M-cadherin regulates the formation of adherens junctions and secondary myofibers during fetal myogenesis to determine adult myofiber number and muscle mass

Summary

Cadherin-mediated adhesion of skeletal muscle stem cells within their niche is necessary for normal adult muscle maintenance and regeneration; however, the role of cadherins in regulating muscle development and growth has not yet been elucidated. Here, we show that M-cadherin protein is localized to adherens junctions in developing muscle and that deletion of Cdh15, the gene encoding M-cadherin, results in the loss of adherens junctions in fetal mouse muscle. This loss of adherens junctions is associated with reduced secondary myofiber formation, ultimately resulting in reduced adult myofiber number and muscle mass. Concordantly, via large-scale exome sequencing, we found that humans with predicted loss-of-function variants in the CDH15 gene had significantly reduced lean mass, indicating that M-cadherin functions to regulate muscle mass in both mice and humans. Overall, these data highlight a previously unrecognized role of M-cadherin in controlling fetal myofiber formation and establishment of adult muscle mass.

Highlights

• •Predicted loss-of-function variants in CDH15 are associated with reduced lean mass
• •Loss of M-cadherin in CDH15^KO mice leads to reduced myofiber number and muscle mass
• •M-cadherin is localized to adherens junctions in developing muscle
• •CDH15^KO fetal muscle lacks adherens junctions and has fewer secondary myofibers

Golann and colleagues found that predicted loss-of-function variants in the CDH15 gene (encoding M-cadherin) were associated with reduced lean mass. Similarly, CDH15^KO mice displayed reduced muscle mass, which was associated with disrupted adherens junctions and impaired secondary myofiber formation during fetal myogenesis. These data implicate M-cadherin as a key regulator of fetal muscle development and establishment of adult muscle mass.

Introduction

The formation and development of skeletal muscle is a highly coordinated process in which myogenic progenitor cells migrate, proliferate, align, differentiate, and fuse to create nascent multinucleated myofibers, ultimately forming somatic muscle comprising ∼40% of adult body mass. Vertebrate skeletal muscle development is initiated by the delamination and migration of progenitor cells from the dermomyotome to positions that eventually develop into trunk and limb musculature (reviewed in Chal and Pourquie, 2017). Upon cessation of migration, myotomal cells express myogenic regulatory factors that induce terminal differentiation into mononucleated myocytes before fusing into syncytial multinucleated myotubes. Myotubes mature into primary myofibers of the myotome during embryonic myogenesis, subsequently serving as cellular scaffolds for secondary myofibers during fetal myogenesis to produce the bulk of fetal muscle. While transcription factors and cellular signaling pathways directing progenitor migration and positioning (e.g., Lbx1 and HGF/c-met) and myogenic regulatory factors mediating commitment and differentiation (e.g., Myf5, MyoD, and myogenin) have been well described (reviewed in Buckingham, 2001, 2017; Chal and Pourquie, 2017), the mechanisms by which primary myofibers physically guide secondary myofiber formation during fetal muscle development have not yet been elucidated.

Given that primary myofibers are in direct contact with myogenic progenitor cells during fetal muscle development, it is likely that cell-cell contact between primary myofibers and progenitor cells may mediate cellular signaling or provide positional cues aiding in the formation of secondary myofibers. Cell-cell contact is essential for proper embryonic development with cadherin-mediated adhesion playing critical roles in maintaining physical contact of cells and regulating development of many tissues. Cadherins are calcium-dependent homophilic adhesion proteins that regulate contact-dependent signaling and the formation and maintenance of adherens junctions via intercellular homotypic interactions. During development, cadherins are required for the proper formation of multiple tissues—for example, N-cadherin is required for proper cortical structure organization in the brain (Kadowaki et al., 2007), as well as proper cardiomyocyte adhesion and establishing left-right symmetry of the heart (Garcia-Castro et al., 2000; Radice et al., 1997). Developing skeletal muscle expresses M-, N-, and R-cadherin (Kaufmann et al., 1999), and certain cadherins have been shown to mediate developmental and regenerative myogenesis in vivo and to regulate myogenesis in vitro (Cortes et al., 2003; Goel et al., 2017; Redfield et al., 1997; Wrobel et al., 2007; Zeschnigk et al., 1995). Interestingly, despite M-cadherin (i.e., muscle cadherin) being named for its high expression in muscle tissue, it is dispensable for muscle development and maintenance in vivo (Hollnagel et al., 2002), potentially due to compensation by additional cadherins also expressed in muscle. It has been shown that cadherin-based adherens junctions are necessary for maintaining muscle stem cell quiescence and are critical for niche retention in adult muscle, and more recently, quiescence maintenance by cadherins has been specifically attributed to N-cadherin and not M-cadherin (Goel et al., 2017; Hung et al., 2024; Kann et al., 2022). However, M-cadherin may still play a role in regulating myogenic processes in different contexts, namely during prenatal development. It has previously been shown that M-cadherin is highly expressed during fetal myogenesis and accumulates at sites of cell-cell contact between primary and secondary myofibers, suggesting a potential role of M-cadherin in regulating secondary myofiber formation during fetal development (Cifuentes-Diaz et al., 1995; Rose et al., 1994).

