Immune cell heterogeneity in response to bone fracture

To investigate the inflammatory phase of bone repair, we generated a single-nucleus RNAseq dataset of the periosteum and hematoma at day 1 post-fracture. We integrated this dataset with datasets from Perrin et al.4 containing the uninjured periosteum and the periosteum and hematoma/callus at days 3, 5, and 7 post-fracture (Fig. 1a). We obtained 23 cell clusters corresponding to 11 cell populations i.e. SSPCs, injury-induced fibrogenic cells (IIFCs), osteoblasts, chondrocytes, adipocytes, Schwann cells, fibroblasts, osteoclasts, immune cells, pericytes and endothelial cells (Fig. 1b, Fig. S1a). The immune cell population represented 30% of the dataset, with a peak in the percentage of immune cells at days 1 and 3 post-fracture before a progressive decrease at days 5 and 7 post-fracture (Fig. 1c, d, Fig. S1b). We isolated the immune cells and obtained a subset dataset containing 10 cell clusters and corresponding to 6 cell populations: macrophages (expressing Adgre1 (F4/80)), monocytes (expressing Cd52), dendritic cells (expressing Traf1), neutrophils (expressing S100a8), NK/T cells (expressing Cd247) and B cells (expressing Ighm) (Fig. 1e, Fig. S1c). The macrophage population was divided into 4 distinct subpopulations: early macrophages were present mostly at day 1 post-fracture and marked by high expression of Arg1, intermediate macrophages were present mostly at day 3 post-fracture and marked by high expression of Cd68, late macrophages 1 were found from day 5 post-fracture and expressed Mrc1 (CD206) and late macrophages 2 were found from day 3 and expressed Taco1 (Fig. 1f, g, Fig. S1d, e).

Fig. 1

The immune cell atlas of bone regeneration. a Experimental design. Nuclei were extracted from the periosteum of uninjured tibia and from the periosteum and hematoma/callus at days 1, 3, 5 and 7 post-tibial fracture of wild-type mice and processed for single-nucleus RNAseq. b UMAP projection of the integration of uninjured, day 1, day 3, day 5 and day 7 datasets. Eleven populations are identified and delimited by black dashed lines. c Feature plot of the immune lineage score in the combined fracture datasets. d Percentage of cells in the immune population per time point. e UMAP projection of the subset of the immune cell clusters of the combined fracture datasets. The six populations are delimited by black dashed lines, and macrophage subpopulations are delimited by green dashed lines. f Feature plots of the expression of Arg1, Cd68, Mrc1 (CD206), and Taco1 in the macrophage subpopulations. g Percentage of cells in the immune cell populations per time point. un. uninjured, mac. macrophages, IIFCs injury-induced fibrogenic cells, Osteob. osteoblasts

Macrophage dynamics during the inflammatory phase of bone healing

To further characterize the macrophage sub-populations in the early fracture healing environment, we performed flow cytometry and immunostaining analyses on uninjured periosteum and near adjacent muscle, and on the hematoma/callus at days 1, 3, 5 and 7 post-fracture from Cx3cr1GFP mice, in which myeloid cells are marked by the GFP fluorescent reporter (Fig. 2, Fig. S2). We confirmed the immune cell dynamics described in the snRNAseq analyses, with an increase of CD45+ and GFP+ cells at day 1 post-fracture, a peak at day 3 and a decrease at days 5 and 7 (Fig. 2b–f). In the uninjured periosteum and muscle, the CD45+, GFP+, F4/80+ macrophage population was Arg1 and CD68 negative. At day 1, we identified a peak of Arg1+ macrophages and the majority of macrophages were Cd68low, corresponding to early macrophages (Fig. 2b–f, Fig. S3). At day 3, macrophages were CD68high while the percentage of Arg1+ cells decreased, corresponding to intermediate macrophages. At day 5, macrophages were mostly CD68low and we observed a continuous decrease of Arg1+ macrophages and a peak of CD206+ macrophages, corresponding to late macrophages 1 and 2. We performed immunostaining to determine the spatial distribution of these macrophage populations in the fracture callus. GFP+ cells localized in the entire callus/hematoma at days 1 and 3 post-fracture, and progressively decreased from day 5, specifically in the regions where cartilage islets and trabecular bone are forming (Fig. 2f). Interestingly, while Arg1+ and CD68+ macrophages were detected within the hematoma/callus, CD206+ macrophages were only detected in the muscle surrounding the fracture site, suggesting that this population is a muscle-specific macrophage subset (Fig. 2f, g). Overall, combined single-nucleus transcriptomic, flow cytometry and histological analyses identified distinct macrophage subsets during the inflammatory phase of bone repair, with specific time and tissue patterning.

