Abstract B036: Systematic Germline and Tumor Genomic Analysis for Gene Discovery in Familial Pancreatic Cancer

Abstract

Despite the identification of several pancreatic ductal adenocarcinoma (PDAC) predisposition genes, more than 75% of familial pancreatic cancer (FPC) families remain genetically unexplained. To address this, we developed a systematic discovery pipeline focused on families with multiple affected individuals, leveraging familial segregation and matched tumor analysis to identify novel hereditary PDAC susceptibility genes. We analyzed germline DNA sequencing samples from 1,628 individuals with PDAC, including 98 patients from 45 FPC families, defined in our study as at least two affected relatives on the same side of the family. Of the 1,628 patients, 462 individuals (65 related cases and 397 probands) were previously reported (Roberts et al., Cancer Discovery). Sequencing comprised 714 germline exome/genome datasets processed through a harmonized variant-calling pipeline (hg19) and 914 matched germline-tumor whole genomes analyzed using patient-level germline and somatic variant calling (hg38). FPC families with multiple affected individuals or available tumor sequencing comprised two discovery sets (Discovery-A1: 81 patients from 38 families; Discovery-A2: 17 patients from 7 families), while the remaining PDAC cohort was used for further candidate gene prioritization. Rare variants were defined as having an allele frequency <0.01 in ExAC v0.3 (hg19) or gnomAD v4.1.1 (hg38). Analyses were restricted to predicted loss-of-function variants, including frameshift indels, stop-gain/loss, and canonical splice-site variants. Discovery-A1 identified rare variants segregating among affected relatives in at least two FPC families. Discovery-A2 evaluated segregating variants in FPC families with matched tumor sequencing showing loss of heterozygosity (LOH) of the gene of interest. Candidate genes were subsequently evaluated in the independent PDAC cohort. Discovery-A1 identified 12 segregating variants in 9 genes among 24 affected individuals from 12 FPC families. Six variants each recurred in two unrelated families, while six additional genes harbored distinct segregating variants in two independent families. Discovery-A2 identified segregating variants in 8 genes among 7 families, with tumor LOH observed in 6 of the 8 genes. Together, these analyses prioritized 17 candidate genes: ALDOB, CCDC105, CRYL1, DNAH9, EME2, ITIH2, MAP3K6, PKHD1L1, RYR1, SETD4, SFXN2, TGFBR2, TNFRSF10B, WDR87, XPOT, ZNF528, and ZNF599. Variant recurrence and tumor LOH in the independent PDAC cohort further prioritized several candidate genes. Combining familial segregation with matched tumor analysis efficiently prioritized candidate hereditary PDAC genes from a large genomic resource. Although several promising candidates have emerged, none currently provides sufficient evidence to establish a causal role in FPC susceptibility. These findings highlight the complexity of hereditary PDAC and support continued genetic and functional evaluation of the prioritized candidate genes.