Musculoskeletal defects associated with myosin heavy chain‐embryonic loss of function are mediated by the YAP signaling pathway

Previously, we characterized the essential functions of MyHC‐embryonic (encoded by the Myh3 gene) in skeletal muscle differentiation during embryonic and early postnatal stages of development (Agarwal et al, 2020). As previously described, we observed 11% lethality of Myh3 germline knockout (Myh3^Δ/Δ) embryos during developmental stages; the remaining 14% completed embryonic development and were born as pups (Agarwal et al, 2020). Although these surviving Myh3 germline knockout mice exhibited severe muscle defects affecting myofiber number, area, fiber type, and satellite cell numbers in embryonic and early postnatal stages, this did not result in increased postnatal lethality compared with controls (Agarwal et al, 2020). On the other hand, additional defects such as scoliosis became apparent in Myh3^Δ/Δ mice at adult stages (Agarwal et al, 2020), indicating that the embryonic loss of MyHC‐embryonic might have effects at later stages or that it might have essential functions in adult life, which we wished to characterize.

Myh3^Δ/Δ mice exhibited significantly reduced total adult body weight consistently, as shown in male mice at 8–10 weeks and 6 months of age (Fig 1A and B). Muscle weights for the tibialis anterior (TA) and gastrocnemius, taken as representative muscles of differing fiber type composition, also showed a significant reduction of > 25% at 8–10 weeks and 6 months, indicating that reduced muscle weight contributed to the reduction in total body weight in Myh3^Δ/Δ male mice (Fig 1C–F). Interestingly, the number of myofibers per unit area of TA showed no significant difference at 8–10 weeks in Myh3^Δ/Δ mice, whereas a significant increase was seen at 6 months (Fig 1G–J). The number of myofibers per unit area showed no significant difference at 8–10 weeks and 6 months in the gastrocnemius of Myh3^Δ/Δ mice (Fig EV1A and B), whereas a significant increase was seen in the soleus muscle of Myh3^Δ/Δ mice at both time points (Fig EV1C and D). These results suggest that some of the consequences of loss of MyHC‐embryonic are dynamic and muscle specific. To characterize this further, myofibers were grouped into discrete bins according to their cross‐sectional area and the number of myofibers falling in each group quantified. We observed a significant increase in the number of smaller‐sized myofibers (1,000–1,500 μm² bin) and a significant decrease in the number of larger‐sized myofibers (2,500–3,000, 3,000–3,500, 3,500–4,000, and 4,000–4,500 μm² bins) in the TA from 6‐month‐old Myh3^Δ/Δ mice (Fig 1I–K). Similar results were observed for 8–10‐week‐old TA (Fig EV1E), 8–10 week and 6‐month‐old gastrocnemius and soleus muscles from Myh3^Δ/Δ mice (Fig EV1F–I). Scoliosis was observed in all Myh3^Δ/Δ mice, which became apparent by 6 weeks of age and persisted thereafter; an example at 8–10 weeks along with a control, imaged using microCT is shown (Fig 1L and M). Vertebral fusion was also seen in Myh3^Δ/Δ mice in the cervical (Fig EV1J and K), thoracic (Fig 1N–P), and lumbar region, with severe reduction in intervertebral disc region observed between fused vertebrae (Fig 1O and P). Upon quantification, approximately 50, 78, and 82% fused vertebrae were seen in the cervical, thoracic, and lumbar regions, respectively, in Myh3^Δ/Δ mice (Figs 1Q and EV1L and M). At the level of muscle function, Myh3^Δ/Δ mice exhibited significantly reduced grip strength normalized to body weight compared with controls, consistently, over 2‐week time points from 12 to 24 weeks of age (Fig 1R). In functional activity assays, Myh3^Δ/Δ mice exhibited reduced latency to fall in rotarod tests (Fig 1S) and reduced time to exhaustion (Fig 1T) as well as distance ran (Fig EV1N) in treadmill tests. The myosin ATPase activity was found to be significantly reduced in embryonic day (E) 16.5 myosin extracts from Myh3^Δ/Δ mice compared with controls (Fig EV1O).

Previously, we reported that loss of MyHC‐embryonic function leads to increased MyHC‐slow+ fiber numbers and elevated MyHC‐slow and ‐IIa protein levels in neonatal stages (Agarwal et al, 2020). Interestingly, although the increased MyHC‐slow+ fiber numbers persisted at postnatal day (P) 15 in Myh3^Δ/Δ mice, there were no significant differences by P30 (Agarwal et al, 2020).

To characterize the effect of MyHC‐embryonic on adult muscle fiber type, we carried out immunofluorescence labeling for adult MyHC isoforms, on cross sections through the TA muscle of control Myh3^+/+ and Myh3^Δ/Δ mice, at early (8–10 week) and late (6‐month) time points. The percentage of MyHC‐IIb+ and MyHC‐IIa+ fiber numbers compared with total fiber numbers were increased significantly in Myh3^Δ/Δ mice at both time points, indicating that MyHC‐embryonic is required for proper adult fiber type maintenance (Fig 2A–H). No significant effect was observed on the percentage of MyHC‐IIx+ fibers in the TA in Myh3^Δ/Δ mice (Fig 2I–L). Since the TA has very few MyHC‐slow+ fibers, we characterized the effect of loss of MyHC‐embryonic on MyHC‐slow fibers in the gastrocnemius, where we observed a significant reduction in the percentage of MyHC‐slow+ fibers in Myh3^Δ/Δ mice, both at 8–10 weeks and 6 months (Fig 2M–P). The soleus is one of the muscles with a high percentage of MyHC‐slow+ fibers, where we observed no significant differences in the percentage of MyHC‐slow+ fibers in Myh3^Δ/Δ mice at the 8–10‐week time point (Fig EV2A). Interestingly, the percentage of MyHC‐slow+ fibers exhibited a significant decrease in Myh3^Δ/Δ mice at the 6‐month time point (Fig EV2B).

