Genome-wide CNV analysis in 221 unrelated patients and targeted high-throughput sequencing reveal novel causative candidate genes for colorectal adenomatous polyposis.

To uncover novel causative genes in patients with unexplained adenomatous polyposis, a model disease for colorectal cancer, we performed a genome-wide analysis of germline copy number variants (CNV) in a large, well characterized APC and MUTYH mutation negative patient cohort followed by a targeted next generation sequencing (NGS) approach. Genomic DNA from 221 unrelated German patients was genotyped on high-resolution SNP arrays. Putative CNVs were filtered according to stringent criteria, compared with those of 531 population-based German controls, and validated by qPCR. Candidate genes were prioritized using in silico, expression, and segregation analyses, data mining and enrichment analyses of genes and pathways. In 27% of the 221 unrelated patients, a total of 77 protein coding genes displayed rare, nonrecurrent, germline CNVs. The set included 26 candidates with molecular and cellular functions related to tumorigenesis. Targeted high-throughput sequencing found truncating point mutations in 12% (10/77) of the prioritized genes. No clear evidence was found for autosomal recessive subtypes. Six patients had potentially causative mutations in more than one of the 26 genes. Combined with data from recent studies of early-onset colorectal and breast cancer, recurrent potential loss-of-function alterations were detected in CNTN6, FOCAD (KIAA1797), HSPH1, KIF26B, MCM3AP, YBEY and in three genes from the ARHGAP family. In the canonical Wnt pathway oncogene CTNNB1 (ß-catenin), two potential gain-of-function mutations were found. In conclusion, the present study identified a group of rarely affected genes which are likely to predispose to colorectal adenoma formation and confirmed previously published candidates for tumor predisposition as etiologically relevant.

Functional testing strategy for coding genetic variants of unclear significance in MLH1 in Lynch syndrome diagnosis.

Lynch syndrome is caused by inactivating mutations in the MLH1 gene, but genetic variants of unclear significance frequently preclude diagnosis. Functional testing can reveal variant-conferred defects in gene or protein function. Based on functional defect frequencies and clinical applicability of test systems, we developed a functional testing strategy aimed at efficiently detecting pathogenic defects in coding MLH1 variants. In this strategy, tests of repair activity and expression are prioritized over analyses of subcellular protein localization and messenger RNA (mRNA) formation. This strategy was used for four unclear coding MLH1 variants (p.Asp41His, p.Leu507Phe, p.Gln689Arg, p.Glu605del + p.Val716Met). Expression was analyzed using a transfection system, mismatch repair (MMR) activity by complementation in vitro, mRNA formation by reverse transcriptase-PCR in carrier lymphocyte mRNA, and subcellular localization with dye-labeled fusion constructs. All tests included clinically meaningful controls. The strategy enabled efficient identification of defects in two unclear variants: the p.Asp41His variant showed loss of MMR activity, whereas the compound variant p.Glu605del + p.Val716Met had a defect of expression. This expression defect was significantly stronger than the pathogenic expression reference variant analyzed in parallel, therefore the defect of the compound variant is also pathogenic. Interestingly, the expression defect was caused additively by both of the compound variants, at least one of which is non-pathogenic when occurring by itself. Tests were neutral for p.Leu507Phe and p.Gln689Arg, and the results were consistent with available clinical data. We finally discuss the improved sensitivity and efficiency of the applied strategy and its limitations in analyzing unclear coding MLH1 variants.

Frequency and phenotypic spectrum of germline mutations in POLE and seven other polymerase genes in 266 patients with colorectal adenomas and carcinomas.

