Neuromuscular Disorders
Volume 8, Issue 2 , Pages 126-130, April 1998

44th ENMC International Workshop: Facioscapulohumeral Muscular Dystrophy: Molecular Studies:

19–21 July 1996, Naarden, The Netherlands

Clinical Genetics Unit, Institute of Child Health, Bristol Children's Hospital, St. Michael's Hill, Bristol BS2 8BJ, UK

Article Outline

 

Back to Article Outline

1. Aims 

The principal aim was:

(i) to collate data from each centre to determine the specificity and sensitivity of a potential diagnostic DNA test for FSHD.Other aims were:

(ii) to confirm correlation between phenotypic severity and genotype;

(iii) to assess the place of mosaicism in new mutation;

(iv) to document cases or families which are exceptions to the rules.

Back to Article Outline

2. Background 

FSHD is one of the most important neuromuscular disorders, demonstrating dominant inheritance and affecting 1/20 000 of the European population. The disease gene is not yet known, but in at least 80% of cases FSHD is associated with deletion of an integral number of copies of a tandemly repeated 3.3 kb sequence in a subtelomeric region at 4q35, producing a shortened DNA fragment detected by probe p13E-11 (locus D4F104S1) [1]. De novo cases of FSHD (which comprise 10% of all cases) are usually associated with de novo occurrence of a deleted fragment, but in some cases a weaker band in one parent suggests parental somatic and germline mosaicism. There are also cases where an `unaffected' parent shares a fragment in full dose with their affected offspring, suggesting either non-penetrance or a premutation phenomenon [2]. Evidence has suggested that the severity of presentation and age at onset correlate with residual fragment size [2]. The possibility of clinical anticipation giving earlier onset in successive generations has also been raised, despite the fragment size remaining constant within a family 2, 3.

The presence of a deleted fragment has the potential to be a specific diagnostic test for 4q35-FSHD. Unfortunately, interpretation is often confounded by a cross-hybridising homologous 3.3 kb tandem repeat sequence at 10q26 which includes polymorphic alleles in the same size range as the deleted 4q35-alleles of FSHD. In October 1995, Deidda and Felicetti from Rome introduced to the FSHD International Working Group a technique of double digest with B1nI in addition to EcoRI, for distinguishing deleted 4q35-alleles from normal polymorphic alleles at 10q26, since only the latter would normally have a B1nI restriction site within each 3.3 kb repeat [4]. Since October 1995, each Group had had the opportunity to test the validity of the `double digest' technique as a specific diagnostic test for 4q35-FSHD. The specificity and sensitivity of this technique could only be assessed reliably from knowing the frequency and size distribution of deleted fragments in a large control population, and through pooling data from FSHD families, and in particular for any exceptions to a very promising general rule.

The disease gene in FSHD is not yet known, but current hypotheses favour a position effect operating on an adjacent (and probably proximally placed) putative FSHD muscle gene. Physical molecular research has concentrated on cloning expressed sequences from this region [5].

Reports of a small number of families demonstrating apparent recombination between the disease gene and 4q35 markers have led to the assumption of a second locus [6]. The proportion of families which appear to be `unlinked' needed to be clarified, and also whether any appear to have a `deleted' 4q35 fragment.

Back to Article Outline

3. Results/pooled data 

3.1. Specificity and sensitivity of deleted 4q35 fragments using double digest (DD) 

3.1.1. Control data 

Normal controls (622) had been studied by five groups, but unfortunately each group had chosen a different upper threshold size for classification of a `small' fragment (Table 1). Overall, 0/622 controls had DD fragments of <28kb, but 3/310 controls (1%) had DD fragments of <35 kb, and a probable higher proportion had fragments <38 kb. It was evident that there is a need for standardisation of the method for sizing fragments, particularly between 30 and 40 kb, where the standard error in measurement is at least 3 kb.

Table 1. Specificity of EcoRI/B1nI fragments as a diagnostic test: European controls
CentreEco/B1n fragment size threshold (kb)Controls below threshold
Cardiff<380/200
Bristol<400/59 (+1)a
Rome<280/135
Leiden<352/50
Tokyo<280/177
Total 2–3/622

aOne additional case (a 31 kb fragment on Eco/B1n double digest) had been identified in a control subject (an unaffected spouse of an FSHD patient), outside the series studied.

3.1.2. Affected subjects 

Of 523 families with FSHD, 491 (94%) showed both a DD fragment <35 kb and compatibility with 4q35 linkage, without recombination. In only 13/523 (2.5%) did no short fragment remain following DD; while apparent recombination with p13E-11 or possible non-linkage to 4q35 was observed in 19/523 families (3.6%) at very most (Table 2). A larger fragment of 38 kb had been observed to co-segregate with a scapulohumeral phenotype (without facial weakness) and with 4q35 markers in one family previously reported [7]. No family had been observed which demonstrated 4q35 linkage of the phenotype but did not have an associated shortened 4q35-linked fragment.