Here, we sought to examine the role of M-cadherin in regulating skeletal muscle development and maintenance. Human genetic evidence supported M-cadherin-mediated regulation of muscle size, as individuals with predicted loss-of-function (pLoF) variants in the CDH15 gene, which encodes M-cadherin, display moderately reduced lean mass. To follow up on these findings, we generated Cdh15 knockout mice (referred to as CDH15^KO), which similarly exhibited reduced skeletal muscle mass, a phenotype rooted in reduced myofiber number during prenatal development. M-cadherin protein is highly abundant at skeletal muscle cell junctions during fetal development in wild-type mice, and deletion of Cdh15 disrupts adherens junctions and reduces secondary myofiber formation. Overall, we have identified a key role of M-cadherin in regulating skeletal muscle development, which is supported by human genetic data.

Results

Predicted loss-of-function variants in the CDH15 gene are associated with reduced lean mass in humans, and Cdh15 deletion is associated with reduced muscle mass in mice

CDH15, the gene that encodes for M-cadherin, is known to be highly expressed in human skeletal muscle tissue and in the cerebellum. While loss-of-function (LoF) variants in the CDH15 gene have been reported to cause intellectual disability (Bhalla et al., 2008), no associations of CDH15 genotype with skeletal muscle phenotypes have been described in the literature to date. We analyzed large-scale exome sequencing and bioelectric impedance analysis-derived lean mass measurements from 460,832 individuals in UK Biobank (UKB) and found that the burden of rare, pLoF and deleterious missense variants in the CDH15 gene was associated with lower whole-body lean mass in adult humans (2.6 kg lower lean mass, p = 0.008), indicating that M-cadherin is a potential regulator of skeletal muscle mass (Figure 1A; full list of included variants is provided in Data S1). To further explore the role of M-cadherin in the development and maintenance of muscle mass, we generated Cdh15-knockout mice, referred to as CDH15^KO, via in-frame insertion of a LacZ reporter cassette to disrupt the transcription of Cdh15 (see methods). LacZ staining of whole E12.5 embryos confirmed the expected M-cadherin expression pattern in developing skeletal muscles (Figure 1B), as well as strong signal throughout adult hindlimb muscles (Figure 1C). Interestingly, we also detected LacZ signal within adult bone cells (Figure S1A), though Cdh15 deletion did not appear to alter markers of bone health, including bone volume, density, or bone mineral content (Figures S1B–S1D).

Figure 1.

CDH15 loss of function is associated with reduced lean mass in humans, and CDH15^KO mice display reduced lean mass and muscle mass

(A) Genetic association between the burden of singleton CDH15 pLoF and deleterious missense variants and lean mass derived from bioelectric impedance in UK Biobank. Missense variants were only included in this analysis if they were deemed to have a high likelihood of deleteriousness. The “participants Ref/Het/Hom” column indicates the number of individuals with the wild-type homozygous, alternative allele heterozygous, or alternative allele homozygous genotypes. A full list of CDH15 pLoF and missense variants included in this burden analysis is provided in Data S1. BIA, bioelectrical impedance analysis; AAF, alternative allele frequency; SD, standard deviation units; CI, confidence interval.

(B and C) LacZ visualization of Cdh15 expression via β-galactosidase staining in E12.5 embryos and adult hindlimb muscle from CDH15^KO mice.

(D) RT-qPCR analysis of Cdh15 gene expression in the gastrocnemius (gastroc.), tibialis anterior (TA), extensor digitorum longus (EDL), and soleus from wild-type and CDH15^KO mice (n = 6 WT mice and n = 7 CDH15^KO mice).

(E) Total body lean mass of 3- to 6-month-old adult mice assessed by echoMRI (n = 16 WT mice and n = 17 CDH15^KO mice).

(F–I) Hindlimb muscle weights of adult mice (n = 6 WT mice and n = 6 CDH15^KO mice). Data are represented as mean ± SEM. ^∗p < 0.05; ^∗∗p < 0.01; ^∗∗∗p < 0.001; ^∗∗∗∗p < 0.0001.

Cdh15 deletion was confirmed in this model via gene expression analysis in multiple adult limb muscles (Figure 1D). Similar to our human genetic findings, adult CDH15^KO mice also displayed a mild reduction in whole-body lean mass (Figure 1E) when compared to wild-type mice that could not be explained by changes in total body weight or fat mass (Figures S1E and S1F). Instead, this reduction in lean mass was found to be driven by reduced muscle mass of multiple hindlimb muscles (by an average of ∼15%) (Figures 1F–1I), while the weight of other organs remained similar to wild-type mice (Figures S1G–S1I). Altogether, these data suggest that M-cadherin plays a role in regulating skeletal muscle mass in both humans and mice.

