Introduction

Hereditary nonpolyposis colorectal cancer (HNPCC) or Lynch's syndrome, which represents the most common form of inherited colorectal cancer, results from germline alterations of the DNA mismatch repair genes MSH2, MLH1, MSH6 and PMS2 (for review see Gruber1). Identification of the causal alteration is essential for an appropriate clinical management of the families as it allows to offer to carriers an efficient tumour detection, based on colonoscopy and hysteroscopy every 2 years from 20 years of age and to avoid an inappropriate follow-up in noncarriers. Point mutations within MSH2, MLH1, and MSH6 account approximately for 25, 25 and less than 5% of the HNPCC families, respectively. Numerous reports, initially based on Southern blot or RT-PCR, then on QMPSF (Quantitative Multiplex PCR of Short Fluorescent Fragments), MLPA (Multiplex Ligation-dependent Probe Amplification) or conversion analysis have now underlined the significant contribution of MSH2 and MLH1 genomic deletions to HNPCC.2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24 At present, 43 MSH2 and 33 MLH1, distinct exonic deletions, removing either a single or multiple exons, have been reported.2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24 We previously estimated that genomic deletions of MSH2 and MLH1 respectively, account for approximately 10 and 4% of HNPCC families fulfilling Amsterdam criteria.6, 13 All the MSH2 genomic deletions detected so far have been shown to result from unequal Alu-mediated recombinations6, 8, 11, 12, 18, 20 and we found that the density of Alu S repeats within the MSH2 5′ region was remarkably high, compared with other human genes.18 In contrast, MLH1 genomic deletions have been found to involve not only Alu repeats, but also nonhomologous sequences.2, 3, 9, 20, 21 As expected, exonic MSH2 and MLH1 duplications, the reciprocal genomic alterations of deletions resulting from unequal crossingover, have also been detected in HNPCC families, but at a much lower frequency. We reported, using QMPSF, the first case of a MSH2 duplication involving exons 9–10 that was confirmed by long-range PCR5 and we subsequently detected another MSH2 duplication involving exons 7–8.13 Only eight other MMR duplications have been reported in studies based either on MLPA or conversion analysis: for MSH2, duplication of exons 5–6,19 exon 7,15 exons 11–16;17 for MLH1, duplication of exon 1,16 exon 3,16 exon 19,16 exons 6–12,19 and exons 7–12.25 We now report the detection of 10 MSH2 and MLH1 exonic duplications, 7 of which are novel. The aim of this report was to highlight the contribution to HNPCC of MMR genomic duplications and to estimate the frequency of these alterations.

Materials and methods

Patients

Patients analysed in this study, either fulfilled Amsterdam (AMS) criteria (at least three relatives with colorectal cancer, cancer of the endometrium, small bowel, ureter or renal pelvis, one of whom is a first-degree relative of the other two; at least two successive generations affected; and at least one cancer diagnosed before age 5026) or were selected because of a suggestive familial history associated to a microsatellite instability (MSI) phenotype in the tumour. A total of 962 and 611 index cases were, respectively, screened for MSH2 and MLH1 genomic rearrangements.

QMPSF analysis

The QMPSF protocol used to screen for genomic rearrangements of MSH2 and MLH1 has been described previously.5, 6 Briefly, short fragments corresponding to the 16 MSH2 and 19 MLH1 exons and to control genes, were simultaneously PCR-amplified using dye-labeled primers. Primer sequences are available upon request. PCR products were separated by electrophoresis on an ABI-377 or 3100 automated sequencer (Applied Biosystems, Foster, USA). Data were analysed using the Gene Scanner Model 672 Fluorescent Fragment Analyzer (Applied Biosystems, Foster, USA). For each index case, QMPSF profile was superposed to that generated from a control-DNA by adjusting to the same level the peaks obtained for the control amplicon. Based on an extensive number of QMPSF analyses (including >50 genes and >1000 tests) performed in our laboratory, we have estimated the variability of gene copy number measurements to be 0.1 among control samples. Thus, copy number ratios between 0.9 and 1.1 indicate two gene copies, and copy number ratios between 1.4 and 1.6 indicate three copies.5, 27, 28, 29 Each positive result was confirmed by another QMPSF analysis performed on an independent sample.

