Identification of rare and penetrant germline mutations for gastric cancer
Final Report Abstract
The aim of the project was to identify genes in which rare and penetrant germline mutations contribute to the pathogenesis of gastric carcinoma (GC). For this, 512 European GC patients with early disease onset (< 65 years) were subjected to whole exome sequencing (WES). After stringent quality control (QC), 468 GC patients and 118,479 cancer-free European controls could be included in the gene-based burden analysis. This revealed a Bonferroni-corrected significant burden of rare loss-of-function (LOF) mutations in 5 genes among patients. The two genes in which LOF mutations were most significantly enriched in GC patients are already known to cause monogenic GC. In CDH1 (E-cadherin), 6 GC patients showed LOF mutations (P = 9.99 x 10^-14, OR = 775) with an average age of onset (AO) of 36.5 years. In ATM (ataxia telangiectasia mutated), 7 GC patients presented LOF mutations (P = 8.53 x 10^-11, OR = 64) with an average AO of 46.43 years. In contrast, we identified 3 genes in which LOF mutations are enriched among GC patients, but which have not been established as GC risk genes so far. In FANCA (Fanconi anaemia complementation group A), we identified LOF mutations in 5 GC patients (P = 7.15 x 10^-10, OR = 183) with an average AO of 43.4 years. As other already established GC risk genes, LOF mutations in FANCA lead to homologous recombination deficiency (HRD) at the cellular level. In addition, LOF mutations in FANCA have previously been found in GC patients. However, the data on this was too sparse to refer to FANCA as an established GC risk gene. In TP53I3 (p53-inducible gene 3), we identified LOF mutations in 9 GC patients (P = 2.34 x 10^-07, OR = 11) with an average AO of 45 years. Germline mutations in TP53I3 have not been identified in any cancer type so far. However, TP53I3 is functionally interesting in the context of cancer development as it is involved in cell apoptosis and DNA repair processes. In ZNF519 (zinc finger protein 519), we identified LOF mutations in 10 GC patients (P = 1.05 x 10^-06, OR = 7.99) with an average AO of 42.7 years. Until now, the function of this gene is poorly understood. Clinically, it is interesting that all 37 patients with LOF mutations in CDH1, ATM, FANCA, TP53I3 and ZNF519 developed GC before the age of 51. Our data thus show that a considerable proportion of GC patients with an AO before the age of 51 are of monogenic pathogenesis. Together with further research data, this could impact future recommendations for molecular genetic diagnostics of GC patients in clinical practice. At present, we are carrying out further analyses to confirm the contribution of the newly identified GC candidate genes. In the latest release of WES data from the UK Biobank, we will analyse whether LOF mutations in FANCA, TP53I3 and ZNF519 are enriched among cancer patients, including GC. In addition, as part of a collaboration, we are currently analysing whether the allele of the respective gene that is not affected by the germline LOF mutation is hit by a somatically acquired mutation in cancer tissue (second hit of a tumour suppressor). This would imply that they have contributed to tumour development in these patients.
