All animal procedures were approved by the Purdue University Animal Care and Use Committee. The wild-type and genetically mutant mice were from C57BL/6 background. MyHC-IIB knockout (Myh4^–/–) mice were kindly provided by Dr. Leslie A. Leinwand (University of Colorado, Boulder, Colorado). PGC-1α transgenic (Ppargc1a^Tg/+) mice were a kind gift from Bruce M. Spiegelman (Harvard University). Myostatin knockout (Mstn^–/–) mice provided by Se-Jin Lee (Johns Hopkins University). PGC-1α transgenic mice and myostatin knockout mice were crossed with MyHC-IIB knockout mice to generate double knockout mice. For clenbuterol treatment, 20 parts per million (ppm) clenbuterol (Sigma, St. Louis, MO) in drinking water was fed to mice ad libitum for 2 weeks. For 5-aminoimidazole-4-carboxamide 1-β-D-ribofuranoside (AICAR, Sigma, St. Louis, MO) study, AICAR was first injected subcutaneously into the mother of the newborn mice for 3 weeks, then directly into the male pups after 3 weeks. The dosage for AICAR was 500 μg per gram body weight for the first week, and 1,000 μg per gram body weight for week 2 and 3. For weeks 4–6, the dosage is 500 μg per g body weight. For each genotype and treatment, six male mice were used for data collection and statistical analysis. To reduce long-term genetic variations, littermates were used for all experiments outlined herein. Specifically, when Myh4^+/– males and females were mated, Myh^+/+ and Myh4^–/– mice each were observed at 25%. These mice were then used as wild-type (WT) and knockouts (KO) in all experiments.

Serial sections from the muscle mid-belly were cut 10 μm thick using a cryomicrotome (Walldorf, Germany) and placed on silane-coated slides. Prior to immunohistochemistry, slides were removed from a –80°C freezer and air dried at room temperature (RT) for 1 h. Slides were blocked for 1 h at RT in phosphate buffered saline (PBS) containing 5% goat serum (blocking buffer). After three 5 min washes in PBST (0.1% Tween 20), serial sections were incubated at 37°C for 1 h with primary antibodies against MyHC-I (A4.840), IIA (SC-71), or IIB (BF-F3) (Developmental Study Hybridoma Bank, Iowa City, IA) diluted at 1:100 in the blocking buffer. To detect IIX fibers, we applied all three antibodies to muscle sections, unstained muscle fibers were identified as pure IIX fibers. After three additional 5 min washes in PBST, sections were incubated at 37°C for 1 h with goat anti mouse IgG and IgM secondary antibody (Jackson ImmunoResearch, 115-065-044, West Grove, PA) diluted 1:1,000 in the blocking buffer. After a 3 × 5 min wash in PBST, sections were incubated at 37°C for 45 min with peroxidase-streptavidin (Jackson ImmunoResearch, 016-030-084, West Grove, PA) diluted 1:1,000 in the blocking buffer. After a 3 × 5 min wash in PBST, MyHCs were visualized using a DAB reaction kit (Vector Laboratories, SK-4100, Burlingame, CA).

Cross sectional area (μm²) of muscle fibers and selected fields on muscle sections were measured using the Adobe Photoshop or NIH ImageJ software. A micrometer image taken at the same magnitude as muscle sections was used as a reference to convert pixel into μm².

To count the total number of the muscle fiber, low-magnitude (40x) images of muscle sections were taken and analyzed using the NIH ImageJ software. For TA and EDL muscle, we counted all the fibers in the muscle. For gastrocnemius muscle, we selected 10 random fields from the whole gastrocnemius muscle with each area containing 200–300 fibers. We then measured the cross-sectional area (μm²) of the selected field using a micrometer image taken at the same magnitude as muscle sections as a reference to convert pixel into μm². We then counted the total number of fibers within that field to obtain the fiber/μm² ratio. The area of the whole gastrocnemius muscle was calculated using a micrometer as a reference. Low magnitude images (40x or 10x) of the whole gastrocnemius muscle section along with a micrometer were taken and analyzed using NIH ImageJ as described above. By multiplying the fiber/μm² ratio with the total μm² of the gastrocnemius muscle, we obtained an estimation of the total number of fibers per gastrocnemius. To assess the accuracy of our estimation model, we counted the total numbers of 10 wild-type gastrocnemius muscles and found this estimation approximated the actual count with a percent error of 0.5%. For Myh4^–/– knockout, we counted all the fibers.

