ANXA11 suppression restores muscular function in the mdx mouse model of Duchenne muscular dystrophy (DMD) Authors: Wen Tang, Bowen Lin, Ming Jin, Tianzhen Liu, Jingyi Zhou, Yanhong Jin, Yi Chen, Yanan Zhu, Gonglu Liu, Ping Hu, Chengyong Shen, Zhuoxian Meng, Guoping Peng, Liqun Lei, Xingxu Huang, Chenxiao Zhang, Ke Chen, Guiheng Zhang, Xufeng Yang, Xue Lv, Adrien Rousseau, Jian Yang, Zhi-Ying Wu, Ge Bai, Zhen Zhong Published: 2026 ABSTRACT AbstractA critical question in Duchenne muscular dystrophy (DMD) research is whether regulatory mechanisms beyond dystrophin loss contribute to impaired muscle regeneration. Through integrative analysis of proteomics and single-nucleus RNA sequencing database, we identify the upregulation of ANXA11, a gene encoding a Ca²⁺-dependent phospholipid-binding protein, in MYH3⁺ regenerative myofibers from both mdx mice and DMD patients. This upregulation disrupts the maturation of regenerative myofibers, preventing adequate compensation for muscle loss in mdx mice due to dysregulation of the mTOR pathway. Suppression of Anxa11 via genetic knockout or AAV9-mediated knockdown significantly enhanced MYH3⁺ myofiber maturation, accompanied by restored S6 phosphorylation and robust functional muscle recovery in mdx mice. These results establish ANXA11 as a key regulator of muscle regeneration and a potential therapeutic target for DMD.Subject terms: Neuromuscular disease, Disease model, Translational research Introduction Duchenne muscular dystrophy (DMD) is a severe inherited X-linked recessive disorder characterized by progressive skeletal muscle degeneration and weakness1. It arises from mutations in the dystrophin gene leading to the absence of dystrophin—a crucial subsarcolemmal protein within the dystrophin-associated protein complex (DAPC)2. This complex connects the intracellular actin cytoskeleton with the extracellular matrix via the sarcolemma in striated muscle cells3. Disruption of the DAPC compromises the protective shield against contractile membrane damage, leading to progressive muscle degeneration3. Moreover, the lack of dystrophin expression in DMD patients diminishes myogenic potential1,4–8. Consequently, muscle regeneration fails to compensate for muscle loss, exacerbating disease progression9. Over the past several decades, considerable efforts have been directed toward restoring dystrophin function, given its central role in therapeutic strategies for single-gene mutation diseases8. However, these approaches encounter some difficulties, primarily due to the diversity of mutations underlying DMD and the variable efficiency of the dystrophin gene delivery to satellite cells1,8. These technical challenges suggest the need to identify new genetic modifiers that regulate muscle regeneration in DMD1,10. A critical, unresolved question in DMD research is whether endogenous regulatory mechanisms beyond the dystrophin mutation itself contribute to impaired muscle repair in DMD. Studies in both animal models and DMD patients have shown an increased number of PAX7⁺ satellite cells and MYH3⁺ regenerative myoblasts in dystrophic muscle6,11–13, suggesting that satellite cell depletion is not the primary cause of regeneration failure in DMD. However, the persistent degeneration of mature myofibers remains a hallmark of the disease. Despite the presence of MYH3⁺ regenerative myofibers12,14, they do not adequately compensate for ongoing muscle loss15,16, indicating that their maturation into fully functional myofibers may be limited. This maturation failure could significantly contribute to disease progression. Therapies targeting this late-stage maturation failure during regeneration may provide substantial benefits beyond restoring sarcolemma integrity alone. While much research has focused on promoting satellite cell proliferation and differentiation in DMD models13,17–19, the transition from immature regenerative myofibers to fully functional myofibers remains poorly understood20,21. Addressing this gap is essential for developing more effective therapies to counteract the progressive muscle degeneration seen in DMD. To explore new therapeutic targets for regeneration dysfunction, we conducted a parallel analysis of the proteome and the transcriptome of muscle tissue from both the mdx mouse model and DMD patients using publicly available databases22–24. Our integrative analysis identified ANXA11 as one of the most consistently upregulated factors in regenerative myofibers across both human DMD samples and the mdx mouse model, suggesting its potential involvement in the mechanisms underlying regeneration failure in DMD. Notably, ANXA11 is highly expressed in skeletal muscle25, yet its role in dystrophic muscle pathology has not been investigated. Here, we report a new role of ANXA11 in muscle pathology and show the upregulation of ANXA11 in MYH3⁺ regenerative myofibers, accompanied by suppressed S6 signaling in dystrophic muscle. Suppression of ANXA11 restored mTORC1-S6K1 activity, enhanced downstream S6 phosphorylation, accelerated MYH3⁺ fiber maturation, and improved motor function in mdx mice. These findings highlight ANXA11 as a novel regulator in the lysosome-dependent mTOR signaling pathway, that constrains myofiber maturation during regeneration, supporting its potential as a therapeutic target to correct regenerative defects in DMD. Results ANXA11 is markedly elevated in the regenerative myofiber of DMD patients and mdx mice To investigate the cellular response contributing to impaired regenerative myogenesis in muscle tissue in human DMD patients and in mdx mice, we conducted a parallel analysis of the proteome from mdx mice and DMD patients from public mass spectrometry data22,23. This analysis identified a cohort of proteins in muscle tissue of both mdx mice and DMD patients that were significantly altered compared to healthy controls and BMD subjects (Fig. 1a). Specifically, we observed an overlap of 30 upregulated proteins and 5 downregulated proteins in the muscle tissue of both DMD patients and mdx mice (Fig. 1a and Supplementary Table 2). ANXA11 is elevated in regenerative myofibers of DMD patients and mdx mice. a Parallel analysis of proteomics data generated from mass spectrometry of DMD patients and mdx mice. b A Venn diagram, based on single-nucleus RNA sequencing data, illustrates the overlap of genes in the regenerative myofibers of DMD patients and the mdx mouse model, with an abs (Log2FC) ≥ 0.85. c Immunoblot results of bicipital muscle tissue from both DMD patients and healthy controls. d Comparison of relative mRNA expression levels of ANXA11 from DMD patients and healthy controls (n = 3 per group). e Immunoblot results of TA muscle tissue from mdx and WT mice. f Comparison of relative mRNA expression levels of Anxa11 from mdx and WT mice (n = 3 per group). g Immunofluorescence staining of TA muscle cross-sections from WT and mdx mice using antibodies against Dystrophin, ANXA11, and MYH3. Scale bar, 50 μm. Representative images from n = 3 biological replicates are shown. Source data are provided as a Source Data file, and the exact p-values have been provided as required in the Source Data. Data are presented as means ± SD. *P < 0.05, **P < 0.01 as determined by unpaired t-test (two-sided). To ensure that the observed differences resulted from gene expression changes in muscle cells rather than an immune response or fibroblast fibrillization during regenerative pathology, we utilized public single-nucleus RNA sequencing (snRNA-seq) data from human and mouse samples24. This approach enabled the identification of distinct gene expression pattern across different cell types within the muscular tissue of both DMD patients and mdx mice, including PAX7+ satellite cells, MYOG+ MYH3+ MYH8+ regenerative myofibers, DCN+ fibroblasts, and PTPRC+ immune cells24. Analysis of human snRNA sequencing data revealed that adult myofibers were identified as either fast myofibers (MYH1/2+ TTN+) or slow myofibers (MYH7+ TTN+), whereas mouse snRNA data revealed the exclusive presence of fast myofibers (Myh1/2+ Ttn+) (Supplementary Fig. 1a, b). Specifically, we identified that upregulation of ANXA11, in addition to another two regenerative markers, MYH3 and MYH814, was common to the regenerative myofibers of both humans and mice (Fig. 1b, Supplementary Fig. 1c, d, Supplementary Table 3, and Supplementary Data 1). Upregulation of these markers was not common in other cell types such as satellite cells, immune cells, or fibroblasts (Supplementary Fig. 1e–g). Single-nucleus RNA sequencing (snRNA-seq) data from both human and mouse skeletal muscle24 were analyzed to evaluate the distribution of ANXA11 expression across transcriptionally distinct cell populations (Supplementary Fig. 1h, i). These plots demonstrate that ANXA11 is specifically enriched in myogenic populations, particularly in regenerating, reinforcing its biological specificity in the DMD context. The upregulation of ANXA11 in bicipital muscle from DMD patients (Supplementary Table 1) was confirmed at both the protein level (~ 50 kDa) (Fig. 1c) and the RNA level (Fig. 1d). Similar findings were observed in the tibialis anterior (TA) muscle tissues at both protein (Fig. 1e) and RNA levels (Fig. 1f) in 12-week-old mdx mice during the chronic regenerative phase26. Upregulation of MYH3 and MYH8 was confirmed in tissue samples from both mdx mice and DMD patients, supporting constant myofiber regeneration in DMD pathology (Fig. 1c, e)12,14. However, persistent degeneration, as evidenced by the reduced expression of adult myosin markers MYH1/2, suggests that MYH3+ regenerative myofibers are unable to fully compensate for ongoing muscle damage (Fig. 1c, e). These findings suggest that insufficient regeneration during the maturation phase contributes to the progressive pathology seen in DMD. We further identified that the upregulated ANXA11 in centronucleated regenerative myofiber was specifically co-localized with MYH3+ myofiber in mdx mice (Fig. 1g), supporting a role of ANXA11 in MYH3+ myofiber