Temporal single-cell sequencing analysis reveals that GPNMB-expressing macrophages potentiate muscle regeneration

Abstract

Macrophages play a crucial role in coordinating the skeletal muscle repair response, but their phenotypic diversity and the transition of specialized subsets to resolution-phase macrophages remain poorly understood. Here, to address this issue, we induced injury and performed single-cell RNA sequencing on individual cells in skeletal muscle at different time points. Our analysis revealed a distinct macrophage subset that expressed high levels of Gpnmb and that coexpressed critical factors involved in macrophage-mediated muscle regeneration, including Igf1, Mertk and Nr1h3. Gpnmb gene knockout inhibited macrophage-mediated efferocytosis and impaired skeletal muscle regeneration. Functional studies demonstrated that GPNMB acts directly on muscle cells in vitro and improves muscle regeneration in vivo. These findings provide a comprehensive transcriptomic atlas of macrophages during muscle injury, highlighting the key role of the GPNMB macrophage subset in regenerative processes. Our findings suggest that modulating GPNMB signaling in macrophages may represent a promising avenue for future research into therapeutic strategies for enhancing skeletal muscle regeneration.

Fig. 1. Identification of five distinct macrophage subsets during skeletal muscle regeneration via scRNA-seq.

a A schematic showing the experimental timeline for CTX-induced skeletal muscle injury in mice, delineating the key analyses performed at various days after injury. b Immunohistochemical detection of Cd68⁺ macrophages within injured muscle at specified time intervals after injury. Scale bar, 100 µm. c Flow cytometric analysis showed a surge in CD11b⁺ cells immediately after injury, which diminishes starting from day 3. d A graphical representation of CD11b⁺ cell percentages, corresponding to flow cytometry results depicted in c. e The transition of macrophage phenotypes from proinflammatory Ly6c^hi early post-injury to anti-inflammatory Ly6c^lo by day 4, as assessed by flow cytometry. f A UMAP plot illustrating the distribution of five identified monocyte/macrophage subsets across the time series. g A bar graph showing the relative frequencies of each monocyte–macrophage subset at the designated time points after injury. h A pseudotime analysis for five monocyte–macrophage subsets, depicting the evolution of cellular states. The arrows indicate the proposed progression order through pseudotime, an analytical construct that aligns cells to represent their developmental continuum during muscle regeneration.

Fig. 2. Temporal single-cell sequencing unveils GPNMB⁺ macrophage signaling as a dominant subset of tissue regeneration.

a Trajectory plots tracing the gene expression profiles mirroring the Gpnmb pattern within the monocyte–macrophage subsets suggest synchronous gene expression events that may underpin shared functional pathways in regeneration. b UMAP series depicting the dynamic expression of the Gpnmb^hiLy6c^lo macrophage subset throughout the regenerative timeline. The UMAPs consecutively illustrate the prevalence and distribution of the Gpnmb^hiLy6c^lo population at each time point after skeletal muscle injury, highlighting the shifts in the macrophage landscape. c Flow cytometric quantification capturing the temporal prevalence of the Gpnmb^hiLy6c^lo population. The analysis delineates the frequency of this subset at progressive intervals post-injury, reflecting its involvement in the regenerative process. d Immunohistochemical staining of TA muscle sections across a temporal spectrum, identifying the presence of GPNMB-expressing cells. The arrows highlight GPNMB-positive macrophages in the injured regions, aiding in visualizing their cellular localization and temporal expression patterns during the regeneration phases. e, f Heatmap analysis summarizing the relative engagement of cell populations in GAS (e) and IGF (f) signaling pathways, offering insights into the predominant cellular functions at each time point (top) and detailing the contribution of specific ligand–receptor pairs to the composite signaling communication network (bottom).

Fig. 3. GPNMB overexpression enhances M2 macrophage polarization by modulating key regulatory genes.

a The experimental scheme for differentiating murine BMDCs (mBMDCs) into M1 and M2 macrophages, followed by LPS, IFNγ and IL-4 treatment. b Quantitative PCR analysis shows significantly higher GPNMB expression in M2 polarized macrophages than in BMDM and M1. c Western blot analysis of GPNMB protein levels in BMDM, M1 and M2 macrophages from two WT mice (#1 and #2) confirmed the increased protein levels of GPNMB in M2 macrophages. d Overexpression of GPNMB in BMDMs leads to the upregulation of M2-associated genes Arg1, Mrc1, IL-4, Mertk, Axl and Igf1r, with no significant effect on the M1 markers IL-6 and Nos2. e Overexpression of GPNMB results in heightened expression of M2-related transcription factors Irf4 and Pparg, suggesting a potential pathway for GPNMB-mediated macrophage polarization.