We found that MyHC‐IIb and ‐IIa protein levels were elevated in Myh3^Δ/Δ mice TA muscle by western blots and densitometry, at 8–10 weeks and 6 months (Fig 2Q–U). MyHC‐slow protein levels were reduced in Myh3^Δ/Δ mice gastrocnemius muscle by western blots and densitometry, at 8–10 weeks and 6 months (Fig 2V–X). Thus, these results validate the immunofluorescence data with respect to the effect of MyHC‐embryonic on adult muscle fiber type. Western blots indicated that MyHC‐IIb protein levels were elevated at both time points, whereas MyHC‐IIa protein levels were increased only at the 8–10‐week time point and not at 6 months, in the gastrocnemius of Myh3^Δ/Δ mice (Fig EV2C–G).

These results demonstrate that MyHC‐embryonic function is crucial for adult muscle fiber type maintenance, with the most consistent effect seen on MyHC‐IIb fibers, which were increased in the TA and gastrocnemius at both time points (Figs 2A–D, Q, R, and T and EV2C, E, and G). Since MyHC‐IIb is the predominant fiber type present in both TA and gastrocnemius, minor differences are unlikely to be statistically significant, suggesting that loss of MyHC‐embryonic has the clearest effect on MyHC‐IIb fiber type. Loss of MyHC‐embryonic had perhaps the strongest effect on MyHC‐slow, with the percentage of MyHC‐slow+ fibers decreased to half and MyHC‐slow protein levels reduced by about fivefold, in the gastrocnemius at both time points (Fig 2M–P and V–X). The percentage of MyHC‐slow+ fibers were not affected in the soleus at 8–10 weeks, but was significantly reduced at 6 months (Fig EV2A and B). This indicates that some of the effects of MyHC‐embryonic on adult muscle fiber type are age‐ and muscle‐dependent. Myh3^Δ/Δ mice also exhibited an increase in MyHC‐IIa fiber type in the TA at both time points, the effect was restricted to the early time point in the gastrocnemius, confirming that loss of MyHC‐embryonic function has distinct effects on adult fiber type in different muscles (Figs 2E–H, Q, S, and U and EV2C, D, and F).

Loss of MyHC‐embryonic function led to an increase in Pax7+ muscle stem (satellite) cell numbers per unit area, in the 8–10‐week‐old TA (Fig 3A–C). Satellite cell numbers were increased at P30, an earlier time point, in the TA of Myh3^Δ/Δ mice (Fig EV3A–C). However, by 6 months of age, Myh3^Δ/Δ mice exhibited a significant reduction in satellite cell numbers in the TA (Fig 3D–F). These results were confirmed by western blots followed by densitometry of the TA, where an increase in Pax7 levels was observed at 8–10 weeks and a reduction at 6 months in Myh3^Δ/Δ mice (Fig 3G–I). In the gastrocnemius, Pax7 levels were decreased in Myh3^Δ/Δ mice at both the 8–10‐week and 6‐month time points (Fig EV3D–F). Previously, we had found that Pax7 levels were significantly reduced upon embryonic and fetal‐specific Myh3 knockout, which was back to normal by P0 (Agarwal et al, 2020). Thus, loss of MyHC‐embryonic leads to a postnatal increase in satellite cell numbers and Pax7 levels in the TA, followed by a decline at the 6‐month later time point (Figs 3A–I and EV3A–C). Unlike the TA, the gastrocnemius exhibited a decrease in Pax7 levels at both the 8–10‐week and 6‐month time points, suggesting that loss of MyHC‐embryonic has distinct effects on satellite cell numbers in different muscles.

The altered satellite cell numbers could be because they are activated early on, leading to exhaustion of the satellite cell pool. It is also possible that satellite cells in Myh3^Δ/Δ mice undergo increased apoptosis, resulting in their depletion. To distinguish between these, protein levels of the differentiation marker MyoD and the apoptotic marker Caspase3 were quantified by western blots of the TA at 8–10 weeks and 6 months (Fig 3G). Interestingly, densitometry showed a significant increase in MyoD levels at the 8–10‐week time point and a significant reduction at the 6‐month time point in Myh3^Δ/Δ mice, indicating that the satellite cells in Myh3^Δ/Δ mice are activated at early time points, leading to their depletion by later stages (Fig 3G, J, and K). Caspase3 levels exhibited a decrease at the 8–10‐week time point and no change at 6 months in Myh3^Δ/Δ mice, suggesting that increased apoptosis does not lead to the altered satellite cell numbers in Myh3^Δ/Δ mice (Fig 3G, L, and M). To further validate this, individual muscle fibers along with resident satellite cells were isolated from the extensor digitorum longus (EDL) muscle of 8–10‐week‐old Myh3^+/+ and Myh3^Δ/Δ mice, cultured, and labeled for MyoD as a marker for satellite cell activation (Fig 3N–O″). Significantly elevated number of activated Pax7+ MyoD+ satellite cells were observed at 6 and 12 h after fiber isolation in Myh3^Δ/Δ mice, confirming that satellite cells are indeed activated faster in Myh3^Δ/Δ mice (Fig 3N–P).

We also observed that the area occupied by connective tissue (marked by red in Sirius red staining) was at least double in the Myh3^Δ/Δ TA compared with controls, at both 8–10 weeks and 6 months (Fig 3Q–V). This suggests that loss of MyHC‐embryonic leads to increased fibrosis, possibly explaining the altered satellite cell activation status and numbers.