maturation. Anxa11 knockout restores muscular regeneration and motor function in mdx mice To understand the physiological role of ANXA11 in regenerative myofiber, we specifically knocked out Anxa11 in wild-type mice (Anxa11mKO) using the MyogCreERT2 mouse line27 (Supplementary Fig. 2a). Since MyoG is crucial for the fusion process during muscle differentiation of myoblasts28, this model allows us to examine the effects of Anxa11 deletion during muscle regeneration. The Anxa11f/f mouse line served as a control. Subsequent to tamoxifen given at 5 weeks of age, we observed that ANXA11 protein levels were diminished in Anxa11mKO mice at 9 weeks of age in various muscles, including the tibialis anterior (TA) and soleus (SOL) (Supplementary Fig. 2b), representative of fast and slow muscle types, respectively29. Expression levels in the brain, spinal cord, liver, spleen, lung, and cardiomyocytes in the KO group were not different to the controls (Supplementary Fig. 2b). The significantly reduced ANXA11 expression observed was likely due to the expression of MyoG during the pubertal growth phase of adult muscle, as reported previously27. Although no significant differences were observed in body weight (Supplementary Fig. 2c) or in the myofiber diameter compared to control mice (Supplementary Fig. 2g), Anxa11mKO mice exhibited a marginally upregulation of MYH1/2 and MYH7 protein levels in the TA muscle (Supplementary Fig. 2e, f). Interestingly, grip strength in Anxa11mKO mice was slightly higher than that in control mice (Supplementary Fig. 2d). These results suggest the ANXA11 has a modest inhibitory effect on adult muscle growth during early development. Next, we investigated whether increased ANXA11 in regenerative myofibers contributes to the myofiber regeneration failure in mdx mice. To this end, mdx; Anxa11mKO mice were successfully generated by crossing mdx mice with Anxa11mKO mice (Fig. 2a). A treadmill fatigue test revealed that mdx; Anxa11mKO mice had significantly higher endurance during treadmill exercise than the mdx; Anxa11f/f group, approaching levels observed in control mice (Anxa11f/f) (Fig. 2b). Furthermore, motor function assessments demonstrated longer latency to fall on the rotarod in mdx; Anxa11mKO compared to mdx; Anxa11f/f mice (Fig. 2c). Additional assessments revealed that mdx; Anxa11mKO mice had significantly greater grip strength (Fig. 2d), despite no noticeable difference in body weight, compared to mdx; Anxa11f/f mice (Fig. 2e). Anxa11 knockout rescues muscular regeneration and muscular dysfunction in mdx mice. a Schematic diagram showing the Anxa11 knockout strategy in mdx mice. Figure was created in BioRender. Wen, T. (2026) https://BioRender.com/f58uh6b. b–e Motor function assessment including exhaustion time of treadmill (b), rotarod test latencies (c), grip strength (d), and bodyweight (e) in Anxa11f/f, mdx; Anxa11f/f and mdx; Anxa11mKO mice (n = 10 per group). f Immunofluorescence staining of TA muscle cross-sections from Anxa11f/f, mdx; Anxa11f/f and mdx; Anxa11mKO mice. Scale bar, 50 μm. Masson trichrome staining of TA muscle from Anxa11f/f, mdx; Anxa11f/f and mdx; Anxa11mKO mice. Scale bar, 200 μm. g Quantification of fibrotic area and diameter of CNFs (Centronuclear Fibers) of TA muscle from Anxa11f/f, mdx; Anxa11f/f and mdx; Anxa11mKO mice (n = 3 per group, 3 images per mouse). Quantification of staining intensity of ANXA11, PAX7, MYH3, and MYH2 of TA muscles from Anxa11f/f, mdx; Anxa11f/f and mdx; Anxa11mKO mice (n = 3 per group, 5 images per mouse). h Immunoblot analysis of TA muscle tissue from Anxa11f/f, mdx; Anxa11f/f and mdx; Anxa11mKO mice. i Quantification of ANXA11, MYH3, MYH1/2 and MYH7 protein expression levels (n = 3 per group). j Serum CK levels of Anxa11f/f (n = 9 per group), mdx; Anxa11f/f (n = 10 per group) and mdx; Anxa11mKO (n = 10 per group) mice. Source data are provided as a Source Data file, and the exact p-values have been provided as required in the Source Data. Data are presented as means ± SD. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, ns not significant. Statistical significance was assessed by unpaired t-test (two-sided) (g and i) or one-way ANOVA (two-sided) followed by Tukey’s multiple comparisons test (b–e and j). To evaluate the effects of Anxa11 knockout on muscular regeneration in mdx mice, immunostaining of the TA muscle further revealed elevated ANXA11 expression in regenerative myofibers with centrally located nuclei in mdx; Anxa11f/f control mice (Fig. 2f, g). This upregulation was markedly diminished following MyogCreERT2-mediated Anxa11 knockout (Fig. 2f, g). Importantly, expression of PAX7, a marker of satellite cell elevated in mdx; Anxa11f/f mice, was not different following Anxa11 knockout, suggesting that Anxa11 does not affect early satellite cell activation but rather plays a role in the late-stage of regeneration (Fig. 2f, g). MYH3 marks the regenerative myofibers in mdx mice model. Analysis of myosin heavy chain isoforms revealed the MYH3 levels were significantly lower following Anxa11 knockout, accompanied by an increased expression of MYH2 in centronuclear fibers compared to mdx; Anxa11f/f control mice, indicating a shift towards a more mature myofiber phenotype (Fig. 2f, g). Number of centronuclear cells in various skeletal muscles were significantly higher in both mdx; Anxa11f/f and mdx; Anxa11mKO mice compared with control mice, indicating ongoing active regeneration (Fig. 2f, g). Notably, the diameter of these cells was larger in mdx; Anxa11mKO mice than in mdx; Anxa11f/f control mice, convincing enhanced maturation of regenerative muscle fibers in the mdx; Anxa11mKO group (Fig. 2f, g). Moreover, analysis of the diaphragm from mdx; Anxa11mKO mice showed that Anxa11 knockout led to reduced MYH3 and increased MYH2 expression, consistent with the changes observed in TA (Supplementary Fig. 2h, i). Furthermore, fibrosis was assessed by Masson’s trichrome staining in TA and diaphragm muscles to determine the broad pathological effects (Fig. 2f and Supplementary Fig. 2h). Consistent with disease progression, various skeletal muscles in mdx; Anxa11f/f mice exhibited increased fibrosis (Fig. 2f, g and Supplementary Fig. 2h, i). Importantly, fibrosis was markedly reduced in Anxa11 knockout mice (Fig. 2f, g and Supplementary Fig. 2h, i), likely as a consequence of improved muscle regeneration. Whole-TA tissue immunoblot analysis corroborated these findings, demonstrating elevated ANXA11 and MYH3 expression in mdx; Anxa11f/f mice, accompanied by diminished MYH1/2 and MYH7 levels, indicative of ongoing muscle degeneration and regeneration. In contrast, we observed a compensatory shift with significantly lower MYH3 levels and higher expression of MYH1/2 and MYH7 in mdx; Anxa11mKO mice than in mdx; Anxa11f/f mice, further confirming enhanced myofiber maturation and regeneration (Fig. 2h, i). Levels of serum creatine kinase (CK), a marker for whole-body sarcolemma homeostasis30, were markedly lower in the mdx; Anxa11mKO group than in the mdx; Anxa11f/f group after treadmill exercise (Fig. 2j), indicating substantial improvement in systemic muscle membrane integrity and stability following regeneration restoration. Lastly, RNA-seq analysis was performed from the TA muscle tissue of the three independent groups, including Anxa11f/f, mdx; Anxa11f/f, and mdx; Anxa11mKO mice. Principal component analysis (PCA) demonstrated clear separation between Anxa11f/f, mdx; Anxa11f/f, while the mdx; Anxa11mKO group exhibited a transcriptional shift toward the Anxa11f/f state, suggesting partial phenotypic rescue (Supplementary Fig. 3a). Genes in Cluster A (n = 295, 59% of the top 500 genes) showed significant restoration toward Anxa11f/f levels in the mdx; Anxa11mKO group (Supplementary Fig. 3b, c). Enrichment analysis revealed that these genes were predominantly associated with immune and inflammatory responses (Supplementary Fig. 3d). Given that Anxa11 deletion was driven in the myogenic lineage, these changes are unlikely to reflect a direct effect of ANXA11 loss on immune pathways, but rather a secondary consequence of improved muscle homeostasis. Consistent with this interpretation, analysis of early myogenic transcriptional regulators (Supplementary Fig. 3e) identified significant alterations in mdx mice, including increased Pax711, Myf5, Myog6,11, and Mymk/Mymx9 along with reduced Fgf2 and Mstn31, which were not restored by Anxa11 knockout. This aligns with the PAX7 staining patterns (Fig. 2f) and supports a model in which ANXA11 primarily influences muscle maturation and fiber-state progression at the post-transcriptional level, rather than directly regulating early transcriptional programs. In conclusion, our results identify ANXA11 as a critical regulator of regenerative myofibers, with its upregulation impairing muscle fiber maturation in mdx mice. Genetic ablation of ANXA11 promoted the transition from an immature, MYH3-positive state to a mature myofiber phenotype, thereby enhancing regenerative capacity. This improved regeneration was associated with restored fibrillar organization and recovery of motor function in mdx mice. ANXA11 plays an inhibitory role in muscle regeneration following BaCl2 induced injury To further investigate the role of ANXA11 in muscle regeneration, we induced acute muscle injury using BaCl220 (Fig. 3a). Muscle regeneration was characterized by MYH3+ regenerative myofibers in centronucleated cells at 5 days post-injection (d.p.i.) (Fig. 3c, f), marking the early phase of muscle repair following injury20. At this stage, ANXA11 expression was notably upregulated (Fig. 3b, e), indicating its involvement in myofiber regeneration. However, Anxa11 knockout did not change the MYH3 expression in regenerative myofibers at 5 d.p.i. (Fig. 3c, f), suggesting that ANXA11 plays a limited role in the early phase of muscle regeneration. ANXA11 plays an inhibitory role in BaCl₂-induced muscle regeneration. a Schematic diagram of BaCl2-induced TA muscle injury of Anxa11mKO mice. b–d Immunofluorescence staining of ANXA11 (b), MYH3 (c), and MYH2 (d) of cross-sectional TA muscle from Anxa11f/f and Anxa11mKO mice after 0, 5, 9 days injected with BaCl2 (Scale bar, 50 μm). e–i Quantification of ANXA11 (e), MYH3 (f), MYH2 (g), CNFs (Centronuclear Fibers) diameter distribution (h), and MYH3+ or MYH3− CNFs diameter (i) (n = 3 per group, 5 images per mouse). Source data are provided as a Source Data file, and the exact p-values have been provided as required in the Source Data. Data are presented as means ± SD. