Frequency of Rearrangements in Lynch Syndrome Cases Associated with MSH2: Characterization of a New Deletion Involving both EPCAM and the 5′ Part of MSH2

See perspective p. 1527

Hereditary nonpolyposis colorectal cancer (HNPCC) or Lynch syndrome is the most frequent autosomal-dominant colorectal cancer (CRC) susceptibility syndrome caused by mutations inactivating one of the genes of the mismatch repair (MMR) system, most frequently in MLH1 and MSH2 and less often in MSH6 and PMS2 (1, 2–4). The phenotype of tumors from these patients is characterized by widespread microsatellite instability (MSI) and loss of protein expression from the affected enzyme detected by immunohistochemical (IHC) staining. This syndrome is characterized by a high risk of early-onset CRC and several other extracolonic malignant tumors, especially endometrial cancer in women (5).

Mutations in 2 of these MMR genes, MSH2 and MLH1, account for the majority (about 40%) of the patients with HNPCC (6). Although the majority of the genetic defects in the human MMR genes responsible for HNPCC are a result of point mutations and small insertions and deletions, a substantial proportion results from gross genomic rearrangements. The role of genomic rearrangements in the etiology of HNPCC has been under investigation because the screening for large deletions [e.g., by multiplex ligation–dependent probe amplification (MLPA; ref. 7) or other techniques] is still not universal in diagnostic laboratories. One mechanism that could originate large genomic rearrangements is the unequal homologous recombination between repeat sequences with a high degree of homology of short interspersed nuclear elements (SINE), including Alu repeats (8). In particular, there is a high incidence of genomic deletion in the MSH2 gene (4). This item has been reported as being due to the presence of a higher percentage of repetitive elements (Alu repeats) in the MSH2 gene. Most of them are present in the first half of the gene, including the 5′ upstream sequence of MSH2 (EPCAM/TACSTD1 and promoter region of MSH2).

It has recently been shown that constitutional 3′ end deletions of the EPCAM gene (OMIM#185535; non-MMR gene) can cause Lynch syndrome through the epigenetic silencing of MSH2 in EPCAM-expressing tissues, resulting in tissue-specific MSH2 deficiency (9). Thus, deletions of the last exon of EPCAM constitute a distinct class mutation associated with Lynch syndrome. Several investigators have reported families with EPCAM deletions (9–13).

In a recent study, Kempers and colleagues (13), established different cancer risks associated with EPCAM deletions depending on whether a deletion affects only the EPCAM gene or both the EPCAM and its neighboring gene MSH2 (EPCAM-MSH2). These risks are then compared with those for Lynch syndrome carriers of a mutation in MMR genes. This is the first study that describes the cumulative cancer risks and cancer profile of EPCAM deletion carriers. They show a profound difference in the frequency of cases of endometrial cancer in this group compared with other Lynch syndrome families with MMR gene mutations.

Strikingly, endometrial cancer was observed only in carriers with large EPCAM deletions that extended close to the MSH2 gene. The authors described the cumulative risk of endometrial cancer at 70 years of age in EPCAM deletion carriers as being 12%. This risk is much lower than that for MSH2 mutation carriers (51%) or combined MSH2-EPCAM deletion carriers (55%). These data suggest that the risk for endometrial cancer in carriers of EPCAM deletions is dependent on the size and location of the deletion. The exact criteria of deletions, conferring a low risk of endometrial cancer, remain to be defined by further assessments of endometrial cancer incidence in carriers of different EPCAM deletions and analyses of the EPCAM-MSH2 intergenic region for transcription-mediating capacity.

These results highlight, on the one hand, the importance of carrying out strategies for defining the exact extent of rearrangements and, on the other hand, that the determination of the tumor spectrum and age-specific cancer risk in families carrying different mutations associated with Lynch syndrome will help to generate optimal recognition and surveillance strategies.

This study was designed to confirm the prevalence of large genomic rearrangements in MSH2 and EPCAM genes in Spanish families. To characterize them, a study was proposed at the molecular level to determine their extent, identify their break points, and characterize the impact of the genomic alteration on the correct splicing of the gene. We also evaluated whether different types of MSH2 gene changes (point mutations or deletion extensions affecting MSH2 or EPCAM-MSH2) were associated with distinct clinical characteristics within the present study series.

Samples from 159 independent families were referred for MMR mutation analysis under the Junta de Castilla y Leon Cancer Genetic Counselling Program (Spain). The criteria for entry into the mutational study were defined in accordance with Amsterdam or Bethesda guidelines. Informed consent was obtained from each patient.

DNA was extracted from blood samples from all of our patients, using the MagNa Pure Systems (Roche).

