Article Text

A common ancestor for COCH related cochleovestibular (DFNA9) patients in Belgium and The Netherlands bearing the P51S mutation

Statistics from

Editor—Hearing impairment is extremely heterogeneous, both phenotypically and genetically. It is the most frequent form of sensory impairment in the western world, affecting approximately 1/1000 newborns and approximately half of the people above the age of 80.1 ,2 In all these cases hereditary factors are a prominent cause. So far, more than 60 loci for monogenic non-syndromal hearing impairment have been described and 14 responsible genes have been identified (Van Camp and Smith, Hereditary Hearing Loss Homepage

Balance problems are also are relatively frequent, but considerably less is known about the causes. Purely genetic forms of vestibular impairment are extremely rare and no genes have been identified yet. However, it is commonly known that many hearing impaired people also suffer from balance problems. Moreover, it is now recognised that many syndromes with genetic hearing impairment also show a dysfunction of the vestibular system.3 The prevalence of vestibular dysfunction may be severely underestimated, as it often remains unnoticed until specialised vestibular tests are performed. Owing to the intimate relationship between the auditory and the vestibular systems, there are probably many genes with a function in both systems.

DFNA9 is the only form of hereditary non-syndromal hearing impairment where strongly marked vestibular involvement has been described. This locus has been mapped to chromosome 14q12-q13.4Progressive sensorineural hearing impairment is present, usually starting between the ages of 35 and 50 in the high frequencies.5-7 This is paralleled by a gradually deteriorating sense of balance, leading to instability (most notably in the dark). Some, but not all patients periodically suffer from vertigo attacks associated with nausea, tinnitus, or aural fullness, which is reminiscent of the symptoms of Menière's disease.5Eventually, after a disease course of approximately 20 years, patients become severely to profoundly deaf across all frequencies and lose their vestibular function.

The identification of the gene responsible for DFNA9 was greatly helped by the availability of a cochlear specific cDNA library.8One of the transcripts in this cDNA library was a novel cochlear gene designatedCOCH-5B2, which was later renamed COCH. COCH was mapped to chromosome 14q12-13, making it a strong candidate gene for DFNA9.9 SubsequentCOCH mutation analysis in three DFNA9 families showed a missense mutation in each of them.10 The predicted COCH protein has a length of 550 amino acids and consists of a signal peptide, a cysteine rich domain with homology to the factor C domain of the horseshoe crabLimulus (FCH domain), and two domains with homology to the von Willebrand factor A domain (vWFA1 domain and vWFA2 domain). All three mutations are located in the FCH domain.

Following the identification of the COCHgene, COCH mutation analysis was performed in several other DFNA9 families. In six Dutch and one Belgian family, a C→T point mutation at base pair 151 was found.5 ,11 This mutation leads to the substitution of a conserved proline for a serine at position 51 of the COCH protein (P51S). Just like the three previously described mutations, this is a missense mutation residing in the FCH domain, and the substituted amino acid is conserved between the human, mouse, and chickenCOCH homologues. In this study, eight additional families (or at least one of their members) were shown to carry the P51S mutation in the COCH gene. Fine mapping of the markers from the COCHregion and haplotype analysis of markers flanking theCOCH gene strongly suggested a single common founder for at least nine of these families.

Details of the families are shown in table 1. The clinical characteristics of family 1,5 family 2,7 ,12 ,13 family 3,14 and families 4-76 ,11 have been described previously. Symptoms similar to those described were observed in families 8, 9, 11, and 12, as well as in the isolated patients 10, 13, 14, and 15. In family 8, a genome search was performed by the Mammalian Genotyping Center (Marshfield, USA), which mapped the disease locus to chromosome 14q12-13 (data not shown).

Table 1

COCH families haplotyped in this study

As the P51S mutation creates an extra site for the restriction enzymeDdeI, the mutation was analysed by PCR amplification followed by restriction enzyme digestion and gel electrophoresis, as described previously.5 In normal controls, the 295 bp PCR product is digested into two bands of 247 and 48 bp, respectively. In heterozygous patients carrying the P51S mutation, two additional bands of 143 and 104 bp are visible on agarose gel electrophoresis.

