Individuals carrying pathogenic germline variants in CDH1 are at high risk of developing diffuse gastric cancer and lobular breast cancer. Preventative measures such as stomach removal are effective in reducing cancer risk in individuals carrying pathogenic variants. The interpretation of CDH1 variants is a significant challenge and currently the majority of CDH1 variants are classified as variants of uncertain significance (VUS). Most CDH1 VUSs are missense mutations, and the functional consequence of these mutations is not obvious.
Here, we used two high-throughput strategies that enable us to systematically define consequences of coding and non-coding variants in CDH1. Specifically:
(A) We developed a pooled CDH1 overexpression library that contains every possible amino acid change within CDH1 (58,432 variants). Using WNT signalling or cell proliferation as readouts, we identified functional and non-functional variants in CDH1 coding regions.
(B) Using base editors and a cell proliferation assay as a readout, we modified the non-coding regions of CDH1 and identify intronic regions in CDH1 that are important in regulating its activity.
Our screens include every possible variant in the coding and non-coding regions of CDH1 and identified three different categories of variants. First, we identified nonsense variants that lead to complete deletion or truncation. Second, we found missense positions that are sensitive to any substitution or positions that are sensitive to specific amino acid changes. Third, we found variants within the coding and non-coding regions that affect splicing and result in loss of CDH1 function. By integrating this dataset with family registries, we found new deleterious CDH1 variants in families with high prevalence of cancer and gain new insights into the biology and functions of CDH1. Our work describes a comprehensive systematic characterisation of CDH1 variants. Furthermore, we describe a general experimental and analytical platform for functional annotation of variants.