SOD1 lactylation impair its enzymatic activity by conformational change to aggravate intervertebral disc degeneration

Yuyao Zhang; Yu Zhai; Chao Liu; Minghang Chen; Yang Zhang; Zhiqun Bian; Xian Chang; Zhilei Hu; Jianmin Li; Chao Zhang; Zhenhong Ni; Yangli Xie; Lin Chen; Minghan Liu; Changqing Li

doi:10.1038/s41467-026-69127-3

. 2026 Feb 28;17:3318. doi: 10.1038/s41467-026-69127-3

SOD1 lactylation impair its enzymatic activity by conformational change to aggravate intervertebral disc degeneration

Yuyao Zhang ^1,^#, Yu Zhai ^1,^✉,^#, Chao Liu ¹, Minghang Chen ¹, Yang Zhang ¹, Zhiqun Bian ¹, Xian Chang ¹, Zhilei Hu ¹, Jianmin Li ^1,², Chao Zhang ¹, Zhenhong Ni ³, Yangli Xie ⁴, Lin Chen ^3,^4,^✉, Minghan Liu ^1,^✉, Changqing Li ^1,^5,^6,^✉

PMCID: PMC13066621 PMID: 41764208

Abstract

Lactate accumulation is a hallmark and contributing factor of intervertebral disc degeneration (IVDD), while the role of protein lactylation caused by lactate accumulation in IVDD remains unclear. Via metabolomics, single-cell RNA-sequencing analysis, and lactylation proteomics, we reveal the lactylome landscape in IVDD and identified superoxide dismutase 1 (SOD1) lactylation at lysine 123 (SOD1^K123la) as crucial for IVDD aggravation. Using in vitro site-directed mutagenesis, in vivo generation of SOD1^K123R mutant male rats, and in silico molecular dynamics simulations, we find that SOD1^K123la alters SOD1 conformation and impairs its enzymatic activity, and induces oxidative damage, and activates p53 pathway in nucleus pulposus cells (NPCs). Notably, we identify a small molecule ZL-01 that inhibits SOD1^K123la. NPC-targeted delivery of ZL-01 via collagen type II-targeted peptide-modified extracellular vesicles alleviated IVDD in male rats. Together, these findings clarify the mechanism by which SOD1^K123la promotes IVDD aggravation and provide a promising therapeutic strategy for IVDD.

Subject terms: Post-translational modifications, Musculoskeletal system, Virtual screening

Lactylation of SOD1 at lysine in nucleus pulposus cells and aggravates intervertebral disc degeneration (IVDD) in male rats.

Introduction

Among musculoskeletal disorders, intervertebral disc degeneration (IVDD) is the primary cause of discogenic pain and associated disability and productivity loss in the workforce worldwide, and it is more prevalent in females, imposing annual costs exceeding $100 billion globally^1,2. Currently, over 400 million individuals suffer from IVDD worldwide, a figure that is expected to increase because of the aging of the population in many countries, thereby imposing substantial economic and social burdens^3,4. Unlike other degenerative diseases, IVDD occurs earlier in human life and often starts during early adolescence. The early onset of IVDD predisposes individuals to various spinal diseases, including disc herniation, spinal stenosis, and spinal deformities. Current clinical interventions for IVDD involve predominantly conservative management strategies and invasive surgical procedures. While these therapeutic approaches may provide symptomatic relief, they do not effectively halt the progression of IVDD.

The intervertebral disc is the largest avascular organ in the human body, with its nucleus pulposus (NP), a unique cartilaginous structure, that naturally exists in a hypoxic and low-nutrient microenvironment due to the lack of direct blood supply⁵. Owing to its avascular nature, the nutritional supply and metabolite clearance of NP rely entirely on passive diffusion through adjacent cartilage endplates; this inefficient transport mechanism not only limits nutrient availability but also leads to the continuous accumulation of metabolic byproducts within NP tissue⁶. During the pathogenesis of IVDD, hypoxia and nutrient deprivation in the NP are further exacerbated, which significantly accelerates the accumulation of various metabolites that may serve as detrimental stimuli in IVDD^6–8. Among these accumulated metabolites, elevated lactate levels are considered a characteristic and potential contributing factor of IVDD⁹. In human IVDD, the lactate concentration can exceed 10 mM, which is more than 10 times higher than the normal concentration in plasma¹⁰. Elevated lactate levels contribute to oxidative damage in nucleus pulposus cells (NPCs), and mitigating lactate accumulation can alleviate the progression of IVDD^11,12. Given the intimate associations among the hypoxic microenvironment of intervertebral discs, lactate metabolism, and NPC dysfunction, investigating the role of lactate in IVDD is critically important for understanding disease mechanisms and developing targeted therapeutic interventions.

Previous studies have shown that lactate participates in the regulation of various physiological and pathological processes via a post-translational modification known as protein lysine lactylation^13–20. Lactate can be converted into lactoyl-CoA and used for lactoyl-modification of specific lysine residues¹⁴. While protein lactylation was initially discovered on histones, a recent study revealed its extensive occurrence on non-histone protein lactylation¹⁵. Protein lactylation is commonly observed in biological processes characterized by lactate accumulation, such as tumor metastasis, tissue fibrosis, and inflammatory responses^16–18. This post-translational modification contributes to various biological alterations, including oxidative stress and cellular senescence^19,20. Although the study has reported the presence of lactylated protein in NPCs²¹, the landscape of the lactylome in IVDD is largely unexplored. Moreover, the definitive role of protein lactylation and its potential involvement in regulating oxidative damage in NPCs warrant further investigation.

In this work, we explored the role and mechanism of lactylation in IVDD. By combining metabolomics, single-cell RNA-sequencing (scRNA-seq), and lactylation proteomics analyses, we found that protein lactylation contributes to oxidative damage in NPCs during IVDD, and we characterized the landscape of protein lactylation in rat degenerative NP tissues. Notably, in the anti-oxidative enzyme superoxide dismutase 1 (SOD1), the lactylation level at lysine 123 (SOD1^K123la) was distinctly increased during IVDD. Genetic mutation mediated inhibition of SOD1^K123la both in vivo and in vitro can alleviate oxidative damage by affecting the p53 pathway. Molecular dynamics simulations indicated that SOD1^K123la impaired SOD1 activity by altering its conformation. Further in silico screening identified ZL-01 as a lactylation inhibitory compound targeting SOD1^K123la. NPCs-targeted ZL-01 delivery achieved effective IVDD alleviation in rats. Collectively, these findings revealed a potentialtherapeutic target for treating IVDD.

Results

Multi-omics analysis reveals the lactylome landscape and the association between oxidative damage and lactylation during IVDD

Two scRNA-seq datasets from the NP tissue of patients with mild (Pfirrmann grade II or III) and severe (Pfirrmann grade IV or V) IVDD were retrieved, including 14 patients who were diagnosed with burst fractures, lumbar disc herniations, and required discectomy and/or intervertebral fusion. Notably, it is difficult to obtain Pfirrmann grade I intervertebral disc tissue. Therefore, the datasets only contain intervertebral disc tissues of grades II to V. Seven main clusters representing different cell types were identified, among which NPCs accounted for the majority (Fig. S1A). Human NPCs are divided into six different clusters: homeostatic, adhesion, effector, hypertrophic and two kinds of fibrotic NPCs (Fig. S1B, C). Homeostatic NPCs account for a dominant proportion in mild-degenerated NP tissues and decrease with the progression of IVDD, while two fibrotic NPCs showed an increasing trend from grade II to grade V in NP tissues (Fig. 1A, B). Gene ontology (GO) analysis suggests that the differentially expressed genes (DEGs) of NPCs from mild and severe IVDD were enriched in several terms (Fig. S1D), and response to oxidative stress, cellular senescence, response to hypoxia, and glycolytic process were upregulated with the IVDD progression as shown by cellular pseudotime analysis (Fig. 1C–F). NP tissue and primary human NPCs were collected from patients with mild and severe IVDD (Fig. 1G). Patient details were presented in Supplementary Table 1. Dihydroethidium (DHE) assays and flow cytometry revealed that the superoxide radical (O₂^-) content and reactive oxygen species (ROS) content in the Severe group were elevated (Fig. S2A–C). DNA damage and cellular senescence markers, including γ-H2AX intensity and β-galactosidase (β-gal) positive cells percentage, were increased in the Severe group (Fig. S2D–G). A metabolomic evaluation was performed using human NP tissue from the Mild and Severe groups, and the results revealed that the lactate content was significantly increased in the Severe group (Fig. 1H, S1H). Degeneration of rat caudal intervertebral discs could be induced by puncture or natural aging. After the puncture-induced disc degeneration (PIDD) and aging-induced disc degeneration (AIDD) rat models were established, the degree of IVDD was evaluated by magnetic resonance imaging (MRI) and histological staining (Fig. S3A–H). The lactate content in the NP tissues of IVDD patients, PIDD rats, and AIDD rats were elevated (Fig. S3I–K). Meanwhile, linear regression analysis revealed that the O₂^- content, ROS content, γ-H2AX intensity, and β-gal-positive cells percentage were positively correlated with the lactate content in human NP tissue (Fig. S3L), suggesting that increased lactate content is related to oxidative damage in NPCs.

Fig. 1 — A UMAP visualization of each subpopulation of human NPCs in the Mild and Severe NP samples (n = 7 individual samples). B Proportion of each subpopulation of NPCs in the Mild and Severe groups. C–F Characterization scores of cell-pseudotime for four biological processes in NPCs. G MRI images of patients with mild or severe IVDD and the morphology of the corresponding NP tissue. Red arrows indicate the intervertebral discs. H Heatmap of the metabolomics results for NP tissue from patients of Mild and Severe groups (n = 12 individual samples). I The expression level of pan-Kla in the NP tissue from patients of Mild and Severe groups (n = 3 individual samples). The samples derive from the same experiment and that gels were processed in parallel. J Scheme of the whole-protein lactylation proteome analysis of NP tissue from Ctrl rats and PIDD rats (n = 3 individual samples). Created in BioRender. Yuyao, Z. (2026) https://BioRender.com/v015adw. K, L Heatmap and histogram showing differentially modified proteins and lysine sites. M Rose diagram showing the subcellular localization of differentially lactylated proteins. N GSEA showing the enrichment of lactylated proteins related to cellular response to oxidative stress. O Radar diagram showing the top 20 differentially lactylated sites associated with cellular response to oxidative stress. GSEA employs a two-tailed non-parametric permutation test to calculate the p value, and multiple comparison adjustment is performed via false discovery rate correction. PIDD puncture-induced disc degeneration, GSEA gene set enrichment analysis. Source data are provided as a Source Data file.

The level of whole lysine lactylation (pan-Kla) was evaluated via western blot, and increased pan-Kla levels were observed in the NP tissues of IVDD patients, PIDD rats, and AIDD rats (Fig. 1I, S3M–Q). Besides, the levels of another two crucial types of acylation, pan-acetylation (pan-Kac) and pan-succinylation (pan-Ksuc), were not significantly altered (Fig. S3R, S). Next, whole-protein lactylation proteomes were generated from the NP tissue of PIDD rats and control rats (Fig. 1J). All replicated samples exhibited ideal reliable correlations and repeatability (Fig. S4A–C). A total of 8168 peptides and 4928 lactylated peptides were identified after passing quality control and were distributed within a reasonable range (Fig. S4D, E). A total of 435 differentially lactylated proteins and 963 differentially lactylated lysine sites were identified (Fig. 1K, L). 54.6% of lactylated proteins were localized in the cytoplasm, 19.1% in the nucleus, and 9.2% in the mitochondria (Fig. 1M). GO analysis revealed that the cellular response to oxidative stress term was enriched (Fig. S4F). Gene set enrichment analysis (GSEA) validated the upregulated expression of lactylated proteins related to cellular response to oxidative stress (Fig. 1N), and the top 20 lactylation sites are displayed in the radar diagram (Fig. 1O). Lactylation motif analysis showed that the K_xxxxxx_KG and K_xxxxx_K sequences exihibited relatively high motif scores (Fig. S4G, H). These results indicate that lactylation-dependent mechanisms may contribute to the oxidative damage to NPCs during IVDD.

Lactylation promotes oxidative damage in NPCs and aggravates IVDD in rats

A lactate injection-induced disc degeneration model (LAIDD) was developed in rats (Fig. 2A). Four weeks after lactate injection, the rats exhibited elevated Pfirrmann grades and aggravated histological alterations (Fig. 2B–E), with elevated O₂^- content and the total ROS level in the NP tissue (Fig. 2F–H). The γ-H2AX intensity and β-gal-positive cells percentage in primary NPCs derived from LAIDD rats were elevated (Fig. 2I, J and S5A, B). Besides, we also observed the upregulation of lactylation level in the NP tissue of LAIDD rats (Fig. 2K and S5C).

The effects of lactate on primary rat NPCs were further investigated in vitro (Fig. 2L). With increasing lactate concentration, the NPCs presented elevated ROS and O₂^- levels (Fig. 2M–O and S5D), with increased γ-H2AX intensity and β-gal-positive cells percentage (Fig. 2P and S5E–G). Similarly, we also observed the upregulation of lactylation level in NPCs treated with lactate (Fig. 2Q and S5H). These results indicate that lactylation is an inducing factor of NPC oxidative damage and subsequent IVDD.

SOD1 lactylation exacerbated oxidative damage in NPCs via the p53 pathway

Among the oxidative stress-related lactylated proteins identified above, cytoplasmic Cu/Zn superoxide SOD1, one of only two enzymes in the body capable of scavenging O₂^- to prevent further oxidative damage, exhibited a prominent increase in lactylation. SOD1 is closely related to oxidative damage resistance during IVDD²². We therefore explored whether lactate induces oxidative damage in NPCs by mediating the lactylation of SOD1. Immunoprecipitation (IP) and immunoblotting revealed that the level of SOD1 protein showed no significant difference in lactate-treated primary NPCs and the NP tissue of PIDD and AIDD rats, while the level of lactylated SOD1 was increased (Fig. 3A–C and S6A–F). Meanwhile, no significant differences were observed in acetylation or succinylation, two reported SOD1 modifications, in the NP tissue of PIDD rats (Fig. S6G–J). SOD1 enzymatic activity was reduced in lactate-treated primary NPCs (Fig. 3D). Moreover, increased malondialdehyde (MDA) levels and decreased total antioxidant capacity (T-AOC) were observed in lactate-treated primary NPCs (Fig. S6K, L). Similar results were observed in the NP tissue of the PIDD and AIDD rats (Fig. 3E, F and S6M–P). These results suggest that SOD1 lactylation may reduce the oxidative damage resistance capacity of NPCs. SOD1 has two lactylation sites at K10 and K123, as shown by the LC-MS/MS spectroscopy (Fig. 3G and S7A), both of which are highly conserved among species (Fig. 3H).

