Altered MEF2 isoforms in myotonic dystrophy and other neuromuscular disorders

Abstract

Because of their central role in muscle development and maintenance, MEF2 family members represent excellent candidate effectors of the muscle pathology in myotonic dystrophy (DM). We investigated the expression and alternative splicing of all four MEF2 genes in muscle from neuromuscular disorder (NMD) patients, including DM1 and DM2. We observed MEF2A and MEF2C overexpression in all NMD muscle, including 12 MEF2-interacting genes. Exon 4 and 5 usage in MEF2A and MEF2C was different between DM and normal muscle, with DM showing the embryonic isoform. Similar splicing differences were observed in other NMD muscle. For MEF2C, missplicing was more pronounced in DM than in other dystrophies. Our data confirm dysregulation of MEF2A and MEF2C expression and splicing in several NMD, including DM. Our findings demonstrate that aberrant splicing in NMD is independent from expression of mutant repeats, and suggests that some aberrant splicing, even in DM, may be compensatory rather than primary.

FIGURE 1

Exon/intron structure and primers for MEF2 genes. Isoforms are from the ‘Known Genes’ track of the UCSC Genome Browser (http://genome.ucsc.edu/cgi-bin/hgGateway?clade=vertebrate&org=Human&db=hg18&hgsid=77028912). White boxes: constant exons; shaded boxes: alternative exons; crosshatched boxes: alternative extensions. Arrows indicate primers for RT-PCR (right arrows: forward; left arrows: reverse). Location of MADS-box is shown with locations of translational starts/stops. To evaluate alternative 5′/3′ exons, because flanking exons are not constant, two reactions were used with one primer common to both: for MEF2A (five isoforms), three cassette exons (2, 5, 9) were interrogated; for MEF2B (eight isoforms), two cassette exons (8, 10) and one alternative 3′ (alt 3′ a); for MEF2C (five isoforms), two cassette exons (1e, 8) and one alternative 5′ exon (1d); and for MEF2D (four isoforms), two cassette exons (8, 11b) and one alternative 5′ exon (1b). For exon 11b, amplification products would be very large compared with those excluding it. This would strongly favor the smaller product. Therefore, a second reaction anchored in exon 11b was necessary to verify its presence in the product.

FIGURE 2

Dysregulation of MEF2-interacting genes. Heatmap and two-way hierarchical cluster analysis showing dysregulated expression of MEF2-interacting genes in all DM and non-DM NMD muscle (compared with N, normal). Blue represents underexpression, and red indicates overexpression. The normal samples cluster to the left, whereas there is no separation of the NMD samples by disease type. Sample names are in the format ‘Category_Sex_Age_Biopsy location_Identifier.’

FIGURE 3

MEF2C assay 2. (A) Assay design showing intron–exon structure of the fragment assayed. The 217-bp isoform was the predominant isoform in normal samples, whereas the 224-bp isoform, predicted to contain exon 5a instead of exons 4a/4b, was predominant in DM and essentially the only isoform present in myoblasts and fetal muscle. (B) Boxplot generated by the Wilcoxon rank sum test for percentage of 217-bp isoform. Comparisons were as follows: all NMDs vs. N, P = 0.00068; DM1 and DM2 vs. N, P = 0.00909; DM1 and DM2 vs. other NMDs, P = 0.00021.

FIGURE 4

MEF2A assay 2. (A) Assay design showing the intron–exon structure of the fragment assayed. The 269-bp isoform was predominant in normal samples, whereas the 264-bp isoform, predicted to contain exon 5a instead of 5b, was predominant in DM skeletal muscle, myoblasts, and fetal muscle. (B) Boxplot generated by the Wilcoxon rank sum test for percentage of 264-bp isoform. Comparisons were as follows: all NMDs vs. N, P = 0.00077; DM1 and DM2 vs. N, P = 0.01212. However, there was no significant difference between DM1/DM2 and the other NMDs.