Myotonic Dystrophy in Transgenic Mice Expressing an Expanded CUG Repeat

See allHide authors and affiliations

Science  08 Sep 2000:
Vol. 289, Issue 5485, pp. 1769-1772
DOI: 10.1126/science.289.5485.1769


Myotonic dystrophy (DM), the most common form of muscular dystrophy in adult humans, results from expansion of a CTG repeat in the 3′ untranslated region of the DMPK gene. The mutantDMPK messenger RNA (mRNA) contains an expanded CUG repeat and is retained in the nucleus. We have expressed an untranslated CUG repeat in an unrelated mRNA in transgenic mice. Mice that expressed expanded CUG repeats developed myotonia and myopathy, whereas mice expressing a nonexpanded repeat did not. Thus, transcripts with expanded CUG repeats are sufficient to generate a DM phenotype. This result supports a role for RNA gain of function in disease pathogenesis.

Myotonic dystrophy (DM, prevalence 1 in 7400 live births) is characterized by dominantly inherited muscle hyperexcitability (myotonia), progressive myopathy, cataracts, defects of cardiac conduction, neuropsychiatric impairment, and other developmental and degenerative manifestations (1). This complex phenotype results from the expansion of a CTG repeat in the 3′ untranslated region (3′UTR) of the DMPK gene, which encodes a serine-threonine protein kinase (2). The transcripts from the mutant allele are retained in the nucleus (3, 4), and levels of DMPK protein are correspondingly reduced (5). The expanded repeat also changes the structure of adjacent chromatin (6) and silences the expression of a flanking gene (7, 8),SIX5, which encodes a transcription factor.

The effects on DMPK and SIX5 expression may account for particular aspects of the DM phenotype. Dmpk knockout mice have reduced force generation in skeletal muscle (9) and abnormal cardiac conduction (10), which suggests that loss of DMPK function may contribute to the muscle weakness and cardiac disease in DM. Six5 knockout mice have an increased frequency of cataracts (11, 12), suggesting that loss ofSIX5 function underlies the development of cataracts in DM. However, neither Dmpk nor Six5 knockout mice have reproduced the myotonia and progressive myopathy (9, 11, 13) that are the most characteristic and severe features of the disease. This suggests a species difference in the requirement for SIX5 or DMPK, or the existence of another independent effect of the expanded repeat.

We investigated the possibility that the pathogenic effect of the DM mutation is mediated by the mutant mRNA—in other words, that the nuclear accumulation of expanded CUG repeats is toxic to muscle fibers. This possibility was suggested by the unusual location (3′ noncoding sequence) of the mutation, the retention of mutant DMPK mRNA in muscle nuclei (3), evidence that expanded CUG repeats form extended hairpins (14, 15), and the observation that transcripts with expanded CUG repeats inhibit the differentiation of myogenic cells in tissue culture (16). We used a genomic fragment containing the human skeletal actin (HSA) gene (17) to express an untranslated CUG repeat in the muscle of transgenic mice. An expanded (∼250 repeats) or nonexpanded (5 repeats) CTG repeat was inserted in the final exon of the HSA gene, midway between the termination codon and the polyadenylation site (Fig. 1A) (18). This placement is similar to the relative position of the CTG repeat within the human DMPK gene, but the repeat tract is shorter than the highly expanded alleles (1 to 4000 CTG repeats) in DM skeletal muscle (19). Except for the repeat, the HSA constructs are devoid of sequences from the DM locus. Transgenic mice expressing a similar HSA fragment without the added CTG repeat have neither increased actin content nor abnormal muscle histology (20, 21), despite having increased levels of actin mRNA. (Human and murine skeletal actin have the same amino acid sequence.)

Figure 1

Expression of HSAtransgenes. (A) Diagram of HSA construct. Filled boxes are the HSA coding sequence. (B) Analysis of skeletal actin expression in vastus (quadriceps) muscle by Northern blot using human- or murine-specific actin cDNA probes. The human probe detects transgene output. The murine probe is a load standard that detects endogenous actin mRNA. The human RNA was isolated from a surgical sample and is partially degraded. (C) Northern blot of total cellular RNA (3 μg), nonpolyadenylated RNA (3 μg), and polyadenylated RNA (100 ng) with human-specific skeletal actin cDNA probe or (CAG)10 oligonucleotide. A small proportion of the HSA LR RNA is present in the nonpolyadenylated [polyA(−)] fraction. It is unclear whether these polyA(−) transcripts are the result of incorrect formation or degradation of the 3′ end, or of incomplete fractionation. (D) Analysis of actin protein in gastrocnemius muscle by protein immunoblot (upper panel) or Coomassie-stained polyacrylamide gels (lower panels) does not show aberrant migration of actin or increased actin mass in short-repeat line SR29 (lane 3) or long-repeat line LR32a (lanes 4 and 5) compared with nontransgenic littermates (lanes 1 and 2). (E) Myotonic discharge in paraspinal muscle of a mouse from line LR32a, elicited by brief movement of the EMG electrode. (F) Frequency histogram of cross-sectional area shows increased variability of muscle fiber size in long-repeat line LR329 relative to short-repeat line SR24.

