Novel candidates in early-onset familial colorectal cancer

In 20-30% of patients suspected of a familial colorectal cancer (CRC) syndrome, no underlying genetic cause is detected. Recent advances in whole exome sequencing have generated evidence for new CRC-susceptibility genes including POLE, POLD1 and NTHL1, but many patients remain unexplained. Whole exome sequencing was performed on DNA from nine patients from five different families with familial clusters of CRC in which traditional genetic testing failed to yield a diagnosis. Variants were filtered by minor allele frequencies, followed by prioritization based on in silico prediction tools, and the presence in cancer susceptibility genes or genes in cancer-associated pathways. Effects of frameshift variants on protein structure were modeled using I-Tasser. One known pathogenic variant in POLD1 was detected (p.S478N), together with variants in 17 candidate genes not previously associated with CRC. Additional in silico analysis using SIFT, PROVEAN and PolyPhen on the 14 missense variants indicated a possible damaging effect in nine of 14 variants. Modeling of the insertions/deletions showed a damaging effect of two variants in NOTCH2 and CYP1B1. One family was explained by a mutation in a known familial CRC gene. In the remaining four families, the most promising candidates found are a frameshift NOTCH2 and a missense RAB25 variant. This study provides potential novel candidate variants in unexplained familial CRC patients, however, functional validation is imperative to confirm the role of these variants in CRC tumorigenesis. Additionally, while whole exome sequencing enables detection of variants throughout the exome, other causes explaining the familial phenotype such as multiple single nucleotide polymorphisms accumulating to a polygenic risk or epigenetic events, might be missed with this approach.