Resident mesenchymal progenitor cells require autocrine IGF-I in homeostatic and regenerating skeletal muscle Authors: Yangyi E Luo, Zoe Abe-Teh, Tarek Y Alsaghir, Li-Ying Kuo, Fahong Yu, Bradley E Stoker, Yumei Zhou, Ambili Bai Appu, Elijah M Dy, Feng Yue, Daniel Kopinke, Elisabeth R Barton Published: 2026 ABSTRACT SummaryFibro-adipogenic progenitors (FAPs) are muscle-resident mesenchymal stromal cells essential for homeostasis and regeneration but can produce fibrosis or intramuscular fat in pathological conditions. Insulin-like growth factor-I (IGF-I) regulates regeneration through actions on muscle fibers, satellite cells, macrophages, and extracellular matrix (ECM) remodeling but has multiple sources. To assess the role of FAP-derived IGF-I, we generated inducible FAP-specific Igf1-deficient (FID) mice. Following BaCl2 injury, FID muscles displayed impaired regeneration, with smaller fibers, fewer Pax7+ and MyoD+ cells, increased CD68+ macrophages, decreased collagen, and suppressed FAP proliferation. After glycerol-induced injury, FID muscles had reduced fat. Primary FID FAPs displayed blunted proliferation, upregulated immune-regulatory genes, and downregulated ECM and growth genes, with delayed fibrogenic and adipogenic differentiation. scRNA-seq of homeostatic muscle revealed reduced protein translation and ECM indices alongside increased senescence markers in FID samples. Taken together, FAP IGF-I is critical for FAP function, with direct and indirect impact on muscle regeneration.Keywords: fibro-adipogenic progenitors, insulin-like growth factor I, mesenchymal progenitor cells, PDGFRα, skeletal muscle regeneration, acute injury, adipogenesis, fibrosis, senescence Introduction Skeletal muscle comprises 40% of human body mass and has extraordinary regenerative capacity. Muscle regeneration is a period of dynamic remodeling that requires coordinated interaction among many cell types, including satellite cells (SCs), immune cells, and fibro-adipogenic progenitor cells (FAPs), to restore functional muscle fibers supported by a healthy extracellular matrix (ECM) (Caballero-Sanchez et al., 2024; Kuang et al., 2007; Relaix et al., 2005). FAPs are PDGFRɑ+ mesenchymal progenitor cells that reside within muscle interstitium; they proliferate upon acute injury to establish the scaffold for nascent fibers and are required for skeletal muscle regeneration and homeostasis (Collins and Kardon, 2021; Joe et al., 2010; Wosczyna et al., 2019). Skeletal muscle lacking FAPs has progressive muscle loss and delayed regeneration (Wosczyna et al., 2019). FAPs can also persist in chronic injury associated with disease, resulting in excessive fibrosis and intramuscular fat deposition (Kopinke et al., 2017; Lemos et al., 2015; Uezumi et al., 2010). Hence, FAPs create a therapeutic challenge to identify strategies where they support efficient regenerative capacity without fibro-fatty accumulation. In addition to multiple cell types, many secreted proteins coordinate muscle regeneration. One such protein is insulin-like growth factor-I (IGF-I), which facilitates skeletal muscle hypertrophy and regeneration (Barton-Davis et al., 1998; Musaro et al., 2001; Vassilakos et al., 2019). IGF-I from muscle fibers is important for post-natal growth and functional recovery from disuse atrophy (Spradlin et al., 2021; Vassilakos et al., 2019). However, other IGF-I sources also have impact on muscle regenerative capacity, including myeloid cell IGF-I that controls macrophage polarization (Tonkin et al., 2015). Single-cell RNA sequencing (scRNAseq) databases implicate FAPs as having one of the highest Igf1 expression per cell in muscle regeneration, yet the biological relevance of IGF-I from FAPs is not known (Kok and Barton, 2021). Thus, it is important to determine if IGF-I from FAPs contributes to FAP autocrine signaling or FAP paracrine effects in homeostatic and regenerating muscle (Kim et al., 2022; Uezumi et al., 2010; Wosczyna et al., 2019). In this study, we hypothesized that IGF-I produced by FAPs regulates muscle regeneration and FAP function. We examined the role of FAP-derived IGF-I on skeletal muscle regeneration and FAP properties using a mouse model with FAP-specific Igf1 deletion. Our results provide multiple lines of evidence that IGF-I from FAPs is a key autocrine regulator during muscle regeneration. Consequently, FAP IGF-I deletion disrupts FAP interactions with other muscle cell populations. Overall, FAP produces an important IGF-I source for muscle. Results Deletion of Igf1 in PGDFRα+ mesenchymal progenitor cells in skeletal muscle moderately reduces IGF-I pool FAPs are marked by the surface marker PDGFRɑ (Contreras et al., 2019; Kopinke et al., 2017; Uezumi et al., 2010; Wosczyna et al., 2019). To understand the significance of IGF-I from FAPs, we generated a conditional FAP-specific knockout mouse line named FAP-IGF-I-deficient (FID, PdgfrαCre-ERT2/+, Igf1fl/fl, Rosa26EYFP/+) (Figure 1A) (Chung et al., 2018; Spradlin et al., 2021; Srinivas et al., 2001). Since one PdgfrɑCRE−ERT2 allele showed no phenotype here and in previous studies (Figures S1A–S1D) (Norris et al., 2025; Uezumi et al., 2021), Igf1fl/fl (Pdgfrα+/+, Igf1fl/fl) served as controls (Figure 1A). Validation of FID mouse model and the efficiency of Igf1 deletion (A) Genetic scheme of Igf1fl/fl control and FID mouse models. Igf1 exon map and floxed exon 4, with primer locations (P) for expression analysis. (B) Immunofluorescent (IF) staining of PDGFRα+, EYFP, and DAPI on ND TA muscle section and FAPs. Scale bars: 200 μm. (C) Endpoint RT-PCR for Igf1 exon 4 deletion (P1 and P4) in FAPs isolated from 3 dpi. The upper band shows amplification of Igf1 cDNA from exon 3 to exon 6 and the lower bands is the result of Cre-recombination. (D) RT-qPCR for Igf1 deletion using primers P1 and P3 targeting exon 3 and exon 4 (region of deletion) and normalized against total Igf1 transcripts (primers P2 and P5, Exon 6-3′-UTR) in cultured primary FAPs, n = 3 per genotype. p = 0.0047 by unpaired t test. (E) Muscle regeneration experimental design. (F) Endpoint RT-PCR for Igf1 exon 4 deletion (P1 and P4) in TA muscles throughout regeneration, as in (C). (G) RT-qPCR for Igf1 deletion as in (D), show Igf1 deletion throughout regeneration. n = 6–8 samples for both strains for 3–14 dpi. The contralateral limb from all mice were used for ND samples (n = 27 and 30 for Igf1fl/fl control and FID mice, respectively). (H) ELISA analysis of muscle tissue IGF-I protein levels during acute muscle regeneration shows strain dependent decreases in content. n = 3∼6 for both strains from ND to 14 dpi. (I) RT-qPCR using primers for exon 6 to 3′ UTR region of the Igf1 to monitor general expression, which increases during regeneration in a strain independent manner. n = 19 for control and n = 25 for FID at ND, n = 6∼8 for both strains from 3 dpi to 14 dpi. An unpaired Student’s t test or two-way ANOVA (strains and injury time) followed by Šídák post-hoc test were used for statistical analysis. #: comparison between time points