Fig. 3 — A–C Western blot evaluating SOD1 lactylation levels after different treatments (n = 3 individual samples). D–F SOD1 activity evaluation after different treatments (n = 6 individual samples). The samples derive from the same experiment and that gels were processed in parallel. G LS-MS/MS spectra analysis of K123 in SOD1. H Sequence homology around the K123 and K10 in SOD1 among different species. I Western blot evaluating SOD1 lactylation levels (n = 6 individual samples). The samples derive from the same experiment and that gels were processed in parallel. J SOD1 enzymatic activity in lactate-treated NPCs with different mutating treatments (n = 6 individual samples). K, L ROS levels in lactate-treated NPCs with different mutating treatments (n = 6 individual samples). M DHE staining in lactate-treated NPCs with different mutating treatments (n = 6 individual samples). Scale bar = 50 μm. N γ-H2AX staining in lactate-treated NPCs with different mutating treatments (n = 6 individual samples). Scale bar = 10 μm. O Heatmap showing DEGs between lactate-treated NPCs with Ctrl or SOD1^K123R mutation (n = 4 individual samples). P KEGG analysis showing the enriched pathway associated with the DEGs. Q GSEA showing the enrichment of DEGs related to p53 signaling pathway. R Cluster heatmap of 37 genes associated with cellular senescence. S Interaction network diagrams for the top 10 hub genes. T–W Western blot results and quantifications of the protein levels of P53, P21 and GADD45α (n = 3 individual samples). The samples derive from the same experiment and that gels were processed in parallel. Data are presented as mean ± SD, and p value were calculated using two-tailed Student’s t-test for two-group comparisons and one-way ANOVA followed by Dunnett’s test for multiple-group comparisons, KEGG analysis employs a Fisher’s exact test to calculate the p value, and GSEA employs a two-tailed non-parametric permutation test to calculate the p value, and multiple comparison adjustment is performed via false discovery rate correction. DHE dihydroethidium, ROS reactive oxygen species. Source data are provided as a Source Data file.

To investigate the effects of SOD1^K123la and SOD1^K10la on SOD1, a lysine-to-arginine mutation (K-to-R) was used to mimic the delactylated state of SOD1 in vitro. Endogenous SOD1 in NPCs was knocked down by short hairpin RNA (shRNA) (Fig. S7B). and then overexpressed with SOD1^WT, SOD1^K10R, or SOD1^K123R, and without causing off-target effects (Fig. S7C–E). SOD1^K123R significantly reduced the SOD1 lactylation level (Fig. 3I and S7F–G). SOD1^K123R, but not SOD1^K10R, reversed the decreased SOD1 enzymatic activity and T-AOC while alleviating the increased MDA level in lactate-treated NPCs (Fig. 3J, S7H, I). Flow cytometry and DHE staining revealed that SOD1^K123R alleviated oxidative damage (Fig. 3K–M and S7J), and γ-H2AX and β-gal staining revealed that SOD1^K123R ameliorated DNA damage and cellular senescence in lactate-treated NPCs (Fig. 3N and S7K–M). Native PAGE results showed that the K123R mutation did not change the monomeric and dimeric contents of SOD1 (Fig. S8A–C).

To further elucidate the mechanism by which SOD1^K123la affects oxidative damage in NPCs, we conducted RNA sequencing (RNA-seq) on lactate-treated SOD1^K123R-NPCs and lactate-treated NPCs. A total of 5716 DEGs were identified, including 3620 upregulated genes and 2096 downregulated genes (Fig. 3O). Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis showed the DEGs were associated with the cellular senescence and p53 signaling pathway (Fig. 3P). GSEA revealed that genes related to the p53 signaling pathway were downregulated (Fig. 3Q). We analyzed the overlap between the identified DEGs and the genes enriched in cellular senescence and identified 37 genes for further investigation (Fig. 3R and S9A). Then, 10 hub genes were selected after using the STRING database and Cytoscape software (Fig. 3S). The transcription factors (TFs) of the 10 hub genes were predicted based on ChEA3 database, and among the top 10 TFs, TP53 exhibited a potent relation with the hub genes (Fig. S9B, C). Western blot revealed significant decreases in the protein levels of P53 and its downstream P21 and GADD45α in the NPCs of the lactate + SOD1^K123R group (Fig. 3T–W). Meanwhile, lactate injection in vivo (Fig. S9D–G) and lactate treatment in vitro (Fig. S9H–K) increaded P53, P21, and GADD45α expression levels. Collectively, these results indicate that SOD1^K123la induces oxidative damage in NPCs by activating the downstream p53 pathway.

Previous studies have reported that G-protein-coupled receptor 81 (GPR81) activation by lactate also regulates cellular biological behaviors through a mechanism different from lactate-mediated intracellular lactylation^23,24. Lentiviruses were used to knockdown monocarboxylate transporter 1 (MCT1) and monocarboxylate transporter 4 (MCT4), which are associated with lactate uptake and lactate efflux, and GPR81 (Fig. S10A). MCT1 knockdown reduced the lactate content and MCT4 knockdown elevated the lactate content within NPCs (Fig. S10B). In lactate-treated NPCs, MCT1 knockdown alleviated oxidative damage and MCT4 knockdown promoted oxidative damage, whereas GPR81 knockdown resulted in no apparent mitigation (Fig. S10C–J), indicating that lactate-induced oxidative damage in NPCs is attributed mainly to the intracellular lactylation.

SOD1^K123la aggravated IVDD by impairing SOD1 enzymatic activity

The specific role of SOD1^K123la in IVDD was further investigated in vivo (Fig. 4A). SOD1^K123R mutant rats were generated by mutating the lysine AAA codon to the arginine AGG codon (Fig. 4B, C). SOD1^K123R rats did not significantly differ from SOD1^WT rats in terms gross phenotype (Fig. S11A–K). A PIDD model was then constructed. The MRI results and histological staining revealed that, compared with SOD1^WT-PIDD rats, SOD1^K123R-PIDD rats presented alleviated degeneration (Fig. 4D–G). A SOD1^K123la-specific antibody that targeted a modified peptide [VVHE-(lactyl)K-QDDLGKGGC] of SOD1 was constructed and exhibited ideal specificity for the SOD1^K123la protein in dot blot assays (Fig. 4H), as verified in lactate-treated NPCs with or without SOD1 knockdown by immunofluorescence and western blot (Fig. S12A–F). In the NP tissue of SOD1^K123R-PIDD rats, the immunofluorescence and western blot results revealed that the SOD1^K123R reduced SOD1^K123la levels (Fig. 4I, J and S13A, B), with declined P53, P21 and GADD45α levels (Fig. 4J and S13C–E). The DHE assay results revealed that the O₂^- content was reduced in the NP tissue of SOD1^K123R-PIDD rats (Fig. 4K), with restored SOD1 enzymatic activity and T-AOC and decreased MDA levels (Fig. 4L and S13F, G). Primary NPCs were extracted from SOD1^K123R- and SOD1^WT-PIDD rats. The ROS content was reduced in SOD1^K123R-PIDD rat-derived primary NPCs (Fig. 4M, N), and the γ-H2AX intensity and percentage of β-gal-positive cells were decreased (Fig. 4O, P and S13H, I). These results suggest that decreasing SOD1^K123la levels can alleviate IVDD by increasing SOD1 enzymatic activity and reducing oxidative damage.

To clarify the effect of SOD1^K123la on SOD1 enzymatic activity, molecular dynamics simulations were performed (Fig. 5A). In the 100 ns simulation, the interaction of the SOD1^K123la protein with O₂^- was not stable compared with SOD1^WT (Fig. 5B). Both the root mean square deviation (RMSD) and radius of gyration (RG) fluctuated significantly after 90 ns of simulation in the SOD1^K123la system, indicating decreased system stability (Fig. 5C and S14A). Similarly, the solvent accessible surface area (SASA) and root mean square fluctuation (RMSF) characterized the enhanced variation in protein flexibility in the SOD1^K123la system (Fig. 5D and S14B, C). Compared with SOD1^WT, the binding energy between O₂^- and SOD1^K123la was lower at 60–100 ns (−251.38 ± 15.66 kJ/mol versus −375.06 ± 18.43 kJ/mol), indicating a weakened enzyme-substrate binding capacity (Fig. 5E). This structural shift may result in structural alterations in the active site and the dislocation of copper ions from the active site and hinder the binding of O₂^- and SOD1. Analysis of the secondary structure of SOD1 revealed an elevated coil structure and a reduced sheet structure and turn structure after lactylation, indicating a more disordered protein structure (Fig. S14D–F). The docking conformations revealed that residues Glu122 and Arg144 are closer to the SOD1 active site (copper-binding regions) in the SOD1^K123la system (Fig. 5F). This proximity promoted hydrogen bond formation between Asn140 and lactylated K123, which then caused significant structural alterations in the adjacent region, especially Arg144. The hydrogen bonds between Arg144 and O₂^-, Arg144 and Cys58, and Arg144 and Gly62 were disrupted, and new hydrogen bonds between Arg144 and Lys136 and between Arg144 and His49 formed (Fig. 5G). These conformational alterations may constitute the molecular basis for the reduction in the enzymatic activity of lactylated SOD1.

ZL-01 acts as a potential specific inhibitor of SOD1^K123la

The results above indicate that the inhibition of SOD1^K123la may alleviate IVDD, suggesting it as a potential therapeutic approach. Previous studies have reported the existence of lactyltransferases, including AARS1, AARS2, GCN5, CBP, P300, TIP60, HBO1, and MYST1. Furthermore, immunoprecipitation-mass spectrometry (IP–MS) analysis revealed variations in the expression of SOD1-interacting proteins before and after lactylation (Fig. 5H, I and S14G). Comparison of the 923 identified proteins with the 8 reported lactyltransferases revealed that AARS1 could interact with lactylated SOD1 in NPCs (Fig. 5J). The expression of the above 8 lactyltransferases were knocked down in NPCs (Fig. S14H), and immunofluorescence and western blot results revealed that AARS1 and P300 may be involved in SOD1^K123la (Fig. 5K, L and S14I, J). HDAC1-3 and SIRT1-3 have been shown to be potential delactylases²⁵, and they were thus overexpressed in NPCs (Fig. S14K). Immunofluorescence and western blot results revealed that SIRT1 and HDAC3 overexpression reduced SOD1^K123la levels (Fig. S14L–O).

To increase the specificity of SOD1^K123la inhibition, virtual screening was performed to identify potential compounds with SOD1^K123la inhibitory activity. A total of 1.6 million commercially available compounds were screened via the K123 residue of SOD1. The top 1000 compounds were docked on the basis of total energy (Fig. 6A). The top 10 compounds that interacted with the pocket of SOD1 containing the K123 residue (Fig. S15A–C). The proper concentrations of the chosen 10 compounds were determined via cell counting kit-8 (CCK-8) assays (Fig. S15D). Compound 3 effectively reduced the SOD1^K123la level and increased SOD1 activity (Fig. S15E, F). The chemical formula of compound 3 is C₁₁H₁₄N₄O₂ and was named ZL-01. Molecular docking further revealed that ZL-01 binds to Thr40, His44, Lys123, and Asn140 of SOD1 (Fig. 6B, C, and S15G). We mutated the aforementioned amino acid residues individually, and microscale thermophoresis (MST) experiments indicated that ZL-01 primarily interacts with SOD1 residues Thr40, His44, and Lys123 (Fig. 6D). DHE staining and flow cytometry revealed that ZL-01 alleviated oxidative damage in lactate-treated NPCs (Fig. 6E–G and S15H) and ameliorated DNA damage and cellular senescence (Fig. 6H, I, K–M, and S15I–K). ZL-01 had no significant effect on the acetylation and succinylation of SOD1 in NPCs (Fig. S16A–F). In contrast, ZL-01 significantly reduced the pan-lactylation induced by lactate in NPCs, which was mainly manifested at the protein molecular weight around 17 kDa (Fig.S S16G–I). ZL-01 effectively reduced SOD1^K123la levels in lactate-treated NPCs (Fig. 6I, J and 6N, O), reversed the decreases in SOD1 enzymatic activity and T-AOC and decreased MDA levels (Fig. 6P and S16J, K). Meanwhile, ZL-01 rescued lactate-induced impairment of SOD1 function in SOD1^WT NPCs, whereas it resulted in no apparent alteration in SOD1 function in SOD1^K123R NPCs. This may have been attributed to the already existing lactylation suppression by SOD1^K123R that protects SOD1 function (Fig. S16L).

Fig. 6 — A Schematic of the experimental process for virtual screening. Created in BioRender. Zhai, Y. (2026) https://BioRender.com/ng5bvmz. B Molecular formula of compound 3 named ZL-01. C Molecular docking showing the binding scheme of ZL-01 and SOD1. D The site at which ZL-01 binds to SOD1 amino acid residues was determined via MST (n = 3 individual samples). E DHE staining of lactate-treated NPCs treated with ZL-01(n = 6 individual samples). Scale bar = 50 μm. F, G ROS content of lactate-treated NPCs treated with ZL-01 (n = 6 individual samples). H γ-H2AX staining of lactate-treated NPCs treated with ZL-01 (n = 6 individual samples). Scale bar = 10 μm. I–M) Western blot and quantifications of SOD1^K123la, P53, P21, and GADD45α protein levels in lactate-treated NPCs treated with ZL-01 (n = 6 individual samples). The samples derive from the same experiment and that gels were processed in parallel. N, O Immunofluorescence results and quantifications of the SOD1^K123la level in lactate-treated NPCs treated with ZL-01 (n = 6 individual samples). Scale bar = 40 μm. P SOD1 activity in lactate-treated NPCs treated with ZL-01 (n = 6 individual samples). Data are presented as mean ± SD, and p value were calculated using one-way ANOVA followed by Dunnett’s test for multiple-group comparisons. ROS reactive oxygen species, FL fluorescence. Source data are provided as a Source Data file.

Targeted delivery of ZL-01 reduced oxidative damage and alleviated IVDD

Different doses of ZL-01 were injected into the NP tissue of PIDD-rats once a week, and assessments of degeneration were performed at 4 weeks post-injury to evaluate the potential therapeutic effects of ZL-01. ZL-01 reduced the histological scores and improve the SOD1 enzymatic activity, but increasing the ZL-01 concentration within a certain range did not significantly increase the therapeutic effect (Fig. S17A–C). Moreover, we found that the local concentration of ZL-01 decreased significantly 48 h after injection (Fig. S17D), which indicated its rapid clearance of ZL-01 from degenerated NP tissue.