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, ns, not significant as determined by unpaired t-test (two-sided). By 9 d.p.i., the previously upregulated ANXA11 exhibited a slight reduction from day 5 to day 9 following BaCl₂ injection (Fig. 3b). At the same time, MYH3 staining diminished compared to 5 d.p.i. in control mice, with the remaining MYH3⁺ centronuclear fibers showing a smaller diameter than MYH3 staining negative fibers (Fig. 3c, i). This decline in MYH3 expression as fibers mature and enlarge aligns with previous observations32. The Anxa11mKO group exhibited a significantly lower proportion of MYH3⁺ fibers compared to the control group, along with a notable increase in muscle fiber diameters than the control (Fig. 3c, f, h). Furthermore, at 14 d.p.i., partial recovery of MYH2 expression was observed in the control group, accompanied by a further reduction in ANXA11 expression. In contrast, the Anxa11mKO group displayed a significantly higher proportion of MYH2⁺ myofibers, indicative of an accelerated maturation process (Fig. 3d, e, and g), despite a slightly elevated baseline at day 0 (Fig. 3d and g). These findings suggest that Anxa11 knockout accelerates the maturation of MYH3⁺ regenerative myofibers, thereby facilitating muscle growth throughout the repair process. In summary, ANXA11 expression peaks during muscle regeneration and diminishes as MYH3⁺ regenerative myofibers mature. This upregulation of ANXA11 expression following injury plays an inhibitory role in myofiber regeneration. ANXA11 inhibits myoblast maturation and regeneration via mTOR-S6K1 axis To understand the mechanism underlying the restoration in muscle regeneration with Anxa11 knockout in mdx mice, we constructed Anxa11 knockout (KO) myoblast cell lines using the C2C12 line (Supplementary Fig. 4a). Consistent with the observations of the mouse model, Anxa11 KO cells exhibited enhanced differentiation potency, as indicated by elevated expression levels of MyHC, including MYH1/2 and MYH7, at day 3 in the differentiation medium (DM) (Supplementary Fig. 4b). More strikingly, Anxa11 KO cells showed hyperfusion potency, evidenced by a higher fusion index, whilst maintaining a state of differentiation index comparable to controls at day 3 in the differentiation medium (Fig. 4a–c). Analysis of expression of the proliferation marker Ki67 in undifferentiated myoblasts (in GM) with flow cytometry revealed no significant difference between Anxa11 KO and control groups (Fig. 4d). However, Anxa11 KO myoblasts in growth medium (GM) had larger cell sizes relative to control myoblasts (Supplementary Fig. 4c, d). Together, these findings suggested that ANXA11 regulates late differentiation and maturation of myoblasts but not the proliferation. Anxa11 plays an inhibitory role during myogenesis and regeneration of muscle tissue. a MyHC immunostaining of Anxa11 knockout C2C12 cells at DM1 and DM3. Scale bar, 50 μm. b, c Fusion index (b) and differentiation index (c) at DM3 (differentiation medium for 3 days) as shown in (a) (n = 3 per group). d Flow cytometry analysis of Ki67 expression ratio in scramble and Anxa11 knockout C2C12 in growth medium (GM) without differentiation (n = 3 per group). e Expression levels of the myogenesis markers MyoD, MYH1/2, MYH7, ANXA11 and key markers in mTOR-S6K1 signaling pathway in Anxa11 knockout cells and scramble C2C12 cells from day 0 to day 5 during differentiation. f, g Quantification of these markers (n = 3 per group) from (e), including MyoD at DM0, MYH2/7 at DM4, and ANXA11 and pS6K1/S6K1 from DM0 to DM5. h, i Immunoblot analysis (h) of ANXA11 and effectors of mTOR signaling from TA muscle tissue of Anxa11f/f, mdx; Anxa11f/f and mdx; Anxa11mKO mice, and (i) quantitation of S6 phosphorylation levels (n = 3 per group). j, k Immunoblot results (j) of bicipital muscle from DMD patients and healthy controls showing expression of ANXA11, Dystrophin, pS6K1, S6K1, and k quantitation of pS6K1/S6K1 ratio (n = 3 per group). Source data are provided as a Source Data file, and the exact p-values have been provided as required in the Source Data. Data are presented as mean ± SD. Statistical significance is indicated as *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001; ns not significant. Unpaired two-sided t tests were used for (b–d, f, i, and k), and two-way ANOVA (two-sided) with Sidak’s multiple comparisons test was used for (g). Kinase-dependent mTOR signaling plays a critical role in regulating cell size33–35, and is also essential for late-stage muscular differentiation36–38. We hypothesized that the mTOR pathway might be involved in the hyperfusion of myofibers (Fig. 4a, b) and the enlarged centronuclear myofiber diameters in the Anxa11 knockout mice (Figs. 2g and 3h). To test this, we examined whether ANXA11 regulates C2C12 cell differentiation via the mTOR-S6K1 axis. Initially, we observed an upregulation of ANXA11 protein expression from days 0 to day 3 during the differentiation of C2C12 cells, which remained constant until day 5. This upregulation coincided with a reduction in the differentiation marker MyoD and an increase in myoblast maturation markers such as MYH1/2 and MYH7 (Fig. 4e, f). Concurrently, mTOR pathway activation, indicated by phosphorylation of S6K1 and S6, was observed from day 1–3 or day 2–4, respectively, in control myoblasts (Fig. 4e, g). Following Anxa11 knockout, we detected significantly higher levels of mature MyHC isoforms, including MYH1/2 and MYH7, which corresponded with the observed hyperfusion. Moreover, Anxa11 KO cells showed a marked elevated pS6K1/S6K1 and pS6/S6 ratios from day 0, peaking at day 2 and day 3, respectively (Fig. 4e, g). The pattern of MyoD expression decline during the C2C12 differentiation from D0 to D5 remained unchanged in Anxa11 KO cells (Fig. 4e, f). These results confirm the involvement of mTORC1-S6K1 activation in regulating late-stage myotube differentiation35–37,39,40. Our findings suggest that deletion of ANXA11 activates mTORC1 pathway, although the reduction of the S6K1 phosphorylation during the late phase (D4 and D5) may depend on other upstream factors. Perturbation of the PI3K/AKT/mTOR signaling pathway has been reported in a dystrophin-deficient myoblast cell line established from the skeletal muscle from the mdx mice41. We further examined the relation between ANXA11 expression and the mTOR pathway across different pathological phases in mdx mice: the pre-weaning phase (0–2 weeks), the acute onset phase (3–6 weeks), and the chronic damage phase (> 7 weeks)26. During the acute degeneration stage (5 weeks), mTOR signaling was strongly activated in TA muscle, while ANXA11 expression remained unchanged (Supplementary Fig. 5a, b). In contrast, during the chronic damage phase (12 weeks to 6 months), S6K1 phosphorylation levels were markedly reduced despite persistent upstream mTOR activation (Supplementary Fig. 5c–f). This reduction coincided with increased ANXA11 expression in the chronic damage phase (Supplementary Fig. 5c–f). A schematic diagram illustrates this dynamic: in the acute onset phase, mechanical stress induces sarcolemma damage, triggering mTOR pathway activation to promote regeneration. However, in the chronic phase, sustained ANXA11 upregulation counteracts mTOR activity by inhibiting downstream S6K1 phosphorylation, thereby impairing muscle regeneration (Supplementary Fig. 5g). Immunoblot analysis confirmed activation of S6 phosphorylation in MyoG-Cre-induced Anxa11 knockout WT mice (Anxa11mKO) (Supplementary Fig. 4e, f). MyoG-Cre-induced Anxa11 knockout in mdx also specifically restored the inhibited S6 phosphorylation (Fig. 4h, i). Similarly, in bicipital muscle tissue collected from DMD patients, we observed an inhibition of the S6K1 phosphorylation pathways alongside elevation of ANXA11 expression (Fig. 4j, k). These findings suggest that the inhibition of signals downstream of the mTOR pathway in the chronic damage phase may serve as a common mechanism underlying myogenesis defects in dystrophic muscle42. ANXA11 selectively regulates S6K1 phosphorylation downstream of mTORC1 To explore the regulation of ANXA11 within the mTORC1-S6K1 axis in regenerative myofibers, we engineered a stable C2C12 cell line that overexpresses ANXA11. In these cells, we observed minimal regulation of S6 kinase phosphorylation (Fig. 5a, b). To highlight this effect, we created Tsc1/2 KO cells, which lack the inhibitory influence of TSC1/2 on mTOR, leading to activation of the mTOR-S6K1 axis, as evidenced by a higher pmTOR/mTOR ratio compared to control cells (Fig. 5a, b), consistent with previous reports43,44. Notably, ANXA11 overexpression almost completely abolished S6K1 phosphorylation induced by Tsc1 KO or Tsc2 KO, resulting in significant suppression of both S6K1 and S6 phosphorylation (Fig. 5a, b). In contrast, ANXA11 overexpression did not affect the elevated mTOR phosphorylation in Tsc1 KO or Tsc2 KO cells (Fig. 5a, b). Similarly, under Tsc KO conditions, we observed activation of 4EBP1 phosphorylation, which was not noticeably modulated by ANXA11 overexpression (Fig. 5a, b). These findings indicate that ANXA11 modulates mTORC1 signaling at the level of downstream complex activation rather than upstream mTOR kinase activation. ANXA11 selectively inhibits phosphorylation of S6K1 in mTOR-S6K1-S6 axis during myogenesis. a, b Immunoblot analysis (a) of key markers in mTOR-S6 signalling pathway in Tsc1 and Tsc2 knockout C2C12 cells (GM, growth medium) with overexpression of GFP or ANXA11-GFP, respectively and b quantification of grey band intensity from immunoblots for the following markers: ANXA11-GFP, S6K1 phosphorylation, TSC1/2, mTOR phosphorylation, 4EBP1 phosphorylation (n = 3 per group). c, d Expression levels of protein (c) involved in the mTOR-S6K1 pathway and myogenesis in rapamycin or vehicle-treated Anxa11 KO C2C12 cells at DM3 (day 3 in differentiation medium) and (d) quantitation of these protein levels from (c) (n = 3 per group). e, f Expression levels of proteins (e) involved in the mTOR-S6K1 pathway and myogenesis in PF-4708671 or vehicle-treated Anxa11 KO C2C12 cells at DM3 and f quantitation of these protein levels (n = 3 per group) from (e). g, h MyHC immunostaining (g) of Anxa11 knockout C2C12 cells treated with PF-4708671 and h fusion index quantification at DM3 (n = 3 per group), Scale bar, 50 μm. i Schematic diagram describing ANXA11 regulation within the mTOR-S6K1-S6 pathway during myogenesis, combined with genetic modification and chemical treatment, including rapamycin and PF-4708671. Figure was created in BioRender. Wen, T. (2026) https://BioRender.com/4i8f72g. Source data are provided as a Source Data file, and the exact p-values have been provided as required in the Source Data. Data are presented as means ± SD. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, ns not significant as determined by unpaired t-test (two-sided). We further employed rapamycin, an mTORC1 inhibitor, to determine if the specific regulatory role of ANXA11 through mTOR in C2C12 cells. Rapamycin treatment almost completely blocked C2C12 differentiation by inhibiting mTORC1-S6 axis, as evidenced by the absence of S6K1 phosphorylation and MYH7 expression (Fig. 5c, d). In contrast, the significant upregulation of pS6K1 and subsequent MYH7 synthesis observed in the KO group at DM3 was abolished upon rapamycin treatment (Fig. 5c, d). This suggests that regulation of S6K1 phosphorylation and myogenesis by ANXA11 depends on mTORC1 activity. Furthermore, PF-4708671, a selective inhibitor of S6 phosphorylation at Ser235/23645, was employed to investigate the role of ANXA11 within the mTORC1-S6K1-S6 signaling pathway (Fig. 5e, f). Notably, the activation of S6 phosphorylation, but not S6K1 phosphorylation, observed in Anxa11 knockout cells was completely reversed by PF-4708671 treatment (Fig. 5e, f). Importantly, Anxa11 knockout failed to differentiate any more following PF-4708671 treatment, as indicated by reduced MYH7 expression, even though there is heightened S6K1 phosphorylation in these cells, as reported previously (Fig. 5e, f)45. Furthermore, the improved fusion in Anxa11 knockout cells at DM3 was abolished by PF-4708671, which confirmed that ANXA11 regulates myogenesis through the mTOR-S6 axis (Fig. 5g, h). Consistent with pharmaceutical manipulation of S6 phosphorylation with PF-4708671, we re-expressed either wild-type S6 or a phospho-inactive S6 mutant (S235A/236A) in Anxa11 KO cells. The dominant negative effects of S6 mutant (S235A/236A) significantly reversed both the hyperphosphorylation of S6 and the excessive fusion observed in Anxa11 KO cells, in contrast with wild-type S6 overexpression (Supplementary Fig. 6a–d). Together, these findings highlight that phosphorylation of S6—rather than S6K1 itself—plays a critical role in regulating the late stages of myogenesis (Fig. 5i). Having established S6 signaling as a critical downstream effector of ANXA11-mediated regulation of myogenic differentiation, we next sought to determine how ANXA11 suppresses mTORC1 activity despite unchanged mTOR phosphorylation. Given that mTORC1 activation requires lysosomal recruitment of mTOR46,47, we investigated whether ANXA11 regulates mTOR subcellular localization. To test this, we examined mTOR correlation with LAMP1-positive lysosomes in cells following Anxa11 knockout. In control cells, mTOR displayed weak and diffuse cytosolic staining (c). Following Anxa11 knockout, mTOR redistributed from the diffuse cytosolic pool to punctate structures associated with lysosomal membranes, resulting in locally brighter staining (Supplementary Fig. 6e). Although the overall LAMP1 staining pattern remained unchanged, Pearson’s correlation analysis demonstrated a significant increase in the association between mTOR and LAMP1, which was further supported by line profile analysis in selected regions of interest (ROIs) (Supplementary Fig. 6f, g). Together, these findings indicate that Anxa11 knockout alters the spatial distribution of mTOR, increasing its association with lysosomes. Functional inhibition of S6 signaling in vivo phenocopies ANXA11 overexpression and impairs regenerative myofiber maturation To further investigate whether ANXA11 regulates myogenesis through the mTOR-S6K1 pathway in vivo, the S6K1 kinase inhibitor PF-4708671 was administered via intraperitoneal injection. PF-4708671 treatment effectively suppressed the enhanced S6 phosphorylation observed in the TA muscle of ANXA11-deficient mdx mice (Fig. 6a, b). This inhibition disrupted the improved regenerative capacity associated with Anxa11 knockout, as indicated by elevated MYH3 expression levels detected via immunoblot analysis (Fig. 6a, b). Additionally, immunostaining revealed a persistent accumulation of small MYH3+ centronuclear cells and a reduction in MYH2 expression in mdx; Anxa11mKO mice following S6 phosphorylation inhibition (Fig. 6c, d), suggesting impaired myofiber maturation. This disruption in muscle regeneration was functionally significant, as PF-4708671-treated mdx; Anxa11mKO mice exhibited markedly reduced grip strength, although body weight remained unchanged (Fig. 6e). Importantly, the same dosage of PF-4708671, which effectively inhibited S6 phosphorylation, did not induce MYH3+ centronuclear cells or alter MYH2 expression in adult myofibers in wild-type mice (Supplementary Fig. 7a–d). Accordingly, body weight and grip strength in wild-type mice with PF-4708671-treated adult myofibers also remained unaffected (Supplementary Fig. 7e). These findings suggest that the accumulation of MYH3+ centronuclear cells observed in mdx; Anxa11mKO mice treated with PF-4708671 is not due to off-target toxicity or injury caused by the inhibitor, but rather reflects of blockage of regenerative myofiber maturation process. S6 inhibition disrupted the enhanced muscle regeneration in mdx mice following Anxa11 knockout. a Immunoblot analysis of myosin isoforms (MYH3, MYH1/2) and mTOR signaling pathway components from PF-4708671 or vehicle-treated TA muscles in mdx mice following Anxa11 knockout. b Quantification of gray band intensities for MYH3, MYH1/2, and phosphorylated S6 (n = 3 per group). c Immunostaining of MYH3 and MYH2 in cross-sections of TA muscles from mdx; Anxa11mKO mice treated with PF-4708671 or vehicle (Scale bar, 50 μm). d Measurement of centronuclear fiber (CNF) diameters and quantification of MYH3 and MYH2 fluorescence intensities in images from panel (c) (n = 3 per group, 5 images per mouse). e Body weight and grip strength of mdx; Anxa11mKO mice treated with PF-4708671 or vehicle (n = 6 per group). Source data are provided as a Source Data file, and the exact p-values have been provided as required in the Source Data. Data are presented as mean ± SD. Statistical significance: *P < 0.05, **P < 0.01, ****P < 0.0001; ns, not significant as determined by unpaired t-test (two-sided). Given the inhibitory role of ANXA11 in the mTOR pathway, similar to the S6K1 inhibitor PF-4708671, we performed comparative in vivo experiments using AAV9-mediated overexpression of ANXA11. ANXA11 overexpression resulted in a marked reduction of S6 phosphorylation in the tibialis anterior (TA) muscle of both wild-type and mdx mice (Supplementary Fig. 8a–g). In mdx mice, ANXA11 overexpression led to a significant increase in MYH3 expression, indicative of impaired regenerative myofiber maturation, accompanied by reduced expression of adult myosin isoforms MYH7 and MYH1/2 (Supplementary Fig. 8a, b). Consistent with these molecular defects, ANXA11 overexpression further significantly reduced motor performance in mdx mice, as assessed by treadmill exhaustion time (Supplementary Fig. 8c). In contrast, overexpression of ANXA11 in adult myofibers of wild-type mice had minimal effects on the expression of myosin heavy chain isoforms, including the adult isoforms MYH1/2 and MYH7 as well as the embryonic marker MYH3 (Supplementary Fig. 8a, b), and did not significantly impair muscle function (Supplementary Fig. 8c). To determine whether these biochemical changes translated into altered myofiber regeneration at the tissue level, we next examined MYH3 expression by immunostaining. In agreement with the immunoblotting results, ANXA11 overexpression in skeletal muscle of uninjured wild-type mice did not induce MYH3 expression, indicating the absence of active muscle regeneration under basal conditions (Supplementary Fig. 8d and g). However, under regenerative stress—either in a BaCl₂-induced injury model at 9 days post-injury (9 d.p.i.) or in the dystrophic mdx background—ANXA11 overexpression resulted in a significantly increased accumulation of MYH3⁺ regenerating myofibers (Supplementary Fig. 8e–g). These findings indicate that ANXA11 overexpression impairs regenerative capacity specifically during active muscle regeneration, supporting a model in which ANXA11 impairs regeneration and worsens disease progression rather than initiating it. In summary, the increased accumulation of MYH3+ regenerative myofibers in mdx background, induced either by ANXA11 overexpression or pharmacological inhibition of S6K1 activity, provides further evidence that the mTOR-S6 signaling is essential for late-stage myogenesis36–38. Notably, upregulation of ANXA11 in wild-type adult myofibers does not alter MYH3/1/2 expression or directly cause muscle atrophy. This underscores the unique regenerative context of mdx muscle, which appears particularly vulnerable to maturation defects resulting from S6K1 inhibition mediated by ANXA11, compared with wild-type muscle. Collectively, our