Relation between extent of myostatin depletion and muscle growth in mature mice

Abstract

Myostatin is a negative regulator of muscle growth and fiber size. Changes in myostatin expression might contribute to changes in muscle mass associated with various conditions, and reducing the amount of active myostatin is a potential strategy for preventing or reversing muscle atrophy. The present study was done to determine the extent to which myostatin levels must decline to induce growth of mature muscles. Myostatin expression was reduced by activating Cre recombinase in adult mice with floxed myostatin genes. The duration of Cre activation varied from 1 to 6 wk, and the residual myostatin mRNA expression after Cre activation varied from 3 to 63% of the normal level. Promyostatin levels declined in parallel with myostatin mRNA. There was no increase in muscle mass over the 3 mo following Cre activation if residual myostatin expression was ≥40% of normal. In mice with <40% of normal myostatin expression, muscle mass increased in proportion to the extent of myostatin depletion. In mice with ≤10% of normal myostatin expression, muscle mass increased ∼25%. Myostatin depletion increased myonuclear domain volumes and the ratio of RNA to myonuclei probably by enhancing DNA transcription rather than by inhibiting RNA decay. There was no evidence that maintenance of the hypertrophy during chronic myostatin deficiency requires altered activity of Akt/mTOR or p38 MAPK signaling pathways. These data suggest that anabolic therapies based on reducing the concentration of active myostatin will be effective only if a very large proportion of the myostatin is removed or inactivated.

myostatin is a potent inhibitor of muscle growth during development and also regulates muscle fiber size after maturity (19). There is much interest in determining whether reducing myostatin activity increases muscle mass in various conditions associated with muscle atrophy, such as cancer, old age, glucocorticoid therapy, and muscular dystrophies. Reducing the levels of active myostatin by RNA interference or antibody neutralization is a potential clinical strategy (4, 16, 23, 25, 36, 42), but it is unclear to what extent myostatin levels must fall to trigger significant muscle growth.

The relation between myostatin levels and muscle size is of interest not only for potential clinical applications but also because variations in myostatin expression might contribute to individual differences in muscularity or muscle growth/atrophy in response to altered physical activity. For example, several studies have demonstrated that resistance training reduces myostatin expression generally in the 20–50% range (8, 12, 14, 15, 24, 31, 37). One of these studies indicated that the increase in muscle mass induced by such exercise correlated with the magnitude of suppression of myostatin mRNA expression (12), whereas another study did not support such a correlation (15).

Although it is known that complete loss of myostatin expression can induce significant muscle fiber hypertrophy even after maturity (39), the quantitative relation between myostatin levels and the size of mature muscles has not been defined. In the present report, we show that muscle mass in mature mice increases in proportion to the extent of myostatin depletion below a certain threshold but does not increase in mice with modest (40–60%) declines in myostatin expression.

A secondary goal of this study was to examine whether myostatin-deficient muscles have altered expression or phosphorylation of a number of molecules that are thought to be important regulators of muscle mass (10): the serine/threonine kinase Akt, which promotes mammalian target of rapamycin (mTOR) activity and inhibits Foxo transcription factor activity when activated; p70 S6 kinase (S6k), an mTOR target that promotes protein synthesis when activated by phosphorylation; ribosomal protein S6 (rpS6), a target of S6k; initiation factor 4E-binding protein-1 (4E-BP1), an mTOR target that inhibits translation when it is not phosphorylated; mRNAs encoding the E3 ubiquitin ligases muscle atrophy F-box (MAFbx; also known as atrogin-1) and muscle RING-finger protein-1 (MuRF1), which are regulated by Akt via Foxo transcription factors; and p38 mitogen-activated protein kinase (MAPK), which can promote expression of MAFbx (13, 21, 43). These molecules were selected for analysis because previous studies have linked the Akt-mTOR and p38 MAPK pathways to myostatin signaling (7, 18, 22, 27, 29, 32, 35, 40). We show in this report that muscle hypertrophy during chronic myostatin deficiency is not associated with altered expression or phosphorylation of these molecules.

