All animal experiments were performed under a project license (No. 2021-X17-69), which was reviewed and approved by the Institutional Animal Care and Use Committee of Chinese PLA General Hospital. All experiments were performed in compliance with the national guidelines for the care and use of animal.

Homozygote Opg ^−/− (Tnfrsf11b^tm1Smoc, NM-KO-00004) mice and WT mice were purchased from the Shanghai Research Center for Model Organisms. All mice were housed in specific-pathogen-free conditions under a 12/12-h light/dark cycle, and movement and feeding were not restricted. Mouse genotypes were determined via PCR amplification of DNA isolated from their tails. Twelve-week-old male WT and Opg ^−/− mice were used in our experiments to exclude the influence of hormones in female mice. A random allocation was performed for each genotype.

Sciatic nerve transection procedures were as follows: Mice were anesthetized using an intraperitoneal injection of 1% sodium pentobarbital. The sciatic nerve of the right hind limb was exposed and transected to a length of 10 mm (den group). The sciatic nerve of the left hind limb was exposed without transection (sham group). Body weight and gait of all mice were evaluated at the indicated time points. The mice were sacrificed before denervation, as well as days 3, 7, and 14 post denervation, and the bilateral gastrocnemius (GAS) muscles of all mice were harvested, weighed, and stored according to standard methods. The wet weight ratio was calculated by dividing the weight of the operative side by that of the sham-operated side.

Information about reagents is listed in Supplementary Tables S1, S2.

The grip strength of mice was measured by allowing the mice to grip the metal mesh attached to the grip strength meter (Chatillon, US) and pulling them continuously. The test was repeated at least three times and the highest force for each test was recorded. The average grip strength of mice was calculated and normalized to body weight.

To evaluate changes in footprint and obtain gait images and parameters, CatWalk XT 9.0 (Noldus Information Technology, Wageningen, Netherlands) was used. Before and after denervation, mice were placed on a 1.3 m walkway and images of the footprints were captured using a high-speed camera. Three consecutive runs were conducted for each test. CatWalk XT software was used to analyze the footprint parameters, including stride length (length between successive placements of the same paw) and sway length (average width of the hind paws).

For HE staining, 10 μm-thick frozen GAS muscle slides were obtained and stained with HE according to the manufacturer’s instructions.

For Masson’s trichrome staining, GAS muscles were fixed with 4% paraformaldehyde (PFA), dehydrated and embedded in paraffin. 4 μm-thick paraffin-embedded GAS muscle slides were stained with Masson’s trichrome reagent according to the manufacturer’s instructions.

Histological slides were scanned using a PRECICE 500 (UNIC, China). ImageJ software (NIH, US) was used to quantify myofiber cross-sectional area (CSA) and minimal Feret diameter in HE slides, and fibrotic area in Masson slides. At least three random areas were analyzed for each slide.

To analyze the bone mass and microarchitecture, micro-CT was performed using the Inveon MM system (Siemens, Munich, Germany), as previously described (Liu et al., 2019). Mouse femur samples were fixed in 4% PFA for 24 h and scanned. The following parameters were calculated using the Inveon Research Workplace (Siemens): bone mineral density (BMD, mg/cm³) and bone volume/total volume (BV/TV, %) in the region of interest on the femur (1-2 mm below the distal growth plate).

Fresh GAS muscles were mounted in optimal cutting temperature medium and snap-frozen in pre-cooled isopentane in liquid nitrogen. Frozen muscle sections (10 μm) were obtained and stored at −80°C. For fiber typing, the muscle sections were permeabilized and blocked with 0.3% Triton X-100 and QuickBlock blocking buffer at room temperature for 30 min. The sections were then incubated overnight at 4°C with the primary antibody solution which consisted of anti-Myhc I (BA-D5, DSHB), anti-Myhc IIa (SC-71, DSHB), anti-Myhc IIb (BF-F3, DSHB) and anti-dystrophin (ab15277, Abcam) diluted in QuickBlock primary antibody dilution buffer. After washing three times, the sections were incubated for 1 h at room temperature with a secondary antibody solution containing Alexa Fluor^® 405 (ab175652, Abcam), Alexa Fluor^® 488 (ab150121, Abcam), Alexa Fluor™ 568 (A-21124, Invitrogen) and Alexa Fluor^® 647 (ab172327, Abcam) diluted in QuickBlock secondary antibody dilution buffer. All sections were captured using confocal laser scanning microscopy (Leica TCS SP5, Germany), and the muscle CSA were analyzed using ImageJ software (NIH, US).