findings highlight the critical role of ANXA11 in regulating muscle regeneration via the mTOR signaling pathway, specifically within the pathological context of mdx mice. AAV9-mediated Anxa11 knockdown restores muscle regeneration and alleviates motor defects in mdx mice To investigate ANXA11 as a potential therapeutic target for Duchenne muscular dystrophy (DMD), AAV9-shAnxa11-GFP constructs, containing short hairpin RNA sequences, were employed for Anxa11 knockdown via intramuscular injection into the TA muscle of mdx mice (Fig. 7a). Anxa11 knockdown was initially confirmed using immunofluorescence staining (Fig. 7d, e) and immunoblots analysis (Fig. 7f, g), which showed significantly lower ANXA11 protein levels in treated mice compared to controls. Anxa11 knockdown with intramuscular injection of AAV9-shRNA rescues muscle dystrophy defects through activation of myoblast regeneration in mdx mice. a Schematic diagram showing the shAnxa11 knockdown strategy in mdx mice. Figure was created in BioRender. Wen, T. (2026) https://BioRender.com/rj6zdwq. b Body weight of AAV9-shAnxa11-GFP or control AAV9 construct injected mdx mice (n = 8 per group). c The maximum isometric tetanic force (mN) of the TA muscles in different groups was measured with Aurora force measurement system (Aurora Scientific, model 1300 A, n = 8 per group). d, e HE (n = 3 per group, 3 images per mouse) and immunofluorescence staining (n = 3 per group, 5 images per mouse) analysis (d) of cross-sectional TA muscle of mdx mice following AAV9-shAnxa11-GFP intramuscular injections (scale bar, 50 μm) and e diameter of CNFs (Centronuclear Fibers) and ANXA11, MYH3 and MYH2 immunostaining intensity in cross-sectional TA muscle of mdx mice following AAV9-shAnxa11-GFP intramuscular injections. f, g Immunoblot analysis (f) of mTOR signaling pathway from AAV9-shAnxa11-GFP injected TA muscle in mdx mice and g quantification of ANXA11, MYH3, MYH7, and S6 phosphorylation from the immunoblot (n = 3 per group). Source data are provided as a Source Data file, and the exact p-values have been provided as required in the Source Data. Data are presented as means ± SD. *P < 0.05, **P < 0.01, ****P < 0.0001, ns not significant as determined by unpaired t-test (two-sided). Despite no detectable differences in body weight between the groups (Fig. 7b), the maximum isometric tetanic force of the TA muscles was significantly higher in the AAV9-shAnxa11-GFP construct-injected mdx group compared to the control, indicating greater muscle function (Fig. 7c). AAV9 delivery effectively targeted nearly all regenerative myofibers with centrally nuclear at the local injection site, as indicated by the GFP tag expression, in both control and gene knockdown group (Fig. 7d). Muscle regeneration was assessed via H&E staining, following which we observed significantly enlarged centronuclear myofibers in the Anxa11 knockdown group than in the control group (Fig. 7d, e). In parallel, we observed downregulation of MYH3 and upregulation of MYH2 in centrally nucleated myofibers of AAV9-shRNA-treated TA muscle, accompanied by constant PAX7 expression (Fig. 7d, e), indicating comparable satellite cell activation between groups. These findings further support enhanced maturation of regenerative fibers following Anxa11 knockdown (Fig. 7d, e). Whole-tissue immunoblots analysis gave further evidence of the activation of the S6K1/S6 phosphorylation pathway activation, accompanied by higher expression of MYH7 and lower expression of MYH3 in the Anxa11 knockdown group (Fig. 7f, g). All these findings suggest improved muscle maturation similar to what was observed in mdx; Anxa11mKO mice (Fig. 2). Discussion There is growing recognition that, alongside myofiber membrane instability, the failure of muscle regeneration plays a significant role in the progression of DMD6. To systematically identify key factors pivotal to muscular regeneration beyond the dystrophin mutation in DMD, we performed a comprehensive analysis from proteome and transcriptome data of muscle tissue in the mdx mouse model and DMD patients22–24. Our analysis revealed that upregulated expression of ANXA11 is one of the foremost genetic signatures in regenerative myofibers common to both DMD patients and mdx mice (Fig. 1). Anxa11, as the critical genetic modifier, is significantly upregulated in the MYH3+ regenerative myofibers of mdx mice (Fig. 1g). MYH3, an intermediate myosin isoform expressed during muscle regeneration, is commonly used as a marker of regeneration12. However, these MYH3+ fibers are often immature and may exhibit structurally fragility compared with fully matured myofibers, which render them more susceptible to damage, especially in the mechanically stressed environment characteristic of dystrophic muscles15. In both the mdx mouse model and in patients with Duchenne muscular dystrophy (DMD), Anxa11 upregulation in regenerative myofibers, a key pathological feature, reflects the constant myogenesis induced with sarcolemma instability. This maladaptive upregulation disrupts regeneration by hindering the transition from MYH3⁺ regenerative fibers to MYH1/2⁺ adult myofibers, thereby failing to compensate for the continuous degeneration of mature fibers. Consistent with this, Anxa11 knockout (or knockdown) in regenerative myofibers markedly accelerated muscle maturation, as evidenced by increased size of centronucleated myofibers and reduced abundance of MYH3⁺ immature fibers in both BaCl2-induced injury mice (Fig. 3) and mdx mice (Figs. 2 and 7) following Anxa11 knockout. Therefore, while MYH3+ regenerative myofibers represent an essential aspect of muscle repair, their inherent fragility underscores the need for therapeutic strategies that facilitate the complete maturation of these fibers to enhance muscle function and resilience. Supporting this model, RNA-seq analysis of mdx mice revealed that Anxa11 knockout significantly attenuated immune- and inflammation-associated transcriptional programs (Supplementary Fig. 3). This finding is consistent with the reduction in fibrosis observed by Masson trichrome staining and with the decreased presence of immature MYH3⁺ myofibers (Fig. 2f, g), which would otherwise promote persistent immune activation and drive fibrosis progression. The kinase-dependent mTORC1-S6K1 pathway has been specifically implicated in regulating the late stages of myogenesis36,38. However, the precise mechanisms governing its selective activation during muscle regeneration remain underexplored. In Anxa11 knockout C2C12 myoblasts, we observed a significant enhancement in S6K1 phosphorylation during day 2 to day 3 of differentiation, which correlated with the formation of hyperfused multinucleated myotubes (Fig. 4a–f). The enhanced myogenesis observed after Anxa11 knockout was abolished by blocking S6 phosphorylation, either through genetic expression of a phospho-inactive S6 mutant or pharmacological inhibition using PF-4708671, a pS6 inhibitor (Fig. 5e–h and Supplementary Fig. 6). The further accumulation of MYH3 in centronucleated myofibers in the mdx background, induced by either ANXA11 overexpression or PF-4708671, identifies that ANXA11 inhibits the MYH3+ regenerative myofiber maturation by suppressing S6 phosphorylation (Fig. 6 and Supplementary Fig. 8). Consistent with this, RNA-seq analysis (Supplementary Fig. 3) supports our central conclusion and aligns with the PAX7 staining patterns (Figs. 2f and 7d), demonstrating that Anxa11 promotes skeletal muscle differentiation and maturation primarily through post-transcriptional regulation via mTOR signaling rather than through direct transcriptional regulation. Notably, the selective effect of ANXA11 on S6K1 phosphorylation highlights a potential therapeutic avenue for enhancing muscle regeneration while limiting the off-target consequences often associated with broad mTORC1 activation48,49. Ultimately, the CRISPR-based gene editing approach has encountered a significant challenge for dystrophin-targeted treatments, partly due to the need for further optimization of AAV delivery in satellite cells50. In our study, nevertheless, intramuscular injection of AAV9-shAnxa11 can effectively target the regenerative myofibers, characterized by centrally nucleated cells in mdx mice (Fig. 7d, e). The AAV9-mediated shRNA delivery achieved a restoration of tissue homeostasis comparable to that of genetic knockout strategies, with similar muscular maturation observed in histochemical analysis in mdx mice (Figs. 2f and 7d). Although the functional restoration in AAV9-injected TA muscles was less pronounced compared to the improvements observed in mdx; Anxa11mKO mice, the maximum isometric tetanic force was significantly enhanced following Anxa11 knockdown (Fig. 7c). This finding is consistent with the effects of intramuscular injections, which effectively enhanced muscle regeneration locally but not systemically. Considering these limitations of intramuscular injection, future efforts are best focused on identifying a more robust muscular delivery vector to enhance the efficacy of shRNA across the entire musculature. In conclusion, we propose that ANXA11 functions as a regulatory factor downstream of muscle regeneration triggered by dystrophin deficiency. ANXA11 has been linked to neuromuscular diseases, including ALS and IBM51–54. Our findings suggest that sustained upregulation of ANXA11 under chronic injury conditions impairs efficient muscle regeneration in DMD. Notably, overexpression of ANXA11 in wild-type muscle does not directly induce dystrophy (Supplementary Fig. 8a–c), indicating that its pathogenic role in dystrophic muscle is mechanistically distinct from the gain-of-function effects caused by ANXA11 mutations in ALS and IBM51–54. Thus, ANXA11 emerges as a disease modifier that exacerbates DMD progression. Targeting ANXA11 may therefore represent a promising therapeutic strategy to enhance muscle maturation and improve regenerative outcomes in DMD (Supplementary Fig. 8h). Future studies will be required to delineate