Somatic aberrations of mismatch repair genes as a cause of microsatellite‐unstable cancers

Abstract

Lynch syndrome (LS) is caused by germline mutations in mismatch repair (MMR) genes, resulting in microsatellite-unstable tumours. Approximately 35% of suspected LS (sLS) patients test negative for germline MMR gene mutations, hampering conclusive LS diagnosis. The aim of this study was to investigate somatic MMR gene aberrations in microsatellite-unstable colorectal and endometrial cancers of sLS patients negative for germline MMR gene mutations. Suspected LS cases were selected from a retrospective Clinical Genetics Department diagnostic cohort and from a prospective multicentre population-based study on LS in The Netherlands. In total, microsatellite-unstable tumours of 40 sLS patients (male/female 20/20, median age 57 years) were screened for somatic MMR gene mutations by next-generation sequencing. In addition, loss of heterozygosity (LOH) of the affected MMR genes in these tumours as well as in 68 LS-associated tumours and 27 microsatellite-unstable tumours with MLH1 promoter hypermethylation was studied. Of the sLS cases, 5/40 (13%) tumours had two pathogenic somatic mutations and 16/40 (40%) tumours had a (likely) pathogenic mutation and LOH. Overall, LOH of the affected MMR gene locus was observed in 24/39 (62%) tumours with informative LOH markers. Of the LS cases and the tumours with MLH1 promoter hypermethylation, 39/61 (64%) and 2/21 (10%) tumours, respectively, demonstrated LOH. Half of microsatellite-unstable tumours of sLS patients without germline MMR gene mutations had two (likely) deleterious somatic MMR gene aberrations, indicating their sporadic origin. Therefore, we advocate adding somatic mutation and LOH analysis of the MMR genes to the molecular diagnostic workflow of LS.

The current value of determining the mismatch repair status of colorectal cancer: A rationale for routine testing

“…There is a significant group of patients (≤35%) whose tumours have an abnormal MMR status by IHC or display MSI on molecular assays, and who have a Lynch syndrome-consistent family history, but in whom genetic testing cannot find an inherited mutation in an MMR gene, and in whom sporadic MLH1 promoter hypermethylation is not demonstrable (Geurts-Giele et al, 2014). "Lynch-like syndrome" (LLS) is a recently proposed term for such cases, and it is thought that it may account for as many as 70% of tumours in which Lynch syndrome is clinically suspected, but genetic testing fails to identify a constitutional MMR gene mutation (Rodriguez-Soler et al, 2013).…”

“…Patients with LLS present with cancer at younger ages, similar to Lynch syndrome (53.7 years vs. 48.5 years of age), promoting the hypothesis that undiagnosed constitutional mutations (aside from constitutional mutations in MMR genes) may also be responsible for tumor phenotypes that closely resemble Lynch syndrome (Geurts-Giele et al, 2014;Haraldsdottir et al, 2014;RodriguezSoler et al, 2013). However, it is probable that some LLS patients could actually have Lynch syndrome, as some tumours may be due to constitutional mutations in DNA MMR genes that are not detectable by standard testing.…”

“…Indeed extended analysis by NGS has demonstrated that in some instances this is the case (Rodriguez-Soler et al, 2013). An alternative cause is somatic mutations in an allele of a DNA MMR gene coupled with loss of heterozygosity (LOH) of the other allele (Geurts-Giele et al, 2014;Mensenkamp et al, 2014). However, up to 50% of LLS CRC exhibits biallelic somatic inactivation of DNA MMR genes within the tumor (Geurts-Giele et al, 2014;Haraldsdottir et al, 2014;Kang et al, 2015;Rodriguez-Soler et al, 2013;Sourrouille et al, 2013).…”

“…An alternative cause is somatic mutations in an allele of a DNA MMR gene coupled with loss of heterozygosity (LOH) of the other allele (Geurts-Giele et al, 2014;Mensenkamp et al, 2014). However, up to 50% of LLS CRC exhibits biallelic somatic inactivation of DNA MMR genes within the tumor (Geurts-Giele et al, 2014;Haraldsdottir et al, 2014;Kang et al, 2015;Rodriguez-Soler et al, 2013;Sourrouille et al, 2013). It has now been demonstrated that many of these are due to constitutional mutations and epimutations in other genes, including POLD1 and POLE (Haraldsdottir et al, 2014;Palles et al, 2013), MUTYH (Castillejo et al, 2014;Syngal et al, 2015), FAN1 (Segui et al, 2015), NTHL1 (Weren et al, 2015), PIK3CA (Cohen et al, 2016), BRCA1/2, 2 PTEN, STK11 (Yurgelun et al, 2015) and other cancer susceptibility genes (SMAD4, BMPR1A, KLLN, AKT1, AXIN2, BUB1 and BUB3) (Hansen et al, 2017) with some resultant tumours sometimes acquiring two somatic mutations in the same MMR gene, and consequently dMMR and/or MSI (Frayling and Arends, 2015;Short et al, 2015).…”

“…For example NGS approaches has helped identify suspected LLS patients that actually have Lynch syndrome, as some tumours may be due to constitutional mutations in DNA MMR genes that are not detectable by standard testing (Rodriguez-Soler et al, 2013). NGS testing has also identified somatic mutations coupled with LOH (Geurts-Giele et al, 2014;Haraldsdottir et al, 2014;Kang et al, 2015;Rodriguez-Soler et al, 2013;Sourrouille et al, 2013) and constitutional mutations in additional genes as causes of dMMR (Adam et al, 2016;Castillejo et al, 2014;Cohen et al, 2016;Haraldsdottir et al, 2014;Palles et al, 2013;Segui et al, 2015;Syngal et al, 2015;Weren et al, 2015).…”