To further characterize the effect of loss of MyHC‐embryonic function, we carried out a mass spectrometric experiment comparing protein expression from the hind limb skeletal muscle of E16.5 Myh3^+/+ and Myh3^Δ/Δ embryos. Upon analysis, 42 proteins were uniquely expressed in Myh3^+/+, 77 in Myh3^Δ/Δ and 394 were common to both genotypes (Fig 4A). Of the 394 proteins expressed in Myh3^+/+ and Myh3^Δ/Δ embryos, 48 were downregulated and 61 upregulated in the Myh3^Δ/Δ muscle (Fig 4A). Further analysis of signaling pathways and processes identified proteins involved in integrin signaling, cytoskeletal regulation by RhoGTPase, ubiquitin‐proteasome, and FGF signaling as the four pathways most affected upon loss of MyHC‐embryonic function (Fig 4B). The presence of FGF signaling validates our previous study, where we had reported the role of FGF signaling in regulating the rate of differentiation upon loss of MyHC‐embryonic function (Agarwal et al, 2020).

Integrins are transmembrane proteins that connect the cellular cytoskeleton to the extracellular matrix, performing a wide range of functions such as adhesion, signaling etc. (Kechagia et al, 2019). Integrins are crucial for skeletal muscle function, with integrin α7 mutations leading to congenital myopathy in humans (Hayashi et al, 1998); integrin α7 knockout mice (Mayer et al, 1997), and integrin α5 mutant chimeric mouse embryos (Taverna et al, 1998) exhibit novel muscular dystrophy symptoms, and β1‐integrin deficient myoblasts show myoblast fusion and sarcomere assembly defects (Schwander et al, 2003). Integrins are localized primarily to the myotendinous junctions and costameres in the skeletal muscle and are crucial for sensing, responding to, and transmitting mechanical forces (Boppart & Mahmassani, 2019). Talin, filamin, and α‐actinin are integrin‐binding cytoplasmic proteins that connect to the actin cytoskeleton (Calderwood et al, 2001). Interestingly, we observed that Talin1 and Talin 2, key cytoplasmic proteins that link integrins to the actin cytoskeleton, are upregulated in Myh3^Δ/Δ muscle (Fig 4C). Previous studies indicate that mouse knockouts for Talin1 or Talin2 exhibit myopathy, and defects in myotendinous junctions, myoblast fusion and sarcomere assembly, phenotypes similar to integrin mutant mice (Conti et al, 2008, 2009). We also find that FLNB, a filamin protein, is detected only in the control Myh3^+/+ samples and not in Myh3^Δ/Δ muscle (Fig 4C). Interestingly, filamins are known to compete with talins to bind β‐integrin cytoplasmic tails, which can influence integrin activation and signaling (Kiema et al, 2006). We also found several cytoskeletal proteins that were dysregulated in the Myh3^Δ/Δ muscle (Fig 4D). Proteins under the category “cytoskeletal regulation by RhoGTPase,” which was the second most affected pathway upon loss of MyHC‐embryonic, are also known to crosstalk with integrins to regulate their signaling, activation, and mechanotransduction (Parsons et al, 2010; Lawson & Burridge, 2014).

Upon observing the dysregulation of key proteins in the integrin and RhoGTPase pathways in the Myh3^Δ/Δ muscle, we hypothesized that loss of MyHC‐embryonic might affect mechanotransduction. Western blots and densitometry confirmed the upregulation of Talin1 and reduction in Filamin B levels in Myh3^Δ/Δ muscle seen in the mass spectrometry experiment (Fig 4E and F). Several studies indicate that the Yes‐associated protein (YAP) signaling pathway functions downstream of integrin signaling in diverse contexts, for instance in skin homeostasis (Elbediwy et al, 2016), skeletal stem cell lineage commitment (Tang et al, 2013), prostate cancer (Varzavand et al, 2016), vascular remodeling (Yamashiro et al, 2020), atherosclerosis (Wang et al, 2016), and cancer cell invasion (Liu et al, 2018). Therefore, we studied YAP expression in skeletal muscle lysates from Myh3^+/+ and Myh3^Δ/Δ E16.5 embryos, where we observed a significant, twofold upregulation of YAP levels upon loss of MyHC‐embryonic (Fig 4E and F). One of the most upregulated candidate proteins in the Myh3^Δ/Δ muscle in our mass spectrometric analysis was neural precursor cell expressed developmentally downregulated 4 (NEDD4), an E3 ubiquitin ligase. We validated the mass spectrometric analysis by western blots, where we found NEDD4 levels to be significantly elevated in the Myh3^Δ/Δ muscle (Fig 4E and F). Interestingly, NEDD4 has been shown to regulate components of the Hippo pathway, such as the serine/threonine kinase large tumor suppressor (LATS), essential for YAP localization and transcriptional activity (Salah et al, 2013; Bae et al, 2015; Jeon et al, 2019). Increased NEDD4 levels lead to LATS degradation and increased YAP activation (Salah et al, 2013; Bae et al, 2015; Jeon et al, 2019). To confirm the increased YAP activity, we carried out a luciferase reporter assay by transfecting the YAP responsive 8xGTIIC‐luciferase plasmid into C2C12 mouse myogenic cells treated with a control siRNA or Myh3 siRNA and allowing the cells to differentiate for 24 h (Dupont et al, 2011). Significantly elevated luciferase activity was observed in the Myh3 siRNA‐treated C2C12 cells, confirming that loss of MyHC‐embryonic function leads to increased YAP activity (Fig 4G). These results were validated by western blots for total YAP which was elevated and phosphorylated YAP (S109), its inactive form, which was significantly decreased upon Myh3 knockdown in C2C12 cells (Fig 4H and I). To test whether this effect on YAP is unique to MyHC‐embryonic, we depleted another developmental MyHC, MyHC‐perinatal encoded by the Myh8 gene. This resulted in increased total YAP and reduced phosphorylated YAP (S109), indicating that in addition to MyHC‐embryonic, other developmental MyHCs are also capable of regulating YAP (Fig 4J and K). Thus, our results suggest that integrin and cytoskeletal pathway dysregulation in the skeletal muscle of Myh3^Δ/Δ embryos result in altered mechanotransduction leading to elevated YAP activation.