The Whitehead database ( was screened with markers D14S262 and D14S1071 flanking theCOCH region at 14q12-13, which yielded YAC contig WC14. YACs 746-F10, 732-A9, 855-C7, 814-E9, 857-D12, 905-B1, 925-C2, 952-D9, and 949-A9 were obtained from the UK Human Genome Mapping Project (HGMP, Harwell, UK). The relative marker order of seven polymorphic Généthon markers, as well as the position of theCOCH gene, were determined by STS content mapping through PCR analysis.

Microsatellite markers were analysed using a standard radioactive PCR assay. Primer sequences and allele frequencies for all markers used in this study were retrieved from the Généthon database.15

The P51S mutation in the COCH gene has been found previously in seven families characterised by a combination of progressive sensorineural hearing impairment and vestibular dysfunction.5 ,11 We analysed the presence of this mutation in a total of 29 additional families or isolated patients with similar symptoms. All families originate from Belgium or the southern part of The Netherlands. In eight of them, the P51S mutation was present, bringing the total number of P51S bearing families to 15. In the remaining 21 families without the P51S mutation, no further mutation analysis has been performed, so it is possible that they have a different, as yet unidentified COCHmutation. General data about the families included in this study are summarised in table 1. Although all P51S bearing families originate from the same region, we found no genealogical evidence for a common ancestor.

To investigate a possible founder effect, we reconstructed the disease haplotypes for polymorphic markers flanking theCOCH locus. This was complicated by the fact that on the Généthon map the order of these flanking markers was not completely resolved (not shown). In addition, the exact position of the COCH gene with respect to these markers was unknown. Therefore, we constructed a YAC contig spanning theCOCH region and determined the marker order through PCR analysis on this contig. As a result of this analysis, the marker order and the position of COCH could be unambiguously determined (fig 1). This marker order is different from the one on the Généthon map, which locates D14S1034 proximal to D14S1021, D14S1040, and D14S1071, and D14S975 distal to D14S257.15

Figure 1

YAC contig spanning the COCH region. YACs were screened for the different markers by PCR. A positive signal is denoted by a black dot, and a stretch of positive signals is connected by a dashed line. The length of the YACs used for the mapping of the markers gave no clear answer to the physical distance between the markers (not shown). Genetic distance between D14S262 and D14S975 is 3.3 cM on the Généthon map,15 whereas the distance between the other markers is unclear. In the Généthon map, markers D14S975, D14S1021, D14S257, D14S1071, D14S1040, and D14S1034 are located within an interval of 0.9 cM. Therefore, the genetic distance between D14S975 and the most distal marker D14S1034 is probably not larger than 1 cM.

The results of the haplotype analysis of the markers flanking theCOCH gene are shown in table 2. Families 1, 4, 7, 8, 9, and 12 show exactly the same haplotype for all seven markers examined: 1-1-2-2-1-6-1 (for allele lengths and frequencies, see table 3).

Table 2

Haplotype sharing around the COCH locus in P51S families

Table 3

Allele lengths of 14q12-13 markers3-150

This common disease haplotype may also be present in family 11 (two brothers) and in patients 10, 13, and 15, but only partly in patient 14. In these cases, the disease haplotypes could not be unequivocally determined. Families 2, 5, and 6 have the same haplotype as families 1, 4, 7, 8, 9, and 12 for the four markers that most closely flank theCOCH gene on both sides. Marker D14S1071 has a different haplotype and therefore the three most distal markers are not shared with the other families. In family 3, only marker D14S975 has the same haplotype as all the other families.

