Noncoding RET variants explain the strong association with Hirschsprung disease in patients without rare coding sequence variant.

The pathogenesis of Hirschsprung disease is complex. Although the RET proto-oncogene is the most frequently affected gene in Hirschsprung disease, rare coding sequence variants explain only a small part of Hirschsprung disease cases. We aimed to assess the genetic background of Hirschsprung disease using a genome-wide association analysis combined with sequencing all RET exons in samples from 105 Hirschsprung disease cases (30 familial and 75 sporadic) and 386 controls. As expected, variants in or near RET showed the strongest overall association with Hirschsprung disease and the most statistically significant association was observed when using a recessive genetic model (rs2435357, NC_000010.10:g.43582056T > C; genotype TT, OR = 17.31, P = 1.462 × 10-21). Previously published associations in variants in SEMA (rs11766001, NC_000007.13:g.84145202A > C; allele C, OR = 2.268, P = 0.009533) and NRG1 (rs4541858, NC_000008.10:g.32410309A > G; allele G, OR = 1.567, P = 0.015; rs7835688, NC_000008.10:g.32411499G > C; allele C, OR = 1.567, P = 0.015) were also replicated in the genome-wide association analysis. Sequencing revealed a total of 12 exonic RET rare variants. Of these, eight amino acid changing rare variants and two frameshift variants caused or possibly caused Hirschsprung disease. Only a minority of the Hirschsprung disease cases (9/30 familial; 7/75 sporadic) carried one of the rare variants. Excluding the rare variant carriers from the genome-wide association analysis did not appreciably change the association of rs2435357 with Hirschsprung disease. We estimate that approximately two thirds of the sporadic cases may be statistically attributed to the recessive action of the common non-coding RET variants. Thus, even though most cases do not carry rare RET variants, combinations of rare variants and the common non-coding RET variant cause the majority of the cases in our population.

Genome-wide association study of Hirschsprung disease detects a novel low-frequency variant at the RET locus.

Hirschsprung disease (HSCR) is a congenital disorder with a population incidence of ~1/5000 live births, defined by an absence of enteric ganglia along variable lengths of the colon. HSCR genome-wide association studies (GWAS) have found common associated variants at RET, SEMA3, and NRG1, but they still fail to explain all of its heritability. To enhance gene discovery, we performed a GWAS of 170 cases identified from the Danish nationwide pathology registry with 4717 controls, based on 6.2 million variants imputed from the haplotype reference consortium panel. We found a novel low-frequency variant (rs144432435), which, when conditioning on the lead RET single-nucleotide polymorphism (SNP), was of genome-wide significance in the discovery analysis. This conditional association signal was replicated in a Swedish HSCR cohort with discovery plus replication meta-analysis conditional odds ratio of 6.6 (P = 7.7 × 10-10; 322 cases and 4893 controls). The conditional signal was, however, not replicated in two HSCR cohorts from USA and Finland, leading to the hypothesis that rs144432435 tags a rare haplotype present in Denmark and Sweden. Using the genome-wide complex trait analysis method, we estimated the SNP heritability of HSCR to be 88%, close to estimates based on classical family studies. Moreover, by using Lasso (least absolute shrinkage and selection operator) regression we were able to construct a genetic HSCR predictor with a area under the receiver operator characteristics curve of 76% in an independent validation set. In conclusion, we combined the largest collection of sporadic Hirschsprung cases to date (586 cases) to further elucidate HSCR's genetic architecture.

Thyroid cancer and co-occurring RET mutations in Hirschsprung disease.

The objective of this study was to assess the occurrence of thyroid cancer and co-occurring RET mutations in a population-based cohort of adult Hirschsprung disease (HD) patients. All 156 patients operated for HD in a tertiary center during 1950-1986 were followed for thyroid malignancies up to 2010 through the nationwide Finnish Cancer Registry. Ninety-one individuals participated in clinical and genetic screening, which included serum calcitonin and thyroid ultrasound (US) with cytology. Exons 10, 11, 13, and 16 were sequenced in all, and all exons of RET in 43 of the subjects, including those with thyroid cancer, RET mutations, suspicious clinical findings, and familial or long-segment disease. Through the cancer registry, two cases (aged 35 and 37 years) of medullary thyroid cancer (MTC) were observed; the incidence for MTC was 340-fold (95% CI 52-1600) compared with average population. These individuals had C611R and C620R mutations in exon 10. One papillary thyroid cancer without RET mutations was detected by clinical screening. Four subjects (aged 31-50 years) with co-occurring RET mutations in exons 10 (C609R; n=1) and 13 (Y791F, n=3) had sporadic short-segment HD with normal thyroid US and serum calcitonin. Three novel mutations and five single-nucleotide polymorphisms were found outside exons 10 and 13 without associated signs of thyroid cancer. MTC-associated RET mutations were restricted to exons 10 and 13 affecting ∼5% of unselected adults with HD. Clinical thyroid assessment did not improve accuracy of genetic screening, which should not be limited to patients with familial or long-segment disease.