We obtained seven lines of transgenic mice expressing the long repeat (LR) and five expressing the short repeat (SR) (Fig. 1B andTable 1) (22). The transgene was expressed only in skeletal muscle (23). Some of the mice from the HSA LR lines carrying the highest number of transgene copies (LR20a and LR21) showed silencing of the transgene. The expanded CTG repeats were fully transcribed, as shown by the appropriate increase in the length of the HSA LRmRNA and its hybridization with a (CAG)10 probe (Fig. 1C).

Table 1

Characteristics of HSA transgenic lines. For relative HSA expression levels, +++ is similar to the level of HSA mRNA in human skeletal muscle. For electromyography, the number of mice showing myotonia per number examined is shown. “−” indicates lines that were not examined. Histologic analysis and electromyography were performed on mice aged 6 to 14 months, except for lines LR32a and LR41, where seven younger mice (1 to 4 months) were also analyzed. Twenty-two mice in line LR32a had abundant myotonia in all regions examined, and seven had myotonia in paraspinal but not forelimb muscles (hindlimbs were not examined). Six hemizygous LR41 mice did not have myotonia up to 14 months of age, but four homozygous animals all developed myotonia before the age of 4 months. CN, central nuclei; +++ indicates CN in more than 25% of fibers.

View this table:

Analysis of the HSA LR mRNA by Northern blot (Fig. 1C) and sequencing of HSA LR cDNAs revealed that the long-repeat transcripts were fully spliced and polyadenylated, and that the actin coding sequence was intact (24). A variable amount of the HSA mRNA in line L32a was shortened (Fig. 1B) because of activation of cryptic splice sites in the 3′UTR, which results in the excision of the CUG repeat tract and 72 nucleotides (nt) of flanking sequence in an intron (25). This splice event was also detected at low levels by reverse transcription–polymerase chain reaction (RT-PCR) in other long-repeat lines, but not in lines with short repeats.

The phenotype of mice in line LR32a was analyzed most extensively because the expression level of the long-repeat transgene was high and silencing was infrequent. These mice showed normal weight gain and histology of nonmuscle tissue, but after weaning they had a mortality of 41% by 44 weeks (versus <5% in nontransgenic orHSA SR mice). Necropsy did not reveal the cause of death. In DM, cardiac arrhythmia is the second leading cause of death. Although we did not detect HSA LRexpression in the heart, the possibility of regional, low-level, or transient expression has not been excluded. There was no evidence of muscle weakness in LR32a mice at 6 months of age (26).

Electromyography in HSA LR lines revealed high-frequency (50 to 200 Hz) runs of muscle action potentials that continued for 1 to 20 s after insertion or repositioning of the recording electrode (Fig. 1E) (27). These repetitive discharges waxed and waned in frequency and amplitude, as is typical of myotonia in DM. Myotonic discharges were observed in six of seven lines that expressed long repeats, but not in short-repeat or wild-type mice (Table 1). The long-repeat mice also showed abnormal hindlimb posture when they initiated movement after a period of inactivity or when they were suspended by the tail. Myotonia was present inHSA LR mice as early as 4 weeks of age, when the muscles had a normal histologic appearance. These observations indicate that HSA LR mice have a true myotonic disorder, rather than nonspecific hyperexcitability associated with muscle necrosis.

Mice that expressed the long-repeat transgene developed histologically defined myopathy, whereas those expressing short repeats did not (Fig. 2 and Table 1) (28). Six of seven lines expressing long repeats showed a consistent pattern of muscle histopathology, including increases in central nuclei and ring fibers and variability in fiber size (Fig. 1F). Higher levels ofHSA LR expression were associated with more severe pathology (Table 1). Although abundant central nuclei, variability in fiber size, and ring fibers can each be observed in other disorders, this constellation of features in the absence of muscle fiber necrosis is suggestive of DM (29). In addition, mice in line LR32a had up-regulated the activity of succinate dehydrogenase (Fig. 2H) and cytochrome oxidase (23), a characteristic feature of oxidative muscle fibers. This alteration may have been triggered by the repetitive myotonic discharges, because a similar oxidative transformation in the muscle ofClcn1 Adr myotonic mice is reversible with anti-myotonia treatment (30). The proportion of oxidative fibers is also increased in human DM (31).

