Serrated polyposis syndrome (SPS) is characterised by the occurrence of multiple serrated polyps in the large bowel. Defined clinically by the 2010 revised WHO, SPS is associated with an increased risk of colorectal cancer (CRC) for affected individuals and their first-degree relatives. Yan et al1 recently reported their findings from whole-exome sequencing (WES) of six individuals with SPS from four families, identifying a single family carrying a germline likely pathogenic variant in RNF43 (c.953-1 G>A, c.953_954delAG, p.E318fs). In this family, six carriers were identified, five of whom met the WHO criteria for SPS. This adds to two previous reports of germline mutations within RNF43 in individuals with SPS or who developed multiple serrated polyps.2 ,3 The current study by Yan et al1 also importantly showed loss of second wildtype allele in 16 serrated polyps, five adenomatous polyps, and in the rectal adenocarcinoma from carriers through somatic single nucleotide variants or loss of heterozygosity, adding further weight to a potential role of RNF43 in the development of colonic serrated neoplasia.

We present here our results from two genetic screens of a large cohort of individuals with SPS recruited to the Genetics of Colorectal Polyposis Study (GCPS). The GCPS has recruited 422 families with at least one SPS-affected individual from Family Cancer Clinics within Australia, Toronto Canada, Ohio, USA and from the New Zealand Familial Gastrointestinal Cancer Service. The first screen comprised 74 individuals with SPS selected based on early age at diagnosis, high numbers of serrated polyps throughout the colon and having a first-degree relative with SPS or CRC (table 1) and consisted of WES (n=58; Agilent XT SureSelect_V4 52 Mb capture, 100 bp paired-end sequencing on a HiSeq2500 to a mean depth of 100×) and whole-genome sequencing (WGS; n=16; 150 bp paired-end reads using Illumina Hi-Seq X Ten sequencer to average 30× coverage). For the RNF43 gene, we filtered for variants with a minor allele frequency of <1% in reference databases and that were (1) likely to cause loss of function (LoF) via a non-sense/stop gain, frameshift or splice site or (2) were non-synonymous variants predicted to be damaging by ≥2 out of four in silico programmes (PolyPhen2, SIFT, FATHMM, CADD). No LoF variants were identified; however, two rare non-synonymous variants predicted to be damaging by both PolyPhen2 and CADD were detected in a single carrier each (table 2).

A second targeted genetic screen was performed specifically testing for the RNF43 p.R113*2 and p.R132*3 variants to determine their prevalence in individuals with SPS. We tested blood-derived DNA from 221 SPS cases from the GCPS, a subset of which have been previously described,4 using a high-resolution melt (HRM) assay for each mutation where samples with aberrant melting profiles were Sanger sequenced to confirm genotype sequence (primers and PCR conditions in online supplementary table). None of the tested individuals with SPS were carriers of either of these two RNF43 germline LoF variants. However, we cannot exclude the possibility that other variants within RNF43 could exist within a proportion of the 221 patients with SPS who were tested.

The study by Yan et al1 adds to the reported number of SPS families with germline mutations in RNF43, now totalling four. The paucity of RNF43 germline LoF variants in 295 patients with SPS from the GCPS suggests that mutations in RNF43 may account for only a small proportion of SPS suggesting that additional genetic risk factors for SPS are yet to be identified. In conclusion, our findings do not support the need for routine germline testing of RNF43 in individuals who meet the criteria for SPS.