results demonstrate that genetic silencing of UCHL3 effectively inhibits ferroptosis, alleviating both the pathological activation of ferroptosis signaling and NPC dysfunction in the context of IVDD.

Synthesis and characterization of PDA@si-UCHL3

Our findings suggest that targeting UCHL3 represents a promising therapeutic strategy for IVDD. PDA nanoparticles have been widely utilized for ROS scavenging in various pathological conditions, including cancer, infections, and inflammatory disorders like periodontitis [44–48]. Notably, PDA effectively chelates excess Fe²⁺, thereby suppressing Fenton reactions, inhibiting the lipid peroxidation cascade, and directly disrupting a core driver of ferroptosis [49,50]. Additionally, the surface of PDA is densely functionalized with catechol and amine groups, which facilitate the high-efficiency loading of nucleic acid therapeutics, such as siRNA, through non-covalent interactions, including π–π stacking, hydrogen bonding, and electrostatic attraction. This ensures the reliable and effective delivery of the therapeutic payload [51–54]. By leveraging these characteristics, particularly the excellent biocompatibility and low cytotoxicity of PDA, this study engineered a nanocomposite platform, designated PDA@si-UCHL3, using PDA as a carrier to deliver UCHL3-targeting siRNA.

PDA was synthesized via oxidative self-polymerization of dopamine under aerobic and mildly alkaline conditions (Figure 8(A)). During the polymerization process, the dopamine solution transitioned from colorless to light brown within 0.1 h, eventually becoming black after 5 h, confirming the successful formation of PDA. As shown in Figure 8(B), free siRNA exhibited a bright fluorescent band. When si-UCHL3 was complexed with PDA at mass ratios of 1:10 and 1:30, visible bands were still observed, but their intensity was markedly reduced compared to the free siRNA, indicating successful loading of si-UCHL3 onto the PDA carrier. At higher PDA mass ratios, the fluorescent band disappeared completely, demonstrating efficient encapsulation of si-UCHL3 by the PDA nanoparticles. Based on these results, a si-UCHL3-to-PDA mass ratio of 1:50 was used in all subsequent experiments for the preparation of PDA@si-UCHL3 nanoparticles. TEM analysis revealed that both PDA and PDA@si-UCHL3 nanoparticles exhibited uniform spherical morphology with excellent dispersity (Figure 8(C)). The PDA nanoparticles displayed a narrow size distribution, with an average diameter of approximately 190 nm (Figure 8(D)), while the PDA@si-UCHL3 nanoparticles showed a slightly broader size distribution, with a mean diameter of approximately 200 nm. The zeta potentials of PDA and PDA@si-UCHL3 were measured as −(25.0 ± 0.7) mV and −(27.8 ± 3.2) mV, respectively (Figure 8(E)). The more negative surface potential of PDA@si-UCHL3 compared to bare PDA suggests successful incorporation of the negatively charged si-UCHL3 into the PDA nanoparticles.

Figure 8.

Preparation and characterization of PDA@si-UCHL3. (A) Schematic illustration of the process to prepare PDA@si-UCHL3. (B) Agarose gel electrophoresis assay of PDA@si-UCHL3 complexes with different si-UCHL3 : PDA ratios. (C) TEM image of PDA and PDA@si-UCHL3. (D) Hydrodynamic size distribution of PDA and PDA@si-UCHL3 nanoparticles. (E) Zeta potential of PDA and PDA@si-UCHL3 determined by DLS.

PDA@si-UCHL3 alleviates ferroptosis of NPCs and progression of IVDD

The therapeutic efficacy of PDA@si-UCHL3 in NPCs subjected to oxidative stress was initially assessed. Compared to the PBS and PDA groups, PDA@si-UCHL3 significantly reduced ferroptosis in NPCs (Figures 9(A-H) and S9A-E). Western blot and IF analyses confirmed that PDA@si-UCHL3 notably increased the expression of ECM anabolic markers (COL2 and ACAN), while decreasing the levels of catabolic enzymes MMP3 and MMP13 (Figure S9F-K), indicating a potent suppression of NPC degeneration. Additionally, mitochondrial functional assays demonstrated that PDA@si-UCHL3 promoted mitochondrial recovery, evidenced by reduced mitochondrial superoxide levels, restored mitochondrial membrane potential, and improved ultrastructural integrity (Figure 9(I-N)).

Figure 9.