Reduced muscle mass of adult CDH15^KO mice is due to impaired myofiber formation during development

To understand the mechanism by which loss of M-cadherin mediates reduced muscle size, we performed histological and molecular profiling of adult CDH15^KO muscles. Interestingly, we observed no differences in myofiber cross-sectional area (CSA) or the number of myonuclei per myofiber in the tibialis anterior (TA) muscle compared to wild-type mice (Figure 2A). Additionally, RNA sequencing (RNA-seq) profiling of the TA and gastrocnemius muscles revealed a very low number of differentially regulated genes (<10) between adult CDH15^KO and wild-type muscle, with the number of differentially regulated genes being similar to that of other organs which do not express Cdh15 (e.g., heart, kidney, liver, lung, and spleen) (Figure 2B; Data S2). We also found that the overall muscle function of CDH15^KO mice appeared normal when assessed via wire hang testing and voluntary wheel running distance and speed, further supporting that reduced muscle mass is not due to atrophy or degeneration in CDH15^KO muscle (Figures S2A–S2C).

Figure 2.

Postnatal CDH15^KO mice have a reduction in myofiber number but not myofiber size

(A) Representative images and quantification of average myofiber size (CSA, cross-sectional area) and the average number of myonuclei per type IIb fiber in adult mice (n = 4 WT mice and n = 4 CDH15^KO mice).

(B) Stacked bar graph depicting the number of differentially expressed genes (DEGs) among different tissues determined by RNA-seq between adult wild-type and CDH15^KO mice (n = 4–5 WT mice and n = 5 CDH15^KO mice).

(C) Representative images of cross-sections of adult TA muscle and quantification of the proportion of fiber types and total number of myofibers within the whole TA muscle (n = 4 WT mice and n = 4 CDH15^KO mice).

(D) Representative images of cross sections of adult soleus muscle and quantification of the proportion of fiber types and total number of myofibers within the whole soleus muscle (n = 6 WT mice and n = 6 CDH15^KO mice).

(E) Representative images of cross sections of P7 lower hindlimb muscles and quantification of average myofiber size and number in the TA and soleus muscles. TA muscle is outlined with a yellow tracing, and soleus is outlined with a white tracing in representative images (n = 3 WT mice and n = 3 CDH15^KO mice). Data are represented as mean ± SEM. ^∗p < 0.05; ^∗∗p < 0.01; ^∗∗∗p < 0.001.

Given that the reduced muscle size in adult CDH15^KO mice could not be explained by myofiber size or activation of atrophic or degenerative pathways, we examined whether alterations in myofiber number could be the cause for reduced muscle size. We found that fast-twitch TA muscle, as well as slow-twitch soleus muscle, from adult CDH15^KO mice had a significantly lower number of myofibers (by ∼16%) compared to wild-type mice (Figures 2C and 2D). This reduction in myofiber number was not due to preferential loss of any specific myofiber type, as proportions of oxidative and glycolytic myofiber subtypes remained similar between wild-type and CDH15^KO muscles (Figures 2C and 2D). Myofiber number is determined during development, with the total myofiber pool being fully established by postnatal day 7 (P7) in mice (Li et al., 2015). In line with adult CDH15^KO muscles, myofiber size was also unaffected as early as P7, and total myofiber number was reduced by ∼15%–30% compared to wild-type muscles (Figure 2E). Altogether, these data indicate that reduced muscle mass in CDH15^KO mice is driven by a reduction in myofiber number, suggesting that M-cadherin plays an important role in regulating myofiber formation during development.

M-cadherin is prevalent at myofiber-myogenic cell-cell junctions but is not required for normal myogenic lineage progression during fetal development

Myofiber formation occurs in two distinct phases during prenatal development: primary and secondary myogenesis. During primary myogenesis (∼E12–E15 in mice), Pax3⁺ myogenic progenitor cells differentiate and fuse to form primary myofibers that act as a scaffold for further rounds of progenitor fusion and myofiber formation. Secondary myogenesis occurs between E15 and E17.5, where myogenic progenitors begin to express Pax7 and fuse either with the primary myofibers or use the primary myofibers as a physical support to align, differentiate, and fuse with each other to form new secondary myofibers. Previous reports have noted the presence of M-cadherin protein at cell-cell junctions during fetal myogenesis (Cifuentes-Diaz et al., 1995; Rose et al., 1994), and here, we also noted distinct localization of M-cadherin between myofibers and at the surface of Pax7⁺ progenitor cells during fetal myogenesis (Figure 3A). While cadherin-based adhesion modulates Pax7⁺ cell number, niche retention, and proliferative capacity in adult mice, we sought to assess any effects of M-cadherin loss in prenatal myogenesis. The total number of Pax7⁺ cells during early (E15) and late (E17.5) stages of fetal myogenesis did not differ between wild-type and CDH15^KO extensor digitorum longus (EDL) muscles (Figure 3A), and the rate of proliferation, assessed by Ki67 staining, was also unchanged (Figure 3B). Additionally, the loss of M-cadherin did not affect the ability of Pax7⁺ progenitors to further expand and contribute to the postnatal pool of muscle stem cells, as wild-type and CDH15^KO EDL muscles contained similar numbers of Pax7⁺ muscle stem cells at P7 (Figure S3A), and the majority of these cells were localized to the myofiber sublaminar space, indicating proper niche entry (Figure S3B).