Next, we wanted to study YAP signaling at postnatal and adult stages to decipher whether the musculoskeletal defects seen in Myh3^Δ/Δ mice are linked to altered YAP pathway activation. Western blots and densitometry for total YAP indicated significantly elevated levels in the Myh3^Δ/Δ muscle at postnatal day (P) 0 and 4‐month‐old adult stages (Fig 5A and B). Serine phosphorylation of YAP is known to inactivate YAP by preventing its nuclear localization or promoting its degradation (Basu et al, 2003; Zhao et al, 2007, 2010; Llado et al, 2015; Meng et al, 2015; Zhang et al, 2015). Therefore, we tested the levels of phosphorylated YAP using the S109‐ and S127‐phospho‐specific YAP antibodies, which upon densitometry showed consistent decrease in the Myh3^Δ/Δ muscle at P0 and 4‐month stages (Fig 5A, C, and D). The reduced levels of phosphorylated YAP indicate that YAP signaling is upregulated upon loss of MyHC‐embryonic function.

Since YAP localization is serine phosphorylation‐dependent, we next isolated the nuclear and cytoplasmic fractions from the muscle of 4‐month‐old Myh3^+/+ and Myh3^Δ/Δ mice, which were then subjected to western blotting for YAP, H3A as nuclear marker and GAPDH as cytoplasmic marker (Fig 5E and F). The nuclear fraction exhibited ~threefold increase in YAP levels whereas the cytoplasmic fraction showed reduced YAP levels in Myh3^Δ/Δ samples, substantiating that the reduced levels of phosphorylated YAP in Myh3^Δ/Δ mice indeed led to increased nuclear accumulation of YAP (Fig 5E–H). To confirm that the increased nuclear YAP levels led to the upregulation of YAP signaling, we checked the transcript expression of two YAP target genes, connective tissue growth factor (CTGF) and cysteine‐rich angiogenic inducer 61 (CYR61) by qPCR (Meng et al, 2015), and found both to be upregulated in the muscle of 4‐month‐old Myh3^Δ/Δ mice (Fig 5I). Western blots for CTGF and CYR61 also showed significantly increased expression for both proteins in muscle protein lysates from 4‐month‐old Myh3^Δ/Δ mice, validating these findings (Fig 5J–M). The transcript levels of Angiomotin (Amot), MOB kinase activator 1A (Mob1a), and MOB kinase activator 1B (Mob1b), which are upstream regulators of YAP in the Hippo signaling pathway (Zhao et al, 2010; Chan et al, 2011; Nishio et al, 2016), were significantly reduced in the muscle of 4‐month‐old Myh3^Δ/Δ mice (Fig 5N).

Next, we tested whether the effect on YAP observed upon loss of MyHC‐embryonic function is recapitulated by inhibitors of cellular contraction. Treatment with para‐aminoblebbistatin, a myosin inhibitor, resulted in increased total YAP and reduced phosphorylated YAP (S109) in C2C12 cells (Fig EV4A and B). Treatment with 2,3‐Butanedione monoxime (BDM), another contractility inhibitor, led to no change in total YAP but increased phosphorylated YAP (S109) levels (Fig EV4C and D). Thus, treatment with inhibitors of cellular contractility leads to altered YAP levels and activity in myogenic cells.

Cell contraction was measured by culturing C2C12 cells treated with control or Myh3 siRNA on a collagen matrix, followed by the release of the matrix (Fig EV4E). A significant reduction in contraction was observed in Myh3 siRNA‐treated cells compared with controls, validating the altered contractility upon loss of MyHC‐embryonic function (Fig EV4F).

Levels of the YAP paralog TAZ were unchanged in the cytoplasmic and nuclear fractions and in the total lysate in muscle protein lysates from 4‐month‐old Myh3^Δ/Δ mice (Fig EV4G–L). Levels of TEAD, the YAP/TAZ interacting transcription factors, were undetectable in the cytoplasmic fraction and unchanged in the nuclear fraction as well as total lysate in muscle protein lysates from 4‐month‐old Myh3^Δ/Δ mice (Fig EV4G–L). Thus, the mechanotransduction‐mediated effects of loss of MyHC‐embryonic function specifically affect YAP and not TAZ or TEAD.

Overall, these results indicate that YAP signaling is upregulated in early postnatal and adult Myh3^Δ/Δ mouse skeletal muscle and may underlie the musculoskeletal defects seen upon loss of MyHC‐embryonic function.

