Rare deleterious germline mutations of candidate genes associated with risk of familial esophageal squamous cell carcinoma (ESCC) by targeted sequencing

Abstract

BACKGROUND AND AIMS: Esophageal cancer (EC) is a deadly disease with poor survival rates of only ∼10%. The EC incidence remains highest in the Shanxi and Henan regions near the Tang Hang Mountain area in Northern China. In endemic esophageal squamous cell carcinoma (ESCC) regions, common familial clustering and historical migration patterns suggest an inherited ESCC genetic susceptibility. We hypothesize that cancer predisposition gene(s) (CPGs) carrying rare deleterious variants play a crucial role in ESCC genetic susceptibility and aim to identify the underlying CPGs of familial ESCC from Henan by applying the targeted sequencing approach. METHODS: Our discovery set utilized NGS technology to capture a comprehensive gene panel critical for ESCC by targeted re-sequencing of 1271 selected genes (11 Mb) and sequenced for the germline variants in the blood DNA from 222 familial ESCC individuals and 124 Henan non-cancer controls. Familial ESCC was defined as at least two family members developed ESCC. The custom designed panel contains a cancer panel of 578 gene list (4Mb) and a custom list of 693 (7Mb) ESCC-specific genes derived from various approaches including microarray profiling in FH+ ESCC primary tissues, homozygosity-by-descent (HBD), SNP array-based analysis by both GWAS and candidate gene approaches of our own set of MassArray SNP association data. The data from 346 individuals were processed using the analysis pipeline following the GATK guideline. Briefly, raw fastq reads were mapped using the Burrows-Wheeler Aligner (BWA). PCR duplicates are marked using Picard. The raw variants are called by GATK HaplotypeCaller using a joint genotyping of all samples. RESULTS: Preliminary data analysis for the 346 samples resulted in the final call set comprised of 52328 rare variants. There are 26486 rare variants present in the 222 FH+ ESCC cases, but absent in the 124 controls. The average sequencing coverage was 88X. Principal component analysis (PCA) and multidimensional scaling (MDS) indicated a homogenous population structure between the case and the control populations. Targeted re-sequencing results of the 1271 selected genes in 222 familial ESCC individuals identified a recurrent missense variant in AIM1L (Absent in Melanoma 1-like) in 14/222 (6.3%) of FH+ ESCC patients but absence in the 124 non-cancer control. Data obtained by a melting curve-based genotyping platform by LC480 and LightSNiP assay for high throughput validation of this recurrent SNP in larger sample cohort are underway and will be presented. CONCLUSIONS: A SNP in AIM1L may be associated with higher risk for ESCC in Henan and validation in larger cohort is required. The data may provide valuable insight for understanding genetic susceptibility to this deadly cancer.