GAS muscles were homogenized in RIPA lysis buffer supplemented with a protease inhibitor cocktail and the total protein content was estimated using a BCA protein concentration determination kit. Equal amounts of protein (20 μg) were separated via sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), transferred to polyvinylidene difluoride membranes (PVDF; Bio-Rad) using a semi-dry transfer system (Bio-Rad) and blocked in EveryBlot Blocking Buffer. The membranes were incubated at 4°C overnight with the following primary antibodies: Anti-OPG, anti-MuRF-1, anti-Fbx32 and anti-GAPDH. The membranes were then incubated at room temperature with anti-rabbit or anti-mouse IgG-conjugated secondary antibodies for 1 h and visualized using Clarity Western ECL substrate. Band grayscale was analyzed using ImageJ software (NIH, US) and normalized to GAPDH.

Total RNA was extracted from cell samples and GAS muscle samples using the FastPure^® Cell/Tissue Total RNA Isolation Kit V2 and cDNA was synthesized using the HiScript^® Ⅲ All-in-one RT SuperMix Perfect for qPCR. Subsequently, qRT-PCR was conducted in a CFX96™ thermal cycler (Bio-Rad, United States) using ChamQ Universal SYBR qPCR Master Mix according to the manufacturer’s instructions. The relative mRNA expression levels were calculated using the 2^−ΔΔCt method, and Gapdh was used as the internal control. Primer sequences used are listed in Supplementary Table S3.

The transcriptomic data were deposited in the GEO database (GSE220152). Total RNA from GAS samples was extracted from the sham and den groups using the TRIzol reagent. The RNA-Sequencing library was constructed in a non-strand-specific manner and sequenced using Illumina Novaseq 6,000 by the Bestnovo Medical Laboratory (Beijing, China). Sequencing reads were mapped to the mouse genome reference build mm10 using STAR aligner. Transcripts per million and differentially expressed genes (DEGs) were calculated using RSEM and ebseq packages. DEGs were defined using the following criteria: FDR <0.05, and absolute log2 fold change >1. Gene ontology (GO) analysis was performed using the clusterProfiler package. Principal component analysis (PCA) plots, heatmaps, and volcano plots were visualized with ggplot2 and heatmap packages.

Satellite cells (SCs) were sorted as previously reported (Ryall et al., 2015). Briefly, hindlimb muscles from 6-week-old male mice were harvested, and blood, tendons, lipid tissues and nerves were removed. The muscles were finely chopped until a slurry was generated. The slurry was digested with 0.2% collagenase II and 0.1% dispase at 37°C for 60 min. The slurry was centrifuged, filtered and incubated with a primary antibody cocktail mixture as follows: APC-labeled α7-integrin, PE-labeled CD11b, CD31, CD45, and Sca1 and 7-AAD. Sorting was conducted on SH800S (SONY) and the gating strategy was based on positive α7-integrin, negative 7-AAD and CD11b/CD31/CD45/Sca1 staining.

