Intramuscular adipose tissue restricts functional muscle recovery

SUMMARY

With age and disease, skeletal muscle is progressively lost and replaced by fibrotic scar and intramuscular adipose tissue (IMAT). While strongly correlated, it remains unclear whether IMAT has a functional impact on muscle. In the present study, we evaluated the impact of IMAT on muscle regeneration by creating a mouse model where the cellular origin of IMAT, fibro/adipogenic progenitors (FAPs), is prevented from differentiating into adipocytes (mFATBLOCK model). We found that blocking IMAT after an adipogenic injury allowed muscle to regenerate more efficiently, resulting in enhanced functional recovery. Our data explain why acute muscle injuries featuring IMAT infiltration, such as rotator cuff tears and acute denervation injuries, exhibit poor regeneration and lead to a loss of muscle function. It also demonstrates the therapeutic importance of preventing IMAT formation in acute injuries in order to maximize regeneration and minimize loss in muscle mass and function.

In brief

Norris et al. demonstrate the detrimental role of intramuscular adipose tissue (IMAT) on muscle regeneration. Following injury, IMAT inhibition enhances muscle regeneration by increasing myofiber density and size. Mechanistically, IMAT restricts the formation and subsequent hypertrophic growth of nascent myofibers, highlighting IMAT as an important therapeutic target.

Graphical abstract

INTRODUCTION

Adipose tissue acts as an energy storage as well as an endocrine organ.¹ However, different fat depots, such as subcutaneous (SAT), visceral (VAT), and intramuscular adipose tissue (IMAT), have stark metabolic and phenotypic differences.^2–4 IMAT, the accumulation of adipocytes between individual myofibers within skeletal muscle, is a pathological hallmark of muscular dystrophies,^5–8 but it is also present in a spectrum of metabolic disorders, including diabetes, obesity, and sarcopenia.^9–14 The progressive infiltration of IMAT within muscle tissue has been closely associated with loss of muscle mass, metabolic dysfunction, disease progression, and impairment of patient mobility.^15–17 In addition to being a feature of chronic diseases, IMAT also arises in response to acute insults, such as spinal cord and rotator cuff injuries, where muscle tissue is also simultaneously lost.^18–20 Despite the strong correlation between IMAT and loss of muscle tissue, its functional impact is still not fully understood.

The cellular origin of IMAT is a population of stem cells located in the muscle interstitium, called fibro-adipogenic progenitors (FAPs).^5,21–25 In a healthy muscle, FAPs are critical in maintaining muscle mass during homeostasis and playing a central role in muscle regeneration.^{21,22,24–30} FAPs secrete pro-myogenic factors to aid the cellular origin of muscle fibers, muscle stem cells (MuSCs), in their differentiation process toward myofibers. With age and disease, however, FAPs can also differentiate into either adipocytes, leading to IMAT formation or myofibroblasts, giving rise to fibrosis.^{5,21–23,25,30–33} While it is not completely understood what drives FAPs to differentiate into IMAT in certain conditions, the downstream adipogenic signaling cascade is governed by Pparγ, the master regulator of adipogenesis.^34,35 Pparγ is a transcription factor that, when active, initiates the signaling cascade required for the commitment and the subsequent differentiation of adipogenic progenitors toward the adipogenic lineage.³⁶

To understand the influence of IMAT on skeletal muscle, we created a conditional mouse model, termed mFATBLOCK that blocked IMAT formation by deleting peroxisome proliferator-activated receptor gamma (Pparγ) from FAPs. This deletion had no effect under normal conditions but successfully prevented IMAT accumulation after an adipogenic injury. Mechanistically, our data argue that IMAT acts as a physical barrier and prevents new nascent myofiber formation during early regeneration, as well as myofiber hypertrophy during the later regenerative phase. Consequently, this results in a functionally weakened muscle that has both fewer and smaller myofibers.

RESULTS

Removal of Pparγ from FAPs has no detectable impact on muscle or overall health during homeostasis in the absence of injury

To determine the functional impact of IMAT on muscle, we created a mouse model where FAPs are prevented from differentiating into adipocytes. For this, we crossed the FAP-specific tamoxifen-inducible Pdgfra ^CreERT2 allele³⁷ to a conditional allele of the adipogenic regulator Pparγ³⁸ (Pdgfra^CreERT Pparγ^lox/lox), referred to as mFATBLOCK or Pparγ^Δ/Δ (Figure 1A). We first assessed the overall health of mFATBLOCK mice under homeostatic conditions in the absence of any injury, 4 weeks after tamoxifen (TMX) administration to induce recombination of Pparγ. We found no difference in total body weight or in the proportion of lean versus fat mass, measured via EchoMRI, between wild-type (WT) and mFATBLOCK mice of either sex, 4 weeks after administration of TMX (Figure 1B). Thus, short-term loss of Pparγ from FAPs has no gross effects on overall weight, nor does it cause significant loss of muscle or fat mass.

Figure 1. FAP-specific deletion of Pparγ has no overall effects during homeostasis.

(A) Experimental design.

(B) Left: whole body weight measurements (females: n = 16; males: n = 8–9). Right: weight of lean and fat mass (females: n = 4–5; males: n = 4–6).

(C) Serum measurements for glucose, insulin, and C-peptide (females: n = 4–5; males: n = 4–6).

(D) Serum concentrations of free fatty acids (FFA), triglycerides (TG), total cholesterol (TC), adiponectin (ADIPOQ), and resistin (RESIS) of female mice (n = 4–5 mice).

(E) Experimental design.

(F) Oil red O staining of liver from WT and mFATBLOCK mice; scale bars: 25 μm. Bottom: immunofluorescence staining for adipocytes (open arrow, perilipin, green) and nuclei (DAPI; magenta). Scale bars: 200 μm. Quantification of liver lipid content (females: n = 6; males: n = 3–5). Quantification of the number of adipocytes normalized to muscle area (females: n = 5–10; males: n = 4–8).

(G) Phalloidin⁺ myofibers were false color-coded according to size. Scale bars: 200 μm. Weight of TA muscle (females: n = 4–12; males: n = 8–14). Quantification of average cross-sectional area (CSA) (females: n = 14–16; males: n = 8–10).

(H) In vivo force production for specific force and power (females: n = 12–20; males: n = 8–14). All data are represented as mean ± SEM. An unpaired two-tailed t test was used (NS, not significant).

We next determined glucose homeostasis by measuring blood glucose concentrations at resting state and after an overnight fast in both sexes of WT and mFATBLOCK mice and found no difference in glucose concentrations between genotypes (Figure 1C). We also detected no differences in serum levels of insulin and C-peptide. Thus, removal of Pparγ has no impact on glucose homeostasis. Next, we evaluated whether mFATBLOCK mice display any defects in their lipid homeostasis (Figure 1D). We failed to detect any significant differences in their serum levels of triglycerides (TG), total cholesterol (TC), or free fatty acids (FFA, Figure 1D). Lastly, we excluded any possible impact of the lack of FAP Pparγ on adipokine production, by measuring the serum levels for the adipose hormones adiponectin and resistin (Figure 1D), both of which displayed no differences between genotypes. Thus, deletion of Pparγ from FAPs does not cause gross alterations in body composition, circulating adipokine levels, or glucose or lipid homeostasis.

To further exclude any possible health impacts of FAP-specific Pparγ removal under non-injury conditions, we also performed histological evaluations of liver and muscle tissues in WT and mFATBLOCK mice (Figure 1E). As systemic lipid imbalance causes lipid accumulation within the liver, known as hepatic steatosis, we first compared the amount of lipid droplet accumulation, marked via the lipophilic dye Oil Red O, in the liver between WT and mFATBLOCK mice. As expected from our serum lipid analysis (Figure 1D), we found no differences between genotypes in both sexes (Figure 1F), confirming that mFATBLOCK mice do not display systemic lipid dysregulation.

We next asked whether FAP-specific Pparγ deletion impacts IMAT formation in muscle under homeostatic conditions by quantifying PERILIPIN⁺ adipocytes normalized to total TA (tibialis anterior) area (Figure 1F). We detected no significant difference between genotypes of either sex (Figure 1F), demonstrating that Pparγ deletion in FAPs does not affect IMAT in healthy muscle. Note that the overall low number of adipocytes in both genotypes is as expected, as murine muscles, in the absence of injury or disease, are almost completely devoid of IMAT.³⁹

To exclude any negative influence of Pparγ deletion from FAPs on muscle health and function, we also assessed overall muscle health by evaluating TA weight, myofiber size, and muscle function. To determine myofiber size (Figure 1G), we utilized our previously published pipeline (see STAR Methods for details and⁴⁰). Muscle function was measured via in situ nerve-mediated contractions of the TA (Figure 1H). We found no difference in TA weights, average cross-sectional area (CSA) of myofibers, peak power, or force production between WT and mFATBLOCK mice of both sexes (Figures 1G and 1H). Thus, deletion of Pparγ in FAPs under homeostatic conditions has no effect on muscle histology or function.

Taken together, these data suggest that inducible deletion of Pparγ within FAPs in the adult mouse has no short-term impact on systemic lipid or glucose balance and IMAT formation or muscle health under homeostatic, non-injury conditions.