within strain and ∗: comparison between strains, with red # indicating comparisons for FID, and black # indicating comparisons for Igf1fl/fl. ∗ or #p < 0.05, ∗∗ or ##p < 0.01, ∗∗∗ or ###p < 0.001, ∗∗∗∗ or ####p < 0.0001 PDGFRɑ immunostaining colocalized with the EYFP reporter post-induction, a pattern replicated in primary FAPs isolated 3 days post-injury (3 dpi) (Figures 1B and S1E), confirming efficient and FAP-specific Cre recombination. Exon 4 deletion was assessed on primary FAPs by endpoint reverse-transcription PCR (RT-PCR) using primers flanking exon 4 (P1 and P4) (Vassilakos et al., 2019); only FID FAPs produced the smaller amplified band (Figure 1C). Quantitative reverse-transcription PCR (RT-qPCR) targeting exons 3 and 4 (P1 and P3), normalized to unaffected exons 6 and 3′UTR (P2 and P5), showed ≥80% of Igf1 transcripts lacked exon 4 in cultured FID FAPs (Figure 1D). Whole tissue expression and protein were also examined in homeostasis and regeneration (Figure 1E), considering multiple IGF-I sources in skeletal muscle (Kok and Barton, 2021). Endpoint RT-PCR confirmed exon 4 excision only in FID muscles (Figure 1F) with a 20%–45% reduction in functional Igf1 transcripts throughout regeneration (Figure 1G). IGF-I protein levels were not different in non-damaged (ND) muscles or in the early burst of IGF-I levels at 3 dpi, but were >40% lower at 5 and 7 dpi, and 15% lower at 14 dpi, significant for both time point and strain by two-way ANOVA (Figure 1H). Despite efficient recombination in FAPs, Igf1 transcription, using primers targeting exon 6 and 3′-UTR shared by all Igf1 isoforms, was not altered by the ablation of FAP Igf1 and retained similar dynamics as controls (Figure 1I). In homeostasis, minimal differences in FID and Igf1fl/fl body or tibialis anterior (TA) muscle masses were found even out to 18 months of age (Figures S1F–S1I and S2A–S2F), supporting that FAP IGF-I does not affect homeostatic muscle or that other sources compensate. IGF-I from FAPs is required for muscle regeneration FAPs are essential for muscle regeneration (Wosczyna et al., 2019). To determine if FAP IGF-I was needed for this process, we examined muscles from mice following BaCl2-induced injury (Figure 2A). While no differences in muscle mass or fiber size were observed in ND muscles, regenerating muscles had significant FID-dependent deficits in recovery, evident for relative muscle mass and fiber size at 7 dpi and 28 dpi in both sexes compared to controls (Figures 2B–2H, S1C, S1D, S1J, and S1K). Thus, FAP IGF-I contributes to the regenerative process. IGF-I produced by FAPs is required for efficient regeneration (A) Muscle regeneration study design. (B and C) Relative muscle weight in both sexes is decreased during regeneration. The muscle relative mass is obtained by normalizing the damaged LTA to the ND RTA is highly significant for day and strain by two-way ANOVA. n = 3∼4 for both strains and sexes from 3 to 28 dpi. (D) IF of Laminin on muscle cryosections during regeneration. Scale bars: 200 μm. (E and F) Median fiber size during regeneration shows time and strain dependent differences. n = 3∼4 for both strains and both sex from ND to 28 dpi. (G and H) Cumulative muscle fiber size distribution at 7 dpi show that FID muscle fibers are smaller in both sexes. Female fiber size 400–1,600 μm p < 0.0001. Male fiber size 800–2,000 μm p < 0.0001. n = 4 for each strain and sex. (I) IF of Pax7, laminin, and DAPI in 5 dpi muscle sections, where overlap between Pax7 and DAPI (pink) indicates a SC (SC). Scale bars: 100 μm. (J) Number of SCs per myofiber. n = 7 Igf1fl/fl; n = 8 for FID. Unpaired t test p = 0.0082. (K) Pax7, MyoD, and DAPI IF on freshly isolated SCs. n = 6–10 per condition. Scale bars: 100 μm. (L) Number of committed and differentiating SCs show FID SCs are delayed in maturation. (M) 3 dpi and 5 dpi EmbMHC and laminin IF. Scale bars: 200 μm. (N) EmbMHC+ quantification. %EmbMHC = # EmbMHC/total fibers. 5 dpi p = 0.0103. n = 6–8 muscles per strain. (O) IF of FAPs, laminin, and CD68+ cells in muscle sections 28 dpi following FAP transplantation or PBS injection at 2 dpi. Scale bars: 50 μm. (P) Quantification of FAPs delineating endogenous and transplanted populations. n = 4–5 per strain. (Q) Median fiber size at 28 dpi following transplantation, with FAP injection vs. PBS injection in FID muscles and PBS vs. no injection in Igf1fl/fl muscles. n = 4–5 per strain. (R) CD68+ cells normalized to myofibers at 28 dpi following transplantation, with FAP injection vs. PBS injection in FID muscles and PBS vs. no injection in Igf1fl/fl muscles. n = 4–5 per strain. An unpaired or paired Student’s t test, or two-way ANOVA (strains and injury time) followed by Šídák post-hoc test were used for statistical analysis. #: comparison between time points within strain and ∗: comparison between strains, with red # indicating comparisons for FID, and black # indicating comparisons for Igf1fl/fl. ∗ or #p < 0.05, ∗∗ or ##p < 0.01, ∗∗∗ or ###p < 0.001, ∗∗∗∗ or ####p < 0.0001. Muscle regeneration requires precise coordination between SCs and FAPs (Kim et al., 2022; Uezumi et al., 2010; Wosczyna et al., 2019). Reduced numbers of Pax7+ cells were evident in FID muscle sections at 5 dpi when SCs are most abundant (Figures 2I and 2J), suggesting that FAP IGF-I supports SC expansion following injury. As SC differentiation into MyoD+/Pax7− myoblasts is a critical lineage transition during regeneration (Bentzinger et al., 2013), primary SCs were isolated at 5 dpi and stained for Pax7 and MyoD after attachment to Matrigel-coated plates. FID SCs had a proportional increase in MyoD+/Pax7+ cells and decrease in MyoD+/Pax7− cells compared to controls (Figures 2K–2L), indicating altered SC dynamics that may impair regenerative efficiency. To evaluate fiber formation, muscle sections were immunostained for embryonic myosin heavy chain (EmbMHC) which marks newly formed fibers (Agarwal et al., 2020). At 3 dpi, EmbMHC+ fibers were sparse in both strains (Figures 2M and 2N). By 5 dpi, EmbMHC+ fiber presence increased dramatically in both strains but was significantly reduced in FID muscles. To determine if regeneration deficits in FID muscles could be corrected by FAP replenishment, 75,000 HRP+ control FAPs, GFP+ FID FAPs, or PBS were injected into FID TA muscles 2 dpi and assessed 28 dpi, with PBS or no injection into regenerating Igf1fl/fl muscles serving as controls. Similar numbers of transplanted control and FID FAPs were found at 28 dpi, when normalized to microscopic field, such that the proportions of transplanted FAPs between strains did not differ (Figures 2O and 2P). However, FID FAP transplantation resulted in smaller fiber cross-sectional areas (CSA) compared to contralateral muscles (Figure 2Q), suggesting that FID FAPs may have an actively negative effect on regenerative capacity. As cell transplantation and regeneration alter inflammatory cell populations, macrophage presence was assessed at 28 dpi by CD68 immunostaining, which marks myeloid cells (Chistiakov et al., 