IHC staining of tumors for MLH1, MSH2, and MSH6 genes was analyzed by a pathologist in the General Yagüe Hospital, Burgos (Spain).

MSI analysis was carried out on matched normal and tumor DNA pairs using the National Cancer Institute/International Collaborative Group on HNPCC reference marker panel (including 2 mononucleotide repeats, Bat-25 and Bat-26, and 3 dinucleotide repeats, D2S123, D5S346, and D17S250). DNA was extracted using the DNeasy Tissue Kit (Qiagen). Fluorescently labeled PCR products were detected using the ABI 3130 Genetic Analyzer and the GeneScan Software. We classified tumors as MSI-positive only if 2 or more markers showed instability (14).

Samples from subjects with MSI were analyzed for the detection of point mutation, using heteroduplex analysis by capillary array electrophoresis (HA-CAE). This method was developed in our laboratory (15), and the validation for MMR genes has been recently published (16).

The selection of genes for analysis was based on IHC results. DNA from peripheral blood leukocytes was used for the analysis. Fragments showing an HA-CAE–altered pattern were sequenced with the BigDye Terminator Sequencing Kit v3.1 (Applied Biosystems) with unlabeled forward and reverse primers on an ABI 3100 DNA sequencer (4 capillaries; Applied Biosystems).

To detect genomic deletions affecting the EPCAM and MSH2 gene loci, the MLPA Assay (MRC-Holland) was used. The test kits used were SALSA MLPA kits P003 and P008 (MRC-Holland) following the manufacturer's instructions. The P003 MLH1/MSH2 Kit contains oligonucleotide probes targeting all exons of MSH2 and an additional probe to test exon 1. The P008 MSH6/PMS2 Kit contains probes targeting EPCAM/TACSTD1 exons 3 and 8, one 27 kb upstream and the other 15 kb upstream from the MSH2 gene. PCR products were analyzed on an ABI 3130 capillary sequencer using GeneMapper software (Applied Biosystems).

The synthesis of cDNA was carried out with the High Capacity cDNA Reverse Transcription Kit (Applied Biosystems), using DNase-treated RNA in the presence of random primers. The cDNA amplification was carried out with specific primers that encompassed the predicted rearrangement designed for the coding sequences flanking the putative mutation. Short amplicons from reverse transcription PCR (RT-PCR) were sequenced with the same primers: TASCTD1-Ex7-FW, 5′-ggttgtggtgatagcagttg-3′; MSH2-Ex4-Rev, 5′-ggttgaggtcctgataaatg-3′.

An array comparative genomic hybridization (aCGH) strategy was used to confirm the extension of deletions identified by MLPA. A specific human aCGH 44K was designed by NIMGenetics for coverage of chromosome 2: 47419322-47580004 (NCBI 36) with a high resolution (500 bp).

On the basis of the information obtained from the aCGH, the interspersed repeats in regions around the break points were examined using the RepeatMasker program (www.repeatmasker.org).

To characterize the rearrangements and determine the exact site of their break points, we carried out long-range PCR of genomic DNA using primers designed to span the putative break points as follows: EPCAM-In6Fw, 5′-TCCCATTTTTAGACCCCAAA-3′; MSH2-In3Rv, 5′ GTGGCTCATGCCTGTAATCC-3′. The Expand Long Template PCR System (Roche Diagnostics GmbH) was used according to the manufacturer's protocol. PCR products were separated on a 0.8% agarose gel and visualized by ethidium bromide staining. PCR products containing the expected rearrangement were cut from the gel and purified using GFX PCR DNA and the Gel Band Purification Kit (Illustra; GE Healthcare UK Limited). Isolated PCR fragments were sequenced by primer walking on both strands using the BigDye V3.1 Terminator Kit (Applied Biosystems) and an automated sequencer.

As a diagnostic tool, we designed a multiplex PCR strategy based on 3 primer sequences (1 forward and 2 reverse) to screen these deletions in patients' first-degree relatives. This analysis of genomic DNA produces a unique band sized 705 bp when primers EPCAM-In6Fw and MSH2-In3Rv are used. This situation happens in a wild-type (WT) case and produces an additional band when the third primer EPCAM-In6′Rev (5′-CAATGTGCAAGACACTGATATGAT-3′) is used in deletion carrier samples (Fig. 3). The experimental conditions are supplied as Supplementary Material S1.

A total of 35 MSI families were tested for point mutation in MMR genes by combining HA-CAE and sequencing analysis. Eighteen of these families showed a deficiency in the expression of MSH2/MSH6 proteins in tumors and the screening began for the MSH2 gene. A total of 8 families with a pathogenic germline mutation were detected in the MSH2 gene (Fig. 1). Clinicopathologic features, molecular findings of the index patients, and sample numbers are listed in Supplementary Material S2.