There is now a great deal of evidence that the muscle growth induced by loss of myostatin signaling after weaning is mediated by fiber hypertrophy rather than hyperplasia and that this hypertrophy is mediated by expansion of myonuclear domain volumes rather than incorporation of extra myonuclei (1, 4, 6, 9, 11, 20, 26, 32, 39, 42, 44). It is noteworthy that myostatin-deficient muscles maintain a normal RNA concentration, although the number of myonuclei per milligram tissue decreases (38, 39). This suggests that RNA production per myonucleus is increased or that the rate of RNA decay is slower in myostatin-deficient muscles. If the decay rate of mature transcripts is reduced, there should be an increased ratio of total RNA to primary transcripts (assuming that posttranscriptional RNA processing does not decline in parallel with RNA degradation rate). In the present study, we examined this ratio for transcripts that are expressed at very high levels in muscle: rRNA, α-actin mRNA, and myosin heavy chain 4 (2b) mRNA. The ratios did not change, suggesting a transcriptional mechanism for maintaining normal RNA levels in large myonuclear domains.

METHODS

Mice.

Male mice with a predominantly C57BL/6 genetic background were studied (6 or more crosses into this background). They were housed two to three per cage, with food and water available at all times. The project was approved by the University of Rochester Animal Research Committee and was compliant with all state and federal regulations on use of animals in research.

The production of mice in which the third exon of the myostatin gene is floxed (Mstn^f/f) was described previously (38). Mice bearing a Cre recombinase/mutant estrogen receptor binding domain transgene knocked in at the Rosa26 locus (Rosa26-CreER^T2) (33) were purchased from Artemis Pharmaceuticals. The CreER^T2 protein produced by this transgene, which is expressed in most tissues, stays in the cytoplasm bound to heat shock protein 90 until tamoxifen displaces the heat shock protein 90 and allows it enter the nucleus and excise floxed DNA. Mstn^f/f mice were crossed with Rosa26-CreER^T2/Mstn^w/w mice, and the Mstn^f/w/Cre+ offspring were bred to establish Mstn^f/f/Cre+ and Mstn^w/w/Cre+ lines.

The Cre genotype was determined by PCR analysis of DNA extracted from tail biopsies obtained at weaning. Three-primer assays were used to distinguish hemizygous and homozygous Cre mice. The forward primer (5′-cttccctcgtgatctgcaactc-3′) matched a sequence in the Rosa26 promoter. One reverse primer (5′-aaactcgggtgagcatgtcttt-3′) matched a sequence in the Rosa26 gene; the ∼300-bp PCR product demonstrated presence of the wild-type Rosa26 locus. The other reverse primer (5′-gtgaaacagcattgctgtcactt-3′) matched a sequence in the Cre coding sequence; the ∼550-bp product demonstrated the presence of the transgene. Mice were excluded from the study if any excision of the floxed DNA was detected in their tails according to a PCR method described previously (39).

Although mice can be bred to homozygosity for the Rosa26-CreER^T2 transgene, we found that Cre^+/+ mice often had some Cre-mediated DNA recombination in the tail at the time of weaning and lost more weight than Cre^+/− mice during tamoxifen feeding. Therefore, we used only Cre^+/− mice in this study. Because all mice in this study had the Cre^+/− genotype, we refer to only the myostatin genotype (Mstn^f/f or Mstn^w/w) in the remainder of this article.

Tamoxifen administration.

When mice were 4 mo of age, they were either euthanized for analysis of baseline (pretamoxifen) muscle mass and myostatin expression (n = 6/genotype) or fed tamoxifen. The tamoxifen was purchased from Sigma-Aldrich, mixed with sucrose, and sent to Harlan-Teklad for production of irradiated food pellets containing 0.025% tamoxifen and 5% sucrose in a low-phytoestrogen base diet (Teklad 2016). Mice were fed tamoxifen for 1 (n = 4/genotype), 2 (n = 6/genotype), or 6 wk (n = 6 per genotype). Because mice lost 10–15% of body weight during 2 wk of tamoxifen feeding, we fed the 6-wk treatment groups regular chow for 1 wk after each 2-wk interval of tamoxifen feeding (2 wk on/1 wk off/2 wk on/1 wk off/2 wk on). The mice were euthanized 3 mo after their last exposure to tamoxifen. At the end of the experiment, mice fed tamoxifen for 6 wk were ∼9 mo old, and those fed tamoxifen for 1 or 2 wk were ∼7.5 mo old.