Table 2. FSHD: exception families (European data)
CentreFamilies where affected members show:
Total familial case familiesNo short fragment with EcoRI/B1nIRecombinationEither: recombination or unlinkedUnlinked to 4q35
Rome55? 10–0–? 2
Leiden13162–0–? 6
Pavia1510––
Cardiff103111–
Tokyo3901–? 3––
Paris1564–––
Marburg24-–? 1? 3
Total52312–134–61–20–11
% 210.40→2

? n, maximum number possibly observed, but data in these individual `exception' families was not yet certain.

3.1.3. Isolated cases 

Notwithstanding that the clinical diagnosis in an isolated case can be less certain, 128/139 (92%) of clinically verified single cases are associated with DD fragment of <35 kb. As a single test, it was concluded that testing for a DD fragment of <35 kb in a person with clinical symptoms compatible with FSHD has a sensitivity of at least 92% and a specificity of 99% as a diagnostic test for the condition (Table 3). The test could not however be applied to an asymptomatic general population at low risk, as the specificity would be low.

Table 3. Proportion of EcoRI/B1nI-resistant fragments in FSHD families, in clinically typical isolated cases and in controls
FSHDControls
Families Isolated cases
Eco/B1n DD fragment <35 kb510 1283
Total 638
No DD fragment <35 kb13 11307
Total 24
Total (both)523 139310
Total 662

As a test for FSHD, a DD fragment of <35 kb gives: test sensitivity=638/662=96% (however, for isolated cases: 128/139=92%). Test specificity=307/310=99%.

3.1.4. DD-sensitive fragments associated with FSHD (4q/10q exchange) 

A proportion (6/139 (4%)) of isolated cases were noted to have a shortened (<35 kb) EcoRI single digest (SD) fragment which was sensitive to DD. This proportion can be compared with the previously noted figure of 10% (8/80) for the proportion of 10q26-linked fragments of <28 kb in the general population [8]. Using pulse field gel studies (PFGE) to show fragments of all sizes detected by p13E-11, the Leiden group had noted that in several such cases the affected person had only one B1n-resistant fragment, and three B1n-sensitive fragments, instead of the expected two of each type. Conversely, in noting that some normal control subjects on conventional Southern analysis have three B1n-resistant fragments, they found on PFGE that these people did indeed have three B1n-resistant fragments but only one B1n-sensitive fragment. The Leiden group therefore proposed that a polymorphism exists in the general population (and in affected FSHD subjects) for chromosomal exchange between the D4F104S1 (D4Z4) repeat at 4q35 with that at 10q26. Ten percent of the population appear to have a B1n-resistant fragment on one of their chromosome 10s, and 10% have a B1n-sensitive fragment on one of their chromosome 4s [9]. FSHD occurs when the fragment on chromosome 4 is truncated, whether it be of B1n-resistant or B1n-sensitive type. Allowing for each person having two copies of chromosome 4, approximately 5% of FSHD cases can be therefore expected to be associated with a shortened EcoRI fragment of <32 kb which is sensitive to DD by B1nI, rather than resistant. The Leiden group also presented evidence in support of a shortened B1n-sensitive fragment co-segregating with FSHD in a small family compatible with 4q35-linkage. The group's findings have been published subsequently [9]. Table 4Table 5, derived from this data, show a postulated distribution of fragment types and sizes causing FSHD, compared with the distribution derived similarly for controls.

Table 4. Postulated proportions (%) of 4q35-attached fragment type and size `causing' FSHD (derived from European combined data)
B1n-resistant (chromosome 4-type)B1n-digested (chromosome 10-type)a
<30 kb31-37 kb>38 kb<30 kb31–37 kb>38 kb
8393.540.5–

aHas only one B1n-resistant fragment on PFGE.

Table 5. Postulated distribution (%) of p13E-11 fragment size (smallest 4q35-fragment; or smallest fragment); derived from European combined data
Smallest fragment on EcoRI single digestSmallest fragment on EcoRI/B1nI double digest
<30 kb31–37 kb>38 kb<30 kb31–37 kb>38 kb
Unequivocal FSHD871038398
Controls52075<0.22a98

aTrisomic, three fragments on PFGE which are B1n-resistant.

Therefore, by using PFGE to count the total number of B1n-resistant and sensitive fragments in those FSHD cases who have a short fragment (<35 kb) which disappears on DD, the sensitivity and specificity of D4F104S1 as a diagnostic test for FSHD is raised to: 96.5% sensitivity, and >99% specificity (Table 6). The double digest technique therefore does provide a highly specific and sensitive test for FSHD in a symptomatic subject.

