Editor—Familial hypertrophic cardiomyopathy (FHC) is a prevalent dominantly inherited disease characterised by unexplained hypertrophy of the heart muscle. The clinical manifestations are heterogeneous and the disease is a leading cause of sudden cardiac death among young, otherwise healthy people.1 More than 120 different mutations have been reported in the following eight genes encoding sarcomeric polypeptides given in parentheses: TNNT2 (troponin T),MYL3 (essential myosin light chain),MYH7 (β myosin heavy chain),MYBPC3 (myosin binding protein C),MYL2 (regulatory myosin light chain),TPM1 (α tropomyosin),ACTC (α cardiac actin), andTNNI3 (troponin I).2 3Furthermore, a disease locus on chromosome 7 has been linked to FHC, but the gene has not yet been identified.4 Additional disease genes probably remain to be discovered since two recent studies found that it was only possible to genotype 50-60% of the FHC population by mutation analyses of seven disease genes.5 6 In order to optimise risk stratification and management of FHC patients, it is important to identify all disease carriers, which is difficult by physical examination because of the age dependent penetrance of the disease. However, disease carriers may be identified by use of genetic diagnosis, although it is laborious because of the large number of disease genes and the pronounced allelic heterogeneity of the disease loci, with the majority of affected families having their own “private” missense mutation.7 In addition, genetic diagnosis is complicated by the fact that several amino acid polymorphisms occur in most of the FHC genes8 9 (unpublished observations). Given this complex genetic background, the use of linkage analysis can be beneficial as it may identify the most likely disease gene and provide evidence for exclusion of some or all of the other candidate disease loci even in small families.10

It was the aim of the present study to develop a firm basis for efficient use of linkage analysis in genetic diagnosis of FHC by a well founded selection of polymorphic markers defining nine FHC loci, including a refined genetic mapping of the troponin T gene in a 4 cM interval. For rapid analysis, multiplex PCR panels were developed comprising all markers selected. The feasibility of the method was evaluated by identification of mutations in three families of varying size.

The genetic mapping of TNNT2 was based on analysis of six informative CEPH pedigrees (obtained from CEPH, France: Nos 102, 884, 1347, 1362, 1413, 1416; 102 subjects) using a previously published intragenic insertion/deletion polymorphism localised in intron 4 (TNNT2-Ins/Del).11 A 340 bp fragment of TNNT2 including exons 3, 4, andTNNT2-Ins/Del (base position 181-520) was amplified using the primers: forward (F) 5′-GTGGCAGGCAGCGTGACTCCAC-3′ (the primer sequence was modified in accordance with our own unpublished sequencing results of intron 3 by omitting Gs in position 184, 190, and 198), reverse (R) 5′-CAGGATTTCCACATTGCTGA-3′. PCR was carried out as previously described3 with primer concentrations given in table 1 and an annealing temperature of 62°C. Multipoint linkage analyses were carried out using chromosome 1 DNA markers in the region previously reported to harbour TNNT2 12 13and the CEPH database version v8.1,14 essentially as previously described.15 A 16 point reference map was chosen in accordance with the Généthon genetic map where the marker loci are ordered with odds of at least 1000:1.16 CMAP analysis was carried out calculating the likelihood for any position ofTNNT2 with respect to the fixed map. The location score curve showed a peak location score of 92.9 between D1S2716 and D1S504, which corresponded to a multipoint lod score of 20.2. The location of TNNT2 in a 4 cM interval between D1S2716 and D1S504 was favoured by odds of more than 100 000:1 over the second most likely interval, that is, between D1S2716 and D1S2622 (complete data set is available upon request). This refined assignment was in accordance with previous localisations ofTNNT2. 12 13 17 18

The remaining FHC loci have all been localised by physical and/or genetic mapping with evidence sufficient to allow a well founded selection of flanking markers. The evidence for the regional localisation of the FHC loci and their relationships to flanking markers were obtained by a thorough integration of mapping information deposited at different databases including the Human Genome Sequencing Project18 and published reports. The results are summarised in fig 1.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Localisation of FHC loci in the human genome. The brackets indicate the most likely locations. Polymorphic markers flanking the disease loci are underlined. The markers used for linkage analyses (in italics) were selected on the basis of the highest achievable odds of their localisation, their HZ indices, and their ability to be incorporated in multiplex PCR panels. All markers depicted are localised on the Généthon genetic map with odds of at least 1000:116 except D11S1993 and D19S254 which are only part of the Marshfield map.30 * indicates that the marker could not be localised relative to the disease locus. The distances are in centiMorgans (cM).