To improve the drug retention and therapeutic efficacy of ZL-01, we designed a NPC-targeting ZL-01 system based on extracellular vesicles (EVs) (Fig. 7A). A collagen type II-targeted peptide (CTP) was genetically engineered onto the outer surface of to generate CTP-EVs (C-EVs). Transmission electron microscopy showed that both the EVs and C-EVs had a typical disc-like structure (Fig. 7B), and highly expressed positive EV markers CD81 and TSG 101, and weakly expressed negative EV marker calnexin (Fig. 7C). Nanoparticle tracking analysis showed that most of the EVs had diameters between 100 and 200 nm (Fig. 7D, E). Immunofluorescence staining revealed that compared with GFP-labeled EVs, GFP-labeled C-EVs were more efficiently taken up by NPCs (Fig. 7F). ZL-01 was loaded into the C-EVs to obtain ZL-01-loaded C-EVs (C-EVs^Z), and sonication at a concentration ratio of 1:6 resulted in better ZL-01 loading efficiency (Fig. S18A). Approximately 76% of ZL-01 could be released within 96 h after loading (Fig. S18B).

Fig. 7 — A Flow chart of the animal experiments in which C-EVs^Z was used to treat IVDD. Created in BioRender. Yuyao, Z. (2026) https://BioRender.com/d75w54p. B Representative transmission electron microscope images of EVs and C-EVs (n = 3 individual samples). Scale bar = 150 nm. C Western blot identification of EVs and C-EVs (n = 3 individual samples). The samples derive from the same experiment and that gels were processed in parallel. D, E Nanoparticle tracking analysis of EVs and C-EVs. F Representative images of DiD-labeled NPCs (red) uptake of EVs and C-EVs (green) after 4 h of treatment (n = 3 individual samples). Scale bar = 20 μm. G, H MRI images of different groups (n = 6 individual rats). Red arrows indicate the intervertebral discs. I, J SO & FG staining of different groups (n = 6 individual rats). Scale bar = 500 μm. K Immunofluorescence evaluating SOD1^K123la level in the NP tissue of different groups (n = 6 individual rats). Scale bar = 250 μm. L Western blot evaluation of relative protein levels in the NP tissue of different groups (n = 6 individual rats). The samples derive from the same experiment and that gels were processed in parallel. M DHE assay evaluation of O₂^- content in the NP tissue of different groups (n = 6 individual rats). N SOD1 activity in the NP tissue of different groups (n = 6 individual rats). O, P MDA and T-AOC levels in the NP tissue of different groups (n = 6 individual rats). Q, R ROS content of primary NPCs extracted from different groups (n = 6 individual samples). S γ-H2AX staining of primary NPCs extracted from different groups (n = 6 individual samples). Scale bar = 10 μm. Data are presented as mean ± SD, and p value were calculated using one-way ANOVA followed by Dunnett’s test for multiple-group comparisons. CTP collagen type II-targeted peptide, EV extracellular vesicles, NP nucleus pulposus, ROS reactive oxygen species, PIDD puncture-induced disc degeneration, DHE dihydroethidium, MDA malondialdehyde, T-AOC total antioxidant capacity. Source data are provided as a Source Data file.

The biocompatibility of the delivery system was comprehensively examined. There were no significant differences in the major indexes of blood routine, liver function, or body weight after the injection of C-EVs, C-EVs^{Z (1 μM)}, or C-EVs^{Z (5 μM)} into the intervertebral discs of the rats (Fig. S18C–K). No apparent toxicity was observed in hearts, livers, spleens, lungs and kidneys at 4 weeks after drug administration (Fig. S18L–Q). C-EVs^Z effectively increased local ZL-01 retention (Fig. S19A, B). MRI and histological staining revealed that the PIDD rats that received C-EVs^{Z (5 μM)} presented alleviated degenerative alterations (Fig. 7G–J). The protein level of SOD1^K123la was reduced in the NP tissue of C-EVs^{Z (5 μM)}-treated PIDD rats (Fig. 7K, L and S19C, D), with decreased P53, P21, and GADD45α protein levels and the O₂^- content (Fig. 7L, M, and S19E–G). Moreover, the SOD1 activity and T-AOC levels in the NP tissue of C-EVs^{Z (5 μM)}-treated PIDD rats increased, and the MDA leveldecreased (Fig. 7N–P). Compared with PIDD model rat-derived NPCs, primary NPCs extracted from C-EVs^{Z (5 μM)}-treated PIDD rats presented lower total ROS levels (Fig. 7Q, R) and alleviated DNA damage and cellular senescence (Fig. 7S and S19H–J). These results suggest that compared with direct injection of ZL-01, C-EVs^Z can maintain drug retention and achieve better therapeutic effects for treating IVDD. A schematic model was drawn to summarize our findings (Fig. 8).

Fig. 8 — Created in BioRender. Yuyao, Z. (2026) https://BioRender.com/z9u42yg. O₂^-: superoxide radical; ROS: reactive oxygen species; LA: lactate.

Discussion

As a naturally hypoxic tissue, the NP relies heavily on glycolysis for cellular energy metabolism. During IVDD, endplate sclerosis exacerbates pathological hypoxia, disrupting cellular homeostasis. High levels of oxidative stress induce cellular senescence in the disc and suppress mitochondrial function, which in turn shifts cellular energy production toward glycolysis²⁶. ScRNA-seq showed that with the progression of degeneration, the scores of the terms response to oxidative stress, cellular senescence, response to hypoxia and glycolytic process increases. The accumulation of metabolites contributes to oxidative damage and cellular senescence in various degenerative diseases^27,28. Our metabolomic analysis revealed that the levels of AMP, lactate, 2,3-DPG, FBP, and GAP/DHAP were significantly elevated during IVDD. Existing studies have not demonstrated the roles of AMP, 2,3-DPG, FBP, or GAP/DHAP in IVDD, osteoarthritis, or degeneration of similar cartilaginous tissues. However, lactate is a stimulator of organismal aging and tissue degeneration, and numerous studies have shown a strong correlation between lactate accumulation and many diseases^29–32. In healthy intervertebral discs, lactate is efficiently transported and cleared by the cartilage endplate. However, dysregulation of endplate clearance mechanisms alongside disruptions in NPC glycolysis during IVDD can result in lactate accumulation^33,34. Lactate can provide a lactyl group to the ε-amino group of a lysine residue, thereby mediating lactylation and regulating the functions of proteins. Our results revealed that the lactylation level increased in degenerated NP tissue. Notably, protein acetylation and succinylation levels were not significantly altered, indicating that these two modifications may not be responsible for IVDD.

Lactylome analyses have revealed that lactylation occurs widely among non-histone proteins^15,35. However, this technique has not been used to detect protein lactylation in IVDD. Lactylation can involve three isomers: L-lactyl-lysine (K^L-la), N-ε-(carboxyethyl)-lysine (K^ce), and D-lactyl-lysine (K^D-la)³⁶. In light of the capability of lactylation proteomes to identify lactylated proteins and their modification sites, we conducted a L-lactyl-lysine-lactylation proteome analysis on both degenerative and healthy NP tissues. The lactylated proteins were enriched in pathways related to oxidative stress. We administered lactate to healthy intervertebral discs and found that lactate administration elevated protein lactylation levels and induced degenerative alterations in the treated discs. In addition, in vitro lactate supplementation directly increased protein lactylation levels and oxidative damage in cultured NPCs. These results confirmed that protein lactylation induces IVDD progression and oxidative damage.

Our study revealed that lactylation occurred in multiple oxidative stress-related proteins. Among them, SOD1 not only showed a relatively high fold change but was also one of the only two enzymes in the body responsible for scavenging O₂^-, playing a crucial role in maintaining redox homeostasis in the intervertebral disc^37,38. Mice deficient in SOD1 or SOD2 demonstrate heightened oxidative stress and expedited senescence^39–41. Maintaining the normal functions of SOD1 and SOD2 is essential for intervertebral disc redox homeostasis^22,42. Unlike SOD2, which is localized exclusively in the mitochondria, SOD1 is distributed across the mitochondria, cytoplasm, and nucleus. Our study revealed that lactate treatment in NPCs led to a significant reduction in SOD1 enzymatic activity, with no significant alteration in SOD1 protein level. K-to-R mutation was used to mimic the delactylated state of the protein⁴³, and these experiments reveals that only SOD1^K123R alleviated oxidative damage and elevated SOD1 activity in lactate-treated NPCs. RNA-seq results revealed that the DEGs in SOD1^K123R mutated NPCs were enriched in cellular senescence and p53 pathway. Elevated ROS level can further activate the p53 pathway to induce cellular senescence in NPCs^44,45. Further in vitro experiments confirmed that SOD1^K123la can activate p53 pathway in NPCs. Additionally, SOD1^K123R mutant rats were generated via the CRISPR-Cas9 technique⁴⁶. The degenerative phenotype of SOD1^K123R rats was alleviated after IVDD induction, with reduced the p53 pathway activation and elevated SOD1 activity. These results suggest that SOD1^K123la impairs SOD1 enzymatic activity, thereby contributing to oxidative damage and cellular senescence in NPCs by activating the p53 pathway and further exacerbating IVDD. In addition to SOD1, other antioxidative proteins that undergo lactylation include PPIA, PRDX2, NQO1, and ANXA1. We believe that their lactylation of these proteins is also associated with the regulation of redox homeostasis in the intervertebral disc, although their specific mechanisms of action remain unclear. Future studies will further explore the effects of lactylation on these proteins and their respective modification sites.

SOD1 is a homodimeric protein with each monomer fold consisting of eight-stranded Greek key β-barrels. The K123 residue lies within the β7/β8 loop of SOD1, also known as the electrostatic loop, which encompasses residues 122–144 and is thought to participate in the shuttling of O₂^- toward the SOD1 active site^47,48. The K123 residue has been reported to be a crucial post-translational residue in SOD1⁴⁹. Molecular dynamics simulations revealed that the electrostatic loop was positioned closer to the copper-binding regions of SOD1 in the SOD1^K123la system. This proximity promoted hydrogen bond formation between Asn140 and lactylated K123 and induced a notable structural alteration in the electrostatic loop, particularly affecting Arg144. Arg144 plays a crucial role in binding O₂^- and directing it towards the SOD1 active site for enzymatic catalysis^50,51. In the SOD1^K123la simulation system, hydrogen bonds between Arg144 and O₂^-, Cys58, and Gly62 were absent, while new hydrogen bonds were observed between Lys136 and His49. This structural alteration resulted in the dislocation of copper from the active site, potentially due to the dislocation of the positively charged Arg144 side chain, which subsequently repels positively charged copper ions. This SOD1^K123la-induced conformational alteration is analogous but different from the well-known conformational pathogenic mechanisms of SOD1 in amyotrophic lateral sclerosis^52,53.

Apart from the mechanism of SOD1^K123la identified in our work, we notice that succinylation and acetylation can occur at the K123 residue, resulting in the suppression of SOD1 enzymatic activity and the inhibition of mitochondrial respiration, respectively^50,54,55. Considering that the succinylation and acetylation levels of SOD1 did not exhibit significant alterations during IVDD, we surmised that SOD1^K123la serves as a main regulator of SOD1 enzymatic function during IVDD. However, the regulatory network of PTMs is complex, and there is crosstalk exists among many PTMs^56,57. Therefore, we cannot rule out the possibility that SOD1^K123la may affect other PTM levels at the K123 site in other tissues or other diseases. Investigating the crosstalk between SOD1^K123la and other PTMs will be of great significance for future research.

Having established the involvement of SOD1^K123la in NPC oxidative damage, we proceeded to explore whether SOD1^K123la could serve as a therapeutic target for the treatment of IVDD. Lactyltransferases were explored first. Through IP–MS, we identified AARS1 and P300 as potential lactyltransferases for SOD1^K123la. AARS1 is an intracellular lactate sensor that global lysine lactylation by binding to lactate in an ATP-dependent manner to form lactate-AMP, which can then transfer lactate to lysine residues^58,59. P300 has been reported to function as a lactyltransferase promotes lactylation^43,60. Our results also showed that SIRT1 and HDAC3 reduced SOD1^K123la levels. Previous studies verified that HDAC3 potential histone delactylases²⁵, with SIRT1 also capable of delactylating specific non-histone proteins⁴³. However, lactyltransferases and delactylases also play roles in regulating other kinds of protein acylation including acetylation, crotonylation, and succinylation, and also regulating aminoacyl-tRNA synthesis. Considering the poor specificity of these enzymes, targeting these enzymes may result in unpredictable effects in IVDD treatment. A more precise therapeutic strategy is necessary to achieve targeted inhibition of SOD1^K123la. Virtual screening techniques have been employed to identify compounds capable of modulating protein function at specific post-translational modification sites, such as phosphorylated and acetylated sites^61–63. The SOD1 pocket containing the K123 residue was selected for high-throughput virtual screening. Among the 1.6 million compounds screened, ZL-01 was identified as capable of reducing SOD1^K123la levels. Molecular docking results showed that ZL-01 binds to Thr40, His44 and Lys123 of the side chain and to Ala124 and Asn140 of the main chain, which may form a placeholder effect and impede the lactylation of Lys123. In vitro experiments demonstrated that ZL-01 reversed the decrease in SOD1 enzymatic activity and mitigated oxidative damage in lactate-treated NPCs.

Intradiscal ZL-01 injection reduced the histologicscore and improved the SOD1 enzymatic activity, but increasing ZL-01 concentration within a certain range did not significantly increase the therapeutic effect, which may be attributed to the rapid clearance of compounds in degenerated NP tissue as previously reported^64,65. To further enhance the bioavailability of ZL-01 in NPCs while prolonging the duration of action following intradiscal administration, we employed an EV-based drug delivery system previously reported. Additionally, leveraging the high expression of collagen type II in NPCs, we modified the system with CTP (WYRGRL) on its surface to enhance targeting specificity^66,67. CTP-EV^Z was designed and isolated and exhibited high affinity for NPCs. To explore its therapeutic potential, low-dose ZL-01, high-dose ZL-01, and CTP-EV^Z were injected into degenerated NP tissue. CTP-EVs^Z effectively reduced SOD1^K123la levels and oxidative damage in NPCs to alleviate experimental IVDD progression. These results demonstrated that ZL-01 was potentially valuable for clinical translation in the treatment of IVDD.

