Muscleblind-like 2 (Mbnl2) -deficient mice as a model for myotonic dystrophy


  • Minqi Hao,

    1. Department of Medicine, David Geffen School of Medicine, University of California at Los Angeles, Los Angeles, California
    Search for more papers by this author
    • Drs. Hao and Akrami contributed equally to this work.

  • Kevan Akrami,

    1. Department of Medicine, David Geffen School of Medicine, University of California at Los Angeles, Los Angeles, California
    Search for more papers by this author
    • Drs. Hao and Akrami contributed equally to this work.

  • Ke Wei,

    1. Department of Medicine, David Geffen School of Medicine, University of California at Los Angeles, Los Angeles, California
    2. Molecular Cellular and Integrative Physiology, David Geffen School of Medicine, University of California at Los Angeles, Los Angeles, California
    Search for more papers by this author
  • Carlos De Diego,

    1. Department of Medicine, David Geffen School of Medicine, University of California at Los Angeles, Los Angeles, California
    Search for more papers by this author
  • Nam Che,

    1. Department of Medicine, David Geffen School of Medicine, University of California at Los Angeles, Los Angeles, California
    Search for more papers by this author
  • Jeong-Hee Ku,

    1. Department of Medicine, David Geffen School of Medicine, University of California at Los Angeles, Los Angeles, California
    Search for more papers by this author
  • James Tidball,

    1. Molecular Cellular and Integrative Physiology, David Geffen School of Medicine, University of California at Los Angeles, Los Angeles, California
    Search for more papers by this author
  • Michael C. Graves,

    1. Department of Neurology, David Geffen School of Medicine, University of California at Los Angeles, Los Angeles, California
    Search for more papers by this author
  • Perry B. Shieh,

    1. Department of Neurology, David Geffen School of Medicine, University of California at Los Angeles, Los Angeles, California
    Search for more papers by this author
  • Fabian Chen

    Corresponding author
    1. Department of Medicine, David Geffen School of Medicine, University of California at Los Angeles, Los Angeles, California
    2. Molecular Cellular and Integrative Physiology, David Geffen School of Medicine, University of California at Los Angeles, Los Angeles, California
    3. Department of Pediatrics, David Geffen School of Medicine, University of California at Los Angeles, Los Angeles, California
    • David Geffen School of Medicine, University of California at Los Angeles, Los Angeles, CA 90095-1760
    Search for more papers by this author


Myotonic dystrophy (DM), the most common adult-onset muscular dystrophy, is caused by CTG or CCTG microsatellite repeat expansions. Expanded DM mRNA microsatellite repeats are thought to accumulate in the nucleus, sequester Muscleblind proteins, and interfere with alternative mRNA splicing. Muscleblind2 (Mbnl2) is a member of the family of Muscleblind RNA binding proteins (that also include Mbnl1 and Mbnl3) that are known to bind CTG/CCTG RNA repeats. Recently, it was demonstrated that Mbnl1-deficient mice have characteristic features of human DM, including myotonia and defective chloride channel expression. Here, we demonstrate that Mbnl2-deficient mice also develop myotonia and have skeletal muscle pathology consistent with human DM. We also find defective expression and mRNA splicing of the chloride channel (Clcn1) in skeletal muscle that likely contributes to the myotonia phenotype. Our results support the hypothesis that Muscleblind proteins and specifically MBNL2 contribute to the pathogenesis of human DM. Developmental Dynamics 237:403–410, 2008. © 2008 Wiley-Liss, Inc.


Myotonic dystrophy (DM) is a dominantly inherited multisystem disorder caused by expansion of CTG or CCTG repeats that results in skeletal muscle hyperexcitability and progressive muscle degeneration (Harper,2001). DM is associated with insulin resistance, cataracts, cardiac conduction defects, hypogonadism, and neuropsychiatric pathology. Type 1 DM has been causally linked to the dystrophica myotonica protein kinase gene where there is a microsatellite (CTG)n expansion at the 3′ untranslated region. Type 2 DM has been localized to the zinc finger protein 9 gene where there is a microsatellite (CCTG)n expansion in intron 1. The number of these CTG/CCTG repeats can vary from less than 100 to several thousand, and the severity of the phenotype correlates with the degree of the microsatellite expansion. mRNAs expressed from these microsatellite repeats bind and adversely affect protein factors important in RNA splicing, most notably CUG-BP and the Muscleblind proteins (Ranum and Cooper,2006).