The sorted SCs populations were cultured in collagen I-coated dishes. The culture medium was F10 basic medium supplemented with 20% fetal bovine serum, 1% penicillin-streptomycin and 10 ng/mL basic fibroblast growth factor. The growth medium was changed every other day. The proliferation ability of SCs was evaluated using the BeyoClick EdU Cell Proliferation Kit following the manufacturer’s instructions. Briefly, EdU was added into the culture medium and incubated for 6 h. The cells were fixed and permeabilized and click additive solution and Hoechst 33,342 were added. The proliferation ability of SCs was calculated by dividing the number of EdU-positive cells by the total number of cells. To induce myogenic differentiation, 20,000 SCs/well were seeded in a 48-well plate coated with collagen I and cultured in a differentiation medium (2% horse serum and 1% penicillin-streptomycin in DMEM) for 48 h. The differentiation ability of SCs was evaluated using myotube staining. Briefly, myotubes were fixed with 4% PFA, permeabilized and blocked with 0.2% Triton X-100 and QuickBlock blocking buffer at room temperature, respectively. The myotubes were incubated with a primary antibody against MyHC (MF20) overnight at 4°C and subsequently incubated with the secondary antibody goat anti-mouse Alexa Fluor 568 at room temperature. The cells were counterstained with DAPI. The differentiation index was calculated by dividing the number of nuclei per myotube by the number of DAPI⁺ nuclei (all MyHC⁺ myotubes were considered). The fusion index was calculated by dividing the number of nuclei per myotube by the number of DAPI⁺ nuclei (only MyHC⁺ myotubes with two or more nuclei were considered).

Statistical analysis was conducted using GraphPad Prism 8.0 (GraphPad) and data were presented as the mean ± standard deviation (SD). Comparisons between multiple groups were evaluated using one-way ANOVA followed by Tukey’s test. A two-tailed unpaired Student’s t-test was used to compare differences between two groups. p < 0.05 was considered statistically significant.

Before we explored the effects of the Opg knockout on denervated muscle atrophy, we first verified the known phenotypes of the Opg ^−/− mice we established. Compared to WT mice, reduced body mass was observed in 3-month-old Opg ^−/− mice (Figure 1A and Supplementary Figure S1B). Further, 3-month-old Opg ^−/− mice developed an osteoporotic phenotype, as indicated by lower BMD (Supplementary Figure S1C) and BV/TV (Supplementary Figure S1D) compared to WT mice. Meanwhile, normalized to body weight, the weight of the GAS (mixed-type muscle) and whole limb grip strength were comparable in 3-month-old Opg ^−/− mice and age-matching WT mice (Supplementary Figures S1E, S1F), which was in line with what was reported previously (Hamoudi et al., 2020).

Next, sciatic nerve transection models were established in WT and Opg ^−/− mice and footprint analysis was performed before denervation and on days 3, 7 and 14 post denervation to evaluate the effects of nerve transection on WT and Opg ^−/− mice (Figure 1B). Representative footprints from the sham and injury conditions were shown (Figure 1C). As motor function was affected by both nerve and muscle, we compared the motor function by using static paw parameters, stride length and sway length (Li et al., 2021). We found that motor function was significantly impaired on day 3 post denervation and gradually recovered on days 7 and 14, which seemed no significant difference between WT and Opg ^−/− mice (Figures 1D, E). Collectively, the above results indicated that denervation severely impaired nerve and muscle function and Opg ^−/− mice displayed the similar functional recovery with WT mice after denervation.