Chemical probes are small-molecule ligands that target specific biomolecular targets, with selectivity and potency being their key characteristics⁶⁸. Through experiments, we identified that ZL-01 can inhibit the lactylation at the SOD1-K123 site in rat NPCs, and this compound exhibits several features of a chemical probe. However, further studies are required to investigate its cross-species applicability, biochemical potency, and compound properties.

In conclusion, our work demonstrated the landscape of the lactylome induced by lactate accumulation in IVDD via whole-protein lactylation proteomics. In addition, we identified a previously unreported mechanism by which SOD1^K123la induces a conformational change in SOD1 to induce oxidative damage and activate the p53 pathway in NPCs in IVDD. Furthermore, we obtained a small molecule ZL-01 with potent SOD1^K123la inhibitory efficacy. Targeted delivery of ZL-01 to NPCs achieved effective IVDD alleviation, indicating that ZL-01 has promising clinical translational potential for treating IVDD in the future.

Limitations

This research has several limitations. First, the in vitro culture conditions for NPCs used in this experiment cannot accurately simulate the hypoxic and nutrient-poor microenvironment within the intervertebral disc, which may have unknown impacts on the experimental results. Second, although our analyzed scRNA-seq datasets and patient samples did not include data from IVDD patients caused by genetic factors, such as those with scoliosis, we hypothesize that the lactylation signature and profiles of key lactylated proteins may differ in patients with IVDD caused by genetic factors. Investigating these potential distinctions represents a valuable and compelling direction for future research. Third, in the present study, we have only preliminarily verified the therapeutic effect of ZL-01, and have not conducted detailed investigations on the optimal administration strategy, potential adverse effects, and comprehensive pharmacokinetic profiles. Fourth, only the male SD rats were used for in vivo experiments, which to a certain extent limits the sex generalizability of the conclusions. Finally, it is challenging to accurately mimic the clinical reality of human intradiscal administration in rat models because of the limitations of minimally invasive procedures and high-volume injection within the rat intervertebral disc. In future studies, we will use additional animal models of IVDD, such as large animal IVDD models and organoids, to validate the molecular mechanism and therapeutic effects we observed.

Methods

Human NP tissue acquisition

All patient sample collections were reviewed and approved by the Ethics Committee of Xinqiao Hospital, Army Medical University (Approval No. 2023-YD184-01; Approval Date: December 28, 2023). NP tissue samples (solid tissue) from 12 male patients and 12 female patients aged 18–65 years were obtained during intervertebral surgery, following the standard surgical protocols of Xinqiao Hospital, Army Medical University. Sex of the patients was determined based on the assigned. Informed consent and clinical data collection protocols were signed prior to surgery. The inclusion criteria were patients with lumbar fracture, lumbar spondylolysis, intervertebral disc herniation, or lumbar spinal stenosis (diagnosed in accordance with the International Classification of Diseases, 11th Revision, ICD-11) and a body mass index between 18 and 24. The exclusion criteria were diabetes, metabolic diseases, severe chronic diseases, hematological diseases, tumors, lumbar infection, or previous lumbar surgery. All patients were alive at the time of NP tissue collection. Immediately after collection from surgical segments, NP tissues were rinsed with sterile 0.9% NaCl solution to remove residual blood and tissue debris. NP tissues free of blood clots or foreign materials were included in the study, with one portion of the tissue used for the extraction of primary NPCs, and the other flash-frozen in liquid nitrogen within 10 min. After stabilization, NP tissue specimens were transferred to an ultra-low temperature freezer at −80 °C for long-term preservation, and subsequent experiments were to be completed within one month of freezing. Patient details were provided in Supplementary Table 1.

Experimental animals

All the animal experiments in the present study were approved by the Laboratory Animal Welfare and Ethics Committee of the Army Medical University (Approval No. AMUWEC20235167; Approval Date: December 1, 2023). Given the anatomical-physiological similarity of rat and human intervertebral discs, plus well-established IVDD modeling methods, rats were selected for animal experiments. To avoid hormonal interference, we used only male rats; hence, male Sprague-Dawley (SD) rats were purchased from the Chengdu Dashuo Experimental Animal Co., Ltd. The rats were housed in a vivarium with a 12 h light/dark cycle at a temperature of 20–24 °C and a humidity of 40–60%, with unrestricted access to food and water. The animal sample size (α error probability: 0.05; power: 0.8) was determined via the online website http://powerandsamplesize.com/. During the experiment, physical condition monitoring of rats was performed twice weekly, with humane endpoints defined as the occurrence of severe tissue damage or infection. There was no exclusion of animal experimental data in the present study.

Generation of SOD1^K123R rats

Male SOD1^K123R rats were designed and generated via CRISPR-Cas9 by Cyagen Biosciences, Inc. (Guangzhou, China). The rat Sod1 gene (GenBank accession number: NM_017050.1; Ensembl: EnsRNOG00000002115) is located on rat chromosome 11. Five exons were identified, with the ATG start codon in exon 1 and the TAA stop codon in exon 5 (transcript: NM_017050.1). p.K123 is located in exon 5, and exon 5 was selected as the target site. gRNA (target sequences: gRNA-A1, CCAAGTCATCTTGTTTCTCG-TGG; and gRNA-B1, CCA CGA GAAA CAAGATGACT-TGG) and a donor oligo (with a targeting sequence flanked by 120 bp homologous sequences on both sides) were designed. Cas9, gRNA, and donor oligos were coinjected into fertilized eggs for mutant rat production. The p.K123R mutation (AAA to AGG) in the donor oligo was introduced into exon 5 via homology-directed repair. The pups were genotyped via PCR followed by sequence analysis.

Open field test

An open field test was conducted to evaluate the locomotor activity and anxiety-like behavior of the rats. All tests were performed in a sound-attenuated, dimly lit room. The apparatus consisted of a square open-field arena (60 cm × 60 cm) made of transparent acrylic, with the floor divided into 25 equal grids. The central zone was defined as the inner grids (9 grids). For each test, a single rat was gently placed in the center of the arena, and its behavior was recorded continuously for 5 min. After the test, the arena was thoroughly cleaned to prevent cross-contamination. A video tracking system (Noldus EthoVision XT) automatically extracted the target parameters.

Experimental animal models

A research protocol was formulated prior to the initiation of each segment of the in vivo experiment. A single rat represented an experimental unit. Three different types of rat IVDD models were utilized in this work. Two-month-old male SD rats, weighing between 250 g and 290 g, were randomly assigned to different groups. Before surgery, the rats were effectively anesthetized. Rats were placed in the prone position and their tails were sterilized using iodophor. For the PIDD model^69,70, after positioning the rat caudal disc by palpation or X-ray, a 21-gauge needle fitted with a depth limiter was inserted vertically into the disc between caudal vertebrae 5 and 6 at a depth of 5 mm. The needle was rotated 360 degrees and held in place for 30 s and removed vertically. The surgical area was disinfected after operation. Rats in the control group did not receive any treatment (Ctrl). For the LAIDD model¹¹, 2 μL of PBS or sodium lactate (2, 6, or 10 mM) was injected into the intervertebral disc between caudal vertebrae 5 and 6 via a 26-gauge needle and a microsyringe (7105KH; Hamilton)⁷¹. Rats in the control group received only microsyringe puncture (Sham). For in vivo experiments with ZL-01 or engineered EVs, after puncturing the intervertebral disc following the aforementioned method, 2 μL of PBS, ZL-01 (1, 5, 10, or 20 mM) or different kinds of C-EVs was injected into the intervertebral disc between caudal vertebrae 5 and 6 via a 26-gauge needle and a microsyringe. Rats in the control group received only microsyringe puncture (Sham). Each rat was administered with analgesics and antibiotics for 3 days. Four weeks after the operation, imaging data were obtained, and the rats were anesthetized and killed to obtain the intervertebral discs for subsequent experiments. For the AIDD model⁷², 20-month-old male rats, weighing between 600 g and 670 g, were used as the natural aging group, and 2-month-old male SD rats weighing between 250 g and 290 g were used as controls (Ctrl). The rats in this group were directly used for imaging data acquisition and subsequent experiments without any treatment. Pfirrmann grade was the primary outcome measure in the rat IVDD model, and histological score, oxidative stress markers, and senescence markers were secondary outcome measures.

Blinding for in vivo experiments

Blinding was intended to reduce bias at different stages of the in vivo experiment, and the personnel aware of the group allocation are described as follows: for randomization, rats to be grouped were numbered, and the number of experimental groups (the number of ranks) was determined. Subsequently, the RAND function and RANK. The EQ function in Microsoft Excel was used to assign a unique rank to each rat, and rats with the same rank were assigned to the same group. During group allocation, only the experimental designer had full knowledge of the complete group allocation details. Surgical operators only received coded labels and were unaware of the actual intervention types corresponding to each group (sham group, IVDD model group, or treatment group). During the conduct of the experiment, due to differences in intervention methods or body size variations among rats in different groups, the surgical operators could not be blinded. Animal feeding and care staff only had access to cage labels with codes and were unaware of the actual group assignments. During outcome assessment and data analysis, the experiment operator first removed all group identifiers from the experimental results and provided the coded data to independent researchers for preliminary analysis. After verifying the accuracy of the data, the research team performed unblinding, matched the codes with the actual groups, and conducted the final result interpretation. The confounders were not controlled in the in vivo experiments.

ScRNA-seq analysis

ScRNA-seq data from human NP tissue were downloaded from the Gene Expression Omnibus (GEO) database (GSE165722 and GSE251686) and processed using the Seurat software package. Low-quality cellular data with fewer than 200 or more than 6000 expressed genes were excluded using the Harmony package, and then batch effects were mitigated using the Runharmony function. Subsequently, cells were categorized according to the expression patterns of classical markers. Samples were combined using the UMAP function and projected into 2D space using the first 30 principal components. Cells were then unsupervised clustered using the FindClusters function at a resolution of 0.5. Highly expressed genes relative to other clusters were identified using the FindAllMarkers function with a threshold of log2FC > 0.25 and P Adj <0.05. The NPCs were clustered using UMAP, an unsupervised clustering. The R package clusterProfiler was used to perform GO enrichment analysis, setting GO terms with P and FDR < 0.05 as significantly enriched. The ggplot2 package was used to visualize some of the GO enrichment results. Pseudo-temporal analysis of NPC timings in each sample was performed using the Monocle 2 software package. Pseudotemporal-dependent gene expression was analyzed in an unsupervised manner in all NPCs. The candidate pathways were then scored using the geom_smooth function in Seurat, and the dynamics of the pathway scores over pseudo-time were displayed using the geom_smooth function in the ggplot 2 software package.

Primary NPC isolation and lactate treatment

NP tissues were isolated from patients’ surgical biospecimen or 2-month-old male rats, and digested with 0.2% type II collagenase (C2-BIOC; Sigma‒Aldrich, USA) for approximately 2 h at 37 °C, followed by centrifugation at 400 × g for 5 min. The cell pellet was suspended, and the cells were cultured in DMEM/F12 (BI, Israel) supplemented with 10% fetal bovine serum (C04001500, VivaCell, China) and 1% penicillin/streptomycin (C0222, Beyotime, China) at 37 °C in 5% CO₂. For lactate treatment, NPCs were treated with different concentrations of sodium lactate for 7 days. The medium was changed every 3 days.

Metabolomic analysis

Metabolomic analysis was conducted by PANOMIX Biomedical Tech Co., Ltd. (Suzhou, China) using an ExionLC UHPLC system and an AB Sciex Triple Quadrupole 6500 plus mass spectrometer (AB Sciex, USA). Fifty milligrams of fresh human NP tissues was mixed with 500 μL of methanol, and then shaken at 1500 rpm, and centrifuged at 12,000 × g for 10 min at 4 °C. A 100 μL supernatant aliquot was combined with a mixed internal standard, dried under nitrogen gas, and reconstituted in 50% acetonitrile. After centrifugation, the supernatant was analyzed using Multiple Reaction Monitoring for quantitative metabolite analysis. Chromatographic peaks were integrated and quantified via standard curves.

Whole-protein lactylation proteomics

The whole-protein lactylation proteome analysis was conducted with the support of PTM Bio, Inc. (Hangzhou, China). Rat NP tissues in lysis buffer (8 M urea, 1% protease inhibitor cocktail, and 3 μM TSA) were sonicated via an ultrasonic processor (Scientz), and digested with trypsin. To enrich the modified peptides, tryptic peptides were incubated overnight at 4 °C with anti-L-lactyllysine antibody-conjugated agarose beads (PTM-1404, PTM Bio). The bound peptides were eluted from the beads with 0.1% trifluoroacetic acid. The eluted fractions were combined and vacuum dried. The peptides were desalted with C18 ZipTips (Millipore). For LC‒MS/MS analysis, the peptides were separated via a nanoElute UHPLC system (Bruker Daltonics) and exposed to a capillary source. The peptides were analyzed via a timsTOF Pro (Bruker Daltonics) mass spectrometer. The resulting MS/MS data were processed via the MaxQuant search engine (v.1.6.15.0). Tandem mass spectra were searched against the human SwissProt database (20422 entries) concatenated with the reverse decoy database, with an FDR < 1%. After centralizing the signal intensity values in different samples, the relative quantitative values of the modified peptides were obtained. The lactylated peptide data were normalized to the respective protein expression levels.

Construction of overexpression and knockdown vectors and transfection

Gene overexpression vectors for rat Sirt1-3, Hdac1-3, Sod1^WT (synonymous mutation), Sod1^K123R, and Sod1^K10R were designed and constructed by Tsingke Biotech (Beijing, China) in pLV4ltr-Puro-CMV-C3 × FLAG plasmids; synonymous Sod1 mutagenesis was performed to avoid shRNA interference. For knockdown, pLKO.1-Puro plasmids with shRNAs targeting Aars1, Aars2, Gcn5, Cbp, P300, Tip60, Hbo1, Myst1, Sod1, Mct1, Mct4, and Gpr81 were created. Lentiviruses were packaged in HEK293T cells (CRL-3216, ATCC) and used to transfect cultured NPCs following the manufacturer’s instructions, with the medium replaced after 12 h. Transfected NPCs were selected with 2 μg/mL puromycin (ST551, Beyotime), and mRNA levels were measured to assess transfection efficiency. Sequences of the shRNA were provided in the Supplementary Table 2.

β-gal staining assay

NPCs that received different treatments were prepared for β-gal staining by aspirating culture medium, washing 3 times with pre-warmed PBS, fixing with 4% PFA at room temperature for 15 min, and washing 3 times again with pre-warmed PBS. β-gal staining solution was prepared following the instructions of a β-gal Staining Kit (C0602, Beyotime) and added to fixed NPCs followed by overnight incubation. After incubation, the staining solution was aspirated, and the cells were washed 3 times with PBS. The percentage of β-gal-positive cells was calculated by examining 300 cells in five microscopic fields. Images were captured with an inverted phase-contrast microscope (IX83, Olympus).