2017). Transplantation of control FAPs significantly reduced CD68+ cells in regeneration FID muscles, whereas transplantation of FID FAPs significantly increased the population of CD68+ cells when normalized to microscopic field (Figures 2O and 2R). Hence, the impaired regeneration following FID FAP transplantation may also be due to persistence of inflammatory cells. Collectively, these measures demonstrate an overall regeneration delay when IGF-I from FAPs is absent, with potentially direct and indirect effects. FAP proliferation requires an autocrine source of IGF-I FAPs express both Igf1 and IGF-I receptor (Igf1r) in developing, regenerating, and aging skeletal muscle (Kok and Barton, 2021), so they likely can respond to IGF-I regardless of its source. However, IGF-I from FAPs may be an important autocrine source. We first examined early phases of regeneration when FAPs proliferate (Figure 3A) (Joe et al., 2010). FID FAPs failed to expand compared to controls, showing fewer PDGFRα+ cells 3–7 dpi (Figures 3B, 3C, and S3A). Sections were immunostained for Ki67 and PDGFRα, revealing a ∼30% decrease in Ki67+ FID FAPs at 5 dpi (Figures 3D and 3E). Primary FAPs were then isolated from ND, 3 and 5 dpi muscles, and stained for Ki67 immediately after attachment (Figure 3F). Although similar total mononuclear cells were collected, fewer FAPs were recovered from FID muscles (Figure S3B). Igf1fl/fl FAP proliferation peaked at 3 dpi, yet FID Ki67+ FAPs were low throughout (Figures 3G and 3H). Next, primary FAPs isolated at 3 dpi were plated at similar density to monitor growth in vitro. By 1 week, FAPs isolated from Igf1fl/fl muscles expanded 7-fold, while FID FAPs only doubled (Figures 3I and 3J). To determine if the IGF-I pathway was intact, FAPs were treated with rIGF-I and/or the IGF-IR inhibitor AEW541. Akt phosphorylation in FID FAPs was increased by rIGF-I but ERK1/2 phosphorylation was not, even though there were no differences in total AKT or ERK1/2; further, FID FAP proliferation was not restored (Figures 3K–3N and S3C). As MAPK signaling is generally linked to mitogenic actions, this suggests that the block of FID FAP proliferation is due to impairments in the MAPK arm of the IGF-IR pathway. IGF-I from FAPs is required for FAPs proliferation (A) Muscle regeneration experimental design. (B) IF of PGDFRɑ, laminin, and DAPI on sections from 3–7 dpi. Scale bars: 100 μm. (C) Number of PDGFRɑ+ cells from 3–7 dpi muscles show significantly decreased FID FAPs. n = 5–6 per strain and time point. (D) IF of PDGFRɑ, Ki67, and DAPI on 5 dpi muscle sections. Scale bar: 200 μm. (E) Number of Ki67+ FAPs on 5 dpi muscle sections show decreased FAP proliferation. Unpaired t test: p = 0.0013. n = 4 per strain. (F) Primary cell isolation experimental design. (G) IF of Ki67 and DAPI on freshly isolated FAPs. Scale bars: 100 μm. (H) Number of proliferating FAPs on freshly isolated FAPS show transient increased in Igf1fl/fl, but not FID FAPs. n = 4–6 per strain per condition. (I) Phase contrast image of cultured primary FAPs. Scale bars: 200 μm. (J) Number of FAPs per field 0–6 days post-plating reveal diminished doubling for FID FAPs. n = 6 per strain. (K) Immunoblotting of IGFIRα, total and pAKT in FAPs exposed to rIGF-I, AEW541, or both for 30 min. GAPDH serves as loading control. (L) P-Akt activation shows IGF-IR pathway is intact in FID FAPs. n = 4 per condition and strain. (M) IF of Ki67 and DAPI of cultured FAPs. Cells were exposed to rIGF-I, AEW541, or both for 24 h. Scale bars: 200 μm. (N) Number of proliferating FAPs show FID FAPs do not respond to IGF-I. n = 4 per strain. (O) IF of Ki67 and DAPI of cultured FAPs following in vitro Cre induction by 4-OHTAM treatment with or without rIGF-I. Scale bars: 100 μm. (P) Number of proliferating primary FAPs show FID FAPs can respond to IGF-I early in deletion but not after 1 week (D11). n = 4 per strain and condition. To determine the emergence of the FID phenotype, we isolated primary FAPs from both strains and performed in vitro recombination using 4-hydroxytamoxifen (4OHTAM) for 4 days to achieve efficient deletion (Figure S7). Exogenous IGF-I was provided 24 h prior to termination at each time point. IGF-I restored cell proliferation at days 2 and 4; however, it failed to restore proliferation in FID FAPs 7 days after recombination (D11) (Figures 3O–3P). Taken together, FAP IGF-I is indispensable for proliferation, and prolonged loss of FAP IGF-I results in FAP cell proliferation impairments. FID FAP gene expression indicates low proliferation and ECM production and a heightened immune signature To expand our understanding of the impact IGF-I loss has on FAPs, we used RNA-seq of primary FAPs isolated from regenerating muscles at 3 dpi (Figures 4A and S4A). A total of 20,937 genes were identified, and with thresholds set to 2-fold changes with p < 0.05, we detected 427 upregulated and 179 downregulated genes in FID FAPs vs. Igf1fl/fl FAPs (Figure 4B). Gene Ontology (GO) analysis revealed that the highest scores for upregulated genes were in immune-related pathways, suggesting that FAPs from FID mice adopted an inflammatory phenotype (Figures 4C, 4D, and S4C). Due to increase in immunoregulatory-related genes, we examined sections from regenerating muscle for macrophage presence, expanding upon the analysis after transplantation (Figure 2R). CD68+ cells were evident in muscles at 7 dpi from both FID and Igf1fl/fl muscles; however, they persisted in FID muscles through 28 dpi in contrast to controls (Figures 4E and 4F). This finding suggests that the endogenous crosstalk between FAPs and macrophages to resolve damage was also disrupted with the loss of IGF-I from FAPs. RNA-seq of isolated FAPs (A) Schematic illustration of bulk RNA sequencing. (B) Volcano plot of all the genes that were captured in RNA-seq. Log2 Fold change of 1 and p value <0.05 were used as cutoffs. (C) Pathway analysis showing the greatest up- and downregulated pathways in FID FAPs. (D) Heatmaps of differential gene expression for immune response, ECM organization, cell proliferation, and adipogenesis. Z score ranges from −2 to 2. Downregulation showed in dark red, and upregulation showed in gray. n = 3 samples per strain. (E) CD68, Laminin, and DAPI IF on muscle sections following injury. Scale bars: 200 μm. (F) Number of CD68+ cells in muscle sections shows reduction of myeloid cells in control muscle but not in FID muscles 28 dpi. n = 4–6 per strain and time point. Out of the 179 genes that were downregulated, most of them clustered into ECM GO terms (Figure 4C), and heatmaps displayed a global shutdown of ECM gene expression in FID FAPs, including multiple collagen genes, fibronectin (Fn1), and laminin 2 (Lama2) (Figures 4D and S4B). Downregulation of Klf4 was also evident in FID FAPs, a key regulator in fibrogenesis and proliferation (Figure 4D) (Chandran et al., 2021). These changes were accompanied by reductions in many genes promoting proliferation and IGF-I signaling, consistent with our in vivo and in vitro measurements but without major