MYL3,19 MYBPC3,20and TNNI3 21 have all been localised at various radiation hybrid (RH) maps with odds of 1000:1, 1000:1, and a lod >3, respectively. Polymorphic markers flanking the genes were all localised with odds of 1000:1 on the RH maps except the marker flanking TNNI3 (D19S218), which is a framework marker localised with odds of 300:1 on the Whitehead RH map.22 The assignment of MYBPC3was in accordance with previous localisations.23-25 MYBPC3 carries an intragenic polymorphic marker (MYBPC3-CA) within intron 20,26 which was characterised by analyses of 50 chromosomes from unrelated white Danish subjects. Five different alleles were identified with the following frequencies: allele 1 (282 bp) 0.12, allele 2 (284 bp) 0.16, allele 3 (286 bp) 0.34, allele 4 (288 bp) 0.32, and allele 5 (290 bp) 0.06. The observed HZ index was 52%. The disease loci on chromosome 7 and TPM1,which carry an intragenic marker (HTM-α-CA), have been localised by linkage analysis with odds of approximately 100:1 and 10 000:1 over the second most likely interval, respectively.4 15

MYH7 and ACTCcarry highly polymorphic markers within intron 1/intron 24 (MYO I/II-CA)27 28 and intron 5 (ACTC-CA),29 respectively. Both genes are part of the Marshfield genetic map.30 MYL2 is part of the chromosome 12 sequence released by the Human Genome Sequencing Project.18 The gene is part of contig NT_000612, sequence AC002351.1. The polymorphic marker D12S1343 is part of the same contig, sequence clone AC002352.1, and localised about 13 kb distally of MYL2.No proximal flanking markers within a reasonable distance ofMYL2 are part of the released sequence.18 However, it was possible to select D12S84 as the proximal flanking marker since the position ofMYL2 has recently been refined by haplotype analysis of a large FHC pedigree carrying aMYL2 mutation.5 The position ofMYL2 is in accordance with previous assignments.18 22

The markers selected for linkage analysis were incorporated in multiplex PCR panels and the outcome of several optimising procedures with respect to primer, template, and deoxynucleotide concentrations, composition of different markers in each multiplex, and number of amplification cycles are shown in table 1. PCRs were carried out using previously reported conditions.3 Before the amplified PCR products were analysed on an ABI prism 377 DNA sequencer (Perkin-Elmer Corp), the DNA concentration of the PCR products were estimated by comparing the intensity of the bands with a known concentration of the reference marker λ Dra by 3% agarose gel electrophoresis. In order to achieve reproducible peaks of the markers it was important to dilute the multiplex PCR products to a final concentration of about 0.4 ng/μl before loading the gel.

The linkage approach developed was evaluated in three FHC families shown in fig 2. Informed consent was obtained from each participant in accordance with local ethical committee guidelines. Echocardiography and standard 12 lead electrocardiograms (ECGs) were obtained and evaluated essentially as previously described.3 In brief, a person was considered to fulfil major diagnostic criteria if the maximal wall thickness (MWT) measured by echocardiography was >13 mm, if ventricular arrhythmia occurred during 48 hours of ambulatory ECG monitoring, or the ECG presented either major Q wave abnormalities, left ventricular hypertrophy, or marked repolarisation alterations. Subjects were considered to fulfil minor diagnostic criteria if their ECG presented minor Q waves in at least two leads. Minor echocardiography criteria were considered to be fulfilled by an MWT=13 mm. The participants in the investigation were classified phenotypically before DNA analysis. All subjects with clinical signs of FHC in families A and B fulfilled major diagnostic criteria while three members of family C fulfilled only minor diagnostic criteria (II.8, III.4, and III.8) and one subject was excluded from all analyses because of hypertension (II.9). The results of the linkage analysis are summarised in table 2. It should be noted that in all three families the locus showing the highest lod score value was found to be the disease causing gene. Although family A consisted of only three affected subjects, it was possible to exclude four FHC loci showing lod score values ⩽−2. The TPM1 locus achieved the highest lod score value of 0.6. Subsequently, all exons ofTPM1 were amplified12 and the PCR products investigated by single strand conformation polymorphism heteroduplex analysis on a precast 12.5% polyacrylamide gel at 4°C or 20°C followed by sequencing of abnormal conformers.31These analyses showed a well established mutation in exon 3 resulting in an Asp175Asn amino acid substitution12 32-35 (data not shown). III.2 was non-penetrant at the age of 18 years. In family B where it was only possible to exclude one FHC locus, theTPM1 locus obtained the highest lod score value of 1.6. The subsequent mutation analysis of the gene identified the same TPM1 mutation as in family A. In the large family C, it was possible to exclude eight FHC loci regardless of whether major or minor diagnostic criteria were applied. Using minor diagnostic criteria a significant maximal lod score of 3.6 was obtained at the ACTC locus and subsequent mutation analyses showed a mutation in exon 5 resulting in an Ala295Ser amino acid substitution.3

