The androgen receptor in mesenchymal progenitors regulates skeletal muscle mass via Igf1 expression in male mice

Significance Androgens are essential not only for the development of male sexual characteristics but also for a range of physiological functions, including the regulation of skeletal muscle growth and function. Understanding the functionality of the androgen receptor (AR) is essential for comprehending the mechanisms through which androgens exert their effects on skeletal muscles, as these effects are mediated through AR binding. We demonstrate that AR is expressed in mesenchymal progenitors, which play a vital role in muscle homeostasis, and regulates the expression of insulin-like growth factor 1 (Igf1)—a key player in skeletal muscle growth—to control muscle mass. Our study provides significant insights into potential therapeutic strategies for addressing muscle atrophy conditions like sarcopenia.


Significance statement
Androgens are essential not only for the development of male sexual characteristics but also for a range of physiological functions, including the regulation of skeletal muscle growth and function.Understanding the functionality of the androgen receptor (AR) is essential for comprehending the mechanisms through which androgens exert their effects on skeletal muscles, as these effects are mediated through AR binding.Our study demonstrates that AR is expressed in mesenchymal progenitors, which play a vital role in muscle homeostasis, and regulates the expression of insulin-like growth factor 1 (Igf1)-a key player in skeletal muscle growth-to control muscle mass.Combining androgens and IGF1 treatments may offer potential therapeutic approaches for addressing muscle atrophy conditions such as sarcopenia.

Introduction
Androgens are crucial regulators of various biological processes, including the development and maintenance of male characteristics.Notably, in a physiological context, the administration of androgens or androgen analogs has been shown to exert anabolic effects on muscles, resulting in increased muscle mass, strength, and body weight (1,2).These physiological effects are mediated by activation of the androgen receptor (AR), a member of the nuclear receptor superfamily (3).However, despite that clinical administration of androgens in men can successfully increase skeletal muscle mass (4), the precise cellular and molecular mechanisms underlying the function of AR in muscle remain largely unexplored.
Skeletal muscle represents a complex tissue comprising multiple cell types, and the effects of androgens are likely mediated by various cell populations expressing AR within skeletal muscles.While male mice systemically lacking AR exhibit low skeletal muscle strength and mass, similar to female levels (5), we previously showed that muscle hypertrophy can be induced by androgen administration even in the absence of AR in muscle fibers, suggesting that non-myofiber AR contributes to the regulation of muscle mass (6).Furthermore, another of our studies demonstrated that deletion of AR in satellite cells, which are muscle stem cells responsible for muscle regeneration, had a limited effect on their proliferation and differentiation (7).Additionally, since limb muscle mass remains unaffected even in mice with AR ablation in both myofibers and muscle stem cells (8), other cells expressing AR within the muscles may play a role in regulating skeletal muscle mass.
Mesenchymal progenitors (9), also known as fibro/adipogenic progenitors (10), have been identified as a cell population distinct from myogenic cells characterized by the expression of platelet-derived growth factor receptor alpha (PDGFRα).These cells are involved in the maintenance and regeneration of skeletal muscle (11,12).Conversely, mesenchymal progenitors have been implicated in pathological fibrosis and fat accumulation in skeletal muscle (13).AR expression has been reported in cells located in the interstitium or outside the basal lamina in both mouse and human skeletal muscles (14,15).Additionally, ablation of AR specifically in embryonic cells of the mesenchyme led to impaired development of the perineal levator ani (LA) and bulbocavernosus (BC) muscles (16), which are particularly responsive to androgen.Furthermore, the reduced proliferation of undifferentiated myoblasts in these mutant mice suggested a potential paracrine regulatory role of mesenchymal AR in controlling skeletal muscles (16).Despite this knowledge, the specific role of AR in adult mesenchymal progenitors in skeletal muscles remains unexplored.
Here we investigate the role of AR expressed in mesenchymal progenitors in muscle maintenance using mice with mesenchymal progenitor-specific AR ablation.We show that AR in mesenchymal progenitors plays a critical role in maintaining skeletal muscle mass via regulation of insulin-like growth factor 1 (Igf1) expression in male mice.

