Article Text

Download PDFPDF
Methylation of the CpG sites in the myotonic dystrophy locus does not correlate with CTG expansion size or with the congenital form of the disease
  1. C Spits1,
  2. S Seneca1,2,
  3. P Hilven1,
  4. I Liebaers1,2,
  5. K Sermon1
  1. 1Department of Embryology and Genetics, Vrije Universiteit Brussel (VUB), Laarbeeklaan 101, 1090 Brussels, Belgium
  2. 2Centre for Medical Genetics, Universitair Ziekenhuis Brussel (UZ Brussel), Laarbeeklaan 101, 1090 Brussels, Belgium
  1. Correspondence to Claudia Spits, Department of Embryology and Genetics, Vrije Universiteit Brussel (VUB), Laarbeeklaan 101, Brussels 1090, Belgium; claudia.spits{at}


We have studied the methylation status of the sequence 152 nucleotides upstream of the CTG repeat of the DM1 locus in patients' peripheral blood. We used the methylation-sensitive endonucleases SacII, HpaII and HhaI, followed by PCR. This allowed to correlate the methylation status of each CTG allele with its size. Contrary to previous findings, only the SacII site is often but not always differentially methylated among expanded CTG alleles. Importantly, this methylation was not restricted to congenital DM1, nor to large expansions, as it was also present in DM1 patients with a classical phenotype and various expansion sizes. On the other hand, we did not find any methylated alleles on the HhaI and HpaII sites, as was reported by Steinbach et al, which is in line with the results of Shaw and collaborators. The size range of the repeat expansions with methylation was from as small as 300 to as large as 2800 repeats.

  • Molecular genetics
  • neuromuscular disease

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Myotonic dystrophy type 1 (DM1, [MIM 160900]) is a dominantly inherited multi-systemic disorder, with an incidence of approximately 1 in 8000 individuals. DM1 has a complex phenotype, with symptoms including progressive skeletal muscle wasting, impaired muscle relaxation (myotonia), cardiac conduction defects resulting in arrhythmias, early-onset iridescent cataracts, insulin resistance and hyperinsulinemia.1 These symptoms are caused by several mechanisms, which include the aberrant processing of the DMPK mRNA, leading to decreased levels of DMPK,2 toxic gain of function of the mRNA,3 decreased expression of the neighbouring SIX5 gene4 5 and disrupted alternative splicing of numerous other genes.6

DM1 is caused by a CTG expansion in the 3′ untranslated region of the DMPK gene located on chromosome 19q13.3. Expanded CTG tracts show both somatic and germinal instability; unaffected individuals have from 5 to 40 CTG repeats, whereas individuals with the disease carry >100 repeats. Disease severity and age of onset correlate with repeat size, and expansions of >1500 repeats often result in a severe congenital form of DM1.

CpG methylation has been suggested to play a role in the stability of repetitive sequences in general7 8 and more particularly in the behaviour of the CTG repeat in the DM1 locus9–11 and in the pathogenesis of the disease.9 12 13

The relationship between the semiology of the disease and the methylation of the CpG sites in the DM1 locus was first investigated by Shaw and collaborators,12 who could not find any clear correlation between the methylation and the age of onset or the severity of the disease. On the other hand, Steinbach and coworkers found that several CpG sites close to the CTG repeat were fully methylated in blood samples containing expansions >1000 repeats. They also reported that since these expansions appeared as single well-defined signals on Southern blot, the methylation of these CpG sites would correlate with triplet repeat stability. This idea was later supported by in vitro and in silico work.7 8 10 11 The work of Filippova built further on the idea that methylation would be related to the pathogenesis of congenital DM1 (CMD). The authors discovered that there are two binding sites for the zinc-finger protein CTCF flanking the CTG repeat and that the hypermethylation of these sequences, which coincide with the region described by Steinbach and collaborators,9 would be incompatible with the binding of the CTCF and its proper function mediating the inhibition of promoter–enhancer interactions by insulator elements.14 Conversely, Libby and collaborators15 found that CpG methylation of CTCF binding sites may rather lead to repeat instability instead of stability as previously suggested.

In this work, we studied 8 of the 18 CpG sites in 152 nucleotides upstream of the CTG repeat of the DM1 locus (see figure 1). These sites are part of the CpG sites investigated by Steinbach and collaborators and did not include the sites affecting the CTCF binding.13 We did not investigate the distal SacII site, located over 1 kb upstream of the CTG repeat, which Steinbach and collaborators reported as constitutively methylated, nor the HpaII and HhaI sites located in this same region. We analysed our samples in duplicate by methylation-sensitive endonuclease digestion, followed by PCR. The PCRs were designed so that they included both the restriction site and the CTG repeat and were performed on small quantities of DNA. Consequently, smears could be resolved into individual bands, and sizing was much more accurate than by using Southern blot on total genomic DNA. Furthermore, it enabled us to correlate the methylation status with the exact length of each allele, which would not have been possible with Southern blot on total genomic DNA or bisulphite sequencing. It is important to notice that in the case of the HpaII restriction enzyme, the studied region contained six sites. This means that we could not distinguish the methylation state of each independent site but could only asses whether at least one of the sites was unmethylated.

