Heterozygous germline MSH3 mutations, and probably MLH3 mutations, act as classical tumour suppressors, leading to excess somatic deletion mutations, signature ID4 and increased colorectal cancer risk
Soriano, I.; Sherwood, K.; Ward, J.; Fernandez-Tajes, J.; Amamullah, A.; Thorn, S.; Gul, G.; Wilson, J.; Isaksson, A.; Glimelius, B.; Sjoblom, T.; Palles, C.; Tomlinson, I.
Show abstract
Background. MSH3 and MLH3 are non-canonical DNA mismatch repair genes, involved in repairing insertion-deletion mutations. Colorectal cancer (CRC) and adenomas have been reported in patients with bi-allelic germline MSH3 mutations, and in a very few bi-allelic MLH3 mutation carriers. Objectives. We hypothesised that germline loss-of-function MSH3 and MLH3 mutations were akin to constitutional mismatch repair deficiency (cMMRd) and Lynch syndrome, such that CRC could result from either bi-allelic germline mutations, or heterozygous germline mutations after second hits. Design. Nearly 12,000 CRC and multiple polyp cases and 460,000 controls were studied. 2,023 patients underwent cancer genome sequencing. Results. One CRC/multiple polyp case had bi-allelic MSH3 mutations and another, bi-allelic MLH3 mutations. MSH3 and MLH3 germline heterozygotes had an increased risk of CRC (respectively 2.2-fold, P=6.6x10-5 and 1.6-fold, P=0.028), owing to somatic second hits that inactivated the wildtype allele. Single second hits sometimes inactivated both MSH3 and the nearby APC gene. All CRCs with MSH3 or MLH3 deficiency were microsatellite-stable, but hypermutant. Deletions of 2 or more base pairs were particularly increased (about 12-fold) and signature ID4 was usually present (P<0.0001). CRCs from heterozygotes without second hits showed no hypermutation. Conclusion. The phenotypes of bi-allelic MSH3 and MLH3 mutation carriers resemble some cMMRd patients. Heterozygous germline MSH3 and MLH3 alleles have incomplete penetrance, but increase CRC risk via hypermutation, phenotypically resembling PMS2-mutant Lynch syndrome. A causal association with specific mutations has not previously been reported for ID4 in human tumours. ID4 probably does not have a single aetiology, but can result from MSH3 or MLH3 deficiency. The findings of this manuscript are provisional and should not be used to guide clinical practice, health-related behavior, or policy.
Matching journals
The top 5 journals account for 50% of the predicted probability mass.
Similar papers in this journal
- Estimating cancer risk in carriers of Lynch syndrome variants in UK Biobank 97%
- Heritable genetic variants in key cancer genes link cancer risk with anthropometric traits 93%
- Functional annotation with expression validation identifies novel metastasis-relevant genes from post-GWAS risk loci in sporadic colorectal carcinomas 92%
Similar papers in this journal
- Profiling diverse sequence tandem repeats in colorectal cancer reveals co-occurrence of microsatellite and chromosomal instability involving Chromosome 8 94%
- Evolving neoantigen profiles in colorectal cancers with DNA repair defects 94%
- Somatic mutational profiles and germline polygenic risk scores in human cancer 94%
Similar papers in this journal
- On Cross-ancestry Cancer Polygenic Risk Scores 92%
- Missense variants causing Wiedemann-Steiner syndrome preferentially occur in the KMT2A-CXXC domain and are accurately classified using AlphaFold2 92%
- Increased ultra-rare variant load in an isolated Scottish population impacts exonic and regulatory regions 92%
Similar papers in this journal
- APC Mutation marks an aggressive subtype of BRAF mutant colorectal cancers 96%
- DNA mismatch repair gene variant classification: evaluating the utility of somatic mutations and mismatch repair deficient colonic crypts and endometrial glands 95%
- Deep sequencing of early T stage colorectal cancers reveals disruption of homologous recombination repair in microsatellite stable tumours with high mutational burdens 93%
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.