Tissue collection and analysis.

Mice were euthanized between 0900 and 1030 by CO₂ inhalation until cessation of breathing, followed by cervical dislocation. Gastrocnemius and quadriceps muscles were immediately removed, weighed, and immersed in liquid N₂. Muscles were stored at −70°C. Quadriceps muscles from some of the mice treated with tamoxifen for 6 wk were frozen in cold isopentane rather than liquid nitrogen. These were used for analysis of muscle fiber cross-sectional areas and myonuclei counts (∼200 fibers/mouse), as described previously (39).

RNA was extracted from the gastrocnemius muscles with Trizol (Invitrogen) as recommended by the manufacturer. The amount recovered was quantified by UV light (260 nm) absorbance with a NanoDrop spectrophotometer. RNA preparations were of high quality, as reflected by A260/A280 ratios of ∼1.9 and intact 18S and 28S ribosomal RNA according to Bioanalyzer (Agilent) analysis. Myostatin and GAPDH mRNA levels were determined by reverse transcription and real-time PCR, as described previously (39). Postdevelopmental myostatin knockout does not affect GAPDH mRNA levels according to real-time PCR assays (39) and microarrays (41). Thus GAPDH cDNA was used as the reference for normalizing myostatin mRNA data across different reverse transcription batches and different thermal cycler runs. Real-time PCR analyses of 18S rRNA, pre-rRNA [5′-external transcribed sequence, (5′-ETS)], mRNAs encoding MAFbx (official gene name/abbreviation: F-box 32), MuRF1 (tripartite motif-containing 63), skeletal muscle α-actin (Acta1), and myosin heavy chain 4 (Myh4; also known as myosin heavy chain 2b), and pre-mRNAs encoding Acta1 and Myh4 were done with RNA from the 12 mice that received tamoxifen for 6 wk. The RNA samples were treated with DNase before reverse transcription to eliminate any traces of genomic DNA that might have been extracted with the RNA. TaqMan gene expression assays for 18S rRNA, Acta1 mRNA, and Myh4 mRNA were done as described previously (39, 40). Primer sequences for other assays were rRNA 5′-ETS, tgtcgttgtcacacctgtcc and aaataaggtggccctcaacc; Acta1 pre-mRNA (intron 1), aaactcacatccagcccatt and tcagacccccaggcatatc; and Myh4 pre-mRNA (intron 1), gattcaagtcggctgctaatgt and cgttggcaaccactgactta. Locked nucleic acid probes (Roche Universal Probe Library) were used to monitor amplicon levels: rRNA 5′-ETS, tggtggag (catalog no. 04686969001); Acta1 intron 1, ggctgctg (catalog no. 04688651001); Myh4 intron 1, ctctggct (catalog no. 04692250001). TaqMan assays for MAFbx (Mm00499523_m1) and MuRF1 (Mm01188690_m1) mRNAs were purchased from Applied Biosystems.

Immunoblots for relative quantification of total and phosphorylated S6k, rpS6, 4E-BP1, and Akt were done as reported previously (40). The same muscle homogenates were used for immunoblots of total and phosphorylated (Thr¹⁸⁰/Tyr¹⁸²) p38 MAPK with antibodies from Cell Signaling Technology (catalog nos. 9212 and 9215). For promyostatin immunoblots, muscles were homogenized in ice-cold PBS with 1% protease arrest reagent (CalBiochem), 0.5 mM EDTA, 0.5% Igepal CA-630, 0.1% SDS, and 0.25% sodium deoxycholate. Promyostatin and other glycoproteins were trapped in Concanavalin A-agarose, nonglycosylated proteins were washed away, and glycoproteins were eluted with methyl α-d-mannopyranoside (2). This step reduced the background from nonspecific antibody binding to proteins other than promyostatin. Glycoproteins were separated by SDS-PAGE with nonreducing conditions (samples heated in Laemmli buffer with no reducing agent for 10 min at 55°C). The anti-myostatin prodomain antibody (AF1539) was obtained from R & D Systems.