Figure 1

Schematic overview of the sequence analysed in this study and its relationship to the genomic regions studied by Shaw et al12 and Steinbach et al.9 The figures are modified versions of the original figures found in the respective publications. (A) Genomic region studied by Shaw et al.12 Three fragments were created using EcoRI and EcoRV and then restricted by HpaII. The results showed that at least some of the HpaII sites contained unmethylated cytosines, independently of the form of the disease or expansion size. (B) Genomic region studied by Steinbach et al.9 The samples were restricted using combinations of SacI and HindIII, and SacII, HpaII or HhaI. The results showed that the upstream SacII site was constitutively methylated and the downstream SacII site only methylated on very largely expanded alleles of patients with congenital DM1. The HpaII and HhaI sites were also all methylated in these patients. (C) Sequence studied in this paper. Genomic DNA was restricted using SacII, HpaII or HhaI and amplified by PCR using the primers indicated on the sequence as arrows. The highlighted sequence corresponds to the CTCF binding site 1.13 The results showed no correlation between the methylation of the SacII site and the disease status of the patient or the expansion size. No methylation of the HhaI site was detected, and the results for HpaII suggested that there was at least always one unmethylated site present in the sequence.

The aim of this study was to establish the threshold for the expansion size necessary for the hypermethylation of the CpG sites close to the CTG repeat and this within one DNA sample. We also aimed to investigate the possibility that hypermethylation is patient- or allele-specific.

DNA samples from peripheral blood from 22 DM1 patients were digested with the methylation-sensitive restriction enzymes SacII, HhaI and HpaII (Biolabs, Westburg, Leusden, The Netherlands). Both digested and undigested samples were analysed by PCR. In the case of the SacII-digested samples, we used the primer set indicated in figure 1 as SacII primer and reverse primer. The DNA input was of 200 pg. For the HhaI and HpaII samples, we used the HhaI–HpaII and the reverse primers. Initially, the DNA input was of 200 pg, but after the first results, DNA input was increased to 2 ng to ensure the detection of possible uncleaved alleles present at a very low quantity. The PCR protocol was further performed as previously described.16 The expansions were visualised by denaturing Southern blot using a GAC probe, and the size calculated by comparison to two molecular weight markers (markers VI and VII, Roche diagnostics, Vilvoorde, Belgium).

Table 1 shows the results for the 22 DM1 patients included in this study. The table outlines the results of the PCR before and after endonuclease cleavage by SacII, along with the disease status and inheritance of the mutation in all those patients for which these data were known. The cohort of patients includes cases of congenital DM1, infantile and classic DM1, both paternally and maternally inherited. The age at sampling ranges from birth to 43 years, and expansions were between 180 to 2800 repeats. Figure 2 shows an example of the Southern blot results after SacII restriction enzyme cleavage and PCR. The results for HhaII and HpaII cleavage and the complete data for the SacII digestion can be found in the supplementary figures 1–3.

Table 1

Details of the studied patients and results after SacII cleavage

Figure 2

Example of the Southern blot results after SacII restriction enzyme cleavage and PCR. Lanes 1, 2, 3, 4 and 5 show restricted samples, lanes 1′, 2′, 3′, 4′ and 5′ are the same DNA samples, but unrestricted. These numbers correspond also to the patient identification numbers found in table 1. The lane marked with MWM contains the molecular weight marker VII (Roche Diagnostics, Vilvoorde, Belgium). Unrestricted samples still contain the wild-type allele (wt), whereas in the restricted samples, the wt is completely cleaved. This can be used as an internal restriction control.

After HhaI endonuclease cleavage, none of the DNA samples showed amplification, whereas the undigested samples worked as a positive control for the PCR. This meant that all alleles were unmethylated, the wild-type alleles and all of the expansions. We verified that the lack of amplification in the digested samples was not due to DNA degradation by amplifying the material with a PCR for four short tandem repeats located on chromosome 19q13.41, 19q13.33, 19p13.3 and 19p13.2 (tandem repeats located using the database The primer sequences and PCR conditions are available upon request. The results showed that the DNA was intact in all samples and proved that the restriction by HhaI had been specific for its target. These results suggest that contrary to the findings of Steinbach and collaborators, this CpG site is constitutively unmethylated and is not necessarily methylated in expansions >1000 repeats or in congenital DM1 patients.