The Muscleblind genes (Mbnl1, Mbnl2, and Mbnl3) contain conserved tandem zinc finger domains, and they are similar to the Drosophila muscleblind (mbl) gene named for its importance in Drosophila muscle and eye development (Begemann et al.,1997; Artero et al.,1998; Miller et al.,2000; Kanadia et al.,2003b). The expression patterns of the three Muscleblind genes are different (Miller et al.,2000; Fardaei et al.,2002; Squillace et al.,2002; Kanadia et al.,2003b), and all colocalize with CUG RNA expansions in DM1 cells (Fardaei et al.,2002). Both MBNL1 and MBNL2 proteins are involved in alternative splicing of the insulin receptor that may be responsible for type 2 diabetes found in DM patients (Dansithong et al.,2004; Ho et al.,2004). A mouse knockout of Mbnl1 recapitulates many but not all of the features of human myotonic dystrophy (Kanadia et al.,2003a). The HSALR mouse model of DM expresses an expanded CTG repeat in the 3′-untranslated region (UTR) of a human skeletal actin transgene (Mankodi et al.,2000). MBNL1 overexpression in the HSALR mouse model reverses RNA missplicing and myotonia, but not the histological abnormalities found in DM (Kanadia et al.,2006). However, more complete reversal of DM can potentially be achieved by inactivating the toxic RNA molecules responsible (Mahadevan et al.,2006). These reports are suggestive of an independent role for MBNL2 in human DM.

Here, we demonstrate that Mbnl2-deficient mice develop myotonia and skeletal myopathy, key features of human DM. We demonstrate that Mbnl2 deficiency affects RNA and protein expression of the chloride channel (Clcn1) that is implicated in myotonia. These results are consistent with the notion that Mbnl1 deficiency alone does not fully replicate the human DM phenotype (Kanadia et al.,2003a). Thus, this Mbnl2 mutant mouse is a valuable model for examining the pathogenesis of human DM.


βgeo neomycin resistance/β-galactosidase DM myotonic dystrophy CSA cross-sectional area DAPI 4′,6-diamidino-2-phenylindole DNA deoxyribonucleic acid EMG electromyography ES embryonic stem kD kilodalton RNA ribonucleic acid PCR polymerase chain reaction SA splice acceptor μm micron wt wild-type


Generation and Characterization of Mbnl2 Mutant Mice

To determine whether deficiency of Mbnl2 contributes to the pathogenesis of DM, we generated a mouse mutant using an embryonic stem cell line with an insertional mutation in Mbnl2 (BayGenomics). The insertion is located in intron 2 of the Mbnl2 gene and results in the production of a fusion molecule encoded by the first Mbnl2 coding exon fused in-frame to a neomycin resistance/β-galactosidase (βgeo) reporter gene (Fig. 1A). We have verified the insertion site of the gene trap–targeting vector to intron 2 using polymerase chain reaction (PCR) analysis, DNA sequencing, and Southern analysis (Fig. 1B). Mbnl2 RNA expression is significantly decreased compared with wild-type and heterozygote in adult skeletal muscle from Mbnl2 homozygous mutant mice as demonstrated by Northern analysis (Fig. 1C). The Northern blot used a cDNA probe containing the entire Mbnl2 coding region. The wild-type Mbnl2 mRNA transcript is approximately 5.3 kb and the predicted mutant transcript is approximately 5.1 kb. The Northern blot did not demonstrate a second RNA transcript corresponding to the mutant Mbnl2 RNA transcript. However, the wild-type and mutant mRNA transcripts are similar in size, and may be difficult to separate on the Northern gel. These results demonstrate that wild-type Mbnl2 expression is significantly reduced in the Mbnl2 homozygous mutants.

Figure 1.

Characterization of Mbnl2 mutant mice. a: Top of figure demonstrates structure and exon organization of Mbnl2. Insertion of the SA (splice acceptor)-Neomycin-LacZ (β-geo cassette) is demonstrated in the intron between exons 2 and 3. Exon 2 is the first Mbnl2 protein coding exon. A fusion molecule encoded by the first Mbnl2 coding exon that is fused in-frame to a β-geo reporter gene is produced. b: Southern Blot of EcoRV-digested genomic DNA using an Mbnl2 intron 2–specific DNA probe. The wild-type (+/+) has a single wild-type (WT) band, and the homozygous mutant (−/−) has a single mutant band. The heterozygous (+/−) has both the wild-type and mutant alleles. c: Northern analysis of adult skeletal muscle RNA with a full-length Mbnl2 cDNA probe. Northern analysis repeated with an α-actin cDNA probe shown below. d: Mbnl2 is expressed in the skeletal muscles of the hindlimb in the adult Mbnl2−/− mouse. Medial view of the left lower extremity demonstrating X-gal staining in the gastrocnemius (arrowhead), tibialis (arrow), and the biceps femoris muscles (asterisk). e: X-gal staining of the vastus muscle that has been dissected from the left lower extremity. Note strong staining in the origin of this muscle.

