Ancient origin of a Japanese xeroderma pigmentosum founder mutation

Xeroderma pigmentosum(XP (MIM278700)) is a rare autosomal recessive disorder [1–3]. XP patients have sun sensitivity, a 10,000-fold increased risk of skin cancer and defective DNA repair [4]. The frequency of XP in Japan is about 1:22,000 [5;6], which is much more common than in the US and Europe (about 1 per million) [2;6]. There are 8 XP DNA repair genes (XPA to XPG and XP variant). The XPA gene is the predominant XP gene in Japan and is defective in about 55% of Japanese XP patients [3]. These Japanese XP patients have a severe form of XP with progressive neurological degeneration [7]. XPA is located on chromosome 9q22.3 and codes for a 273 amino acid protein that is involved in nucleotide excision repair[8]. More than 90% of the mutant alleles in Japanese XP-A patients have the same G to C base changemutation [7–9]. This founder mutationat the 3′ splice acceptor site of intron 3 (IVS3-1G>C) results in no detectable protein production and markedly reduced DNA repair. Approximately 1% of the Japanese general population are heterozygous carriers of this mutation [5]. In order to estimate the age of the most recent common ancestorof this founder mutationin Japan, we used haplotype analysis. We studied DNA samples from XP-A patients who were homozygous for the founder mutation. Using Sanger sequencing, initially we measured 70 single nucleotide polymorphisms (SNP’s) that were located up to 2.5 MB upstream and 2.5 MB downstream of the mutation on chromosome 9(Table 1) as indicated on the HapMap http://www.hapmap.org. We then selected SNP’s that were highly polymorphic in the Japanese population as indicated by the Japanese HapMap subjects. Table 1 Haplotype analysis of region surrounding Japanese XPA founder mutationon chromosome 9q22.3 To estimate the age of the most recent common ancestorof the XPA Japanese founder mutation, we used a likelihood-based method [10] that uses multilocus marker data. This method uses multilocus marker data to estimate the age of the most recent common ancestorof the mutation from a small number of patients. This method was originally tested through simulations and shown to be well suited to estimating the age of rare mutations. The basic assumption for the method is that the N affected individuals, all carrying the same mutation at disease locus D, descend from a common ancestor who introduced the mutation ngen generations ago. The likelihood was written as a function of the recombination fraction (θ) between D and each marker, ngen, and the mutation rate and allele frequencies at each marker locus. Twenty-five HapMap-based SNPs flanking D were genotyped for this analysis. Japanese HapMap subjectsa llele frequencies for these SNPs were between 50:50and 71.6:28.4. Since SNPs have a very low mutation rate, the mutation rate was fixed to 0 for this analysis. The closest short tandem repeat markers flanking D had a θ/Mb ratio of approximately 1, thus physical distance (Mb) between each marker and D was converted to θ using this ratio. ngen was estimated from the size of the haplotype shared by the N affected individuals on each side of D and 95% confidence intervals were computed. We received de-identified samples from XP-A patients in 39 Japanese families located throughout Japan except for the northern areas of Tohoku and Hokkaido. At NIH, DNA was extracted from 43 XP-A patients in the 39 families. Forty one XP-A patients in 37 of the families were confirmed to be homozygous for the XPA founder mutation(red column) (Table 1). DNA from 2 XP patients was found to be heterozygous for this splice mutation as well as heterozygous for nearby SNP loci and was not included in the analysis (data not shown). There were 3 affected siblings in one family (family 1) and 2 affected siblings in two other families (families 18 and 28) (Table 1 – first column, blue shading). Six of the families (families 32–37) had a history of consanguinity (4 were first cousins, 1 was second cousinsand 1 unspecified) (Table 1 – first column, yellow shading). DNA was tested for 25 SNPs on chromosome 9in the region of the XPA gene (Table 1 and data not shown). In families 1, 10 and 28 we tested DNA from multiple affected siblings. The siblings in each family had the same haplotypes. We found a small region of SNP homozygosity in all the XP patients extending about 150 kb upstream and 50 kb downstream from the mutation (bright green area). The size of this region is a reflection of the relationship among the patients: a large region indicates a close relationship and a small region indicates a more ancient relationship. Using the method described above, we determined the age of the most recent common ancestorof the XPA founder mutationto be 120 generations (95% CI, 71–205 generations). Assuming a 20-year generation interval this corresponds to 2400 years (95% CI, 1420–4100 years) or 3600 years (95% CI, 2130– 6150 years) based on a generation length of 30 years. The Japanese archipelago was completely separated from the Eurasian continent about 12,000 years ago. Thus this mutation occurred after this separation occurred and spread throughout the isolated Japanese population. XP is recessive and the carriers of this XPA founder mutationdo not have overt clinical symptoms. All of the XP-A patients we studied were homozygous for the founder mutationindicating that both parents had the same mutation. However, only 6 of the families were aware of a close relationship between the parents of the affected XP patients. This study indicates that the common ancestor of the other 31 families could have occurred thousands of years ago. Since this mutation is present in about 1 million Japanese carriers [5], genetic counseling for this ancient founder mutationmay be considered in the Japanese population.