Article Text


Screening for genomic rearrangements of the MMR genes must be included in the routine diagnosis of HNPCC
  1. F Di Fiore1,
  2. F Charbonnier1,
  3. C Martin1,
  4. S Frerot1,
  5. S Olschwang2,
  6. Q Wang3,
  7. C Boisson2,
  8. M-P Buisine4,
  9. M Nilbert5,
  10. A Lindblom6,
  11. T Frebourg1
  1. 1Department of Genetics, University Hospital, and INSERM EMI 9906 -IFRMP, Faculty of Medicine, Rouen, France
  2. 2INSERM U434 and Saint Antoine Hospital, Paris, France
  3. 3Department of Molecular Oncology, Centre Léon Bérard, Lyon, France
  4. 4Laboratory of Biochemistry and Molecular Biology, University Hospital, Lille, France
  5. 5Department of Oncology, University Hospital, Lund, Sweden
  6. 6Department of Clinical Genetics, Karolinska Hospital, Stockholm, Sweden
  1. Correspondence to:
 Dr T Frebourg
 Department of Genetics, University Hospital, and INSERM EMI 9906-IFRMP, Faculty of Medicine, 22 Bd. Gambetta, 76183 Rouen, France;

Statistics from

In hereditary non-polyposis colorectal cancer (HNPCC), the most common form of inherited colorectal cancer, detection of the causal alteration of the mismatch repair (MMR) gene involved is essential for proper management of the families. This will allow the identification of relatives with high risk for colorectal or endometrial cancer, who require the appropriate screening and, conversely, will avert useless surveillance in non-carrier relatives. Mutational studies,1 based on conventional screening methods, have indicated that point mutations of MSH2, MLH1, or MSH6 can be detected in approximately 55% of the families, fulfilling the Amsterdam (AMS) criteria. These stipulate:

  • at least three relatives with colorectal cancer, or 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 the age of 50 years.2

In a recent study, we showed that genomic rearrangements of MSH2 are involved in approximately 20% of the AMS+ HNPCC families without detectable point mutations within MSH2 or MLH1.3 This study was performed using quantitative multiplex PCR of short fluorescent fragments (QMPSF), which can easily detect heterozygous genomic deletions and duplications.3–7 This method is based on the simultaneous amplification of short genomic sequences under quantitative conditions, using dye labelled primers, and the superimposition of the electropherograms of patients and controls.

Key points

  • In hereditary non-polyposis colorectal cancer (HNPCC), point mutations of MSH2, MLH1, or MSH6 are detected in approximately half of the families involved, which therefore fulfil the Amsterdam criteria (AMS).

  • We analysed MSH2 in 120 AMS+ and 212 AMS− HNPCC families without MSH2 or MLH1 point mutations, using quantitative multiplex PCR of short fluorescent fragments(QMPSF). We identified in 22% of the AMS+ and in 8% of the AMS− families 19 distinct exonic deletions and two cases of duplication of MSH2. We detected seven distinct 5′ breakpoints in the deletions removing exon 1. Specific QMPSF analysis of the MSH2 promoter in 65 AMS+ families, without MSH2/MLH1 point mutations or MSH2 exonic deletion, revealed only one case of promoter deletion. Among 86 AMS+ and 106 AMS− families, we detected seven distinct MLH1 exonic deletions in 7% of the AMS+ families and in 3% of the AMS− families. We found that the selective extinction of MMR protein in the tumours was highly predictive of an MMR rearrangement.

  • We conclude that MSH2 rearrangements are involved in at least 10% of the AMS+ families, which justifies screening for these in the routine diagnosis of HNPCC. The presence of MLH1 rearrangements should be considered in AMS+ HNPCC patients, when there is a selective loss of MLH1 expression in the tumours.