Fig. 2

Dynamics of macrophage subsets during bone regeneration. a Experimental design. Cells were isolated from uninjured muscle and periosteum and from the fracture environment (hematoma/callus, activated periosteum and skeletal muscle surrounding the fracture site) at 1, 3, 5 and 7 days post-fracture from Cxc3cr1GFP mice and analyzed by flow cytometry (n = 3–7 per group). b Percentage of CD45+ cells and macrophages (CD45+ GFP+ F4/80+ cells) in the uninjured periosteum/muscle and fracture environment at 1, 3, 5 and 7 days post-fracture of Cx3cr1GFP mice. c Percentage of Arg1+ macrophages (CD45+ GFP+ F4/80+ Arg1+ cells). d CD68 signal in the macrophage population (CD45+ GFP+ F4/80+ cells). e Percentage of CD206+ macrophages (CD45+ GFP+ F4/80+ CD206+ cells). f Left. Safranin’O staining and fluorescent images of GFP signal in longitudinal sections of tibial fracture at 1, 3, 5 and 7 days post-fracture. Middle and right: immunofluorescence of CD68, Arg1 and CD206 on callus hematoma/fibrosis at 1, 3, 5 and 7 days post-fracture in Cx3cr1GFP mice showing cells labeled by Arg1 or CD68 (white arrowheads) (n = 3 sections from 3 mice). g Left: Safranin’O staining of tibial fracture section of Cx3cr1GFP mice at 7 days post-fracture. Right: Immunofluorescence of CD206 and CD68 on adjacent sections shows that CD206+ macrophages are only localized in the skeletal muscle surrounding the fracture site. Scale bars: f high magnification: 1 mm, low magnification: 25 μm. g High magnification: 250 μm, low magnification: 100 μm

Distinct secretome and paracrine roles of the macrophage subsets during bone healing

Additional analyses of single-nucleus transcriptomics defined the paracrine roles of macrophage subsets in bone healing. Connectome analyses showed that the paracrine role of immune cells is transient and mostly occurs at days 1 and 3 post-fracture (Fig. 3a). Among the immune cell populations, monocytes, early macrophages, intermediate macrophages, and late macrophages 1 had the highest interaction strengths (Fig. 3b). We identified that macrophages express factors from several key signaling families, like interleukins, cytokines and PDGFs (Fig. S4a). Early macrophages expressed genes encoding pro-inflammatory factors including Cxcl2, Ccl2, Ccl7, Ccl9 and Thbs1, and predominantly at day 1 post-fracture (Fig. 3c).29,30,31 Intermediate macrophages expressed pro-repair factors such as Tgfb1, Apoe, and Pf4 mostly at day 3 post-fracture (Fig. 3c).31,32,33,34 Finally, late macrophages 1 expressed anti-inflammatory factors like Igf1, Gas6 and Pdgfc, mostly at day 5 post-injury (Fig. 3c).30,35,36 We observed the progressive shift from the pro-inflammatory phase at day 1 post-fracture to the anti-inflammatory phase at day 5 post-fracture. We also found that some pro-inflammatory genes like Spp1 and Nampt were expressed by both early and intermediate macrophages, and some anti-inflammatory genes like Il15 and Ly86 were expressed by both intermediate and late macrophages, suggesting that intermediate macrophages represent the population at the interface between the pro- and anti-inflammatory phase of bone healing (Fig. S4b). Using CellChat package,37 we identified the factors, including Nampt, Spp1, Tgfb1, Pdgfc and Igf1, expressed by early, intermediate, and late macrophages that can signal to SSPCs, IIFCs, chondrocytes and osteoblasts (Fig. 3d), unraveling multiple and dynamic interactions between macrophages and the SSPC, IIFC, chondrocyte and osteoblast populations.