AR ablation in mesenchymal progenitors was validated in PDGFRα-CreER;AR L2/Y mice
AR expression in fibroblasts of the perineal LA muscles and in mesenchymal cells of the fetal LA muscles has been reported (14,16,17); however, it is unknown whether mesenchymal progenitors in adult limb muscle express AR.To examine the expression pattern of AR in mesenchymal progenitors, immunofluorescence staining of PDGFRα, a marker of mesenchymal progenitors (9), and AR was performed in adult (12 weeks old) male mice.We found that AR was expressed in mesenchymal progenitors staining positive for PDGFRα (Fig. 1A), as well as in muscle fibers surrounded by laminin-expressing basal lamina, consistent with previous findings (17,18).
To investigate the role of AR in mesenchymal progenitors, PDGFRα-CreER (19) and (20) mice were crossed to selectively delete the AR gene in mesenchymal progenitors expressing PDGFRα.Tamoxifen (TMX) was administered to Cre control male mice (PDGFRα-CreER) and mutant male mice (PDGFRα-CreER;AR L2/Y ) for 5 consecutive days to induce AR ablation (Fig. 1B).To assess the efficiency of AR ablation, immunofluorescence staining of PDGFRα and AR was performed in Cre control and mutant mice.Quantification of PDGFRα and AR expression in forelimb biceps brachii (BB) and hindlimb tibialis anterior (TA) muscles demonstrated a significant decrease of 90% in AR protein in mutant mice compared with Cre control mice, confirming successful ablation of AR expression in mesenchymal progenitors from PDGFRα-CreER;AR L2/Y mice (Fig. 1C).Taken together, these findings demonstrate successful and specific loss of AR expression in PDGFRα+ cells, rendering them an optimal model for investigating AR in mesenchymal progenitors in skeletal muscles.

Lack of AR in mesenchymal progenitors had a limited effect on steady-state limb skeletal muscles
Because ablation of PDGFRα+ cells in intact muscles results in skeletal muscle atrophy (11,12), we first analyzed the muscles of Cre control and mutant mice under steady-state conditions at a young age (14 weeks old) (Fig. S1A).Body weight and limb skeletal muscle mass were not affected in mutant compared with control mice at 2 weeks after TMX administration (Fig. S1B).
Moreover, the grip strength of mutant mice was comparable with that of control mice (Fig. S1C).
The minimum Ferret diameter, which is the robust parameter in order to measure myofiber size, and number of myofibers in BB and TA muscles remained unchanged in mutant mice (Fig. S1D,    E).As relayed signaling between mesenchymal progenitors and muscle stem (satellite) cells is critical for muscle hypertrophy (21), we also quantified the number of muscle stem cells.
Immunofluorescence staining of Pax7, a marker of muscle stem cells (22), revealed no difference in the number of muscle stem cells between control and mutant BB and TA muscles (Fig. S1F).
We next investigated the fiber types of slow soleus muscles, which are affected by muscle fiberand muscle stem cell-specific AR ablation (8,23).Analysis of soleus muscles fiber types by immunofluorescence staining of type I (Myh7), type IIa (Myh2), and type IIx/IIb (without signals) muscle fibers demonstrated no difference in the fiber-type parameters (minimum Ferret diameter and frequency of each fiber-type) between control and mutant mice (Fig. S1G, H).Thus, AR expression loss in mesenchymal progenitors has a limited effect on the steady-state limb muscles of young mice.

The impact of AR deficiency in mesenchymal progenitors on adipogenesis in muscles was limited
We also examined the adipogenesis of mesenchymal progenitors in mutant mice because adipocytes originate from mesenchymal progenitors in skeletal muscles (9).To address this, both the control and mutant mice were subjected to an 8-week regimen of high fat diet (HFD) feeding (Fig. S2A).No significant differences were observed between the control mice and mutant mice in terms of total body weight, hindlimb muscle weights, epididymal white adipose tissue (eWAT) mass, and subcutaneous white adipose tissue (sWAT) mass (Fig. S2B), and grip strength (Fig. S2C).Immunofluorescence staining of perilipin, a marker of adipocytes (24), along with laminin, demonstrated no significant difference in the minimum Ferret diameter of myofibers or the area occupied by adipocytes between the control and mutant TA muscles (Fig. S2D, E).
To further investigate the impact of AR deficiency in muscle adipogenesis, we employed a muscle injury model by administering glycerol injections into the TA muscle of both control and mutant mice (Fig. S2F).No significant differences were observed between the control and mutant mice in terms of the body and TA muscle weights at 2 weeks after injection (Fig. S2G).
Immunofluorescence staining of perilipin in conjunction with laminin (Fig. S2H) revealed no significant difference in the minimum Ferret diameter of myofibers or the area occupied by adipocytes between the control and mutant TA muscles (Fig. S2I).These findings suggest that the absence of AR in mesenchymal progenitors does not significantly affect the adipogenic properties of TA muscles.