Figure 3.

M-cadherin is not required for normal myogenic lineage progression during development

(A) Representative images and quantification of Pax7⁺ progenitor cells in EDL muscle during E15 and E17.5 developmental time points (n = 4 WT mice and n = 3–4 CDH15^KO mice).

(B) Representative images and quantification of the proportion of proliferating (Ki67⁺) Pax7⁺ progenitor cells in EDL muscle at E17.5 (n = 4 WT mice and n = 4 CDH15^KO mice).

(C) Representative images and quantification of myogenin⁺ cells in EDL muscle at E15 and E17.5 (n = 4 WT mice and n = 3–4 CDH15^KO mice).

(D) RT-qPCR analysis of Myomixer and Myomaker gene expression in E15 hindlimb muscle (n = 4 WT mice and n = 3 CDH15^KO mice). Data are represented as mean ± SEM.

While the loss of M-cadherin did not appear to affect the proliferative ability of Pax7⁺ progenitors during fetal development, the ability of muscle progenitor cells to differentiate and fuse to form myofibers may have been compromised, leading to reduced myofiber number. Immunostaining for M-cadherin revealed distinct surface localization on committing/differentiating myogenin⁺ cells during fetal development in wild-type muscles (Figure 3C). However, there was no difference in the total number of myogenin⁺ cells between wild-type and CDH15^KO muscles at E15 or E17.5 (Figure 3C), indicating normal myogenic lineage progression. Additionally, there was no difference in expression of the fusogenic genes Myomixer and Myomaker at E15, demonstrating that myogenic cell fusion remained unaltered during fetal development (Figure 3D). Overall, these data indicate that while M-cadherin protein is localized at myogenic progenitor cell junctions during fetal development, its absence does not impair myogenic progenitor proliferation, differentiation, or fusion during the formation of myofibers.

M-cadherin is required for the establishment of adherens junctions and formation of secondary myofibers during fetal myogenesis

Since alterations in myogenic lineage progression could not account for reduced myofiber number in CDH15^KO muscle, we hypothesized that M-cadherin localization at myofiber-myofiber or myofiber-myogenic progenitor cell junctions may regulate myofiber formation via M-cadherin’s role in mediating cell-cell adhesion. Surprisingly, when we stained muscles with a pan-cadherin antibody that recognizes multiple type I cadherins (including N- and R-cadherins, previously shown to be expressed early during muscle development; Cifuentes-Diaz et al., 1994; Rosenberg et al., 1997), we found that there was a notable absence of cadherin proteins at the myofiber surface in the EDL of CDH15^KO muscle at E15 and E17.5 (Figure 4A). Additionally, staining for non-phosphorylated (i.e., stable) and total β-catenin, the intracellular partner of membrane-localized cadherins, revealed that there was essentially a complete loss of membrane-localized cytoplasmic β-catenin (except small areas with interstitial expression of β-catenin, likely in non-myogenic cells) (Figures 4B and S4). Interestingly, the loss in cytoplasmic β-catenin in CDH15^KO myofibers/progenitors was not due to nuclear accumulation, and expression analysis for canonical β-catenin genes (e.g., Gsk3b, Axin2, Dkk1, and Sp5) did not show any change between wild-type and CDH15^KO muscle at E15 (Figure S5). Thus, it appears that M-cadherin’s presence at the surface of myofibers and myogenic cells is necessary for proper cell-cell adhesion during fetal development, as loss of its expression fully disrupts the assembly of adherens junctions, evidenced by loss of membrane-localized β-catenin.

Figure 4.

M-cadherin is required for the formation of adherens junctions and secondary myofibers during fetal development

(A) Representative images and quantification of the area of the EDL muscle containing positive signal for pan-cadherin at E15 and E17.5 (n = 3–4 WT mice and n = 3–4 CDH15^KO mice).

(B) Representative images and quantification of the area of the EDL muscle containing positive signal for non-phosphorylated β-catenin at E15 and E17.5 (n = 3–4 WT mice and n = 3–4 CDH15^KO mice).

(C) Representative images and quantification of average myofiber size of the EDL muscle and the number of primary and secondary myofibers comprising the whole EDL muscle at E17.5 (n = 4 WT mice and n = 4 CDH15^KO mice). Asterisks denote WT primary myofibers with multiple adjacent secondary myofibers or CDH15^KO primary myofibers with 0–1 adjacent secondary myofibers. Data are represented as mean ± SEM. ^∗p < 0.05; ^∗∗∗p < 0.001; ^∗∗∗∗p < 0.0001.