In a number of families with colorectal adenomatous polyposis or suspected Lynch syndrome/HNPCC, no germline alteration in the APC, MUTYH, or mismatch repair (MMR) genes are found. Missense mutations in the polymerase genes POLE and POLD1 have recently been identified as rare cause of multiple colorectal adenomas and carcinomas, a condition termed polymerase proofreading-associated polyposis (PPAP). The aim of the present study was to evaluate the clinical relevance and phenotypic spectrum of polymerase germline mutations. Therefore, targeted sequencing of the polymerase genes POLD1, POLD2, POLD3, POLD4, POLE, POLE2, POLE3 and POLE4 was performed in 266 unrelated patients with polyposis or fulfilled Amsterdam criteria. The POLE mutation c.1270C>G;p.Leu424Val was detected in four unrelated patients. The mutation was present in 1.5% (4/266) of all patients, 4% (3/77) of all familial cases and 7% (2/30) of familial polyposis cases. The colorectal phenotype in 14 affected individuals ranged from typical adenomatous polyposis to a HNPCC phenotype, with high intrafamilial variability. Multiple colorectal carcinomas and duodenal adenomas were common, and one case of duodenal carcinoma was reported. Additionally, various extraintestinal lesions were evident. Nine further putative pathogenic variants were identified. The most promising was c.1306C>T;p.Pro436Ser in POLE. In conclusion, a PPAP was identified in a substantial number of polyposis and familial colorectal cancer patients. Screening for polymerase proofreading mutations should therefore be considered, particularly in unexplained familial cases. The present study broadens the phenotypic spectrum of PPAP to duodenal adenomas and carcinomas, and identified novel, potentially pathogenic variants in four polymerase genes.

Germline variants in the SEMA4A gene predispose to familial colorectal cancer type X.

Familial colorectal cancer type X (FCCTX) is characterized by clinical features of hereditary non-polyposis colorectal cancer with a yet undefined genetic background. Here we identify the SEMA4A p.Val78Met germline mutation in an Austrian kindred with FCCTX, using an integrative genomics strategy. Compared with wild-type protein, SEMA4A(V78M) demonstrates significantly increased MAPK/Erk and PI3K/Akt signalling as well as cell cycle progression of SEMA4A-deficient HCT-116 colorectal cancer cells. In a cohort of 53 patients with FCCTX, we depict two further SEMA4A mutations, p.Gly484Ala and p.Ser326Phe and the single-nucleotide polymorphism (SNP) p.Pro682Ser. This SNP is highly associated with the FCCTX phenotype exhibiting increased risk for colorectal cancer (OR 6.79, 95% CI 2.63 to 17.52). Our study shows previously unidentified germline variants in SEMA4A predisposing to FCCTX, which has implications for surveillance strategies of patients and their families.

Evaluating the performance of clinical criteria for predicting mismatch repair gene mutations in Lynch syndrome: a comprehensive analysis of 3,671 families.

Carriers of mismatch repair (MMR) gene mutations have a high lifetime risk for colorectal and endometrial cancers, as well as other malignancies. As mutation analysis to detect these patients is expensive and time-consuming, clinical criteria and tumor-tissue analysis are widely used as pre-screening methods. The aim of our study was to evaluate the performance of commonly applied clinical criteria (the Amsterdam I and II Criteria, and the original and revised Bethesda Guidelines) and the results of tumor-tissue analysis in predicting MMR gene mutations. We analyzed 3,671 families from the German HNPCC Registry and divided them into nine mutually exclusive groups with different clinical criteria. A total of 680 families (18.5%) were found to have a pathogenic MMR gene mutation. Among all 1,284 families with microsatellite instability-high (MSI-H) colorectal cancer, the overall mutation detection rate was 53.0%. Mutation frequencies and their distribution between the four MMR genes differed significantly between clinical groups (p < 0.001). The highest frequencies were found in families fulfilling the Amsterdam Criteria (46.4%). Families with loss of MSH2 expression had higher mutation detection rates (69.5%) than families with loss of MLH1 expression (43.1%). MMR mutations were found significantly more often in families with at least one MSI-H small-bowel cancer (p < 0.001). No MMR mutations were found among patients under 40-years-old with only colorectal adenoma. Familial clustering of Lynch syndrome-related tumors, early age of onset, and familial occurrence of small-bowel cancer were clinically relevant predictors for Lynch syndrome.