IMAT formation is blocked in mFATBLOCK mice

While IMAT infiltration is a major pathological hallmark of the human muscle, murine muscles are very resistant to IMAT formation.³⁹ In order to force murine muscles to develop IMAT, we induced IMAT formation through a glycerol (GLY) injury, a widely accepted adipogenic model.^{21,22,25,41,42} In order to better understand the temporal dynamics of FAP-to-FAT transition, we first evaluated adipocyte differentiation at different time points post-GLY injury (Figure 2A). The differentiation of pre-adipocytes into adipocytes involves the accumulation of fusion of small lipid droplets into one large droplet, which expands and shrinks based on energy demands. The first signs of adipogenesis are visible on day 3 when platelet-derived growth factor receptor alpha (PDGFRα) positive FAPs started to accumulate PERILIPIN⁺ lipid droplets signifying the beginning of their differentiation into adipocytes. By day 4, most adipocytes are still immature and carry multiple small lipid droplets, which start to fuse and become larger, as well as more numerous, between days 5 and 7. By 21 days post-injury (dpi), adipocytes have fully matured and carry a single large lipid droplet.

Figure 2. mFATBLOCK mice display limited IMAT formation following a GLY injury.

(A) Differentiation from pre-adipocyte to adipocyte. Quantification of lipid droplet diameter in females after a GLY injury (n = 60–80 lipid droplets). Quantification of the number of adipocytes normalized to the injured area after a GLY injury (females: n = 4–20). Immunofluorescence 3 days post-GLY injury (dpi) of FAPs (PDGFRα, gray) beginning to differentiate. Lipid droplets stained with PERILIPIN (green) at 3, 4, 5, 7, and 21 dpi. Nuclei were visualized through DAPI (magenta). Scale bars: 10 μm for 3 dpi and 20 μm for the remaining time points.

(B) For all subsequent experiments, all mice were administered tamoxifen (TMX) through oral gavage. Experimental design. RT-qPCR of whole TA lysate for Pparγ and C/epbα at 3 and 5 days after a GLY injury (females: n = 6–8).

(C) Experimental design. Immunofluorescence staining for adipocytes from WT and mFATBLOCK mice 21 days after a GLY injury (PERILIPIN, green). Nuclei were visualized with DAPI (magenta). Scale bars: 250 μm. Quantification of the number of adipocytes normalized to the injured area 21 dpi (females: n = 9–20; males: n = 10).

(D) RT-qPCR of Plin1 expression 5 and 7 dpi in whole TA lysate (females: n = 3–8).

(E) Experimental design. Fluorescence image of primary FAPs isolated from WT and mFATBLOCK mice 3 days after a GLY injury, submitted to differentiation in vitro. After 5 days of differentiation, lipids were visualized through a LipidTOX staining (green) and nuclei through DAPI (magenta). Scale bars: 100 μm. Quantification of adipogenic differentiation between FAPs from WT and mFATBLOCK mice (n = 8 wells; n = 2 mice) (data normalized to WT and set to 1). See also Figure S1. All data are represented as mean ± SEM. An unpaired two-tailed t test was used. *p = 0.05, **p = 0.01, ***p = 0.001, and ****p ≤ 0.0001.

To determine whether Pparγ deletion within FAPs prevents adipogenic differentiation, we evaluated early differentiation in mFATBLOCK mice at 3 and 5 dpi. Adipogenesis is initiated through the transcription factors CCAAT/enhancer-binding proteins (C/EBPs).^43,44 Activation of these factors leads to downstream activation of PPARγ, which in turn determines cellular commitment to the adipogenic lineage and subsequently activates C/EBPα (Figure 2B). Following TMX administration through oral gavage, WT and mFATBLOCK mice TAs were injured with GLY, and expression of these transcription factors was evaluated at 3 and 5 dpi (Figures 2B and S1A). We found no difference in C/ebpβ or C/ebpδ expression between genotypes or timepoints, indicating that deletion of Pparγ does not cause upstream changes to the adipogenic signaling cascade (Figure S1A). However, as expected, we found strong repression in Pparg expression, including its downstream target Cebp/α, in mFATBLOCK mice compared to WTs at both 3 and 5 dpi (Figure 2B). Thus, our mouse model successfully removes Pparγ from FAPs and prevents the adipogenic differentiation signaling cascade from being initiated.

We next tested whether preventing the adipogenic signaling cascade leads to a decrease in adipocyte formation at 7 and 21 dpi, when IMAT has fully formed (Figure 2A). IMAT was determined through quantification of PERILIPIN⁺ cells normalized to the total injured area of the TA, as previously described⁴⁵ (Figure 2C). Excitingly, we found a significant decrease in IMAT formation at 7 (Figure S1B) and 21 dpi (Figure 2C) in both female and male mFATBLOCK mice compared to WT littermates. To note, while we routinely achieve around 80–85% recombination with the Pdgfrα^CreERT/+ allele, around 15–20% of FAPs still possess functional Pparγ, and can therefore still differentiate into adipocytes, most likely explaining the residual amount of IMAT we observe in mFATBLOCK mice. Next, we corroborated this decrease in IMAT by assessing the expression of mature adipocyte markers through reverse transcription quantitative polymerase chain reaction (RT-qPCR), Perilipin1 (Plin1) (Figure 2D) and adiponectin (AdipoQ) (Figure S1C), in whole muscle lysate. We found a strong repression of both genes at 5 and 7 dpi in mFATBLOCK mice compared to WT mice, demonstrating successful repression of adipocyte differentiation within the whole muscle. Last, we quantified the total area of muscle occupied by PERILIPIN-expressing adipocytes (Figure S1E). Deleting Pparγ from FAPs effectively blocked IMAT formation, resulting in a substantial reduction in the percentage of muscle area occupied by IMAT from 12% to 2% in mFATBLOCK mice. Taken together, we establish, using multiple time points and evaluation methods, that loss of Pparγ from FAPs prevents their differentiation into adipocytes, thereby blocking IMAT infiltration.

In addition to Cre negative controls, we initially included Cre positive but Pparγ heterozygous animals (Pdgfrα^CreERT/+ Pparγ^lox/+). In these mice, TMX administration should result in the removal of only one allele (Pparγ^Δ/+). When we analyzed WT, Pparγ^Δ/+, and Pparγ^Δ/Δ littermates for their ability to form IMAT, we discovered that Pparγ^Δ/+ mice displayed IMAT repression as an intermediate phenotype between WT and mFATBLOCK mice (Figure S1E), as previously reported in vitro³⁵ and in human patients.⁴⁶ Therefore, to optimize our breeding strategy, we focused on comparing WT and mFATBLOCK for all subsequent studies.

To elucidate whether deletion of Pparγ within FAPs represses their differentiation through a cell autonomous or non-autonomous effect, we performed two experiments. First, we ran an in vitro adipogenesis assay in isolated primary FAPs with and without Pparγ. For this, we injured mFATBLOCK and WT littermates with GLY, harvested TAs at 3 dpi and isolated FAPs through differential plating (Figure 2E). Once in culture, FAPs were induced to differentiate, and adipocyte formation was measured by quantification of lipid droplets through LipidTOX staining (Figure 2E). We found that FAPs isolated from mFATBLOCK mice had significantly lower adipocyte formation compared to WT FAPs (Figure 2E). Second, we genetically labeled FAPs before injury through an EYFP reporter⁴⁷ and traced their differentiation into EYFP⁺ adipocytes after a GLY injury. We utilized control (Pdgfrα^CreERT/+ Pparγ^+/+ Rosa^EYFP) and mFATBLOCK mice (Pdgfrα CreERT/+ Pparγ lox/lox Rosa EYFP), where, upon TMX administration, FAPs indelibly expressed EYFP (Figure S2F).²² Because fixation destroys endogenous EYFP fluorescence, we counterstained EYFP with an anti-GFP antibody, as previously described.^22,45 After TMX administration via oral gavage, TAs were injured with GLY. At 21 dpi, we quantified total PERILIPIN⁺ adipocytes and total GFP⁺ adipocytes and calculated the percentage of GFP⁺ PERILIPIN⁺ adipocytes over total adipocytes in both groups (Figure S1F). We found a strong decrease in total PERILIPIN⁺ and GFP⁺ adipocytes as well as the percentage of GFP⁺ adipocytes in mFATBLOCK mice compared to controls (Figure S1F). Thus, combined, our data demonstrate that deletion of Pparγ within FAPs causes a strong repression in IMAT formation in a cell-autonomous fashion, devoid of the influence of other cell types.

In addition to differentiating into adipocytes, FAPs are also progenitors of myofibroblasts, the cellular origin of fibrotic scar tissue.^25,30 As injectable injury models cause remodeling of the extracellular matrix (ECM) leading to a transient increase in collagen deposition,⁴⁸ we evaluated whether prevention of FAP differentiation into adipocytes causes a shift toward a more fibrotic fate upon loss of Pparγ. We determined collagen deposition in mFATBLOCK mice and WT littermates 21 dpi through the histological staining using Sirius Red^24,49 (Figure S2G). We found no significant difference in collagen deposition between genotypes in either sex (Figure S2G). We confirmed this by quantifying the early fibrotic response at 5 dpi, through expression of known fibrotic genes: the fibrotic inducer, transforming growth factor β (Tgf-β), and the myofibroblast marker smooth muscle α-actin (Acta2). We found no difference in Tgf-β or Acta2 expression between mFATBLOCK mice and WT littermates (Figure S1H), indicating that deletion of Pparγ within FAPs neither redirects their fate toward a fibrotic lineage nor impacts collagen deposition.