Adherens junctions potentially play a key role in regulating the formation of secondary myofibers during fetal development, similar to previously described roles of specific adhesion molecules (e.g., NCAM) that are spatially concentrated between primary and secondary myofibers (Fazeli et al., 1996). To determine whether disrupted adherens junctions caused by Cdh15 deletion may lead to altered secondary myogenesis, we stained E17.5 fetal muscles with F-actin and laminin to visualize myofibers, and type I myosin, a marker for primary myofibers at this time point, to distinguish between primary and secondary fibers. We found that similar to the phenotype observed in P7 and adult mice, the EDL muscle from CDH15^KO mice at E17.5 had myofibers of normal size but a reduced overall number of myofibers compared to wild-type mice (Figure 4C). CDH15^KO muscle contained a similar abundance of type I⁺ primary myofibers; however, reduced total myofiber number was ultimately driven by a significant reduction in secondary myofiber number. Overall, these data support that failure to form adherens junctions in the absence of M-cadherin impairs secondary myofiber formation during fetal development, ultimately leading to reduced muscle size.

Discussion

The development of skeletal muscle is a tightly coordinated process regulated by the spatiotemporal activation of transcription factors and signaling networks that control myogenic progenitor cell migration, proliferation, alignment, and differentiation to form mature myofibers. An important process during fetal muscle development is the formation of secondary myofibers, which physically associate with primary myofibers, using them as a scaffold for muscle tissue formation. Among the adhesion molecules found to be localized to primary and secondary myofiber junctions, we have found here that M-cadherin protein is abundant at myofiber adherens junctions during fetal development and is necessary for the proper formation of secondary myofibers and the subsequent establishment of normal muscle mass.

Although M-cadherin is highly and somewhat selectively expressed in skeletal muscle, previous studies utilizing CDH15^KO mice have not found it to play any major role in muscle development or maintenance (Goel et al., 2017; Hollnagel et al., 2002), though these studies did not perform comprehensive in vivo phenotypic analysis. Here, we identified a phenotype of mildly reduced lean body mass in individuals with pLoF variants of the CDH15 gene, implicating M-cadherin as a potential regulator of muscle mass in humans. To follow up on this observation, we generated CDH15^KO mice that supported the human phenotype, displaying reduced lean mass and muscle mass. Despite their reduced muscle mass, we found that similar to previous reports (Goel et al., 2017; Hollnagel et al., 2002), CDH15^KO mice did not have any gross muscle abnormalities, and upon more detailed phenotyping, they did not display alterations in myofiber type distribution, myofiber size, or exercise performance, and their skeletal muscle transcriptome was not significantly different than that of wild-type mice. Instead, we found that the reason CDH15^KO mice have reduced muscle size was due to a reduction in myofiber number, confirming a role of M-cadherin in mediating myofiber formation during development.

To determine the mechanism by which M-cadherin regulates myofiber formation during development, we assessed myogenic processes during embryonic and fetal development in CDH15^KO mice. Surprisingly, although M-cadherin was clearly present at the apical surface of Pax7⁺ progenitors and myogenin⁺ cells (as well as at myofiber-myofiber junctions) during late embryonic and fetal development stages (E15 and E17.5, respectively), the loss of M-cadherin in CDH15^KO mice did not affect the abundance of these cell types, nor did it affect the proliferation of Pax7⁺ progenitors or expression of the fusogenic genes Myomixer and Myomaker. These data suggesting that M-cadherin does not affect myogenic processes during development are in contrast with several reports of M-cadherin regulating myogenesis in vitro. For example, multiple groups have shown that knockdown of M-cadherin or blockade of M-cadherin-mediated adhesion via antibodies or peptides that target its extracellular domain can inhibit myoblast differentiation in vitro (Charrasse et al., 2007; Wang et al., 2013; Wrobel et al., 2007; Zeschnigk et al., 1995), and one report has shown that M-cadherin knockdown or blockade via antibodies inhibits myoblast proliferation in vitro (Marti et al., 2013). A likely reason for the discrepancy between these in vitro data and our in vivo findings is that the homotypic interactions of cadherins between myogenic cells and mature myofibers cannot be fully recapitulated in 2D culture (Redfield et al., 1997), which complicates the translatability of these assays. In vitro, M-cadherin primarily localizes to sites of myoblast fusion to potentially promote myogenesis, whereas in vivo, M-cadherin is localized at the apical surface of myogenic progenitors as part of the adherens junction complex (and sometimes with other cadherins in a context-dependent manner), and it has not been found to play a role in myogenesis during in vivo development or regeneration (Goel et al., 2017; Hollnagel et al., 2002).

Although we did not observe a clear effect of loss of M-cadherin on myogenic progenitor proliferation or myogenesis during development, we found that reduced myofiber number in CDH15^KO mice was resultant from impaired formation of secondary myofibers. As we also observed a loss of adherens junctions in fetal muscle of CDH15^KO mice, these data suggest that formation and maintenance of adherens junctions by M-cadherin is essential for establishing proper adhesion between primary and secondary myofibers as well as between primary myofibers and myogenic cells for the formation of secondary myofibers. Previous reports have alluded to M-cadherin’s potential role in secondary myogenesis, as it has been shown to accumulate at sites of cell-cell contact during fetal myogenesis in mice (Cifuentes-Diaz et al., 1995; Rose et al., 1994) and is enriched in myogenic cells at late embryonic and fetal development stages in humans, although not during earlier embryonic stages (Xi et al., 2020). Our data show that spatiotemporal localization of M-cadherin at myofiber-myofiber and progenitor-myofiber junctions during secondary myogenesis is biologically relevant, as M-cadherin mediates the formation of adherens junctions and likely proper myocyte adhesion to regulate secondary myofiber formation.