Reduced migration of MLH1 deficient colon cancer cells depends on SPTAN1.

INTRODUCTION: Defects in the DNA mismatch repair (MMR) protein MLH1 are frequently observed in sporadic and hereditary colorectal cancers (CRC). Affected tumors generate much less metastatic potential than the MLH1 proficient forms. Although MLH1 has been shown to be not only involved in postreplicative MMR but also in several MMR independent processes like cytoskeletal organization, the connection between MLH1 and metastasis remains unclear. We recently identified non-erythroid spectrin αII (SPTAN1), a scaffolding protein involved in cell adhesion and motility, to interact with MLH1. In the current study, the interaction of MLH1 and SPTAN1 and its potential consequences for CRC metastasis was evaluated. METHODS: Nine cancer cell lines as well as fresh and paraffin embedded colon cancer tissue from 12 patients were used in gene expression studies of SPTAN1 and MLH1. Co-expression of SPTAN1 and MLH1 was analyzed by siRNA knock down of MLH1 in HeLa, HEK293, MLH1 positive HCT116, SW480 and LoVo cells. Effects on cellular motility were determined in MLH1 deficient HCT116 and MLH1 deficient HEK293T compared to their MLH1 proficient sister cells, respectively. RESULTS: MLH1 deficiency is clearly associated with SPTAN1 reduction. Moreover, siRNA knock down of MLH1 decreased the mRNA level of SPTAN1 in HeLa, HEK293 as well as in MLH1 positive HCT116 cells, which indicates a co-expression of SPTAN1 by MLH1. In addition, cellular motility of MLH1 deficient HCT116 and MLH1 deficient HEK293T cells was impaired compared to the MLH1 proficient sister clones. Consequently, overexpression of SPTAN1 increased migration of MLH1 deficient cells while knock down of SPTAN1 decreased cellular mobility of MLH1 proficient cells, indicating SPTAN1-dependent migration ability. CONCLUSIONS: These data suggest that SPTAN1 levels decreased in concordance with MLH1 reduction and impaired cellular mobility in MLH1 deficient colon cancer cells. Therefore, aggressiveness of MLH1-positive CRC might be related to SPTAN1.

Genome-wide analysis associates familial colorectal cancer with increases in copy number variations and a rare structural variation at 12p12.3.

Colorectal cancer (CRC) is one of the most common cancer worldwide. However, a large number of genetic risk factors involved in CRC have not been understood. Copy number variations (CNVs) might partly contribute to the 'missing heritability' of CRC. An increased overall burden of CNV has been identified in several complex diseases, whereas the association between the overall CNV burden and CRC risk is largely unknown. We performed a genome-wide investigation of CNVs on genomic DNA from 384 familial CRC cases and 1285 healthy controls by the Affymetrix 6.0 array. An increase of overall CNV burden was observed in familial CRC patients compared with healthy controls, especially for CNVs larger than 50kb (case/control ratio = 1.66, P = 0.025). In addition, we discovered for the first time a novel structural variation at 12p12.3 and determined the breakpoints by strategic PCR and sequencing. This 12p12.3 structural variation was found in four of 2862 CRC cases but not in 6243 healthy controls (P = 0.0098). RERGL gene (RERG/RAS-like), the only gene influenced by the 12p12.3 structural variation, sharing most of the conserved regions with its close family member RERG tumor suppressor gene (RAS-like, estrogen-regulated, growth inhibitor), might be a novel CRC-related gene. In conclusion, this is the first study to reveal the contribution of the overall burden of CNVs to familial CRC risk and identify a novel rare structural variation at 12p12.3 containing RERGL gene to be associated with CRC.

Role of the oxidative DNA damage repair gene OGG1 in colorectal tumorigenesis.