Identical results were obtained after cleavage with HpaII. The wild-type and the expanded alleles were unmethylated (for at least one of the six HpaII sites in the sequence) in all the samples studied. These results are in concordance with the findings of Shaw and collaborators12 and contradict those of Steinbach and co-workers.9

From our results, it seems that of the CpG sites we studied, the SacII site immediately upstream of the repeat is the only one in the region that can be differentially methylated between wild-type alleles, which are never methylated, and expanded alleles, which may be methylated. Our data are difficult to harmonise with the previously published study.9 We did not find a correlation between the methylation status of this site on one hand and the allele size on the other hand with the presence of a smear on the Southern blot—indicative of tripet repeat instability—as was suggested by Steinbach et al. Alleles as small as 400 repeats could be methylated (patient 20) and alleles of 1600 repeats unmethylated (patient 8), indicating that there is no correlation between the allele size and the methylation status. We found several patients showing intermediate methylation patterns (some alleles were methylated, while others were not) and even patients without any methylation at this site. A patient could present a smear in which all alleles appeared to be methylated (ie, patient 7), indicating that the methylation status apparently does not correlate with the stability of the repeat. However, it is clear that the presence of CTG length mosaicism either with or without methylation does not necessarily make one the cause or effect of the other. The time during which one occurred may or may not have coincided with the other. Furthermore, the link between methylation of this site and congenital DM1 does not seem to be as strict as previously assumed. It is true that congenital DM1 cases analysed here showed methylation of the SacII site in all the expanded CTG alleles (patients 4 and 6), but non-congenital DM1 patients also presented it (patients 5 and 7). Furthermore, we did not find the correlation previously described9 between the methylation of the SacII site and of the HhaI and HpaII sites, since we never found methylation in the latter two, not even in the congenital DM1 patients. The fact that we did not find a methylation pattern specific to the CMD patients contradicts the hypothesis that differential methylation may be responsible for the distinct features of CMD by modulating the binding of CTCF in the DM1 locus.13 It is, however, important to bear in mind that we did not study the methylation of all the CpG sites present in the CTCF binding sites and that methylation may still play an important role in the pathogenesis of DM1.

These discordances do not seem attributable to incomplete restriction or technical differences. In the works of Shaw et al12 and Steinbach et al9 as in ours, incomplete restriction would have been detected by the incomplete cleavage of the wild-type allele. To ensure that the results for HpaII and HhaI were not due to the degradation of the samples after restriction, we confirmed the presence of intact DNA by PCR for microsatellite markers. The fact that, in our work, we analysed PCR products by Southern blot whereas Shaw et al12 and Steinbach et al9 performed direct Southern blotting of the restricted DNA only adds to the resolution of the expanded alleles. Furthermore, it is possible that our method is more sensitive in picking up a few unrestricted alleles. Nevertheless, for the HhaI and HpaII digests, we never found any uncleaved allele.

In the work of Shaw et al,12 it is suggested that the patient's age might play a role on the methylation levels in CMD. The authors suggest that since their CMD patients were rather of mature age, the methylation could have been lost. It is important to bear in mind that these authors were investigating a possible relationship between imprinting of the DM1 locus and CMD and hypothesised that this imprinting would only be of importance in utero and would be subsequently lost after birth. In Steinbach's paper, the oldest CMD patient is 24 years and shows full methylation of the expanded allele for all the SacII, HhaI and HpaII sites. Shaw et al12 did not mention the exact age of the patients, but if they were >24 years, their hypothesis would harmonise the difference in results between Shaw's and Steinbach's papers, but not with ours, where the two CMD patients were very young and showed no methylation for the HhaI and HpaII sites. An age effect on the classical DM1 patients can be ruled out, as Steinbach's work as ours has patients of all ages (Steinbach's range was 19–62, ours 4–45 years). Furthermore, there does not seem to be a parent of origin effect.

In conclusion, we have studied the methylation status of the proximal sequence upstream of the CTG repeat of the DM1 locus. From our results, we establish that, contrary to previous findings, only one of the studied sites is differentially methylated in wild-type and expanded CTG alleles and that this methylation does not strictly correlate with the length of the allele, nor with the disease status of the patient. Further research on this topic is necessary to elucidate the reasons for the discrepancy with some of the previously published results.


The authors wish to thank their colleagues at the Centre for Medical Genetics for the fruitful collaboration and Prof. C.E. Pearson for his constructive comments on the manuscript.



  • Funding This work has been supported by grants from the Fund for Scientific Research Flanders (Fonds voor Wetenschappelijk Onderzoek Vlaanderen (FWO Vlaanderen), Egmontstraat 5, 1000 Brussels, Belgium. CS is a postdoctoral fellow at the FWO Vlaanderen.

  • Competing interests None.

  • Patient consent Obtained.

  • Ethics approval This study was conducted with the approval of the Commissie Medische Ethiek Universitair Ziekenhuis Brussel.

  • Provenance and peer review Not commissioned; externally peer reviewed.