We postulated that β-catenin may be responsible for mediating M-cadherin regulation of secondary myofiber formation, as its signaling activity is canonically inhibited by cadherin-mediated sequestration to the cell membrane, or alternatively, its nuclear translocation and transcriptional activation in the absence of cadherins. This cadherin-mediated catenin signaling has been shown to play an important role in adult muscle stem cell activation and myogenesis, where loss of N-cadherin, or both N-cadherin and M-cadherin together, causes translocation of β-catenin from the membrane to the nucleus (as there is a lack of cadherins to sequester it at the membrane), activation of downstream β-catenin target genes, and activation of myogenesis (Goel et al., 2017). Here, we found that fetal muscle from CDH15^KO mice had a complete loss of membrane-localized β-catenin, suggesting that M-cadherin is the main cadherin protein that stabilizes adherens junctions during development, while other cadherins may compensate in adult muscle cells. Surprisingly, with the loss of membrane-localized β-catenin, we did not observe its nuclear translocation or activation of β-catenin downstream target genes; rather, it resulted in a complete loss of non-phosphorylated β-catenin protein within myogenic cells. While this is in contrast to what is observed in adult muscle stem cells upon the loss of cadherins (Goel et al., 2017), it is possible that fetal muscle cells express higher levels of destruction complex proteins (e.g., GSK3β) that lead to rapid degradation of liberated β-catenin, thus preventing it from nuclear accumulation. Further, our data showing reduced overall myofiber number are supported by a previous study of myogenic progenitor-specific deletion of β-catenin yielding a similar phenotype during development (Hutcheson et al., 2009). Loss of M-cadherin did not result in β-catenin nuclear localization or signaling; rather, it was required for β-catenin’s membrane localization and the formation of adherens junctions, suggesting that M-cadherin’s role in fetal development and formation of secondary myofibers can be attributed to its ability to establish and maintain proper cell-cell adhesion. These findings are supported by previous data demonstrating that cadherin/catenin-dependent adhesion is an important regulator of adult muscle stem cell quiescence, irrespective of nuclear β-catenin signaling (Hung et al., 2024).

In summary, we have identified that pLoF variants in the CDH15 gene are associated with reduced lean mass in humans and that CDH15^KO mice have a concordant phenotype, displaying reduced muscle mass, which can be attributed to impaired myofiber formation during fetal development. Our data suggest that M-cadherin plays an essential role in establishing adherens junctions between myofibers and myogenic cells of developing muscle and that this process is necessary for the proper formation of secondary myofibers and, thus, for the regulation of muscle mass. These findings, demonstrating a context-dependent role of M-cadherin specifically during fetal muscle development, provide additional insight into the role of cadherin-based adhesion in regulating skeletal muscle development and growth.

Methods

Human genetic analysis

Human genetic analysis was performed in the UKB cohort (Sudlow et al., 2015). In UKB, bioelectrical impedance (BIA) was assessed at baseline using a Tanita BC 418ma Body Fat Analyzer device. Bioelectrical impedance-derived whole-body lean mass was used as reported in UKB field 23101. Sample collection and preparation for whole-exome sequencing and data processing performed by the Regeneron Genetics Center have been previously described (Backman et al., 2021; Gaynor et al., 2024).

We identified variants predicted to affect protein function using SnpEff 4.3 (Cingolani et al., 2012), by selecting the most severe consequence for each variant. Frameshift variants, splice donor/acceptor variants, and stop-gain variants were considered pLoF variants, and five variant annotation algorithms were used to predict deleteriousness for missense variants, as previously described (Akbari et al., 2022). We aggregated ultra-rare (singleton) pLoF and likely deleterious missense variants (predicted as deleterious by all five algorithms) in CDH15 into a CDH15 rare variant burden exposure. The association between this CDH15 rare variant burden and BIA-derived lean mass was assessed using linear regression implemented in Regenie, as previously described (Akbari et al., 2022; Mbatchou et al., 2021).

Animal generation and tissue harvesting

All animal procedures were performed in accordance with the Regeneron Institutional Animal Care and Use Committee (IACUC). CDH15^KO mice on a C57BL/6 background were generated and maintained under pathogen-free conditions at Regeneron Pharmaceuticals Inc. using the VelociGene technology. Briefly, a TM-LacZ cassette was inserted in-frame with the 24th amino acid residue followed by a selection cassette that disrupted the transcription of the Cdh15 gene resulting in a null allele. Heterozygous mice were interbred to produce homozygous knockout and wild-type littermates. M-cadherin expression pattern was confirmed by β-galactosidase staining, and Cdh15-targeted deletion was measured by quantitative reverse-transcription PCR (RT-PCR) and RNA-seq of the mouse tissue.