differences in MAP-ERK signaling (Figures 4D and S4C). Because FAPs can adopt fibrogenic or adipogenic lineages (Joe et al., 2010; Uezumi et al., 2010), we examined genes regulating FAP fate. There were no consistent patterns in FAP regulatory or adipogenic genes in FID FAPs, indicating an absence of trans-differentiation (Figures 4G and S4C). Igf1 was also evaluated using primers P2 and P5, which was not statistically different between strains with this primer pair (Qval = 0.33). Overall, the FID FAP gene expression profile exhibited deficits in multiple pathways needed for efficient regeneration. Loss of FAP IGF-I suppresses ECM deposition and slows fibroblast and adipogenic differentiation During muscle regeneration, FAPs deposit ECM as a scaffold for fiber formation. When FAPs are depleted, ECM content is reduced (Kotsaris et al., 2023; Lemos et al., 2015). RNA-seq of FID FAPs prompted assessment of collagen deposition via Picro-Sirius red staining. The transient collagen increase seen in Igf1fl/fl muscles was absent in FID muscles (Figures 5A, 5B, and S5A), consistent with suppressed ECM gene expression (Figure 4D), though this could reflect reduced FAP numbers. To test fibroblast differentiation, primary FAPs were stimulated with TGF-β (5 ng/mL) at matched confluence (Figure 5C) (Contreras et al., 2019). Collagens (Col1a4, Col4a1) were similarly upregulated at 24 h, indicating intact differentiation, whereas fibronectin (Fn1) remained low in FID FAPs (Figure 5D) (Jones et al., 2019). Immunoblotting showed comparable Smad2 phosphorylation (Figures 5E and S5), but αSMA induction was delayed in FID FAPs and PDGFRα downregulation was slower, suggesting retention of the progenitor state (Figures 5E, 5F, and S5). No perilipin accumulation was observed, excluding adipogenic trans-differentiation (Figure 5E). IGF-I from FAPs is required for efficient fibrogenic and adipogenic differentiation (A) Picro-Sirius-red (PSR) staining of TA muscles from ND to 14 dpi. Scale bars: 200 μm. (B) Collagen content by % total area that is PSR positive shows the regenerating FID muscles have blunted accumulation of collagen. n = 5–8 per strain and time point. (C) Phase-contrast images of primary FAPs during fibrogenic differentiation. Scale bars: 100 μm. (D) RT-qPCR analysis of Col4a1, Col1a4, and Fn1 in FAPs from pre and 24 h post TGF-β exposure. n = 3 per strain and time point. (E) Immunoblotting analysis of protein isolated from primary FAPs following TGF-β treatment. Total protein quantification was obtained from ponceau staining. (F) αSMA and PDGFRα protein levels show delay in αSMA accumulation and persistence of PDGFRα in FID FAPs. n = 3 per strain and time point. (G) Perilipin and DAPI IF in muscle sections 7 and 21 dpi after glycerol injection. Scale bars: 200 μm. Perilipin+ area is an indication of adipocytes at 7 and 21 dpi. n = 6–10 per strain and time point. (H) Oil red O staining of primary FAPs 7 days after exposure to adipogenic media. Right column shows black and white 8-bit high contrast image for better visualization. ORO+ area at 7 dpi with FAPs from control, FID, and FID with supplemental IGF-I shows FID FAPs do not respond to IGF-I. n = 4 per condition. (I) RT-qPCR analysis of FAPs 24 h post-adipogenic induction. Adipoq p = 0.0013, Pparg p = 0.5175, Fn1 p = 0.032 by unpaired t tests. n = 3 per strain. (J) Immunoblotting of primary FAPs at 3 days post-adipogenic induction for PDGFRα, perilipin, and αSMA. Ponceau was used a total protein loading control. (K) Perilipin, PDGFRα, and αSMA protein levels after adipogenic induction. Perilipin p = 0.0002, PDGFRα p = 0.0824, αSMA p = 0.0868 by unpaired t tests. n = 6 per strain. As FAPs are primary also adipocyte progenitors (Kopinke et al., 2017; Uezumi et al., 2010), we tested adipogenic differentiation using 50% glycerol injections in female mice, which have greater adipogenic potential (Kopinke et al., 2017). By 7 dpi, perilipin+ adipocytes appeared, and by 21 dpi, fat accumulation was higher in Igf1fl/fl than FID muscles (Figure 5G). Minimal differences in muscle mass were observed (Figures S5A and S5B), suggesting loss of FAP-derived IGF-I suppresses adipocyte accumulation without affecting myogenic regeneration. To examine adipogenesis further, primary FAPs from both strains were cultured to 95%–100% confluency and treated with adipogenic induction media (0.25 μM dexamethasone, 0.5 mM IBMX, 5 μg/mL insulin, 5 μM troglitazone). Lipid droplets accumulated in Igf1fl/fl FAPs by 1 day post-induction, whereas in FID FAPs this occurred at 5 days (Figures S5D and S5E). GFP lineage tracing confirmed that differentiation was delayed in all FID GFP+ FAPs. By 7 days, ∼70% of Igf1fl/fl FAPs were oil red O+, while far fewer FID cells stained positive (Figure 5H). Although IGF-I can promote adipogenesis in mesenchymal stem cells (Scavo et al., 2004; Smith et al., 1988), exogenous 10 nM IGF-I failed to rescue differentiation in FID FAPs (Figure 5H). Evaluation of key markers of adipogenesis was performed early in induction. Delays were evident, with lower expression of adiponectin (Adipoq) and Pparg at 3 dpi in FID FAPs while the fibrogenic marker Fn1 remained unchanged (Figure 5I). Immunoblotting at 3 dpi showed robust perilipin levels in Igf1fl/fl FAPs but not in FID FAPs, with trends of higher FID PDGFRα and comparable decreases in αSMA (Figures 5J and 5K). These findings indicate that FID FAPs need autocrine IGF-I for efficient adipogenesis and fibrogenesis. Single cell expression analysis reveals loss of fibrotic FAPs and adoption of a senescent-like phenotype in FID FAPs To examine the molecular signature of homeostatic FAPs without IGF-I, scRNA-seq was performed on TA muscles 21 days post-TAM induction, enriching for mononuclear cell types by eliminating muscle fiber nuclei. Approximately 20,000 cells per strain were assessed (∼40,000 reads/cell). Uniform manifold approximation and projection (UMAP) clustering identified 9 cell populations based on unique markers (Figures S7A–S7C) (Oprescu et al., 2020b). FID samples had reduced proportions of FAPs, fibroblasts, monocytes, and a modest decrease in SCs; the proportional increase in endothelial cells was likely the result of the diminished populations other cell types (Figure S7D). FAPs and fibroblasts had the most pronounced Igf1 expression in the cells, and loss of FAP-derived IGF-I did not induce compensatory Igf1 expression in other cell types (Figures S7B and S7C). FAP sub-clustering revealed four populations: fibrotic FAPs (high collagen), progenitor FAPs (Dpp IV+), activated FAPs (Sox6+), and microenvironmental FAPs (critical ECM genes) (Figure 6A, S7F-G) (Babaeijandaghi et al., 2023; Krebsbach and Villa-Diaz, 2017). Fibrotic FAPs were most affected in number and gene expression, consistent with global collagen reduction (Figures 6A, 4C, and 4D). Top 20 DE genes showed elevated Thbs1 and p53, which promote cell-cycle arrest and senescence (Figures 6B and 6C) (Lawler et al., 2001; Sundaram et al., 2011). Downregulated genes included proliferation