To evaluate muscle loss, GAS muscles of each mouse were harvested before denervation and on days 3, 7 and 14 after denervation (Figure 2A) and the wet weight ratio was calculated. During the first 2 weeks of denervation, the wet weight ratio of GAS muscle decreased rapidly in both WT and Opg ^−/− mice, whereas the wet weight ratio of Opg ^−/− mice was higher than that of WT mice on days 3, 7 and 14 after denervation, which suggested that Opg ^−/− mice showed delayed GAS muscle atrophy (Figure 2B). HE staining was performed to further evaluate the effects of denervation on muscle pathology (Figure 2C). Firstly, there are increased nuclear volume, heteromorphic nuclei, vacuolated nuclei and increased number of nuclei in the course of denervation in WT and Opg ^−/− mice. Moreover, we evaluated the muscle fiber CSA of WT and Opg ^−/− mice. Regardless of denervation, the CSA and mean minimal Feret diameter of Opg ^−/− mice were significantly lower than those of the WT mice, suggesting a muscle atrophy phenotype in Opg ^−/− mice, which was consistent with previous report (Hamoudi et al., 2020) (Supplementary Figures S2A, S2B). After denervation, the GAS CSA ratio was significantly reduced in both WT and Opg ^−/− mice, but the ratio in Opg ^−/− mice was significantly higher than that in WT mice, demonstrating a delayed muscle atrophy (Figure 2D). Myofiber distribution analysis showed that the peak of CSA distribution in WT mice was higher than that in Opg ^−/− mice before denervation and Opg ^−/− mice had a higher percentage of small myofibers (<1,500 μm²) (Figure 2E). Denervation reduced myofiber CSA and the CSA distribution was not significantly different between the WT and Opg ^−/− mice (Figure 2E), which further supported the delayed muscle atrophy in Opg ^−/− mice. Denervation induced muscle fibrosis 14 days after nerve transection, as indicated by Masson’s trichrome staining (Supplementary Figure S2C). However, the fibrotic area observed in Opg ^−/− mice was significantly lower than that in WT mice (Supplementary Figure S2D). In general, GAS muscle atrophied rapidly upon denervation and Opg knockout delayed muscle atrophy and alleviated fibrosis caused by denervation.

Ubiquitin-proteasome system is an important proteolytic system with a significant role in muscle atrophy (Parveen et al., 2021). To further explore the effects of Opg knockout on denervated muscle, we measured the expression of E3 ubiquitin ligases MuRF-1 and Atrogin-1 in GAS muscles. The results showed the higher expression of MuRF-1 and Atrogin-1 in Opg ^−/− sham group than WT sham group (Figures 3A–C), suggesting the undergoing muscle atrophy in Opg ^−/− mice, which was the same as previous reports (Hamoudi et al., 2020). After denervation, compared to respective sham group, the expression of MuRF-1 and Atrogin-1 was upregulated in WT mice, whereas that in Opg ^−/− mice was discrepant (Figures 3A–C). MuRF-1 was downregulated and Atrogin-1 was upregulated in Opg ^−/− mice, whereas the increase fold of Atrogin-1 in Opg ^−/− mice was lower than that in WT mice after denervation relative to sham group (Figures 3A–C). These results suggested that the ubiquitin-proteasome pathway was activated upon denervation and Opg knockout mice exhibited milder activation compared to WT mice, which further supported the protective role of Opg knockout.

To further explore the effects of Opg knockout on specific muscle fiber type, we performed muscle fiber phenotyping of the GAS muscle types I, IIA, IIB and IIX, the percentage of which could change due to physical exercise, disuse, aging and injury (You et al., 2021) (Figures 4A, B). In summary, denervation promoted a muscle fiber type switch from slow-to fast-twitch, reflecting the change in the metabolic states of myofibers. This was consistent for both denervated WT and Opg ^−/− mice (Figures 4A, B). Type I fibers (slow-twitch myofibers) increased in Opg ^−/− mice compared to WT mice (Figure 4C). Denervation induced the loss of type I fibers (%) in both WT and Opg ^−/− mice but was especially significant in Opg ^−/− mice (Figure 4C), whereas the CSA of type I fibers showed no significant difference (Figure 4D). In GAS muscles, type IIB fibers were the primary fiber type making up more than 60% of the total myofibers and type IIB fibers increased following denervation in both WT and Opg ^−/− mice (Figure 4C). Interestingly, the CSA of WT mice was significantly reduced upon denervation, whereas Opg knockout delayed the loss of CSA of type IIB fibers (Figure 4D). The percentage of other non-type IIB fibers fluctuated in both WT and Opg ^−/− mice, which was likely to be influenced by hybrid muscle fibers, such as types IIB/IIX, IIA/IIX and IIA/IIB fibers, but the percentage of other non-type IIB fibers decreased (Figure 4C). Moreover, the change in the CSA of other non-type IIB fibers was not significant, except for type IIX fibers in WT mice, which appeared as smaller fibers (Figure 4D). Taken together, denervation promoted muscle fiber type from slow-fast transformation and Opg knockout delayed denervated muscle atrophy in a specific type IIB fiber.