Biallelic inherited mutations in the oxidative DNA damage repair gene MUTYH predispose to colorectal adenomas and colorectal carcinoma (CRC) with high penetrance. We investigated whether rare inherited variants in other oxidative DNA damage repair genes predisposed to CRC. Single marker association analyses were assessed under an allelic model with Bonferroni correction for multiple testing. All statistical tests were two-sided. A rare inherited nonsynonymous variant in OGG1 (Gly308Glu), the functional partner of MUTYH, was over-represented in case patients with advanced CRC compared with population-based control subjects (n = 36 of 2142 case patients vs n = 15 of 2175 control subjects in the training phase, P = 1.8×10(-3); and n = 22 of 1005 case patients vs n = 8 of 1389 control subjects in the validation phase, P = 4.8×10(-4); P = 1.4×10(-5) combined; odds ratio = 2.92, 95% confidence interval = 1.80 to 4.74). Glycine at residue 308 was highly conserved through evolution, and the glutamic acid substitution was predicted as likely to interfere with function. Biallelic inherited and somatic OGG1 mutations were rarely observed in OGG1 (Gly308Glu) carriers, nor did we find any associated somatic mutator phenotype. These data suggest that OGG1 (Gly308Glu) may act as a low-penetrance allele that contributes to colorectal tumorigenesis.

Hereditary nonpolyposis colorectal cancer (HNPCC)/Lynch syndrome.

BACKGROUND: Hereditary nonpolyposis colorectal cancer HNPCC, Lynch syndrome) is a genetic disease of autosomal dominant inheritance. It is caused by a mutation in one of four genes of the DNA mismatch repair system and confers a markedly increased risk for various types of cancer, particularly of the colon and the endometrium. Its prevalence in the general population is about 1 in 500, and it causes about 2% to 3% of all colorectal cancers. Lynch syndrome is diagnosed in two steps: If it is suspected (because a patient develops cancer at an unusually young age or because of familial clustering), the tumor tissue is analyzed for evidence of deficient mismatch repair (microsatellite instability, loss of mismatch repair protein expression). If such evidence is found, a genetic mutation is sought. The identification of a pathogenic mutation confirms the diagnosis in the patient and enables predictive testing of other family members. Diagnostic evaluations for Lynch syndrome should be carried out with appropriate genetic counseling. METHOD: Selective literature review. RESULTS: Prospective cohort studies from Germany, Finland and the Netherlands have shown that colorectal cancers detected by systematic colonoscopic surveillance tend to be at an earlier stage than those that are discovered after the patients present with symptoms. The Finnish study also showed an overall reduction in cancer risk from colonoscopic polypectomy at regular intervals. CONCLUSION: The studies conducted so far have not yet clearly documented the putative benefit of an individualized, risk-adapted surveillance strategy. Until this is done, patients with Lynch syndrome and healthy carriers of causative mutations should be monitored with annual colonoscopy and (for women) annual gynecological examination.

Risks of less common cancers in proven mutation carriers with lynch syndrome.

PURPOSE: Patients with Lynch syndrome are at high risk for colon and endometrial cancer, but also at an elevated risk for other less common cancers. The purpose of this retrospective cohort study was to provide risk estimates for these less common cancers in proven carriers of pathogenic mutations in the mismatch repair (MMR) genes MLH1, MSH2, and MSH6. PATIENTS AND METHODS: Data were pooled from the German and Dutch national Lynch syndrome registries. Seven different cancer types were analyzed: stomach, small bowel, urinary bladder, other urothelial, breast, ovarian, and prostate cancer. Age-, sex- and MMR gene-specific cumulative risks (CRs) were calculated using the Kaplan-Meier method. Sex-specific incidence rates were compared with general population incidence rates by calculating standardized incidence ratios (SIRs). Multivariate Cox regression analysis was used to estimate the impact of sex and mutated gene on cancer risk. RESULTS: The cohort comprised 2,118 MMR gene mutation carriers (MLH1, n = 806; MSH2, n = 1,004; MSH6, n = 308). All cancers were significantly more frequent than in the general population. The highest risks were found for male small bowel cancer (SIR, 251; 95% CI, 177 to 346; CR at 70 years, 12.0; 95% CI, 5.7 to 18.2). Breast cancer showed an SIR of 1.9 (95% CI, 1.4 to 2.4) and a CR of 14.4 (95% CI, 9.5 to 19.3). MSH2 mutation carriers had a considerably higher risk of developing urothelial cancer than MLH1 or MSH6 carriers. CONCLUSION: The sex- and gene-specific differences of less common cancer risks should be taken into account in cancer surveillance and prevention programs for patients with Lynch syndrome.