AR deficiency in mesenchymal progenitors affected limb skeletal muscles in mature adult but not aged mice
To examine the long-term impact of AR ablation on mesenchymal progenitors, we analyzed mature adult (6 months old) mice and compared the Cre control group and mutant group (Fig. 2A).In the mature adult mutant mice, we observed decreases in body weight, TA, and gastrocnemius muscle weights (Fig. 2B), while there were no significant changes in the weights of eWAT and sWAT (Fig. 2B), which are affected by systemic AR ablation (25).However, grip strength (Fig. 2C) was comparable between the control and mutant mice.
Additionally, the weights of the entire body, hindlimb muscles, and white adipose tissues were compared between aged (28 months old) control and mutant mice (Fig S2J).The aged mutant mice did not exhibit any significant differences from the aged control mice (Fig. S2K and    S2L).Therefore, our findings demonstrate that AR deficiency in mesenchymal progenitors affects skeletal muscle mass in mature adult mice only, not young or aged mice, primarily in the limbs.

Mesenchymal progenitor-specific AR ablation induced atrophy of LA/BC muscles
The perineal LA/BC skeletal muscle complex exhibits increased responsiveness to androgens compared with other striated muscles (26).Because this perineal muscle complex also has higher AR expression (14), we examined the impact of AR ablation on mesenchymal progenitors of LA/BC skeletal muscles.We confirmed efficient ablation of AR in PDGFRα+ cells in LA muscles as well as limb muscles (Fig. 2D, E).At 14 weeks, 6 months, and 28 months of age, the LA/BC mass was consistently lower in the mutant mice than control mice (Fig. 2F).Because AR ablation in myofibers (HSA-Cre;AR L2/Y mice) also leads to reduced LA/BC mass (27), we generated HSA-Cre;PDGFRα-CreER;AR L2/Y mice to examine whether the weight reduction depended on AR in myofibers.Interestingly, the LA/BC weight was lower in HSA-Cre;AR L2/Y mice than in PDGFRα-CreER;AR L2/Y mice (Fig. 2G).Moreover, HSA-Cre;PDGFRα-CreER;AR L2/Y mice had the lowest LA/BC mass compared with HSA-Cre;AR L2/Y and PDGFRα-CreER;AR L2/Y mice (Fig. 2G).These findings suggest that AR expression in both myofibers and mesenchymal progenitors independently regulates the mass of perineal muscles.We also found that AR depletion in mesenchymal progenitors in LA muscles led to a decrease in the number of PDGFRα+ cells (Fig. 2H).Examination of the cross-sectional area of LA muscles by laminin staining (Fig. 2I) revealed a lower myofiber size in the mutant mice than control mice (Fig. 2J).
In summary, AR expression in mesenchymal progenitors independently regulates the mass of perineal muscles, separate from its influence on myofibers.

AR regulated genes related to extracellular matrix organization in mesenchymal progenitors
To determine the effects of AR on gene expression, we performed RNA-seq in PDGFRα+ cells from the LA/BC muscles of both control and mutant mice (Fig. 3A).Using a previously established gating strategy for isolating mesenchymal progenitors (CD45− CD31− Vcam1− PDGFRα+) (12)(Fig.S3A), we isolated 300 PDGFRα+ cells from LA/BC muscles for RNA-seq.
We detected more than 100,000 genes in these cells, indicating sufficient read depth for further analysis (Fig. S3B).The PCA plot showed separation between the control and mutant samples (Fig. S3C).We identified 80 upregulated and 76 downregulated genes, including AR, in PDGFRα+ cells from the mutant mice (Fig. S3D).The upregulated genes were enriched in the Gene Ontology (GO) terms including regulation of cell death and cell motility (Fig. 3B).On the other hand, the downregulated genes were enriched in the GO terms related to extracellular matrix organization and regulation of proteolysis (Fig. 3C).
To determine the AR target genes, we used CUT&RUN method for AR and trimethylated lysine 4 from histone H3 (H3K4me3), which is enrich in active promoters, in mesenchymal progenitors isolated from C57BL/6J male mice treated with DHT 2 hours before sampling (Fig 3D).We detected total 6542 peaks of AR in mesenchymal progenitors in skeletal muscles.The AR peaks were primarily distributed in intron or intergenic regions, similar to previous reports with AR ChIP-seq in muscle and non-muscle tissues (28,29).We also found the androgen response elements (ARE) in these peaks as the most enriched motif as well as AR half-site (Fig. 3E, S3F).Moreover, AR CUT&RUN was validated by peaks found in the Fkbp5 locus (Fig. S3G), where canonical AR binding sites are located for several tissues (28).By profiling AR peaks with AR and H3K4me3 reads, we found that AR binding sites were present mainly not in promoter region (Fig. 3F).Furthermore, AR binding events were enriched for regulation of locomotion and extracellular matrix organization GO terms (Fig. 3G).These results indicate that AR in mesenchymal progenitors directly regulate the transcriptions including extracellular matrix organization genes.