Fig. 4. Impaired muscle regeneration in GPNMB-KO mice after CTX injury.

a Western blot analysis confirmed the absence of GPNMB expression in GPNMB-KO mice compared with WT mice. b A schematic of the experimental design for assessing muscle regeneration in GPNMB-KO and C57BL/6 control mice after CTX injection. Mice were injected with CTX on day 0, and the TA muscles were collected for histopathological analysis on days 4 and 7 after injury. c Representative images of TA muscles from GPNMB-KO and WT mice that are uninjured and at days 4 and 7 after injury, demonstrating differences in muscle morphology. The GPNMB-KO mice exhibited impaired muscle regeneration with less striated muscle fibers than WT mice. d H&E staining of TA muscle sections revealed histological changes during regeneration, with GPNMB-KO mice showing reduced tissue repair compared with controls. e Quantitative analysis of CSA of muscle fibers in uninjured, day 4 post-injury and day 7 post-injury muscle sections confirmed statistically significant impairment of regeneration in GPNMB-KO mice.

Fig. 5. Diminished macrophage efferocytosis and impaired muscle regeneration after GPNMB KO and MERTK inhibition.

a The experimental setup and subsequent flow cytometry analysis for evaluating macrophage efferocytosis. On day 3 after CTX-induced muscle injury, mononuclear cells were isolated from both WT and GPNMB-KO mice and cocultured with CFSE-labeled, staurosporine (STS)-induced apoptotic C2C12 myoblasts for 24 h. Flow cytometry was then used to assess the phagocytosis of apoptotic cells by CD11b⁺F4/80⁺ macrophages. The rightmost graph provides a quantitative comparison between the WT and GPNMB-KO groups, showing a significant decrease in efferocytosis efficiency in the GPNMB-KO macrophages, as reflected by their reduced uptake of CFSE-labeled apoptotic bodies. b Quantitative PCR analysis reveals that overexpression of GPNMB in primary macrophages significantly increases the mRNA levels of key efferocytosis genes, Mertk and Axl. c Left: a schematic of the experimental design depicting the treatment of C57BL/6 mice with a MERTK inhibitor after CTX injury to evaluate its effect on muscle regeneration. Right: the gross morphology of TA muscles from mice treated with MERTK inhibitor doses shows dose-dependent effects on muscle appearance. d Histological examination of muscle regeneration by H&E staining at days 4 and 7 after CTX injury with or without MERTK inhibitor treatment, highlighting the impact of MERTK signaling on tissue repair. e, f Statistical analysis of muscle fiber CSA and length at days 4 (e) and 7 (f) after injury, with MERTK inhibition leading to a marked reduction in both parameters, signifying compromised regenerative capacity.

Fig. 6. GPNMB stimulation promotes muscle regeneration and myogenic differentiation of murine myoblasts.

a After GPNMB overexpression, quantitative real-time PCR was utilized to monitor the temporal expression of key myogenic regulatory factors in C2C12 cells, including Myf5, Myf2a, Myod and Myog. b Immunofluorescent staining for MyHC of C2C12 myoblast cultured in differentiation medium with (100 ng/ml or 200 ng/ml) or without rGPNMB. C2C12 myoblasts were induced to differentiate for 3 days. c, d Fusion indices were calculated by expressing the number of nuclei within MyHC-positive myotubes with ≥2 nuclei as a percentage of the total nuclei (c), and the myotube width was measured at 3 different points on the cell (d). The average width per myotube was calculated. Data are presented as mean ± s.d. of three independent experiments. e H&E staining of injured mouse TA muscle on days 4 and 7 with (10 µg/kg or 20 µg/kg) or without rGPNMB. Scale bar, 100 µm. f, g The CSA of myofibers from the injured-only group and the groups treated with rGPNMB (10 µg/kg and 20 µg/kg) on days 4 and 7. Data are presented as mean ± s.d. (n = 3) of each time point.

Fig. 7. GPNMB-expressing macrophage contributes to skeletal muscle regeneration.

The dynamic response of skeletal muscle to injury can be broadly categorized into three primary stages: the inflammatory, the inflammatory-to-regenerative and the regenerative stage. After muscle injury, muscle tissues recruit monocyte-derived macrophages, categorized into five distinct clusters (C1, C2, C3, C4 and C5) based on their inflammatory profiles. Our study identifies that macrophages with high GPNMB expression during the transitional phase from inflammation to regeneration are crucial in activating and expanding muscle progenitor cells. This process aids in the transition of macrophages toward an anti-inflammatory phenotype, which is critical for the resolution of inflammation and the subsequent muscle repair and regeneration. The dynamic shifts in macrophage subsets, highlighted by changes in the proportions of cluster 2, underscore the importance of GPNMB as an effector in muscle regeneration. Created with BioRender.com.