Since we observed increased YAP pathway activation in Myh3^Δ/Δ mice, we next tested whether inhibiting the YAP pathway might rescue the defects seen in these mice. We used the small molecule CA3, which has been reported to be a potent inhibitor of YAP activity (Song et al, 2018; Kandasamy et al, 2020). We found that CA3 treatment at 1 mg/kg body weight, starting at P15, twice a week for 6 weeks, followed by analysis at 12–14 weeks of age led to a significant increase in the grip strength in Myh3^Δ/Δ mice, compared with vehicle‐treated mice (Figs 6A and EV5A). Control Myh3^+/+ mice did not exhibit any significant change in grip strength upon CA3 treatment (Fig EV5B). The number of myofibers per unit area (mm²) exhibited a significant decrease, suggesting an increase in the proportion of larger myofibers upon CA3 treatment (Fig 6B). This turned out to be the case when we grouped the myofibers according to their cross‐sectional area, where the proportion of myofibers with > 2,500 mm² significantly increased upon CA3 treatment (Fig 6C). The elevated fibrosis measured by Sirius red staining seen in Myh3^Δ/Δ mice was significantly reduced in the TA muscle upon CA3 treatment compared with vehicle (Fig 6D–F). Next, we studied the levels of MyHC and Pax7 proteins upon CA3 treatment in the TA; MyHC‐IIa levels which were increased in the TA of Myh3^Δ/Δ mice showed a significant reduction upon CA3 treatment (Fig 6G–H), MyHC‐IIb levels which were increased in the TA of Myh3^Δ/Δ mice showed no change upon CA3 treatment (Fig 6G and I), whereas levels of Pax7, which was increased at early time points in the TA showed a significant reduction upon CA3 treatment (Fig 6G and J). MyHC‐slow levels which were downregulated in the gastrocnemius of Myh3^Δ/Δ mice exhibited a significant increase upon CA3 treatment (Fig 6K and L). Interestingly, levels of Pax7 which were decreased consistently in the gastrocnemius of Myh3^Δ/Δ mice showed the opposite trend upon CA3 treatment, where Pax7 levels were elevated (Fig 6K and M). The proportion of MyHC‐slow+ fibers also exhibited a significant increase in the soleus upon CA3 treatment (Fig 6N). Expectedly, CA3 treatment did not alter the levels of YAP in the TA or gastrocnemius, since it has been reported to inhibit YAP activity rather than levels (Song et al, 2018; Fig 6O–R). Levels of the YAP targets CTGF and Cyr61 were significantly reduced in both TA and gastrocnemius, validating that CA3 treatment indeed led to decreased YAP signaling (Fig 6S–V). Treatment with CA3 led to a significant increase in MyHC‐IIa protein levels but did not significantly affect the levels of MyHC‐IIb, Pax7, YAP, or CTGF in control Myh3^+/+ mice (Fig EV5C and D). Remarkably, a robust reduction in scoliosis was seen in Myh3^Δ/Δ mice treated with CA3, compared with vehicle‐treated mice (Fig 6W and X). Upon quantification of the Cobb angle which measures the severity of scoliosis (Wang et al, 2018), a significant reduction was seen in CA3‐treated Myh3^Δ/Δ mice (Fig 6Y).

To test whether administration of CA3 starting in adult life had similar effects on the musculoskeletal abnormalities exhibited by Myh3^Δ/Δ mice, we performed identical experiments as detailed above, except that CA3 dosing was initiated at 6 weeks of age (adult dosing), as opposed to the P15 time point. Thereafter, a similar dosing strategy was followed until 12 weeks of age when the mice were tested (Fig EV5E). Adult CA3 dosing regimen did not result in any significant changes in grip strength, number of myofibers per unit area (mm²), proportion of larger myofibers, or fibrotic area (Fig EV5F–I). A slight increase in grip strength is observed in the vehicle‐treated Myh3^Δ/Δ mice (Figs 6A and EV5F) compared with untreated ones (Fig 1R), which could suggest an effect of the vehicle DMSO on muscle function. No significant change in protein levels of MyHC‐IIa or MyHC‐IIb in the TA or MyHC‐slow in the gastrocnemius was observed upon adult CA3 dosing (Fig EV5J–M). However, Pax7 protein levels were significantly decreased in the TA but unaffected in the gastrocnemius upon adult CA3 dosing (Fig EV5J–M). Total YAP protein levels were unaffected in the TA and gastrocnemius, while the levels of the YAP target CTGF were decreased in the TA but unchanged in the gastrocnemius following adult CA3 dosing (Fig EV5N–Q). Adult CA3 administration also did not have any significant effect on the severity of scoliosis in Myh3^Δ/Δ mice (Fig EV5R).

Thus, our results indicate that most defects seen in Myh3^Δ/Δ mice such as reduced grip strength and fiber size, elevated muscle fibrosis, altered levels of MyHC‐IIa and ‐slow, changed Pax7 levels and scoliosis are mediated by increased YAP pathway activation, which are rescued by treatment with the YAP inhibitor CA3 during early postnatal stages. Administration of CA3 initiated in adult stages did not have much effect on the musculoskeletal abnormalities exhibited by Myh3^Δ/Δ mice. This indicates that the severity of the Myh3^Δ/Δ musculoskeletal abnormalities progresses with age and early administration of CA3 possibly arrests or reverts the progression of these defects, which may not be the case in adult dosing regimen.

Based on our findings, we propose that MyHC‐embryonic in the developing skeletal muscle sarcomeres is required for normal contractility, mechanical cues, and integrin signaling. This leads to activation of the upstream kinases in the Hippo signaling pathway which phosphorylate YAP with the help of adaptors such as MOB1A/1B. Phosphorylated YAP is retained in the cytoplasm or degraded by the proteasome, resulting in reduced levels of nuclear YAP, preventing the activation of YAP target genes such as CTGF and CYR61 and reduced fibrosis (Fig 7A). In Myh3^Δ/Δ mice, loss of MyHC‐embryonic in the sarcomeres leads to aberrant contractility, mechanical cues, and integrin signaling. This in turn causes reduction in levels of the Hippo pathway adaptors such as MOB1A/1B preventing YAP phosphorylation leading to its stabilization, nuclear entry, binding to TEAD, activation of downstream targets such as CTGF and CYR61 and increased fibrosis (Fig 7B). Inhibiting the YAP signaling pathway by CA3 during early postnatal stages reduces YAP target gene activation and normalizes most of the musculoskeletal defects exhibited upon loss of MyHC‐embryonic function. Thus, YAP signaling is a crucial therapeutic target in MYH3‐associated musculoskeletal diseases such as spondylocarpotarsal synostosis.