Fig. 3

Distinct paracrine roles of macrophage subsets in the bone fracture environment. a Number of paracrine interactions from immune cells per time point determined by Connectome package. b Incoming and outgoing interaction plot of the immune cell populations showing that early macrophages, intermediate macrophages and late macrophages 1 are the main macrophage populations with a paracrine role after fracture. c Feature plots (left) and violin plots per macrophage cluster (in all time points, middle) and per time point (in all macrophage clusters, right) of the score of the pro-inflammatory, pro-repair and anti-inflammatory secretome in the subset of macrophages. d Dot plot of the significant interactions from early, intermediate and late macrophages 1 to SSPCs, IIFCs, chondrocytes and osteoblasts determined by CellChat package. eMac early macrophages, iMac intermediate macrophages, lMac1 late macrophages 1, lMac2 late macrophages 2, un. uninjured, Ch. chondrocytes, Ob. osteoblasts

Fibrogenic and inflammatory response of SSPCs to bone fracture

Next, we focused our analyses on the fate of SSPCs during the inflammatory phase of bone repair. We performed a subset analysis of SSPCs, IIFCs, osteoblasts and chondrocytes (Fig. 4A). As previously reported, we observed that SSPCs differentiate in 2 consecutive steps by first engaging into an injury-induced fibrogenic state before undergoing either osteogenesis or chondrogenesis (Fig. 4a, b, Fig. S5).4 We observed that injury-induced fibrogenic cells (IIFCs) were divided in 2 subpopulations, one predominant at day 1 and 3 post-fracture (cluster 2) and one predominant at day 5 and 7 post-fracture (clusters 3–5) (Fig. 4c, d). We performed gene regulatory network (GRN) analyses using SCENIC package (Single Cell rEgulatory Network Inference and Clustering) to identify regulons (transcription factors and their target genes) specifically active in the 2 IIFC subpopulations.38 IIFCs from cluster 2 had active regulons linked to the stress response/cell activation (Junb, Fos), and to several pro-inflammatory signaling pathways including JAK-STAT (Stat1, Stat3), NFκB (Nfκb1, Rela) and Interferon regulatory factors (Irf1, Irf3). IIFCs from clusters 3 to 5 had active regulons linked to tissue regeneration, such as Six1, Meis1, Pax9 and to the resolution of inflammation, such as Pparg and Etv1. (Fig. 4e).39,40,41,42 These analyses revealed that IIFCs first exhibit a pro-inflammatory profile before switching to a pro-repair/anti-inflammatory profile. Cell interaction analyses showed that the secretome of IIFCs also changes during the steps of bone healing (Fig. S4c). Pro-inflammatory IIFCs (pIIFCs) expressed genes coding for pro-inflammatory factors including Cxcl2, Cxcl5, Ccl2, Ccl5, and Csf1 while anti-inflammatory IIFCs (aIIFCs) expressed several pro-repair/anti-inflammatory genes such as Igf1, Bdnf, Tgfb3, Gas6, Hgf and Inhba (Fig. 4f). CellChat analyses established SSPCs and IIFCs as the main sources of signals for the immune cells after injury and more importantly for macrophages (Fig. S4d), indicating that signaling from IIFCs can also modulate the inflammatory response to bone fracture. Using our previously published scRNAseq datasets,8 we confirmed that muscle SSPCs also adopt a pIIFC profile followed by an aIIFC profile, associated with secretion of pro-inflammatory and anti-inflammatory factors respectively (Fig. S6a–d). Overall, during their fibrogenic response to bone fracture, SSPCs adopt successive inflammatory profiles that parallel the temporal dynamic of macrophages.

