Genetic localization of Bethlem myopathy

Bethlem myopathy (early-onset benign autosomal dominant limb-girdle myopathy with contractures) is a hereditary myopathy with slowly progressive muscular atrophy and weakness, and contractures of multiple joints. ^[1,2] Onset of weakness is usually either in the neonatal period, presenting with hypotonia, or in early childhood. ^[3] There is generalized, slight atrophy of the musculature and diffuse, mild weakness, proximal more severe than distal. Involvement of the facial muscles is absent. Nearly all patients have flexion contractures of the interphalangeal joints of the last four fingers, elbows, and ankles. ^[2,4] Contractures of the metacarpophalangeal joints, wrists, knees, hips, and shoulders are less common. In contrast with X-linked Emery-Dreifuss muscular dystrophy, contractures of the neck and spine are rarely present. ^[2] The contractures, present from onset of weakness onwards, ^[2,4] give little or no functional impairment in most patients. ^[2]

In spite of slowly ongoing deterioration of muscle power, ability to work is preserved until old age. ^[2] Many patients remain ambulant, often with the help of a cane, whereas others become wheelchair-dependent for outdoor transportation (manuscript in preparation). There is no cardiac involvement ^[2,5] and life expectancy is unaffected. ^[1] Ancillary investigations such as electrophysiologic studies and muscle biopsy indicate a myopathic origin of the disease. ^[2]

Bethlem myopathy has an autosomal dominant mode of inheritance with complete penetrance. A previous linkage effort failed to detect genetic linkage. ^[6] In the current study we carried out a genomewide search with highly polymorphic microsatellite markers in six Dutch families.

Methods.

Results.

Initially, we analyzed apparent candidate regions like the dystrophin-like protein (utrophin) gene on chromosome 6q24, the limb-girdle muscular dystrophy 1 locus (LGMD1) on 5q22.3-31.3, LGMD2A on chromosome 15q15.1-q21.1, LGMD2B on chromosome 2p16-p13, and LGMD2C on chromosome 13q12 without detecting linkage. A total of 52 microsatellite markers from all autosomes were examined without finding linkage, thereby excluding roughly 30% of the autosomal genome (two-point lod scores Z less than equals minus 2.0, data not shown).

Significant linkage was found with the 21q22.3 marker D21S171 Table 2. Additional markers (ordered from telomere to centromere) COL6A1, CD18, PFKL, D21S1259, D21S49, D21S212, D21S1260, D21S266, HMG14, and D21S211, all showed linkage Table 2. PFKL yielded the highest pairwise lod score, Z_max equals 6.86 at theta equals 0.03. The maximum two-point lod score obtained within one family (Family B) with this marker was 3.76 at theta equals 0.00. Each marker displayed at least one recombination.

The families were analyzed for genetic heterogeneity by means of the HOMOG program. To obtain maximal informativeness of each analysis, the following tightly linked markers were constructed into haplotypes: D21S171 and PFKL (genetic distance plus minus 2 cM), D21S1259 and D21S49 (plus minus 1 cM), and D21S1260 and D21S266 (0 cM). ^[13-15] In each analysis the overall maximum likelihood of either homogeneity or heterogeneity was obtained with alpha equals 1.00 (alpha is the proportion of families showing linkage to the marker). For instance, for COL6A1, the maximum likelihood for alpha equals 0.80 was 9.69 times 10⁴ at theta equals 0.00, whereas the overall maximum likelihood was 1.86 times 10⁶ (theta equals 0.05 and alpha equals 1.00). For COL6A1, D21S171 plus PFKL, D21S1259 plus D21S49, and D21S212 the overall maximum likelihoods were found at theta equals 0.05, and for D21S1260 plus D21S266 at theta equals 0.10. These data suggest genetic homogeneity; however, the families are not sufficiently large to prove that only one gene is involved.

The alleles of all tested 21q22.3 markers were constructed into a haplotype to obtain the most probable localization of each crossover event. Telomeric markers (COL6A1, CD18, and PFKL) showed fewer recombinations than the more centromeric markers (see Table 2). From this haplotype construction it was apparent that in one patient (in Family A) all markers including the most distal marker, COL6A1, recombined with the disease. Clinical reassessment affirmed that this individual was indeed affected; he had flexion contractures of the interphalangeal joints. A second blood sample was obtained and DNA extracted from this sample was retested for the COL6A1 polymorphisms, yielding identical results to the first sample.

Discussion.

The linkage data of the present study indicate that the Bethlem myopathy locus is located on the distal part of chromosome 21q. A two-point lod score exceeding 3.0 in one family is compelling evidence for linkage, as is the case for Family B with COL6A1, CD18, PFKL, and D1S212. Other individual family lod scores were lower than 3.0. Because the statistical power of a homogeneity test with linkage data like ours is limited, ^[12] we cannot conclude definitely that all pedigrees are linked to chromosome 21q22.3. However, the HOMOG results indicate that genetic homogeneity is more likely than heterogeneity.

From haplotype constructions all families contribute to linkage. Distal markers give rise to fewer recombinations than proximal markers. Taken together with the recombination between the markers and the disease in one patient from Family A, the most likely localization is distal to COL6A1. Located on a telomere, markers are only available at one side of the locus, thus making it difficult to confine precisely by genetic analysis the candidate region. COL6A1 is located on the most distal NotI fragment of chromosome 21q with an estimated size of 0.8 Mb. ^[13]

The collagen type VI genes might be considered good candidate genes for Bethlem myopathy because of the involvement in binding of the basal lamina to extracellular matrix glycoproteins, ^[16,17] the expression in skeletal muscle, ^[18] and the widespread occurrence of contractures in Bethlem myopathy. The collagen type VI protein consists of three peptides: collagen 6a1 (COL6A1), 6a2 (COL6A2), and 6a3 (COL6A3). ^[19] The COL6A1 and COL6A2 genes on the distal end of chromosome 21q, ^[19] have probably arisen from a duplication event. ^[20] On the NotI contiguous map of the IVth International Workshop on Human Chromosome 21 COL6A1 is shown as the most distal gene. ^[13] This orientation and the recombination with COL6A1 observed in one patient probably exclude the COL6A1-2 genes as disease genes of Bethlem myopathy.

The only identified gene telomeric of the COL6A1-2 cluster on chromosome 21q is the beta subunit of protein S-100 gene (S100B). ^[13] As it is mainly expressed in glial cells and not in skeletal muscle, ^[21] this gene is not a good candidate for Bethlem myopathy.

In conclusion, the present study has assigned Bethlem myopathy to the telomere of chromosome 21q. Most likely, the disease gene is an as yet unidentified gene distal of COL6A1. Bethlem myopathy displays typical clinical characteristics with a virtually constant expression. Nonetheless, there are some reports of limb-girdle myopathies with contractures displaying some atypical features. ^[22,23] The identification of the genetic defect of Bethlem myopathy will help to delineate the nosologic classification of those families.

Acknowledgments

We are indebted to the family members who participated in this study. We would like to thank Dr. G.J. te Meerman, Department of Medical Genetics, University of Groningen, The Netherlands, for help with the HOMOG program.