Data analysis.

P values (2-tailed) for differences between conditions were determined by t-tests.

RESULTS

Myostatin gene expression.

Expression of myostatin mRNA in Mstn^w/w control mice was similar under all conditions, so data from all of these mice were pooled to define the mean “normal” expression level. At 4 mo of age, Mstn^f/f mice that had not been exposed to tamoxifen had myostatin mRNA levels that were 81–126% of normal, all within the range of values observed in Mstn^w/w mice of the same age (Fig. 1). Mean myostatin exon 3 expression in Mstn^f/f mice fed tamoxifen for 6 wk was 9% of normal (range 3–14%). Mice fed tamoxifen for 1 or 2 wk had variable reductions in myostatin exon 3 expression (residual expression 19–63% of normal mean). There was no significant difference in the mean myostatin exon 3 expression in mice fed tamoxifen for 2 wk (36% of normal) and those fed tamoxifen for 1 wk (33% of normal). Even after an outlier in the 2-wk group (63% of normal) was excluded, the mean values were not significantly lower in this group (31% of normal) than in the 1-wk tamoxifen group.

Fig. 1.

Myostatin mRNA expression in gastrocnemius muscles before exposure to tamoxifen (tmx) and 3 mo after 1, 2, or 6 wk of tmx ingestion. There were no significant differences across conditions in the Mstn/GAPDH gene expression ratios in Mstn^w/w mice, so data from these mice were pooled to define the normal ratio (open bar). Gray bars show values from Mstn^f/f mice. P refers to comparison with mean value of Mstn^w/w mice. Bars represent mean values. Error bars represent 1 SE.

Translation of myostatin mRNA produces a propeptide with an NH₂-terminal extracellular localization domain, a prodomain that inhibits the activity of the active domain, and a COOH-terminal active domain. Dimers of the propeptide are formed by a cystine bridge between two active domains. Proteolytic cleavage of the promyostatin dimer to release the active dimer must occur before myostatin can activate its receptors. The predominant form of myostatin in mouse muscle is the propeptide dimer (2). In mice with myostatin mRNA expression levels that were ∼10–100% of normal, propeptide dimer concentrations were proportional to myostatin mRNA levels (Fig. 2).

Fig. 2.

Loss of promyostatin is proportional to loss of myostatin mRNA expression. Immunoblots of samples taken from muscles of tmx-treated Mstn^f/f (f) and Mstn^w/w (w) mice after extraction from equal amounts of total protein with concanavalin A. The %mRNA and %protein values were calculated by defining 100% as the average value for Mstn^w/w samples on a particular blot. Scatter plot with line of identity summarizes data shown under blots.

Body and muscle mass.

Before tamoxifen administration, at 4 mo of age, Mstn^f/f and Mstn^w/w mice did not differ in body mass (Table 1). At the termination of the experiment, mean body weights of all groups fed tamoxifen were 15–25% heavier than the average baseline weight. There were no statistically significant differences in final body mass of Mstn^f/f and Mstn^w/w mice under any conditions.

Table 1.

Body and muscle mass

Condition	Mstnw/w	Mstnf/f	%Δ	P
Body mass, g
Pretamoxifen	27.2±0.5	26.9±0.5	−1	0.69
1–2 Wk tamoxifen	31.7±0.6	31.7±0.8	0	0.98
6 Wk tamoxifen	32.7±1.0	33.8±0.5	+3	0.33
Gastrocnemius mass, mg
Pretamoxifen	292±4	284±9	−3	0.42
1–2 Wk tamoxifen	313±4	334±5	+7	0.004
6 Wk tamoxifen	301±8	376±5	+25	0.00001
Quadriceps mass, mg
Pretamoxifen	398±4	406±8	+2	0.41
1–2 Wk tamoxifen	435±8	473±6	+9	0.002
6 Wk tamoxifen	398±18	516±10	+30	0.0002

Values are means ± SE. P values (2-tailed) based on t-tests.