For adult mouse experiments, 3- to 6-month-old male wild-type and Cdh15 knockout (CDH15^KO) mice were used. Body composition and bone health parameters were assessed using echoMRI (EchoMRI-100H) and MicroCT (Quantum Fx System, Caliper Life Sciences) analyses, respectively. The TA, gastrocnemius/plantaris complex (gastroc.), soleus, and EDL were dissected for muscle weights and then either frozen in liquid nitrogen-cooled isopentane in OCT for histology or stored in RNALater (AM7020, Thermo Fisher Scientific) for subsequent gene expression analysis. Additional organs (e.g., heart, liver, and spleen) were dissected for organ weights and gene expression analysis. The hindlimbs of mouse pups were harvested at P7 and frozen in OCT for histology. Embryos were harvested at E15 and E17.5, with hindlimbs being dissected and either frozen in OCT for histology or stored in RNAlater for subsequent gene expression analysis.

Hanging wire test

Mice were allowed to grasp onto a wire mesh grid with all four limbs, and the wire grid was then inverted over a padded surface. Mice were tested 3 times for how long they could hang upside down, with at least 3 min between trials. The max cutoff time was 10 min, and no additional trials were done if a mouse reached that time.

Voluntary wheel running

Mice were single housed in Promethion metabolic cages (Sable Systems) for 2 weeks for acclimation prior to wheel running. Mice had voluntary wheel access for 3 consecutive weeks where running distance and speed were tracked.

Immunofluorescence and LacZ staining

For immunofluorescence staining, OCT-embedded samples were sectioned at 12 μm thickness on a cryostat (CM3050S, Lecia Microsystems) at −20°C onto micro superfrost plus slides (48311-703, VWR). Sections were fixed with 4% paraformaldehyde (PFA) for 15 min. Sections were washed three times with PBS and then incubated in blocking solution (20% goat serum and 0.3% Triton X-100) for 1 h. Sections were incubated in primary antibody solution (1:100, unless otherwise stated, in blocking solution) overnight at room temperature. Sections were then washed three times for 5 min with PBS and incubated in filtered secondary antibody solution (1:250, in blocking solution). Sections were counterstained with Hoechst 33342 (1:1,000; H3570, Invitrogen) for 5 min and mounted in Fluoromount G (00-4958-02, Invitrogen). A modified protocol was used when staining for fiber type. In brief, sections were permeabilized with 0.1% Triton X-100 for 10 min and then incubated in 4% BSA for 1 h. Then sections were washed three times with PBS and were then incubated for 1 h in primary antibody solution. After three additional PBS washes, sections were incubated in secondary antibody solution for 1 h and then fixed with 4% PFA. Sections were then washed three times for 5 min with PBS and mounted in Fluoromount G. Images were taken on a Zeiss Axioscan microscope. Quantification of myofiber size, myofiber type, cell abundance, and cadherin/catenin area was performed using HALO software (Indica Labs) and ImageJ (National Institutes of Health).

Primary antibodies for immunofluorescence staining were as follows: myosin heavy-chain isoforms (MHC type I, BA-D5; MHC type IIa, SC-71; MHC type IIx, 6H1; MHC type IIb, BF-F3; Developmental Studies Hybridoma Bank [DSHB], University of Iowa), laminin (L9393, Sigma or 1:500 ab11576, Abcam), dystrophin (ab15277, Abcam), myogenin (ab124800, Abcam), M-cadherin (611101, BD Biosciences), Pax7 (Pax7, DSHB), F-actin (1:500; a2228, Thermo Fisher Scientific), pan-cadherin (ab51034, Abcam), non-phosphorylated β-catenin (8814S, Cell Signaling Technology), and Ki67 (ab15580, Abcam). The primary antibodies were detected with the appropriate Alexa Fluorophore 488, 546, or 647 secondary antibodies (Thermo Fisher Scientific).

LacZ staining was performed in embryos and adult sections as previously described (Adams and Gale, 2007). In brief, embryos were fixed in 4% PFA for 30–45 min, rinsed with cold PBS, washed three times for 20 min at 4°C, and incubated in LacZ staining solution (0.1% Xgal in DMF, 2 mM MgCl₂, 0.01% deoxycholic acid, 0.02% IGEPAL, and 5 mM potassium ferrocyanide, in PBS) up to 48 h in the dark on a rocker at 4°C. Embryos were then rinsed in cold PBS, post-fixed with PFA at 4°C overnight, incubated in 50% glycerol overnight at 4°C, and cleared/stored in 70% glycerol. For adult tissues, mice were perfused with 2% PFA, and dissected tissues were fixed for 30 min, rinsed, embedded in OCT and frozen, and subsequently sectioned at 10 μm for LacZ staining.