MLPA

MLPA was performed using the SALSA P003 (MLH1/MSH2) MLPA kit (MRC Holland, Amsterdam, The Netherlands) according to the manufacturer's protocol. Fragments were separated and analyzed with a 377XL genetic analyzer and Genescan 3.1 software as well as with a 3100-Avant capillary genetic analyzer and GeneMapper 3.5 software (Applied Biosystems). Peak areas and heights were then exported to a Microsoft Excel spreadsheet (http://leedsdna.info/science/dosage/Conventional_MLPA/Conventional_MLPA.htm) and calculations were carried out according to the method described by Taylor et al.10

Results

Using QMPSF, we detected 10 distinct genomic duplications of MSH2 or MLH1 genes in 11 families: seven duplications involved MSH2 exons 1–2–3, exons 4–5–6, exon 7, exons 7–8, exons 9–10, exon 11, and exon 15, and three duplications involved MLH1 exons 2–3, exon 4 and exons 6–7–8 (Table 1). As illustrated by Figure 1, exonic duplications could easily be detected by QMPSF by a 1.5 increase of the corresponding peaks. Only the duplication of MSH2 exons 9–10 detected in family L14 had previously been confirmed by long-range PCR.5 MLPA was therefore used in the present study to confirm the nine other duplications. In each case, MLPA confirms the presence of the exonic duplication(s) detected by QMPSF. These duplications generate in MLPA an increase of the corresponding peak area(s) above 1.3. In one family (G1), we observed a discrepancy between QMPSF and MLPA concerning the extent of duplication. Indeed, QMPSF revealed the presence of a MSH2 duplication involving exons 5–6, whereas MLPA suggested that this duplication also involved exon 4. In order to clarify the situation, we designed new QMPSF assays exploring specifically intron 3, exon 4 and intron 4 of MSH2. These complementary QMPSF analyses revealed that, within the G1 family, the duplication included exon 4, but did not extend to nucleotide c.646–121 corresponding to the sense primer used in the first QMPSF assay, and therefore did not involve the first 1920 nucleotides of intron 3.

Table 1 Duplications of MSH2 or MLH1 in HNPCC families
Full size table
Figure 1
figure 1

Detection by QMPSF of duplications of the MSH2 and MLH1 genes. The left panels show the QMPSF electropherograms in four families and the right panels the results of the corresponding MLPA analyses. In each QMPSF panel, the electropherogram of the patient (in red) was superimposed to that of a normal control (in blue) by adjusting to the same level the peaks obtained for the control (C) amplicon. The Y-axis displays fluorescence in arbitrary units, and the X-axis indicates the size in bp. MSH2 or MLH1 exons are indicated. Lines indicate the top of the normal blue peak. Duplications are detected by a 1.5 increase of the peaks compared with a normal control. (a) Detection of an MSH2 exon 11 duplication in family R2; (b) Detection of an MSH2 exon 15 duplication in family M2; (c) detection of an MLH1 exons 2 and 3 duplication in family M1; (d) detection of an MLH1 exons 6, 7 and 8 duplication in family G2.

Full size image

Discussion

One hypothesis that might explain the difference in the numbers of MMR genes genomic deletions and duplications reported so far is the technical difficulty to detect duplications versus deletions. Indeed, the largest series of HNPCC patients screened for MMR genomic rearrangements have been analysed using QMPSF13 or MLPA.19 In both methods, heterozygous deletions result in a 0.5 reduction of the corresponding peaks, whereas duplications generate a 1.5 increase. This could explain why duplications are more difficult to detect than deletions, and this might represent a technical bias in the estimation of MMR duplications frequency. QMPSF analysis allowed us in this large series of suspected HNPCC patients to detect 10 distinct MSH2 or MLH1 partial genomic duplications, which were all confirmed by another independent method. This shows, as previously illustrated for the BRCA1,27 BRCA2,28 and APP genes,29 that QMPSF is a reliable method for the detection of genomic duplications. Although we cannot provide in this study an accurate estimation of the contribution of MSH2 and MLH1 duplications to HNPCC, as the population analysed was heterogeneous including both AMS+ patients and AMS− patients with an MSI phenotype in the tumour, the frequency of MSH2 and MLH1 duplications appears low. In our series, 962 and 611 index cases were, respectively, screened for MSH2 and MLH1 genomic rearrangements. This allowed us to detect, in 115 families, 30 and 18 distinct exonic deletions of MSH2 and MLH1, respectively (Charbonnier et al.18 and unpublished results). In contrast, in the same series, we detected in 11 families, 7 and 3 exonic duplications of MSH2 and MLH1. Therefore, MSH2 and MLH1 duplications account probably for approximately 1% of the HNPCC cases. Although genomic duplications of MSH2 and MLH1 are less frequent that genomic deletions, their presence should nevertheless be considered in HNPCC families without detectable point mutations.