Duchenne muscular dystrophy (DMD), the most common inherited myopathy affecting approximately one in 3,500 males, is characterized by muscle dystrophin deficiency. Dystrophin deficiency is caused by translational reading frame shifts or nonsense mutations in the dystrophin gene 1. DMD is a rapidly progressive disease occurring during childhood that causes affected individuals to lose their ability to walk by the age of 12 years old before they succumb during their twenties due to either respiratory or cardiac failure.

DMD is known to progress with a rather uniform pattern of muscle weakness, However, the existence of a modifying gene has been suggested due to the identification of unusually mild DMD phenotypes 234. Some phenotypic variability has been explained by the precise locations of the mutations and their effects on the dystrophin-dystroglycan complex 56 or by the identification of aberrant splicing products from the dystrophin gene 78910. In some cases, however, the same dystrophin mutation has been reported to result in different phenotypes 1112.

Although some phenotypic variability may arise due to environmental factors, such as diet or exercise, there are likely to be contributions from genetic components. In fact, differences in genetic backgrounds have been shown to influence the phenotypes of mice with a dystrophin-glycoprotein complex disorder caused by a mutation in the σ-sarcoglycan gene 13. Therefore, it is highly plausible that unknown genetic factors modify the phenotype of DMD.

Myostatin, also known as growth and differentiation factor 8 (GDF8), is a muscle-specific secreted peptide that functions to limit muscle growth 14. Several studies analyzing mutations of the myostatin gene have been conducted in the Western worlds 151617. To date, six polymorphisms and one intronic mutation have been identified in the myostatin gene. One of the identified polymorphisms (p.153K>R) has been associated with a hypertrophic response in muscles due to strength training 18. Recently, an infant was identified with the first homozygous disruption of the myostatin gene, which resulted in the child being exceptionally muscular at birth and unusually strong with increased muscle mass at four years of age 19. Remarkably, a single nucleotide change creating a potential illegitimate micro RNA target site in the 3' untranslated region of the sheep myostatin gene was disclosed to cause translation inhibition leading to the increase of muscularity 20.

Furthermore, disruption of endogenous myostatin by gene or RNA targetings was shown to result in anatomic, biochemical, and physiologic improvements in the dystrophic phenotype of mdx mice, a mouse model of DMD with a nonsense mutation in the dystrophin gene 2122, including particularly prominent enlarged fiber diameters and greatly reduced fatty fibrosis. These results suggest that blocking endogenous myostatin is a potential strategy for treatment of DMD 23.

We hypothesized that genetic variation in the myostatin gene modifies the phenotype of DMD. Therefore, nucleotide changes in the myostatin gene were investigated in Japanese DMD patients, resulting in the identification of novel mutations.

All three coding regions and a part of the 3' untranslated region of the myostatin gene were successfully PCR amplified from genomic DNA; 102 DNA samples were subjected to direct sequencing. Sequencing results of the exon 1-encompassing region disclosed completely normal sequences in all of the samples except for one nucleotide change in one sample. In this case (case 712), direct sequencing revealed overlapping G and C peaks at the 283^rd position of the myostatin cDNA (c.283G>G/C) (Fig. 2). Because this nucleotide position is a G in the wild-type sequence, the presence of a C at this position was determined to be a mutation (c.283G>C).

c.283G>C changed the codon corresponding to the 95^th amino-acid residue of myostatin from GAT to CAT, which substituted an Asp residue to a His residue (p.95D>H). This missense mutation was located at a conserved amino-acid residue in the propeptide region (Fig. 1) 24 and was predicted to affect the function of myostatin. Clinical examination of muscle strength, however, failed to reveal a clear difference between the DMD patient carrying this nucleotide change and the other DMD patients. At 8 years old, the patient could not stand up by himself, but was able to walk independently with a waddling gait.

In the exon 2-encompassing region, overlapping G and A peaks at the 490^th nucleotide of the myostatin cDNA (c.490G>G/A) were uncovered in four samples. c.490G>A corresponded to a known polymorphism that changes a GAG codon for Glu to a AAG codon for Lys at the position corresponding the 164^th amino-acid residue of myostatin (p.164E>K) 15. The allele frequency of c.490G>A was 2.0% (4 of 204 alleles). Interestingly, in one of the four samples with c.490G>A (case 549), overlapping C and A peaks were also found at the 466^th nucleotide of the myostatin cDNA (c.466C>C/A) (Fig. 3). Subsequent subcloning sequencing disclosed the presence of two clones; one was identical to the wild-type sequence (Fig. 3), whereas the second clone carried the two nucleotide changes observed with direct sequencing: c.466C>A and c.490G>A (Fig. 3). c.466C>A, which has not been previously described, changed a CTA codon for Leu to a ATA codon for Ile (p.156L>I). Because the novel missense mutation was located at a conserved residue, this mutation was predicted to disrupt myostatin function (Fig. 1), particularly when it was combined with the second amino acid substitution located only eight amino-acid residues away. Clinical examination of the patient carrying the allele encoding two nucleotide changes did not disclose a clear mild DMD phenotype in the 14 year old. He was wheel-chair bound at 7 years old. The phenotypes of the other three DMD cases carrying c.490G>A in one allele were clinically indistinguishable from the other DMD patients.