Mutation and association analyses of the candidate genes ESR1, ESR2, MAX, PCNA, and KAT2A in patients with unexplained MSH2-deficient tumors.

Lynch syndrome (Hereditary non-polyposis colorectal cancer/HNPCC) is a cancer susceptibility syndrome which is caused by germline mutations in DNA mismatch repair (MMR) genes, in particular MLH1 and MSH2. A pathogenic germline mutation in the respective MMR gene is suggested by the finding of a loss of a mismatch repair protein in tumor tissue on immunohistochemical staining combined with an early age of onset and/or the familial occurrence of colorectal cancer. Pathogenic germline mutations are identifiable in around 60% of patients suspected of Lynch syndrome, depending on the familial occurrence. The aim of the present study was to identify novel susceptibility genes for Lynch syndrome. 64 Healthy controls and 64 Lynch syndrome patients with no pathogenic MSH2 mutation but a loss of MSH2 expression in their tumor tissue were screened for rare and disease causing germline mutations in the functional candidate genes ESR1, ESR2, MAX, PCNA, and KAT2A. Thirty variants were identified, and these were then genotyped in an independent sample of 36 mutation negative Lynch syndrome patients and 234 controls. Since a trend towards association was observed for KAT2A, an additional set of 21 tagging SNPs was analyzed at this locus in a final case-control sample of 142 mutation negative Lynch syndrome patients and 298 controls. The mutation analysis failed to reveal any rare disease-causing mutations. No association was found at the single-marker or haplotypic level for any common disease-modifying variant. The present results suggest that neither rare nor common genetic variants in ESR1, ESR2, MAX, PCNA, or KAT2A contribute to the development of Lynch syndrome.

FHL2 expression in peritumoural fibroblasts correlates with lymphatic metastasis in sporadic but not in HNPCC-associated colon cancer.

Four and a half LIM domain protein-2 (FHL2) is a component of the focal adhesion structures and has been suggested to have an important role in cancer progression. This study analyses the role of FHL2 in peritumoural fibroblasts of sporadic and hereditary non-polyposis colorectal cancer (HNPCC). Tissue specimens of 48 sporadic and 49 hereditary colon cancers, respectively, were stained immunohistochemically for FHL2, transforming growth factor (TGF)-ß1 ligand and α-SMA. Myofibroblasts at the tumour invasion front co-expressed α-SMA and FHL2. Sporadic colon cancer but not HNPCC cases showed a correlation between TGF-ß1 expression of the invading tumour cells and FHL2 staining of peritumoural myofibroblasts. Overexpression of FHL2 in peritumoural myofibroblasts correlated to lymphatic metastasis in sporadic colon cancer but not in HNPCC. In cultured mouse fibroblasts, TGF-ß1 treatment induced myofibroblast differentiation, stimulated FHL2 protein expression and elevated number of migratory cells in transwell motility assays, suggesting that FHL2 is regulated downstream of TGF-ß. Physical contact of colon cancer cells and myofibroblasts via FHL2-positive focal adhesions was detected in human colon carcinoma tissue and in co-culture assays using sporadic as well as HNPCC-derived tumour cell lines. Our data provide strong evidence for an important role of FHL2 in the progression of colon cancers. Tumour-secreted TGF-ß1 stimulates FHL2 protein expression in peritumoural fibroblasts, probably facilitating the invasion of tumour glands into the surrounding tissue by enhanced myofibroblast migration and tight connection of fibroblasts to tumour cells via focal adhesions. These findings are absent in HNPCC-associated colon cancers in vivo and may contribute to a less invasive and more protruding tumour margin of microsatellite instable carcinomas.