Mbnl2 expression was examined with X-gal staining for β-galactosidase activity. The β geo gene is fused with the truncated Mbnl2 in the gene trap vector and should be expressed wherever Mbnl2 is present. As demonstrated, β-galactosidase activity was strong in the skeletal muscles of the hindlimb (Fig. 1D). Note that, in the vastus muscle, β-galactosidase activity was not uniform with notably stronger expression near the origin of this muscle (Fig. 1E).

Mbnl2 Deficient Mutant Mice Are Viable and Have Myotonia

The Mbnl2 heterozygous mutant embryonic stem cells were used to generate chimeric male mice that were bred to C57Bl6 female mice. Resulting heterozygote mice were intercrossed. From 154 adult mice produced, there were 38 +/+ : 84 +/− : 32 −/− genotypes. This ratio of genotypes does not demonstrate a significant mutant lethal phenotype. At 6 months of age, many Mbnl2 homozygous mutant mice displayed a phenotype that is similar to the clinical features found in DM patients. Clinical myotonia can be defined as delayed relaxation of a muscle following a voluntary contraction. After an individual with DM closes his eyes or performs a handgrip maneuver, relaxation of the skeletal muscle used is delayed and remains contracted for a few seconds. With repeated muscle contractions, the myotonia progressively decreases, and this is called the “warm up” phenomenon (Logigian et al.,2005). This myotonia was apparent with observation and manipulation of the Mbnl2 homozygous mutant mice. There appeared to be delayed muscle relaxation from rest that improved with motor activity (Supplementary Movie S1, which can be viewed at

Electromyography (EMG) of the paraspinal, gastrocnemius, vastus, and biceps skeletal muscles was used to further assess the myotonia, a hallmark feature of human DM. Myotonic discharges on EMG are characterized by prolonged bursts of repetitive activity. The amplitude of the electrical activity can rise rapidly and decline gradually. This electrical activity can have a characteristic auditory effect that is demonstrated in Supplementary Movie S2 (Harper,2001). All Mbnl2−/− mice tested (n = 13) exhibited myotonia (Fig. 2A, Supplementary Movie S2). Mbnl2 wild-type littermate controls (n = 9) all demonstrated normal EMG patterns (Fig. 2A).

Figure 2.

Mbnl2 deficiency results in myotonia, lordosis, and skeletal muscle pathology. a: Electromyography (EMG) of wild-type paraspinal skeletal muscle demonstrates normal EMG electrode insertional activity. Insertion of electrode triggers myotonic discharges in Mbnl2−/− mutant paraspinal muscle. b: Mbnl2−/− mice display lordosis, a feature of muscular dystrophy that is thought to be due to myopathy of the paraspinal muscles. An index of spinal curvature was calculated by dividing line AB (caudal border of the last cervical vertebra to caudal border of the lumbar vertebra adjacent to the cranial border of the ileum) by line CD (point of greatest vertebral curvature perpendicular to line AB). WT, wild-type. c: Average muscle fiber area was found to be smaller in Mbnl2−/− mutant compared with littermate wild-type controls. Mutant fibers exhibited a higher percentage of central nuclei compared with wild-type controls. d: Frequency histogram of skeletal muscle cross-sectional areas (CSA) showing decreased fiber CSA in Mbnl2−/− mutant compared with littermate wild-type controls. e,f: Masson trichrome staining of vastus muscle (objective magnification ×40) in wild-type (e) and Mbnl2−/− mutant mice (f). Mbnl2 mutant skeletal muscle demonstrates central nuclei (arrow), decreased myofiber size (arrow head), and increased fibrosis (blue staining) compared with a littermate wild-type. g,h: A higher magnification (objective magnification, ×100) view of vastus skeletal muscle demonstrates central nuclei (arrows) in Mbnl2−/− mutant compared with littermate wild-type mice. Scale bars = 30 μm.