Myh3 knockout mice (Myh3^Δ/Δ) were generated and genotyped as described previously (Agarwal et al, 2020). Mice were euthanized at selected time points (E16.5, P0, 4 weeks, 8–10 weeks, 12–14 weeks [4 months] and 6 months), as required for the study. All muscles of the hind limb were used at E16.5 and P0, whereas specific muscles (TA, gastrocnemius and soleus) were used at all other stages. Mice were maintained and all animal experiments performed according to the protocols approved by the institutional animal ethics committee (protocol numbers: RCB/IAEC/2016/005 and RCB/IAEC/2019/056).

DMSO (Sigma; D8418) or CA3 (Selleck chemicals; CIL56) resuspended in DMSO at 25 mg/ml, was administered intraperitoneally to control Myh3^+/+ or Myh3^Δ/Δ mice at a dose of 1 mg/kg body weight, starting at 15‐day postbirth, twice a week, for 6 weeks. To study its effect during adult life, 6–7‐week‐old mice were administered CA3. The treated animals were allowed to age until 12–14 weeks. Post‐treatment, x‐ray images of the mice were captured using an in vivo imaging system (Lago X, Spectral instruments imaging) to measure the Cobb angle. To determine scoliosis, the Cobb angle was measured manually using the vertebrae tilt method (Wang et al, 2018; Chen et al, 2021). The hind limb muscles of the treated mice were harvested following euthanasia and used for further experiments. MicroCT imaging of control and Myh3^Δ/Δ mice was carried out using an in vivo imaging system (Perkin Elmer; Quantum GX2 microCT). Fore and hind limb grip measurements were recorded using a grip strength meter (GT3, Panlab, Harvard Apparatus). For the rotarod test, mice were trained on a standard accelerating rotarod (Harvard Apparatus) for 3 days. In the final experiment, each mouse was placed on the rotating drum of the apparatus, rotating at a ramp‐up speed of 0–50 rpm for 50 s. The falling mice land on a platform connected to a timer which records the duration spent on the rotating drum. Results were represented as three technical readings per mouse. For the treadmill assay, mice were acclimatized to running on the treadmill apparatus (Columbus Instruments) at a speed of 10 m/min for 5 min, for 3 days. For the final experiment, each mouse was allowed to run on the treadmill at 10 m/min for 5 min, following which the speed was increased by 2 m/min every 2 min until exhaustion, when the duration and distance were noted. Male mice were used for body weight, muscle weight, grip strength, rotarod, and treadmill assays, where there are known differences between sexes. For all other parameters and analyses, gender segregation was not done. Control Myh3^+/+ and Myh3^Δ/Δ littermates were randomly selected for treatments. For grip strength assay, the investigators were blind to the treatment given.

The muscle tissues were harvested and embedded in cryomatrix (Thermo Scientific; 6769006) cryo‐sectioning media and flash‐frozen in cooled 2‐methyl butane (Sigma; M32631). The embedded tissue was sectioned at 10 μm thickness using a cryomicrotome (Thermo Scientific; Microm HM 550) and adjacent sections collected on charged glass slides (Avantor; 631‐0108). Adjacent sections were processed for the detection of adult myosin heavy chain isoforms (MyHC‐IIa, ‐IIb, ‐IIx), laminin, and Pax7, respectively. To stain for adult MyHC isoforms, sections were blocked with 10% goat serum (BioAbChem, 72‐0480) and 1% Tergitol solution (Sigma; NP‐40) in ice‐cold phosphate‐buffered saline (PBS), after acetone fixation (MyHC‐IIa, ‐IIx) or no fixation (MyHC‐IIb), incubated with the cocktail of adult MyHC isoform‐specific sera and laminin antibody in appropriate dilutions in 10% goat serum‐PBS, at 4°C overnight. Slides were rinsed with ice‐cold PBS twice and 0.5% PBS‐Triton X‐100 (PBST) once, incubated in biotin‐conjugated secondary antibody for amplification at 4°C overnight, followed by incubation in streptavidin‐coupled fluorophore for 2 h at 4°C. The sections were then rinsed in PBS, stained with Hoechst (ThermoFisher Scientific; H3570), fixed in 4% paraformaldehyde (PFA), and mounted using Fluoromount‐G (Southern Biotech; 0100‐01). To stain for Pax7, sections were fixed with 4% PFA and the antigens were retrieved using an antigen retriever (2100 PickCell Retriever; Aptum Biologics) in citrate buffer (1.8 mM citric acid and 8.2 mM sodium citrate in water). The sections were blocked using TNB blocking buffer (Perkin Elmer; FP1012) for 1 h and incubated with Pax7 sera in appropriate concentration diluted in TNB buffer. The signals were amplified using vectastain (Vector Laboratories; PK‐6100) and TSA reagent kits (Perkin Elmer; NEL741B001KT, NEL744B001KT). Details of all antibodies used are provided in Appendix Table S1. The sections were imaged using an Olympus fluorescence microscope (Olympus BX63F), with a Hamamatsu ORCA‐Flash 4.0 camera. Each image was maximally projected, and the number of MyHC+ fibers and fiber area were quantified using SMASH (Semiautomatic muscle analysis using segmentation of histology software; Smith & Barton, 2014). The cross‐sectional area was measured using CellSens Dimension 1.16 software and statistically analyzed (GraphPad Prism 5).