Fig. 4

Inflammatory profile of SSPCs in response to bone fracture. a UMAP projection of the subset of SSPCs, IIFCs, osteoblasts and chondrocytes from the combined fracture dataset. b Schematic representation of the differentiation trajectory of pSSPCs after bone fracture. c Percentage of cells in the different populations per time point. d UMAP projection of the cells from the day 1 and day 5 post-fracture datasets in the integrated dataset. e Heatmap of the regulon activity upregulated in pro-inflammatory IIFCs (pIIFCs, left) and upregulated in anti-inflammatory IIFCs (aIIFCs, right). f Feature plots and violin plots per time point of the score of the expression of pro-inflammatory factors (left) and anti-inflammatory factors (right). pSSPCs periosteal skeletal stem progenitor cells

Prolonged inflammatory response in musculoskeletal trauma

We previously showed that musculoskeletal traumatic injury (i.e. combined fracture and adjacent muscle injury) causes fibrotic accumulation in the fracture callus and subsequent bone nonunion (Fig. 5a).8 At 21 days post-fracture, the callus was fully ossified with almost complete resorption of fibrous tissue8 and we did not detect CD68+ and CD206+ macrophages in callus fibrosis. However, GFP+ cells were detected in the newly formed bone marrow tissue surrounding bone trabeculae (Fig. 5a–c). After MTI, we observed accumulation of GFP+ macrophages expressing CD68 and CD206 in the persistent fibrous tissue within the callus (Fig. 5b, c). This suggested that the macrophage dynamic is perturbed in MTI and could contribute to fibrotic accumulation. To investigate the impact of MTI on the immune cell response to injury, we performed flow cytometry analyses of the cells from the fracture hematoma/callus and surrounding muscles after MTI and compared with the fracture data from Fig. 2 (Fig. 5d). At day 1 post-injury, we observed similar percentages of CD45+ cells in MTI and fracture, but decreased percentages of macrophages, Arg1+ and CD68low macrophages in MTI, indicating a delay in macrophage recruitment or activation after MTI (Fig. 5e, f). At day 3 post-injury, no differences were observed between fracture and MTI. At days 5 and 7, we detected a decrease in immune cells in the fracture callus compared to day 3, while in MTI the percentages of CD45+, GFP+ cells, total macrophages, Arg1+, and CD68high macrophages remained high and were significantly increased compared to fracture (Fig. 5e, f, Fig. S2c). In addition, the percentage of CD206+ macrophages continued to increase and remained high in MTI compared to fracture. Analyses on tissue sections confirmed the accumulation of macrophages in the callus and skeletal muscle surrounding the fracture site (Fig. 5g), showing an altered immune response marked by a delay in the resolution of inflammation at days 5 and 7 post-MTI.

Fig. 5

Musculoskeletal trauma alters macrophage dynamic during bone repair. a Left: microCT images of day 21 post-fracture (top) or post-MTI (bottom) callus showing absence of bone union in MTI (white arrowheads). b Low magnification of picrosirius staining and GFP signal on adjacent sections of 21-days post-injury callus of Cx3cr1GFP mice and high magnification of the callus showing GFP+ CD206+ and GFP+ CD68+ cells in the fibrosis after MTI (white arrowheads). c Number of GFP+ and CD206+ cells in the callus fibrosis at day 21 post-fracture or MTI (n = 3–4 per group). d Experimental design. Cells were isolated from muscles and hematoma/callus at days 1, 3, 5 and 7 post-fracture or MTI of Cx3cr1GFP mice and analyzed by flow cytometry (n = 3–8 per group). e Percentage of CD45+ cells, macrophages (CD45+ GFP+ F4/80+), and percentage of Arg1+ and CD206+ macrophages in uninjured tissue and at 1, 3, 5 and 7 days post-fracture or MTI of Cx3cr1GFP mice. f CD68 signal in the macrophage population (CD45+ GFP+ F4/80+ cells). Statistical difference was determined between the median of fluorescent signal per sample. g Top: Safranin’O staining of day 5 post-fracture or MTI callus. Middle-bottom: high magnification of the callus and muscle sections of Cx3cr1GFP mice showing increased amount of GFP+ CD206+ and GFP+ CD68+ cells in the fibrosis after MTI. P-value: *P < 0.05, **P < 0.01. bm bone marrow, fib fibrosis. Scale bars: a 1 mm. b low magnification: 1 mm, high magnification: 10 μm. g low magnification: 1 mm, high magnification: 100 μm