In addition to myotonia, the Mbnl2 homozygous mutant mice have abnormal spinal curvature or lordosis (Fig. 2B) that was palpable and visually noticeable. We quantified the spinal curvature using X-ray radiographs in a manner similarly used to quantify kyphosis in children with Duchenne muscular dystrophy (Laws and Hoey,2004). A line (AB) was drawn between the caudal border of the last cervical vertebra to the caudal border of the lumbar vertebra adjacent to the cranial border of the ileum. Another line (CD) was drawn from the point of greatest vertebral curvature perpendicular to line AB. An index of spinal curvature was calculated by dividing AB by CD. Mbnl2−/− mice had a spinal curvature index of 3.7 ± 0.2 compared with wild-type littermates that had a score of 5.2 ± 0.6 (n = 3 for each genotype). Spinal deformity is a feature of muscular dystrophy. Similarly, the Mdx mutant mouse model of Duchenne muscular dystrophy has thoracolumbar kyphosis that worsens with increasing age. This thoracolumbar kyphosis or lordosis found in boys with Duchenne muscular dystrophy is thought to be due to dystrophy of skeletal muscles that support the spine.

Mbnl2-Deficient Mice Have Skeletal Muscle Pathology and Altered Chloride Channel Expression

Histological examination of Mbnl2−/− skeletal muscle revealed pathologic abnormalities often found in DM. We present the following histological analyses on the vastus muscle that demonstrated myotonia by EMG. The vastus muscle is affected in human DM and is a large skeletal muscle that can be easily divided for the RNA and protein analysis presented further below. There were more central nuclei in skeletal muscle from Mbnl2−/− compared with litter mate wild-type mice (Fig. 2C,E–H). Increased number of central nuclei is a pathologic feature found in human myotonic dystrophy and chronic neuropathic or myopathic processes. In addition, there was increased heterogeneity and an overall reduction in skeletal muscle fiber size in Mbnl2−/− as assessed by measuring cross-sectional area of muscle fibers from the vastus muscles (Fig. 2C,D). There was also increased collagen and fibrosis in Mbnl2−/− skeletal muscle compared with wild-type examined with trichrome staining. We calculated the percentage of trichrome staining in the microscopic field to be 4.8% ± 1.9% for Mbnl2−/− skeletal muscle compared with 0.9% ± 0.2% for wild-type (n = 3 for each genotype; Fig. 2E,F). This increase in collagen and fibrosis is a feature often seen in muscular dystrophy.

Although microsatellite mRNA repeat expansions in the DMPK and ZNF9 genes are causative for DM, the pathogenesis may be more directly attributable to their dominant effect on RNA binding proteins important in mRNA alternative splicing, and possibly RNA localization. In particular, aberrant splicing or mutation of the chloride channel gene (CLCN1) can cause myotonia in congenital human myotonia (Koch et al.,1992), and this finding has been implicated in DM (Mankodi et al.,2002; Kanadia et al.,2003a). We first examined alternative RNA splicing changes in genes previously shown to be altered with mouse Mbnl1 deficiency (Lin et al.,2006). Of interest, we found a modest defect in Clcn1 mRNA splicing in Mbnl2 mutant mice (Fig. 3A,B). Alternative splicing of Zasp (Z-band alternative spliced PDZ-motif-containing protein), Serca1 (fast twitch skeletal muscle Ca+2 ATPase), Tnnt3 (fast skeletal muscle troponin T), and M-titin mRNA that can be affected in human DM were not significantly affected. DNA sequencing of the Clcn1 reverse transcriptase-PCR (RT-PCR) products indicated a modest increase in inclusion of the cryptic exon 7a (Fig. 3B), resulting in an alternative RNA splice form of Clcn1 (Charlet et al.,2002; Mankodi et al.,2002). This RNA splicing defect was also described and more pronounced in the Mbnl1 mutant (Kanadia et al.,2003a) and HSALR mouse models of DM (Mankodi et al.,2000).

Figure 3.

Mbnl2 mutant mice have defective expression of chloride channel (Clcn1). a: Alternative RNA splicing of Clcn1 was affected in the Mbnl2−/− mouse, while m-Titin, ZASP, Serca1, and Tnnt3 were normally spliced (indicated exons are shown). There was increased expression of the cryptic exon 7a that can result in a truncated nonfunctional chloride channel. b: The percentage of Clcn1 RNA expression of exons 5–8 that contain exon 7a was increased in the Mbnl2−/− mouse (*P = 0.046, significant difference between mutant and wild-type; n = 3 for each genotype;). c,d: Clcn1 mRNA expression is decreased in skeletal muscle from Mbnl2−/− mice compared with control wild-type mice as assessed by real time reverse transcriptase-polymerase chain reaction (RT-PCR) specific for Clcn1 exons 10–14 (c) and exons 16–19 (d). The relative Clcn1 mRNA expression levels are shown. mRNA levels of Clcn1 are normalized to β2 Microglobulin2M) RNA levels. Each bar represents the mean ± standard error (n = 3 each for Mbnl2−/− and wild-type control).

