- Top of page
- Materials and Methods
Ataxia-telangiectasia (A-T) is an early onset autosomal recessive ataxia associated with characteristic chromosomal aberrations, cell cycle checkpoint defects, cancer susceptibility, and sensitivity to ionizing radiation. We utilized the protein truncation test (PTT), and single strand conformation polymorphism (SSCP) on cDNA, as well as denaturing high performance liquid chromatography (dHPLC) on genomic DNA (gDNA) to screen for mutations in 24 Polish A-T families. Twenty-six distinct Short Tandem Repeat (STR) haplotypes were identified. Three founder mutations accounted for 58% of the alleles. Three-quarters of the families had at least one recurring (shared) mutation, which was somewhat surprising given the low frequency of consanguinity in Poland. STR haplotyping greatly improved the efficiency of mutation detection. We identified 44 of the expected 48 mutations (92%): sixty-nine percent were nonsense mutations, 23% caused aberrant splicing, and 5% were missense mutations. Four mutations have not been previously described. Two of the Polish mutations have been observed previously in Amish and Mennonite A-T patients; this is compatible with historical records. Shared mutations shared the same Single Nucleotide Polymorphism (SNP) and STR haplotypes, indicating common ancestries. The Mennonite mutation, 5932 G>T, is common in Russian A-T families, and the STR haplovariants are the same in both Poland and Russia. Attempts to correlate phenotypes with genotypes were inconclusive due to the limited numbers of patients with identical mutations.
- Top of page
- Materials and Methods
Ataxia-telangiectasia (A-T; MIM # 208900) is an autosomal recessive, neurological disorder with a frequency of 1/40 000–1/100 000 (Gatti, 2002). Cerebellar ataxia, immunodeficiency, oculocutaneous telangiectasia, and radiation sensitivity are characteristic findings in A-T patients. These patients also have a greatly increased risk of cancer (Gatti & Good, 1971; Swift et al. 1986). They typically manifest premature aging, degeneration of the cerebellum, thymus and gonads, growth retardation, and telomere shortening (Gatti, 2002; Chun & Gatti 2004). Carrier frequencies of ATM mutations have been estimated as 1–1.8% and are proving significant with regard to breast cancer susceptibility (Swift et al. 1987; Easton, 1994; Gatti et al. 1999; Concannon, 2002; Buchholz et al. 2004).
- Top of page
- Materials and Methods
Due to the large size of the ATM gene and the broad spectrum of ATM mutations, mutation detection is not yet cost-effective for establishing a diagnosis of A-T. In this study, the diagnosis was confirmed by a lack of ATM protein western blotting and radiosensitivity on by CSA, in all patients. Serum AFP levels were elevated in all patients. Thus, these aspects of the A-T phenotype were not influenced by genotype in any apparent way.
We observed that 58% of the A-T families in Poland shared one of three founder mutations (Haplotype [A], [B], and [D]), and 83% of the families carried at least one of eight Polish founder haplotypes. We were surprised to find this degree of genetic homogeneity, considering that Polish population migrations have not been restricted by geographical features such as large bodies of water or high mountain ranges. Nonetheless, our previous studies of ATM haplotypes and mutations strongly suggest that shared, recurring mutations predate modern ethnicities and nationalities (Campbell et al. 2003), and STR haplotypes such as [A], [B], and [D] may reflect influences on ancient migrations rather than on modern ones. Eleven Polish ATM mutations have also been found in other ethnic groups (Table 2).
|Mutation||Also found in|
|1563_1564delAG||Amish, Turkish, Italian, German, Brazilian|
|5932G>T||Norwegian, Danish, Mennonite, American-Hispanic, German, Russian|
|6095G>A||Swedish, German, French|
|IVS53-2A>C||Danish, American-Hispanic, Brazilian, Portuguese|
Splicing mutations comprised 23% of the mutations found in this study, a proportion not unlike those in previous studies (Teraoka et al. 1999; Mitui et al. 2003). Splicing mutations typically involve the highly conserved canonical 3′ or 5′ splice sites, as is the case for IVS 53-2A>C on Polish haplotype [A]. Three other splicing mutations were noted on non-recurring Polish haplotypes. IVS20–597delAAGT is a ‘masked’ mutation that causes Type II splicing with pseudoexon formation (Eng et al. 2004). The mutation occurs deep within intron 20, and disrupts the U6 portion of a U1 snRNA binding site (Pagani et al. 2002). It has also been observed in German, Turkish and Hispanic-American patients (Mitui et al. 2003; Eng et al. 2004); the standardized STR haplotypes of the Hispanic-American families differ slightly from those of the Old World (Polish, German, Turkish) A-T families (Eng et al. 2004), providing further evidence that many ATM mutations predate STR haplotypes, but not the common SNP haplotypes (Thorstenson et al. 2001; Campbell et al. 2003)
Most ATM mutations are associated with specific STR and SNP haplotypes (Campbell et al. 2003; Mitui et al. 2003; Eng et al. 2004). This held true without exception for the SNP haplotypes associated with Polish mutations. In general, this was also true for the association of these mutations with STR haplotypes, with two exceptions: in WAR33 [A][D] and WAR 19 [B][E]. WAR33 carries the Haplotype [A] mutation, IVS53-2A>C; however, the S2179 allele appears to have changed from ‘141’ to ‘139’. The H3 SNP haplotype background remains the same as that observed for all [A] haplovariants in this study. Haplovariants were also observed for the 5932 G>T mutation on Haplotype [D] (see below) and for the 6095G>A mutation on Haplotype [B]. Both long and short forms of this Haplotype [B] were observed (Fig 1A), with only a single allele (S2179 ‘137’) shared by all four chromosomes (WAR 6, 19, 22, and 49–3). Taken together, these data suggest that the longest variant (eg: WAR 49–3) is the older, ancestral haplotype for this mutation, although alternative interpretations are possible.