PDA@si-UCHL3 attenuates oxidative stress-induced ferroptosis in NPCs. (A-B) Representative western blot images and statistical quantification of ferroptosis-related proteins in NPCs (n = 3). (C-D) Quantification of intracellular ROS levels by flow cytometry with corresponding statistical analysis (n = 3). (E-F) Measurement of GSH and MDA concentrations in NPCs (n = 3). (G-H) Detection of intracellular Fe²⁺ using FerroOrange and quantitative analysis of relative fluorescence density (n = 3). (I-J) MitoSOX-based immunofluorescence staining and quantification of mitochondrial ROS levels (n = 3). (K-L) Analysis of mitochondrial membrane potential using JC-1 aggregates/monomers ratio (n = 3). (M) Mitochondrial function of NPCs evaluated by MitoTracker. (N) TEM images illustrating mitochondrial morphology in NPCs. *p < 0.05; **p < 0.01; ***p < 0.001; ns = not significant. Note that n denotes the number of independent biological replicates.

To assess the in vivo protective effects of PDA@si-UCHL3, the nanocomplex was injected into the NP region of rat caudal IVDs immediately following needle puncture-induced IVDD (Figure 10(A)). Radiological evaluations revealed that PDA@si-UCHL3 treatment significantly attenuated major degenerative features, including loss of disc height, decreased disc hydration, and subchondral bone destruction (Figure 10(B-E)). Histological and western blot analyses further demonstrated that PDA@si-UCHL3 significantly downregulated MMP3 and MMP13, while upregulating COL2 and ACAN, thereby mitigating IVDD progression (Figures 10(F-M) and S10A, B). Ferroptosis-related assays showed that PDA@si-UCHL3 markedly reduced MDA and Fe²⁺ levels, increased GSH content, and suppressed the expression of key ferroptosis markers (4HNE and ACSL4) (Figure S10C-I). These results highlight that PDA-mediated delivery of UCHL3-targeting siRNA effectively mitigates IVDD progression.

Figure 10.

PDA@si-UCHL3 significantly attenuates IVDD progression. (A) Schematic illustration of the experimental design (n = 6). (B-C) Representative X-ray images of rat caudal IVDs treated with intradiscal injection of PBS, PDA, or PDA@si-UCHL3 (n = 6). (D-E) Representative MRI images of rat caudal IVDs treated with intradiscal injection of PBS, PDA, or PDA@si-UCHL3 (n = 6). (F-G) HE, SO-FG, and Alcian blue staining of rat caudal IVDs (n = 6). (H-M) IHC staining of UCHL3, COL2, and MMP3 in rat coccygeal IVDs, with quantitative analysis of UCHL3, COL2, and MMP3 immunostaining (n = 6). *p < 0.05; ***p < 0.001. ns = not significant. Note that n denotes the number of independent biological replicates.

Discussion

Accumulating evidence firmly establishes ferroptosis as a principal driver of NPCs depletion in IVDD, catalyzing the development of targeted therapeutic interventions. Recent advances have introduced multifaceted approaches to intercept this pathway, including pterostilbene-loaded nanoliposomes that suppress ferritinophagy via the NFE2L2/NRF2-NCOA4-FTH1 axis [55], ADGRG1-targeted hypoxic extracellular vesicles delivering taurine to disrupt oxidative stress-driven ferroptotic feedback [56], and FBXW7-mediated mitophagy regulation that mitigates ferroptotic damage in NPCs [57]. Collectively, these studies underscore the therapeutic promise of targeting ferroptosis while revealing a critical knowledge gap: the upstream signaling networks that initiate ferroptosis within the unique disc microenvironment remain poorly defined. This study identified a novel UCHL3-HMGB1-NCOA4 signaling axis that modulates ferritinophagy and subsequent ferroptosis in NPCs. Targeted silencing of UCHL3 via PDA-based nanoparticle-mediated siRNA delivery exerts a robust protective effect against IVDD progression in vivo (Figure 11). These findings not only deepen the mechanistic understanding of oxidative stress-induced NPC ferroptosis but also position UCHL3 as a promising therapeutic target for developing disease-modifying treatments for IVDD.

Figure 11.

Schematic illustration of the UCHL3-HMGB1-NCOA4 axis regulating ferroptosis in NPCs. Under oxidative stress, upregulated UCHL3 deubiquitinates HMGB1 by cleaving K48-linked polyubiquitin chains, stabilizing HMGB1 and promoting its cytoplasmic translocation. Cytoplasmic HMGB1 activates NCOA4-mediated ferritinophagy, triggering ferroptosis. The polydopamine-based nanoplatform (pda@si-UCHL3) delivers UCHL3-targeting siRNA to inhibit this pathway, thereby alleviating IVDD in a rat model.