and ribosomal genes (Cdk8, Rps28, Rps29, and Rpl29) (Figure 6C) (Kim et al., 2019; Kirn-Safran et al., 2007; Szilagyi and Gustafsson, 2013). GO and pathway analysis highlighted ECM organization, p53 regulation, and protein translation as the most affected processes (Figures 6D and 6E). Single-cell gene expression analysis of FAPs (A) Overlay of FAPs from FID and Igf1fl/fl muscle reveal a separation in the profiles of fibrotic FAPs from FID samples away from those from Igf1fl/fl muscle. Identification of four FAP sub-populations: (1) fibrotic FAPs with the highest collagen gene expression, (2) progenitor FAPs expressing Dpp IV, (3) activated FAPs with high Sox6 expression, and (4) microenvironmental FAPs with ECM gene expression. There are ∼30% fewer FAPs in FID muscles, with active and fibrotic FAPs being the most affected. TA muscles from n = 5 mice pooled per strain. (B) Heatmap comparison of the top 20 up- and downregulated genes in FID FAPs. (C) UMAPs for representative changed genes. Thbs1 is most pronounced in the fibrotic FAPs of FID muscles. Protein synthesis genes (Rpl39 and Rps29) are distributed throughout the cluster. High expression of Trp53 is also distributed throughout the cluster. (D) Biological process analysis shows decreased protein translation, increased cellular senescence, and decreased ECM organization. (E) KEGG analysis reveals the FAPs from FID muscles have suppression of translation, ECM organization, RNA metabolism and IGF-I pathways, and enhancement of p53 and cellular senescence. These results suggest that loss of FAP-derived IGF-I shifts FAPs toward an inactive, potentially senescent state. FID FAPs displayed transcriptional profiles resembling some aspects of senescent FAPs (Moiseeva et al., 2023) with elevated inflammatory signaling and reduced proliferation markers (Figures 4 and 6). In homeostatic muscle, 22% of FID FAPs were PDGFRα+/β-Gal+ versus 5% in Igf1fl/fl FAPs (Figures 7A and 7B). Following BaCl2 injury, >60% of FAPs lacking IGF-I were β-Gal+ at 7 dpi versus 20% in controls, coinciding with minimal FAP expansion. By 28 dpi, total FAP numbers normalized, but β-Gal+ proportions in FID muscles remained >60% (Figures 7A and 7B). Cultured FID FAPs also had elevated β-Gal but only modestly increased p53 (Figures 7C, 7D, and S7B) (Ben-Porath and Weinberg, 2004; Moiseeva et al., 2023; Saito et al., 2020). In complementary experiments, primary FAPs were examined for β-Gal during in vitro 4-OH-TAM treatment (Figure S7A). By day 11, FID FAPs were ∼80% β-Gal+ (Figures 7E and 7F). Continuous IGF-I treatment during the 11-day period normalized β-Gal+ staining, but acute 24 Hr IGF-I was ineffective (Figures 7E and 7F). Because FID FAPs have lower proliferation than Igf1fl/fl FAPs (Figures 3M–3P), this may have affected which cells were present in the endpoint and their β-Gal staining, but tracing of the cell populations was not pursued. Collectively, FAPs from FID muscles exhibit senescent-like features both in vitro and in vivo under homeostatic conditions, which may underly the lack of FID FAP participation in muscle regeneration. IGF-I ablation induces senescence phenotype in FAPs (A) PDGFRα, β-galactosidase (β-Gal), and DAPI IF in muscle sections from ND, 7 and 28 dpi. Scale bars: 100 μm. (B)β-Gal+ FAPs serve as an index of cellular senescence with a heightened and persistent senescence phenotype in FID FAPs. n = 6 per strain and time point. (C) Brightfield image of plated primary FAPs after β-Gal staining. Scale bars: 200 μm. (D) β-Gal+ FAPs are more prominent in FID FAPs. Unpaired t test: p < 0.0001. n = 4 for Igf1fl/fl, n = 5 for FID. (E) Brightfield images of β-gal stained FAPs during in vitro 4-OHTAM recombination, with lower panels displaying effects of 24 hr or continuous IGF-I treatment. Scale bars: 200 μm. (F) β-Gal+ FAPs show progressive increase of β-gal staining in FID FAPs that is reduced only with continuous IGF-I treatment. n = 3–4 per strain and time point. Discussion Research in the past decade has established that FAPs play a critical role in muscle homeostasis, regeneration, aging, and even neuromuscular junction integrity (Joe et al., 2010; Kim et al., 2022; Lukjanenko et al., 2019; Uezumi et al., 2021; Wosczyna et al., 2019). FAPs also secrete IGF-I, which is a well-documented factor contributing to skeletal muscle biology (Barton-Davis et al., 1998; Kok and Barton, 2021; Musaro et al., 2001; Vassilakos et al., 2019). Both the sources and timing of IGF-I production are central to skeletal muscle growth and regeneration. This motivated the central questions of this study, to determine if FAP IGF-I was needed for efficient muscle regeneration in general, and if FAP actions were specifically affected. Our results support that IGF-I from FAPs plays a direct autocrine role for FAPs, removing them from effectively participating in the regeneration process, mirroring a phenotype associated with the complete ablation of FAPs (Wosczyna et al., 2019). Several other pathways have been identified that alter FAP fate. For example, Odd-skipped related transcription factor 1 (Osr1) is an important regulator of FAP activity during muscle regeneration. Loss of Osr1 impedes FAP proliferation, enhances FAP apoptosis, impairs immune interactions, and upregulates ECM-related genes (Kotsaris et al., 2023). We have delineated a contrasting phenotype with reduced ECM production and enhanced immune pathway activity, with modest downregulation in Osr1 expression. This suggests an independent pathway, potentially acting in opposition to Osr1-regulated mechanisms. Upon activation following acute injury, FAPs develop primary cilia, a crucial event for FAP function in muscle tissue, as cilia-deficient FAPs exhibit resistance to adipogenic differentiation, in part, through the disruption of Hedgehog signaling (Kopinke et al., 2017). The expression of two primary cilia formation genes, Arl13b and Ift88, remained unaffected in FID FAPs, suggesting that IGF-I does not alter primary cilia. IGF1R is localized on the primary cilia of mesenchymal stem cells (Dalbay et al., 2015), and it remains unclear if signaling crosstalk between IGF-I and other pathways is disrupted, such as Hedgehog signaling, enriched in primary cilia. We observed reduced regenerative capacity in FID muscles unlike enhanced myogenesis found in the absence of primary cilia (Kopinke et al., 2017), suggesting that primary cilia formation is intact, or that the impact of IGF-I loss from FAPs overrides the potential benefit of primary cilia loss. The senescent-like hallmarks of FID FAPs in vivo and in vitro were surprising. While a transient increase of senescent cells has been documented in muscle regeneration, increases in senescent cell abundance are apparent in aging and diseased muscles, and impair efficient regeneration (Moiseeva et al., 2023). Senescence can occur in any of the cell types involved in regeneration; however, in our model, few, if any, cells other than FAPs displayed these hallmarks. Delayed muscle regeneration by senescent cells is caused