Barrett esophagus (BE) occurs in 1% to 10% of the general population and is believed to be the precursor of esophageal adenocarcinoma (EAC). The incidence of EAC has increased 350% in the last 3 decades without clear etiology. Finding predisposition genes may improve premorbid risk assessment, genetic counseling, and management. Genome-wide multiplatform approaches may lead to the identification of genes important in BE/EAC development.
To identify risk alleles or mutated genes associated with BE/EAC.
Design, Setting, and Patients
Model-free linkage analyses of 21 concordant-affected sibling pairs with BE/EAC and 11 discordant sibling pairs (2005–2006). Significant germline genomic regions in independent prospectively accrued series of 176 white patients with BE/EAC and 200 ancestry-matched controls (2007–2010) were validated and fine mapped. Integrating data from these significant genomic regions with somatic gene expression data from 19 BE/EAC tissues yielded 12 “priority” candidate genes for mutation analysis (2010). Genes that showed mutations in cases but not in controls were further screened in an independent prospectively accrued validation series of 58 cases (2010).
Main Outcome Measures
Identification of germline mutations in genes associated with BE/EAC cases. Functional interrogation of the most commonly mutated gene.
Three major genes, MSR1, ASCC1, and CTHRC1 were associated with BE/EAC (all P<.001). In addition, 13 patients (11.2%) with BE/EAC carried germline mutations in MSR1, ASCC1, or CTHRC1. MSR1 was the most frequently mutated, with 8 of 116 (proportion, 0.069; 95% confidence interval [CI], 0.030–0.130; P<.001) cases with c.877C>T (p.R293X). An independent validation series confirmed germline MSR1 mutations in 2 of 58 cases (proportion, 0.035; 95% CI, 0.004–0.120; P=.09). MSR1 mutation resulted in CCND1 up-regulation in peripheral-protein lysate. Immunohistochemistry of BE tissues in MSR1-mutation carriers showed increased nuclear expression of CCND1.
MSR1 was significantly associated with the presence of BE/EAC in derivation and validation samples, although it was only present in a small percentage of the cases.