As a member of the deubiquitinase (DUB) superfamily, UCHL3 regulates protein stability and intracellular signal transduction by catalyzing the removal of ubiquitin moieties from substrate proteins [58]. While substantial evidence has established the oncogenic role of UCHL3 in various cancers by stabilizing key oncoproteins [59], its biological functions in degenerative musculoskeletal disorders, particularly IVDD, have not been investigated. In this study, transcriptomic profiling of ferroptotic NPCs first identified UCHL3 as a significantly upregulated gene. Subsequent analysis of clinical human NP tissues confirmed that UCHL3 expression correlated positively with both the PG of IVDD and the levels of classical ferroptosis markers (4HNE, ACSL4). Functional gain- and loss-of-function experiments revealed that UCHL3 overexpression exacerbated oxidative stress-induced ferroptosis in NPCs, evidenced by increased intracellular labile iron accumulation, enhanced lipid peroxidation, and GSH depletion. In contrast, UCHL3 knockdown alleviated these ferroptosis phenotypes and preserved mitochondrial structural and functional integrity. Mitochondria, being both the primary source of ROS and a key target of ferroptotic damage, were found to be modulated by UCHL3 in terms of superoxide production and membrane potential. This establishes a direct link between UCHL3-mediated mitochondrial dysfunction and ferroptosis, a critical pathological process in IVDD.

The biological functions of DUBs are primarily mediated through their specific deubiquitination of downstream substrate proteins [60]. To elucidate the molecular mechanism by which UCHL3 regulates ferroptosis in NPCs, IP/MS and molecular docking analyses were conducted, identifying HMGB1 as a direct and specific binding partner of UCHL3. HMGB1 is a highly conserved dual-functional protein that serves as a nuclear architectural factor involved in DNA repair, transcription, and chromatin dynamics under normal conditions. Upon cellular stress, HMGB1 translocates to the cytoplasm, where it functions as a DAMP [61,62]. Co-IP and immunofluorescence colocalization assays further confirmed the specific physical interaction between UCHL3 and HMGB1, and identified their functional binding domains within the C-terminal 81–230 amino acids of UCHL3 and the 80–163 amino acid region of HMGB1. Notably, UCHL3 regulates HMGB1 at the post-transcriptional level. UCHL3 knockdown reduced HMGB1 protein stability without affecting its mRNA expression, and this effect was reversed by treatment with the proteasome inhibitor MG132, suggesting a proteasome-dependent regulatory mechanism. Mechanistically, WT UCHL3, but not the catalytically inactive C95A mutant, significantly inhibited HMGB1 ubiquitination and degradation. Further experiments revealed that UCHL3 specifically cleaves Lys48-linked polyubiquitin chains on HMGB1, a canonical signal for proteasomal degradation. These findings establish UCHL3 as a specific DUB for HMGB1, stabilizing HMGB1 protein in NPCs through deubiquitination, which facilitates the progression of ferroptosis. While recent studies identify DUBs such as USP11 as ferroptosis suppressors through stabilization of antioxidant proteins [10], our work reveals UCHL3 as a pro-ferroptotic DUB that operates via a distinct HMGB1-NCOA4-ferritinophagy cascade. This functional divergence underscores the substrate-specific and pathway-selective nature of DUB-mediated ferroptosis control.

The cytoplasmic translocation of HMGB1 is essential for its role in regulating autophagic and cell death signaling pathways [42]. Our results showed that oxidative stress induced by TBHP or erastin promoted the translocation of HMGB1 from the nucleus to the cytoplasm in NPCs. The stabilized cytoplasmic HMGB1 activated NCOA4-mediated ferritinophagy, a selective autophagic process that degrades ferritin to release labile iron, a key initiator of ferroptosis [63]. HMGB1 overexpression upregulated NCOA4 and LC3-II expression, downregulated FTH1 and SQSTM1, and enhanced the formation of autophagosomes/autolysosomes in NPCs, whereas HMGB1 knockdown exhibited the opposite effects. These findings confirm that HMGB1 is a positive regulator of ferritinophagy during ferroptosis in NPCs, extending the previously identified HMGB1-NCOA4-ferritinophagy axis – known from hepatic and neuronal models [40,64]—to the IVD microenvironment. Rescue experiments further validated the specificity of the UCHL3-HMGB1-NCOA4 axis: NCOA4 knockdown reduced the ferroptosis induced by UCHL3 overexpression in NPCs, alleviating iron overload, lipid peroxidation, mitochondrial damage, and restoring ECM homeostasis. These results establish a clear causal relationship between UCHL3-mediated HMGB1 stabilization, NCOA4-dependent ferritinophagy activation, and ferroptosis in NPCs, unveiling a novel upstream regulatory pathway of ferritinophagy in IVDD.