None of these Japanese DMD cases carried p.55A>T or p.153K>R, frequently observed polymorphisms in the United States, or the single nucleotide change in intron 2 that has been reported to result in gross muscular hypertrophy (Table 1) 15. Moreover, examination of the exon 3-encompassing region (Fig. 1) did not disclose any nucleotide changes in the genomic samples. This was compatible with results obtained in previous studies.

Sequencing of the 396-bp fragment of the 3' untranslated region revealed no nucleotide change, especially in the stretch of AC

G

Myostatin is a negative regulator of muscle growth that is attracting attention as a novel target for increasing muscle growth in cases of DMD 23. In this study, we conducted extensive sequence analysis of the myostatin genes in 102 Japanese DMD patients. As a result, two novel missense mutations (p.95D>H and p.156L>I) were identified (Figs. 2 and 3). In addition, a known polymorphism (p.164E>K) was identified in four of the DMD patients. Because one of the DMD patients carried p.156L>I and p.164E>K in the identical allele (Fig. 3), the overall mutant allele frequency was 2.5% (5 of 204 alleles). No truncation mutations in the myostatin gene, however, were identified in our study. In particular, the intron mutation that introduces a premature stop codon in myostatin mRNA resulting in marked muscle hypertrophy 1926 was not observed. Although a single nucleotide change in the 3' untranslated region of the sheep myostatin gene was shown to lead translational inhibition 20, any nucleotide change creating the octamer motif for the micro RNA target site was not disclosed in DMD patients. Exceptionally, one case was disclosed to harbor two nucleotides deletion in the 3' untranslated region (c.2264-2265delAT) in one myostatin gene (Fig. 4). It needs further study to clarify the meaning of this deletion.

Both of the novel mutations (p.95D>H and p.156L>I) were predicted to be pathogenic because they are located at conserved amino-acid residues at the propeptide region of myostatin 24. The phenotypes of the DMD cases heterozygous for p.95D>H, p.164E>K, or p156L>1 and p164E>K, however, were not significantly different from the phenotypes of the other DMD cases. It has been reported that a mother carrying a truncation mutation in intron 2 of the myostatin gene appeared muscular, whereas her son, who was homozygous for the same mutation, showed remarkable muscle hypertrophy 19. Therefore, the muscle volume and strength of individuals that are heterozygous for myostatin mutations may not be markedly affected by these mutations. Considering that women with one missense polymorphism in the propeptide region of myostatin exhibited increase in muscle volume in response to strength training 18, it is supposed that the case with two amino acid substitutions can be phenotypically modified by providing proper muscle rehabilitation. Future studies will address this supposition.

In order to clarify the roles of the two novel mutations, it may be necessary to identify cases in which the mutations are homozygous. In previous studies, homozygous polymorphisms in the myostatin gene have been reported to cause no clear changes in muscle volume or strength 151627. In this study, some of the DMD patients had mild phenotypes, such as an ability to walk independently past the age 12 years old (data not shown). Although we hypothesized that in these cases the mild phenotypes were a result of a modifier of the DMD phenotype, these patients did not have mutations in their myostatin genes. Particularly we have reported that aberrant splicing products of the dystrophin gene are a modifier of DMD 810.

The variability of the human myostatin gene has been studied in Western countries (Table 1) 15161727. To date, six nucleotide changes (two are common and four are uncommon) have been identified in the myostatin gene (Table 1 and 2). Two polymorphisms, a G to A change at codon 55 in exon 1 (p.55A>T) and an A to G substitution in exon 2 (p.153K>R), were represented in 6.6% and 9.8% of the examined alleles in the USA 15. In an Italian study 16, the p.55A>T and p.153K>R were identified in 0.2% and 3.3% of the examined alleles, respectively. In a study in Belgium, only one individual from 57 males was found to be heterozygous for p.153K>R 17. None of the Japanese patients in this study, however, carried p.55A>T or p.153K>R (Table 1). These differences in the incidences of the polymorphisms demonstrate that there is a racial difference in the spectra of these polymorphisms.

It is remarkable that four of the 102 patients were heterozygous for p164E>K, which corresponds an allele frequency of 2.0% in the Japanese patients (Table 1). The p164E>K allele was identified in only 2 out of 189 individuals in the USA, and was not observed in 120 Italians and 57 Belgians 151617. In one of the Japanese DMD patients carrying p.164E>K, a second p.156L>I mutation was found in the same allele (Fig. 3). Considering that p.164E>K was identified in both the Americans and Japanese populations, p.164E>K may be an old polymorphism that originated in a common ancestor of the two populations or the polymorphism is located at hot spot for nucleotide changes.

The present study, although limited to DMD cases, showed the rare occurrence of mutations in the myostatin gene in Japanese subjects. Our results indicate that heterozygous missense polymorphisms including two novel mutations did not produce an apparent increase in muscle volume or strength in Japanese DMD cases, even in a patient carrying two amino acid substitutions.

DMD: Duchenne muscular dystrophy

The author(s) declare that they have no competing interests.

AN carried out the molecular genetic studies and drafted the manuscript. YT carried out the clinical and molecular genetic studies. KS and AN participated molecular genetic study. YO and MY participated in clinical examinations. MM conceived of the study, and participated in its design and coordination. All authors read and approved the final manuscript.