the precise molecular pathways regulated by ANXA11 and to evaluate its therapeutic potential in preclinical models. Methods Study participants All analyses of human DMD tissue were conducted with the approval of the Branch of Medical Research and Clinical Technology Application, Medical Ethics Committee, the First Affiliated Hospital of Fujian Medical University (Approval: 2018[019]). Written informed consent was obtained from all participants or their legal guardians, in accordance with the Declaration of Helsinki. No personally identifiable information is included in this manuscript, and all data are presented in an anonymized and aggregated manner to prevent identification of individual participants. Muscle samples were obtained from male pediatric patients diagnosed with Duchenne muscular dystrophy (DMD) and from male healthy controls. The DMD cohort (n = 3) had a mean age of 5.7 ± 3.0 years at sampling, while the control group (n = 3) had a mean age of 27 ± 10 years. All DMD patients carried pathogenic mutations in the dystrophin gene, including exon deletion, frameshift, or nonsense variants, consistent with DMD diagnosis. No family history of DMD was reported in either group. Control subjects showed no detectable pathogenic variants in dystrophin-related sequencing analyses. Due to ethical constraints, age-matched pediatric healthy muscle samples were unavailable, and control samples encompassed a wider age range. No age-associated variation was observed in the analyzed parameters among healthy controls. Animals All experiments involving mice were approved and performed in accordance with the guidelines from the Animal Research and Ethics Boards of Zhejiang University (Approval: 29800), and performed in compliance with the ARRIVE guidelines. All mice were housed in Zhejiang University’s Animal Facility, and were raised in standard conditions including a 12 h light and 12 h dark cycle with free access to water and food (Xietong Pharmaceutical Bio-Engineering Co., Ltd (XTI01FZ-010). Before experimental operations, mice were anesthetized with 2.5% isoflurane, and body temperature was maintained at 37 °C using a homeothermic heating pad. Mice were euthanized at the conclusion of the study in accordance with the guidelines of the Animal Research and Ethics Committee of Zhejiang University. For all animal experiments, mice were randomized into treatment and control groups, and the experimenters assessing the outcomes were blinded to both genotype and intervention. Unless otherwise indicated, control mice were relevant flox littermates. Control mice were Anxa11-floxed mice (Anxa11f/f) on a C57BL/6J background (Cyagen). mdx mice carrying a 4-base deletion in exon 4 (Strain No. T003035, GemPharmatech) were crossed with Anxa11f/f mice to generate mdx; Anxa11f/f control mice. We first generated the myoblast-specific Anxa11 knockout (Anxa11mKO) mice by crossing Anxa11f/f mice with a myoblast-specific Cre line (MyogCreERT2) (Strain NO. 025668, The Jackson Laboratory). We then generated myoblast-specific Anxa11 knockout mdx mice (mdx; Anxa11mKO) by crossing Anxa11mKO mice with homozygous mdx mice. Knockout of Anxa11 was induced by intraperitoneal injection of tamoxifen (75 mg/kg) for 5 consecutive days at 5 weeks of age. Comparisons were performed between Anxa11f/f, Anxa11mKO, mdx; Anxa11f/f and mdx; Anxa11mKO mice, as indicated in the figure legends. Wild-type (WT, C57BL/6J) mice were used (SLAC ANIMAL) only for ANXA11 overexpression and PF4708671 treatment experiments presented in the Supplementary Information. Male and female animals were assigned to experimental groups equally and randomly. The numbers relating to sex and all biological repeats are provided in detail in the figure legends and Source Data. Our study included both homozygous female and hemizygous male mdx mice in comparable numbers. Analyses were performed without distinguishing between sexes, as no significant differences were observed between sex (Source Data). Bioinformatics analysis Public mass spectrum data process Mass spectrum data of mdx mice and DMD patients were obtained from label-free LC-MS/MS analysis in published research22,23. No new mass spectrometry experiments, sample preparation, or wet-lab procedures were performed as part of this study. All experimental details, including sample preparation, LC–MS/MS acquisition parameters, instrument configuration, database search settings, and peptide/protein identification criteria, are described in the original publications and are not repeated here. Protein identification and quantification were performed in the original studies using Mascot-based workflows, and the criteria for peptide and protein identification followed the parameters described therein. For the present study, differential protein expression was re-evaluated using the statistical frameworks reported in the original publications. For the mouse dataset, differential protein expression was assessed using ANOVA (P ≤ 0.05), with additional quality filters (≥ 2 unique peptides, MASCOT score ≥40) and an adjusted P-values < 0.05. For the human dataset, proteins with adjusted P-values < 0.01 (Tukey’s post hoc test) were considered as differentially expressed. To ensure consistency across both datasets, a |Log2FC| threshold of 1 was uniformly applied for human and mouse analyses. Proteins with absolute Log2FC value > 1 and adjusted P-value < 0.05 (n = 4 vs 4 for mouse) or adjusted P-value < 0.01 (n = 15 vs 15 for human) were considered as significantly changed proteins. In light of the less severe clinical manifestations observed in Becker muscular dystrophy (BMD) and the role of muscle regeneration12, we focused on retaining the noteworthy alterations seen in DMD compared to the control group. Conversely, we omitted factors in which there was no significant differences between DMD and BMD. A scatter plot was made using the ggplot2 (version 3.4.1) package in R software (version 4.0.2) with Log2FC in the mouse dataset as the y-axis and in the human dataset as the x-axis. Single nuclei RNA-sequencing (snRNA-seq) analysis. Human and mouse DMD data were downloaded from PRJNA772047 and PRJNA771932, respectively24. 10× Genomics Cell Ranger software (version 6.1.2) was applied to map the raw sequencing data to the mouse (refdata-gex-mm10-2020-A) and human (refdata-gex-GRCh38-2020-A) genome provided by 10× Genomics. The gene expression matrix generated by Cell Ranger was then processed by Seurat (version 5.0.3). For all the samples, we removed cells expressing less than 200 genes. Low-quality or dying cells frequently display elevated mitochondrial gene expression. The distribution of mitochondrial gene percentages (%MT) across all samples were assessed as previously described, cells expressing more than 20% mitochondria genes were removed in our subsequent analysis55–58. Harmony (version 0.1.0) was used to integrate the samples59. The differential expression analysis and the fold change (FC) between samples were conducted by FindMarkers function with the parameters “min.pct = -Inf, logfc.threshold = -Inf, min.cells.feature = 1, min.cells.group = 1”. P-values were calculated by the Wilcoxon rank sum test and adjusted for multiple testing using the Bonferroni method. All downstream visualizations were implemented using Seurat. The area-proportional venn diagrams were generated using R package Vennerable (v3.0). Candidate gene screening was performed iteratively: we began with a conventional fold-change threshold (|Log2FC| > 1) (primarily recovered well-established regeneration markers MYH3 and MYH8) and progressively lowered it until a previously unreported and biologically meaningful factor was robustly detected in both human and mouse snRNA-seq. The final threshold used for hit selection was |Log2FC| ≥ 0.85. Bulk RNA-sequencing (RNA-seq) analysis AG RNAex Pro Reagent (Accurate Biology, 20 μL Trizol per 1 mg tissue) was used to extract total RNA. 1000 ng total RNA was used to construct RNA-seq libraries using the VAHTS Universal V6 RNA-seq Library Prep Kit for Illumina according to the manufacturer’s instructions. Library samples were then sequenced by the Illumina NovaSeq 6000 platform. Raw sequencing reads were trimmed to remove adapters by Trim Galore! (version 0.6.4_dev) and then aligned to the mouse transcripts (assembly mm10) using Hisat2 (version 2.2.1)60. The gene expression matrix (counts per gene) was quantified to the exon level (-t exon) by featureCounts (version 2.0.1)61. The average gene expression and differential expression analysis were performed using DESeq2 (version 1.28.1)62. The expression matrix generated by DESeq2 was used as the input for heatmap visualization. Principal component analysis (PCA) was performed by the “plotPCA” function in DESeq2. Gene Ontology (GO) enrichment analysis was conducted using Enrichr63. Cell culture and genetic modification Cell culture and compounds treatments 293T cells (ATCC Cat#CRL-3216) were cultured in DMEM medium (BasalMedia) supplemented with 10% fetal bovine serum (HAKATA) and 1% penicillin-streptomycin (HAKATA) at 5% CO2 and 37 °C. Mouse myoblast C2C12 cells (ATCC Cat#CRL-1772) were cultured in DMEM medium supplemented with 20% fetal bovine serum and 1% penicillin-streptomycin at 5% CO2 and 37 °C. The cells were routinely checked for morphology and growth characteristics and tested negative for mycoplasma contamination. For the myogenic differentiation experiment, myoblasts were induced with differentiation medium (DMEM medium supplemented with 2% horse serum (HAKATA)) when cells had reached the contact inhibition stage. Differentiation medium was changed every day. After 3 days of C2C12 differentiation, when the myotubes were clearly visible, the cells were collected for subsequent experiments. To regulate the mTORC1 pathway, C2C12 myotubes were treated with 20 nM rapamycin or 20 μM PF4708671 for 24 h in differentiation medium before sample collection. Plasmid DNA transfection in C2C12 cell line C2C12 cells were seeded onto 6-well plates and