The translational potential of our mechanistic findings was validated using the rat caudal needle puncture IVDD model, a well-established in vivo model that closely mimics the key pathological features of human IVDD [65]. Intradiscal injection of AAV-shUCHL3 effectively silenced UCHL3 expression in the NP, significantly mitigating IVDD progression. This was evidenced by preserved disc height and hydration, improved histological structure, and restored ECM anabolic-catabolic balance, marked by upregulation of COL2 and ACAN and downregulation of MMP3 and MMP13. UCHL3 knockdown inhibited ferroptosis in rat IVD tissue, as indicated by reduced levels of MDA and Fe²⁺, increased GSH content, and downregulation of ferroptosis markers. These findings confirm the critical in vivo role of UCHL3 in driving IVDD through ferroptosis and provide direct preclinical evidence supporting the therapeutic potential of targeting UCHL3. However, clinical application of viral vector-mediated gene therapy faces several challenges, including immunogenicity, off-target effects, and low delivery efficiency in the avascular IVD tissue [66]. To overcome these limitations, this study engineered a PDA-based nanoplatform (PDA@si-UCHL3) for the targeted delivery of UCHL3 siRNA, leveraging the unique biological properties of PDA: efficient ROS scavenging, iron chelation, high-efficiency siRNA loading via non-covalent interactions (π–π stacking, hydrogen bonding, electrostatic attraction) [67], and favorable biocompatibility with low cytotoxicity [68]. Characterization results showed that PDA@si-UCHL3 nanoparticles exhibited uniform spherical morphology, stable physicochemical properties, and complete siRNA encapsulation at an optimal mass ratio of 50:1 (PDA:siRNA). Both in vitro and in vivo experiments confirmed that PDA@si-UCHL3 exerted a dual therapeutic effect: it silenced UCHL3 to inhibit NPCs ferroptosis via the UCHL3-HMGB1-NCOA4 axis, and it directly scavenged ROS and chelated excess iron through PDA, thereby inhibiting lipid peroxidation and ferroptosis at the cellular level. This nanotherapeutic strategy effectively attenuated IVDD progression in rats and demonstrated superior therapeutic efficacy, presenting a promising new nanodelivery system for IVDD gene therapy.

Despite the significant findings of this study, several limitations warrant further investigation. First, while our experiments were primarily conducted in human primary NPCs and rodent IVDD models, the translational relevance of the UCHL3-HMGB1-NCOA4 axis should be validated in larger animal models (e.g., rabbits, pigs), which possess IVD anatomy more akin to humans, as well as in clinical samples from patients with different IVDD subtypes (e.g., degenerative disc disease, disc herniation). Second, our research focused on the role of UCHL3 in regulating ferritinophagy and ferroptosis, but the potential crosstalk between UCHL3 and other cell death pathways (e.g., apoptosis, necroptosis, pyroptosis) in the context of IVDD remains unexplored. Third, despite the promising therapeutic efficacy of PDA@si-UCHL3, its long-term biosafety requires rigorous evaluation before clinical translation. Prolonged retention of polydopamine nanoparticles in the avascular disc microenvironment may raise concerns regarding melanin-like pigment deposition and chronic inflammatory responses. Comprehensive long-term toxicological profiling is therefore essential to systematically assess nanoparticle clearance kinetics, tissue-specific accumulation, and immunogenicity. Fourth, the upstream signaling pathways regulating UCHL3 expression under oxidative stress in NPCs remain unidentified; investigating these pathways will help uncover additional therapeutic targets for IVDD and provide a more comprehensive understanding of the molecular regulation of ferroptosis in the IVD microenvironment.

In conclusion, this study identified UCHL3 as a novel and critical regulator of NPC ferroptosis and elucidated a previously unrecognized UCHL3-HMGB1-NCOA4 signaling axis that drives IVDD progression by activating ferritinophagy and ferroptosis. Targeted silencing of UCHL3 via PDA-based nanoparticle-mediated siRNA delivery effectively inhibits NPC ferroptosis and attenuates IVDD in vivo, offering new mechanistic insights into IVDD pathogenesis and providing a promising therapeutic strategy for IVDD intervention. These findings open new avenues for the development of DUB-targeted and nanomedicine-based therapies for degenerative IVD diseases and lay a solid foundation for future disease-modifying treatments for chronic low back pain associated with IVDD.

Materials and methods

Collection of human intervertebral disc specimens and ethical approval

Human nucleus pulposus (NP) samples were collected from 30 patients undergoing discectomy at Xuanwu Hospital of Capital Medical University, for conditions such as lumbar disc herniation, lumbar spinal stenosis, lumbar vertebral fractures, or scoliosis. Written informed consent was obtained from all participants before enrollment. The study protocol received approval from the Institutional Medical Ethics Committee of Xuanwu Hospital (Approval No. XW-2024–010-001). Preoperative T2-weighted magnetic resonance imaging (MRI) was used to assess the degenerative status of IVDs according to the Pfirrmann grading (PG) system (Grades I – V). Detailed information regarding these patients is provided in Table S1.