Single fibers were isolated from the EDL muscle as described previously (Agarwal et al, 2022) of 8–10‐week‐old Myh3^+/+ and Myh3^Δ/Δ mice and fixed in 4% PFA for 15 min. The fibers were blocked in 5% goat serum in 1x PBS containing 0.1% Triton X‐100 (MP Biomedicals; 194854) for 1 h and incubated in primary antibodies overnight at 4°C. The fibers were then rinsed in PBS, incubated with secondary antibodies for 2 h at room temperature, rinsed in PBS, and mounted in fluoromount‐G. The fibers were imaged using a Nikon Eclipse Ti microscope with a DS‐Qi2 camera.

To study fibrosis, muscle sections were stained with Sirius red. Briefly, muscle sections were fixed for 1 h in Bouin's fixative (75% saturated picric acid, 25% Formaldehyde, and 5% Glacial acetic acid) at 56°C, washed with tap water, and stained with Picro‐Sirius red (Sigma; 365548) for 1 h, followed by washing with 0.5% acetic acid. The sections were dehydrated in 50, 70, and 100% ethanol, equilibrated with xylene, and mounted in DPX (Sigma; 06522). The sections were imaged using a Nikon Eclipse Ti microscope with a DS‐Fi2 camera.

To document spinal fusion, whole‐mount skeletal staining of bone and cartilage was done using Alizarin red and Alcian blue as described (Rigueur & Lyons, 2014). Briefly, 4‐month‐old Myh3^+/+ and Myh3^Δ/Δ mice were euthanized, deskinned and all organs removed. The samples were dehydrated in 95% ethanol and incubated in acetone to remove fats and permeabilize. This was followed by incubation in Alcian blue (Loba Chemie Pvt. Ltd; 0083000005), destaining in 70% ethanol and 95% ethanol. The specimens were cleared in 1% potassium hydroxide, transferred to Alizarin red (Loba Chemie Pvt. Ltd; 0086000025) prepared in potassium hydroxide and destained in 1% potassium hydroxide. The samples were stored in 100% glycerol. Imaging of cervical, thoracic, and lumbar vertebrae was performed on a bright‐field stereozoom microscope (Nikon; SMZ 745T). The percentage of vertebral fusion was calculated by manually counting the number of fused vertebrae as a fraction of vertebrae for that group.

Hind limb muscles were harvested from Myh3^+/+ and Myh3^Δ/Δ embryos or mice and flash‐frozen in liquid nitrogen. For protein lysate preparation, the muscles were transferred to ice‐cold radioimmunoprecipitation assay (RIPA) buffer (Sigma; R0278‐500ml) containing protease/phosphatase inhibitor cocktail (Cell Signaling Technology; 5872S). The muscles were homogenized, vortexed intermittently for 30 min on ice, followed by centrifugation at 16,000 g for 15 min at 4°C. The supernatant was collected and protein samples were quantified using the Pierce BCA Protein Assay kit (Thermo Scientific; 23225). Protein samples were separated on SDS–PAGE followed by western blots using standard protocols. Blots were blocked in 5% BSA (phospho‐antibodies) or 5% milk, incubated with primary antibodies at 4°C, overnight, and subsequently with HRP‐conjugated secondary antibodies at room temperature for 2 h, followed by detection using the HRP substrate (Millipore; WBLUF0100). The blots were imaged on an ImageQuant LAS 4000 (GE) and densitometric quantification was performed using ImageJ, where protein levels were normalized to GAPDH levels.

The nuclear and cytoplasmic protein fractions were prepared from the hind limb muscles of Myh3^+/+ and Myh3^Δ/Δ mice according to published protocols (Dimauro et al, 2012). Briefly, tissues were homogenized in ice‐cold STM buffer (250 mM sucrose, 50 mM Tris–HCl pH 7.4, 5 mM magnesium chloride), vortexed, incubated on ice for 30 min, and centrifuged at 800 g for 15 min. The pellet obtained was processed for nuclear fraction and the supernatant for cytoplasmic fraction. The pellet was resuspended in STM buffer, vortexed, and centrifuged again at 800 g to get rid of debris. Finally, the pellet was resuspended in NET buffer (20 mM HEPES pH 7.9, 1.5 mM magnesium chloride, 0.5 M sodium chloride, 0.2 mM EDTA, 20% glycerol, and 1% triton x‐100), vortexed, and incubated on ice for 30 min. The nuclei were lysed by sonication and centrifuged to obtain pure nuclear lysate fraction. The supernatant for the cytoplasmic fraction was centrifuged twice at 11,000 g for 10 min and the pellet was discarded. The proteins in the supernatant were precipitated using 100% acetone and resuspended in STM buffer after centrifugation. The quantity of protein was estimated as mentioned above, and western blots were performed. H3A and GAPDH were used to normalize the quantity of proteins in the nuclear and cytoplasmic fractions, respectively. Details of all antibodies used are provided in Appendix Table S1.

Actin was isolated from mouse hind limb skeletal muscle using standard protocol (Racusen & Thompson, 1997; Das et al, 2019). Myosin heavy chain proteins were isolated in 300 μl of myosin extraction buffer (1.0 M KCl, 0.15 M potassium phosphate pH 6.8, 10 mM sodium pyrophosphate, 5 mM MgCl₂, 0.5 mM EGTA, 8 mM DTT and 1x protease inhibitor cocktail). The extracted myosin was loaded onto a 100 kDa cut‐off column and washed in myosin storage buffer (0.5 M KCl, 20 mM MOPS pH 7.0, 2 mM MgCl₂, and 8 mM DTT) with 5 mM ATP thrice followed by thrice without ATP. The extracted myosin was quantified using BCA kit, and ATPase hydrolysis assay was performed using the EnzCheck Phosphate Assay kit (E6646 ThermoFisher Scientific) according to the manufacturer's protocol. The absorbance was measured at 360 nm at 10‐s intervals for 1 h.