Blocking FAP adipogenic differentiation changes their cellular dynamics

FAP numbers are tightly regulated during regeneration with rapid early proliferation followed by apoptosis to prevent uncontrolled expansion that could lead to intramuscular fat or fibrosis.^48,50 Due to the block in adipogenic differentiation upon removal of Pparγ from FAPs, we investigated how this manipulation in fate affects the overall dynamic of FAPs during early regeneration (Figure 3A). We found no difference in the total number of PDGFRA⁺ FAPs at 3 dpi (Figure 3B), indicating that, during the early regenerative phase, loss of Pparγ has no discernable impact on FAPs. In contrast, we detected an increase in PDGFRA⁺ FAPs lacking Pparγ compared to WT FAPs, starting on day 5 and persisting through day 7 until 21 dpi.

Figure 3. Prevention of FAP differentiation causes an increase in FAP population.

(A) Experimental design.

(B) Immunofluorescence of FAPs 7 dpi (PDGFRα, magenta) and nuclei (DAPI, white). Arrowheads indicate examples of FAPs. Scale bars: 100 μm. Quantification of total FAPs per 20× field (female: n = 8–14) at 3, 5, 7, and 21 dpi.

(C) Immunofluorescence of FAPs (PDGFRα, magenta) and Ki67 (green) at 3 dpi. Yellow arrowheads indicate examples of proliferating FAPs. Scale bars: 100 μm. Quantification of proliferating FAPs over total FAPs (%) at 3 and 5 dpi (female: n = 8).

(D) Immunofluorescence of FAPs (PDGFRα, magenta) and cleaved caspase 3 (cC3, green) at 5 dpi. Yellow arrowheads show examples of FAPs undergoing apoptosis. Scale bars: 100 μm. Quantification of percent of FAPs undergoing apoptosis in WT and mFATBLOCK (female: n = 7–13) mice 5 and 7 dpi. All data are represented as mean ± SEM. A multiple unpaired two-tailed t test followed by a Holm-Šídák post-hoc test was used. NS = not significant; *p = 0.05, **p = 0.01, ***p = 0.001, and ****p ≤ 0.0001. See also Figure S2.

To determine whether lack of Pparγ results in increased proliferation, we evaluated FAP proliferation in WT and mFATBLOCK mice by quantifying the number of PDGFRA⁺ FAPs expressing the proliferation marker antigen Kiel 67 (Ki67) at 3 and 5 dpi (Figure 3C). While the proliferation rates of WT FAPs remained constant between days 3 and 5, FAP lacking Pparγ displayed increased proliferation on day 3 before returning to WT levels on day 5.

We also asked whether loss of Pparγ results in enhanced survivability by quantifying apoptosis in FAPs through cleaved Caspase 3 (PDGFRA⁺ cC3⁺ cells as percent of all PDGFRA⁺ FAPs) at 5 and 7 dpi (Figure 3D). We found a small increase in FAPs undergoing apoptosis at 5 dpi in mFATBLOCK mice compared to WT controls, with normal apoptosis rates at 7 dpi (Figure 3D), potentially compensating for the increase in FAP numbers.

Thus, loss of Pparγ leads to an increase in FAP numbers, possibly due to an early but transient burst of proliferation. An alternative, and equally possible, explanation is that FAPs without Pparγ remain a PDGFRA⁺ FAPs and do not differentiate into an adipocyte compared to WT FAPs, which lose PDGFRA expressing and turn into fat cells.

Muscle regeneration is a process that requires orchestration of multiple cell types that leads to efficient tissue regeneration with FAPs being a major influencer. Due to the observed alterations in FAP numbers, we evaluated two other cell populations that crosstalk with FAPs: immune and endothelial cells. We found no difference in expression of genes via RT-qPCR of whole muscle lysates related to hematopoietic cells (Cd45), leukocytes (Cd11b), or mature macrophages (F4/80) at 5 dpi (Figure S2A) arguing against any impact of FAPs lacking Pparγ on the regenerative immune response. Next, we analyzed endothelial cells at 7 and 21 dpi to evaluate whether the persistent increase in FAP numbers impacted vascularization after injury. We visualized endothelial cells through fluorescently-tagged Isolectin in WT and mFATBLOCK mice and found no difference in the percentage of muscle occupied by endothelial cells at either 7 or 21 dpi between genotypes (Figure S2B), indicating that the increase in FAP population does not impact the vasculature following a glycerol injury.

Taken together, preventing FAP differentiation into adipocytes via removal of Pparγ leads to an increase in FAP numbers, likely due to an initial and transient increase in proliferation, without causing any detectable effects on cellular dynamics during regeneration.

IMAT impacts myofiber density and muscle function

There is a strong inverse correlation between IMAT formation and muscle mass and strength in human patients. While we and others have shown that this inverse relationship is also present in mice,^39,51–54 it is still unclear what the direct effects of IMAT are on muscle regeneration. Therefore, we sought to utilize our mFATBLOCK mouse model to understand the direct impact of IMAT on muscle regeneration and function.

Following TMX administration through oral gavage and a GLY injury, we assessed whole muscle function in mFATBLOCK mice and WT littermates. At 21 dpi, nerve-mediated in situ force measurements were conducted on the TA muscle (Figure 4A). Excitingly, we found a significant increase in submaximal and maximal isometric specific force and peak power output at 35% of maximal force in both sexes of mFATBLOCK mice compared to their WT littermates (Figures 4B and 4C). Thus, IMAT severely compromises the functional recovery of muscle.

Figure 4. IMAT negatively impacts muscle function.

(A) Experimental outline.

(B and C) In vivo force production of the TA 21 days after a GLY injury (dpi). (B) Specific force generated over a range of frequencies (1–150 Hz) (females: n = 7–17; males: n = 8–12). (C) Power production (females: n = 7–12; males: n = 7–9).

(D) PHALLOIDIN⁺ myofibers were false color-coded according to size 21 dpi. Scale bars: 500 μm.

(E) Quantification of whole TA area cross-section 21 dpi (females: n = 10–20; males: n = 9–10).

(F) Quantification of the number of fibers normalized to injured TA area 21 dpi (females: n = 9–19; males: n = 9–10).

(G) Quantification of average cross-sectional area of muscle fibers normalized to body weight (females: n = 10–21; males: n = 8).

(H) Distribution of myofibers based on the amount of centrally located nuclei 21 dpi (females: n = 10–22; males: n = 10).

(I) Immunofluorescence of fiber types (2B, blue; 2A, cyan) and outlined fibers (Laminin, orange). Quantification of fiber type composition, as a percentage of total fibers (females: n = 5–7). Scale bars: 100 μm. All data are represented as mean ± SEM. An unpaired two-tailed t test (C, F, and G) or two-way ANOVA followed by a Tukey’s multiple comparison (B, H, and I) was used. *p = 0.05, **p = 0.01, ***p = 0.001, and ****p ≤ 0.0001. See also Figure S3.

To determine if the functional improvement in our mFATBLOCK mice is due to changes in myofiber regeneration, we histologically assessed muscle regeneration 21 days post-GLY injury (Figure 4D). False colored-coded muscle fibers were visualized through phalloidin staining, segmenting, and measuring using our previously published pipeline.⁴⁰ We found no difference in total TA area between genotypes (Figure 4E), indicating that prevention of IMAT formation is not causing any gross phenotypic differences. However, we discovered a significant increase in the density of myofibers in mFATBLOCK mice compared to WT littermates in both sexes (Figure 4F), indicating that IMAT is negatively affecting the number of fibers that are able to regenerate after an injury. Next, we assessed the average CSA of myofibers normalized to body weight and found that female mFATBLOCK mice had a higher average CSA compared to WTs, with male mice approaching significance (Figure 4G).

As shown before, Pparγ heterozygous animals displayed an intermediate phenotype with reduced but not blocked IMAT formation (Figure S1E). To determine whether a modest reduction in IMAT can improve muscle regeneration, we also analyzed the muscles of Pparγ heterozygous mice. The only significant improvement we detected was increased myofiber density in heterozygous females compared to WT females, similar to the mFATBLOCK phenotype (Figure S3A). This suggests that enough IMAT can still form in the heterozygous animals to impair muscle regeneration, especially in males.

As we detected an increase in average CSA in mFATBLOCK mice, we asked whether this is due to defective myoblast fusion. For this, we quantified the distribution of fibers based on the number of centrally located nuclei. We found no difference in the ratio of myofibers carrying one or more centrally located nuclei between genotypes of either sex (Figure 4H), indicating that IMAT has no effect on myoblast fusion.

To explore whether IMAT might cause a shift in fiber type distribution in our model, we analyzed the composition of fiber types in WT and mFATBLOCK mice 21 days after a GLY injury (Figure 4I). Visualized through the myosin-specific expression of each fiber type, we found no difference in fiber types 2B, 2X, or 2A (Figure 4I), indicating that IMAT does not cause fiber type shifts.

Additionally, we evaluated IMAT and muscle regeneration following a cardiotoxin (CTX) injury model, where IMAT formation is limited in comparison to a GLY injury.³⁹ Following TMX administration to mFATBLOCK and WT littermates, TAs were injured with CTX and harvested 21 dpi (Figure S4A). We found a significant decrease in IMAT formation in both sexes of mFATBLOCK mice compared to WT littermates, although the total IMAT amount was significantly less (~20-fold) compared to a GLY injury (Figure S4B). As IMAT infiltration is often accompanied by fibrosis, we also evaluated collagen deposition 21 days following a CTX injury and found that the area occupied by collagen was significantly reduced in mFATBLOCK mice compared to WT controls within both sexes (Figure S4C). While we did not detect changes in collagen deposition after a GLY injury, these observations suggest that, at least with CTX, IMAT formation might be linked to excessive collagen deposition. Importantly, the data from both GLY- and CTX-induced injuries demonstrate that FAPs lacking Pparγ are not adopting a fibrogenic fate as an alternative. Lastly, we evaluated muscle regeneration by assessing myofiber density and average CSA, as described above (Figure S4D). We found no difference in any of these measurements between genotypes in either sex, demonstrating that muscle regeneration after a CTX injury is not affected in mFATBLOCK mice.