Table 6. Testing for FSHD by p13E-11
TestSensitivitySpecificity
EcoRI fragment <35 kbVery high (97%)Low (in 25% of controls)
Eco/B1n resistant fragment <35 kbNot so high (92%)High (in 1% of controls)
Eco/B1n digest with PFGE for fragments >31 kb or for B1n-sensitive fragments <31 kb, to detect 1 or 3 copiesVery high (96.5%)Very high (<0.2% in controls)
3.1.5. Non-linkage and recombination 

From 523 families with clinical FSHD, there were 11 at most, and yet possibly zero (0–2%) which could be unlinked to 4q35; otherwise recombination between FSHD and a shortened D4F104S1 DD fragment had been seen in only four to six families (0.8-1.1%). Within the context of a 4q35-linked family such cases remain difficult to explain unless clinical over-interpretation of minor signs is occurring, or in the laboratory interpretation, there is a confounding 10q26-linked shortened B1n-resistant fragment. There were a further one to two families where recombination or non-linkage could account for aberrant findings. However, overall no more than 19/523 families (3.6%) either showed recombinants or were unlinked, and in several of these there were still uncertainties over some of the data.

3.1.6. New mutation in FSHD 

Data was presented on 128 clinically new mutation cases of FSHD, defined as having a B1n-resistant fragment of ≤35 kb. Only in 87/128 (68%) was this fragment truly de novo, although in 26/128 (20%) it was seen at reduced density in one parent, who was presumed to be a somatic and germline mosaic. Predominantly this occurred in the mother (21/26=81%) of `mosaic' cases. In the remaining 12% (15/128) of de novo cases defined as above, the fragment was shared with one or other parent, who in all cases was believed to show no clinical sign of FSHD. In the 7/15 of these cases where the sex of the relevant parent was recorded, in three the fragment was shared with the father, in four with the mother. In none of these cases had both grandparents been available to determine if the fragment might be de novo in the parent, but the possibility of a pre-mutation seems enticing. Therefore it seems that 30% of clinical `new mutation' cases of FSHD are associated with somatic mosaicism or possible pre-mutation. The possibility that the mutation mechanism operates during mitosis as well as meiosis is intriguing, particularly in view of a possible role of 4q35/10q26 interaction /interchange.

3.2. Other clinical aspects 

3.2.1. Correlation between deletion size and severity measured by age at onset 

This correlation had been noted by all groups who had studied this, and therefore was confirmed as a real phenomenon. In general, most groups had also noted that new mutation cases tended to have the larger deletions (smaller residual fragment size).

3.2.2. Anticipation 

The possibility of clinical anticipation remained unresolved. At least two groups were aware of families where this did not appear to be observed.

3.2.3. Modifying factors 

The possibility of non-genetic or epigenetic factors involved in the control of clinical expression was furthered by the description from Dr. Rosella Tupler of a pair of monozygous twins who were discordant for severity of FSHD, but both sharing a shortened D4F104S1 fragment.

Back to Article Outline

4. Conclusions of phenotype/genotype aspects 

1.The presence of a short (<35 kb) B1n-resistant fragment alone provides a diagnostic test for FSHD, for which the specificity and sensitivity are both high, but if combined additionally with PFGE, the specificity and sensitivity are improved even further.

2.Recombinant cases (with p13E-11) or families unlinked to 4q35 comprise only a very small proportion of all cases.

3.Around 20% of new mutation cases may arise from mosaicism in one parent, but a further 12% may be due to parental non-penetrance or pre-mutation.

Back to Article Outline

5. Technical developments 

5.1. Use of polymerase chain reaction (PCR) 

Current genetic testing with p13E-11 uses Southern blot techniques. Arahata presented a PCR technique which uses primers for chromosome 4-specific sequences flanking the total 3.3 kb repeats, and can amplify up to four repeats (equivalent to an approximately 23 kb fragment on Southern blotting). Therefore only the repeats at 4q35 are seen. Additionally, use of B1nI determines whether the repeats are of chromosome 4-type or chromosome 10-type (or a hybrid of both). At present the technique is limited by the size limit for the PCR, but could already be useful in family studies and prenatal tests for the smallest fragment sizes.

Back to Article Outline

6. Efforts at gene cloning 

The 3.3 kb repeat does not contain an open-reading frame and any mRNA produced is assumed not to be translated. Jane Hewitt had cloned the telomeric region beyond the 3.3 kb repeats, but this has proved to have a lot of repeat sequences which cross-hybridize with chromosomes 14,22 and other acrocentrics, but does not exhibit open reading frames. One gene (FRG1) mapping proximal to the 3.3 kb repeats had already been cloned by van Deutekom [5]. Sequencing of the coding region in a typical FSHD patient who did not show a small fragment, did not unfortunately reveal any mutation in FRG1. Another gene clone with high homology to beta-tubulin genes had been identified, but tissue expression studies revealed no expression in adult muscle, and only low levels in fetal tissues. This did not seem to be promising as a candidate for the FSHD gene. The search continues, but it may be that the mechanism involved is more complex than we presently hypothesise, in terms of how the deletion is affecting the subtelomeric region.