AR in mesenchymal progenitors regulated Igf1 expression
To further investigate the gene regulation by AR in mesenchymal progenitors, we combined the RNA-seq and CUT&RUN analysis.We found 39 genes which were downregulated in ARdeficient mesenchymal progenitors and have AR peaks annotated by a nearest transcription start sites (TSS) analysis (Fig. 4A).Top enriched GO term of these genes was extracellular matrix organization (Fig. 4B), including Collagen family (Col1a1, Col3a1, Col4a1, Col4a2, Col5a1, Col5a2, and Col6a2, Fig. S4A and S4B), and Igf1 (Fig. 4C), which is critical for maintaining muscle mass (30).We also found the bindings of AR with ARE and peaks of H3K4me3 in upstream of Igf1 TSS (Fig. 4D).Moreover, AR peaks were found with ARE in Ar gene locus (Fig. S4C), indicating the regulation of Ar by AR in mesenchymal progenitors as other cells (31).
To further investigate the effects of androgen and AR on Igf1 expression, we isolated PDGFRα+ cells from C57BL/6J male mice treated with DHT (Fig. 4E).Igf1 expression was upregulated in PDGFRα+ cells after DHT administration (Fig. 4F).We also found reduction of IGF1 protein in the mutant LA muscles (Fig. 4G and 4H).Taken together, AR in mesenchymal progenitors regulates Igf1 expression controlling skeletal muscle mass.