Taken together, our results demonstrate that IMAT, if allowed to form, causes a decline in myofiber density and size, resulting in functionally impaired muscle. In addition, the fact that IMAT only had an impact on myofiber regeneration post-GLY injury, which causes dramatically more accumulation of adipocytes than CTX, indicates the presence of an IMAT threshold that is needed for IMAT to have a negative impact.

Myogenesis is not influenced by IMAT or FAPs lacking Pparγ

FAPs are crucial producers of many pro-myogenic cues.^22,31,55,56 To determine whether the functional recovery in mFATBLOCK mice is due to loss of Pparγ itself or the increase in FAPs in the mFATBLOCK mice, we screened multiple known myogenic cues, produced and secreted by FAPs, via RT-qPCR at 3 and 5 days post-GLY injury but detected no differences between genotypes (Figures 5A and S5A, respectively). We chose these early timepoints as we detected the first changes to FAP dynamics and blocked adipogenic differentiation starting at 3 dpi. We also asked whether FAPs lacking Pparγ have any effect on the immune response by evaluating cytokine secretion (Figure 5B). For this, we ran muscle-specific protein lysates through a 32-Plex Cytokine assay but found that neither loss of Pparγ nor IMAT repression caused any significant changes. Thus, deletion of Pparγ does not result in a discernible secretory phenotype.

Figure 5. IMAT does not cause aberrant signaling during regeneration.

(A) RT-qPCR of whole muscle TA lysate (female: n = 6–8) 3 days post-GLY injury for FAP-derived signals.

(B) Cytokine ELISA of whole protein TA lysate (female: n = 5) 5 days post-GLY injury.

(C) Immunofluorescence of C2C12-derived myofibers (MF20, red) and nuclei (DAPI, blue) 5 days after differentiation and co-cultured with 10,000 or 20,000 3T3L1 cells. Scale bars: 250 μm. Quantification of differentiation and fusion indices, normalized to the control C2C12 condition.

(D) Concentration of LEPTIN in serum (“circulating”) and tibialis anterior (TA, “muscle”) lysate (females: n = 4–6) 5 days post-GLY injury.

(E) Experimental design. Immunofluorescence of nuclei (DAPI, blue) and C2C12-derived myofibers (MF20, red) 3 days post-differentiation induction, cultured with control media or fat-derived conditioned media. Scale bars: 100 μm. Quantification of differentiation and fusion indices, normalized to control media.

(F) Experimental design. Heatmap of RNA sequencing results of whole muscle TA lysates 5 dpi (females: n = 4–5). Volcano plot of differentially expressed genes with log2 fold-change ≥1 and an adjusted p value of <0.05 (genes in red are downregulated and in green are upregulated).

(G) Gene ontology enrichment results for differentially expressed genes. All data are represented as mean ± SEM. An unpaired two-tailed t test (D and E), a multiple unpaired two-tailed t test followed by a Holm-Šídák post hoc test (A), or a one-way ANOVA followed by a Dunnett’s multiple was used (C). NS = not significant; **p = 0.01 and ***p = 0.001. See also Figure S5.

To further exclude that FAP-specific changes cause functional recovery in mFATBLOCK mice, we performed a series of in vitro experiments. First, we assessed whether deletion of Pparγ influences myogenesis by isolating primary FAPs from WT and mFATBLOCK mice 3 days after a GLY injury through differential plating⁵⁶ and co-cultured them with the myoblast cell line, C2C12 (Figure S5B). After 5 days, we visualized mature myofibers through expression of the heavy chain of myosin II (MF20) and calculated differentiation and fusion indices between WT, mFATBLOCK, and a C2C12-alone control. We found a significant increase in both differentiation and fusion indices when FAPs were co-cultured with C2C12s, independent of their genotype (Figure S5B). Therefore, while FAPs promote myogenic differentiation as previously demonstrated,²⁵ this effect is not influenced by expression of Pparγ within FAPs. As we observed an increase of ~20% in the number of FAPs (Figure 3B), we next determined whether an increase in the FAP population could have an additive effect on myogenic differentiation (Figure 5C). Upon induction, C2C12s were co-cultured with 10,000 versus 20,000 cells of the pre-adipocyte cell line, 3T3L1, which we and others have shown to be a suitable proxy for primary FAPs.^5,22,56–58 We found that co-culturing 3T3L1 cells with C2C12s significantly promoted their differentiation and fusion, as seen with primary FAPs (Figure 5C). However, this effect was independent of the number of cells used, indicating that the 20% increase in FAPs we observed in vivo is unlikely to explain the functional improvement in the mFATBLOCK mice.

In addition to FAPs, adipocytes have also been shown to influence muscle by secreting anti-myogenic signals.^4,59–62 Thus, it is possible that IMAT acts as a signaling center that negatively influences muscle regeneration. One crucial adipogenic signal that influences myofiber size is leptin.⁶² To test whether IMAT affects systemic and/or local leptin levels, we measured circulating leptin concentrations in serum and local levels in muscle lysates 5 days post-GLY injury by enzyme-linked immunosorbent assay (ELISA) (Figure 5D). We found no differences in leptin concentration in the serum or muscle lysate between mFATBLOCK and WT mice (Figure 5D).

To further evaluate whether adipocytes could directly affect myogenesis, we performed a conditioned media experiment. We first cultured 3T3L1 cells, a widely used pre-adipocyte cell line, induced their adipogenic differentiation for 7 days, and collected their conditioned media. Next, we cultured C2C12 myoblasts and, upon myogenic differentiation, supplemented the myogenic induction media daily with conditioned media from adipocytes (1:1). After 3 days, we evaluated myogenic differentiation and myoblast fusion by quantifying MF20⁺ myotubes and DAPI⁺ myonuclei. We found no difference between the conditions, indicating that factors secreted by adipocytes do not affect myogenesis in vitro (Figure 5E).

Last, we queried the muscle transcriptome of mFATBLOCK mice at 5 days post-GLY injury, a time when most adipocytes have started to form (Figure 2A), to ask whether we can detect any significant changes upon loss of Pparγ compared to control mice. For this, we isolated whole-muscle RNA from TAs 5 days post-GLY injury and performed bulk RNA sequencing (Figure 5F). Surprisingly, we only detected ~200 down- and ~50 upregulated genes (log2 fold-change >1 and padj <0.05). Besides Pparγ, most of the top downregulated genes, such as Fabp4, Lpl, Tusc5, Tmem120a, Aqp7, and Enpp2, have known roles in adipogenesis.^63–67 Fittingly, a gene ontology enrichment analysis highlighted that most processes, which changed upon Pparγ deletion, are related to adipogenesis and fat metabolism (Figure 5G). One of the hits suggested that brown fat cell differentiation might be affected. Upon a closer look, 16/17 of these differentially expressed genes represented common adipogenic genes shared with white adipose tissue, such as Pparγ, C/ebpα, and Adipoq. Fittingly, we failed to confirm any differences in multiple brown fat-specific genes by RT-qPCR at 7 days post-GLY (Figure S5C).

Combined, our data highlight that removal of Pparγ from FAPs and the subsequent repression of IMAT formation does not cause any detectable changes in adipokines, cytokines, or myogenic factors.

IMAT restricts the formation of new fibers

To determine why mFATBLOCK mice display enhanced functional recovery, we evaluated the early myogenic processes in mFATBLOCK mice and WT littermates. Following TMX administration through oral gavage, TAs were injured with a GLY injection and harvested at 3, 4, 5, and 7 dpi (Figure 6A). We first analyzed the expression of the MuSCs marker, Pax7, and myoblast marker, MyoD, by RT-qPCR (Figure 6B). We found no difference in the expression of Pax7 or MyoD between genotypes 3 and 5 days post-GLY injury. We also quantified the total number of myoblasts, characterized by Myogenin expression (MyoG⁺ cells). At 3 and 5 days post-GLY injury, MyoG⁺ myoblasts displayed no difference between genotypes at either time point (Figure 6C). Thus, IMAT or FAPs lacking Pparγ do not impact early myogenic differentiation.

Figure 6. IMAT limits myofiber density but not myogenic differentiation.

(A) Experimental outline.

(B) RT-qPCR of whole muscle lysate (females: n = 6–8) 3 and 5 days post-GLY injury (dpi) for Pax7 (MuSCs) and MyoD (myoblasts).

(C) Immunofluorescence for myoblasts (MYOG, green; open arrow), and nuclei (DAPI, magenta) at 3 dpi. Scale bars: 100 μm. Quantifications of MYOG⁺ nuclei per 20× field of view (females: n = 7–8) at 3 and 5 dpi.

(D) Immunofluorescence detection of newly formed myofibers (embryonic myosin; MYH3, red), adipocytes (PERILIPIN, green), and nuclei (DAPI, blue) 4 dpi. Scale bars: 100 μm. Quantification of MYH3⁺ myofibers normalized to injured area 4 and 5 dpi (females: n = 4–8).

(E) Distribution of myofibers based on the amount of centrally located myonuclei at 5 dpi (females: n = 6–8).

(F) Immunofluorescence of myofibers (PHALLOIDIN, magenta) and adipocytes (PERILIPIN, green) 7 dpi. Scale bars: 500 μm. Zoom scale bars: 150 μm. Quantifications of the number of myofibers normalized to the injured area 5 and 7 dpi (females: n = 5–8).