RNA-seq and bioinformatics analyses

RNA-seq was performed by Maiwei Biotechnology Co., Ltd. Total RNA from NPCs was isolated using TRIzol and quantified using a NanoDrop 2000 (NanoDrop, U.S.A.). cDNA libraries were constructed with SuperScript™ II reverse transcriptase. Then, 2 × 150 bp paired-end RNA-seq was performed using the Illumina NovaSeq platform. The significantly DEGs were filtered with the thresholds of a |log2(fold change, FC)| > 2 and a P value < 0.05 using the R package DESeq2. Transcription factor prediction was performed using the STRING database, CHEA3 database or Cytoscape software, and the data were filtered using the cytoHubba plugin.

MST test

Purified SOD1^WT and SOD1^T40A, SOD1^H44A, SOD1^K123A, and SOD1^N140A mutant proteins were prepared by Tsingke Biotech Co., Ltd. (Beijing, China). pGEX-4T2 plasmids expressing GST-tagged SOD1 variants were transformed into BL21CodonPlus (DE3)-RIL cells, induced with IPTG, lysed, and purified via chromatography and filtration, followed by removal of the GST tag. The purified proteins were labeled with MST fluorescent dye (MO-L011, NanoTemper) per the manufacturer’s protocol. ZL-01 was serially diluted in 50 mM HEPES buffer (0.05% Tween 20; pH 7.4) and incubated with labeled SOD1 for 30 min at room temperature. Affinity was measured using three replicates, and Kd values were determined using Monolith software.

IP–MS analysis

The SOD1 protein was purified by IP, followed by 12% SDS–PAGE at 125 V for 1 h. The gel slices were rinsed with decolorizing solution for 30 min, reduced, alkylated, and digested with trypsin at 37 °C for 20 h. The peptides were desalted, vacuum-dried, resuspended in 0.1% formic acid, and then analyzed using an Easy-nLC 1000 system (Thermo Fisher) eluted with 0.1% aqueous formic acid (mobile phase A) and 0.1% formic acid in 84% acetonitrile (mobile phase B). The column was equilibrated with 95% mobile phase A, and samples were loaded via an autosampler. IP–MS was performed in positive ion mode (scan range, 300–1800 m/z; resolution at 200 m/z, 70,000; AGC, 1e6; max IT, 50 ms; and dynamic exclusion, 30.0 ms), acquiring 20 MS2 spectra per scan with HCD activation (isolation, 2 m/z; resolution at 200 m/z, 17,500; collision energy, 27 eV; and underfill, 0.1%). Data were analyzed with Proteome Discoverer 2.5, and the spectra were searched against the UniProt Rattus norvegicus database (93045_20240115) using trypsin, carbamidomethyl C as a fixed modification, protein/methionine oxidation as a variable modification, a maximum of 2 missed cleavages, and high-confidence peptide filtering.

MRI evaluation

A 3.0 T animal MRI system (Bruker Pharmascan, Germany) was used to assess the signal and structural changes in rat intervertebral discs via sagittal T2-weighted images. The parameters for the T2-weighted sections were set as follows: TR time, 2000 ms; TE time, 80 ms; incentive time, 2; scan time, 3 min 20 s; fat reduction technology, SPAIR; scan matrix, frequency encoding 368 and phase encoding 288; layer thickness, 2.5 mm; echo chains, 12; and spin echo sequence, TSE. The classic Pfirrmann classification criteria⁷³ were used to evaluate MRI images of patients. The modified Pfirrmann classification criteria⁷⁴ were used to evaluate MRI images of rats. To ensure the reliability, the grading was performed by three spinal surgeons (observers A, B, and C). All of them had more than 8 years of experience in spinal imaging diagnosis and received standardized training prior to grading. Each observer was allowed to interpret only 20 sets of images per day to avoid fatigue. During grading, each observer could refer to the classification criteria. In the case of any discrepancy among them, the final decision was based on the majority opinion. The weighted Kappa coefficient was used to assess intra- and inter-observer consistency, and the results were presented in Supplementary Tables 3 and 4.

Histological evaluation

Decalcified rat intervertebral discs or organs were dehydrated, embedded in paraffin, and sliced into 5 μm paraffin sections. The prepared sections were stained with a modified SO & FG Staining Kit (G1371, Solarbio) or a Hematoxylin–Eosin (HE) Staining Kit (G1120, Solarbio) according to the manufacturer’s instructions. Images were captured with an inverted phase-contrast microscope (CKX53, Olympus). The histological grade was calculated following the previous methods⁷⁵. The grading was performed by three researchers (observers D, E, and F), and all of them had over 3 years of experience in the establishment and evaluation of the rat caudal vertebra IVDD model and received standardized training prior to grading. Each observer was allowed to interpret only 20 sets of images per day to avoid fatigue. During grading, each observer could refer to the classification criteria. In the case of any discrepancy among them, the final decision was based on the majority opinion. The intra-class correlation coefficient (ICC) was used to assess the intra- and inter-observer consistency, and the results were presented in Supplementary Table 5.

Real-time PCR

Total RNA was extracted from NPCs with TRIzol reagent and reverse transcribed into cDNA via a PrimeScript RT Kit (RR047A, Takara) according to the manufacturer’s protocol. A Premix Ex Taq II kit (RR390A, Takara) was used to amplify and detect the relative mRNA expression levels of target genes with a CFX Connect Real-Time System (Bio-Rad). Target gene mRNA expression was normalized to β-actin mRNA expression, and relative mRNA expression was calculated via the 2^-ΔΔCt method. Information of the primers used for real-time PCR was provided in Supplementary Table 6.

Immunofluorescence staining assay

NPCs were fixed with 4% paraformaldehyde and permeabilized with immunostaining permeabilization buffer. Tissue sections were deparaffinized, rehydrated, and subjected to antigen retrieval. Samples were blocked with QuickBlock immunostaining buffer and incubated overnight at 4 °C with primary antibodies, including anti-γ-H2AX (AP0687, ABclonal), anti-SOD1 (MA1-105, Thermo Fisher), anti-pan lactylation (PTM-1401RM, PTM BIO), and anti-SOD1^K123la (customized, PTM BIO) antibodies. Samples were incubated with fluorescently labeled rabbit and mouse (ab150113, ab150080, Abcam) secondary antibodies for 1 h at 37 °C. Hoechst 33358 or DAPI was used to counterstain nuclei. Fluorescence images were acquired using a confocal microscope (LSM880, ZEISS, Germany) and quantified with ImageJ.

O₂⁻ and total ROS detection

Intracellular O₂^- production was evaluated using 10 μM DHE (HY-D0079, MCE) and incubating cultured NPCs at 37 °C for 30 min in the dark, after which images were captured with a confocal microscope. The excised NP tissue was homogenized on ice and centrifuged at 4 °C, and the supernatant was treated with 10 μM DHE for 30 min at 37 °C in the dark, after which the fluorescence intensity was measured by a spectrophotometer (M2, Molecular Devices) and normalized to that of protein mass (mg). ROS levels were assessed using a total ROS detection kit (88-5930-74; Invitrogen) and flow cytometry. NPCs and fresh NP tissue were digested into single-cell suspensions stained at 37 °C, and analyzed using a flow cytometer (Gallios, Beckman).

Quantification of L-lactate levels

The concentration of L-lactate content in NP tissue and NPCs was assessed using a L-lactate assay kit (ab65331, Abcam). Briefly, NPCs or milled NP tissues were mixed with assay kit buffer while a gradient concentration of lactate standard was prepared. The analytical buffer was added to the above samples and incubated at room temperature for 30 min, then the optical density value at 450 nm was measured using a spectrophotometer, and a standard curve was plotted between the concentration of the lactate standard and the optical density value, and the concentration of lactate in the samples was calculated by obtaining the equation of the standard curve.

Measurement of the MDA level, T-AOC, and SOD1 activity

MDA levels, T-AOC, and SOD1 activity in NPCs were measured using an MDA assay kit (S0131M, Beyotime), a T-AOC assay kit (S0121, Beyotime), and a Superoxide Dismutase Assay Kit (E-BC-K022-M, Elabscience), following the manufacturers’ instructions. Briefly, protein concentration in lysates was quantified, mixed with assay buffers and working enzymes, incubated under specified conditions, and measured for optical density at the respective wavelengths using a spectrophotometer (M2, MOLECULAR).

Molecular dynamics simulations

The SOD1 conformation was sourced from the Protein Data Bank (PDB ID: 1PU0). Homology modeling was performed using Modeller v9.19, with a lactyl group added to the K123 residue via YASARA. Molecular docking of O₂^- with SOD1 was conducted using AutoDock 4.2.6. Molecular dynamics simulations were executed with Gromacs 2018.4, employing the Amber14SB force field and TIP3P water model under periodic boundary conditions. Long-range electrostatics were calculated using the particle mesh Ewald method (1.0 nm cutoff), and van der Waals interactions used a 1.0 nm cutoff, with hydrogen bonds constrained by the LINCS algorithm. Temperature was maintained at 300 K via the V-rescale method, and pressure at 1 bar with the Parrinello-Rahman method. SOD1^WT and SOD1^K123la were energy minimized, equilibrated for 100 ps under NVT at 300 K, and run for 100 ns under NPT conditions with a 2 fs time step, saving trajectories every 10 ps (5,000 snapshots). Analysis was performed by Mode Science and Technology Co., Ltd. (Chengdu, China).

High-throughput virtual screening

Schrödinger Maestro version 11.4 was used for structure-based virtual screening. The X-ray crystal structure of SOD1 (PDB ID: 1PU0) was obtained from the RCSB Protein Data Bank and further processed via the Protein Preparation Wizard module of Schrödinger. A ChemDIV screening library containing 1.6 million compounds was utilized for virtual screening after energy optimization by the LigPrep module of Schrödinger. The optimized compounds were subjected to a virtual screening workflow for molecular docking. All the compounds were screened in Glide HTVS mode, and the top 10% of the scored compounds were obtained to perform second-round Glide Standard Precision docking. The top 10% of the Glide Standard Precision docking results was docked in the third round using Glide Extra Precision. The top 10 compounds with the highest absolute docking scores were selected as candidates for further experimental validation. The 2D and 3D compound‒protein docking modes were created via PyMol. The analysis was performed by TargetMol Chemicals, Inc. (Shanghai, China), and the solid compounds were also purchased from this company. A summary of virtual screening and validation of candidate compounds was provided in Supplementary Table 7.

CCK-8 assay

A CCK-8 reagent (C0038, Beyotime) was employed to evaluate the cytotoxic effects of various compounds on NPCs. NPCs were seeded in 96-well plates at a density of 5×10³ cells per well. Following cell adhesion, 10 candidate compounds were individually added to the respective wells and incubated for 48 h at final concentrations of 0.5, 1, 2.5, 5, 10, 20, and 50 μM. After replacing the culture medium with fresh complete medium, 10 μL of CCK-8 reagent was pipetted into each well, followed by incubation at 37 °C for 2 h. A spectrophotometer was utilized to measure the optical density (OD) values at a wavelength of 450 nm.

Western blot, Native PAGE, and IP

For western blot electrophoresis, NP tissues or NPCs were lysed in Western and IP cell lysis buffer (P0013, Beyotime) supplemented with protease inhibitors (P1008, Beyotime) and trichostatin A (Selleck). The protein concentration was determined using a BCA assay kit (P0010s, Beyotime). Proteins were resuspended in SDS–PAGE loading buffer (P0015, Beyotime), boiled, and separated by 10% or 12% SDS–PAGE, then transferred to PVDF membranes (Merck, Germany). Membranes were blocked with QuickBlock Western solution (P0252, Beyotime) and incubated overnight at 4 °C with primary antibodies, including anti-P21 (ab109199, Abcam), anti-GADD45α (ab54210, Abcam), anti-SOD1 (ab13498, Abcam), anti-pan lactylation (PTM-1401RM, PTM), anti-SOD1^K123la (customized, PTM), anti-pan acetylation (PTM-101, PTM), anti-pan succinylation (PTM-419, PTM), and anti-β-actin (66009-1-Ig, Proteintech) antibodies. After being washed with TBST, the membranes were incubated with secondary antibodies (SA00001-2 or SA00001-1, Proteintech). Bands were visualized using an ECL kit (1705060, Bio-Rad) and imaged with a ChemiDoc system (Bio-Rad), and the gray values were analyzed using ImageJ software and β-actin as the internal control.

For Native PAGE electrophoresis, NPC lysates were prepared using the NativePAGE Sample Preparation Kit (BN2008, Thermo Fisher). After addition of Blue Native PAGE Sample Buffer (P0761, Beyotime), electrophoresis was performed on 4-13% Blue Native PAGE precast gels (P0546S, Beyotime) using cathodic electrophoresis buffer. The other procedures were the same as western blot.For IP electrophoresis, the supernatant was collected, and the lysate was retained as the input group. For the IP group, IgG or anti-SOD1 antibody was conjugated to protein A + G magnetic beads. The IgG-conjugated beads are used to remove nonspecific proteins from the sample that bind to the anti-SOD1 antibody. Proteins in the lysate were precipitated via IgG antibody-bound magnetic beads or SOD1 antibody-bound magnetic beads, and proteins in the input group and the IP group were denatured and separated via SDS‒PAGE loading buffer (P0015, Beyotime). Boiled protein samples were separated by SDS-PAGE, and then the Coomassie brilliant blue staining (P0003M, Beyotime) was performed to visualize proteins in immunoprecipitated samples. Relative quantification of proteins was then performed via Western blot. Uncropped scans of blots were provided in the Source Data file.

Construction of CTP-EGFP-Lamp2b lentivirus and transfection

The Lamp 2b-CTP lentivirus was constructed and synthesized by OBiO Technology Co., Ltd. (Shanghai, China). Briefly, the Lamp2b + EGFP + CTP sequence was constructed, and after digestion by XmaI and EcorI restriction enzymes, Lamp2b-EGFP-CTP and the lentiviral vector psLenti-CMV-MCS-3xFLAG-PGK-Puro-WPRE were purified on gel to construct the CTP-EGFP-Lamp2b lentivirus. The CTP-EGFP-Lamp2b lentivirus was transfected into NPCs after determining the viral titer, and then the stably transfected cells were screened using G418.

