Comparison between germline and somatic loss-of-function <i>RNF43</i> mutations reveals different genotype-phenotype associations and provides insights into the genetic mechanisms of colorectal tumourigenesis.

Abstract

Germline <i>RNF43</i> mutations cause a dominantly inherited syndrome of colorectal cancer (CRC) and serrated polyps. However, these data originate from highly selected families. We assessed germline <i>RNF43</i> variants in patients more representative of the general population and compared these with somatic <i>RNF43</i>mutations in CRCs. We studied 49 823 CRC and/or polyp cases from the CORGI study, 100 000 Genomes (100kGP) and UK Biobank (UKB), alongside 165 250 controls. Somatic mutations were analysed in 2722 CRCs. Consistent with the literature, a germline loss-of-function <i>RNF43</i> variant (p.Thr158ProfsTer6) was found in a multigenerational CORGI family with early-onset CRC and serrated and/or filiform polyps. However, while 23 CRC/polyp cases and 47 controls from 100kGP or UKB had germline <i>RNF43</i> mutations, cases often lacked multiple polyps or a notable family history. Sometimes, CRCs developed independently of the germline <i>RNF43</i> mutation. In case-control analyses, germline <i>RNF43</i> variants were associated with CRC risk (OR=2.696, p=0.010), but penetrance was much greater for germline mutations in the N-terminal half of the gene. Germline C-terminal mutations conferred no increased CRC risk. However, somatic C-terminal mutations were pathogenic, perhaps because their relatively weak effects are supplemented by accompanying mutations in Wnt genes, including <i>ZNRF3</i> and a new driver, <i>SFRP4</i>. <i>RNF43</i> is a CRC predisposition gene, but risks are moderate, the reported polyposis phenotype is often absent and molecular phenocopies can occur. N-terminal germline <i>RNF43</i> variants confer higher risk, although weak effects of C-terminal variants cannot be excluded. Genetic testing and patient management should incorporate these factors.