To measure the percentage of Type I, IIA, IIX, and IIB fibers in the gastrocnemius muscle, we counted the total number of each fiber type and divided this number by the sum of the number of all four types.

An Myh4-luciferase reporter containing Myh4 gene sequence from –2,560 to + 13 bp was kindly provided by Dr. Steven J. Swoap (Williams College, Williamstown, MA, United States) and described elsewhere (Swoap, 1998; Swoap et al., 2000). Plasmid pRL-SV40 (Promega, Madison, WI) encoding Renilla luciferase, was co-injected with the reporter to normalize transfection efficiency. Plasmids were purified using an EndoFree plasmid mega kit (Qiagen, Valencia, CA) diluted to a final concentration of 0.5 μg/μl in sterile 0.45% NaCl. Electroporation was performed as follows: two spatula electrodes were placed on each side of the muscle belly, and eight pulses (1 Hz, 200 v/cm) at 20 ms intervals were applied using a BTX ECM electroporator (Genetronics, San Diego, CA). Two weeks after electroporation, gastrocnemius muscles were collected and homogenized in passive lysis buffer (Promega), and luciferase activity was measured in a TD-20/20 luminometer (Turner Designs, Sunnyvale, CA) using a Dual-Luciferase assay kit (Promega).

Gastrocnemius muscles were snap frozen in liquid nitrogen and homogenized in ice-cold RIPA lysis buffer (1% Nonidet P-40, 0.1% SDS in 50 mM NaCl, and 20 mM Tris, pH 7.6) in the presence of 1 mM PMSF, aprotinin (10 μg/ml), leupeptin (10 μg/ml), 50 mM NaF, and 1 mM Na₃VO₄. Homogenates were centrifuged at 13,000 × g at 4°C for 10 min, and supernatants were analyzed for protein concentration with the Bio-Rad Protein Assay Kit II (Bio-Rad, #5000002, Hercules, CA). An aliquot (100 μg) was separated using SDS-PAGE (10%). Proteins were transferred to PVDF membranes, which were blocked in 5% non-fat milk at 4°C overnight. After three 5 min washes, membrane was incubated with primary antibodies followed by secondary antibody incubation. Where indicated, blots were stripped and re-probed with appropriate antibodies. Primary antibodies against ERK1/2 (#4696), P38 (#9212), and JNK (#9252) and their respective phosphorylated forms pERK1/2 (#4376, T202/Y204), pP38 (#9216, T180/Y182), and pJNK (#9255, T183/Y185) were purchased from Cell Signaling Technology (Danvers, MA). Antibody-antigen complexes were visualized using an ECL kit (Amersham, Piscataway, NJ).

Numerical data were graphed using Microsoft Excel software. Data are reported as mean ± standard error of mean (SEM). Statistical analysis was performed using the mixed model of JMP (SAS Institute, Cary, NC). Body weight, muscle weight, total fiber number, cross sectional area, fiber type percentage, and densitometry of Western blots were compared between genotypes and treatments. Asterisks or letters that differ indicate significance level at P < 0.05.