The TA muscle was harvested and flash‐frozen in liquid nitrogen from 12 to 14‐week‐old Myh3^+/+ and Myh3^Δ/Δ mice. Muscle samples were homogenized, RNA isolated using the RNeasy Lipid Tissue Mini Kit (Qiagen; 74804), and cDNA synthesized using SuperScript III reverse transcriptase (Invitrogen; 18080‐044) and oligo (dT) (Invitrogen; 58862) as per manufacturer's instructions. qPCR was performed with TB Green Premix Ex Taq (Takara; RR420A) on an ABI 7500 Fast Real Time PCR system (Applied Biosystems) and normalized to Gapdh transcript levels. Details of all primers used are provided in Appendix Table S2. A minimum of three biological replicates were used for the expression analysis.

Timed matings were set up and Myh3^+/+ and Myh3^Δ/Δ embryos were collected at E16.5, hind limb muscles isolated, homogenized, and sonicated in RIPA buffer followed by centrifugation at 13,000 g at 4°C for 10 min. The supernatant was subjected to acetone (Sigma; 179124) precipitation, incubated first at −20°C for 10 min, and then at −80°C for 30 min. After centrifugation at 3,500 g for 45 min at 4°C, the pellet was washed twice with acetone and dissolved in 8 M urea (Sigma; U5378). Following quantification by BCA assay, 100 μg of protein was processed further. Briefly, after adding Dithiothreitol (Sigma; 53819), the samples were incubated for 1 h at 56°C, which was followed by incubation in dark at room temperature for 30 min, with the addition of cysteine blocking agent (IAA; Sigma; 15161) to the samples. Samples were then digested with trypsin (Thermo Scientific; 90057) at 37°C overnight. Peptide extraction was performed by ultrasonication for 20 min in an extraction solution (50% Acetonitrile, 0.1% Formic acid) and the extract was dried in a speed vacuum concentrator at 25°C. The samples were then desalted using Pierce C18 Zip‐tips (Thermo Scientific; 87782) and run on an LC–MS/MS mass spectrometer (Triple TOF 5600 SCIEX).

Mascot generic files (mgf) were generated by using the raw files of the MS–MS run, which were then conferred protein identification after database searches using Search GUI. Sample peptides were scored in the form of emPAI (exponentially modified Protein Abundance Index) using MASCOT software, run with a false discovery rate of 1%.

C2C12 mouse myoblasts (ATCC; CRL‐1722) were maintained according to ATCC guidelines in culture media containing DMEM‐Dulbecco's Modified Eagle Medium (Gibco; 12800‐017) supplemented with 10% (v/v) fetal bovine serum (Sigma; F2442) and 2% penicillin–streptomycin (Gibco; 15140122).

For luciferase assay, ~ 30,000 C2C12 cells were seeded in each well of a 24‐well plate (Nunc; 142485) and the cells were reverse transfected with control or Myh3 siRNA using Lipofectamine RNAiMAX (Thermo Fisher; 13778150). After 24 h, growth media containing siRNA was removed and the cells were transfected with an 8XGTIIC‐luciferase (Addgene; 34615) reporter plasmid (1 μg/well; Dupont et al, 2011) and pHRL‐CMV renilla luciferase (Promega; E6271) plasmid (20 pg/well) using Lipofectamine 2000 (Thermo Fisher; 11668019). Post‐24 h, the media was replaced with fresh media for 24 h, thereafter replaced with differentiation media (DMEM containing 2% [v/v] horse serum [BioAbChem; 72‐0460] and 1% penicillin–streptomycin). Luciferase assay was performed after 24 h of culture in differentiation media, using a luminometer (Promega) with the Dual‐Glo Luciferase assay system (Promega; E2920).

For protein lysates, ~ 1,20,000 cells were seeded in each well of a 6‐well plate and transfected with control or Myh3 siRNA using Lipofectamine RNAiMAX as described above. The cells were shifted to complete media after 24 h, cultured for another 24 h, following which the complete media was changed to differentiation media. The cells were allowed to differentiate in this media for 24 h, harvested, protein lysates prepared in RIPA buffer and processed for western blots.

For cell contractility assay, ~ 50,000 C2C12 cells were reverse transfected with control or Myh3 siRNA in each well of a 24‐well plate. After 24 h, the cells were trypsinized and mixed with a collagen matrix as per manufacturer's protocol (Cell Biolabs, Cell contraction assay, Cat No. CBA‐201). The cells were supplemented with complete media, which was replaced with differentiation media after 24 h. The cells were cultured in differentiation media for 48 h, and the collagen was gently released by a pipette tip and monitored for 24 h.

The effect of cell contraction modulators on YAP signaling in C2C12 cells was studied by plating ~ 30,000 cells in each well of a 24‐well plate. The cells were cultured in complete media for 2 days, followed by replacement with differentiation media supplemented with the contraction modulators. The contraction modulators used were Para‐aminoblebbistatin (10 μM; Cayman Chemicals, Cat. No. 22699) and 2,3 Butanedione monoxime (BDM; 10 mM; Cell Biolabs, Cat. No. 20105). Cells were treated with the contraction modulators or DMSO for 24 h and lysates prepared for western blots as described above.

Experimental data were analyzed using GraphPad Prism 8 software (GraphPad Software Inc., CA, USA) and plotted as mean ± standard error of the mean. Parametric, two‐tailed unpaired t‐test was used to test the probability of significant differences, and P‐values ≤ 0.05 were considered significant (asterisk). All the animal studies were conducted using at least three biological replicates. Control Myh3^+/+ or Myh3^Δ/Δ littermates were randomly selected for treatments. For grip strength tests, the investigator was blind to the treatments given.