We examined Clcn1 mRNA expression further by real-time RT-PCR, specifically examining the expression of two distinct Clcn1 distal coding regions, exons 10–14 and exons 16–19 (Fig. 3C,D). There was a significant decrease in expression of these two Clcn1 distal coding regions. Thus, these results demonstrate a significant reduction in Clcn1 mRNA expression with Mbnl2 deficiency. Clcn1 protein expression was examined in the vastus muscle using Western blotting for Clcn1 (Fig. 4). There was no significant difference in Clcn1 protein expression between Mbnl2−/− compared with wild-type skeletal muscle. Tubulin expression served as a protein loading control. However, we further examined Clcn1 expression by immunohistochemistry to determine whether there was abnormal localization or segmental regions of Clcn1 deficiency that could be responsible for the myotonia. In fact, there were large segments of decreased Clcn1 expression in Mbnl2−/− compared with wild-type skeletal muscle (Fig. 5A,B). Control immunohistochemistry was performed using an antibody for dystrophin (Fig. 5C,D) and this demonstrated no significant difference between Mbnl2−/− and wild-type. The immunohistochemistry for dystrophin was performed on the same tissue specimens as used for Clcn1 immunohistochemistry. Thus, there were areas of skeletal muscle in the Mbnl2−/− mice with a significant decrease in Clcn1 protein expression skeletal muscle that also had an intact skeletal muscle membrane as demonstrated by the expression of dystrophin. These results support the notion that defective Clcn1 expression associated with Mbnl2 deficiency contributes to the myotonia in Mbnl2−/− mice.

Figure 4.

Western analysis of adult vastus skeletal muscle chloride channel (Clcn1) expression. There was no significant difference in Clcn1 protein expression in Mbnl2 homozygous mutant mice compared with wild-type when the entire vastus muscle was examined. Western analysis for tubulin expression was used to control for protein loading shown below.

Figure 5.

Mbnl2 deficiency causes segmental loss of chloride channel (Clcn1) expression in skeletal muscle. a,b: Immunohistochemistry for Clcn1 demonstrates normal Clcn1 expression in wild-type skeletal muscle (a) and decreased Clcn1 expression in regions of Mbnl2−/− mutant skeletal muscle (b). c,d: Comparable expression of dystrophin is present in wild-type (c) and Mbnl2−/− mutant mice (d; same specimen as a,b). Vastus muscle from an 8-month-old mouse. Waviness of the skeletal muscle membrane is an artifact of cryosectioning. 4′,6-Diamidine-2-phenylidole-dihydrochloride (DAPI) staining for nuclei is present. Scale bars = 30 μm.


This report demonstrates that mouse Mbnl2 deficiency causes myotonia that can serve as a model for examining human DM. Given that DM is a multisystem disorder, multiple mechanisms likely contribute to the pathogenesis of DM. Although there may be functional redundancies present between Mbnl1, Mbnl2, and Mbnl3, there are undoubtedly many differences in their functional capacity and their expression patterns. The severity of the myotonia, muscular dystrophy, and other systemic features of DM are dependent on the extent of the expression of the CTG/CCTG repeat expansions that cause the disease. MBNL1 and MBNL2 both bind these microsatellite repeat expansions, and this negatively affects their function (Ranum and Cooper,2006). The importance of Mbnl1 in the pathogenesis of DM was beautifully shown with a mouse Mbnl1 homozygous mutant that has myotonia and defective Clcn1 expression (Kanadia et al.,2003a). However, as Kanadia et al. point out, the entire spectrum of defects found in DM is not represented in the Mbnl1 mutant mouse. It is likely that Mbnl2 provides both redundant and independent functions compared with Mbnl1.

To determine the cause of the myotonic phenotype, we examined specific Clcn1 RNA splicing changes described in the Mbnl1 mutant mouse, and found a modest, but significant change in inclusion of exon 7a. This in contrast to the ∼50% inclusion of exon7a that was found with mouse Mbnl1 deficiency (Kanadia et al.,2003a). We performed quantitative real-time PCR using two different regions of the Clcn1 gene and demonstrated a marked reduction in Clcn1 mRNA expression in Mbnl2 homozygous mutant compared with wild-type skeletal muscle. Thus, compared with Mbnl1 deficiency, it is likely that Mbnl2 deficiency uniquely affects mRNA metabolism or causes splicing changes that involve other regions of Clcn1 pre-mRNA. Incorrect RNA splicing can lead to premature translation termination codons that cause loss of Clcn1 protein expression, or can result in dominant-negative truncated Clcn1 proteins that can bind and inhibit wild-type Clcn1 (Berg et al.,2004).