Mice lacking functional MyHC-IIB fibers have reduced muscle mass (Acakpo-Satchivi et al., 1997; Allen and Leinwand, 2001; Desgeorges et al., 2017), however, it is unknown whether the loss of muscle mass happens in utero or postnatally. We found that the body weights of newborn wild-type and Myh4 knockout mice were comparable, at approximately 1.3 g (P > 0.05; Figure 1B). We then measured the cross-sectional area (Figures 1A,C) and the total fiber number (Figures 1A,D) of the EDL and TA from wild-type and Myh4 knockout mice. Similar to weight, no differences in these muscle parameters were detected across genotypes (P > 0.05). These data suggest loss of the ability to generate MyHC-IIB fibers had no effect on muscle mass development suggesting loss of muscle mass in Myh4 null mice occurs postnatally. To narrow down the time window of muscle loss postnatally in Myh4 knockout mice, we measured the body weight (Figure 1E), muscle weight (Figure 1F), and muscle fiber number (Figure 1G) of gastrocnemius (GS) muscle 2, 3, 4, and 6 weeks postnatally. We chose gastrocnemius muscle because MyHC-IIB fibers are the most predominant fiber type in that muscle, and as such, knockout Myh4, and thus MyHC-IIB fibers would have a dramatic influence on the GS. The body weight of Myh4 null mice were significantly lower than that of the wild-type mice starting at week 3 of age (Figure 1E). Myh4^–/– gastrocnemius muscle weight followed the same pattern as the body weight (Figure 1F). Interestingly, the reduction of muscle fiber number in gastrocnemius muscle occurred between week 2 and 4, stabilizing after week 4, although between week 4 and 6, muscle weight continued declining. Together, these data suggest the loss of MyHC-IIB fibers reduces the capacity of muscle to increase in size and this effect follows a tight temporal pattern postnatally.

To further dissect the causes of muscle loss in Myh4 knockout mice, we examined muscle fiber type composition and the cross-sectional area of each fiber in the gastrocnemius muscle. We stained serial muscle sections with anti-myosin heavy chain I, IIA, and IIB antibodies to identify type I, IIA and IIB fibers (Figure 2A). For IIX fibers, we stained the section with all three myosin heavy chain antibodies and counted the non-stained fibers as IIX fibers (Figure 2A). In wild-type mice, the percentage of type I fibers in the GS muscle gradually decreased in frequency between 2 and 6 weeks. Type IIA and X(D) fibers remaining relatively constant over the same timeframe (Figure 2B). Conversely, IIX(D) fibers predominated the GS muscle in Myh4 null mice over the same period (Figure 2B). In the GS muscle of wild-type mice type IIB was the dominate fiber type, accounting for some 80% of all fibers postnatally (Figure 2C). In contract, type IIX(D) accounted for more than 80% of fibers in null mice postnatally (Figure 2C). Common across both wild-type and knockout gastrocnemius muscles was the observation that the percentage of type I fibers continuously decreased consistently between week 2 and 6. We then examined the cross-sectional area of each fiber types in wild-type and Myh4 knockout mice. In wild-type mice, all the four fiber types increased postnatally, reaching a plateau approximately 4 weeks postnatally (Figure 2D). Whereas in Myh4 knockout gastrocnemius, type I and IIA reached a plateau at week 4, type IIX(D) fibers continued to enlarge through 6 weeks of age (Figure 2D). Together, these findings show that muscle attempts to compensate for the lack of functional type IIB fibers by increasing the size and frequency type IIX(D) fibers as well as increasing the sizes of the remaining fiber types.