Mean gastrocnemius and quadriceps muscle mass was similar in Mstn^f/f and Mstn^w/w mice that did not receive tamoxifen. Three months after tamoxifen administration for 6 wk, muscles of Mstn^f/f mice were significantly larger than those of Mstn^w/w mice (25% gastrocnemius, 30% quadriceps; P < 0.001). The muscle enlargement cannot be explained by edema; water content of the muscles was similar in Mstn^w/w (75.0 ± 0.7%) and Mstn^f/f mice (75.6 ± 0.3%). The mean ratio of muscle mass to body mass (average of values for gastrocnemius and quadriceps) was 23% greater in the knockout mice (P < 10⁻⁷). In randomly selected microscope fields of quadriceps cross-sections from three mice of each genotype, mean fiber size was 18% larger in the myostatin-deficient muscles (Fig. 3). This hypertrophy occurred without an increase in the number of myonuclei per fiber [mean myonuclear frequency of 1.06 (Mstn^f/f) vs. 1.10 (Mstn^w/w) nuclei/fiber in 10 μm sections], so mean myonuclear domain size was 22% larger in the sections from myostatin-deficient muscles (Fig. 3).

Fig. 3.

Muscle fiber cross-sectional areas and myonuclear domain size in quadriceps muscles of myostatin-deficient (Mstn^f/f; gray bars) and control mice (Mstn^w/w; open bars) at 9 mo of age. The mice were fed tmx to activate Cre recombinase activity for three 2-wk periods starting at 4 mo of age. Cryosections of quadriceps muscles were obtained from 3 mice of each genotype, and then 5–7 microscope fields/section (each with ∼30–50 fibers, >200 fibers/mouse) were selected by a technician who did not know the genotypes of the mice. The box plots show the distribution [median, 25th and 75th percentiles (box borders), 10th and 90th percentiles (whiskers), and highest/lowest values (•)] of mean fiber cross-sectional areas and mean myonuclear domain sizes in the individual microscope fields.

In mice fed tamoxifen for 1–2 wk, there was a modest increase in mean gastrocnemius (7%) and quadriceps mass (9%) in Mstn^f/f mice vs. Mstn^w/w controls (P < 0.01). Feeding tamoxifen for 1–2 wk had a variable effect on myostatin gene expression, and the magnitude of muscle growth in Mstn^f/f mice correlated with residual myostatin gene expression rather than duration of tamoxifen feeding. Figure 4 shows the relation between myostatin mRNA expression and relative muscle mass (average of quadriceps and gastrocnemius).

Fig. 4.

Relation between myostatin expression and muscle mass 3 mo after induction of Cre recombinase activity with tmx. Open symbols illustrate the variability in muscle mass (normalized to body mass) and myostatin gene expression (normalized to GAPDH gene expression) in Mstn^w/w mice, in which Cre recombinase should have no specific effect on myostatin gene expression. Filled symbols illustrate relation between decline in myostatin mRNA induced by Cre recombinase-mediated knockout of floxed myostatin exon 3 and subsequent muscle growth in Mstn^f/f mice. Circles (Rosa26-CreER^T2 transgene) represent data from the present study, and triangles (CAGG-CreER transgene) represent data published previously (39).

Molecular phenotypes.

In mice treated with tamoxifen for 6 wk, there was no significant effect of long-term myostatin depletion on the total amount or phosphorylation of Akt, S6k, rpS6, 4E-BP1, and p38 MAPK (Table 2). Expression of MAFbx and MuRF1 mRNAs also was not affected by myostatin deficiency. The muscles of myostatin-deficient mice maintained a normal concentration of RNA per milligram tissue, most of which is rRNA, although the number of myonuclei per milligram muscle declined. Tissue concentrations of abundant muscle-fiber-specific mRNAs (Acta1 and Myh4) also were normal after myostatin depletion [levels of other abundant muscle fiber-specific transcripts also are normal in relation to Acta1 or Myh4 mRNA levels in myostatin-deficient muscles (41)]. Myostatin depletion did not alter the ratio of primary transcripts (as reflected by 5′-ETS of rRNA, intron 1 of Acta1 RNA, and intron 1 of Myh4 RNA) to primary plus processed transcripts (as reflected by assays for 18S rRNA and exons of Acta1 and Myh4 transcripts). These results suggest that increased transcription per nucleus rather than reduced degradation of RNA is responsible for the maintenance of normal RNA levels in myostatin-deficient muscles.