The mutation on Haplotype [F], 1563_1564delAG, is perhaps the most commonly observed ATM mutation worldwide and always occurs on a SNP H2 background. It was observed in three Polish families in association with SNP haplotype (H2), but with several STR haplovariants. As previously described (Campbell et al. 2003), 1563_1564delAG is associated with STR haplovariant 1 (in Turkish, Polish and Amish A-T patients), haplovariant 2 (in a Brazilian patient), and haplovariant 3 (in Turkish and Italian patients). In all of these families, the allele for S1818 was ‘160’, as is also observed in two of the Polish families; however, in WAR 46, a new haplovariant 4 was defined by allele S1818 ‘158’ (instead of ‘160’). These findings are compatible with the historical origins of the Amish of Pennsylvania (U.S.A) from Germanic settlers, descendants of an Anabaptist movement in northern Europe (1525–1536) (Hostetler, 1983a).
The mutation 5932G>T, found on Polish Haplotype [D], has also been observed in Mennonites, another Germanic Christian sect of Anabaptist roots that settled in Kentucky and Pennsylvania. In the early 1500s, Mennonites from the Netherlands and North Germany migrated to the Vistula Detta (now Poland) and later continued their migrations to Canada and the United States (1873–74; 1922–30) (Hostetler, 1983b). This mutation has also been found in A-T patients from Denmark, England and Guatemala, always in association with a SNP (H2) haplotype as previously described (Campbell et al. 2003). It has recently been reported to be the most common founder mutation (44%) in a group of Russian A-T families, and the STR haplotypes are the same (Birrell et al. 2005). Alleles at S1819 vary as either ‘131’ or ‘135’, thereby defining two haplovariants that are found in both Russia and Poland.
We attempted to correlate clinical data with specific mutations. Such genotype/phenotype correlations are difficult to achieve unless several important criteria are met: (1) the classical diagnosis must be confirmed on a molecular basis to distinguish variant A-T phenotypes from other phenotypically similar diseases; (2) a sufficient number of patients must be homozygous for a mutation so that the genotypic effects can be isolated from other ATM alleles. Unless one postulates that a heterozygous mutation will have a dominant interfering effect (in which case one parent should manifest symptoms), comparing compound heterozygous patients who share only a single mutation is not likely to reveal significant genotype/phenotype correlations. Despite this, a dominant interfering effect for at least one ATM mutation (2546delSRI) has been demonstrated in the mouse both in vivo and ex vivo, with some suggestion that parent carriers of this mutation may manifest an increased incidence of cancer (Concannon, 2002; Scott et al. 2002; Spring et al. 2002). The Polish data set contained only two homozygous patients (WAR 12 and WAR 31) and neither of these haplotypes was observed in other families. Only Haplotypes [A], [D], and [B] were observed repeatedly in 7, 5, and 4 patients, respectively, and no significant genotype/phenotype correlations were noted. The entire clinical dataset is included in Table 1 so that it might later contribute to a meta-analysis of a larger cohort of A-T patients. This Table also suggests parameters for planning such analyses.
Mutation detection for the Polish A-T population was performed with the hope that it would be of assistance in counselling Polish families for family planning, prenatal diagnosis, and identification of heterozygote carriers. This information may also help in diagnosing A-T patients at a younger age by SNP and STR haplotype prescreening for the eight recurring Polish founder haplotypes. Certain types of mutations (asterisked in the ‘consequence’ column of Figure 1B) may be amenable to therapeutic intervention with aminoglycosides or other compounds (Lai et al. 2004). Lastly, understanding the spectrum of ATM mutations in Polish patients with A-T allows these mutations to be sought in breast cancer and other diseases.