(G) CSA myofiber quantification (females: n = 8–11; males: n = 8–9).

(H) Correlations at 21 dpi between IMAT (% muscle occupied by IMAT, data from Figure S1E) and specific force (data from Figure 4B), myofiber density (data from Figure 4F), and myofiber size (data from Figure 4G).

(I) Experimental design. Immunofluorescence of adipocytes (PERILIPIN, magenta) and co-cultured C2C12-derived myotubes (MF20, yellow) 5 days after myogenic differentiation. Correlation plot of % area occupied by adipocytes versus myotubes of randomly selected fields. All data are represented as mean ± SEM. An unpaired two-tailed t test and a Pearson’s correlation were used. NS = not significant; **p = 0.01 and ***p = 0.001. See also Figure S5.

We next evaluated newly formed fibers, characterized by expression of embryonic myosin (MYH3⁺ fibers), in mFATBLOCK mice and WT littermates at 4 and 5 days post-GLY (Figure 6D). Here, we detected a robust increase in the number of MYH3⁺ fibers 4 days post-GLY in mFATBLOCK mice compared to WT littermates (Figure 6D). By 5d pi, only a few MYH3⁺ fibers could still be detected with no difference between genotypes, indicating a dramatic but transient increase in newly formed fibers in mFATBLOCK mice. To determine whether this increase in MYH3⁺ fibers is due to increased myoblast fusion, we quantified the number of nuclei within regenerating fibers (Figure 6E). No differences between genotypes were observed, indicating that IMAT does not impact myoblast fusion.

To determine whether the dramatic increase in MYH3⁺ fibers at 4 dpi (Figure 6D) translates into the increased myofiber density present at 21 dpi (Figure 4F), we quantified myofiber density on days 5 and 7 (Figure 6F). While we found no difference between genotypes at 5 dpi, myofiber density was almost two-fold greater in mFATBLOCK compared to WT mice 7 dpi. We also evaluated myofiber size (CSA) at 7 dpi, when most myofibers have regenerated but still need to undergo hypertrophy to reach pre-injury size. However, we found no difference in CSA between genotypes of either sex (Figure 6G).

Interestingly, PERILIPIN⁺ adipocytes start to appear by day 4 (Figures 6D and 2A) with massive IMAT infiltration on day 7 (Figures 6F and 2A), which is potently inhibited in the mFATBLOCK mice (Figure S1B). As shown above (Figure 2), deletion of Pparγ from FAPs effectively blocked IMAT formation, resulting in ~10% reduction in the percentage of muscle area occupied by IMAT in mFATBLOCK mice (Figure S1E). As the total muscle area remained unchanged (Figure 4E), the absence of IMAT in the Pparγ-deficient muscles created approximately 10% more space within the muscle tissue, potentially allowing the formation of a higher density of myofibers. Combined with the fact that early myogenesis is unaffected by either IMAT or FAPs lacking Pparγ (Figure 5), our data strongly argue that IMAT acts as a physical barrier, which inhibits a portion of available muscle volume, thereby preventing new myofiber formation. Considering that changes in myofiber size are only present at 21 dpi (Figure 4G) but not 7 dpi (Figure 6G), our data also indicate that IMAT impacts hypertrophic myofiber maturation between 7 and 21 dpi. If our hypothesis is true, we would expect a close inverse correlation between the muscle area occupied by IMAT versus myofibers. To test this, we plotted the area occupied by IMAT against maximum specific force and myofiber density and size (CSA) from mFATBLOCK and control mice at 21 dpi (data from Figures 2 and 4). A simple linear regression analysis confirmed the strong negative correlation between IMAT and muscle: the more space within muscle that is occupied by IMAT, the fewer and smaller the fibers are, ultimately resulting in reduced force production (Figure 6H).

To test this hypothesis further, we turned to an in vitro experiment, where we artificially created areas covered by versus devoid of adipocytes and asked whether myotube formation would be physically restricted by fat (Figure 6I). We again cultured 3T3L1 cells and induced their differentiation into adipocytes. After 10 days, we scratched off some adipocytes to create a checkerboard-like pattern of areas devoid of adipocytes. We then seeded C2C12 cells on top of the adipocytes and induced their differentiation into myotubes for 5 days. We found that C2C12 cells readily and homogenously differentiated into myotubes in the scratched-off areas devoid of adipocytes. However, in areas where adipocytes were present, myotube formation was limited. Thus, the only restriction on where myotubes were able to form was whether the area, at time of seeding C2C12, was devoid of adipocytes. We confirmed this by quantifying random areas, where we could detect a clear negative correlation between the area occupied by adipocytes versus myotubes. To note, this experiment also demonstrates that adipocytes do not directly impair myogenesis either via secreted factors or through direct cell contact, confirming our previous experiments (Figure 5E). Thus, similar to our in vivo data, adipocytes physically restrict the area in which myotubes are allowed to form rather than through the production of an anti-myogenic signal.

Taken together, our results demonstrate that IMAT does not affect myogenic differentiation but rather points to the exciting possibility that IMAT acts as a physical barrier that blocks myofiber formation. Our data also argue that IMAT physically prevents myofiber hypertrophy by limiting the space the myofibers can expand into.

DISCUSSION

Accumulation of IMAT is a shared feature of age and disease and is strongly associated with decreased muscle strength and function.^{5–14,31,39,68} However, it is still unresolved whether IMAT is actively repressing muscle regeneration and/or function. In this study, we created an inducible IMAT prevention mouse model, called mFATBLOCK, where we can inhibit the cellular origin of IMAT, FAPs, from differentiating into adipocytes. Deletion of the key adipogenic regulator, Pparγ, from FAPs, blocked their adipogenic differentiation, in a cell autonomous manner with no detectable shift in their cellular fate, thereby effectively preventing IMAT formation. Interestingly, the repressive impact of IMAT was independent of post-regenerative MuSC-driven myogenesis. In addition, IMAT did not act as a major signaling center or broadly influence intercellular communication. Instead, our data strongly argue that IMAT functions as a physical barrier: during early regeneration, to prevent nascent myofibers from forming, and during later regeneration, to restrict their post-regenerative hypertrophic growth. Together, our results define the negative impact of IMAT on regenerating muscle and emphasize the importance of developing novel approaches to target IMAT shortly after an adipogenic injury to prevent long-term muscle loss.

IMAT: Innocent bystander or active participant?

Several studies have tried to address whether IMAT might have a direct influence on muscle regeneration with mixed results. The Feige group took advantage of mice with a whole-body knockout of Pparγ, the master regulator of adipogenic fate.⁶⁹ These mice are incapable of developing adipose tissue, including IMAT 14 days post-GLY injury. Importantly, Pparγ^−/− mice displayed compromised myofiber recovery, suggesting that IMAT is important for muscle regeneration.⁷⁰ However, whole-body Pparγ^−/− mice are lipodystrophic, as they are completely devoid of all fat, and develop liver steatosis, resulting in severe metabolic defects. In addition, PPARG functions in other cell types, including MuSCs and macrophages.^71,72 Thus, the observed phenotype could be due to severe metabolic distress and/or off-target effects.

Using a clever and rigorous cell ablation approach to kill adipocytes, the Kuang group found that fewer myofibers regenerated after a CTX injury when IMAT formation was repressed.³² However, the Cre line used to ablate adipocytes is also active in endothelial cells,^73–76 macrophages,^77–79 and adipogenic progenitors⁸⁰ and, thus, would also have ablated these cell types. In addition, they only used a CTX injury, where we found myofiber regeneration to be independent of IMAT (Figure S4).

Using another elegant fat-ablation model, the Meyer group showed that IMAT causes contractile tension deficits of TA muscles 21 days post-GLY injury.⁸¹ However, this fat-ablation model also causes lipodystrophy, like the germline Pparγ KO model. Interestingly, mice devoid of all adipose tissues developed muscles with lower cross-sectional area and peak tetanic tension at baseline prior to an injury, indicating a general pro-myogenic role for adipose tissue. Recently, the Meyer group demonstrated through clever functional reconstitution experiments that these baseline deficits are due to disrupted leptin levels, a potent adipokine secreted by adipocytes. Leptin may also be required for post-regenerative myofiber hypertrophy, as, despite functional rescue, myofiber size was reduced post-injury in mice devoid of adipose tissue. Thus, adipose tissue, via leptin, is critical for the development of normal muscle mass and strength.⁶²

Here, we describe our innovative mFATBLOCK model in which we genetically deleted Pparγ specifically and postnatally in FAPs to successfully inhibit their adipogenic differentiation into IMAT without causing major phenotypic changes. In the absence of injury, 15-week-old mFATBLOCK mice displayed normal body weights and showed no indication of lipodystrophy 5 weeks post-TMX administration. Indeed, our model had no gross effects on overall health, glucose handling, insulin secretion, lean or fat mass, or lipid imbalance. Importantly, postnatal deletion of Pparγ specifically in FAPs did not cause massive cell death nor had any impact on muscle homeostasis, as assessed by muscle weight, function, and myofiber size prior to injury (Figure 1), most likely due to preserved adiposity (Figure 1B) and resulting leptin levels (Figure 5D). Taken together, these data suggest that inducible deletion of Pparγ within FAPs in the adult mouse has no short-term impact on systemic lipid or glucose balance, IMAT formation, or muscle health under homeostatic, non-injury conditions.