To further understand the underlying molecular mechanism of how Opg knockout delayed denervated muscle atrophy, we performed transcriptome sequencing of GAS muscles 7 days post denervation, the peak of phenotypic differences between WT and Opg ^−/− mice (Figure 5A). PCA and Pearson correlation analysis revealed clear sample clusters and correlations (Supplementary Figure S3). In WT mice, compared to the sham group, the expression of 3,028 genes was significantly changed in den group in which 1,495 were upregulated and 1,533 were downregulated (Figure 5B). In Opg ^−/− mice, compared to the sham group, the expression of 3,126 genes was significantly altered in den group in which 1,551 genes were upregulated and 1,575 genes were downregulated (Figure 5B). We also performed GO analysis of DEGs after denervation in WT and Opg ^−/− mice. Upregulated pathways were mainly involved in rRNA processing, endoplasmic reticulum stress, ubiquitin-dependent catabolic process, apoptotic signaling pathway, neuron death, muscle organ development and autophagy in both WT and Opg ^−/− mice (Figure 5C). Downregulated pathways involved synapse organization, vasculature development, oxidative phosphorylation, muscle system process and mitochondrion organization in both WT and Opg ^−/− mice (Figure 5C). There was no significant difference in terms of the GO terms. However, we did find some genes associated with muscle function differentially expressed between WT den group and Opg ^−/− den group, Inpp5k, Deptor, Rbm3 and Tet2, and performed further validation using qRT-PCR (Figures 5D, E). We found that the expression of Inpp5k and Rbm3 increased upon denervation in both WT and Opg ^−/− mice; however, Opg ^−/− mice showed higher expression levels (Figure 5D), which exerted protective effects against muscle atrophy (Hettinger et al., 2021; McGrath et al., 2021). In addition, Deptor expression was only elevated in WT mice (Figure 5D), knockdown of which could ameliorate disused muscle atrophy (Kazi et al., 2011). More importantly, the expression of Tet2, coding for an important DNA demethylase (Ostriker et al., 2021; Wang et al., 2021; Zhang et al., 2022), only increased in denervated Opg ^−/− mice (Figure 5D) Therefore, we speculated that TET2 might regulate the key transcription factors to further exert protective effects on muscles.

We sorted primary SCs from WT and Opg ^−/− mice via FACS to explore the effects of Opg knockout. According to previous report, we harvested hindlimb muscles from 6-week-old male WT and Opg ^−/− mice and obtained α7-integrin⁺, 7-AAD^- and Lin⁻ (CD11b/CD31/CD45/Sca1) SCs (Supplementary Figure S4) (Ryall et al., 2015). First, we evaluated the proliferation ability of SCs from WT and Opg ^−/− mice using EdU staining and found that the number of EdU⁺ cells in WT mice was significantly higher than that in Opg ^−/− mice, suggesting that Opg knockout inhibited cell proliferation (Figures 6A, B), which was reported that OPG was a potent mitogen (Arnold et al., 2019). Moreover, we evaluated the differentiation ability of SCs via myotube staining and found that Opg knockout SCs showed better differentiation ability, as evidenced by a higher differentiation index and fusion index compared to WT SCs (Figures 6C–F). Lastly, we verified the expression of Tet2 during SCs differentiation and found the increasing expression of Tet2 in SCs differentiation and Opg ^−/− SCs had higher Tet2 expression than WT ones (Figures 6G, H). In general, OPG deficiency inhibited SC proliferation and promoted SC differentiation accompanied by higher Tet2 expression.