Isolation and characterization of EVs, C-EVs, and C-EVs^Z

NPCs stably expressing CTP-EGFP-Lamp2b were cultured in DMEM/F12 supplemented with 10% exosome-free serum (C3801-0100, Viva Cell, China) and 1% penicillin/streptomycin. C-EVs or EVs were extracted using Total Exosome Isolation Reagent (4478359, Invitrogen, USA). EV particle size was analyzed by nanoparticle tracking analysis (ZetaView PMX 110, Particle Metrix, Germany), and their morphology was visualized by transmission electron microscopy (JSM-1400PLUS, JEOL, Japan). EV positive markers CD81, TSG101, and negative marker calnexin were detected by western blot.

ZL-01 was loaded into C-EVs by sonication at ratios of 1:1 (1 mg/mL C-EVs to 1 μM ZL-01), 1:3 (1 mg/mL to 3 μM), and 1:6 (1 mg/mL to 6 μM), with 10 cycles of 30 s of sonication and 30 s in an ice bath, followed by a 2 h incubation at 37 °C to restore EV membrane integrity. ZL-01 concentration was quantified by UV spectrophotometry (Infinite M200 Pro, Tecan, Shanghai, China), and loading efficiency was calculated as the ratio of loaded to total ZL-01. For release testing, 5 mL of C-EVs^Z in a dialysis bag was immersed in 25 mL of release medium, with 500 μL sampled at designated times (replaced with fresh medium), and the release rate was determined as the ratio of ZL-01 in the medium to the total amount. NPCs labeled with DiD (HY-D1028, MCE) were washed with PBS, mixed with EVs, and incubated at 37 °C. Uptake of DiD-labeled NPCs by EGFP-carrying EVs was observed by confocal microscopy.

High-performance liquid chromatography (HPLC) detection of ZL-01 in the NP tissue of rats

All standards and samples were analyzed using an HPLC system (UltiMate 3000 RS, Thermo Fisher, USA) with a triple quadrupole mass spectrometer (TSQ Quantum, Thermo Fisher, USA). Chromatographic separation was performed with a 4.6 × 150 mm, 5.0 μm column (Welch, China) using a mobile phase of 0.1% formic acid in HPLC-grade water and methanol (20:80, v/v) at a flow rate of 0.8 mL/min being filtered through a 0.22 μm Millipore filter. The calibration samples were ZL-01 standard solutions prepared in methanol at concentrations of 20, 50, 200, 500, 2000, and 5000 nM. NP tissue was treated with methanol and hexane and centrifuged at 12,000 g for 10 min at 4 °C, after which the organic layer of the supernatant was collected, the solvent was removed by evaporation with 50% nitrogen, and the residue was diluted with methanol, filtered, and analyzed by HPLC.

Quantification and statistical analysis

Statistical assessment and analysis were performed using GraphPad Prism 9 and SPSS v26.0. The Shapiro‒Wilk test was used to test for data normality, and Levene’s test for homogeneity of variance. If the normality test was passed, statistical significance was determined by unpaired, two-tailed Student’s t-test with a confidence level of 95% for comparisons of two groups or one-way analysis of variance (ANOVA) followed by Dunnett’s test for comparisons between multiple experimental groups and a single control group, or Tukey’s test for pairwise comparisons among all groups for comparisons of more than two groups. If the normality test was not passed, statistical significance was determined by the Mann-Whitney test for comparisons of two groups or the Kruskal-Wallis test with Bonferroni-corrected Dunn’s post hoc test for comparisons of more than two groups. Inter- and intra-observer consistency for ordinal variables was verified by weighted Kappa coefficient or intra-class correlation coefficient (ICC). All experiments were performed with at least three individual biological samples and three technical replicates. All error bars represent the mean ± standard deviations derived from at least three independent experiments. Researchers were blinded to the evaluation of experimental outcomes and analysis of the raw data.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Supplementary information

Supplementary Information^{(5.9MB, pdf)}

Reporting summary^{(3.6MB, pdf)}

Transparent Peer Review file^{(4.4MB, pdf)}

Source data

Source data^{(6.1MB, xlsx)}

Acknowledgements

This work was supported by the National Natural Science Foundation of China (82472517, 81972113 to C. Li; 82472445 to M. Liu), National Science Foundation for Excellent Young Scholars (82122044 to Y.X.), the Natural Science Foundation of Chongqing in China (cstc2024ycjh-bgzxm0031 to M. Liu), the Special Foundation of Army Military Medical University to Enhance Scientific and Technological Innovation Ability (2022XQN32 to Y. Zhai), and the Subject Competence Enhancement Program of Army Military Medical University (2024C050 to C. Li).

Author contributions

C.Q.L., M.H.L., L.C., and Y.Zhai. designed the experiments. Y.Y.Z., Y.Zhai., Y.Zhang., and Z.Q.B. performed the experiments. J.M.L., Z.L.H., and X.C. obtained and analyzed clinical samples. M.H.C. performed the scRNA-seq analysis and bioinformation analysis. C.L., Y.Zhang., and Y.Zhai. analyzed the raw data. C.Z., Z.H.N., Y.L.X., and L.C. reviewed the data and made substantial contributions to improving the studies. Y.Zhai, Y.Y.Z., and M.H.L. wrote the manuscript, which was reviewed by all authors.

Peer review

Peer review information

Nature Communications thanks Patrick Lüningschrör, and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Data availability

The lactylation proteome data have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository with the dataset identifier PXD072520 [http://www.ebi.ac.uk/pride]. The raw data of RNA-seq have been deposited to the GEO repository with the dataset identifier GSE315354 [https://www.ncbi.nlm.nih.gov/geo/]. The metabolite data are available at Figshare [10.6084/m9.figshare.30971824]. The mass spectrometry data are available at Figshare [10.6084/m9.figshare.30993925]. The scRNA-seq data used in this study are available in the GEO database under accession code GSE165722 and GSE251686 [https://www.ncbi.nlm.nih.gov/gds/]. Source data generated in this study are provided in the Source Data file. Source data are provided with this paper.

Competing interests

The authors declare no competing interests.

Footnotes

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

These authors contributed equally: Yuyao Zhang, Yu Zhai.

These authors jointly supervised this work: Yu Zhai, Lin Chen, Minghan Liu, Changqing Li.

Contributor Information

Yu Zhai, Email: zhaiyu9501@tmmu.edu.cn.

Lin Chen, Email: linchen70@tmmu.edu.cn.

Minghan Liu, Email: liuminghan@tmmu.edu.cn.

Changqing Li, Email: changqli@tmmu.edu.cn.

Supplementary information

The online version contains supplementary material available at 10.1038/s41467-026-69127-3.

References

1.Knezevic, N. N., Candido, K. D., Vlaeyen, J. W. S., Van Zundert, J. & Cohen, S. P. Low back pain. Lancet398, 78–92 (2021). [DOI] [PubMed] [Google Scholar]
2.Gautschi, O. P. et al. Sex differences in subjective and objective measures of pain, functional impairment, and health-related quality of life in patients with lumbar degenerative disc disease. Pain157, 1065–1071 (2016). [DOI] [PubMed] [Google Scholar]
3.Vijay, M. et al. Degenerative lumbar spine disease: estimating global incidence and worldwide volume. Global Spine J.8, 10.1177/2192568218770769 (2018). [DOI] [PMC free article] [PubMed]
4.Ferreira, M. L. Global, regional, and national burden of low back pain, 1990-2020, its attributable risk factors, and projections to 2050: a systematic analysis of the Global Burden of Disease Study 2021. Lancet Rheumatol.5, e316–e329 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
5.Nerlich, A. G., Schaaf, R., Wälchli, B. & Boos, N. Temporo-spatial distribution of blood vessels in human lumbar intervertebral discs. Eur. Spine J.16, 547–555 (2007). [DOI] [PMC free article] [PubMed] [Google Scholar]
6.Grunhagen, T., Shirazi-Adl, A., Fairbank, J. C. & Urban, J. P. Intervertebral disk nutrition: a review of factors influencing concentrations of nutrients and metabolites. Orthop. Clin. North Am.42, 465–477 (2011). [DOI] [PubMed] [Google Scholar]
7.Francisco, V. et al. Metabolomic signature and molecular profile of normal and degenerated human intervertebral disc cells. Spine J.23, 1549–1562 (2023). [DOI] [PubMed] [Google Scholar]
8.Pacholczyk-Sienicka, B., Radek, M., Radek, A. & Jankowski, S. Characterization of metabolites determined by means of 1H HR MAS NMR in intervertebral disc degeneration. Magma28, 173–183 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
9.Shi, J. et al. Increased lactic acid content associated with extracellular matrix depletion in a porcine disc degeneration induced by superficial annular lesion. BMC Musculoskelet. Disord.20, 551 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
10.Bartels, E. M., Fairbank, J. C., Winlove, C. P. & Urban, J. P. Oxygen and lactate concentrations measured in vivo in the intervertebral discs of patients with scoliosis and back pain. Spine23, 1–7 (1998). [DOI] [PubMed] [Google Scholar]
11.Shen, J. et al. Exhausted local lactate accumulation via injectable nanozyme-functionalized hydrogel microsphere for inflammation relief and tissue regeneration. Bioact. Mater.12, 153–168 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
12.Zhang, Y. et al. Lactic acid promotes nucleus pulposus cell senescence and corresponding intervertebral disc degeneration via interacting with Akt. Cell Mol. Life Sci.81, 24 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
13.Zhang, D. et al. Metabolic regulation of gene expression by histone lactylation. Nature574, 575–580 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
14.Zhu, R. et al. ACSS2 acts as a lactyl-CoA synthetase and couples KAT2A to function as a lactyltransferase for histone lactylation and tumor immune evasion. Cell Metab.37, 361–376.e367 (2025). [DOI] [PubMed] [Google Scholar]
15.Wan, N. et al. Cyclic immonium ion of lactyllysine reveals widespread lactylation in the human proteome. Nat. Methods19, 854–864 (2022). [DOI] [PubMed] [Google Scholar]
16.Tong, H. et al. Dual impacts of serine/glycine-free diet in enhancing antitumor immunity and promoting evasion via PD-L1 lactylation. Cell Metab.36, 2493–2510.e2499 (2024). [DOI] [PubMed] [Google Scholar]
17.Fan, W. et al. Global lactylome reveals lactylation-dependent mechanisms underlying T(H)17 differentiation in experimental autoimmune uveitis. Sci. Adv.9, eadh4655 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
18.Wang, L. et al. Serpina3k lactylation protects from cardiac ischemia reperfusion injury. Nat. Commun.16, 1012 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]
19.Wei, L. et al. H3K18 lactylation of senescent microglia potentiates brain aging and Alzheimer’s disease through the NFκB signaling pathway. J. Neuroinflammation20, 208 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
20.Ma, X. M. et al. MCT4-dependent lactate transport: a novel mechanism for cardiac energy metabolism injury and inflammation in type 2 diabetes mellitus. Cardiovasc. Diabetol.23, 96 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
21.Shi, Y. et al. Integrating Bulk RNA and Single-Cell RNA Sequencing Identifies and Validates Lactylation-Related Signatures for Intervertebral Disc Degeneration. J. Cell Mol. Med.28, e70262 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
22.Zhou, T. et al. Prussian blue nanoparticles stabilize SOD1 from ubiquitination-proteasome degradation to rescue intervertebral disc degeneration. Adv. Sci. (Weinh.)9, e2105466 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
23.Liu, X. et al. Activation of GPR81 by lactate drives tumour-induced cachexia. Nat. Metab.6, 708–723 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
24.Lundø, K., Trauelsen, M., Pedersen, S. F. & Schwartz, T. W. Why Warburg Works: Lactate Controls Immune Evasion through GPR81. Cell Metab.31, 666–668 (2020). [DOI] [PubMed] [Google Scholar]
25.Moreno-Yruela, C. et al. Class I histone deacetylases (HDAC1-3) are histone lysine delactylases. Sci. Adv.8, eabi6696 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
26.Luo, J., Mills, K., le Cessie, S., Noordam, R. & van Heemst, D. Ageing, age-related diseases and oxidative stress: What to do next? Ageing Res. Rev.57, 100982 (2020). [DOI] [PubMed] [Google Scholar]
27.Han, X. et al. Integrating genetics and metabolomics from multi-ethnic and multi-fluid data reveals putative mechanisms for age-related macular degeneration. Cell Rep. Med.4, 101085 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
28.Wang, Z. et al. Integrative single-cell metabolomics and phenotypic profiling reveals metabolic heterogeneity of cellular oxidation and senescence. Nat. Commun.16, 2740 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]
29.Ross, J. M. et al. High brain lactate is a hallmark of aging and caused by a shift in the lactate dehydrogenase A/B ratio. Proc. Natl. Acad. Sci. USA107, 20087–20092 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]
30.Jia, L. et al. Rheb-regulated mitochondrial pyruvate metabolism of Schwann cells linked to axon stability. Dev. Cell56, 2980–2994.e2986 (2021). [DOI] [PubMed] [Google Scholar]
31.Dou, X. et al. PDK4-dependent hypercatabolism and lactate production of senescent cells promotes cancer malignancy. Nat. Metab.5, 1887–1910 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
32.Huang, Y. F. et al. Lactate-upregulated NADPH-dependent NOX4 expression via HCAR1/PI3K pathway contributes to ROS-induced osteoarthritis chondrocyte damage. Redox Biol.67, 102867 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
33.Silagi, E. S. et al. Lactate Efflux From Intervertebral Disc Cells Is Required for Maintenance of Spine Health. J. Bone Min. Res.35, 550–570 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
34.Wang, Z. et al. Inhibition of aberrant Hif1α activation delays intervertebral disc degeneration in adult mice. Bone Res.10, 2 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
35.Yang, Z. et al. Lactylome analysis suggests lactylation-dependent mechanisms of metabolic adaptation in hepatocellular carcinoma. Nat. Metab.5, 61–79 (2023). [DOI] [PubMed] [Google Scholar]
36.Zhang, D. et al. Lysine L-lactylation is the dominant lactylation isomer induced by glycolysis. Nat. Chem. Biol.10.1038/s41589-024-01680-8 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
37.Ezer, S. et al. Infantile SOD1 deficiency syndrome caused by a homozygous SOD1 variant with absence of enzyme activity. Brain145, 872–878 (2022). [DOI] [PubMed] [Google Scholar]
38.Hwang, J. et al. SOD1 suppresses pro-inflammatory immune responses by protecting against oxidative stress in colitis. Redox Biol.37, 101760 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
39.Zhang, Y. et al. A new role for oxidative stress in aging: the accelerated aging phenotype in Sod1(-/)(-) mice is correlated to increased cellular senescence. Redox Biol.11, 30–37 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
40.Watanabe, K. et al. Sod1 loss induces intrinsic superoxide accumulation leading to p53-mediated growth arrest and apoptosis. Int J. Mol. Sci.14, 10998–11010 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
41.Breyer, V., Weigel, I., Huang, T. T. & Pischetsrieder, M. Endogenous mitochondrial oxidative stress in MnSOD-deficient mouse embryonic fibroblasts promotes mitochondrial DNA glycation. Free Radic. Biol. Med.52, 1744–1749 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
42.Tamagawa, S. et al. SOD2 orchestrates redox homeostasis in intervertebral discs: a novel insight into oxidative stress-mediated degeneration and therapeutic potential. Redox Biol.71, 103091 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
43.Zhang, N. et al. α-myosin heavy chain lactylation maintains sarcomeric structure and function and alleviates the development of heart failure. Cell Res.33, 679–698 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
44.Chencheng, F. et al. The matrikine N-acetylated proline-glycine-proline induces premature senescence of nucleus pulposus cells via CXCR1-dependent ROS accumulation and DNA damage and reinforces the destructive effect of these cells on homeostasis of intervertebral discs. Biochim. Biophys. Acta Mol. Basis Dis.1863, 10.1016/j.bbadis.2016.10.011 (2016). [DOI] [PubMed]
45.Yu, S. et al. Rescuing nucleus pulposus cells from senescence via dual-functional greigite nanozyme to alleviate intervertebral disc degeneration. Adv. Sci. (Weinh)10, 10.1002/advs.202300988 (2023). [DOI] [PMC free article] [PubMed]
46.Guan, Y., Shao, Y., Li, D. & Liu, M. Generation of site-specific mutations in the rat genome via CRISPR/Cas9. Methods Enzymol.546, 297–317 (2014). [DOI] [PubMed] [Google Scholar]
47.Getzoff, E. D. et al. Electrostatic recognition between superoxide and copper, zinc superoxide dismutase. Nature306, 287–290 (1983). [DOI] [PubMed] [Google Scholar]
48.Getzoff, E. D. et al. Faster superoxide dismutase mutants designed by enhancing electrostatic guidance. Nature358, 347–351 (1992). [DOI] [PubMed] [Google Scholar]
49.Banks, C. J. & Andersen, J. L. Mechanisms of SOD1 regulation by post-translational modifications. Redox Biol.26, 101270 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
50.Lin, Z. F. et al. SIRT5 desuccinylates and activates SOD1 to eliminate ROS. Biochem. Biophys. Res. Commun.441, 191–195 (2013). [DOI] [PubMed] [Google Scholar]
51.Fisher, C. L., Cabelli, D. E., Tainer, J. A., Hallewell, R. A. & Getzoff, E. D. The role of arginine 143 in the electrostatics and mechanism of Cu,Zn superoxide dismutase: computational and experimental evaluation by mutational analysis. Proteins19, 24–34 (1994). [DOI] [PubMed] [Google Scholar]
52.Daryl, A. et al. Wild-type and mutant SOD1 share an aberrant conformation and a common pathogenic pathway in ALS. Nat. Neurosci.13, 10.1038/nn.2660 (2010). [DOI] [PMC free article] [PubMed]
53.Benjamin, G. et al. Altered SOD1 maturation and post-translational modification in amyotrophic lateral sclerosis spinal cord. Brain145, 10.1093/brain/awac165 (2022). [DOI] [PMC free article] [PubMed]
54.Kaliszewski, M. et al. SOD1 Lysine 123 Acetylation in the Adult Central Nervous System. Front Cell Neurosci.10, 287 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
55.Banks, C. J. et al. Acylation of superoxide dismutase 1 (SOD1) at K122 governs SOD1-mediated inhibition of mitochondrial respiration. Mol. Cell Biol.37, 10.1128/mcb.00354-17 (2017). [DOI] [PMC free article] [PubMed]
56.Ma, Q. et al. Crosstalk between lysine lactylation and acetylation regulates lactate dehydrogenase in Streptococcus mutans. Genomics Proteom. Bioinforma.10.1093/gpbjnl/qzaf073 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]
57.You, Y. et al. Analysis of a macrophage carbamylated proteome reveals a function in post-translational modification crosstalk. Cell Commun. Signal21, 241 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
58.Zong, Z. et al. Alanyl-tRNA synthetase, AARS1, is a lactate sensor and lactyltransferase that lactylates p53 and contributes to tumorigenesis. Cell187, 2375–2392.e2333 (2024). [DOI] [PubMed] [Google Scholar]
59.Ju, J. et al. The alanyl-tRNA synthetase AARS1 moonlights as a lactyltransferase to promote YAP signaling in gastric cancer. J. Clin. Invest.134, 10.1172/jci174587 (2024). [DOI] [PMC free article] [PubMed]
60.Li, F. et al. Positive feedback regulation between glycolysis and histone lactylation drives oncogenesis in pancreatic ductal adenocarcinoma. Mol. Cancer23, 90 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
61.Fan, S. et al. Inhibition of autophagy by a small molecule through covalent modification of the LC3 protein. Angew. Chem. Int. Ed. Engl.60, 26105–26114 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
62.Kwon, A. et al. Potent small-molecule inhibitors targeting acetylated microtubules as anticancer agents against triple-negative breast cancer. Biomedicines8, 10.3390/biomedicines8090338 (2020). [DOI] [PMC free article] [PubMed]
63.Chen, H. et al. Selectively targeting STAT3 using a small molecule inhibitor is a potential therapeutic strategy for pancreatic cancer. Clin. Cancer Res. 29, 815–830 (2023). [DOI] [PubMed] [Google Scholar]
64.Imke, R.-J. et al. Drug retention after intradiscal administration. Drug Deliv.31, 10.1080/10717544.2024.2415579 (2024). [DOI] [PMC free article] [PubMed]
65.Colella, F. et al. Drug delivery in intervertebral disc degeneration and osteoarthritis: Selecting the optimal platform for the delivery of disease-modifying agents. J. Control Release328, 985–999 (2020). [DOI] [PubMed] [Google Scholar]
66.Ren, K. et al. Zwitterionic polymer modified xanthan gum with collagen II-binding capability for lubrication improvement and ROS scavenging. Carbohydr. Polym.274, 118672 (2021). [DOI] [PubMed] [Google Scholar]
67.Vedadghavami, A., Zhang, C. & Bajpayee, A. G. Overcoming negatively charged tissue barriers: drug delivery using cationic peptides and proteins. Nano Today34, 10.1016/j.nantod.2020.100898 (2020). [DOI] [PMC free article] [PubMed]
68.Arrowsmith, C. H. et al. The promise and peril of chemical probes. Nat. Chem. Biol.11, 536–541 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
69.Cai, W. et al. Multiscale mechanical-adapted hydrogels for the repair of intervertebral disc degeneration. Bioact. Mater.48, 336–352 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]
70.Ohnishi, H. et al. Anti-inflammatory effects of adiponectin receptor agonist AdipoRon against intervertebral disc degeneration. Int. J. Mol. Sci.24, 10.3390/ijms24108566 (2023). [DOI] [PMC free article] [PubMed]
71.Barcellona, M. N., McDonnell, E. E., Samuel, S. & Buckley, C. T. Rat tail models for the assessment of injectable nucleus pulposus regeneration strategies. JOR Spine5, e1216 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
72.Liu, L. et al. Dynamics of N6-methyladenosine modification during aging and their potential roles in the degeneration of intervertebral disc. JOR Spine7, e1316 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
73.Pfirrmann, C. W., Metzdorf, A., Zanetti, M., Hodler, J. & Boos, N. Magnetic resonance classification of lumbar intervertebral disc degeneration. Spine26, 1873–1878 (2001). [DOI] [PubMed] [Google Scholar]
74.Grunert, P. et al. Assessment of intervertebral disc degeneration based on quantitative magnetic resonance imaging analysis: an in vivo study. Spine39, E369–E378 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
75.Lai, A. et al. Development of a standardized histopathology scoring system for intervertebral disc degeneration in rat models: an initiative of the ORS spine section. JOR Spine4, e1150 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Information^{(5.9MB, pdf)}