Table 2.

Molecular phenotypes 3 mo after Cre recombinase activation for three 2-wk periods

Measurement	Mstnw/w	Mstnf/f	P
Relative protein concentration/μg total extracted protein
Akt	100±15	97±9	0.84
Phospho-Thr308-Akt	100±15	66±18	0.19
Phospho-Ser473-Akt	100±19	66±9	0.13
S6k	100±14	101±12	0.96
Phospho-Thr389-S6k	100±8	100±13	0.98
rpS6	100±15	119±9	0.30
Phospho-Ser235/236-rpS6	100±26	86±21	0.69
4E-BP1	100±3	102±6	0.81
Phospho-Thr36/45-4E-BP1	100±8	95±8	0.64
Phospho-Ser64-4E-BP1	100±9	98±8	0.87
Phospho-Thr69-4E-BP1	100±8	89±5	0.26
p38 MAPK	100±5	88±3	0.06
Phospho-Thr180/Tyr182-p38 MAPK	100±21	106±14	0.81
Relative rRNA or mRNA expression
18S rRNA*	100±5	102±9	0.86
Pre-rRNA (5′-ETS)†	100±5	110±6	0.20
Acta1 mRNA*	100±8	98±3	0.83
Acta1 pre-mRNA (intron 1 RNA)‡	100±11	109±20	0.64
Myh4 mRNA*	100±17	91±14	0.68
Myh4 pre-mRNA (intron 1 RNA)§	100±8	101±22	0.97
MAFbx mRNA*	100±22	85±14	0.59
MuRF1 mRNA*	100±15	84±9	0.40
Total RNA, mg/g tissue	0.65±0.06	0.61±0.04	0.61

Values are means ± SE. S6k, p70 S6 kinase; rpS6, ribosomal protein S6; 4E-BP1, 4E-binding protein-1; 5′-ETS, 5′-external transcribed sequence; Myh4, myosin heavy chain 4; MAFbx, muscle atrophy F-box; MuRF1, muscle RING-finger protein-1. Except for total RNA, all values are expressed as %mean value of Mstn^w/w control mice. P values (2-tailed) based on t-tests.

^*Relative to total RNA.

^†Per 18S rRNA.

^‡Per total Acta1 RNA (exon 3–4 assay).

^§Per total Myh4 RNA (exon 39 assay).

DISCUSSION

In the present study, myostatin expression had to be reduced more than 60% to induce muscle growth in mature mice. These results suggest that adult muscles have ∼2–3 times the minimum concentration of myostatin that is needed to restrain muscle growth. Thus, a modest reduction in myostatin expression, such as the reduction often observed after exercise (8, 12, 14, 15, 24, 31, 37), might not be sufficient to promote muscle hypertrophy by itself. This conclusion is supported by the observation that many subjects with reduced (∼50%) myostatin expression after strength training do not have muscle fiber hypertrophy (15). These data also are consistent with the lack of correlation between circulating myostatin levels and muscle mass in normal men, in whom the lowest myostatin levels were approximately threefold lower than the mean values (17). Thus, anabolic interventions based on reducing the amount of active myostatin in muscle, e.g., by RNA interference or neutralizing antibodies, probably will not be successful unless a very large proportion of the myostatin is removed or inactivated.