The 10 families tested negative for point mutations in MSH2 were screened for the presence of large genomic rearrangements in EPCAM/MSH2 using MLPA (Fig. 1). MLPA detected 2 different rearrangements in MSH2, involving the deletion of exon 7 and exons 4 to 8, in 3 and 4 unrelated families, respectively, using the SALSA MLPA Kit P003 MSH2/MLH1 (Table 1). Also, a new deletion encompassing EPCAM-MSH2 in 2 members from an additional family was detected using both the SALSA MLPA Kit P003 and Kit P008 (Table 1). No cases involving EPCAM deletion alone (without MSH2 5′ involvement) were detected.

Two different rearrangements in MSH2, involving the deletion of exon 7 and exons 4 to 8, were previously confirmed through an RT-PCR analysis and sequencing (16). The exact break points were determined in a companion article by Pérez-Cabornero and colleagues (in this issue). Both of them are positioned within Alu elements (Table 1).

To confirm the extension of multiexonic MSH2 deletion involving exons 1 to 3 in the 2 carriers (cases C43 and C132), we used the SALSA MLPA Kit P008 PMS2/MSH6. A reduction of the peak area at the probe was observed for the MSH2 exon, which also exhibited an aberrant hybridization signal for 1 or 2 EPCAM probes (the one located in exon 8), which was confined to the deletion beginning in the 3′ region of the EPCAM gene, located upstream of MSH2 (Fig. 2A). RT-PCR on RNA from index subject C43, in which exon 7 of EPCAM is fused to exon 4 of MSH2, was detected (Fig. 2C). An aCGH analysis was used to determine the extension of the deletion. The results obtained indicate a deletion extension in chromosome 2: 47464677-47492513 (NCBI 36).

Primer pairs were designed to obtain a patient-specific junction fragment, which produced a length of approximately 550 bp (Fig. 3A). Sequence analysis of the junction fragment confirmed a 28.9-kb deletion (c. 859-1860_MSH2:646-254del; Fig. 3B). The break points are located within 2 interspersed elements, one Alu-Y and an Alu-Sx in inverse orientation (Fig. 3A). Interestingly, the crossover site lies within a 3-bp sequence of perfect identity. This deletion has not been reported in previous studies, but many deletions involving exons 1to 3 of MSH2 have been described in the Leiden Open Variation Database (http://www.chromium.liacs.nl/LOVD2/colon_cancer/home.php) and their break points have still not been investigated.

We designed a PCR test based on the discrimination of deletion or WT alleles to screen this complex deletion in first-degree relatives of the VA-25 family (Fig. 3C). This procedure is faster, cheaper, and easier than MLPA.

Combined EPCAM-MSH2 deletion was detected in a 34-year-old man diagnosed with a primary CCR at the same age (C43 case). An additional asymptomatic 43-year-old woman carrier was also detected (C132 case). The pedigree of the family is shown in Figure 2B. This family fulfilled the Amsterdam I criteria for Lynch syndrome. There were 6 cancers associated to the Lynch syndrome spectrum in 2 consecutive generations, 2 of which were diagnosed before the age of 50. Individuals C43 and C132 were carriers of the combined EPCAM-MSH2 deletion. Family members are currently undergoing analysis for this mutation to predict their risk of developing Lynch syndrome.

The present patients with deletions in MSH2 were divided into groups (with vs. without) involving the EPCAM 3′ region gene, and they were compared with index individuals from our regional registry known to carry MSH2 point mutations (Table 2).

We compared clinical information of individuals in the study in whom a point mutation in MSH2 (n = 19) was identified with those in whom such a deletion in the same gene was identified, depending on the extension [only MSH2 gene (n = 15) or EPCAM-MSH2 gene (n = 1); Table 2]. The average age of CCR diagnosis for those cases with a point mutation was approximately 47 years compared with approximately 38 and 34 years for those in the study with deletion abnormalities in males. In females, the average age of CCR diagnosis for patients with a point mutation was 50 years as compared with 33 and approximately 38 years for MSH2 or EPCAM-MSH2 deletion–positive cases, respectively. The sex distribution (male/female) was approximately 6:5 (MSH2 point mutation) and 1:1 for both cases MSH2 or EPCAM-MSH2 deletion. The average age of endometrial diagnosis for those cases with a point mutation was 43 years compared with approximately 51 years for those in the study with an MSH2 deletion.

When compared with patients with point mutations, MSH2 deletion patients had a 2.5 times higher chance of developing endometrial cancer. EPCAM-MSH2 deletion patients did not present any endometrial cancer in women in the family described.