We have now integrated QMPSF into the routine diagnosis of HNPCC. We first analysed, as previously described,3 the 16 exons of MSH2 in 332 families, without point mutations within MSH2 and MLH1 (table 1). These families corresponded to 120 families fulfilling AMS criteria and 212 AMS- families. Immunohistochemical (IHC) staining of the tumours was performed in 19 AMS+ and 12 AMS− patients, and revealed a selective extinction of the MSH2 protein. Among the AMS+ patients without IHC information, an MSH2 genomic rearrangement was detected in 16% of the cases; the detection rate reached 58% in AMS+ patients, showing a selective loss of expression of MSH2 in their tumours. Among the AMS− patients negative for MSH2 and MLH1 mutations, we found an MSH2 genomic rearrangement in 4% of the cases, when IHC analysis had not been performed, and in 58% of the cases with IHC MSH2 extinction. QMPSF analysis of the 19 exons of MLH1 was then carried out in 192 families, corresponding to 86 AMS+ families (including 11 families in which IHC staining of the tumours had been undertaken and revealed a selective extinction of MLH1) and 106 AMS− families. Among the AMS+ patients, we found an MLH1 genomic deletion in 3% of the cases when IHC had not been performed, and in 36% of the cases when IHC had revealed a selective MLH1 extinction. Although the patients for whom IHC staining of the tumours revealed a selective loss of expression were low in number, our results indicate that the selective extinction of an MMR protein within a tumour is predictive, in HNPCC families, of the genetic alteration.

Table 1

 Frequency of MSH2 and MLHI exonic rearrangements detected by QMPSF in HNPCC families according to their status

We detected a total of 21 distinct MSH2 exonic rearrangements, including 19 deletions and two duplications, in 43 families; and seven exonic deletions of MLH1 in nine families (table 2). In families with an MSH2 rearrangement removing exon 1, QMPSF scanning of 50 Kb of genomic sequences upstream of the MSH2 transcription initiation site (table 3) revealed at least seven distinct 5′ breakpoints. Furthermore, this analysis showed that the recurrent exonic deletions that we detected (deletions of exon 1, exons 1–2, 1–6, 1–7) had been independently generated (tables 2 and 3), thus excluding a founder effect. The numerous breakpoints within the 5′ MSH2 region led us to screen for rearrangements affecting the promoter selectively, which would have escaped the initial QMPSF analysis of MSH2 restricted to the 16 exons. We therefore performed a specific QMPSF assay for the 4.4 kb promoter region, using the promoter amplicons indicated in table 3, and reanalysed 65 AMS+ families without MSH2/MLH1 point mutations or MSH2 exonic deletion. We identified in a single family a 1.7 kb partial deletion of the promoter removing the -1770–60 region.

Table 2

 Summary of MSH2 and MLH1 exonic rearrangements detected by QMPSF in HNPCC families

Table 3

 QMPSF scanning of the 5′ MSH2 region in HNPCC families with exonic deletions removing exon 1

This study confirms the following.

  • The frequency of MSH2 exonic rearrangements in AMS+ HNPCC families without detectable point mutations of MSH2 or MLH1 can be estimated to approximately 20% (table 1).

  • We identified, on the basis of exonic and promoter rearrangements, 30 distinct genomic alterations that demonstrate the remarkable heterogeneity of MSH2 rearrangements (tables 2 and 3).

  • The rearrangements that affect the MSH2 promoter selectively occur in less than 2% of AMS+ families.

  • MLH1 rearrangements are involved in 7% of AMS+ families without point mutations.

Two recent papers,8,9 have reported higher detection rates of MSH2 and MLH1 rearrangements, respectively, but these remarkable percentages are probably due to the differences between the populations analysed. The first study,8 performed on 24 AMS+ families without point mutations, using Southern Blot analysis, documented the detection of an MSH2 genomic deletion in 50% of the cases. This remarkable detection rate is probably explained by the fact that the exons 1–6 deletion, detected in seven families, were shown to be associated to a founder effect.8 The second study,9 based on the QMPSF analysis of 52 AMS+ families without point mutations, reported the detection of genomic deletions of MSH2 and MLH1 in 12% of the families, for each gene. In this work, the existence of a common haplotype, in four families harbouring a deletion of MLH1 exons 1–10, suggested a founder effect and may also have led to an overestimation of the relative contribution of MLH1 deletions in HNPCC.

In conclusion, we recommend that investigation for MSH2 rearrangements be included systematically in the routine diagnosis of HNPCC. The contribution of these alterations to HNPCC is higher than that of MSH6 mutations.10–12 Considering the lower frequency of MLH1 rearrangements, except in certain populations where they are associated with a founder effect,8,9 it is probably more efficient to search in HNPCC families only when IHC staining of the tumours has revealed a selective loss of MLH1 expression.


We are grateful to Mario Tosi for critical review of the manuscript, and to the clinicians who referred HNPCC families.


View Abstract


  • This work was supported by L′ Association pour la Recherche sur le Cancer, and La Ligue Nationale de Lutte contre le Cancer.

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.