The most common genes in which germline mutations may cause this disease are MSH2 on chromosome 2p16, MLH1 on chromosome 3p21 which account for approximately 90 % of cases of Lynch syndrome, MSH6 on 2p15, and much less commonly, PMS2 on chromosome 7p22, which can be a rare cause of childhood brain cancer, when biallelic mutations are present (De Vos et al. 2004). Mutations in MSH3 are very rarely reported in patients with colorectal cancer (Nicoliades et al. 1994; Akiyama et al. 1997; de la Chapelle 2004). Interpretation of reported mutations in PMS2 is complicated by the presence of pseudogenes (Hayward et al. 2004). Up to one-half of individuals diagnosed with colorectal cancer before the age of 35 years may have a germline mutation in one of these genes (Liu et al. 1995), but only a minority of patients with colorectal cancer diagnosed between the ages of 35 and 45 years without a family history of colorectal cancer have such mutations. The overall contribution of Lynch syndrome to colorectal cancer is probably 2–3 % (de la Chapelle 2004), but only extensive, multimodal mutation analysis can find more than 75 % of all mutations. Alu-mediated exonic deletions are particularly common in MSH2.
All these genes are involved in the same pathway for repair of mismatches in DNA, and MSI may be demonstrated as DNA replication errors detectable in Tumours relative to genomic DNA (Parsons et al. 1993; Liu et al. 1995; Shia et al. 2005). Frameshift mutations of the transforming growth factor (TGF) type II receptor gene (TGFBR2) and other growth-/apoptosis-related genes such as BAX are common and are due to the runs of mono- or dinucleotide repeats present in these target genes. Analysis of a series of families with Lynch syndrome suggests that perhaps half of the cases are due to mutations in the MSH2 gene, 30–40 % to mutations in the MLH1 gene, less than 10 % are attributable to mutations in MSH6, and only a handful is due to mutations in other genes. When multiple analysis techniques are used, most families that fulfill the Amsterdam Criteria are found to have germline Lynch syndrome mutations (DiFiore et al. 2004; Wagner et al. 2003). Founder mutations in the genes causing Lynch syndrome are known; for instance, there is an MLH1 deletion of exon 16 in the Finns (Nystrom-Lahti et al. 1995), and a founder mutation in MSH2, known as p.Ala636Pro probably accounts for a third of all Lynch syndrome cases in the Ashkenazi Jewish population (Foulkes et al. 2002). Attenuated forms of FAP may rarely masquerade as Lynch syndrome, usually due to unusual mutations in the APC gene (Spirio et al. 1993). Other genes that may be involved in colorectal cancer predisposition, with single reports of germline mutations reported in cases, include AXIN2 and TGFBR2.
Subsequent data have suggested that the latter mutation is most likely to be a neutral polymorphism (Lu et al. 1998; de la Chapelle 2004).
A small minority of cases of Lynch syndrome can be caused by constitutional MLH1 epimutations, with soma-wide allele-specific promoter methylation and transcription silencing of this gene. These epimutations are reversible, resulting in non-Mendelian inheritance of the phenotype. These were initially identified in cases with early-onset colorectal cancer showing loss of staining for MLH1 and MSI in the Tumours, but with no identified germline mutation in the MLH1 gene. Deletion of the unmethylated MLH1 gene was demonstrated in the Tumours (Hitchins and Ward 2009). In many cases the somatic methylation was mosaic, and the case sporadic, but some familial cases have been described, with apparent variability in the degree of somatic methylation of the MLH1 gene. The risk of a Lynch syndrome phenotype in the first-degree relatives of such cases is less than 50 %, as the methylation defect appears to be unstable, leading to non-Mendelian inheritance patterns, but it is recommended that first-degree relatives of cases should be screened as for Lynch syndrome.
Constitutional epimutations in MSH2 have been detected in some cases of Lynch syndrome with loss of staining for MSH2 and MSH6, and MSI, in Tumours samples, but with no detectable germline mutation in MSH2. In these cases a germline deletional mutation has been detected in the final exons of the upstream gene EPCAM (TACSTD2). This deletion disrupts the transcription termination signal from EPCAM, resulting in read-through into MSH2. This results in methylation of the MSH2 gene particularly in epithelial tissues where EPCAM expression is upregulated, causing a mosaic somatic methylation pattern (Kovacs et al. 2009). The risk of endometrial cancer in EPCAM deletion carriers appears to be significantly lower than in carriers of MSH2 mutations, although the risk of colorectal cancer is comparable (Lightenberg et al. 2013).
Heterozygous cells repair DNA normally in Lynch syndrome (or at levels that generally do not have clinical consequences), but in cells where the second allele has been inactivated, for instance, by deletion, in the colonic epithelium, a mutator phenotype is generated and MSI is seen. Genes with microsatellite sequences in their coding regions (such as TGFBR2) are prone to somatic mutation, which is thought to promote the carcinogenic progression (de la Chapelle 2004). It has been suggested that the DNA MMR defect in Lynch syndrome generates increased numbers of novel and potentially immunogenic mutations, which result in an increased immunologic response to Tumours in affected individuals. MSI can be demonstrated in hyperplastic endometrium in women without Lynch syndrome, but not in normal endometrium, and is also detectable in about 20 % of endometrial cancers (Baldinu et al. 2002). Similar findings have been described in individuals with Lynch syndrome with colorectal cancer and also even before the diagnosis of cancer.