Reporting summary^{(3.6MB, pdf)}

Transparent Peer Review file^{(4.4MB, pdf)}

Source data^{(6.1MB, xlsx)}

Data Availability Statement

[CR1] 1.Knezevic, N. N., Candido, K. D., Vlaeyen, J. W. S., Van Zundert, J. & Cohen, S. P. Low back pain. Lancet398, 78–92 (2021). [DOI] [PubMed] [Google Scholar]

[CR2] 2.Gautschi, O. P. et al. Sex differences in subjective and objective measures of pain, functional impairment, and health-related quality of life in patients with lumbar degenerative disc disease. Pain157, 1065–1071 (2016). [DOI] [PubMed] [Google Scholar]

[CR3] 3.Vijay, M. et al. Degenerative lumbar spine disease: estimating global incidence and worldwide volume. Global Spine J.8, 10.1177/2192568218770769 (2018). [DOI] [PMC free article] [PubMed]

[CR4] 4.Ferreira, M. L. Global, regional, and national burden of low back pain, 1990-2020, its attributable risk factors, and projections to 2050: a systematic analysis of the Global Burden of Disease Study 2021. Lancet Rheumatol.5, e316–e329 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR5] 5.Nerlich, A. G., Schaaf, R., Wälchli, B. & Boos, N. Temporo-spatial distribution of blood vessels in human lumbar intervertebral discs. Eur. Spine J.16, 547–555 (2007). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR6] 6.Grunhagen, T., Shirazi-Adl, A., Fairbank, J. C. & Urban, J. P. Intervertebral disk nutrition: a review of factors influencing concentrations of nutrients and metabolites. Orthop. Clin. North Am.42, 465–477 (2011). [DOI] [PubMed] [Google Scholar]

[CR7] 7.Francisco, V. et al. Metabolomic signature and molecular profile of normal and degenerated human intervertebral disc cells. Spine J.23, 1549–1562 (2023). [DOI] [PubMed] [Google Scholar]

[CR8] 8.Pacholczyk-Sienicka, B., Radek, M., Radek, A. & Jankowski, S. Characterization of metabolites determined by means of 1H HR MAS NMR in intervertebral disc degeneration. Magma28, 173–183 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR9] 9.Shi, J. et al. Increased lactic acid content associated with extracellular matrix depletion in a porcine disc degeneration induced by superficial annular lesion. BMC Musculoskelet. Disord.20, 551 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR10] 10.Bartels, E. M., Fairbank, J. C., Winlove, C. P. & Urban, J. P. Oxygen and lactate concentrations measured in vivo in the intervertebral discs of patients with scoliosis and back pain. Spine23, 1–7 (1998). [DOI] [PubMed] [Google Scholar]

[CR11] 11.Shen, J. et al. Exhausted local lactate accumulation via injectable nanozyme-functionalized hydrogel microsphere for inflammation relief and tissue regeneration. Bioact. Mater.12, 153–168 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR12] 12.Zhang, Y. et al. Lactic acid promotes nucleus pulposus cell senescence and corresponding intervertebral disc degeneration via interacting with Akt. Cell Mol. Life Sci.81, 24 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR13] 13.Zhang, D. et al. Metabolic regulation of gene expression by histone lactylation. Nature574, 575–580 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR14] 14.Zhu, R. et al. ACSS2 acts as a lactyl-CoA synthetase and couples KAT2A to function as a lactyltransferase for histone lactylation and tumor immune evasion. Cell Metab.37, 361–376.e367 (2025). [DOI] [PubMed] [Google Scholar]

[CR15] 15.Wan, N. et al. Cyclic immonium ion of lactyllysine reveals widespread lactylation in the human proteome. Nat. Methods19, 854–864 (2022). [DOI] [PubMed] [Google Scholar]

[CR16] 16.Tong, H. et al. Dual impacts of serine/glycine-free diet in enhancing antitumor immunity and promoting evasion via PD-L1 lactylation. Cell Metab.36, 2493–2510.e2499 (2024). [DOI] [PubMed] [Google Scholar]

[CR17] 17.Fan, W. et al. Global lactylome reveals lactylation-dependent mechanisms underlying T(H)17 differentiation in experimental autoimmune uveitis. Sci. Adv.9, eadh4655 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR18] 18.Wang, L. et al. Serpina3k lactylation protects from cardiac ischemia reperfusion injury. Nat. Commun.16, 1012 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR19] 19.Wei, L. et al. H3K18 lactylation of senescent microglia potentiates brain aging and Alzheimer’s disease through the NFκB signaling pathway. J. Neuroinflammation20, 208 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR20] 20.Ma, X. M. et al. MCT4-dependent lactate transport: a novel mechanism for cardiac energy metabolism injury and inflammation in type 2 diabetes mellitus. Cardiovasc. Diabetol.23, 96 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR21] 21.Shi, Y. et al. Integrating Bulk RNA and Single-Cell RNA Sequencing Identifies and Validates Lactylation-Related Signatures for Intervertebral Disc Degeneration. J. Cell Mol. Med.28, e70262 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR22] 22.Zhou, T. et al. Prussian blue nanoparticles stabilize SOD1 from ubiquitination-proteasome degradation to rescue intervertebral disc degeneration. Adv. Sci. (Weinh.)9, e2105466 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR23] 23.Liu, X. et al. Activation of GPR81 by lactate drives tumour-induced cachexia. Nat. Metab.6, 708–723 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR24] 24.Lundø, K., Trauelsen, M., Pedersen, S. F. & Schwartz, T. W. Why Warburg Works: Lactate Controls Immune Evasion through GPR81. Cell Metab.31, 666–668 (2020). [DOI] [PubMed] [Google Scholar]

[CR25] 25.Moreno-Yruela, C. et al. Class I histone deacetylases (HDAC1-3) are histone lysine delactylases. Sci. Adv.8, eabi6696 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR26] 26.Luo, J., Mills, K., le Cessie, S., Noordam, R. & van Heemst, D. Ageing, age-related diseases and oxidative stress: What to do next? Ageing Res. Rev.57, 100982 (2020). [DOI] [PubMed] [Google Scholar]

[CR27] 27.Han, X. et al. Integrating genetics and metabolomics from multi-ethnic and multi-fluid data reveals putative mechanisms for age-related macular degeneration. Cell Rep. Med.4, 101085 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR28] 28.Wang, Z. et al. Integrative single-cell metabolomics and phenotypic profiling reveals metabolic heterogeneity of cellular oxidation and senescence. Nat. Commun.16, 2740 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR29] 29.Ross, J. M. et al. High brain lactate is a hallmark of aging and caused by a shift in the lactate dehydrogenase A/B ratio. Proc. Natl. Acad. Sci. USA107, 20087–20092 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR30] 30.Jia, L. et al. Rheb-regulated mitochondrial pyruvate metabolism of Schwann cells linked to axon stability. Dev. Cell56, 2980–2994.e2986 (2021). [DOI] [PubMed] [Google Scholar]