Cell culture and treatments

NPCs were isolated from IVD tissue samples of patients with Pfirrmann grades I – II using collagenase type II digestion. Following enzymatic digestion, the cell suspension was centrifuged at 300 × g for 5 min. The harvested NPCs were resuspended in DMEM/F-12 medium (Gibco, Thermo Fisher Scientific 11,330,057) supplemented with 10% fetal bovine serum (FBS, Gibco 10,099-141C) and 1% penicillin – streptomycin (Beyotime, C0222). Cells were then seeded into culture dishes and maintained in a humidified incubator at 37°C with 5% CO₂. When cells reached 80–90% confluence, they were passaged using 0.25% trypsin (Beyotime, C0201), and second-passage cells were used for all subsequent experiments. HEK293T cells (Pricella Life Science & Technology Co., Ltd, CL-0005) were cultured in DMEM supplemented with 10% FBS and 1% penicillin – streptomycin under the same conditions.

For cellular treatments, NPCs were stimulated with 10 μM erastin (MedChemExpress, HY-15763) for 12 h or 100 μM tert-butyl hydroperoxide (TBHP; Sigma-Aldrich, 458,139) for 3 h to induce ferroptosis [10]. To assess HMGB1 protein stability, 10 μM cycloheximide (CHX; MedChemExpress, HY-12320) was applied to NPCs at time points 0, 2, 4, and 8 h to block protein synthesis. Additionally, NPCs were treated with 20 μM MG132 (MedChemExpress, HY-13259) to inhibit proteasomal degradation.

Plasmid transfection and RNA interference

Flag-tagged UCHL3, its catalytically inactive C95A mutant, and MYC-tagged HMGB1 were generated by cloning the respective coding sequences into expression vectors with N-terminal epitope tags. Plasmids encoding HA-tagged ubiquitin mutants (K48R and K63R) were also obtained from Genebay Biotech (Shanghai, China). Small interfering RNAs (siRNAs) targeting UCHL3, HMGB1, and NCOA4 were employed to silence their expression in NPCs and HEK293T cells. The siRNAs were synthesized by GenePharma (Shanghai, China) and transfected into NPCs or HEK293T cells using Lipofectamine™ 2000 (Thermo Fisher Scientific 11,668,019) for 48–72 h to achieve efficient gene knockdown.

Western blot analysis

NPCs were lysed on ice for 20 min using RIPA lysis (Beyotime, P0013B) buffer supplemented with a 1% protease and phosphatase inhibitor cocktail (Beyotime, P1049). The lysates underwent brief sonication, followed by centrifugation at 14,000 × g for 30 min to collect the supernatant. Protein concentrations were determined using a BCA assay kit (Beyotime, P0012S). Samples were mixed with SDS loading buffer and heated for 10 min. Protein separation was performed using SDS-PAGE, and the proteins were transferred onto polyvinylidene difluoride (PVDF) membranes. Membranes were blocked with 5% nonfat milk in TBST (Beyotime, ST673) for 2 h at room temperature, followed by overnight incubation with primary antibodies at 4°C. After three washes with TBST, membranes were incubated with horseradish peroxidase (HRP)-conjugated secondary antibodies for 1 h at room temperature, followed by three additional TBST washes (10 min each). Chemiluminescent detection was performed by applying a chemiluminescent substrate directly to the membrane, corresponding to the target protein. Immunoreactive bands were visualized using a chemiluminescence imaging system, and band intensity was quantified as grayscale density using ImageJ software. Detailed information on the antibodies used is provided in Table 1.

Table 1.

The primary antibodies used for immunoblotting.

Target	Manufacturer	Catalog
UCHL3	Proteintech	12384–1-AP
GPX4	Abcam	ab252833
FTH1	Affinity	DF6278
ACSL4	Affinity	DF12141
PTGS2	Affinity	AF7003
Flag-Tag	Proteintech	20543–1-AP
HA-Tag	Proteintech	51064–2-AP
MYC-Tag	Proteintech	60003–2-Ig
ACTB	Cell Signaling Technology	3700T
4HNE	Abcam	ab46545
COL2	Servicebio	GB11021-100
ACAN	Affinity	DF7561
MMP3	Affinity	AF0217
MMP13	Affinity	AF5355
NCOA4	Proteintech	83394–4-RR
HMGB1	Proteintech	10829–1-AP
SQSTM1	Proteintech	18420–1-AP
LC3	Proteintech	14600–1-AP
Ub	Proteintech	10201–2-AP

RNA isolation and RT-qPCR

Total RNA was extracted from cells using a commercial RNA isolation kit (Beyotime, R0017S) and reverse transcribed into cDNA. Quantitative real-time PCR (qRT-PCR) was conducted using SYBR Green PCR Master Mix (Qiagen, 330,502) on a LightCycler® 96 system, following the manufacturer’s instructions. The mRNA expression levels of target genes were normalized to ACTB/β-actin as an endogenous reference, and relative transcript abundance was calculated using the 2^−ΔΔCt method. Primer sequences are listed in Table 2.

Table 2.

The primer sequences used for RT-qPCR.