transfected PBTRE-ANXA11-GFP, PB-TET, and HyPBase (PiggyBac Transposon system) (1:1:2) with Hieff Trans® Liposomal Transfection Reagent (Yeasen Biotechnology), following the manufacturer’s instructions. Transfected cells were selected with 2.5 μg/mL blasticidin (MCE) at each passage to enrich for positive cells. The ANXA11-GFP and GFP coding sequence was cloned into the PBTRE donor plasmid. Wild-type S6 and its phospho-modified variants were similarly cloned into PBTRE, with S235 and S236 codons mutated to Ala (S235A/S236A, phospho-inactive) as indicated. S6 fragments were synthesized by GENEWIZ (Azenta Life Sciences). All constructs were verified by Sanger sequencing. Primers used for amplification of ANXA11-GFP and GFP were: ANXA11-F: 5′-gatcggccggatatcgaattcATGAGCTACCCTGGCTATCCCC-3′, ANXA11-R: 5′-gcttgggtctcTCAGTCATTGCCACCACAGATC-3′; GFP-F: 5′-aatgactgaGAGACCCAAGCTTGGTACCGA-3′, GFP-R: 5′-tattggatccatgtccccgggTTACTTGTACAGCTCGTCCATGCC-3′. CRISPR-Cas9 mediated gene knockout in cultured cells The following gRNAs that target the exon 3 of mouse Anxa11 gene were individually cloned into the Lenti-CRISPRv2 vector (Addgene, Cat#52961): Anxa11-gRNA1-F: CACCGATAACGTAGCCAACTATGCG; Anxa11-gRNA2-F: GGATAACGTAGCCAACTATG; Anxa11-gRNA3-F: ACGTTATCCAGCCCGATAGG. To knockout mouse Tsc1 or Tsc2, the following gRNAs were individually cloned into the Lenti-CRISPRv2 vector: Tsc1-gRNA-F: CACCGGGAGAGCAGCTCCCCAATGT; Tsc2-gRNA-F: CACCGGACTGAGTTTATCATCACAT. For lentivirus production, 293T cells were cultured in DMEM medium. Lenti-CRISPRv2-gRNA and two packaging plasmids: psPAX2 (Addgene, Cat#12260) and pMD2.G (Addgene, Cat#12259) were co-transfected into 293 T. Transfection was performed using Hieff Trans® Liposomal Transfection Reagent. Two days after transfection, lentivirus supernatant was harvested by centrifugation at 13,800 × g (Thermo Fisher Scientific Inc). C2C12 cells were infected by lentivirus in growth medium. The infected cells were screened with 2 μg/mL puromycin (MCE). Single clones were isolated and then genotyped by PCR and sequencing. Primers of genotyping: Anxa11-gDNA-F: CTTCCTCTGCTTTCCCTGGCC; Anxa11-gDNA-R: CATATGCTCACCTCCCTGCAGG. Tsc1-gDNA-F: GGACCCCGAAGCAGAGAGAAC; Tsc1-gDNA-R: TGCCTTGGTCATGGTGTCTGT. Tsc2-gDNA-F: GTGGCTGAGTTCCTCCTGCA; Tsc2-gDNA-R: AGGAGAGGAGGACCCAGTCTC. Quantitative RT-qPCR Total RNA was isolated with AG RNAex Pro Reagent (Accurate Biology, 20 μL Trizol per 1 mg tissue) according to the manufacturer’s instructions. Reverse transcription was performed with HiScript II Q RT SuperMix for qPCR kit (Vazyme), and 1 µg of total RNA was used for reverse transcription per sample in a 20 µL reaction volume. RT-qPCR reactions were performed with NovoStart®SYBR qPCR SuperMix plus kit (Novoprotein). Samples were prepared under the same conditions for the detection of each gene. Equal amounts of cDNA (1 µL per reaction) and primers (0.5 µM final) were used across samples, and reactions were run using the same cycling conditions on a BioRad-CFX96. The cycling conditions were as follows: 95 °C for 1 min (initial denaturation), followed by 40 cycles of 95 °C for 20 s (denaturation), 60 °C for 30 s (annealing/extension). Gene expression was normalized to Gapdh or GAPDH as a housekeeping gene. Values were calculated with the second derivative method and normalized to reference gene expression. The relative levels of mRNAs were calculated according to the 2−(ΔΔCt) equation. All quantitative results are presented as mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, ns, not significant. Primers (designed from Primer Bank database) for qPCR reactions were as follows: ANXA11 (forward: AGTTCAATGCGGTTCTGTGCT and reverse: GTCCCGGCCTGTCATTCTCT), Anxa11 (forward: TGGATAACGTAGCCAACTATGCG and reverse: GCGGATACATTCCATAAGGAGG). GAPDH (forward: GGAGCGAGATCCCTCCAAAAT and reverse: GGCTGTTGTCATACTTCTCATGG), Gapdh (forward: AGGTCGGTGTGAACGGATTTG and reverse: TGTAGACCATGTAGTTGAGGTCA). Cell proliferation detection C2C12 cell proliferation in growth medium was detected with PerCP-Cy5.5 mouse anti-Ki-67 antibody (BD Pharmingen) and Zombie APC-Cy7 (viability dye) (BD Pharmingen) following the manufacturer’s instructions. Fluorescence signal was detected on the Flow Cytometry Analyzer (CytoFLEX LX, Beckman). FlowJo was used for data analysis. For cell size analysis, the median FSC-A value was quantified from live, singlet-gated C2C12 cells. For proliferation analysis, a separate aliquot of cells was fixed and permeabilized for intracellular Ki-67 staining and analyzed using an independent gating strategy. Cells were first gated on a forward scatter-area (FSC-A) (Median FSC) versus side scatter-area (SSC-A) dot plot to exclude debris, followed by singlet discrimination using a side scatter-height (SSC-H) versus side scatter-area (SSC-A) plot. Ki-67–positive cells were subsequently identified based on Ki-67 fluorescence intensity relative to negative controls, using Ki-67 versus SSC-H plots. Finally, the percentage relative to the total cell counts within the single-cell gate of Ki-67 PC5.5 fluorescence intensity was calculated. Immunohistochemistry and related techniques Immunoblotting After lysing cell sample or muscle tissues in ice-cold RIPA buffer (Beyotime Biotechnology) with protease inhibitor, phosphatase inhibitors, and EDTA (Beyotime Biotechnology). Protein concentration was measured by BCA Protein Assay Kit (Coolaber). 10 to 40 μg of total protein aliquots were loaded into 7.5% polyacrylamide gels (EpiZyme Biotechnology) and then transferred to nitrocellulose membranes via Trans-Blot Turbo Transfer System (Bio-Rad). The membranes were blocked in 5% (w/v) skimmed milk in TBS with 0.1% Tween 20 (Biomed) for 1 h and then incubated overnight in the primary antibody (1:1000) diluent (4% BSA in TBS with 0.1% Tween 20) at 4 °C. The primary antibodies used were as follows: anti-ANXA11 (sc-46686, Santa Cruz Biotechnology), anti–GAPDH (AF7021, Affinity), anti-MYH7 (SAB4200670, Sigma-Aldrich), anti-β-Tubulin (M1305-2, HUABIO), anti-MyoG (sc-52903, Santa Cruz Biotechnology), anti-Myosin (Skeletal, Fast) (M4276, Sigma-Aldrich), anti–mTOR (sc-517464, Santa Cruz Biotechnology), anti-pmTOR (2974 T, Cell Signaling Technology), anti-p70 S6 kinase (sc-8418, Santa Cruz Biotechnology), anti-p-p70 S6 kinase (9234S, Cell Signaling Technology), anti-Ribosomal Protein S6 (sc-74459, Santa Cruz Biotechnology), anti-pS6 Ribosomal Protein (4858S, Cell Signaling Technology), anti-TSC1 (sc-377386, Santa Cruz Biotechnology), anti-TSC2 (4308 T, Cell Signaling Technology), anti-Dystrophin (sc-47760, Santa Cruz Biotechnology),anti-MYH8 (sc-53097, Santa Cruz Biotechnology) and anti-MYH3 (sc-53091, Santa Cruz Biotechnology). Peroxidase-conjugated anti-rabbit or anti-mouse secondary antibodies (1:10000) (EarthOx Life Sciences) were applied at room temperature for 1 h, followed by NcmECL Ultra Enhanced Chemiluminescent (New Cell & Molecular Biotech). Bands were quantitated using ImageJ software. Hematoxylin and Eosin (H&E) staining H&E staining was performed in 4 μm-thick sections that were prepared from freshly frozen tissues and fixed in 4% paraformaldehyde. Hematoxylin staining followed with 0.5% eosin solution in each section provides nuclei counterstain and cytoplasmic components, including muscle fibers, consequently. Next, sections were treated with 80% ethanol, 95% ethanol, anhydrous ethanol, xylene carbolate, and xylene consequently. Masson trichrome staining Tissue paraffin sections are deparaffinized and rehydrated through a series of xylene and graded alcohol washes as what mentioned before. Next, nuclei are stained with Weigert’s iron hematoxylin for 5–10 min, followed by rinsing in water. The sections are then treated with Biebrich scarlet-acid fuchsin solution for 5–10 min to stain cytoplasm and muscle fibers. After washing, the slides are differentiated in phosphomolybdic-phosphotungstic acid solution for 5–10 min to remove excess red stain. Subsequently, aniline blue solution is applied for 5–10 min to stain collagen fibers. Finally, sections are rinsed in distilled water, dehydrated through graded alcohols, cleared in xylene, and mounted with a resinous medium. In the final staining, collagen fibers appear blue. Immunofluorescence C2C12 cultured on coverslips (differentiated for 3 days) or Confocal dish (without differentiation), then cells were fixed in 4% paraformaldehyde for 10 min at room temperature, permeabilized with 0.25% Triton ×-100 in PBS, and blocked with 5% bovine serum albumin for 1 h at room temperature. Cells were incubated with MyHC antibody (for differentiated C2C12) (MF20, Developmental Studies Hybridoma Bank) or mTOR (2983S, Cell Signaling Technology) and LAMP1 (1D4B, Developmental Studies Hybridoma Bank) antibody (for mTOR and lysosomes colocalization analysis) overnight at 4 °C, followed by incubation with Alexa Fluor–conjugated secondary antibodies (1:200) (EarthOx Life Sciences). Nuclei were stained with Hoechst (1:1000) (Coolaber). Differentiation index was measured as the percentage of MyHC-positive cells to total cells. Fusion index was measured as the percentage of MyHC-positive cells that had three or more nuclei relative to the total MyHC-positive cells. Muscle tissues harvested and fixed in 4% paraformaldehyde overnight at 4 °C, then dehydrated in 30% sucrose at 4 °C for 3 days. After dehydration, muscle tissues were cryo-sectioned at 15 μm. After being washed with PBS, sections were blocked with 5% BSA 5% goat serum 0.5% Triton ×-100 in PBS at room temperature for 2 h. Muscle tissues were also incubated with primary antibody (1:100) overnight at 4 °C, followed by incubation with Alexa Fluor–conjugated secondary antibodies (1:200) (EarthOx Life Sciences). The primary antibodies used were as follows: anti-ANXA11 (sc-46686, Santa Cruz Biotechnology), anti-MYH7 (SAB4200670, Sigma-Aldrich), anti-MYH2 (sc-71, DSHB), anti-Ribosomal Protein S6 (sc-74459, Santa Cruz Biotechnology), anti-p-S6 Ribosomal Protein (4858S, Cell Signaling Technology), anti-Dystrophin (MANDYS105, DSHB), anti-PAX7 (PAX7, DSHB) and