Contribution of Large Genomic Rearrangements in Italian Lynch Syndrome Patients: Characterization of a Novel Alu-Mediated Deletion

Lynch syndrome is associated with germ-line mutations in the DNA mismatch repair (MMR) genes, mainly MLH1 and MSH2. Most of the mutations reported in these genes to date are point mutations, small deletions, and insertions. Large genomic rearrangements in the MMR genes predisposing to Lynch syndrome also occur, but the frequency varies depending on the population studied on average from 5 to 20%. The aim of this study was to examine the contribution of large rearrangements in the MLH1 and MSH2 genes in a well-characterised series of 63 unrelated Southern Italian Lynch syndrome patients who were negative for pathogenic point mutations in the MLH1, MSH2, and MSH6 genes. We identified a large novel deletion in the MSH2 gene, including exon 6 in one of the patients analysed (1.6% frequency). This deletion was confirmed and localised by long-range PCR. The breakpoints of this rearrangement were characterised by sequencing. Further analysis of the breakpoints revealed that this rearrangement was a product of Alu-mediated recombination. Our findings identified a novel Alu-mediated rearrangement within MSH2 gene and showed that large deletions or duplications in MLH1 and MSH2 genes are low-frequency mutational events in Southern Italian patients with an inherited predisposition to colon cancer.


Introduction
Hereditary nonpolyposis colorectal cancer (HNPCC; also known as Lynch syndrome) is an autosomal dominant disorder characterised by colorectal cancer [1] that accounts for 3-5% of all colorectal cancers. Affected individuals have approximately 60-80% lifetime risk of developing colorectal cancer and women with Lynch syndrome have 54% risk of developing endometrial cancer [2]. It is associated with germline mutations in the DNA mismatch repair (MMR) genes, mainly MLH1 and MSH2 [3]. Mutations in MSH6 [4], PMS2 [5], and MLH3 [6] are less common. Recently, a germ-line point mutation in MSH3 was found to be associated with the Lynch syndrome phenotype [7]. Inactivation of the MMR complex manifests microsatellite instability (MSI), which is detected in tumour tissue [8].
e majority of mutations in the MMR genes so far identi�ed are missense, nonsense, or small insertions/deletions [http://www.insight-group.org/mutations mutations]. Depending on the population studied, large genomic rearrangements of the MMR genes constitute various proportions of the germ-line mutations that predispose to HNPCC [9][10][11]. Moreover, it seems that large genomic rearrangements occur more frequently in some populations than in others [11,12]. e relative incidence of genomic rearrangements among Lynch Syndrome families appears to vary from 5-20% [13]. A systematic study on genomic rearrangement in Lynch Syndrome showed that MLH1 and MSH2 are the most frequently targeted MMR genes for this type of mutation [14]. Furthermore, molecular characterisation of the breakpoints involved in large rearrangements within MLH1 and MSH2 genes showed that the majority are caused by homologous recombination between Alu repeats [15][16][17]. ese mutations are not usually detected by conventional methods of mutation analysis, such as denaturing high-performance liquid chromatography (DHPLC) and direct DNA sequencing, but they are detectable by a simple and robust technique such as the Multiplex Ligation-Probe Dependent Ampli�cation (MLPA) [18,19] assay.
As little is known about the frequency of large rearrangements in the MLH1 and MSH2 genes to Lynch syndrome in Italian population, the aim of our study was to assess the contribution of large genomic rearrangements in these two genes in a well-characterised series of 63 Southern Italian patients affected by Lynch Syndrome.  [20], without germ-line pathogenic point mutations in the MLH1, MSH2, or MSH6 genes, were recruited from several health centres in Campania (Southern Italy).

Materials and Methods
All patients received genetic counselling and gave their written informed consent to participate in this study.

Isolation of Genomic DNA.
Total genomic DNA was extracted from 4 mL peripheral blood lymphocytes using a Nucleon BACC2 Kit (Amersham Life Science) and from tumour tissues and surgical margins by standard methods [21]. (MLPA). MLPA was performed using the SALSA MLPA P003-B1 ML H1/MSH2 kit (MRC-Holland, e Netherlands) according to the manufacturer's instructions. Fragment analysis was conducted on an ABI Prism 3130 Genetic Analyser using GeneMapper soware (Applied Biosystems, Foster City, CA, USA). Migration of fragments was calculated by comparison to the GeneScan LIZ-500 size standard (Applied Biosystems, Foster City, CA, USA). Peak areas were then exported to a Microso spreadsheet (www.MLPA.com) and calculations were done according to the method described by Taylor and colleagues [22]. A 30-50% decrease in the peak area(s) indicated a deletion of the corresponding exon(s), while a 30-50% increase in the peak area(s) indicated a duplication of the corresponding exon(s). MLPA results were con�rmed in at least two independent experiments.
2.5. RNA Analysis of MSH2 Gene. RNA was extracted from 4 mL peripheral blood lymphocytes using a Trizol reagent by standard methods (Quiagen). cDNA was synthesised using SuperScript II RT (Invitrogen by Life Technologies) and ampli�ed with primers that produced a 598-bp fragment (2cFP 5 � -GGCTCTCCTCATCCAGATTG and 2cRP 5 � -AAGATCTGGGAATCGACGAA) spanning exons 4-7 of the messenger RNA. e PCR products were analysed on a 2% agarose gel and visualised by ethidium bromide staining.