Target	Forward	Reverse
UCHL3	CTGTGTCAATGAGCCCTGAAGAAC	TCATCTATACTTGGTGCCTCAGTCTG
HMGB1	TCTTCCTCTTCTGCTCTGAGTATCG	TTCATAAGGCTGCTTGTCATCTGC
ACTB	CCACGAAACTACCTTCAACTCCATC	AGTGATCTCCTTCTGCATCCTGTC

X-ray and MRI

Radiological evaluations of IVDs in experimental animals were performed using X-ray and MRI. Prior to imaging, rats were anesthetized with an intraperitoneal injection of sodium pentobarbital (40 mg/kg, 10% solution) and positioned prone for both X-ray and MRI scans. X-ray imaging was carried out at a tube current of 160 mA and a tube voltage of 50 kV. MRI was performed using a 3.0-T clinical magnet (Siemens) with T2-weighted sequences to assess IVD morphology, signal intensity, and hydration status. IVDD severity was graded using the Pfirrmann and Disc Height Index (DHI) scoring systems based on X-ray and MRI findings.

Double-labeled adenovirus mRFP-GFP-LC3 transfection

NPCs were seeded at a density of 1 × 10⁵ cells per well in confocal imaging dishes and cultured for 24 h. Cells were then transduced with the mRFP-GFP-LC3 adenovirus (Hanbio, LV109112701) according to the manufacturer’s instructions. After treatment, cells were washed with PBS (Beyotime, C0221A), fixed with 4% paraformaldehyde, and imaged using a confocal microscope.

Analysis of cell viability

NPCs were seeded in 96-well plates at a density of 1 × 10⁴ cells per well. Following the designated treatments, the culture medium was replaced with fresh medium, and Cell Counting Kit-8 (CCK-8) reagent (Beyotime, C0038) was added as per the manufacturer’s instructions. After 3 h of incubation at 37°C, absorbance was measured at 450 nm using a microplate reader to assess cell viability.

Immunofluorescence (IF)

NPCs were seeded into 24-well plates and subjected to designated treatments. Cells were washed twice with PBS, fixed with 4% paraformaldehyde for 15 min, permeabilized with 0.1% Triton X-100 (Beyotime, P0096) for 5 min, and then blocked with 10% goat serum (Solarbio, SL038) at 37°C for 1 h. Primary antibodies were applied and incubated overnight at 4°C. After two additional PBS washes, cells were incubated with fluorophore-conjugated secondary antibodies at 37°C for 1 h. Nuclei were counterstained with DAPI (Beyotime, C1006) for 5 min. Imaging was carried out using fluorescence microscopy, and quantitative analysis was conducted using ImageJ. Comprehensive antibody information appears in Table 3.

Table 3.

The primary antibodies used for immunofluorescence.

Target	Manufacturer	Dilution ratio
UCHL3	Proteintech	12384–1-AP
HMGB1	Proteintech	10829–1-AP
LC3	Proteintech	14600–1-AP
NCOA4	Proteintech	83394–4-RR

Calcein-AM/PI staining

NPCs were seeded into 24-well plates at an appropriate density. After experimental treatments, 250 μL of calcein-AM/PI (Beyotime, C2015M) working solution was added to each well, and the cells were incubated in the dark at 37°C for 30 min. Following incubation, live and dead cells were visualized under a fluorescence microscope (PI: red; calcein-AM: green). Fluorescence intensity was quantified using ImageJ software.

Determination of MDA, GSH, and iron content

Levels of MDA, GSH, and Fe²⁺ were measured using commercial assay kits (MDA: Westang, F01963; GSH: Westang, G02170; Fe²⁺: Westang, G01670), according to the manufacturers’ instructions. For the analysis of MDA, GSH, and Fe²⁺ levels in rat IVD tissue, samples from each experimental group were homogenized and processed per the assay protocols. The resulting homogenates were quantified using the respective kits, with absorbance measured using a microplate reader. Concentrations of MDA, GSH, and Fe²⁺ were determined by comparing the measured absorbance to standard curves generated for each analyte.

FerroOrange Fe²⁺ staining

Intracellular labile Fe^{2 +} levels were assessed using the FerroOrange probe (Dojindo, F374) according to the manufacturer’s instructions. Briefly, cells were incubated with 10 μM FerroOrange for 30 min at 37°C in a 5% CO₂ atmosphere, followed by fluorescence imaging.

JC-1 measurement

Cells were incubated with JC-1 dye (Beyotime, C2006) at 37°C in the dark for 20 min, followed by two washes with JC-1 assay buffer. JC-1 aggregates emit red fluorescence in mitochondria with high membrane potential, while the monomeric form emits green fluorescence under depolarized conditions. Mitochondrial membrane potential was assessed by quantifying the red-to-green fluorescence intensity ratio using a fluorescence microscope.

