Integrating genetics and transcriptomics to decipher susceptibility genes for risk stratification of gastric cancer and effect modification of Helicobacter pylori treatment.

Abstract

In a transcriptome-wide association study, we deciphered susceptibility genes that may predict gastric cancer (GC) risk and modify the effects of Helicobacter pylori (H. pylori) treatment. Genetically predicted expression models of 4518 genes were developed based on the GTEx and applied to a nested case-control study (935 GCs and 1869 controls) of the Mass Intervention Trial in Linqu, Shandong Province (MITS), with genes associated with GC risk further validated in BioBank Japan (7921 GCs and 159,201 controls). Transcriptome risk scores (TRSs) integrating key genes were constructed, utilizing imputed transcriptomes from the Shandong Intervention Trial (SIT) and UK Biobank, and observed transcriptomes from the National Upper Gastrointestinal Cancer Early Detection (UGCED) program. We also examined whether TRS may modify the association of H. pylori infection and anti-H. pylori treatment with GC risk. Integrating 11 independent GC-associated genes identified based on the MITS (FDR-q < 0.05) and BioBank Japan (P < 0.05), the TRS demonstrated a dose-dependent association with an elevated risk of incident GC in both the SIT (P-trend = 0.003) and UK Biobank (P-trend = 0.008), and exhibited an upward trend as gastric lesions progressed based on the UGCED program (P-trend = 5.01 × 10^-4). In the SIT, the increased risk of GC associated with H. pylori infection (P-interaction = 0.03) and beneficial effect of successful H. pylori eradication (P-interaction = 0.05) were significant for individuals with high TRSs. We identified a gene panel which may predict GC risk across populations of multiple ancestries, which offers important insights into GC risk stratification and presents a precision approach to primary prevention. Funders are listed in the Acknowledgement.