Abstract and Introduction
Abstract
Background: Lynch syndrome, an autosomal-dominant disorder characterised by high colorectal and endometrial cancer risks, is caused by inherited mutations in DNA mismatch repair (MMR) genes. Mutations fully abrogating gene function are unambiguously disease causing. However, missense mutations often have unknown functional implications, hampering genetic counselling. We have applied a novel approach to study three MSH2 unclassified variants (UVs) found in Dutch families with suspected Lynch syndrome.
Methods: The three mutations were recreated in the endogenous Msh2 gene in mouse embryonic stem cells by oligonucleotide-directed gene modification. The effect of the UVs on MMR activity was then tested using a set of functional assays interrogating the main MMR functions.
Results: We recreated and functionally tested three MSH2 UVs: MSH2-Y165D (c.493T>G), MSH2-Q690E (c.2068C>G) and MSH2-M813V (c.2437A>G). We observed reduced levels of MSH2-Y165D and MSH2-Q690E but not MSH2-M813V proteins. MSH2-M813V was able to support all MMR functions similar to wild-type MSH2, whereas MSH2-Y165D and MSH2-Q690E showed partial defects.
Conclusions: Based on the results from our functional assays, we conclude that the MSH2-M813V variant is not disease causing. The MSH2-Y165D and MSH2-Q690E variants affect MMR function and are therefore likely the underlying cause of familial cancer predisposition. Since the MMR defect is partial, these variants may represent low penetrance alleles.
Introduction
Lynch syndrome (LS) is an autosomal-dominant disorder caused by inherited defects in the mismatch repair (MMR) system. Individuals with LS have an up to 10 times elevated risk of colorectal (CRC) and endometrial cancer compared with the general population. They furthermore have an increased risk of developing tumours of the small bowel, stomach, ovaries, pancreas, ureter and renal pelvis, bladder, biliary tract, brain and sebaceous glands.
The main function of MMR is recognition and repair of replication errors. MSH2 protein can form a heterodimeric complex with MSH6 (hMutSα) or MSH3 (hMutSβ), which can initiate the MMR process by binding to erroneously copied DNA. hMutSα binds single base–base mismatches or single unpaired bases, whereas hMutSβ preferentially binds loops of multiple unpaired bases. Mismatch-bound hMutS dimer recruits hMutLα, a dimer of MLH1 and PMS2, which directs further repair.
Most inherited MMR mutations in LS are found in MLH1 and MSH2 (70–80% of cases); mutations in MSH6 and PMS2 account for the remaining 20–30%. When LS is suspected, tumour material is analysed for microsatellite instability (MSI) and/or expression of MMR proteins by immunohistochemistry (IHC). If a lack of expression is found, the corresponding gene(s) are analysed for mutations. When IHC shows no aberrancies or cannot be done, but the family history is strongly indicative for LS, mutation analysis of all four MMR genes is warranted. The majority of mutations found in LS families are abrogating mutations, such as frameshifts, truncations and deletions, which are unambiguously disease causing. However, also missense mutations are found, whose potential for pathogenicity is much more difficult to predict. These so-called unclassified variants (UVs) pose a serious problem to the clinic. As long as it is unclear whether a mutation affects the MMR system, the clinical suspicion of LS cannot be confirmed and no presymptomatic testing is available to identify at-risk family members.
The assess the pathogenicity of UVs, integrative approaches have been proposed to calculate likelihood ratios based on clinical data (segregation analysis, IHC, MSI, loss of heterozygosity) and in silico analysis considering evolutionary conservation and physicochemical differences between amino acids. In addition, functional assays have been developed based on in vitro reconstituted MMR reactions or cellular complementation of MMR deficiency by overexpression of variant MMR proteins, often in distantly related species such as yeast and Escherichia coli. While such approaches can accurately identify fully deleterious mutations, partial MMR defects may remain unnoticed.
We have therefore developed an alternative method to study the functional effects of missense mutations in MSH2 and MSH6 by recreating these variants at the endogenous genes in mouse embryonic stem cells (ESC) using oligonucleotide-directed gene modification. This allows us to study whether and to which extent the variant protein supports the most relevant MMR functions in a highly homologous system (murine MSH2 and MSH6 proteins share 94 and 96% homology with their human counterparts, respectively).
A defect in the MMR system is manifested by length alterations of mononucleotide or dinucleotide repeat sequences (MSI). Additional functions of the MMR system include the prevention of recombination between homologous but not identical DNA sequences and the cellular response to methylating agents such as 6-thioguanine (6-TG) and N-methyl-N'-nitro-N-nitrosoguanidine (MNNG) that add methyl groups to bases in the DNA. MMR prohibits the incorporation of nucleotides opposite methylated guanines during DNA replication, which ultimately leads to double-strand breaks. As a result, MMR-proficient cells are sensitive to methylating agents, while MMR-deficient cells are tolerant.
Addressing these functions in ESCs expressing the variant allele provides an accurate and quantitative readout for MMR activity. Here we applied this method to study the functional consequences of three MSH2 missense mutations that have been found in Dutch suspected LS families and found evidence for partial MMR defects.