anti-MYH3 (sc-53091, Santa Cruz Biotechnology). Nuclei were stained with Hoechst (1:1000) (Coolaber). Fluorescence images were captured using a camera on the Olympus FV1000 Microscope Confocal Laser Scanning System or ZEISS LSM 900 with Airyscan. Mean fluorescence intensity was calculated using ImageJ software. Pearson’s r was calculated by ZEN Blue software. Line scan analysis of the pixel intensity was conducted with ImageJ. In vivo AAV9 administration AAV9-ANXA11-GFP, AAV9-GFP, or AAV9-shAnxa11-GFP and control were generated by the BrainVTA (Wuhan) Co., Ltd. AAV9-ANXA11-GFP, AAV9-GFP virus injection was performed on both side of the TA muscle of 6-week-old wild-type or mdx mice with 2.5 μL per leg at the titer of 5 × 1012 vg/mL. AAV9-shAnxa11-GFP and control virus injection was performed on both side of the TA muscle (5 × 1012 vg/mL, 2.5 μL per leg) of 6-week-old mdx mice. The sequences were as follows: shAnxa11-1: GGAAACTTTGAGAAGACTATC; shAnxa11-2: GGATTTGATCAAAGACCTAAA. Tissue samples were collected for analysis three weeks following virus injection. Muscle function test Grip strength test For each mouse, grip strength from all four limbs together were measured on a pull-down grid assembly connected to the grip strength meter (LE902, Panlab). Each individual mouse was drawn along a straight line parallel to the grid until the grip was finished, providing the peak force in grams. This was repeated three times with 5 min intervals between each measurement. Rotarod test A rotarod machine with automatic timers and falling sensors (YLS-4D, Jinan Yiyan Technology Development Co., LTD) was used in this test. Each mouse was placed on a horizontal rod that rotated about its long axis. For training: mice were placed on the rod, and the rod were accelerated to 50 rpm in accelerated mode. The test ended after all the mice had fallen off. The procedures were repeated three times per day, each time at an interval of 20 min for 2 days. To test falling time latency, the same procedure was used as in training, but the time from when each mouse was put on until it fell off was recorded. Treadmill exercise Before running, mice were acclimated to the treadmill chamber for 2 min. To measure exhaustion time, the procedures was as follows: each mouse was acclimatized at a speed of 4 m/min for 2 min, and then the speed was immediately increased to 8 m/min for 8 min. Then, the training phase began, which lasted for 30 min at 12 m/min. During the 30 min training, if a mouse could no longer run, the treadmill was turned off and a 2 min rest was enforced. Following this, the treadmill was set to a speed of 4 m/min for 2 min, followed by 12 m/min for 30 min. This process was repeated if fatigue occurred again (up to 5 times for an individual mouse). After all the training finished, the mice were rested for 2 min, then the speed was increased to 18 m/min and the time at which fatigue set in was measured for each mouse. Aurora force measurements Mice were anesthetizing with tribromoethanol, and then the tibialis anterior (TA) muscle was carefully exposed using surgical scissors, ensuring preservation of an intact connection between the TA muscle and the proximal tendon. The distal end of the TA muscle was then severed and secured firmly with sutures. The sutures were then attached to the mechanical arm of the Aurora force measurement system (Aurora Scientific, model 1300A). The knee joint was stabilized to maintain the muscle at its natural physiological length. Two sensor probes were positioned at the proximal and distal ends of the TA muscle for accurate measurements. The mechanical arm was adjusted to achieve a baseline tension of 50–60 mN. Once this setup was completed, force testing was conducted, the maximum force generated was recorded for each mouse. Serum creatine kinase (CK) level detection Tail vein blood of mice was collected at 8-week-old after treadmill exercise, and serum was separated by 400 × g centrifugation for 10 min. Serum CK values were measured using VITROS Chemistry Products CK Slides (VITROS–8478034 ORTHO-CLINICAL DIAGNOSTICS, INC.). Compounds treatment in vivo To induce muscle injury, 8-week-old TA muscle of mice were injected with 50 μl BaCl2 (50 μM MCE; HY-W111524). Muscles were then harvested at different time points post injection for the subsequent experiment. mTOR-S6 pathway regulation in vivo: Mice were intraperitoneally injected with S6 kinase inhibitor PF-4708671 (Target Mol, T2002) at a concentration of 10 mM at a daily dose of 25 mg/kg for 6 consecutive days. Statistics All data are presented as mean ± SD. Behavioral and functional data were analyzed using one-way ANOVA (two-sided), followed by Tukey’s multiple comparisons test. For Western blot and immunofluorescence quantifications, two-sided unpaired t-tests were used for predefined pairwise comparisons. For experiments involving two independent variables, two-way ANOVA (two-sided) was applied followed by Sidak’s multiple comparisons test. Statistical analyses were performed using GraphPad Prism 10 (GraphPad Software). Statistical significance was defined as *P < 0.05 as significantly different, and **P < 0.01, ***P < 0.001, ****P < 0.0001; ns, not significant (P ≥ 0.05). The statistical test used, sample size (n), and exact P values are specified in the figure legends and Source Data files. Use of AI tools ChatGPT (OpenAI) was used only to improve language clarity; all scientific content and conclusions were determined by the authors. Reporting summary Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article. Supplementary information Supplementary Information Description of Additional Supplementary Files Supplementary Data 1 Reporting Summary Transparent Peer Review file Source data Source Data 1 Source Data 2 Acknowledgements We thank Prof. Guan Kunliang (Westlake University), Prof. Zhu Dahai (Guangzhou Laboratory), Prof. Wang Xiaonan (Shanghai Jiao Tong University), Prof. Wang Qingqing, Prof. Yang Wei, Prof. Liu Wei, Prof. Zhou Yiting, Prof. Wang Liang, Prof. Li Chen (Zhejiang University), for valuable discussion. We thank Prof. Xu Haoxin’s laboratory (Zhejiang University) for their technical assistance. We thank National Natural Science Foundation of China (82071424 to Z.Z.; 82325016 to G.B.; 31871491 to J.Y.; 82450113 to X.H.), the Fundamental Research Funds for the Central Universities (2024ZFJH01-01 to Z.Z./G.B.) for funding support and startup Foundation for Hundred-Talent Program of Zhejiang University (to Z.Z.). We thank Prof. Wang Lie and the animal facility staff for their support. We would like to acknowledge Dr. Fang Sanhua, Dr. Ma Cheng, Yin Wei, and Huang Qiong in the Core Facilities, Zhejiang University School of Medicine for their technical assistance. Author contributions W.T. conducted the majority of the experiments, while B.L. performed the informatics analyses. M.J. handled the patient-related aspects. Methodology development was carried out by J.Z., T.L., Y.C., Y.Z., P.H., X.H., C.S., X.L., and Z.M. Data analysis and visualization were contributed by T.L., J.Z., Y.J., G.P., Y.C., G.L., X.Y., L.L., G.Z., K.C., and C.Z. Project conceptualization and experimental design were led by Z.Z., G.B., J.Y., and Z.W.; Z.Z. and W.T. drafted the manuscript, with G.B., A.R., P.H., J.Y., and C.S. providing review and edits. Z.Z. supervised and managed the overall project. Peer review Peer review information Nature Communications thanks Yvan Torrente, and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available. Data availability All data sets generated or analyzed during this study are included in the published article and its supplemental material. The supporting data values are available in the Supporting Data Values file, and Source data are provided with this manuscript. The raw RNA-seq data generated in this study have been deposited in the Genome Sequence Archive (GSA) in National Genomics Data Center, China National Center for Bioinformation/Beijing Institute of Genomics, Chinese Academy of Sciences database under accession code CRA031280. Human and mouse proteomic data used in this study were obtained from Tables in [D Capitanio et al. 2020] and [S Murphy et al. 2015], respectively, as described in the “Methods”. Human and mouse RNA-seq datasets (accession codes PRJNA772047 and PRJNA771932) were reused to identify cell-specific gene regulation associated with disease. Source data are provided with this paper. Competing interests Z.Z., G.B., Z.W., W.T., T.L., and J.Z. have a patent related to the data reported in this paper. The other authors declare that they have 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: Wen Tang, Bowen Lin, Ming Jin. Contributor Information Ge Bai, Email: gebai@zju.edu.cn. Zhen Zhong, Email: zhenzhong@zju.edu.cn. Supplementary information The online version contains supplementary material available at 10.1038/s41467-026-72824-8. References Associated Data Supplementary Materials Supplementary Information Description of Additional Supplementary Files Supplementary Data 1 Reporting Summary Transparent Peer Review file Source Data 1 Source Data 2 Data Availability Statement All data sets generated or analyzed during this study are included in the published article and its supplemental material. The supporting data values are available in the Supporting Data Values file, and Source data are provided with this manuscript. The raw RNA-seq data generated in this study have been deposited in the Genome Sequence Archive (GSA) in National Genomics Data Center, China National Center for Bioinformation/Beijing Institute of Genomics, Chinese Academy of Sciences database under accession code CRA031280. Human and mouse proteomic data used in this study were obtained from Tables in [D Capitanio et al. 2020] and [S Murphy et al. 2015], respectively, as described in the “Methods”. Human and mouse RNA-seq datasets (accession codes PRJNA772047 and PRJNA771932) were reused to identify cell-specific gene regulation associated with disease. Source data are provided with this paper.