in part by their senescence-associated secretory phenotype (SASP) (Basisty et al., 2020), a host of secreted factors that are pro-inflammatory and pro-fibrotic. FID FAPs displayed some features of senescent cells, including reduction in genes for cell proliferation and enhancement of genes for inflammation (Moiseeva et al., 2023), but they lacked upregulation of ECM related genes. The reduction of ECM remodeling and deposition may prevent normal fiber formation during regeneration, as physical and biochemical cues from the matrix would be lost. Alternatively, this may be an advantage in instances of excess fibrosis, as the “quasi”-senescent FAPs could not produce increased ECM. Understanding the balance between the beneficial and detrimental consequences of FAPs lacking autocrine IGF-I actions is likely context dependent, and worthy of future investigation. A larger question is how consequences to muscle homeostasis and regeneration differ between the FID model, where FAPs lack IGF-I production, and the complete ablation of FAPs as previously studied (Uezumi et al., 2021; Wosczyna et al., 2019). The models share regeneration defects, with loss of FAPs having slower regenerative capacity and more negative impact on muscle SC than loss of FAP IGF-I. Our examination of regenerating muscle spanned the 4 week period in which FAPs are known to contribute to resolving damage, but we examined FAPs most thoroughly at the height of expansion, at 3 dpi, which may not reflect their properties throughout regeneration. The comparison of transplantation of healthy and FID FAPs reveal differential effects on their abilities to resolve damage in FID muscles, but compensation by other circulating factors may have also helped. FAPs isolated from FID muscles were still able to respond to fibrogenic and to adipogenic media, albeit more slowly, suggesting that the loss of IGF-I does not block all cytokine mediated communication to the FAPs. The contribution of other pathways may partially rescue IGF-I dependent deficits leading to a milder phenotype than when FAPs are completely absent. In the current study, we focused on acute muscle regeneration in healthy young adult mice to afford a controlled environment for evaluating the contribution of IGF-I from FAPs on regenerative capacity. During the early stages of muscle regeneration, FAPs interact with SCs to promote myogenesis (Joe et al., 2010). Previous work has shown that FAPs secrete WNT1-inducible signaling pathway protein 1 (WISP1), which regulates SC proliferation by activating the Akt signaling pathway (Lukjanenko et al., 2019). Additionally, FAPs produce HGF that may activate quiescent SCs and/or regulate their basal proliferation (Allen et al., 1995; Kim et al., 2022). Although our RNA-seq data indicated that expression levels of Wisp1 and Hgf were unchanged in the absence of IGF-I, we did not examine the FAP secretome in this study to determine if release of these proteins was affected by the loss of FAP IGF-I. Alternatively, the reduction in SC number could theoretically result solely from a decline in FAP numbers, as FAP ablation also results in reduced SCs in regenerating muscle (Uezumi et al., 2021; Wosczyna et al., 2019). Although Pax7+ SCs are an important part of muscle development and regeneration (Kuang et al., 2007; Relaix et al., 2005), it also requires supporting FAPs for them to thrive in their environment (Joe et al., 2010; Kim et al., 2022; Lukjanenko et al., 2019). If SCs acquire deficiencies in their ability to proliferate or differentiate, the consequences will be compounded by the inability of IGF-I deficient FAPs to promote their activation via factors such as HGF (Kim et al., 2022). Similarly, a heightened inflammatory environment may go unchecked without the ability for FAPs to interact effectively with the macrophage population. However, the actions of polarized macrophages on FAPs to enhance fatty-fibrotic infiltration in diseased muscles (Liu et al., 2023) may also be disrupted by loss of IGF-I from FAPs. Future studies are warranted to determine the potential benefits or disadvantages of removing this source of IGF-I from skeletal muscle. Understanding how the interplay among these cell populations changes when FAPs are in a state of torpor will be important to pursue in light of the reduced ECM deposition in FID muscles, which may be beneficial in disease, but at the expense of reduced regenerative capacity, which may be detrimental. The tissue environment associated with genetic neuromuscular disease or aging may exacerbate the phenotypes we have observed in regenerating FID muscles. Overall, this study provides multiple lines of evidence that IGF-I produced by FAPs in muscle is a required autocrine factor for FAP actions during muscle regeneration. The consequences of its loss alter FAPs to an extent that they are removed from performing their normal interactions with other cell populations in the muscle, ultimately stalling regenerative capacity. While IGF-I from FAPs may also have paracrine actions on neighboring cells, these potential direct actions cannot be resolved from those mediated indirectly via its required beneficial actions on FAPs themselves. Resource availability Lead contact Requests for further information and resources should be directed to and will be fulfilled by the lead contact, Elisabeth R. Barton (erbarton@ufl.edu). Materials availability Mouse lines were generated from genetic crosses of commercially available lines. The specific cross generated in this study are available from the lead contact with a completed materials transfer agreement. Data and code availability •Bulk RNA sequencing data have been deposited in GEO: GSE282309. Bulk RNA sequencing data have been deposited in GEO: GSE282309. •Single-cell RNA sequencing data have been deposited in GEO: GSE332757. Single-cell RNA sequencing data have been deposited in GEO: GSE332757. Acknowledgments This work was supported by the Spinal Muscular Atrophy Foundation to E.R.B., and by US National Institute of Health (NIH) P50 AR052646 to E.R.B. supporting the Physiological Assessment Core. D.K. is supported by NIH R01 AR079449. Y.E.L. is supported by the UF College of Health and Human Performance Graduate Fellowship. T.Y.A. was supported by the UF University Scholars Program. The authors acknowledge University of Florida Research Computing for providing computational resources and support that have contributed to the research results reported in this publication. Author contributions Y.E.L., D.K., F. Yue, and E.R.B. conceived and designed the study. Y.E.L., Z.A.-T., T.Y.A., L.