Clcn1 RNA expression abnormalities associated with Mbnl2 deficiency should affect Clcn1 protein expression. Western blot for Clcn1 protein expression in the vastus muscle did not reveal a significant difference in Mbnl2 protein expression. However, immunohistochemistry demonstrated that segments of skeletal muscle had markedly decreased Clcn1 protein expression. Patchy areas of Clcn1 deficiency may not result in a gross change in Clcn1 protein expression levels detectable by Western blotting. In addition, Clcn1 expression by Western blot detects expression of Clcn1 in cell types other than skeletal muscle. The segmental decrease in Clcn1 expression demonstrated in immunohistochemistry is likely due to RNA splicing alterations associated with Mbnl2 deficiency. Regional differences in expression or function of Mbnl2 in skeletal muscle may explain these segmental defects in Clcn1 protein expression.

Defective Clcn1 expression likely contributes to the myotonia in Mbnl2−/− mice. From a physiological perspective, myotonic discharges that are responsible for clinical myotonia may be due in part to recurring single-fiber potentials from a defective area of the skeletal muscle membrane (Bertorini,2002). Even a small area of abnormal action potential “re-entry” in a muscle fiber due to segmental defects in Clcn1 function could be all that is required for myotonia.

The ability of MBNL2 to regulate Clcn1 expression is consistent with the multiple functions of RNA binding molecules that can include regulating RNA splicing, RNA localization, translation, and degradation (Ule and Darnell,2006). Mbnl2 likely influences Clcn1 expression by regulating RNA splicing. Whether and how Mbnl2 directly influences Clcn1 by affecting other aspects of RNA processing requires further examination. The lack of splicing defects in other genes examined may be due to functional redundancy provided by Mbnl1.

The phenotype of this Mbnl2 mutant mouse is in contrast to an earlier report of a different Mbnl2 mouse mutant (Lin et al.,2006). Unlike our results, that Mbnl2 mutant mouse did not exhibit RNA splicing defects, skeletal muscle pathology, and physiologic features of DM such as myotonia. Unlike this Mbnl2 mutant that has a gene trap inserted in intron 2, the previously reported Mbnl2 mutant has a gene trap insertion in intron 4. This Mbnl2 mutant encodes a truncated Mbnl2 protein that only contains one complete zinc finger. The previously reported Mbnl2 mutant encodes a molecule that contains additional coding regions and two zinc fingers. We demonstrate using Northern analysis that this Mbnl2 homozygous mutant expresses significantly less Mbnl2 mRNA. Our Northern analysis used a full-length cDNA probe for Mbnl2 expression that includes the entire protein-coding region. Lin et al. used a 3′-UTR probe for their Northern analysis, which may underestimate the amount of residual wild-type Mbnl2 expression present in the homozygous mutant mouse skeletal muscle. Differences in the skeletal muscle phenotype may also be due to genetic background effects resulting from our using a 129/Ola X C57BL/6 mixed genetic background. However, Lin et al. also reported that their mice were also maintained in a similar genetic background. In addition, we also focused our attention on mice after 6 months, rather than 4 months of age, because the neuromuscular phenotype was more severe after that time point.

Given the nature of the gene trap mutation, it is possible that the Mbnl2 mutant mouse expresses low levels of endogenous Mbnl2 due to RNA transcripts that bypass the gene trap splice acceptor. This was demonstrated by the presence of a low level of the 5.3-kb full-length RNA transcript in the Northern analysis of Mbnl2−/− mouse skeletal muscle. This low level of wild-type Mbnl2 expression could attenuate the severity of a phenotype found with complete Mbnl2 deficiency. The gene trap should result in the production of a truncated Mbnl2 RNA that is fused with a neomycin resistance/β-galactosidase (β geo) reporter. By necessity, this mutant RNA transcript was expressed in the mouse embryonic stem cells to permit neomycin selection. In addition, skeletal muscle from Mbnl2 mouse stained blue with X-gal demonstrating that β-galactosidase was expressed. The absence of significant detectable truncated mutant Mbnl2 RNA on the Northern analysis is likely due to its expression at low levels. It is possible that a truncated Mbnl2 molecule could have a dominant-negative effect on wild-type Mbnl2 and potentially Mbnl1 function. We believed that this possibility was less of a concern given the absence of significant mutant truncated Mbnl2 mRNA on the Northern analysis. A dominant-negative effect of the truncated Mbnl2 mutant on Mbnl1 function was also less likely given the absence of eye cataracts and RNA splicing defects that were present in the Mbnl1 mutant (Kanadia et al.,2003a).