Preliminary data showed that after a mouse is born, its muscle fiber shifts from type I, oxidative type to type IIB, glycolytic types within the first 4 weeks. As such, we hypothesized that as muscle fibers attempt and fail to transition to the type IIB phenotype, fibers may be lost because the slower contracting programs are blunted. In an attempt to address this possibility and mitigate the loss in fibers, we crossed Myh4^–/– mice with Ppargc1^Tg/+ mice to generate double mutant mice. We chose Ppargc1^Tg/+ mice because they exhibit a fast-to-slow fiber type switching when PGC-1α is overexpressed in skeletal muscle (Lin et al., 2002); on the other hand, knockout of PGC-1α promotes fiber type switching from oxidative type I and IIA fibers to glycolytic type IIX and IIB fibers (Handschin et al., 2007). We found the body weight of Ppargc1a^Tg/+ mice were slightly lower than that of wild-type mice, though not at a significant level (Figure 3A). On the other hand, Ppargc1a transgenics increased the relative gastrocnemius mass (muscle weight/BW) in both Ppargc1a^Tg/+ and double mutant mice (Figure 3B). Although Ppargc1a^Tg/+ mice did not exhibit increased muscle fiber number in gastrocnemius, it partially rescued gastrocnemius fiber number in double mutant mice (Figure 3C). We then examined the percentage of each fiber type in gastrocnemius and found that Ppargc1a^Tg/+ muscle had reduced type IIB fibers and correspondingly increased oxidative type I and IIA fibers (Figure 3D). Overexpressing PGC-1α greatly reduced IIX fibers and increased IIA fibers in the MyHC-IIB deficient gastrocnemius (Figure 3D). Additionally, Ppargc1a^Tg/+ transgenics significantly increased type I and IIA fiber cross sectional area in both Ppargc1a^Tg/+ and Ppargc1a^Tg/+/Myh4^–/– mice (Figure 3E). Together, these data show that shifting muscle metabolism and presumably its signaling from glycolytic to oxidative phenotype partially rescued muscle fiber and mass loss caused by Myh4 ablation.

To confirm our transgene efforts, we used 5′-aminoimidazole-4-carboxamide 1-β-D-ribofuranoside (AICAR), a widely used compound to shift muscle metabolism to a more oxidative phenotype, to examine whether we could replicate the PGC-1α effect. AICAR is a pharmacological nucleoside, once transported into cells, it is metabolized to 5′-aminoimidazole-4-carboxamide 1-β-D-ribofuranosyl monophosphate (ZMP). Like endogenous AMP, ZMP is a potent AMP-activated kinase (AMPK) agonist (Ljubicic and Jasmin, 2013). Once activated, AMPK functions through silent mating-type information regulator 2 homolog 1 (SIRT1)-mediated PGC-1α activation to promote mitochondrial biogenesis (Canto et al., 2010; Viollet, 2017). Additionally, chronic AICAR treatment has been shown to significantly reduce type IIB fibers (Desgeorges et al., 2017), yet increases IIX fibers in rat extensor digitorum longus (EDL) muscle (Suwa et al., 2003). We administrated AICAR for 6 weeks, first through maternal milk, then directly into the weaned pups, and found AICAR had similar effects on body weight (Figure 4A), and gastrocnemius weight (Figure 4B), fiber number (Figure 4C), percent fiber type composition (Figure 4D), and fiber size (Figure 4E), as Ppargc1a^Tg/+ transgenics, though to a different extent. These data suggest that although Ppargc1a^Tg/+ transgenics and AICAR may function through different mechanisms, shifting muscle metabolism signaling to a more oxidative phenotype helps rescue or overcome some of muscle fiber losses in MyHC-IIB deficient mice.

By pushing muscle metabolism toward more oxidative phenotype, we showed that muscle mass and fiber loss can be partially rescued. Next, we asked whether IIB fibers are required for maximal muscle hypertrophy. In other words, could the existing muscle physiology created by the loss of Myh4^–/– be stimulated to compensate for the loss of muscle mass and fiber number in these mice. To answer this, we chose the well-studied “double muscling” mice lacking myostatin (Mstn^–/–) and crossed them with Myh4^–/– mice to generate double knockout mice (Mstn^–/–; Myh4^–/–). Inactivation of myostatin in mice (McPherron et al., 1997) cattle (Grobet et al., 1997; Kambadur et al., 1997; McPherron and Lee, 1997), humans (Schuelke et al., 2004), sheep (Clop et al., 2006), dogs (Mosher et al., 2007), as well as horses (Bower et al., 2012; Petersen et al., 2013), dramatically increases muscle mass and/or muscle strength/performance. In addition, Mstn^–/– mice also show increased total fiber number and contractility properties of muscles (Mendias et al., 2006). Mstn^–/– mice had significantly heavier body weights than the wild-type mice, however, lack of myostatin failed to rescue the lowered body weights of Myh4^–/– mice (Figure 5A). Similarly, Mstn^–/– gastrocnemius exhibited profound hypertrophy (Figure 5B), however, myostatin deficiency failed to rescue muscle loss caused by MyHC-IIB deficiency (Figure 5B). Moreover, the loss of myostatin was incapable of increasing the cross-sectional area of any of the fiber types in MyHC-IIB deficient gastrocnemius (Figure 5C), raising the possibility that part of the increases in muscularity noted in myostatin null animals may be related to shifts in muscle fiber composition.