-Y.K., F. Yue, F. Yu, Y.Z., B.E.S., A.B.A., and E.M.D. performed all experiments. All authors analyzed the data. Y.E.L. and E.R.B. wrote the manuscript. Declaration of interests The authors have stated explicitly that there are no conflicts of interest in connection with this article. STAR★Methods Key resources table REAGENT or RESOURCESOURCEIDENTIFIERAntibodiesRabbit Anti-LamininInvitrogenCat# MA1-06100; RRID: AB_559896Goat Anti-PDGFRaR&D systemCat# AF1062; RRID: AB_2236897Mouse Anti-Alpha-Smooth Muscle ActinR&D systemCat# MAB1420; RRID: AB_262054Rat Anti-CD68BioLegendCat# 137001; RRID: AB_2044003Mouse Anti-Pax7(IgGγ1)DSHBCat# PAX7-S; RRID: AB_2299243Rabbit Anti-Perilipin-1(D1D8) Cell Signaling 9349Cell SignalingCat# 9349; RRID: AB_10829911Rabbit Phospho-Smad2Cell SignalingCat# 8828; RRID: AB_2631089Rabbit Anti-Ki67AbcamCat# Ab16667; RRID: AB_302459Mouse Anti-P53Santa CruzCat# Sc-126; RRID: AB_628082Mouse Anti-IGF1RαSanta CruzCat# Sc-271606; RRID: AB_10650003Mouse Anti-Erk1/2Cell SignalingCat#4696s; RRID: AB_390780Rabbit Anti-p-Erk1/2Cell SignalingCat#9101s; RRID: AB_331646Mouse Anti-AktCell SignalingCat#2920s; RRID: AB_1147620Rabbit Anti-p-AktCell SignalingCat#9271s; RRID: AB_329825Mouse Anti-GAPDHSanta CruzCat#Sc-32233; RRID: AB_627679Mouse Anti-MyoD(IgG2b)Santa CruzCat#Sc-377460; RRID: AB_2813894DAPIThermo FisherCat#D1306; RRID: AB_2629482555 Donkey Anti-GoatThermo FisherCat#A-32816; RRID: AB_2762839488 Donkey Anti-RabbitThermo FisherCat#A-21206; RRID: AB_2535792488 Goat Anti-RatThermo FisherCat#A-11006; RRID: AB_2534074488 Goat Anti-Mouse IgG2bThermo FisherCat#A-21141; RRID: AB_141626594 Goat Anti-Mouse IgG1Thermo FisherCat#A-21125; RRID: AB_141593IRDye 800CW Donkey Anti-GoatLi-CorCat#926-32214: RRID: AB_621846RDye 680 Goat Anti-MouseLi-CorCat#926-68070: RRID: AB_10956588IRDye 680 Goat Anti-RabbitLi-CorCat#926-68071; RRID: AB_10956166IRDye 800 Goat Anti-RabbitLi-CorCat#926-32211; RRID: AB_621843Chemicals, peptides, and recombinant proteinsG2 Master MixPromegaCat#M7882TamoxifenSigmaCat#T5648-5GGlycerolSigmaCat#G5516-500MLBarium ChlorideSigmaCat#342920-50GTissue-Tek O.C.T. CompoundSakuraCat#45832-methylbutaneSigmaCat#277258-1LTrizol ReagentThermo FisherCat#15596026DNase IThermo FisherCat#18047019High Capacity cDNA Reverse Transcription KitThermo FisherCat#4368814SYBR Green Master Mix StandardThermo FisherCat#A46109RNeasy Micro KitQiagenCat#74004Fluoromount-GSouthern BiotechCat#0100-01Pierce BCA Protein AssayThermo FisherCat#23227NuPAGE Bis-Tris Mini GelThermo FisherCat#10780345Intercept Blocking BufferLicorCat#927-60001Debris Removal SolutionMiltenyi BiotecCat#130-109-398PDGFRa Microbead KitsMiltenyi BiotecCat#130-101-502Satellite Cell isolation Kit, MouseMiltenyi BiotecCat#130-104-268Matrigel Basement Membrane MatrixCorningCat#354230Human InsulinSigmaCat#12643-25MGTroglitazoneSigmaCat#T2573DexamethasoneSigmaCat#D4902-25MG3-Isobutyl-1-methylxanthineSigmaCat#I5879-100MGTGF-β1PeproTechCat#AF-100-21C-2UG4-HydroxytamoxifenSigmaCat#ML1666Critical commercial assaysMouse/Rat IGF-I/IGF-1 Quantikine ELISA KitR&DCat# MG100Senescence β-Galactosidase Staining KitCell SignalingCat#9860SDeposited dataBulk RNA Seq raw data on primary FAPs isolated 3dpiThis paperGEO: GSE282309Single Cell Analysis on cells isolated from ND muscleThis paperGEO: GSE332757Experimental models: Organisms/strainsMouse: B6.129S-Pdgfratm1.1(cre/ERT2)Blh/JThe Jackson LaboratoryRRID:IMSR_JAX:032770Mouse: FVB.129(B6)-Igf1tm1Dlr/JThe Jackson LaboratoryRRID:IMSR_JAX:012663Mouse: B6.129X1-Gt(ROSA)26Sortm1(EYFP)Cos/JThe Jackson LaboratoryRRID:IMSR_JAX:006148Mouse: Gt(ROSA)26Sortm13(CAG-HRP)Luo/JThe Jackson LaboratoryRRID:IMSR_JAX:037697OligonucleotidesPrimersSee Table S1 for sequences–Software and algorithmsImageJSchneider et al., 2012https://imagej.nih.gov/ij/SMASHSmith and Barton, 2014https://figshare.com/articles/dataset/SMASH_semi_automatic_muscle_analysis_using_segmentation_of_histology_a_MATLAB_application/1247634STAR v2.7.9aDobin et al., 2013https://hbctraining.github.io/Intro-to-rnaseq-hpc-O2/lessons/03_alignment.htmlCell Ranger10X Genomicshttps://www.10xgenomics.com/support/software/cell-ranger/latestScanpyWolf et al., 2018https://scanpy.scverse.org/en/stable/ Experimental model and study participant details Mouse models All animal procedures followed protocols approved by the University of Florida IACUC. Mice were on a C57BL/6 background under a 12-h light/dark cycle with ad libitum access to food and water. FAP-Igf1-Deficient (FID) mice were made by crossing Pdgfra Cre-ERT2 mice (Chung et al., 2018) (JAX#032770) with Igf1fl/fl mice (Liu et al., 1998; Spradlin et al., 2021) (JAX#012663) and Rosa26EYFP mice (Srinivas et al., 2001) (JAX#006148). FID mice were homozygous for Igf1fl/fl and Rosa26EYFP, and hemizygous for PdgfraCre-ERT2; littermates lacking the Cre allele served as controls (Kang et al., 2010; Kopinke et al., 2017). FAP-HRP mice were generated by crossing PdgfraCre-ERT2 mice with Rosa26HRP-iPEEL mice (Luo et al., 2026; Shuster et al., 2022). Genotyping was performed with tail DNA and G2 Master Mix (Promega) followed by agarose gel electrophoresis. Cre recombination was induced by IP injections of tamoxifen (20 mg/mL in corn oil; 100 μg/g body weight) for 5 days in ∼12-week-old mice and both sexes are included in the majority of the experiments. Only females were used in glycerol injury since female exhibits greater adipogenic response (Norris et al., 2023; Norris et al., 2024, 2025). Experiments were performed ≥1 week after final injection. Method details Muscle injury and tissue processing Acute muscle injury was induced by injecting 50 μL PBS containing 1.2% BaCl2 into the tibialis anterior (TA) muscle. Adipogenic injury was induced by 50 μL of 50% glycerol in PBS; only females were used due to their higher adipogenic response (Norris et al., 2024; Pisani et al., 2010). Tissues were collected 3–28 days post-injury (dpi) and weighed before processing. Muscles for histology were embedded in OCT and frozen in 2-methylbutane cooled by liquid nitrogen. Samples for biochemical analysis were frozen in liquid nitrogen and stored at −80°C. Primary cell culture Primary FAPs were isolated from TA muscles as previously described with minor modifications (Oprescu et al., 2020a). Dissected TA muscles were minced and digested with collagenase II and dispase for 1 h at 37°C on a rotator. Cell suspensions were cleared using Debris Removal Solution (Miltenyi Biotec) and counted (Countess 3). FAPs were purified using the PDGFRα Microbead Kit and MACS system (Miltenyi Biotec) per manufacturer’s protocol. Cells were plated at 30–35% confluency in proliferation medium (DMEM, 10% bovine calf serum, 1% penicillin/streptomycin, and 1% GlutaMAX), changed every 2–3 days. Primary FAPs were plated at 30–35% confluency in proliferation medium (DMEM, 10% bovine calf serum, 1% penicillin/streptomycin, and 1% GlutaMAX), changed every 2–3 days. Adipogenic differentiation was induced at 95–100% confluency with 5 μg/mL insulin or 10 nM IGF-I, 5 μM troglitazone, IBMX, and dexamethasone in media (Low et al., 2017). Fibrogenic differentiation was induced at 65–70% confluency with 5 ng/mL TGF-β(Contreras et al., 2019). Differentiation media were refreshed daily. In vitro recombination was induced by 2 μM 4-hydroxytamoxifen for 4 days, with ethanol vehicle as control. Freshly isolated FAPs or SCs (Miltenyi Satellite Cell Isolation Kit) were also plated on 24-well or Matrigel-coated coverslips, allowed to adhere 4–6 h, then fixed