A novel pathogenic MLH1 missense mutation, c.112A > C, p.Asn38His, in six families with Lynch syndrome

About 3% of all colorectal cancers is due to Lynch syndrome, an autosomal dominant condition caused by germline mutations in one of the DNA mismatch repair (MMR) genes, MLH1, MSH2, MSH6 and PMS2 [1]. Carriers of a mutation in one of these MMR genes have a high risk of developing colorectal cancer, endometrial cancer and also an increased risk of specific other malignancies including ovarian, upper urinary tract, gastric, small intestinal and biliary tract cancer and adenoma or carcinoma of the sebaceous gland [1].

In families who meet the Amsterdam and/or the revised Bethesda criteria [2], tumor examination is indicated including an assay for microsatellite instability (MSI) and immunohistochemistry (IHC) for mismatch repair (MMR) protein expression [2]. When a tumor shows MSI, with or without alterations in immunohistochemical staining of these proteins, mutation analysis of the MLH1, MSH2, MSH6 and PMS2 genes is offered.

Missense mutations comprise about 20% of all pathogenic mutations associated with Lynch syndrome [3, 4]. For most missense mutations, convincing evidence for pathogenicity is lacking, and these are called unclassified variants (UVs) or variants of uncertain clinical significance [3, 5, 6]. To gain more insight in the nature of such an UV, it is useful to study clinical, morphological and molecular features of affected patients and their families.

In this study, we describe six Dutch families with Lynch syndrome and a previously described UV in the MLH1 gene [7‐10]. We have combined data from literature with several parameters studied in these families like co-segregation of the MSI/IHC results and the UV with disease, germline mutation analysis of MMR genes, haplotype analysis, geneaology, and germline mutation testing of healthy controls to gain more insight into the clinical significance of this UV.

Criteria for pathogenicity of missense mutations include difference in the chemical properties of the amino acid (Grantham score), evolutionary conservation of the amino acid, absence in the normal population, co-segregation with the disease and association between MSI and/or absence of immunohistochemical staining for the MMR protein in the tumor [4]. From literature and available databases, it is known that this mutation lies in a functional domain of the MLH1 gene and the Asn-38 codon has a full evolutionary conservation in the MLH1 protein [3, 8, 15]. Furthermore, another missense mutation in this codon, p.Asn38Asp, c.112A > G, is proven to be pathogenic and affects the activity of the MLH1 protein: the protein cannot bind Mg++, does not have ATPase activity and cannot correct for a MLH1 deficiency [15]. Lack of mismatch repair activity has also been described recently for this p.Asn38Asp mutation [10]. The variant described in our study induces a greater shift in Grantham score than the variant described by Takahashi [15]. Also, according to SIFT and MAPP-MMR scores this UV should be considered pathogenic [9, 10].

The results of our study show a complete co-segregation between the cancer cases (colorectal, endometrial and other Lynch syndrome related cancers) and the mutation in six families. All tumors show an MSI high phenotype, mostly without alterations in immunohistochemical protein expression. Pathogenic mutations are known to induce MSI and absence of immunohistochemical staining of one or more MMR proteins. In general, immunohistochemical staining is considered to be more sensitive in detecting MMR gene mutations than MSI analysis [16‐18]. However, missense mutations can have subtle effects on the protein products of the MMR genes, which can result in loss of function while maintaining intact protein that is able to bind antibodies [19]. Therefore, immunochemistry can show normal staining of the proteins, as seen in the majority of the tumors in this study. In accordance to our findings, other studies have also described tumors showing an MSI high phenotype but normal expression of the MMR proteins [7, 10]. We also observed loss of immunostaining of PMS2 in some tumors. Loss of PMS2 protein expression is often seen in tumors from MLH1 mutation carriers, since PMS2 and MLH1 form heterodimeric complexes in mismatch repair [19]. Surprisingly, we also observed loss of MSH2 and/or MSH6 protein expression. An artefact in the methods used in this study is unlikely since staining in normal tissue of the same patients was normal and others also have reported loss of MSH2 and MSH6 protein expression in a tumor of a MLH1 missense mutation carrier [19] and control tissue showed normal staining. Immunostaining can be dependent on a second hit which can differ for each tumor [20].

Taken together, in the case of missense mutations, which might lead to presence of abnormal protein products that still have binding capacity for antibodies, MSI might be more sensitive for detecting patients with MMR gene mutations than immunohistochemical staining [21, 22].

All tested family members share a common allele and their ancestors originate from the same region in The Netherlands. Genealogical studies showed that three families have a common ancestor. The variant in our families is reported earlier by Van Puijenbroek [7], who studied genome-wide copy neutral loss of heterozygosity (LOH) in a cohort of Dutch familial and sporadic colorectal carcinomas. To our knowledge, this mutation has not been reported in families outside The Netherlands [3, 10, 23, 24]. Together with the genealogy and haplotype analysis, this makes the mutation likely to be a Dutch founder mutation.

In the described families, a Lynch syndrome phenotype is clearly present. In addition to colorectal and endometrial cancer, other cancers were also present in mutation carriers: skin cancers (sebaceous gland carcinoma and squamous cell carcinoma), sarcoma, malignant histiocytoma, ureteral cell cancer, astrocytoma and bone tumors not otherwise specified. Most of these cancers fit within the tumor spectrum of Lynch syndrome. However, sarcoma and histiocytoma are not common tumor types seen in Lynch syndrome. There are some studies that suggest these tumors can be part of the tumor spectrum and a causal relationship with MSH2 has been shown [25‐27]. In our study, we confirmed that the patient with the sarcoma is a carrier of the MLH1 mutation. We were not able to confirm that the patient affected with histiocytoma has been a carrier of the MLH1 gene mutation. In Family 1 (VUmc C198), the family history is consistent with Muir-Torre syndrome [14]. Sebaceous gland carcinomas are part of the Lynch syndrome tumor spectrum [28, 29]. Our patients affected with sebaceous gland carcinoma and sebaceous gland adenoma were proven to be carriers of the MLH1 gene mutation and tumor specimens showed a MSI-high phenotype, which confirms the diagnosis Lynch/Muir-Torre syndrome.