In this study, we carried out MSH2/EPCAM MLPA analyses (10) of high-risk Lynch syndrome cancer patients with loss of MSH2 protein expression in the tumor. We have found copy number alterations in about 23% of high-risk families for HNPCC with MSI (8 of 35) who were previously screened negative for point mutations in MLH1, MSH2, and MSH6 genes. MLPA detected 2 different rearrangements in MSH2 in 7 unrelated families, and a new deletion encompassing EPCAM-MSH2 was identified in an additional family. Therefore, all of them were found to involve the MSH2 gene. Particularly, 2 new rearrangements encompassing exon 7 and exon 4 to 8 deletion in MSH2 were detected in 3 and 4 nonrelated families, respectively, and an additional deletion affecting both EPCAM and its neighboring gene MSH2 (EPCAM-MSH2). However, the 2 rearrangements involving the MSH2 gene occurred in multiple kindred, and, in an article to be published in parallel, we will show evidence of a common origin in the 2 deletion of MSH2 (data published in Pérez-Cabornero and colleagues (in this issue).

Our results provide the first evidence that, as in many other studied populations, large genomic changes involving the MSH2 gene is an important event in our HNPCC family series (almost 50% of pathogenic mutations). The frequency of large rearrangements in MSH2, as compared with MLH1, depends on the studied population. Several studies have shown that these rearrangements correspond between 15% and 55% of the mutations in MMR genes (17). In a study of the Spanish population, an exceptionally low frequency of rearrangements in MLH1/MSH2 genes (<1.5%) was reported (18), although a higher frequency of rearrangements was found in a Basque Country population (∼25%; ref. 19). Our data imply that the high frequency of deletions in this study is caused by strong founder effects in our population.

On the other hand, deletions in the EPCAM gene have been reported in several populations with a different frequency, from 19% (10) to 40% (9) in Hungarian and Dutch populations, respectively. In the Spanish cohort, one family carrier of this kind of mutation has been identified (∼10% incidence; ref. 11). All the subjects of EPCAM deletion carriers were selected from patients with tumors with MSH2 loss and MSI who lacked an MSH2 or MSH6 point mutation. No deletion confined to the EPCAM gene alone was identified in our patient series, but we have identified and characterized a new combined deletion EPCAM-MSH2 by break point analysis, encompassing from the 3′ end region of EPCAM (TACSTD1; exons 8 and 9) to the 5′ initial sequences of the MSH2 (exons 1–3). From this, one expressed EPCAM-MSH2 fusion transcript was identified and it was predicted to be in frame. The tumors of the carriers show high levels of MSI and MSH2 protein loss.

Knowing the frequency of deletion in the population can significantly influence the screening algorithms for patients at risk of HNPCC. Considering these results, as well as the rapid and easy to carrying out techniques for the characterization of these mutations (such as MLPA, RT-PCR, aCGH, and sequencing), we proposed that the mutation screening algorithm should begin with MLPA and not with conventional screening/scanning methods, especially in cases in which the protein expression pattern of the tumor shows a loss of MSH2 protein or is unknown. Also, our PCR-based assay could be useful for rapid cost-effective HNPCC screening of EPCAM-MSH2 deletion first-degree relatives.

The possibility that cancer risks vary, depending on the type of MMR gene mutation, may have significant implications for cancer screening recommendations. The proportion of pathogenic point mutations versus rearrangements in MSH2 versus EPCAM-MSH2 deletions identified in this set of samples is of 50% versus 43.8% versus 6.3%. To the best of our knowledge, the clinical features of families carrying the detected rearrangements were not different from those of families exhibiting other types of mutations, despite results published by other groups such as Kempers and colleagues (13), who showed that endometrial cancer was observed only in carriers with large EPCAM deletions that extended close to or into the MSH2 gene.

In conclusion, our data show that large genomic rearrangements occur in MSH2 with a high frequency and genetic evidence has been provided that a certain proportion of these deletions involve the EPCAM gene. The need to incorporate techniques to routinely detect large genomic rearrangements and confirm the extension of the deletions involving the MSH2 gene is emphasized, as it could be involved in the predisposing to Lynch syndrome.

No potential conflicts of interest were disclosed.

The authors thank the National DNA Bank of Salamanca and the Science Foundation AECC. The authors also thank Noemy Martínez and Lara Hernández for their excellent technical support and Alan Hynds for his critical reading of the manuscript.

This work was supported by the Regional Government of Castilla y León (Spain). L. Perez-Cabornero and A. Acedo were supported by a predoctoral FPI fellowship from the Government of Castilla y León (Spain).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.