[CR31] 31.Dou, X. et al. PDK4-dependent hypercatabolism and lactate production of senescent cells promotes cancer malignancy. Nat. Metab.5, 1887–1910 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR32] 32.Huang, Y. F. et al. Lactate-upregulated NADPH-dependent NOX4 expression via HCAR1/PI3K pathway contributes to ROS-induced osteoarthritis chondrocyte damage. Redox Biol.67, 102867 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR33] 33.Silagi, E. S. et al. Lactate Efflux From Intervertebral Disc Cells Is Required for Maintenance of Spine Health. J. Bone Min. Res.35, 550–570 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR34] 34.Wang, Z. et al. Inhibition of aberrant Hif1α activation delays intervertebral disc degeneration in adult mice. Bone Res.10, 2 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR35] 35.Yang, Z. et al. Lactylome analysis suggests lactylation-dependent mechanisms of metabolic adaptation in hepatocellular carcinoma. Nat. Metab.5, 61–79 (2023). [DOI] [PubMed] [Google Scholar]

[CR36] 36.Zhang, D. et al. Lysine L-lactylation is the dominant lactylation isomer induced by glycolysis. Nat. Chem. Biol.10.1038/s41589-024-01680-8 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR37] 37.Ezer, S. et al. Infantile SOD1 deficiency syndrome caused by a homozygous SOD1 variant with absence of enzyme activity. Brain145, 872–878 (2022). [DOI] [PubMed] [Google Scholar]

[CR38] 38.Hwang, J. et al. SOD1 suppresses pro-inflammatory immune responses by protecting against oxidative stress in colitis. Redox Biol.37, 101760 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR39] 39.Zhang, Y. et al. A new role for oxidative stress in aging: the accelerated aging phenotype in Sod1(-/)(-) mice is correlated to increased cellular senescence. Redox Biol.11, 30–37 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR40] 40.Watanabe, K. et al. Sod1 loss induces intrinsic superoxide accumulation leading to p53-mediated growth arrest and apoptosis. Int J. Mol. Sci.14, 10998–11010 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR41] 41.Breyer, V., Weigel, I., Huang, T. T. & Pischetsrieder, M. Endogenous mitochondrial oxidative stress in MnSOD-deficient mouse embryonic fibroblasts promotes mitochondrial DNA glycation. Free Radic. Biol. Med.52, 1744–1749 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR42] 42.Tamagawa, S. et al. SOD2 orchestrates redox homeostasis in intervertebral discs: a novel insight into oxidative stress-mediated degeneration and therapeutic potential. Redox Biol.71, 103091 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR43] 43.Zhang, N. et al. α-myosin heavy chain lactylation maintains sarcomeric structure and function and alleviates the development of heart failure. Cell Res.33, 679–698 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR44] 44.Chencheng, F. et al. The matrikine N-acetylated proline-glycine-proline induces premature senescence of nucleus pulposus cells via CXCR1-dependent ROS accumulation and DNA damage and reinforces the destructive effect of these cells on homeostasis of intervertebral discs. Biochim. Biophys. Acta Mol. Basis Dis.1863, 10.1016/j.bbadis.2016.10.011 (2016). [DOI] [PubMed]

[CR45] 45.Yu, S. et al. Rescuing nucleus pulposus cells from senescence via dual-functional greigite nanozyme to alleviate intervertebral disc degeneration. Adv. Sci. (Weinh)10, 10.1002/advs.202300988 (2023). [DOI] [PMC free article] [PubMed]

[CR46] 46.Guan, Y., Shao, Y., Li, D. & Liu, M. Generation of site-specific mutations in the rat genome via CRISPR/Cas9. Methods Enzymol.546, 297–317 (2014). [DOI] [PubMed] [Google Scholar]

[CR47] 47.Getzoff, E. D. et al. Electrostatic recognition between superoxide and copper, zinc superoxide dismutase. Nature306, 287–290 (1983). [DOI] [PubMed] [Google Scholar]

[CR48] 48.Getzoff, E. D. et al. Faster superoxide dismutase mutants designed by enhancing electrostatic guidance. Nature358, 347–351 (1992). [DOI] [PubMed] [Google Scholar]

[CR49] 49.Banks, C. J. & Andersen, J. L. Mechanisms of SOD1 regulation by post-translational modifications. Redox Biol.26, 101270 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR50] 50.Lin, Z. F. et al. SIRT5 desuccinylates and activates SOD1 to eliminate ROS. Biochem. Biophys. Res. Commun.441, 191–195 (2013). [DOI] [PubMed] [Google Scholar]

[CR51] 51.Fisher, C. L., Cabelli, D. E., Tainer, J. A., Hallewell, R. A. & Getzoff, E. D. The role of arginine 143 in the electrostatics and mechanism of Cu,Zn superoxide dismutase: computational and experimental evaluation by mutational analysis. Proteins19, 24–34 (1994). [DOI] [PubMed] [Google Scholar]

[CR52] 52.Daryl, A. et al. Wild-type and mutant SOD1 share an aberrant conformation and a common pathogenic pathway in ALS. Nat. Neurosci.13, 10.1038/nn.2660 (2010). [DOI] [PMC free article] [PubMed]

[CR53] 53.Benjamin, G. et al. Altered SOD1 maturation and post-translational modification in amyotrophic lateral sclerosis spinal cord. Brain145, 10.1093/brain/awac165 (2022). [DOI] [PMC free article] [PubMed]

[CR54] 54.Kaliszewski, M. et al. SOD1 Lysine 123 Acetylation in the Adult Central Nervous System. Front Cell Neurosci.10, 287 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR55] 55.Banks, C. J. et al. Acylation of superoxide dismutase 1 (SOD1) at K122 governs SOD1-mediated inhibition of mitochondrial respiration. Mol. Cell Biol.37, 10.1128/mcb.00354-17 (2017). [DOI] [PMC free article] [PubMed]

[CR56] 56.Ma, Q. et al. Crosstalk between lysine lactylation and acetylation regulates lactate dehydrogenase in Streptococcus mutans. Genomics Proteom. Bioinforma.10.1093/gpbjnl/qzaf073 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR57] 57.You, Y. et al. Analysis of a macrophage carbamylated proteome reveals a function in post-translational modification crosstalk. Cell Commun. Signal21, 241 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR58] 58.Zong, Z. et al. Alanyl-tRNA synthetase, AARS1, is a lactate sensor and lactyltransferase that lactylates p53 and contributes to tumorigenesis. Cell187, 2375–2392.e2333 (2024). [DOI] [PubMed] [Google Scholar]

[CR59] 59.Ju, J. et al. The alanyl-tRNA synthetase AARS1 moonlights as a lactyltransferase to promote YAP signaling in gastric cancer. J. Clin. Invest.134, 10.1172/jci174587 (2024). [DOI] [PMC free article] [PubMed]

[CR60] 60.Li, F. et al. Positive feedback regulation between glycolysis and histone lactylation drives oncogenesis in pancreatic ductal adenocarcinoma. Mol. Cancer23, 90 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR61] 61.Fan, S. et al. Inhibition of autophagy by a small molecule through covalent modification of the LC3 protein. Angew. Chem. Int. Ed. Engl.60, 26105–26114 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR62] 62.Kwon, A. et al. Potent small-molecule inhibitors targeting acetylated microtubules as anticancer agents against triple-negative breast cancer. Biomedicines8, 10.3390/biomedicines8090338 (2020). [DOI] [PMC free article] [PubMed]

[CR63] 63.Chen, H. et al. Selectively targeting STAT3 using a small molecule inhibitor is a potential therapeutic strategy for pancreatic cancer. Clin. Cancer Res. 29, 815–830 (2023). [DOI] [PubMed] [Google Scholar]

[CR64] 64.Imke, R.-J. et al. Drug retention after intradiscal administration. Drug Deliv.31, 10.1080/10717544.2024.2415579 (2024). [DOI] [PMC free article] [PubMed]

[CR65] 65.Colella, F. et al. Drug delivery in intervertebral disc degeneration and osteoarthritis: Selecting the optimal platform for the delivery of disease-modifying agents. J. Control Release328, 985–999 (2020). [DOI] [PubMed] [Google Scholar]

[CR66] 66.Ren, K. et al. Zwitterionic polymer modified xanthan gum with collagen II-binding capability for lubrication improvement and ROS scavenging. Carbohydr. Polym.274, 118672 (2021). [DOI] [PubMed] [Google Scholar]

[CR67] 67.Vedadghavami, A., Zhang, C. & Bajpayee, A. G. Overcoming negatively charged tissue barriers: drug delivery using cationic peptides and proteins. Nano Today34, 10.1016/j.nantod.2020.100898 (2020). [DOI] [PMC free article] [PubMed]

[CR68] 68.Arrowsmith, C. H. et al. The promise and peril of chemical probes. Nat. Chem. Biol.11, 536–541 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR69] 69.Cai, W. et al. Multiscale mechanical-adapted hydrogels for the repair of intervertebral disc degeneration. Bioact. Mater.48, 336–352 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR70] 70.Ohnishi, H. et al. Anti-inflammatory effects of adiponectin receptor agonist AdipoRon against intervertebral disc degeneration. Int. J. Mol. Sci.24, 10.3390/ijms24108566 (2023). [DOI] [PMC free article] [PubMed]

[CR71] 71.Barcellona, M. N., McDonnell, E. E., Samuel, S. & Buckley, C. T. Rat tail models for the assessment of injectable nucleus pulposus regeneration strategies. JOR Spine5, e1216 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR72] 72.Liu, L. et al. Dynamics of N6-methyladenosine modification during aging and their potential roles in the degeneration of intervertebral disc. JOR Spine7, e1316 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR73] 73.Pfirrmann, C. W., Metzdorf, A., Zanetti, M., Hodler, J. & Boos, N. Magnetic resonance classification of lumbar intervertebral disc degeneration. Spine26, 1873–1878 (2001). [DOI] [PubMed] [Google Scholar]

[CR74] 74.Grunert, P. et al. Assessment of intervertebral disc degeneration based on quantitative magnetic resonance imaging analysis: an in vivo study. Spine39, E369–E378 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR75] 75.Lai, A. et al. Development of a standardized histopathology scoring system for intervertebral disc degeneration in rat models: an initiative of the ORS spine section. JOR Spine4, e1150 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

PERMALINK

SOD1 lactylation impair its enzymatic activity by conformational change to aggravate intervertebral disc degeneration

Yuyao Zhang

Yu Zhai

Chao Liu

Minghang Chen

Yang Zhang

Zhiqun Bian

Xian Chang

Zhilei Hu

Jianmin Li

Chao Zhang

Zhenhong Ni

Yangli Xie

Lin Chen

Minghan Liu

Changqing Li

Abstract

Introduction

Results

Multi-omics analysis reveals the lactylome landscape and the association between oxidative damage and lactylation during IVDD

Fig. 1. The lactylome landscape during IVDD and the association between oxidative damage with lactylatioJn revealed by multi-omics analysis.

Lactylation promotes oxidative damage in NPCs and aggravates IVDD in rats

Fig. 2. Lactylation aggravates IVDD in rats and promotes oxidative damage in NPCs.

SOD1 lactylation exacerbated oxidative damage in NPCs via the p53 pathway

Fig. 3. SOD1K123la exacerbates oxidative damage in NPCs via p53 pathway.

SOD1K123la aggravated IVDD by impairing SOD1 enzymatic activity

Fig. 4. SOD1K123R alleviates IVDD by reversing SOD1 enzymatic activity.

Fig. 5. Molecular dynamics simulations of SOD1K123la with O2− and identification of potential SOD1K123la lactyltransferases.

ZL-01 acts as a potential specific inhibitor of SOD1K123la

Fig. 6. Virtual screening of ZL-01 and its role in reducing SOD1K123la and inhibiting oxidative damage in NPCs.

Targeted delivery of ZL-01 reduced oxidative damage and alleviated IVDD

Fig. 7. Construction of C-EVsZ and its role in reducing oxidative damage and alleviating IVDD.

Fig. 8. Schematic diagram of the role and mechanism of lactate accumulation-induced SOD1K123la in aggravating IVDD.

Discussion

Limitations

Methods

Human NP tissue acquisition

Experimental animals

Generation of SOD1K123R rats

Open field test

Experimental animal models

Blinding for in vivo experiments

ScRNA-seq analysis

Primary NPC isolation and lactate treatment

Metabolomic analysis

Whole-protein lactylation proteomics

Construction of overexpression and knockdown vectors and transfection

β-gal staining assay

RNA-seq and bioinformatics analyses

MST test

IP–MS analysis

MRI evaluation

Histological evaluation

Real-time PCR

Immunofluorescence staining assay

O2− and total ROS detection

Quantification of L-lactate levels

Measurement of the MDA level, T-AOC, and SOD1 activity

Molecular dynamics simulations

High-throughput virtual screening

CCK-8 assay

Western blot, Native PAGE, and IP

Construction of CTP-EGFP-Lamp2b lentivirus and transfection

Isolation and characterization of EVs, C-EVs, and C-EVsZ

High-performance liquid chromatography (HPLC) detection of ZL-01 in the NP tissue of rats

Quantification and statistical analysis

Reporting summary

Supplementary information

Source data

Acknowledgements

Author contributions

Peer review

Peer review information

Data availability

Competing interests

Footnotes

Contributor Information

Supplementary information

References

Fig. 3. SOD1^K123la exacerbates oxidative damage in NPCs via p53 pathway.

SOD1^K123la aggravated IVDD by impairing SOD1 enzymatic activity

Fig. 4. SOD1^K123R alleviates IVDD by reversing SOD1 enzymatic activity.

Fig. 5. Molecular dynamics simulations of SOD1^K123la with O₂⁻ and identification of potential SOD1^K123la lactyltransferases.

ZL-01 acts as a potential specific inhibitor of SOD1^K123la

Fig. 6. Virtual screening of ZL-01 and its role in reducing SOD1^K123la and inhibiting oxidative damage in NPCs.

Fig. 7. Construction of C-EVs^Z and its role in reducing oxidative damage and alleviating IVDD.

Fig. 8. Schematic diagram of the role and mechanism of lactate accumulation-induced SOD1^K123la in aggravating IVDD.

Generation of SOD1^K123R rats

O₂⁻ and total ROS detection

Isolation and characterization of EVs, C-EVs, and C-EVs^Z