ROS detection

Intracellular ROS levels were detected using DCFH-DA (Beyotime, S0033M) according to the manufacturer’s instructions. After treatment, cells were incubated with DCFH-DA at 37°C in the dark for 20 min. After thorough washing to remove excess probe, fluorescence intensity was measured by flow cytometry and analyzed using FlowJo software.

Mitochondrial superoxide assay

Mitochondrial superoxide production was evaluated using MitoSOX Red (Beyotime, S0061S), a mitochondria-targeted fluorescent probe. Cells cultured in 24-well plates were incubated with 5 μM MitoSOX Red at 37°C in the dark for 30 min. Following incubation, cells were washed three times with pre-warmed PBS to remove unbound dye. Fluorescent signals were visualized under a fluorescence microscope, with red fluorescence intensity serving as a quantitative indicator of mitochondrial superoxide levels. All images were captured under identical exposure settings to ensure consistent and comparable assessments across experimental groups.

Lipid peroxidation assay

Lipid peroxidation was quantified using the fluorescent probe BODIPY 581/591 C11 (Beyotime, S0043S) as per established protocols. Briefly, cells were incubated with 2 μM BODIPY 581/591 C11 at 37°C for 20 min. Excess probe was removed by two sequential washes with PBS, followed by immediate fluorescence imaging. BODIPY 581/591 C11 exhibits a fluorescence shift from red (indicative of non-oxidized lipids) to green (reflecting oxidized lipids). The extent of lipid peroxidation was quantitatively assessed by calculating the ratio of green to red fluorescence intensity.

Mitochondria staining

The culture medium was aspirated and replaced with a pre-warmed MitoTracker Green (Beyotime, C1048) working solution, prepared according to the manufacturer’s instructions. Cells were incubated with the dye at 37°C for 30 min. After incubation, the staining solution was removed, and cells were gently washed once with pre-warmed PBS before replenishing with fresh culture medium equilibrated to 37°C. Mitochondrial morphology and distribution were visualized using a confocal microscope.

Mitochondrial morphology observed by TEM

Specimens were fixed in 2.5% glutaraldehyde at 4°C overnight, followed by post-fixation with 1% osmium tetroxide at room temperature for 60 min. Samples were dehydrated through a graded ethanol series, infiltrated, and embedded in epoxy resin. Ultrathin sections (85 nm) were cut with an ultramicrotome, mounted on copper grids, and doubly stained with uranyl acetate and lead citrate (2% w:v) for contrast enhancement. High-magnification ultrastructural imaging was performed using a transmission electron microscope (TEM: HT7800, HITACHI, Japan), and representative micrographs were captured for analysis.

Co-immunoprecipitation (co-IP)

NPCs were lysed on ice for 30 min using NP-40 lysis buffer (Beyotime, P0013F) supplemented with 1% protease and phosphatase inhibitor cocktail and 1% PMSF (Beyotime, ST507). The lysates were centrifuged at 15,000 × g for 10 min at 4°C. The supernatants were then incubated overnight at 4°C on a rotating platform with Protein A/G magnetic beads (MedChemExpress, HY-K0202). The immunoprecipitated complexes were washed with 1 × SDS loading buffer and subjected to western blot analysis as previously described.

Immunoprecipitation – mass spectrometry (IP/MS)

Eluates from immunoprecipitation were resolved by SDS-PAGE. Target protein bands (approximately 1 mm³) were excised from the gel and subjected to destaining, reduction, alkylation, and in-gel tryptic digestion. The resulting peptides were separated by capillary high-performance liquid chromatography (HPLC) and analyzed by tandem mass spectrometry using a Q Exactive HF-X mass spectrometer (Thermo Scientific). Raw mass spectrometry data were processed and searched against a reference protein database using Proteome Discoverer 3.0 software for peptide and protein identification.

Animal IVDD model

Six- to eight-week-old Sprague-Dawley (SD) rats weighing 250–300 g were housed under specific pathogen-free conditions at the Experimental Animal Center of Xuanwu Hospital of Capital Medical University, maintained at 22°C with a 12-h light/dark cycle. Under fluoroscopic guidance, a 22-gauge needle was inserted 5 mm into the IVD space between the seventh and eighth caudal vertebrae (Co7/8). The needle was rotated 360 degrees and held in place for 30 seconds to induce standardized disc injury. All animal procedures were approved by the Institutional Animal Care and Use Committee of Xuanwu Hospital of Capital Medical University (Approval No. XW-20251022-1).

Histological analysis

Eight weeks post-surgery, rat IVD samples were harvested and rinsed twice with PBS. The tissues were fixed in 4% paraformaldehyde for 48 h, followed by decalcification in 10% ethylenediaminetetraacetic acid (EDTA) solution at room temperature for 8 weeks. After dehydration through a graded ethanol series, the specimens were embedded in paraffin and sectioned into 5-μm-thick slices. Histological evaluation was performed using hematoxylin and eosin (HE), Safranin O – Fast Green (SO-FG), and Alcian blue staining. The stained sections were examined under a light microscope for structural and compositional analysis.