The more noticeable myotonia in mice greater than 6 months may be due to an independent role of Mbnl2 in mature or remodeling skeletal muscle that is not completely compensated by Mbnl1. In addition, the expression of Mbnl1 and Mbnl2 may differ with regard to skeletal muscle distribution, developmental stage, response to injury, or aging. We chose to examine 6-month-old mice because we found that older mice had an obvious decrease in motor activity apparent by observation of mice in the cage. This finding is similar to other reported mouse mutant models of skeletal muscular dystrophy where the abnormal phenotype manifests at older ages. For example, transgenic overexpression of human DMPK at young age displays only mild myopathic findings (Jansen et al.,1996). However, older 11- to 15-month-old mice demonstrate more severe skeletal muscle pathology, including myotonia that models human DM (O'Cochlain et al.,2004).

The late-onset phenotype noted in the Mbnl2−/− mutant mice is similar to the adult onset phenotype that is commonly seen in human DM. The classic form of DM is the adult-onset type where symptoms present in middle age and where signs of DM are detectable in the second decade (Machuca-Tzili et al.,2005). Many DM patients become severely affected in their fifth and sixth decades of life, and many die of respiratory failure or sudden cardiac death. The mechanism for the late presentation is not clear, and the examination of a late adult phenotype such as found in our Mbnl2 mutant mouse may be important for understanding this phenomenon. The importance of RNA splicing transitions in postnatal remodeling has been demonstrated in the mouse heart. For example, cardiac-specific knockout of the RNA splicing factor ASF/SF2 causes defective postnatal alternative RNA splicing transitions that results in a cardiomyopathy and death at 6 to 8 weeks after birth (Xu et al.,2005). Postnatal remodeling of skeletal muscle may explain the late-onset phenotype in mice and humans.

Our data demonstrate that Mbnl2 deficiency can lead to myotonia and skeletal myopathy, cardinal features of human DM. This further supports the hypothesis that sequestration of MBNL proteins contributes significantly to the pathogenesis of DM. In general, human DM is an adult-onset, heterogeneous, multisystem disorder that most likely results from expression of abnormal amounts of microsatellite repeat expansions. Silencing expression of these RNA repeat expansions may be a useful therapeutic strategy (Mahadevan et al.,2006). The pathogenic RNA found in DM likely affects multiple molecules, including MBNL2 that are important for the entire disease. With regard to the Muscleblind family members, analysis of MBNL1, MBNL2, and possibly MBNL3, will be important for understanding the pathogenesis and finding new treatments for DM.


Ethical Approval

The Animal Research Committee of the Office of Protection of Research Subjects at the University of California, Los Angeles approved all animal protocols.

Production of Mbnl2 Mutant Mice

Mbnl2 mutant mice were generated from the BayGenomics embryonic stem (ES) cell line RRK149 (Stryke et al.,2003; Skarnes et al.,2004). Clone RRK149 was injected into C57BL/6J blastocysts, and the mice were maintained on a 129/Ola × C57BL/6J genetic background. Tail biopsies (5 mm) for genotyping were obtained using sharp, sterile scalpel blades and while mice were under brief anesthesia with isoflurane. Genotype was determined by Southern blot analysis and PCR with specific primers for the wild-type and mutant alleles. Primers used for amplification of an intron fragment for Southern probe are as follows: forward primer 5′ TCTGGCTTTTGACTGAGAGC 3′, reverse primer 5′ CCCCAAGAGAATGACAGCACAT 3′. Mbnl2 genotyping primers: wild-type forward primer 5′ TTCCTTGTGCTAATGACCTGGCGAGCCTAT 3′, wild-type reverse primer 5′ CAGGAAGAGATACAGCAGAAATTATAGTGCC 3′, mutant forward primer 5′ CCATCTCATTTTCACACACACACACACCCC 3′, mutant reverse primer 5′ GATAGGTTACGTTGGTGTAGATGGGCGCAT 3′.