Association between TAS2R38 gene polymorphisms and colorectal cancer risk: a case-control study in two independent populations of Caucasian origin.

Molecular sensing in the lingual mucosa and in the gastro-intestinal tract play a role in the detection of ingested harmful drugs and toxins. Therefore, genetic polymorphisms affecting the capability of initiating these responses may be critical for the subsequent efficiency of avoiding and/or eliminating possible threats to the organism. By using a tagging approach in the region of Taste Receptor 2R38 (TAS2R38) gene, we investigated all the common genetic variation of this gene region in relation to colorectal cancer risk with a case-control study in a German population (709 controls and 602 cases) and in a Czech population (623 controls and 601 cases). We found that there were no significant associations between individual SNPs of the TAS2R38 gene and colorectal cancer in the Czech or in the German population, nor in the joint analysis. However, when we analyzed the diplotypes and the phenotypes we found that the non-taster group had an increased risk of colorectal cancer in comparison to the taster group. This association was borderline significant in the Czech population, (OR = 1.28, 95% CI 0.99-1.67; P(value) = 0.058) and statistically significant in the German population (OR = 1.36, 95% CI 1.06-1.75; P(value) = 0.016) and in the joint analysis (OR = 1.34, 95% CI 1.12-1.61; P(value) = 0.001). In conclusion, we found a suggestive association between the human bitter tasting phenotype and the risk of CRC in two different populations of Caucasian origin.

Biallelic MLH1 SNP cDNA expression or constitutional promoter methylation can hide genomic rearrangements causing Lynch syndrome.

BACKGROUND: A positive family history, germline mutations in DNA mismatch repair genes, tumours with high microsatellite instability, and loss of mismatch repair protein expression are the hallmarks of hereditary non-polyposis colorectal cancer (Lynch syndrome). However, in ~10-15% of cases of suspected Lynch syndrome, no disease-causing mechanism can be detected. METHODS: Oligo array analysis was performed to search for genomic imbalances in patients with suspected mutation-negative Lynch syndrome with MLH1 deficiency in their colorectal tumours. RESULTS AND CONCLUSION: A deletion in the LRRFIP2 (leucine-rich repeat flightless-interacting protein 2) gene flanking the MLH1 gene was detected, which turned out to be a paracentric inversion on chromosome 3p22.2 creating two new stable fusion transcripts between MLH1 and LRRFIP2. A single-nucleotide polymorphism in MLH1 exon 8 was expressed from both alleles, initially pointing to appropriate MLH1 function at least in peripheral cells. In a second case, an inherited duplication of the MLH1 gene region resulted in constitutional MLH1 promoter methylation. Constitutional MLH1 promoter methylation may therefore in rare cases be a heritable disease mechanism and should not be overlooked in seemingly sporadic patients.

Polymorphisms in CTNNBL1 in relation to colorectal cancer with evolutionary implications.