Back to Article Outline

7. List of participants 

Sponsored and hosted by the European Neuromuscular Centre (ENMC), this workshop brought together 18 participants from 9 European groups and one from Japan who form part of the International Working Group involved in researching the molecular basis of facioscapulohumeral muscular dystrophy (FSHD).

K. Arahata (Tokyo, Japan)

E. Bakker (Leiden, The Netherlands)

O. Brouwer (Leiden, The Netherlands)

C. de Toma (Paris, France)

G. Deidda (Rome, Italy)

L. Felicetti (Rome, Italy)

R. Frants (Leiden, The Netherlands)

J. Hewitt (Manchester, UK)

P. Jardine (Bristol, UK)

J. Köhler (Marburg, Germany)

P. Laforet (Paris, France)

P. Lunt (Bristol, UK)

G. Padberg (Nijmegen, The Netherlands)

M. Rogers (Cardiff, UK)

R. Tupler (Pavia, Italy)

M. Upadhayaya (Cardiff, UK)

J. van Deutekom (Leiden, The Netherlands)

O. Vogels (Nijmegen, The Netherlands)

Back to Article Outline

Acknowledgements 

We are grateful to ENMC for financial support and to Professor Alan Emery (Research Director of ENMC) for his scientific advice and Mr. M. Rutgers and Ms. Janine de Vries for organisational help. This Workshop was made possible thanks to the financial support of the European Neuromuscular Centre (ENMC) and ENMC main sponsors: Association Française contre les Myopathies (France); Italian Telethon Committee (Italy); Muscular Dystrophy Group of Great Britain and Northern Ireland (UK); Vereniging Spierziekten Nederland (Netherlands); Muskelsvindfonden (Denmark); Deutsche Gesellschaft für Muskelkranke (Germany); Schweizerische Stiftung für die Erforschung der Muskelkrankheiten (Switzerland); Prinses Beatrix Fonds (Netherlands); Verein zur Erforschung von Muskelkrankheiten bei Kindern (Austria); Finnish Muscular Dystrophy Association (Finland) and ENMC associate members: Unione Italiana Lotta alla Distrofia Muscolare.

Back to Article Outline

References 

  1. van Deutekom JC, Wijmenga C, van Tienhoven EA, et al.  FSHD associated DNA rearrangements are due to deletions of integral copies of a 3.2 kb tandemly repeated unit. Hum Mol Genet. 1993;2:2037–2042
  2. Lunt PW, Jardine PE, Koch MC, et al. Correlation between fragment size at D4F104S1 and age at onset or at wheelchair use, with a possible generational effect, accounts for much phenotypic variation in 4q35-facioscapulohumeral muscular dystrophy (FSHD). Hum Mol Genet 1995;4:951–958; and Erratum; 4:1243–1244.
  3. Zatz M, Marie SK, Passos-Bueno MR, et al.  High proportion of new mutations and possible anticipation in Brazilian facioscapulohumeral muscular dystrophy families. Am J Hum Genet. 1995;56:99–105
  4. Deidda G, Cacurri S, Piazzo N, Felicetti L. Direct detection of 4q35 rearrangements implicated in facioscapulohumeral muscular dystrophy (FSHD). J Med Genet. 1996;33:361–365
  5. van Deutekom JCT, Lemmers RJLF, Grewal PK, et al.  Identification of the first gene (FRG1) from the FSHD region on human chromosome 4q35. Hum Mol Genet. 1996;5:581–590
  6. Gilbert JR, Stajich JM, Wall S, et al.  Evidence for heterogeneity in facioscapulohumeral muscular dystrophy (FSHD). Am J Hum Genet. 1993;53:401–408
  7. Jardine PE, Upadhyaya M, Maynard J, et al.  A scapular onset muscular dystrophy without facial involvement: possible allelism with facioscapulohumeral muscular dystrophy. Neuromusc Disord. 1994;4:477–482
  8. Bakker E, Wijmenga C, Vossen RHAM, et al. The FSHD linked locus D4F104S1 (p13E-11) on 4q35 has a homologue on 10qter. Muscle Nerve 1995;Suppl. 2:S39–S44.
  9. van Deutekom JC, Bakker E, Lemmers RJ, et al.  Evidence for subtelomeric exchange of 3.3 kb tandemly repeated units between chromosomes 4q35 and l0q26: implications for genetic counselling and etiology of FSHD1. Hum Mol Genet. 1996;5:1997–2003

PII: S0960-8966(98)00012-1

doi:10.1016/S0960-8966(98)00012-1

Neuromuscular Disorders
Volume 8, Issue 2 , Pages 126-130, April 1998

Article Tools