Identifying candidate Hirschsprung disease-associated RET variants

Abstract

Patients with sporadic Hirschsprung disease (HSCR) show increased allele sharing at markers in the 5' region of the RET locus, indicating the presence of a common ancestral RET mutation. In a previous study, we found a haplotype of six SNPs that was transmitted to 55.6% of our patients, whereas it was present in only 16.2% of the controls we used. Among the patients with that haplotype, 90.8% had it on both chromosomes, which led to a much higher risk of developing HSCR than when the haplotype occurred heterozygously. To more precisely define the HSCR-associated region and to identify candidate disease-associated variant(s), we sequenced the shared common haplotype region from 10 kb upstream of the RET gene through intron 1 and exon 2 (in total, 33 kb) in a patient homozygous for the common risk haplotype and in a control individual homozygous for the most common nonrisk haplotype. A comparison of these sequences revealed 86 sequence differences. Of these 86 variations, 8 proved to be in regions highly conserved among different vertebrates and within putative transcription factor binding sites. We therefore considered these as candidate disease-associated variants. Subsequent genotyping of these eight variants revealed a strong disease association for six of the eight markers. These six markers also showed the largest distortions in allele transmission. Interspecies comparison showed that only one of the six variations was located in a region also conserved in a nonmammalian species, making it the most likely candidate HSCR-associated variant.

Figure 1

MULAN plots displaying the percentages of similarity of the 5′ region of the RET locus in the chicken, rat, mouse, and chimpanzee genomic sequences, compared with the human sequence as a reference. The horizontal gray line indicates 70% identity. The top panel of the figure depicts the positions of the 31 SNPs that are located within putative transcription factor binding sites. Those indicated in green fall in one of the evolutionarily conserved regions. The upper scale refers to RET ATG (0 kb). The lower scale represents the length of the whole region. TFBSs = transcription factor binding sites.

Figure 2

Haplotypes consisting of 14 SNPs—6 from our previous risk haplotype and 8 newly selected ones—of four homozygous individuals. Individuals homozygous for one of the haplotypes identified in our previous study (the corresponding SNPs are indicated in plain typeface) were typed for eight new SNPs (indicated in bold typeface). They were also homozygous for all eight new SNPs. Risk alleles are in dark gray boxes; nonrisk alleles are in light gray boxes. freq T = frequency of transmitted haplotype; freq NT = frequency of nontransmitted haplotype.

Figure 3

Results of allelic and genotypic association tests and TDT for the 5′ RET locus SNPs. Results presented regard markers characterized in the previous study (Burzynski et al. 2004) as well as six new loci. For the single-locus allelic and genotypic association analyses, the frequencies of the alleles and the genotypes, respectively, were compared between patients and pseudocontrols by use of a χ² test. Transmission distortion of each allele versus all other alleles together was tested and combined in a multiallelic TDT (Spielman and Ewens 1996). All P values and CIs were corrected for multiple testing for independent tests by use of a Bonferroni correction.