Immunohistochemistry (IHC) staining

IHC was conducted to assess the expression levels of UCHL3, 4HNE, ACSL4, COL2, and MMP3 in human NP tissues and rat IVDs. Briefly, 5-μm-thick paraffin sections were deparaffinized in xylene and rehydrated through a graded ethanol series. Endogenous peroxidase activity was quenched by incubation with 3% (v/v) hydrogen peroxide for 10 min. Antigen retrieval was performed by treating the sections with 0.4% pepsin (Servicebio, GC305004) at 37°C for 30 min. To block nonspecific binding, sections were incubated with 10% goat serum at 37°C for 30 min. The slides were then incubated overnight at 4°C with primary antibodies against the target proteins, followed by incubation with appropriate secondary antibodies. Finally, immunoreactivity was visualized using a standard detection system and examined under a light microscope.

Adeno-associated virus (AAV) transfection in vivo

Six- to eight-week-old SD rats were randomly assigned to three experimental groups (n = 6 per group) and housed under standard laboratory conditions. After anesthesia with 2% sodium pentobarbital (40 mg/kg, intraperitoneally), 1 μL of AAV-shUCHL3, AAV-shCtrl, AAV-UCHL3, and AAV-Con (both at a titer of 1.5 × 10¹² viral genomes/mL) was injected into the NP region of the caudal IVDs between the Co7/8 once weekly for four consecutive weeks using a 33-gauge Hamilton syringe. Two months post-injection, the animals were euthanized, and the NP tissues were harvested for subsequent radiological and histological analyses. The AAV vectors were provided by GenePharma Co., Ltd. (Shanghai, China).

Synthesis of PDA

Dopamine hydrochloride (9 mg) was precisely weighed and dissolved in 90 mL of deionized water. The pH of the solution was adjusted to 9–10 by adding NaOH, followed by stirring in the dark at 60°C for 5 h using a thermostatic magnetic stirrer. The resulting suspension was then centrifuged at 15,557 × g for 30 min at 4°C. The supernatant was discarded, and the pellet was resuspended in deionized water and re-centrifuged under identical conditions. This washing procedure was repeated three times to ensure thorough purification. The final black precipitate was resuspended in PBS to form a stable dispersion.

Synthesis of PDA@si-UCHL3

An aqueous solution of PDA nanoparticles was mixed with si-UCHL3 in buffer at mass ratios ranging from 10:1 to 100:1 (PDA:siRNA). The mixture was gently vortexed or pipetted to ensure uniformity while minimizing bubble formation. The suspension was incubated at 4°C for 1–2 h to facilitate efficient siRNA loading. After incubation, the mixture was centrifuged at 12,857 × g for 10 min at 4°C, and the supernatant, containing unbound siRNA, was carefully discarded. The pellet was washed twice with RNase-free 10 mM PBS (pH 7.4), with each wash followed by centrifugation under identical conditions (12,857 × g, 10 min, 4°C) to remove any residual free siRNA.

Characterization of PDA@si-UCHL3

The morphology of PDA and PDA@si-UCHL3 nanoparticles was examined using TEM. Particle size and zeta potential were measured using a Malvern Zetasizer Nano-ZS90 (Malvern Instruments, UK).

Statistical analysis

Data are presented as mean ± standard deviation. Statistical analyses were performed using GraphPad Prism 10.0 software. For comparisons between two independent groups, an unpaired two-tailed Student’s t-test was employed. For experiments involving three or more groups with normally distributed data, two-way analysis of variance (ANOVA) was conducted, followed by Tukey’s post hoc multiple comparisons test. A p value < 0.05 was considered statistically significant.

Funding Statement

The work was supported by Beijing Natural Science Foundation [L254007, 7264285]; National Natural Science Foundation of China [82572800, 12502375]; Beijing Hospital Management Center ‘Yangfan’ Program for Clinical Technology Innovation Projects [ZLRK202517]; Anhui Provincial Natural Science Foundation [2508085Y056]; Anhui Provincial Health and Health Research Project [AHWJ2024Aa30336]; Anhui Provincial Traditional Chinese Medicine Inheritance and Innovation Research Project [2025CCCX211]; Clinical and Translational Research Project of Anhui Province [202527c10020068]; National Key Research Development Program of China [2025YFC2511600]; [AOS-Startup-22–009]; National Key Research & Development Program of China [2024YFC3607001].

Disclosure statement

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

Data availability statement

The data supporting the findings of this study are available from the corresponding author upon reasonable request.

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/15548627.2026.2679172.