Discussion
In the present study, we demonstrated that mice with specific AR ablation in mesenchymal progenitors showed lower skeletal muscle mass in the limbs only at the mature adult stage, but not at a younger stage (14 weeks old).Taken together with previous reports demonstrating that mice with AR ablation specifically in fast-twitch fibers lack muscle mass reduction at 13 weeks old but exhibit this reduction at 12-13 months old (23), there may be a potential age-dependent role of AR or an effect that requires time to develop.Additionally, we demonstrated that the number of Pax7+ muscle stem cells was similar between adult control and mutant mice.This is in contrast to the observed reduction in myoblast proliferation following deletion of AR in the mesenchyme of BC muscles during development (16), providing evidence that AR in mesenchymal progenitors plays distinct roles during development, adulthood, and aging in mice.
We also showed that specific ablation of AR in mesenchymal progenitors does not affect the fiber types of slow-twitch muscles.Depletion of mesenchymal progenitors results in a slight increase in the proportion of slow fibers (type I) predominantly in fast-twitch TA muscles (12), suggesting that factors other than AR in mesenchymal progenitors influence muscle fiber types.
PDGFRα+ cells are the origin of ectopic adipocytes in skeletal muscles (9,10).However, the specific molecular mechanisms underlying fatty infiltration of muscles remain largely unexplored (32).In general, androgen and AR signaling seem to have inhibitory effects on adipogenesis (33).Serum androgens are linked to reduced levels of fatty infiltration of skeletal muscles (34).In contrast, deprivation of androgens leads to an increase in fatty infiltration of skeletal muscle in men (35).Moreover, transgenic mice expressing AR induced by collagen type I alpha 1, which is also expressed in mesenchymal progenitors (36), have reduced body weight and fat mass (37).Additionally, HFD-induced fat mass deposition in skeletal muscles was increased by a low androgen level caused by orchiedectomy (38).However, our results indicate that AR deficiency in mesenchymal progenitors does not affect adipogenesis in skeletal muscle even after HFD feeding or glycerol injection.An indirect androgen pathway may also regulate adipogenesis of mesenchymal progenitors in skeletal muscles.Indeed, recent research indicates that androgens can modulate macrophage polarization, subsequently affecting adipocyte differentiation (39).Further studies are needed to gain a comprehensive understanding of the mechanisms by which androgens and AR signaling influence fatty infiltration of muscle.
We observed a modest reduction in the muscle weight of the mutant mice at the mature adult age (24 weeks old), but this effect was not observed in aged mice (28 months old).Because selective ablation of AR in myofibers does not lead to reduced muscle mass even in 40-week-old mice (27), the role of AR in mesenchymal progenitors is more important than its role in myofibers in the context of muscle maintenance, at least at the mature adult age.A decrease in testosterone concentration, even in 1-year-old mice compared with young 12-week-old mice (40), may contribute to muscle mass loss, potentially masking any subtle differences between the aged control and mutant mice.Due to the reduction in muscle mass observed in 2-year-old mice with AR depletion specifically in fast-twitch fibers (23), AR expression in muscle fibers may play a crucial role in regulating muscle metabolism and quality, which help maintain muscle mass in aged mice.
The perineal skeletal muscles LA and BC are highly androgen sensitive.Muscle fiberspecific AR knockout results in a more pronounced mass reduction in LA/BC muscles compared with limb muscles (26,27), which is consistent with our present study demonstrating a consistent reduction in LA/BC muscles in mice with AR ablation in mesenchymal progenitors.We also demonstrated that AR deficiency in mesenchymal progenitors resulted in reduced LA/BC weight, although the extent of the reduction was milder compared with mice in which AR was ablated in muscle fibers.Furthermore, the double Cre mice (HSA-Cre;PDGFRα-CreER;AR L2/Y ), with ablation of AR in both myofibers and mesenchymal progenitors, exhibited a greater reduction in LA/BC mass compared with single Cre mice (HSA-Cre;AR L2/Y or PDGFRα-CreER;AR L2/Y ).Taken together, AR expression in mesenchymal progenitors independently regulates the mass of perineal muscles, separate from the influence of AR in myofibers.The distinct reactions observed in perineal and limb muscles could be attributed to a greater abundance of myonuclei expressing AR in perineal muscles compared with limb muscles (14).
Nevertheless, our findings revealed that the majority of PDGFRα+ cells exhibited AR expression in both limb and perineal muscles, suggesting an additional mechanism by which AR in mesenchymal progenitors regulates muscle mass and other mesenchymal progenitor functions.
The signaling pathway involving IGF1 plays a crucial role in both the developmental growth and steady-state maintenance of skeletal muscle mass (41,42).IGF1 is produced primarily in the liver (43) but local production, including in skeletal muscles, is induced by androgen treatment.The administration of testosterone to older men led to an increase in IGF1 protein expression in the limb (vastus lateralis) muscles (44).Igf1 expression is decreased in the LA/BC muscles of mice with skeletal muscle-specific AR ablation (27).Dubois et al. has reported indirect androgen regulation of IGF1 via non myocytic AR because Myod-Cre-driven AR-deficient mice showed no reduction in the Igf1 transcript level, and Igf1 expression was decreased after castration in skeletal muscles (17).We demonstrated that AR in mesenchymal progenitors in skeletal muscles regulated Igf1 expression, governing muscle mass.Moreover, activation of Akt signaling via the expression of Bmp3b in mesenchymal progenitors (12) suggests a connection with Igf1, highlighting the complex interplay of signaling pathways in the maintenance of skeletal muscle mass.
In summary, our results show that AR in mesenchymal progenitors regulates skeletal muscle mass via Igf1 expression.Combining treatment with androgens and IGF1 could be useful to treat muscle atrophy such as sarcopenia.