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Pedigrees of three FHC families investigated. Index patients are indicated by arrows. Open symbols are unaffected subjects, filled symbols are affected subjects fulfilling major FHC diagnostic criteria, shaded symbols are affected subjects fulfilling minor FHC diagnostic criteria. A cross indicates a subject excluded from linkage analyses because of hypertension. Families A and B carry a TPM1 mutation, while pedigree C carries an ACTC mutation. Alleles defined by markers at the TPM1 and ACTC loci in families A/B and C respectively are shown below symbols. Clinical details of families A and B will be reported elsewhere while details of family C have already been reported.3

The heterogeneous clinical manifestations and the genetic complexity of FHC make the achievement of genetic diagnosis difficult and resource demanding. Linkage analysis is a potentially valuable tool in reducing the number of laborious mutation analyses. However, it is necessary to examine and assess the phenotypes of the entire FHC family in question before linkage analysis can be accomplished. It is important to use stringent phenotypic definitions of the disease in order to identify the correct disease gene within the family, although it is often difficult to define the precise phenotype of FHC. Naturally, the diagnostic criteria change and develop in accordance with the current knowledge of the disease. Even minor clinical abnormalities are more likely to represent FHC within a multiply affected family than they would be outside the context of familial disease. Therefore, it is a sensible strategy to use alternative diagnostic criteria with a subgroup of patients fulfilling conventional disease criteria and, if present, another group of patients fulfilling minor diagnostic criteria. This balanced strategy was used in family C where clinical investigations indicated three subjects with minor signs of FHC (II.8, III.4, III.8). Instead of complete exclusion of the three subjects from the analyses, two lod score calculations were performed using major and minor diagnostic criteria, respectively. As seen from table 2, both lod score values appointed ACTC as the most likely disease gene. However, the use of minor diagnostic criteria should be used with caution since there is a greater risk that a subject will be falsely categorised as having the phenotype resulting in no identification of the disease gene. Consequently, minor diagnostic criteria should only be used in context with results obtained by major diagnostic criteria as an additional help in prioritising the order of mutation analyses and certainly not for exclusion of candidate disease genes. The main value of the linkage approach in small families is the ability to prioritise the order of mutation analyses and exclude FHC loci as disease responsible genes, since it is often difficult to achieve a significant lod score because of a limited number of affected subjects. However, once the identification of all FHC genes has been accomplished, significant linkage might be obtained even in small pedigrees as the relative distribution of lod scores among the different FHC loci would determine the significance. For instance, a lod score of 1 would be considered significant if the second most likely candidate locus showed a maximal lod score of –1. Thus, in this future scenario, the linkage approach would play an even greater role in genetic diagnosis of FHC. In the two small pedigrees, A and B, the TPM1 locus showed the highest lod score value of all FHC loci tested and subsequent mutation analysis identified a previously publishedTPM1 mutation. The fact that the pathogenic nature of the TPM1 mutation identified has been well described,12 32-35 including expression studies in transgenic mouse models,36 led us to the conclusion that further mutation analyses of the remaining non-excluded FHC loci were needless. Naturally, the pathogenic impact of any amino acid variation identified needs to be based on substantial evidence since a genetic diagnosis may have profound consequences for disease carriers. It is rarely sufficient to rely only on hypotheses regarding the gene product's functional domains, phylogenetically conserved regions, and three dimensional structure. Similarly, absence of the sequence variation in control chromosomes indicates a possible pathogenic impact, but it is not proof since the sequence variation may still be a rare polymorphism. Linkage analysis may indicate the most likely disease locus, based on quantifiable evidence, and thereby substantiate that a sequence variation identified in this gene is the disease causing agent. Nevertheless, the ultimate proof of disease responsibility is difficult to achieve without studies aiming at expressing the sequence variation in biological test systems. Thus, it is difficult to assess the pathogenic nature of novel missense mutations identified especially in small pedigrees showing insignificant lod score values. Consequently, genetic counselling in small pedigrees associated with novel missense mutations should be cautious and an offer of clinical follow up of all family members without considering their genotype should be considered.

In conclusion, it is apparent that clinical and genetic investigations in FHC are indissolubly connected. The clinical investigations meet the demand of the FHC families in having their disease status clarified and provide the phenotypic assessment necessary for the performance of the genetic investigations. Genetic diagnosis is desirable in order to identify as many of the disease carriers as possible and hopefully diminish the number of sudden deaths associated with the disease by improvement of risk stratification and management. The results of the present study provide a firm basis for linkage analysis in FHC pedigrees and thereby a feasible approach, based on quantifiable evidence, for prioritising the order of mutation analyses and for detection of the disease causing agent among nine disease associated loci.