Clenbuterol is a β2-adrenergic receptor agonist, and increases muscle mass and function (Shi et al., 2007; Kim et al., 2018; Ito et al., 2019). To test whether the adult muscle fiber in Myh4^–/– muscle responds to the growth promotant clenbuterol to the degree the wild-type mice fibers do, we fed the wild-type and Myh4^–/– mice 20 part per million clenbuterol in drinking water for 2 weeks. Clenbuterol significantly increased the body weight of the wild-type mice, whereas the body weight of the knockout mice remained unchanged (Figure 6A). Similarly, clenbuterol had a hypertrophic effect on muscle weight of gastrocnemius, whereas in Myh4^–/– muscle, such effect disappeared (Figure 6B). Of note, clenbuterol promoted I to IIA to IIX and to IIB conversion in wild-type muscle, whereas we saw a decrease in IIX fiber composition, though the proportion of type I and IIA fibers increased in Myh4^–/– muscle (Figure 6C). Accordingly, all the fiber types in the wild-type muscle increased significantly in size (Figure 6D). In contrast, IIX fibers in Myh4^–/– muscle decreased, though type I and IIA fibers increased in size (Figure 6D). Overall, these findings show clenbuterol cannot elicit a hypertrophic effect in mice lacking a functional Myh4 gene compared to their wild-type counterparts and suggest the added lean growth accretion caused by beta-adrenergic agonist feeding may be requisite on changes in muscle fiber types, especially including type IIB fibers.

To test whether the signals driving Myh4 gene expression had changed in Myh4^–/– muscle, we electroporated a reporter gene carrying a Myh4 promoter region, linked to luciferase into gastrocnemius (Figure 7A). The reporter gene activity remained unchanged in Myh4^–/– muscle, compared to the wild-type muscle (Figure 7A). Since the signaling of the MAPK family, ERK1/2 signaling in particular, is preferentially expressed in glycolytic muscles (Shi et al., 2008; Oishi et al., 2019), we measured the expression of ERK1/2 in wild-type and Myh4^–/– muscles postnatally. We observed that the protein level of ERK1/2 remained unchanged in weeks 2 and 3 postnatally, remaining at a high level at week 4 despite a decrease in its expression level at week 4 in wild-type muscle (Figure 7B). Furthermore, the phosphorylation level of ERK1/2 was significantly higher in Myh4^–/– muscle as compared to wild-type at week 4 (Figures 7C,D). Interestingly, the total levels of both p38 and JNK in Myh4^–/– muscle significantly increased compared to the wild-type, whereas only phosphorylated JNK increased in Myh4^–/– muscle (Figures 7C,E,F). To examine potential changes in signaling, we measured the ratio of all three phosphorylated MAPKs to their respective total MAPK proteins and found that pERK/ERK signaling significantly increased (Figure 7G). In contrast, pP38/P38 MAPK ratio decreased, whereas that of pJNK/JNK remained unchanged (Figure 7G). Together, these findings suggest that the signaling that drives the formation of fast-twitch fiber program remains unchanged even when Myh4 gene is ablated.