Leveraging tissue-specific regulatory maps and network-assisted analysis to identify novel genetic risk loci for esophageal adenocarcinoma

点击翻译按钮获取中文摘要

PROJECT SUMMARY / ABSTRACT Esophageal adenocarcinoma (EAC) is a rare yet lethal cancer with median survival <1 year. Genome-wide association studies (GWAS) have estimated a substantial heritable component of risk (25-35%) for EAC and its precursor, Barrett’s esophagus (BE). Nearly 20 novel genetic risk loci have been discovered, but most heritability remains unexplained. ‘Missing heritability’ hinders the power of GWAS to illuminate molecular pathways underlying disease risk and identify novel targets for intervention. In this proposal, we seek to overcome inherent limits on sample sizes for BE/EAC and identify novel susceptibility loci by integrating advanced network-based methods and tissue-specific regulome resources into a biologically-motivated discovery framework. Several lines of evidence implicate transcriptional regulatory networks in BE/EAC biology and motivate use of network-based approaches to probe undiscovered genetic underpinnings of this cancer. These findings include reactivation of key embryonic transcriptional regulators in BE/EAC tissues; somatic genomic alterations in transcription factor (TF) genes in EAC tumors; and genome-wide-significant GWAS signals in close proximity to genes encoding esophageal/foregut TFs. Building on these observations, and the prevailing view that disease-linked genetic variation functionally converges on a limited set of core biological pathways, we hypothesize that genetic signals embedded in developmental transcriptional networks represent an important source of ‘missing heritability’ for BE/EAC. Using customized disease-relevant reference networks overlaid with GWAS-derived node weights, we will screen for gene-level and enhancer/promoter-level genetic associations missed by prior genome-wide scans. Our multi-disciplinary MPI team draws on a strong track record in BE/EAC genetics, leveraging access to the largest available GWAS datasets, and extensive omics data from GTEx, RoadMap/ENCODE, and promoter-capture HiC. In Aim 1, we will identify co-expressed genes enriched in risk-associated genetic variation, using transcriptional regulatory networks derived from RNA-seq profiles. Co-expression networks assembled via mutual information and graphical lasso methods applied to transcriptomes of 330 gastro-esophageal junction tissues will be populated with weights from gene-level GWAS tests, and analyzed using Hierarchical Hotnet (HHN). In Aim 2, we will identify linked promoters and enhancers with concentrated GWAS signal using regulatory maps from 3D chromatin interaction profiles. Enhancer-target reference networks built using promoter-capture-HiC data in normal esophagus will be loaded with weights from custom SNP-set-based tests and evaluated via HHN. Our proposed research will help elucidate the genetic architecture of EAC and its only known precursor (BE). Candidate risk genes and enhancers/promoters will be advanced to functional validation studies currently underway for known loci through an ongoing collaboration, with the goal of defining new preventive/therapeutic targets for BE/EAC.