RNA Analysis

Total RNA was prepared with Trizol (Life Technologies). Tissue was obtained after euthanasia performed with isoflurane inhalation followed by cervical dislocation. Quantitative RT-PCR comparison of gene expression was determined with the Bio-Rad I-cycler. For analysis of alternative pre-mRNA splicing, the PCR was performed for 27 cycles, annealing temperature of 52°C, and the reaction was spiked with 10 μCi (α32-p)-dCTP. Primers used were as follows: Zasp, forward primer 5′ GGAAGATGAGGCTGATGAGTGG 3′, reverse primer 5′ TGCTGACAGTGGTAGTGCTCTTTC 3′; M-titin, forward primer 5′ GTGTGAGTCGCTCCAGAAACG 3′, reverse primer 5′ CCACCACAGGACCATGTTATTTC 3′; Clcn1, forward primer 5′ GGAATACCTCACACTCAAGGCC 3′, reverse primer 5′ CACGGAACACAAAGGCACTGAATGT 3′; Serca1, forward primer 5′ GCTCATGGTCCTCAAGATCTCAC 3′, reverse primer 5′ GGGTCAGTGCCTCAGCTTTG 3′; Tnnt3, forward primer 5′ TCTGACGAGGAAACTGAACAAG 3′, reverse primer TGTCAATGAGGGCTTGGAG 3′. RT-PCR products were electrophoretically resolved on 8% nondenaturing polyacrylamide gels and subsequent autoradiography. Blots were scanned and examined using ImageJ software. Statistical analyses used the Student's two-tailed t-test.

Quantitative real time PCR comparison of gene expression was determined using the Bio-Rad I-cycler. β2 Microglobulin2M) was used as RNA input control. PCR primers used include the following: Clcn1 1212 exon 10 Forward CCGAAAGCACAAGTGTCTCA; Clcn1 1593 exon 14 Reverse GGCCATGATCTCTCCTACCA; Clcn1 2047 exon 16 Forward TCAGAACTGCAGTCCCTCCT; Clcn1 2411 exon 19 Reverse CGTTGGAAGGTGGATGATCT; β2M Forward ATTCACCCCCACTGAGACTG; β2M Reverse TGCTATTTCTTTCTGCGTGC.

Immunohistochemistry, Antibodies, Western Analysis, and Histology

Vastus muscles of mice were dissected and fixed in 10% formalin overnight, paraffin embedded, and sectioned (6 μm). Hematoxylin/eosin staining and Masson trichrome staining followed the manufacturer's protocol (Sigma-Aldrich). We quantified the trichrome staining using video microscopy and ImageJ software. Trichrome-positive areas were calculated as a percentage of the total area of the microscopic field (vastus muscle from three mice of each genotype were analyzed). Skeletal muscles for cryosections were embedded in OCT and cut at 10 μm. For analysis of fiber size, all fibers in the microscopic fields were examined for that experiment, and diameters were examined at the midrange of the muscle. Cells were immunostained and/or Western blots were performed with antibodies against Clcn1 (ClC11-A, Alpha Diagnostic, San Antonio, TX; 1:100 dilution), Tubulin (Sigma, St. Louis, MO; 1:2,000 dilution), and Dystrophin (BD Transduction laboratories, 1:20 dilution). Secondary antibody was used at a 1:500 dilution (Molecular Probes, Eugene, OR). DAPI (1:1,000 dilution, Molecular Probes) was used to stain nuclei. Single-fiber cross-sectional area (CSA) was measured using Spot optical software on calibrated images of vastus muscles with dystrophin staining (×200 magnification). All microscopic images were obtained using either a Zeiss Axiovert 200 microscope or a Leica TCS-SP MP confocal microscope with fluorescent attachments. Images were captured with a CCD camera and presented using Photoshop 7.0.

Vastus skeletal muscle proteins were resolved on 10% sodium dodecyl sulfate polyacrylamide gels and transferred onto nitrocellulose. Membranes were blocked with 5% nonfat milk and immunoblotted with antibodies for Clcn1 and tubulin. Signals were detected using the ECL detection system (Amersham).

X-Gal Staining

Tissues were fixed for 15–45 min in a solution containing 1% formaldehyde, 0.2% glutaraldehyde, and 0.02% NP-40 in phosphate-buffered saline (PBS; pH 7.4). Fixed tissues were washed twice in PBS and subjected to staining for 4–16 hr in a solution containing 1 mg/ml X-Gal, 5 mM K3Fe(CN)6, 5 mM K4Fe(CN)6, and 2 mM MgCl2 in PBS.


Paraspinal, gastrocnemius, vastus, and biceps skeletal muscles were examined at rest following a gentle movement of the EMG needle to provoke muscle contraction. Approximately three needle insertions were performed in each muscle group. Mice were anesthetized for the brief procedure with inhaled isoflurane. Viking Select (Viasys, 2005) and a Sierra Wave 2 channel (Caldwell, 2006) electromyography machines were used. Recordings were taken from the muscle using a sterile disposable 25 × 0.3 mm needle.


We thank D. Black, D. Lang, J. Chen, and Y. Wang for review of this manuscript; Tom Vondriska, Samuel Lo Chen, and Stan Chou for technical assistance.