To elucidate the altered inflammatory response in MTI, we focused our analyses on the day 5 post-injury by generating a snRNAseq dataset of the fracture callus 5 days post-MTI (Fig. 6a). We integrated this dataset with the day 5 post-fracture dataset and identified the same cell populations as in the bone repair atlas (Fig. 6a). The percentages of IIFCs, chondrocytes and osteoblasts were reduced in the MTI dataset compared to fracture, while the percentage of immune cells was increased (Fig. 6b, c), confirming the persistent immune response at day 5 in MTI. Detailed analyses of the immune cells showed an overall increase of immune cells, but specifically pro-inflammatory early macrophages (Fig. 6d, e). Gene ontology (GO) analyses indicated enrichment in functions related to macrophage activation and proinflammatory activity (Fig. 6f). More precisely, while eMac in fracture showed upregulation of GO related to the regulation of cellular processes, eMac in MTI showed upregulation of GO linked to macrophage function, revealing their increased activity (Fig. S7). We observed an increased expression of pro-inflammatory signals by immune cells and a decrease in pro-repair and anti-inflammatory factors expression (Fig. 6g). Therefore, MTI alters the inflammatory response to fracture, with a delay in macrophage recruitment, a prolonged pro-inflammatory phase and a delayed anti-inflammatory phase and resolution of inflammation.

Fig. 6

Prolonged pro-inflammatory state of macrophages and SSPCs after musculoskeletal trauma. a Experimental design. Nuclei were extracted from the fracture callus at day 5 post-fracture or post-MTI and processed for snRNAseq. UMAP projection of the integration of the day 5 post-fracture and post-MTI datasets. b UMAP projection of the integrated dataset separated by dataset of origin. c Percentage of cells in the IIFC, immune cell, chondrocyte and osteoblast populations per dataset. d UMAP projection of the subset of immune cells from the day 5 post-fracture and MTI datasets. e Percentage of early, intermediate and late macrophages in the total dataset from day 5 post-fracture and MTI. f Gene ontology (GO) of the upregulated genes in macrophages (clusters 1–5) from MTI dataset compared to fracture dataset. g Violin plots of the score of pro-inflammatory, pro-repair and anti-inflammatory secretome in the subset of immune cells. h Top: UMAP projection of the subset of IIFCs, chondrocytes, and osteoblasts from the day 5 post-fracture and post-MTI datasets. Bottom: UMAP projection per dataset of origin. i Dot plot of genes markers in IIFCs, chondrocytes, osteoblasts and fibrosis genes specifically upregulated in MTI. j Percentage of pro-inflammatory IIFCs (pIIFCs) and anti-inflammatory IIFCs (aIIFCs) in the fracture and MTI datasets. k Violin plots of the expression of pro-inflammatory and anti-inflammatory factors in the subset dataset. l Dot plot of genes encoding for pro-inflammatory factors upregulated in MTI and genes encoding for anti-inflammatory factors upregulated in fracture. pIIFC pro-inflammatory injury induced fibrogenic cells, aIIFC anti-inflammatory injury induced fibrogenic cells, eMac early macrophages, iMac intermediate macrophages, lMac1 late macrophages 1, Chond. chondrocyte. Osteob. Osteoblast