RNA-seq preparation and analysis

Tissues were homogenized in TRIzol, and chloroform was used for phase separation. The aqueous phase, containing total RNA, was purified using MagMAX-96 for Microarrays Total RNA Isolation Kit (Ambion by Life Technologies) according to manufacturer’s specifications. Genomic DNA was removed using RNase-Free DNase Set (QIAGEN). Strand-specific RNA-seq libraries were prepared using KAPA mRNA-Seq Library Preparation Kit (Kapa Biosystems). Twelve-cycle PCR was performed to amplify libraries. Sequencing was performed on Illumina HiSeq2000 (Illumina) by multiplexed single-read run with 33 cycles. Raw sequence data (BCL files) were converted to FASTQ format via Illumina Casava 1.8.2. Reads were decoded based on their barcodes, and read quality was evaluated with FastQC (http://www.bioinformatics.babraham.ac.uk/projects/fastqc/). Differentially expressed genes were identified by DESeq2 package, and significantly perturbed genes were defined with fold changes no less than 1.5 in either an up or down direction and with p < 0.01.

Taqman gene expression analysis

Total RNA was purified using MagMAX-96 for Microarrays Total RNA Isolation Kit (AM1839, Thermo Fisher Scientific) according to manufacturer’s specifications. Genomic DNA was removed using MagMAXTurboDNase Buffer and TURBO DNase from the MagMAX kit listed earlier. mRNA (up to 2.5 μg) was reverse-transcribed into cDNA using SuperScript VILO Master Mix (11755500, Thermo Fisher Scientific). cDNA was diluted to 0.5–5 ng/μL. 2.5–25 ng cDNA input was amplified with the SensiFAST Hi-ROX MasterMix (1 × 100 mL) (CSA-01113, Bioline) using the ABI 7900HT Sequence Detection System (Applied Biosystems). The following TaqMan assays from Thermo Scientific were used for gene expression analysis: Cdh15, Fwd-5′CCAGAGCCTTGGCCTGTC, Rev-5′-CTTGATCTGCACAAGTGGGTAGG, probe-5′-CTGGGCAGTCCCTGAGCCCAAA, Myomixer, Fwd-5′- CCGACTGAGCTCCCAAGACAT, Rev-5′-CTGTTGCTGGCTGAGGACAAAG, probe-5′-TCTGCTGAGCTGTCTGCT, Myomaker, Fwd-5′-CACTGGCCGACTTTGATGAAC, Rev-5′-AGTCCGCACAGCGATGGTAAG, probe-5′-CCCAGAGATCGACCTTCACAATGC, Gsk3b, Fwd-5′-GACACTCCCTGTGATCTATGTCAA, Rev-5′-GTCTCGATGGCAGATTCCAAA, probe-5′-ATGTATCAGCTGTTCAGAAGTCTAGCC, Axin2, Fwd-5′- CAGCACCTCTCCTCTGTTACCTT, Rev-5′-CGCTGGCAGACAGGACATAC, probe-5′-CCTTGGCGATCGCTCGGGTTT, Dkk1, Fwd-5′- GCCCAAGCAAGTGATTCCAG, Rev-5′-AAGTTGGCAGGCTTCCTTC, probe-5′-TGATGGGCAGAATTCAGAAGAGATTATGCA, Sp5, Fwd-5′-CGTGTGCAACTGGCTCTTCTG, Rev-5′-CTTGGCGAGGTGATCGCTTC, probe-5′-CAAGAGCTTCACGCGCTCGGACG.

Statistical analyses

Data are expressed as mean ± SEM. Statistical analyses were performed using GraphPad Prism version 9. Unpaired two-tailed Student t tests were performed for statistical comparison between groups, with p < 0.05 being considered statistically significant.

Resource availability

Lead contact

Requests for further information and resources should be directed to and will be fulfilled by the lead contact, Michael J. Stec (mstec3@its.jnj.com).

Materials availability

This study did not generate new unique reagents.

Data and code availability

UK Biobank individual-level genotypic and phenotypic data used in this study may be accessed by approved investigators via the UK Biobank study (www.ukbiobank.ac.uk/). RNA-seq data have been deposited in the Gene Expression Omnibus (GEO) managed by the National Center for Biotechnology Information under GEO accession number: GSE315987. Any additional information required to replicate the data reported in this paper is available from the lead contact upon request. This paper does not report original code.

Author contributions

D.R.G. and M.J.S. conceived and designed the research herein. D.R.G., V.H., S.N., J.B., P.D., P.S., Q.S., Y.T., and M.J.S. completed the experiments. D.R.G., Q.S., and M.J.S. analyzed and interpreted results. D.R.G. and M.J.S. wrote the manuscript. D.R.G., P.J.F., N.P.K., M.H., V.H., S.N., J.B., L.L., M.W.S., and M.J.S. edited and revised the manuscript.

Declaration of interests

D.R.G., P.J.F., N.P.K., M.H., V.H., S.N., J.B., P.D., L.L., P.S., Q.S., Y.T., M.W.S., and M.J.S. are current/former employees and shareholders of Regeneron Pharmaceuticals.

Published: February 12, 2026

Supplemental information

Data S1. Human genetic variants in the CDH15 gene and associations with lean body mass

CDH15 gene variants detected in the UK Biobank dataset.

Data S2. Differentially expressed genes between wild-type and CDH15^KO mouse tissues

RNAseq data from skeletal muscles, spleen, lung, liver, kidney, and heart, with differential expression cutoff of fold change no less than 1.5 in either an up or down direction and with P < 0.01.