Thus, our mFATBLOCK model allowed us to determine the direct impact of IMAT on muscle regeneration. Surprisingly, our data argue that IMAT acts as a physical barrier, initially hindering nascent myofiber formation and later restricting their hypertrophic growth. Without IMAT, the functional recovery of muscle was dramatically enhanced (Figure 4). Surprisingly, the improved recovery was independent of post-regenerative MuSC-driven myogenesis and significant intercellular communication (Figure 5). Instead, our in vitro and in vivo data argue that the early, post-injury formation of IMAT hinders nascent myofiber formation and later restricts their hypertrophic growth, ultimately leading to enhanced force production (Figure 6).

IMAT threshold: What are the implications?

We recently compared IMAT formation to myofiber regeneration between CTX and GLY injuries. We found a strong negative correlation between IMAT and myofiber size: the more IMAT had formed, the smaller the newly regenerated myofibers were.³⁹ Fittingly, here we only detected the negative influence of IMAT on myofiber regeneration after a GLY but not a CTX injury (Figure 4 versus Figure S4). Given the significant difference in IMAT levels between the two injuries, our results strongly argue for the existence of an IMAT threshold, and that, only when exceeded, will IMAT become detrimental to muscle regeneration. The existence of an IMAT threshold could have major implications for murine models that try to mimic human conditions, which display massive IMAT infiltration as a key pathological phenotype. For example, in our previous study, we demonstrated that IMAT formation is dependent on the genetic background (i.e., strain) and sex of the mouse.³⁹ Female mice across different strains and injuries produced more IMAT, while C57/BL6 mice were extremely lean and formed very little IMAT. Therefore, if a specific IMAT threshold is indeed necessary for IMAT to influence myofiber regeneration, careful consideration must be given to the sex, strain, and injury type when studying the effects of IMAT on muscle health.

Removal of Pparγ has no major impact on FAPs

FAPs are crucial for the repair of damaged muscles by secreting beneficial factors.^{21,22,24–30} Therefore, we carefully assessed any potential impact of the loss of Pparγ on the health of FAPs. Overall, we detected a ~20% increase in the overall FAP population in mFATBLOCK mice starting at day 5 and persisting until our endpoint of 21 dpi (Figure 3). While we noticed a modest increase in FAP proliferation at day 3, we believe that the increase in FAP numbers is rather due to the fact that FAPs lacking Pparγ fail to differentiate into adipocytes and thereby maintain their FAP fate. If true, it would also suggest the interesting possibility that up to 20% of all WT FAPs, based on the ~20% difference in FAP numbers, differentiate into adipocytes upon injury. It will be interesting to determine whether the 20% increase in FAPs represents a specific subpopulation inherently predisposed to forming adipocytes, or if all FAPs possess this capability, with only a fraction (i.e., 1 out of 5) actually undergoing adipogenesis. Importantly, this increase in FAP numbers is not detrimental to the muscle as we did not observe increased collagen deposition post-GLY (Figure S1G) but rather reduced levels 21 days post-CTX (Figure S4C).

As FAPs are essential for regenerative myogenesis,^27,28 the observed increase in FAPs could increase in pro-myogenic factors, thereby providing an alternative explanation for why loss of Pparγ could improve muscle regeneration. After carefully assessing multiple known factors, we failed to detect any differences in mFATBLOCK mice compared to controls. In fact, we were unsuccessful in detecting any changes in a variety of signals (adipokines and cytokines) or global transcriptional changes beyond genes involved in adipogenesis (Figure 5). We also failed to detect any direct impact of FAPs lacking Pparγ or number of cells on C2C12-driven in vitro myogenesis (Figures S5B and 5C). Co-culturing FAPs with C2C12s significantly increased myogenic differentiation and fusion, regardless of FAP genotype, confirming that FAPs promote myogenesis.²⁵ However, increasing the number of 3T3L1 cells, a FAP proxy, did not further enhance this effect, suggesting the observed 20% increase in FAPs in mFATBLOCK mice is unlikely the primary reason for the improved functional recovery of muscle.

Besides adipocytes,^{5,21–23,25,30,32,33} FAPs can also differentiate into myofibroblasts, the cellular origin of tissue fibrosis.^25,30 Moreover, an increase in FAP numbers is often associated with an increase in fibrosis.^25,48,50 We failed to detect any increase in collagen deposition post GLY injury despite the increase in FAPs. Interestingly, mFATBLOCK mice displayed elevated collagen content 21 days post-CTX, although overall collagen levels were significantly lower compared to GLY. Thus, while repressing IMAT formation might cause minor ECM remodeling in an injury-dependent fashion, these data argue against FAPs differentiating into myofibroblasts as an alternative fate upon Pparγ deletion.

IMAT’s role in acute injuries versus chronic conditions

One important aspect to consider is how IMAT arises and whether it would have the same effect on muscle depending on the etiology of the disease. In this study, we have shown that in an injury setting that allows IMAT to form while muscle is regenerating, the regenerative process is hindered, leading to a loss in muscle fiber density and, ultimately, muscle function. One example of this type of acute but adipogenic injury is rotator cuff tears in the supraspinatus muscle. The greater the tear, the more FAPs expand, leading to massive IMAT infiltration.^82–86 Consequently, muscle regeneration is compromised, resulting in muscle atrophy and a declined functional output. Thus, preventing this rapid IMAT infiltration could have significant benefits.

However, the effects of IMAT on muscle could be substantially different in other diseases devoid of injury, where metabolic defects lead to progressive IMAT infiltration. For example, during aging and obesity, IMAT progressively infiltrates muscle over time, and its effects on muscle could be substantially different from an acute injury.^9–11,13 Adipose tissue is a potent endocrine organ, and IMAT in particular has been shown to modulate muscle insulin sensitivity, possibly via secretion of inflammatory cytokines.⁴ There is also some evidence that adipocytes could impair myogenesis in vitro via secretion of certain cytokines such as IL6, IL-1β, and TNF-α,^87–89 or myofiber size in vivo through leptin.⁶² In our IMAT prevention model, we found no differences in these targets or other inflammatory markers. In addition, we failed to detect any impact of IMAT, either via secreted molecules or through direct cell-cell interactions, on C2C12-driven in vitro myogenesis (Figure 5). Combined, this suggests that injury-induced IMAT in a metabolically healthy organism results in a significantly different response.

Neuromuscular diseases such as Duchenne muscular dystrophy (DMD) are known to have bouts of degeneration followed by regeneration of muscle (mimicking injury), while also having dysfunctional signaling and metabolism,^90,91 seen by abnormal amino acid, energy, and lipid metabolism, as well as calcium balance, mitochondrial function, and insulin resistance that is independent of corticosteroid use.⁹² Thus, DMD and similar diseases might represent a scenario where IMAT might act both as a physical barrier to prevent the regeneration of damaged myofibers and a potent immune modulator. In the future, by combining mouse models which display progressive but complete replacement of muscle with IMAT⁶ with our mFATBLOCK model will allow us to define the long-term impact of IMAT on muscle health in chronic conditions.

In summary, this work demonstrates that IMAT formation following acute muscle injury limits the formation of new myofibers and restricts their post-regenerative hypertrophic growth. Importantly, genetically suppressing IMAT formation increased the number of myofibers and muscle strength without evidence of altered myogenesis. Together, these observations are consistent with a scenario where IMAT acts as a physical impediment to functional muscle regeneration, although additional experiments may be required to prove this definitively. Our work, therefore, highlights the importance of limiting IMAT formation as a potent strategy to improve the healing of acute but adipogenic injuries such as rotator cuff tears.

Limitations of the study

While our findings strongly indicate that IMAT impairs muscle regeneration primarily through physical obstruction, without definitive evidence of signaling interference with myogenesis, we cannot formally exclude the possibility of IMAT exerting some direct influence on myogenic processes. One such possibility could be IMAT serving as a local energy reservoir. However, to our knowledge, no functional evidence exists for this idea. In contrast, lipid droplets, called intramyocellular lipids, within muscle fibers can act as an energy reservoir. These organelles are at the core of the “athlete’s paradox,” where trained athletes as well as obese/T2D patients experience high levels of intramyocellular lipids.⁹³ It will be interesting to study a potential relationship between intramyocellular lipids and white adipocytes that make up IMAT. Similarly, we cannot formally exclude that the increase in FAPs we observe in the mFATBLOCK mice might be involved in the functional recovery of muscle. One major limitation to studying IMAT is the lack of suitable murine models. Mice, particularly those with C57/BL6 genetic background, are generally resistant to IMAT formation, limiting the options to study IMAT.^39,94 For example, physiologically relevant models like mdx mice, a model for DMD, do not develop sufficient IMAT. Only when combined with additional genetic alleles, such as ApoE, will mdx mice display extensive IMAT formation.⁶ Therefore, we used GLY, which causes rapid and reproducible IMAT formation,^{21,22,25,41,42} to study the impact of IMAT on muscle regeneration. Our findings are readily translatable to acute injuries like rotator cuff tears, common in individuals without metabolic conditions and prone to IMAT infiltration, making the GLY injury model a suitable tool for this specific context.

RESOURCE AVAILABILITY

Lead contact

Further information and requests for resources and reagents should be directed to the lead contact, Daniel Kopinke (dkopinke@ufl.edu).

Materials availability

This study did not generate new, unique reagents.

Data and code availability

• RNA-seq data have been deposited at NCBI GEO at GEO: GSE292861 and are publicly available as of the date of publication.

• No original code was generated in this study.

• Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

STAR★METHODS

EXPERIMENTAL MODEL AND STUDY PARTICIPANT DETAILS

Mice

The alleles utilized have been previously described. To genetically target fibro-adipogenic progenitors (FAPs), we utilized the Pdgfrα^CreERT2 allele.³⁷ PDGFRA is the “gold-standard” marker in the field to identify FAPs within murine and human muscle.^{5,21,22,25,26,28–30,95} To prevent FAPs from differentiating into adipocytes, we genetically deleted the master regulator of adipogenesis, Pparγ (Pparγ^lox/lox38 Based on our previous work,³⁹ the 129S1/SvlmJ strain has a higher adipogenic potential compared to the C57BL/6J strain. Thus, mFATBLOCK mice were backcrossed onto a 129S1/SvlmJ strain. To induce deletion, 10–14 week old mice, regardless of genotype, received Tamoxifen (TRC, T006000; 200–250 mg/kg) dissolved in corn oil via oral gavage on two consecutive days. Muscle injuries were performed a week after the last tamoxifen treatment to allow metabolic clearance of tamoxifen from mice. Littermates were used for each timepoint and experiment; mice lacking the Cre allele were used as controls. Littermates with the Cre allele and heterozygous or homozygous for the floxed allele were analyzed separately. Upon observation of a haploinsufficiency of the Pparγ gene for IMAT formation, we utilized littermates lacking the Cre allele (Pdgfrα^+/+; Pparγ^fl/+ or Pparγ^fl/fl) as controls, and mice with the Cre allele that are homozygous for the floxed allele (Pdgfrα^CreERT2/+; Pparγ^fl/fl) as the experimental group. Whenever sex was not reported, females mice were used since (a) we failed to detect any meaningful sex differences in both major phenotypes and (b) females had the strongest phenotype. An EchoMRI whole body composition analyzer (EchoMedical Systems) was used to assess body composition by nuclear magnetic resonance. Serum concentrations of insulin, c-peptide, FFA, TG, TC and adipokines were determined through the Mouse Metabolism & Phenotyping core at Baylor College of Medicine. Glucose measurements were taken, via tail vein blood sampling, at fed (9a.m.) and fasting (8hr fast) states using the ONE Touch Ultra 2 Blood Monitoring System. Mice were housed in standard ventilated cages at a controlled temperature (22°C–23°C), with 40–50% humidity and ad libitum access to food and water. All animal work was approved by the Institutional Animal Care and Use Committee (IACUC) of the University of Florida.

Cells

For in vitro studies, all cells were initially plated at 526 cells/mm², unless specified differently, media was changed every other day and were maintained at 37°C in 5% CO₂. The myoblast cell line, C2C12 (ATCC, CRL-1772), was grown in DMEM (Gibco, 11965–092) with 10% FBS (Invitrogen, 10438026) and 1% GlutaMax (Gibco, 35050–061). For myogenic differentiation, cells were allowed to expand for 2 days and then media was switched from 10% FBS to 2% horse serum (Gibco, 26050–070). Differentiation was assessed 2–5 days after induction of differentiation. The pre-adipocyte cell line, 3T3-L1 (Zenbio, SP-L1-F) and primary FAPs were grown in DMEM (Gibco, 11965–092), 10% newborn calf serum (Gibco, 16010–159) and 1% GlutaMax (Gibco, 35050–061). For adipogenic differentiation, cells were allowed to expand for 2 days and induced with media containing DMEM, 10% FBS, 1% GlutaMax, Insulin (0.862 μM; Sigma, I2643) and Troglitazone (5 μM; Sigma, T2573) for 3 days. On the 4^th day, media was switched to maintenance media previously outlined with Insulin but lacking Troglitazone. For primary FAP isolation, WT and mFATBLOCK mice were given tamoxifen through oral gavage and allowed 7 days of washout. TAs and Gastrocnemius muscles were injured with Glycerol (GLY; 50- and 80 μL respectively) and harvested 3 days post injury (dpi). Upon harvesting, muscles were mechanically and enzymatically dissociated in digest media containing DMEM, 1% Pen/Strep, 50mg/mL collagenase IV (Worthington, LS004188) and 6 U/mL Dispase (Gibco, 17105041) for 40–60min at 37°C in constant agitation. The cellular suspension was then filtered through a 70 μm sized pore (Fisher, 22363548), plated in DMEM and allowed 30min for FAPs to adhere at 37°C in 5% CO₂. After, the supernatant was removed and the plate was washed 3 times with warmed PBS, obtaining isolated FAPs and maintenance media was added.⁵⁶ The female C2C12 and male 3T3L1 cell lines were purchased from the respective vendors as indicated above and maintained by careful expansion of early passage stocks in our laboratory. For each experiment, one fresh aliquot of these expanded stocks were used. All cell lines were tested biannually for mycoplasma contamination.

METHOD DETAILS

Muscle injuries

All injuries were performed on adult 10–14 week old mice. Mice were put under anesthesia with isoflurane and the Tibialis Anterior (TA) was injected with 30–50 μL of 50% glycerol (GLY; Acros Organics, 56–81-5) or with 10 μM cardiotoxin (CTX; Naja Pallidum, Lotaxan, L8102–1MG), both diluted in sterile saline.

In situ force production

Functional testing of the Tibialis Anterior muscle (TA) was measured in situ by stimulation of the sciatic nerve as previously described.^96–99 Briefly, Mice were anesthetized with an intraperitoneal injection of xylazine (1 mg/kg) and ketamine (10 mg/kg) and subsequent doses of ketamine were given as needed for maintenance. The TA was isolated from its distal insertion and the distal tendon was secured with a 4–0 silk suture and attached to the lever arm of a force transducer (Cambridge Technology, Model No. 2250). Muscle contractions were induced by stimulating the sciatic nerve with bipolar electrodes using 0.02 ms square wave pulses (Aurora Scientific, Model 701 A stimulator). Data collection and servomotor control were managed using the Lab View-based DMC program (version 615A.v6.0, Aurora Scientific Inc). After determining the optimal muscle length through twitch contractions, a force frequency curve was performed. Isometric contractions were elicited using 500ms train (current 2mA, pulse width 0.2ms) at stimulation frequencies of 1Hz, 15Hz, 30Hz, 45Hz, 60Hz, 75Hz, 100Hz, 125Hz, and 150Hz with one minute of rest between contractions. Peak twitch and tetanic force levels were reported as absolute and specific (normalized to wet muscle weight) force levels. Maximum isometric force was determined from the peak tetanic force elicited from the force-frequency analysis. Given most practical uses of skeletal muscle require an aspect of muscle shortening to generate power, a single after-loaded isotonic contraction was performed against a predetermined, submaximal load. This involved electrically stimulating the nerve with supramaximal voltage (2mA, 0.2ms pulse width, 100ms train duration) at 150Hz allowing the muscle to shorten once it reaches 35% of maximum isometric tension. This allows for the quantification of muscle performance characteristics including force (newtons), displacement (mm), shortening velocity (m/s), and mechanical power (watts). Shortening velocity was calculated as the change in distance (mm) from a 10ms period which began 20ms after the initial length change. Peak power was calculated as the product of the shortening velocity (m/s) and corresponding force (N/kg).

In vitro studies

To test whether FAP genotype influences myogenesis, C2C12 cells were plated, allowed to expand and upon differentiation, primary FAPs from WT or mFATBLOCK females were added, obtained as described above. Cells were fixed in 4% PFA 5 days post induction to evaluate myogenesis. When testing whether FAP density influences myogenesis, C2C12 cells were plated, allowed 2 days to expand and upon differentiation, either 10K (526 cells/mm²) or 20K (1053 cells/mm²) 3T3-L1 cells were added. Cells were then fixed with 4% PFA and myogenesis was assessed.

We evaluated the indirect effects of fat on myogenesis by first culturing and differentiating 3T3-L1 into adipocytes, as described above. Once fully differentiated 7 days post induction, 24hr conditioned media was collected, mixed at a 1:1 ratio with C2C12 induction media and added to C2C12 culture daily. 3 days post induction cells were fixed with 4% PFA and myogenesis was evaluated. We evaluated the direct effects of fat on myogenesis by first culturing and differentiating 3T3-L1 cells into adipocytes, and 10 days post induction, we created a grid on the plate by scratching off adipocytes using a pipet tip, After several washes, C2C12 cells were seeded on top. After 2 days, C2C12 cells were induced to differentiate and myogenesis was assessed 5 days post induction after fixing in 4% PFA.

Histology and immunofluorescence

Upon harvesting, TAs were processed as previously described,⁴⁵ where each TA was cut in two parts: for RNA isolation and histology. For histology, TAs were fixed in 4% PFA (paraformaldehyde) for 2.5hrs at 4°C, washed with PBS and cryopreserved overnight in a 30% sucrose solution. TAs were embedded in OCT-filled cryomolds (Sakura; 4566) and frozen in liquid N₂-cooled isopentane. Sections were obtained with a Leica cryostat, collecting 3–4 sections that were 10 μm thick every 250–350 μm of the TA. For immunofluorescence, sections were blocked and incubated with primary antibodies in blocking solution (5% donkey solution in PBS with 0.3% Triton X-100) overnight at 4°C. Primary antibodies used were rabbit anti-Perilipin (1:1000; Cell Signaling, 9349 S), rabbit anti-MyoG (1:250; Proteintech Group 14688–1-AP), rabbit anti-cleaved Caspase 3 (1:500, Millipore Sigma AB3623), goat anti-PDGFRα (1:250, R&D Systems #AF1062), rabbit anti-Ki67 (Abcam ab15580), chicken anti-GFP (1:1000, Avis lab), rabbit anti-Laminin (1:1000; Sigma-Aldrich, L9393), mouse anti-Myosin heavy chain Type IIA (1:50; DSHB, SC-71), mouse anti-Myosin heavy chain Type IIB (1:50; DSHB, BF-F3), mouse anti-myosin heavy chain (1:250; DSHB, MF20). For the primary antibody rabbit anti-MYH3 (1:250, Proteintech 22287–1-AP), antigen retrieval was required, using a Sodium Citrate Buffer (10 mM Sodium Citrate, 0.05% Tween 20, pH 6.0). Secondary antibodies used were Alexa Fluor-conjugated from Life Technologies (1:1000), as well as directly conjugated dyes Phalloidin-Alexa 568 and 647 (1:200, Molecular Probes #A12380 & A22287) and isolectin (1:200; Invitrogen; I21411), and incubated for 45min at room temperature. Slides were mounted (SouthernBiotech; 0100–01) before imaging. DAPI (Invitrogen, D1306) was used to visualize nuclei. To visualize fibrillar collagen, we performed the histological stain, Sirius Red. To detect lipid droplets in vitro, cells were stained with HCS LipidTOX Green Neutral Lipid Stain (Invitrogen H34475).