Altered immune response of SSPCs in musculoskeletal trauma

We next focused the snRNAseq analyses on the impact of MTI on SSPC differentiation. We generated a subset of the IIFCs, chondrocytes and osteoblasts (Fig. 6h). We observed a strong reduction of IIFCs, chondrocytes and osteoblasts marker gene expression in MTI compared to fracture, and identified upregulated ECM genes in MTI (Col5a1, Col5a3, Col3a1) (Fig. 6i). Specifically, the percentage of pro-inflammatory IIFCs (pIIFCs) was increased in MTI and the percentage of anti-inflammatory IIFCs (aIIFCs) was decreased, indicating a delay in the switch from pIIFCs to aIIFCs (Fig. 6j). In parallel, we observed increased expression of pro-inflammatory factors, such as Csf1, Ccl2 and Cxcl5, and decreased expression of anti-inflammatory factors (Igf1, Gas6, Inhba) (Fig. 6k, l). We confirmed these observations using scRNAseq datasets of SSPCs in uninjured muscles and muscles adjacent to the fracture site at days 3 and 5 post-fracture or MTI (Fig. S6e). We also identified pIIFC and aIIFC populations at day 3 post-injury with a delay in the switch from pIIFCs to aIIFCs at day 5 post-injury in the MTI dataset, correlated with increased expression of pro-inflammatory factors and decreased expression of anti-inflammatory factors (Fig. S6f–h). Overall, MTI causes an altered immune response at the fracture site, with delayed macrophage recruitment, delayed resolution of inflammation in SSPCs and macrophages, and accumulation of macrophages in the fibrotic tissue at late stages of repair.

Macrophages promote fibrotic accumulation in musculoskeletal trauma

We then assessed functionally the contribution of macrophages to bone fibrosis and nonunion in MTI. We first assessed bone healing in Ccr2−/− mice where systemic recruitment of macrophages is impaired.43 We induced MTI in Ccr2−/− mice and Ccr2+/+ controls and observed 75% of bone union in Ccr2−/− mice compared to 20% in controls 35 days post-injury, suggesting that reducing macrophage recruitment in MTI can ameliorate bone healing (Fig. 7a, b). Improved healing in Ccr2−/− mice was associated with a drastic reduction in callus fibrosis 21 days post-fracture and fewer CD206+ and CD68+ cells in the callus (Fig. 7c, d). We then used an inducible approach to deplete macrophages from days 2 to 4 post-MTI to prevent the prolonged inflammation observed in MTI. We used LysMCre; R26tdTom mice that efficiently mark all macrophage populations including CD68+ and CD206+ cells accumulating in the fibrotic tissue of MTI (Fig. S8a). Diphteria toxin injection in LysMCre; R26tdTom/IDTR mice from days 2 to 4 post-MTI decreased the percentage of macrophages (F4/80+ CD45+ Tomato+ cells) and the percentage of Arg1+ and CD68High macrophages at day 5 post-fracture (Fig. S8b). By day 35, we observed improved bone healing, with bone union/semi-union observed in 50% of LysMCre; R26tdTom/IDTR mice compared to 0% in LysMCre; R26tdTom control mice (Fig. 7e, f). Macrophage depletion led to a significant decrease in fibrosis volume and fewer CD206+ and CD68+ macrophages in the callus fibrosis of depleted mice compared to control mice at 21 days post-MTI (Fig. 7g, h).