In contrast to the present results, it was reported that a 25% decrease in myostatin mRNA levels induced by myostatin antisense RNA caused muscle growth in mice (23). However, the antisense RNA might have reduced efficiency of myostatin mRNA translation; myostatin protein levels were not determined. Electroporation into rat tibialis anterior muscle of an anti-myostatin small hairpin RNA reduced mean myostatin mRNA expression 27%, reduced mean protein expression 48%, and increased mean tibialis anterior mass 10% (25). Myostatin protein appeared to be absent in some of the muscles but was close to normal in other treated muscles. The relation between myostatin protein expression and muscle mass across the individual rats treated with the anti-myostatin shRNA was not shown.

We did not examine muscle function in this study, but previous research has indicated that the hypertrophy associated with blocking myostatin signaling can improve muscle strength or endurance in normal mice and in mouse models of muscular dystrophies (3–6, 9, 11, 18, 28, 30, 34, 42).

A potential limitation of the tamoxifen-inducible knockout model is that tamoxifen has some toxicity, which is reflected by weight loss during tamoxifen administration. Although both Mstn^f/f mice and the Mstn^w/w controls had identical exposures to tamoxifen, the possibility that this toxicity influences the effects of myostatin depletion cannot be ruled out. External signs of tamoxifen toxicity are transient. Within 2 wk after the end of tamoxifen administration, body weights and voluntary wheel running distances are normal (unpublished data).

The increase in muscle mass after postnatal loss of myostatin activity in mice develops because of muscle fiber hypertrophy rather than addition of new muscle fibers and does not require addition of new myonuclei (1, 4, 6, 9, 11, 20, 26, 32, 39, 42, 44). The present study supports the prior evidence that the fiber hypertrophy associated with myostatin deficiency develops without addition of myonuclei. In our previous study of postdevelopmental myostatin knockout, fiber hypertrophy fully accounted for the increase in muscle mass (39). The mean percent increase in fiber size based on randomly selected microscope fields was less than the percent increase in muscle mass in the present study. Given the overwhelming evidence from several studies (cited above) that postnatal myostatin deficiency does not induce hyperplasia, this discrepancy probably reflects sampling variation (note the large variation in mean fiber size across microscope fields shown in Fig. 3).

Although the number of myonuclei per milligram of muscle is reduced after myostatin depletion, there is not a corresponding reduction in the amount of rRNA or muscle-specific mRNAs. The mechanisms that are involved in increasing the RNA/myonuclear DNA ratio in myostatin-deficient muscle must have a global impact on RNA production or stability, because the overall mRNA/rRNA ratio does not change (38) and there is not a subset of highly expressed genes (e.g., those encoding contractile or metabolic proteins) that is selectively upregulated (41). The normal ratio of primary transcripts to total RNA in the myostatin-deficient muscles suggests that maintenance of RNA concentrations in the expanded myonuclear domains depends on increased transcription per myonucleus rather than slower RNA decay.

Because it has become clear that postdevelopmental myostatin deficiency leads to muscle fiber hypertrophy rather than hyperplasia and that hypertrophy occurs without addition of new myonuclei, we did not examine markers of satellite cell activation, proliferation, differentiation, or fusion. Instead, we examined several molecules in pathways (Akt/mTOR/S6k, p38 MAPK) that can regulate cell size. We did not detect a significant effect of long-term myostatin depletion on expression or phosphorylation of any of these molecules, although it is conceivable that effects might have been observed had the mice been euthanized during the dark cycle, when food intake and activity levels are high. Moreover, these results do not necessarily reflect the early responses to myostatin depletion. For example, we recently reported that inhibition of myostatin signaling for 4 days with an anti-myostatin antibody increased phosphorylation of S6k and rpS6 (40). It is possible that, once muscle fibers enlarge to a certain point, feedback signals dampen anabolic pathways to restrain further growth. At steady state, myostatin-deficient muscles have a normal fractional rate of protein turnover (38), which is consistent with the apparent lack of changes in pathways that regulate protein synthesis and breakdown. However, there must be some permanent molecular changes to maintain the hypertrophic state. Perhaps the key to understanding this will be to determine why the myonuclei are producing more RNA.

GRANTS

This work was supported by National Institutes of Health grants AR-054366 and RR-024160.