Refined Expression Quantitative Trait Loci Analysis On Adenocarcinoma At The Gastroesophageal Junction Reveals Susceptibility Genes And Prognostic Markers

Background and objective:Many single nucleotide polymorphisms(SNPs)have been linked to complex diseases including diabetes,Alzheimer disease and multiple cancers.Genome-wide association studies(GWASs)reveal hundreds of risk loci,most of which are located in non-coding regions of DNA,suggesting deciphering their biological functions remains challenging.Expression quantitative trait locus(eQTL)is a commonly used method to explain potential functions of risk loci identified by GWAS.This approach is based on the concept that a GWAS variant,in some tissues,affects expression at a nearby gene and that both the gene and the tissue might play a role in the disease mechanism.However,previous studies mostly investigated eQTLs at the tissue level,leading to ever growing sample sizes but partially successful in prioritizing disease risk eSNPs and eGenes.Recent studies have shifted focus on cell-type specific eQTL because many diseaseassociated eQTLs have been observed in specific cell types and these eQTLs show strong enrichment for heritability across complex traits.Moreover,despite non-coding RNAs especially long non-coding RNA(lncRNAs)have been regarded as tissue-specific prognosis markers and therapeutic targets of many kind of tumors,eQTL analyses focusing on non-coding eGenes still remain exiguous.Adenocarcinomas at gastroesophageal junction(ACGEJ)have had rapidly increasing incidences in western countries during the past few decades.In Asia where squamous cell carcinoma is the predominant type of esophageal cancer,rising incidence of ACGEJ have also been reported.The 5-year survival rates of this cancer are 20-25%,lower than that of esophageal or gastric cancers.Previous tissue-level studies have identified susceptibility loci and therapeutic genes of ACGEJ.However,most eQTL studies of tumors lack enough resolution and paired adjacent normal samples.Therefore deciphering tumor-specific eQTLs may help to elucidate carcinogenic mechanisms and inform broadly applicable risk assessment efforts.Methods:In the present study,we conduct an eQTL study on ACGEJ using whole genome sequencing(WGS),RNA-seq and public single cell RNA-seq(scRNA-seq)data.To increase the resolution of eQTLs on cell type level,we use proportion of epithelial cells multiplying genotypes as an interaction term.Epithelial cell proportion is generated by computational deconvolution using reference gene expression profiles produced by scRNA-seq data.To increase the resolution of eQTLs on disease level,we estimate effects of epithelial cell-specific eQTLs on ACGEJ and normal cells.Results:We show substantial ACGEJ-specific effects in the genetic control of gene expression,revealing 1,009 ACGEJ-loss and 1,027 ACGEJ-gain expression quantitative pairs(ePairs).By integrating GWAS and clinical data,we finally identify 5 potential susceptibility loci including rs4947311-PSORS1C1,rs13134812-LOC391674,rs2240191-RASAL1,rs4236599-FOXP2 and rs17508585-CDK13-DT and 1 prognosis marker rs309483-LINC01355.Futhermore,we reveal that LINC01355 promotes progression of ACGEJ by down-regulating expression of cell cycle checkpoint genes including p27 and p53.We also proposed a hypothesis that initiation and progression of ACGEJ need irregularly expressed transcription factors as the driving force and germline alterations as the precondition,which were validated by experimental validation of top ACGEJ-specific regulatory axis rs658524-KLF5-CTSW.These findings further our understanding on the susceptibility mechanisms of ACGEJ and illustrate the importance of investigating eQTLs in non-coding regions.Conclusion:We reveal 2,036 substantial ACGEJ-specific effects(ACGEJ-gain/-loss eQTLs)in the genetic regulation of gene expression.Moreover,we integrated cell typespecific eQTLs with GWAS and clinical data to identify susceptibility and prognosis markers of ACGEJ and elucidated that ACGEJ-specific eQTLs regulated expression of target genes by influencing binding capacity of transcription factors.The present study provides insights into tumor-specific eQTLs and identifies risk loci more precisely and specifically than tissue-level studies.These novel findings improved our knowledge to pathogenic mechanism of ACGEJ and highlighted crucial roles of risk loci at non-coding regions to tumor studies from the perspective of eQTL study.