Image analysis

Images were acquired with a Leica DMi8 microscope equipped with an SPE confocal and high-resolution color camera. Images of the whole TA were obtained with the navigator function within the Leica LSA software. All images were quantified through ImageJ Software (v1.552p). To quantify percent of GFP⁺ adipocytes, total and proliferating FAPs, FAPs undergoing apoptosis and total MyoG cells, 6–7 random areas of one TA were imaged with a 20× objective and the total number of cells in the areas were averaged per sample.

Gene expression analysis

RNA was isolated as previously described (44). Briefly, muscle was homogenized using a bead beater (TissueLyser LT, Qiagen, 69980) in TRIzol (ThermoFisher Scientific, 15596026). To isolate RNA, chloroform was added, and RNA was purified following the QIAGEN RNeasy kit (74104) per manufacturer’s instructions. 500–800 ng of RNA was used to transcribe into cDNA using the qScript Reverse Transcriptase kit (Quanta bio; 84003). RT-qPCR was performed in technical triplicates or quadruplets on a QuantStudio 6 Flex Real-Time 384-well PCR System (Applied Biosystems, 4485694) using PowerUp SYBR Green Master Mix (ThermoFisher Scientific, A25742). Fold changes were normalized to the house keeping genes Sra1 and Pde12. See Table S1 for qPCR primer details.

For RNAseq, a total of 5 control and 4 mFATBLOCK samples were used. For each sample, RNA was isolated 5 days post GLY injury from the tibialis anterior muscles. RNA samples were sent to Novogene for processing. Briefly, the TruSeq Stranded Total RNA Library Prep Kit (Ilumina) was used to generate the library, which was subsequently sequenced using an Illumina NovaSeq X Plus platform and aligned to the GRCm38 mouse genome. Individual read counts were normalized to the geometric mean read count across all samples using DEseq2. Using a cutoff of p < 0.05 and neglecting genes with less than 100 reads, 258 genes were identified. Raw data are available at NCBI GEO under GSE292861.

Cytokine assay

After injury, TAs were harvested and lysed with a TissueLyser LT (Qiagen, 85600) as previously described, in 1x Phosphate buffered saline (PBS) with proteinase inhibitor (Pierce, Cat #A32953). Once centrifugated, the supernatant was collected, and protein concentration determined through a BCA assay (Pierce, Cat #232225). Samples were then run for an enzyme-linked immunosorbent assay (ELISA; 32-plex mouse cytokine assay, Eve technology).

QUANTIFICATION AND STATISTICAL ANALYSIS

The experimenter was blinded until data were collected. TAs with injury less than 50% of total area were excluded from this study. All data was graphed using GraphPad Prism with data presented as mean ± SEM. For comparing two samples with one variable, an unpaired two-tailed t test or a multiple unpaired two-tailed t test followed by a Holm-Šídák post hoc test was used. For more than two samples with one variable, a one-way ANOVA followed by a Dunnett’s multiple comparison test was used. For two variables, a two-way ANOVA followed by Tukey’s multiple comparison was carried out. A Pearson correlation were used to for all correlation studies. A p value less than 0.05 was considered statistically significant and denoted as follows: *<0.05, **<0.01, ***<0.001 and ****<0.0001.

Supplementary Material

SUPPLEMENTAL INFORMATION

Supplemental information can be found online at https://doi.org/10.1016/j.celrep.2023.113061.

KEY RESOURCES TABLE.

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Rabbit anti-Perilipin	Cell Signaling	Cat #: 9349; RRID: AB_10829911
Rabbit anti-MyoG	Proteintech Group	Cat #: 14688-1-AP
Rabbit anti-cleaved Caspase 3 (cC3)	Millipore Sigma	Cat #: AB3623; RRID: AB_91556
Goat anti-PDGFRα	R&D Systems	Cat #:AF1062; RRID: AB_2236897
Rabbit anti-Ki67	Abcam	Cat #: ab15580; RRID: AB_443209
Chicken anti-GFP	Avis lab	Cat #: GFP-1010; RRID: AB_2307313
Rabbit anti-MYH3	Proteintech	Cat #: 22287-1-AP; RRID: AB_2879060
Rabbit anti-Laminin	Sigma-Aldrich	Cat #: L9393; RRID: AB_477163
Mouse anti-Myosin heavy chain type IIA (2A)	DSHB	Cat #: SC-71; RRID: AB_2147165
Mouse anti-Myosin heavy chain type IIB (2B)	DSHB	Cat #: BF-F3; RRID: AB_2266724
Mouse anti-Myosin heavy chain type I (1)	DSHB	Cat #: BA-D5; RRID: AB_2235587
Phalloidin-Alexa 568	Invitrogen	Cat #: A12381; RRID: AB_2315633
Phalloidin-Alexa 647	Invitrogen	Cat #: A22287; RRID: AB_2620155
DAPI	Invitrogen	Cat #: D1306
HCS LipidTOX Green Neutral Lipid Stain	Invitrogen	Cat #: H34475
Isolectin-488	Invitrogen	Cat #: I21411
Isolectin-647	Invitrogen	Cat #: I32450
Chemicals, peptides, and recombinant proteins
Tamoxifen (TMX)	TRC	Cat #: T006000
Dulbecco’s Modified Eagle’s Medium (DMEM)	Gibco	Cat #: 11965-092
Fetal Bovine Serum (FBS)	Invitrogen	Cat #: 10438026
GlutaMax	Gibco	Cat #: 35050-061
Horse serum (HS)	Gibco	Cat #: 26050-070
Newborn calf serum (NCS)	Gibco	Cat #: 16010-159
Glycerol (GLY)	Acros Organics	Cat #: 56-81-5
Cardiotoxin (CTX)	Lotaxan	Cat #: L8102-1MG
Collagenase IV	Worthington	Cat #: LS004188
Dispase	Gibco	Cat #: 17105041
Insulin	Sigma	Cat #: I2643
Troglitazone	Sigma	Cat #: T2573
Ketamine	Patterson	Cat #: 07-894-8462
Xylazine	Patterson	Cat #: 07-895-0792
TRIzol	ThermoFisher Scientific	Cat #: 15596026
PowerUp SYBR Green Master Mix	ThermoFisher Scientific	Cat #: A25742
Proteinase i nhibitor	Pierce	Cat #: A32953
Critical commercial assays
QIAGEN RNeasy kit	QIAGEN	Cat #: 74104
qScript Reverse Transcriptase kit	Quanta bio	Cat #: 84003
TruSeq Stranded Total RNA Library Prep Kit	Illumina	Cat #: RS-122-2302
BCA assay	Pierce	Cat #: 232225
ELISA; 32-plex mouse cytokine assay	Eve technology	Cat #: MD32
ONE Touch Ultra 2 Blood Monitoring System	ONE Touch	Cat #: B004JC993E
Deposited data
Raw and analyzed data	This study	NCBI GEO: GSE292861
Experimental models: cell lines
Mouse: C2C12	ATCC	Cat #: CRL-1772; RRID: CVCL_0188
Mouse: 3T3-L1	Zenbio	Cat #: SP-L1-F
Experimental models: organisms/strains
Mouse: B6.129S-Pdgfratm1.1(cre,ERT2)Blh/J (PdgfraCreERT2)	Jackson Laboratories	Strain #: 032770; RRID: IMSR_JAX:032770
Mouse: B6.129-Ppargtm2Rev/J (PPARγloxP)	Jackson Laboratories	Strain #: 004584; RRID: IMSR_JAX:004584
Mouse: B6.129X1-Gt(ROSA)26Sortm1(EYFP)Cos/J	Jackson Laboratories	Strain #: 006148; RRID: IMSR_JAX:006148
Mouse: 129S	Jackson Laboratories	Strain #: 002448; RRID: IMSR_JAX:002448
Oligonucleotides
Primers for qRT-PCR (Table S1)	This study	N/A
Software and algorithms
Fiji (ImageJ)	N/A	RRID: SCR_002285
DEseq2	Bioconductor	RRID: SCR_015687
GraphPad Prism 9	GraphPad Software	RRID: SCR_002798
Cellpose	N/A	RRID: SCR_021716
Leica LSA software	Leica	RRID: SCR_013673
Aurora Scientific (v.615A.v6.0)	Aurora Scientific	N/A
Other
70 μm Cell Strainer	Fisher Scientific	Cat #: 22-363-548

Highlights.

• A murine model prevents intramuscular adipose tissue (IMAT) accumulation

• Post-injury, inhibiting IMAT enhances skeletal muscle regeneration

• Mechanistically, IMAT prevents the formation of nascent myofibers

• IMAT also suppresses the subsequent hypertrophic growth of these nascent fibers