Animals
PDGFRα-CreER mice (19), HSA-Cre mice (45), and C57BL/6J male mice were purchased from Jackson Laboratories (Strain #: 018280, 006149, and 000664, respectively).AR L2/Y mice have been described (20).PDGFRα-CreER male mice were crossed with AR L2/+ female mice to generate PDGFRα-CreER;AR L2/Y male mutant mice.Male littermates with the PDGFRα-CreER;AR +/Y genotype were used as control mice.To generate HSA-Cre;PDGFRα-CreER;AR L2/Y male mice, HSA-Cre male mice were crossed with PDGFRα-CreER;AR L2/+ female mice.All mice were housed in a specific pathogen-free facility under climate-controlled conditions and a 12-hour light/dark cycle and were provided water and a standard diet ad libitum.In both control and mutant mice, 100 µL of 20 mg/mL tamoxifen (TMX, Sigma-Aldrich, Cat# T5648), dissolved in corn oil (Sigma-Aldrich, Cat# C8267), was injected intraperitoneally for 5 consecutive days at 8 or 12 weeks of age.A high-fat diet (HFD, CLEA Japan, Cat# HFD32) was given ad libitum for 8 weeks to induce obesity in mice.The diet was refreshed every 2 days.To induce fatty infiltration of skeletal muscle, 100 µL of 50% v/v glycerol was injected into tibialis anterior (TA) muscles under anesthesia using isoflurane (46).To induce AR signaling, dihydrotestosterone (DHT or stanolone, 2.5 mg/mouse; Tokyo Chemical Industry, Cat# A0462), dissolved in absolute ethanol and diluted in 0.3% w/v hydroxypropyl cellulose (Wako, Cat# 085-07932) in PBS, was injected subcutaneously at 2 hours before sampling.All animal experiments were approved by the Animal Experiment Committee of Ehime University (Approval No. 37A1-1•16) and were conducted in full compliance with the Guidelines for Animal Experiments at Ehime University.

Grip strength
The grip strength of the forelimb was measured using a strain gauge (Melquest, Cat# GPM-100B).The maximal force was determined after 10 measurements.The mean value recorded by two independent experimenters was reported as the grip strength of each mouse.

Immunofluorescence staining and microscopy
For immunofluorescence staining, muscles were snap frozen in liquid nitrogen-chilled isopentane.Cryosections of 10 µm thickness were prepared for immunofluorescence staining.

Quantification of muscle sections
To quantify the number, cross-sectional area (CSA), and minimum Ferret diameter of the myofibers, captured images of entire cross sections were analyzed using Fiji (v2.3.0)(https://imagej.net/Fiji).The myofiber area was identified by the laminin+ signals.The region of interest (ROI) was set within the laminin+ areas.The frequency of each muscle fiber type in the ROI was recorded by Trainable Weka Segmentation in Fiji after manual training.Small (< 150 µm 2 ) or large (> 10,000 µm 2 ) fibers were excluded for analyzing.The number of AR+PDGFRα+ mesenchymal progenitors, Pax7+ satellite cells, and the perilipin+ area was manually counted or measured using Fiji.

RNA-seq and data analysis
For isolation of RNA from mesenchymal progenitors in LA/BC muscles, 300 PDGFRα+ cells were sorted into the wells of a 96-well plate containing 10 µL lysis buffer at 4°C.The lysis buffer was prepared by mixing 1 µL RNase inhibitor (TaKaRa, Cat# 2313A) and 19 µL 10X lysis buffer (TaKaRa, Cat# 635013) diluted to 1X with nuclease-free water.3' UTR RNA-seq was conducted using the CEL-Seq2 protocol (48), with the exception of employing the Second Strand Synthesis Module (New England Biolabs, Cat# E6111L) for double-stranded cDNA synthesis.The library was then amplified via 10 cycles of PCR without any sample pooling.Subsequently, the library was sequenced on the Illumina NovaSeq 6000, and quantitative analysis was performed using 81 bp of insert reads (Read2).

RNA isolation and quantitative real-time PCR
Total RNA was extracted from freshly sorted 30,000 PDGFRα+ cells from LA/BC muscles using the RNeasy Plus Micro Kit (Qiagen, Cat# 74034) following the manufacturer's protocol.
Reverse transcription was performed using PrimeScript (Takara, Cat# RR036A) to synthesize cDNA from total RNA.Quantitative PCR (qPCR) was conducted in technical duplicate samples utilizing TB Green Premix Ex Taq II (Takara, Cat# RR820S) and the Thermal Cycler Dice (Takara, Cat# TP850).The expression levels of the target genes were normalized to that of Rpl13a.The primer sequences are listed in Table S1.

Statistical analysis
Prism 9 (GraphPad Software) or R (v4.0.5) (https://www.r-project.org) was used for the statistical analysis.Welch's t-test was used to compare two groups.The two-sample test for equality of proportions was performed using the test of equal or given proportions.To compare variables among three or more groups, Brown-Forsythe and Welch's ANOVA test with Dunnett's T3 multiple comparison test was used.

Figure 3 .Figure 4 .
Figure 3. Functional and genomic targets of AR in mesenchymal progenitors of skeletal