The Lynch syndrome, caused primarily by germ-line point mutations within MMR genes, is also associated with large rearrangements that account for 5–20% of all mutations. Here, we report the results of our screening for large rearrangements in the MLH1 and MSH2 genes in a cohort of 63 Southern Italian patients who were negative for pathogenic point mutations in the MLH1, MSH2, and MSH6 genes. We identified one large rearrangement in the MSH2 gene and none in the MLH1 gene. Therefore, large rearrangements in the MLH1 and MSH2 genes occur at a low frequency in our patient cohort (1.6%).
The rearrangement in MSH2
identified in this study caused a large deletion that removed exon 6 and was detected in two patients from the same family who met the Amsterdam-1 criteria. The two affected brothers presented colorectal cancer with early-onset, before 40 years of age. Other family members were also affected (not tested in this study) and presented with the same phenotype (). DNA extracted from the tumour tissues of the two patients showed an MSI-H status, with instability at all markers analysed. The novel deletion is 9,655
bp long and extends from a region 346
bp downstream of exon 5 to 5323
bp upstream of exon 7. The exact breakpoints could not be ascertained because of the presence of identical 11-bp sequences at both ends; in fact using the RepeatMasker program, the breakpoints of this deletion were found to lie within the 26-bp core of two AluSx sequences that share 96% homology. As these two AluSx sequences were found to differ by only one nucleotide, it is possible that recombination could occur at this sequence. Therefore, we speculated that the MSH2
rearrangement is most likely an Alu-Alu homologous recombination event that deletes approximately 9.5 kb of the MSH2
genomic region encompassing exon 6.
The complete deletion of exon 6 has been previously reported to cause Lynch syndrome in a Dutch family [26
], however the deletion was classified as resulting from nonhomologous recombination, as the breakpoints did not fall in Alu sequences. The breakpoint characterised in this study therefore demonstrates that we have identified a novel deletion.
genes are known to have a high density of Alu sequences, 34% and 21%, respectively, several large rearrangements in this gene have been reported [16
]. However, given the high frequency with which these repetitive sequences occur within these two genes, we would expect the overall incidence of large rearrangements in our cohort to be much higher than that identified. Therefore, it is reasonable to hypothesise that Alu-mediated homologous recombination could also cause intragenic rearrangements, such as translocations or inversions, that are not always detectable with the MLPA assay used in this study. MLPA is used for detecting copy number changes in genomic DNA and can only detect large deletions or duplications. Inability to detect intragenic rearrangements could in part explain the low frequency of these molecular alterations in our cohort. Moreover, it is noteworthy that an exceptionally low frequency of large rearrangements in the MLH1
genes (<1.5%) was also reported in a study of the Spanish population [11
]; indeed, due to historical inheritage Spaniards share a common genetic pool with the Southern Italian population. In contrast, other studies performed especially on populations of Northern-Europe (including Northern Italy population) have reported an increasingly higher frequency of large rearrangements in these two genes [28
], with a recent study of Slovak HNPCC [12
] reporting a frequency of 25%. Moreover, differences in the frequency of large rearrangements are also seen in other Alu-rich genes that are responsible for hereditary diseases, such as BRCA1,
and BRCA2, STK11
, depending on the population analysed [30
]. Therefore, based on these informations the Alu sequences may be regarded as passive elements that serve as favourable substrates for recombination and the molecular mechanism that promotes recombination events remains to be clarified.
Beyond possible explanations about the low frequency of large rearrangements in our population, it should be highlighted that the majority of patients with Lynch syndrome tested in this study do not have a mutation in the MMR
genes most frequently mutated. It is also important to emphasize that our families were selected on the basis of the Amsterdam clinical criteria and MSI-H, thus there is good evidence that all affected have a strong genetic component to early development of cancer. We therefore suggest that some undiscovered genetic mechanism in Lynch syndrome patients is yet to be investigated. Recently, it has been shown that unclassified genetic variants in MMR
genes can behave as low-risk alleles that contribute to the risk of colon cancer in Lynch syndrome families when interacting together or with other low-risk alleles in other MMR
]. Furthermore, it is also possible that the existence of other as yet undiscovered genes may confer susceptibility to colon cancer in Lynch syndrome families. The EPCAM
gene in addition to MMR
genes has already been associated HNPCC phenotype [33
] as well as MYH
in addition to APC
gene has been associated FAP phenotype [34
]. Recently, association studies have identified a number of loci that appear confer more increases in colon cancer risk [35
]. Further studies are needed to better identify the underlying genetic risk factors associated with disease in these families.