The P51S mutation in the COCH gene, which is responsible for progressive cochleovestibular impairment (DFNA9), had previously been found in seven families. In this study we have found the same mutation in eight novel families and isolated patients. This could indicate that either all 15 P51S bearing families share a common ancestor or that the P51S mutation represents a mutation hotspot that gave rise to several independent mutational events. Analysis of the disease associated haplotype of the polymorphic markers aroundCOCH showed significant haplotype sharing in nine families and six of them (1, 4, 7, 8, 9, 12) shared haplotypes for all seven markers examined. The probability that six independent families would share a seven allele haplotype merely by chance is negligible (table 3). Therefore, families 1, 4, 7, 8, 9, and 12 must be related to and originate from a common ancestor.

Most probably, families 2, 5, and 6 also originate from this common affected ancestor. These latter families have four alleles in common with families 1, 4, 7, 8, 9, and 12, which are the ones that most closely flank the COCH gene (table 2). One ancestral recombination between D14S257 and D14S1071 can explain the different haplotypes of the distal markers D14S1071, D14S1040, and D14S1034. Taken together, a total of nine families have the identical haplotype (1-1-2-2) for markers D14S262, D14S975, D14S1021, and D14S257 and it can be concluded that they all share a common ancestor.

Only family 3 has a haplotype that is distinct from the other families for the markers flanking COCH, with the exception of marker D14S975. However, as allele 1 is a very common allele for D14S975 (66%), sharing of it could be a coincidence. It cannot be excluded that the mutation in family 3 arose independently from the other families, but, on the other hand, this family may be more distantly related to the others and the shared interval around theCOCH gene may be very narrow.

Families 10, 13, 14, and 15 represent isolated patients with the P51S mutation, whereas family 11 consists of two affected brothers. Here, the disease haplotype could not be unambiguously determined. However, with the exception of patient 15, who is homozygous for allele 4 of marker D14S1021, the haplotypes observed in these isolated patients indicate that it cannot be excluded that they, too, are a carrier of the 1-1-2-2-1-6-1 haplotype. This question can only be resolved if more family members become available.

The 10 families in which the exact disease associated haplotype could be determined may allow an estimate of the number of generations passed since the common ancestor. Formulae to calculate this age rely on the probability of having recombinations in an interval with a known size.16 ,17 The accuracy of these calculations, however, are heavily dependent upon θ and thus upon the accuracy of the genetic maps. However, the data obtained here indicate that the Généthon genetic map around the COCHlocus is not reliable. Therefore, the intermarker distances around theCOCH gene are unknown and even a rough estimate of the age of the mutation is not feasible. This question may be resolved in the future by an improvement of the map or by the sequencing of the COCH region as part of the Human Genome Project.

The geographical spreading of the P51S mutation may be another clue to the age of the mutation. The 15 families in which the P51S mutation was found contain more than 200 patients in Belgium and The Netherlands and there must be many more unidentified mutation carriers (table 1). This makes the P51S mutation a frequent cause of late onset hearing impairment combined with vestibular dysfunction, at least in the Dutch and the Belgian populations. A systematic mutation analysis in other populations would indicate whether this mutation is also present elsewhere. A more general occurrence of the P51S mutation would indicate an ancient origin of the mutation or a mutational hotspot. This would imply that COCH is a major gene for cochleovestibular impairment in many different populations harbouring a recurrent mutation that can easily be screened for in a diagnostic setup.


The authors wish to thank all the families who contributed to this study. Gerard J te Meerman is acknowledged for his critical remarks on the manuscript. This project was funded by a research grant from the Flemish Fund for Scientific Research (FWO-Vlaanderen) to GVC and PVdH, a grant from the University of Antwerp to GVC and PVdH, and by a grant from The Netherlands Organisation for Scientific Research to FPMC. The clinical studies in the six Dutch (Nijmegen) DFNA9 families were funded by grants from the Heinsius Houbolt foundation and the Nijmegen ORL-Research Foundation. The genome search in family 8 was performed at the Mammalian Genotyping Service (Marshfield, USA). EF and GVC hold research positions at the FWO-Vlaanderen.


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.