Colorectal cancer (CRC) is a complex disease related to environmental and genetic risk factors. Several studies have shown that susceptibility to complex diseases can be mediated by ancestral alleles. Using RNAi screening, CTNNBL1 was identified as a putative regulator of the Wnt signaling pathway, which plays a key role in colorectal carcinogenesis. Recently, single nucleotide polymorphisms (SNPs) in CTNNBL1 have been associated with obesity, a known risk factor for CRC. We investigated whether genetic variation in CTNNBL1 affects susceptibility to CRC and tested for signals of recent selection. We applied a tagging SNP approach that cover all known common variation in CTNNBL1 (allele frequency >5%; r(2)>0.8). A case-control study was carried out using two well-characterized study populations: a hospital-based Czech population composed of 751 sporadic cases and 755 controls and a family/early onset-based German population (697 cases and 644 controls). Genotyping was performed using allele specific PCR based TaqMan® assays (Applied Biosystems, Weiterstadt, Germany). In the Czech cohort, containing sporadic cases, the ancestral alleles of three SNPs showed evidence of association with CRC: rs2344481 (OR 1.44, 95%CI 1.06-1.95, dominant model), rs2281148 (OR 0.59, 95%CI 0.36-0.96, dominant model) and rs2235460 (OR 1.38, 95%CI 1.01-1.89, AA vs. GG). The associations were less prominent in the family/early onset-based German cohort. Data derived from several databases and statistical tests consistently pointed to a likely shaping of CTNNBL1 by positive selection. Further studies are needed to identify the actual function of CTNNBL1 and to validate the association results in other populations.

Missense variants in hMLH1 identified in patients from the German HNPCC consortium and functional studies.

Missense mutations of the DNA mismatch repair gene MLH1 are found in a significant fraction of patients with Lynch syndrome (hereditary nonpolyposis colorectal cancer, HNPCC) and their pathogenicity often remains unclear. We report here all 88 MLH1 missense variants identified in families from the German HNPCC consortium with clinical details of these patients/families. We investigated 23 MLH1 missense variants by two functional in vivo assays in yeast; seven map to the ATPase and 16 to the protein interaction domain. In the yeast-2-hybrid (Y2H) assay three variants in the ATPase and twelve variants in the interaction domain showed no or a reduced interaction with PMS2; seven showed a normal and one a significantly higher interaction. Using the Lys2A (14) reporter system to study the dominant negative mutator effect (DNE), 16 variants showed no or a low mutator effect, suggesting that these are nonfunctional, three were intermediate and four wild type in this assay. The DNE and Y2H results were concordant for all variants in the interaction domain, whereas slightly divergent results were obtained for variants in the ATPase domain. Analysis of the stability of the missense proteins in yeast and human embryonic kidney cells (293T) revealed a very low expression for seven of the variants in yeast and for nine in human cells. In total 15 variants were classified as deleterious, five were classified as variants of unclassified significance (VUS) and three were basically normal in the functional assays, P603R, K618R, Q689R, suggesting that these are neutral.

Recurrence and variability of germline EPCAM deletions in Lynch syndrome.

Recently, we identified 3' end deletions in the EPCAM gene as a novel cause of Lynch syndrome. These truncating EPCAM deletions cause allele-specific epigenetic silencing of the neighboring DNA mismatch repair gene MSH2 in tissues expressing EPCAM. Here we screened a cohort of unexplained Lynch-like families for the presence of EPCAM deletions. We identified 27 novel independent MSH2-deficient families from multiple geographical origins with varying deletions all encompassing the 3' end of EPCAM, but leaving the MSH2 gene intact. Within The Netherlands and Germany, EPCAM deletions appeared to represent at least 2.8% and 1.1% of the confirmed Lynch syndrome families, respectively. MSH2 promoter methylation was observed in epithelial tissues of all deletion carriers tested, thus confirming silencing of MSH2 as the causative defect. In a total of 45 families, 19 different deletions were found, all including the last two exons and the transcription termination signal of EPCAM. All deletions appeared to originate from Alu-repeat mediated recombination events. In 17 cases regions of microhomology around the breakpoints were found, suggesting nonallelic homologous recombination as the most likely mechanism. We conclude that 3' end EPCAM deletions are a recurrent cause of Lynch syndrome, which should be implemented in routine Lynch syndrome diagnostics.