for immunostaining. FAP transplantation Transplantations were performed in FID mice following bilateral BaCl2-induced injury to TA muscles. At 2 dpi, 75,000 freshly isolated control FAPs (HRP+, from FAP-HRP mice) or FID FAPs (GFP+, from FID mice) were resuspended in 50 μL PBS and injected into the left TA. The right TA received PBS only. At 28 dpi, TAs were collected, weighed, and processed for histology. RNA-seq analysis Total RNA was extracted from primary FAPs isolated from 3 dpi muscles. Sequencing reads generated by the Illumina NovaSeq 6000 were trimmed with Cutadapt (Martin, 2011) to remove adapters and bases with quality scores <20; reads <40 bp were excluded. Clean reads were aligned to the Mus musculus GRCm38 (mm10) reference genome using STAR v2.7.9a (Dobin et al., 2013). Mapped reads were processed with HTSeq v0.11.2 (Anders et al., 2015) samtools, and in-house scripts (UF ICBR) to remove PCR duplicates and count unique reads. Gene expression was analyzed using a DESeq2-based R pipeline (Love et al., 2014). Differentially expressed genes were defined by |log2FC| > 1 and adjusted p < 0.05 Volcano plots were generated in R (v4.1.3), and Bowtie2 v2.5.4 was used for gene and exon searches (Langmead and Salzberg, 2012). Single Cell sample preparation and sequencing Tissue dissociation followed a published protocol (Oprescu et al., 2020a). Ten TA muscles from five male mice per strain were minced in wash medium (F-10 with 10% FBS and 1% penicillin–streptomycin) and digested for 1 h at 37°C in wash medium containing Collagenase II, Dispase II, and 1% BSA. The suspension was filtered twice through 40 μm strainers, treated with RBC lysis buffer for 3 min, and stained with Propidium Iodide (Invitrogen) to exclude dead cells. Viable cells were sorted by FACS (UF ICBR Core), yielding ∼200,000 cells per sample. Single-cell libraries were prepared using the 10x Genomics Chromium GEM-X Single Cell 3′ v4 platform with a target recovery of 20,000 cells and sequenced on an Illumina NovaSeq X. Reads were aligned and annotated using Cell Ranger v9.0.0 on HiPerGator (UF Research Computing), yielding ∼40,000 reads per cell and ∼15,000 retained cells per sample. Downstream analyses were performed in Scanpy (Wolf et al., 2018). Gene expression matrices from control and FID samples were merged for joint analysis. Cells with <200 genes, >6,000 genes, or >5% mitochondrial content, and genes expressed in <3 cells, were excluded. Predicted doublets were removed using Scrublet (Wolock et al., 2019). Counts were normalized to 10,000 per cell, and batch effects were corrected with Harmony (Korsunsky et al., 2019). Dimensionality reduction was performed using PCA with 50 neighbors for downstream analyses. DGE, cell clustering, pathway analysis, and cell-cell communication Cell clusters were identified using differential gene expression (DEG) with the Wilcoxon rank-sum test (p < 0.0001) and annotated based on marker genes using Scanpy. FAP clusters were extracted, normalized to 10,000 counts per cell, and re-clustered using DEG analysis (Wilcoxon, p < 0.0001). Functional enrichment within FAPs was performed using Gene Set Enrichment Analysis (GSEA) with GO m5 terms, and canonical pathway analysis was conducted using m2 pathways (version 2024.1) (Smith and Eppig, 2009). Real-time PCR Frozen tissues were pulverized, and total RNA was extracted using Trizol (Thermo Fisher). RNA from primary FAPs was isolated with the Qiagen RNeasy Micro Kit, per manufacturer protocol. RNA was DNase-treated, quantified, and checked for integrity. Equal amounts were reverse-transcribed with the High-Capacity cDNA Kit (Thermo Fisher). Gene expression was measured in duplicate with QuantStudio-3 and SYBR Green, normalized to Gapdh and/or Sra-1. Igf1 deletion was assessed as described (Spradlin et al., 2021) by comparing qPCR amplification of exon3–4 (deleted) to exon6–UTR (unaffected). Primer sequences are listed in Table S1 and illustrated in Figure 1. Histological staining and analysis Muscle cryosections were fixed in 4% paraformaldehyde for 10 min at RT and washed in PBS. Sections were blocked 1 h in 5% BSA, 0.02% Triton X-100, 0.03% sodium azide, then incubated overnight at 4°C with primary antibodies (Key Resources). After washing (3×PBS), secondary antibodies with DAPI were applied for 1 h at RT, followed by PBS-T (0.5% Tween 20) washes. Slides were air-dried, mounted with Fluoromount-G, and sealed. Cultured cells were treated similarly. Picro-Sirius Red staining assessed collagen (Smith et al., 2020). Cross-sectional areas were quantified using SMASH and CellPose (Smith and Barton, 2014; Waisman et al., 2021). PDGFRα+, CD68+ cells, and adipocytes were quantified in ImageJ. Images were acquired using Keyence X-BZ and EVOS imaging systems. Immunoblotting Proteins from tissues and cells were extracted using RIPA buffer (150 mM NaCl, 50 mM Tris-HCl, 1% Triton X-100, 0.5% sodium deoxycholate, 0.1% SDS). Equal amounts of protein (determined with Pierce BCA Protein Assay (Thermo Fisher)) were mixed with loading buffer containing 2-mercaptoethanol, denatured at 95°C for 5 min, and loaded onto 4–12% Bis-Tris gels (Invitrogen). Proteins were transferred to nitrocellulose or PVDF membranes at 60 V for 2.5 h. Membranes were stained with Ponceau S, washed in TBST, and blocked (Licor) for 1 h at RT. They were incubated with primary antibodies (Key Resources) overnight at 4°C, followed by TBST washes and incubation with secondary antibodies at RT. After final washes, images were captured on Li-Cor Odyssey CLX. Band intensities were quantified using Image Studio. IGF-I ELISA IGF-I levels in TA muscles were measured using the Quantikine ELISA kit (R&D Systems) per manufacturer’s instructions. Skeletal muscle lysates were diluted 1:3, and optical density was read at 450 and 570 nm (SpectraMax M5, Molecular Devices). Wavelength correction was performed by subtracting 570 nm from 450 nm readings. IGF-I levels were calculated from duplicates using a 4-parameter standard curve and normalized to total protein measured by Pierce BCA Protein Assay (Thermo Fisher). Results are expressed as pg IGF-I/μg total protein. Quantification and statistical analysis Comparisons were performed by unpaired Student’s t test, one-way and two-way ANOVAs, following by post hoc testing where appropriate. P-values <0.05 were considered significant. Published: June 25, 2026 Footnotes Supplemental information can be found online at https://doi.org/10.1016/j.stemcr.2026.102973. Supplemental information Document S1. Figures S1–S7 and Table S1 Document S2. Article plus supplemental information References Associated Data Supplementary Materials Document S1. Figures S1–S7 and Table S1 Document S2. Article plus supplemental information Data Availability Statement •Bulk RNA sequencing data have been deposited in GEO: GSE282309. Bulk RNA sequencing data have been deposited in GEO: GSE282309. •Single-cell RNA sequencing data have been deposited in GEO: GSE332757. Single-cell RNA sequencing data have been deposited in GEO: GSE332757.