Figure 2

Muscle histology of short-repeat (line SR29) (A to D) or long-repeat (line LR32a) (E to H) transgenic mice. Representative images are transverse frozen sections of vastus muscle obtained from 6-month-old mice. Hematoxylin and eosin–stained muscle is normal in line SR29 (A) but shows increased variability in fiber size, split fibers, and central nuclei in line LR32a (E). Fluorescence microscopy using stains for nuclei (DAPI, blue) and basement membrane (anti-laminin, green) shows increased central and peripheral muscle nuclei in line LR32a (F) compared to line SR29 (B). FISH using CAG repeat oligonucleotide probe shows multiple discrete foci of expanded CUG repeats (green) in muscle nuclei (blue) of line LR32a (G), but not in line SR29 (C). These results are representative of five separate FISH experiments in line LR32a and two experiments in lines LR41 and LR20b. Histochemical stains for succinate dehydrogenase show increased activity and loss of fiber-type distinctions in line LR32a (H) relative to line SR29 (D). Scale bars, 5 μm in (C) and (G), 100 μm in other panels.

To quantitate the changes in myonuclear number and location, we performed morphometry with antibodies to laminin to outline the basement membrane of muscle fibers (Fig. 2, E and F) (32) and to distinguish muscle nuclei from the nuclei of interstitial cells. Relative to mice expressing short repeats, mice in line LR32a had more than twice the number of nuclei per muscle fiber and a much higher proportion of central nuclei (Table 2). In human DM there is a similar up-regulation of myonuclear number and a marked increase in central nuclei (33).

Table 2

Morphometric analysis of muscle in HSAtransgenic lines. Values are means ± SD for measurements of >100 muscle fibers per mouse, n = 4 mice per group. *P< 0.05, **P ≤ 0.001.

View this table:

The intracellular location of expanded CUG repeats was determined by fluorescence in situ hybridization (FISH) (34). The long-repeat transcripts were retained in the nucleus in multiple discrete foci (Fig. 2G) reminiscent of those seen in fibroblasts and myoblasts from DM patients (3). Because the expanded CUG repeat is the only sequence shared by the HSA LRand DMPK mRNAs, it appears that this sequence is sufficient to trigger the nuclear retention of a mature mRNA.

These results are consistent with the idea that transcripts with expanded CUG repeats are deleterious in muscle fibers. A direct effect by the CTG repeat tract in DNA is unlikely, becauseHSA LR mice that do not express the mRNA appear normal. An effect by actin protein is unlikely because (i)HSA SR lines had no myopathy or myotonia, (ii) the levels of actin protein were not increased in long- or short-repeat lines (Fig. 1D) (35), (iii) mutant actin was not detected (Fig. 1D), (iv) nuclear retention would limit the translation ofHSA LR transcripts, and (v) the protein product of the HSA LR mRNA, if it were translated, would be identical to murine skeletal actin. Current formulations for the mechanism of genetic dominance, which posit effects solely at the level of proteins encoded by mutant genes (36), may need to be revised.

The mechanism by which transcripts with expanded CUG repeats induce myotonia and muscle degeneration is unclear. Models involving trans-interference with polyadenylation (37) or splicing (38), sequestration of a CUG binding protein (39), or interactions with double-stranded RNA binding proteins (15) have been proposed. The deleterious effects of expanded CUG repeats are probably not restricted to skeletal muscle or DM, because expansion of an untranslated CTG repeat in a brain-expressed gene was recently associated with autosomal dominant cerebellar degeneration (40).

Muscle wasting and weakness is a frequent feature of DM.HSA LR mice, however, have not developed obvious weakness or muscle wasting by the age of 6 months. It is possible that muscle regeneration and repair can compensate for the myopathy inHSA LR mice. Alternatively, theHSA LR model may be incomplete because of factors related to the CUG repeat (its length, developmental expression, and flanking sequences) or a requirement for other effects of theDM mutation, such as deficiency of DMPK or SIX5.

  • * To whom correspondence should be addressed. E-mail: charles_thornton{at}


View Abstract

Navigate This Article