Fig. 7

Macrophage depletion prevents fibrotic accumulation in musculoskeletal trauma. a Experimental design. MTI was induced in Ccr2+/+ and Ccr2−/− mice, in which macrophages fail to infiltrate tissues during inflammation. Representative microCT scan of 35 days post-MTI callus from Ccr2+/+ and Ccr2−/− mice, showing nonunion in Ccr2+/+ mice (white arrowheads) and bone bridging in Ccr2−/− mice. b Percentage of healed calluses from Ccr2+/+ and Ccr2−/− mice showing bone union (white), semi-union (gray) or nonunion (black) on microCT scan at day 35 post-MTI (n = 4-5 mice per group). c Volume of fibrosis in the callus of Ccr2+/+ and Ccr2−/− mice at 21 days post-MTI. d Immunofluorescence on callus sections of Ccr2+/+ and Ccr2−/− mice at 21 days post-MTI showing the presence of CD206+ and CD68+ macrophages only in the fibrosis of Ccr2+/+ mice. e Experimental design. MTI was induced to LysMCre; R26tdTom and LysMCre; R26tdTom/IDTR mice, and diphteria toxin (DT) was injected at days 2, 3 and 4 post-injury to induce macrophage depletion. Representative microCT scan of 35 days post-MTI callus from LysMCre; R26tdTom and LysMCre; R26tdTom/IDTR mice, showing nonunion in LysMCre; R26tdTom mice (white arrowheads) and bone bridging in LysMCre; R26tdTom/IDTR mice. f Percentage of healed calluses from LysMCre; R26tdTom and LysMCre; R26tdTom/IDTR mice showing bone union (white), semi-union (gray) or nonunion (black) on microCT scan at day 35 post-MTI (n = 7 mice per group). g Volume of fibrosis in the callus of LysMCre; R26tdTom and LysMCre; R26tdTom/IDTR mice at 21 days post-MTI. h Immunofluorescence on callus sections of LysMCre; R26tdTom and LysMCre; R26tdTom/IDTR mice at 21 days post-MTI showing the presence of CD206+ and CD68+ macrophages only in LysMCre; R26tdTom mice. Scale bars: a–e: 1 mm, d–h: 100 μm. P-value: *P < 0.05

Pharmacological CSF1R inhibition improves bone healing in musculoskeletal trauma

Given the direct involvement of macrophages in the MTI phenotype, we sought to test a pharmacological approach by targeting macrophages to improve MTI outcome. The FDA-approved CSF1R inhibitor Pexidartinib was previously shown to efficiently deplete Tumor-associated macrophages (TAMs) in osteosarcomas.44 In fracture healing, Csf1r expression was limited to macrophages and osteoclasts and increased in MTI (Fig. S8c–f). Further, Csf1 was upregulated in IIFCs in MTI suggesting that inhibiting CSFR1 may be an efficient strategy to decrease macrophage function in MTI (Fig. 6l). We induced MTI in wild type mice and treated them daily with oral gavage of Pexidartinib or vehicle from days 3 to 14 post-injury (Fig. 8a). At 21 days post-fracture, we observed improved bone healing with a significant increase in the number of bridged sides in Pexidartinib-treated compared to vehicle-treated mice (Fig. 8b, c). CSF1R inhibition led to increased mineralized bone and cartilage in the callus and reduced fibrosis accumulation compared to control (Fig. 8d, e). In addition, we observed fewer CD206+ and CD68+ cells in the callus of Pexidartinib-treated mice (Fig. 8f). This shows that pharmacological CSF1R inhibition is a promising strategy to improve bone healing after musculoskeletal trauma.

Fig. 8

CFS1R inhibition prevents fibrous nonunion after musculoskeletal trauma. a Experimental design. Wild-type mice were treated by oral gavage of Pexidartinib or vehicle from days 3 to 14 post-MTI. b Representative microCT scan of 21 days post-MTI callus from mice treated with Pexidartinib or vehicle, showing nonunion in vehicle-treated mice (white arrowheads) and bone bridging in Pexidartinib-treated mice (n = 6–9 mice per group). c Number of bridged sides of day 21 post-MTI callus of mice treated with Pexidartinib or vehicle. d Volume of mineralized bone in 21 days post-MTI callus of mice treated with Pexidartinib or vehicle (n = 6–9 mice per group). e Percentage of cartilage and fibrosis in day 21 post-MTI callus of mice treated with Pexidartinib or vehicle. f Immunofluorescence on callus sections of vehicle- and Pexidartinib-treated mice at 21 days post-MTI showing fewer CD206+ and CD68+ macrophages in Pexidartinib-treated mice. Scale bars: b: 1 mm, f: 100 μm. P-value: *P < 0.05, **P < 0.01