Long-Range Polymerase Chain Reaction and Breakpoint
Analysis. 500 ng of genomic DNA was ampli�ed in a 50 Lreaction volume using 2.75 mM Mg 2+ , 500 M of each dNTP, 2 U of Expand Long Template PCR System (Expand Long Template Buffer 2; Roche Diagnostics), and 300 nM of each primer. Primers were designed between exon 5 and intron 7 of the MSH2 gene. is region was ampli�ed in four PCR fragments. e same forward oligonucleotide (5FP) was used in each reaction with a different reverse oligonucleotide, each approximately 1000 bp apart (Table 1). Cycling conditions were as follows: 94 ∘ C for 2 min, followed by 10 cycles consisting of 94 ∘ C for 10 sec, 60 ∘ C for 30 sec (−0.5 ∘ C/cycle) and 68 ∘ C for 15 min, followed by 25 cycles consisting of 94 ∘ C for 15 sec, 57 ∘ C for 30 sec, and 68 ∘ C for 15 min (+20 sec/cycle), and �nishing with one cycle at 68 ∘ C for 7 min.

Sequencing
Analysis. e PCR products were sequenced in both the forward and reverse directions using an ABI Prism 3100 Genetic Analyser (Applied Biosystems, Foster City, CA, USA).

In Silico
Analysis. e nucleotide sequences of the genomic MSH2 region (NG_007110.1) were analysed with the RepeatMasker program (http://repeatmasker.org/) using the default settings. Sequence comparisons in RepeatMasker were performed by the program cross_match [24].

Detection of Large Genomic Rearrangements in the MSH2
and MLH1 Genes by MLPA. MLPA analysis on 63 unrelated patients identi�ed a deletion in the MSH2 gene in one patient only (1.6%) (Figure 1). is deletion removed exon 6, which is located between the small intron 5 and the large intron 6. e exon 6 deletion was con�rmed at the RNA level by RT/PCR sequencing of a fragment with a lower molecular weight. e deletion was identi�ed in a 39-year-old man with a family history of colorectal cancer, who had developed a tubulovillous adenoma with small fragments of mucinous adenocarcinoma in the rectum, approximately 75 cm from the anus. e same deletion was also detected in his 33year-old brother. Although the brother was asymptomatic, endoscopy revealed an adenocarcinoma located proximal to the hepatic �exure ( Figure 2).

Microsatellite
Analysis. MSI analysis was performed on DNA extracted from tumour tissues (adenocarcinoma), and surgical margins of both patients (the proband and his brother) carrying the MSH2 exon 6 deletion. Both patients were found to have an MSI-H status, with instability at all markers analysed (data not shown).

Breakpoint Characterisation of the MSH2 Exon 6 Deletion.
e breakpoints of the exon 6 deletion within the MSH2 gene were characterised by analysing the intragenic regions between exon 5 and exon 7. is region was ampli�ed using region-speci�c oligonucleotides, as described in the Materials and Methods section. One forward primer located in exon 5, and different reverse primers starting in exon 7 were used. Abnormal fragment products of 3804, 2731, 1673 and 690 bp were ampli�ed from the patient�s DNA but not from the DNA of the healthy control using the primer pairs 5FP/6RPl, 5FP/6RPi, 5FP/6RPh, and 5FP/6RPg, respectively. No ampli�cation products were obtained using the primer pair 5FP/7 RP. As shown in Figure 3, sequence analysis of the 690bp ampli�cation product obtained using the primer pair 5FP/6RPg revealed the loss of a 9655-bp genomic region. e 5 � breakpoint is located in intron 5, in a strech of 11 nucleotides located 1,535-1,525 nt before the �rst nucleotide of exon 6. e 3 � breakpoint is located in intron 6, in an identical sequence of 11 nucleotides located 5,325-5,315 nt before the �rst nucleotide of exon 7. e exact breakpoints could not be ascertained because of the presence of an identical 11-bp sequences at both ends. is deletion c.942+(346-356)_1077-(5323-5313)del, alternatively NC_000002.11:g.47641903_47651558del, is named in accordance with the mutation nomenclature instructions provided by the HGVS (http://www.hgvs.org/); it creates a premature stop codon and the formation of a truncated protein.

In Silico Analysis.
Using the RepeatMasker program, the 5 � and 3 � breakpoints of the 9655-bp deletion were found to lie within the 26-bp core sequence of two Alu elements, which share 96% homology and differ by only one nucleotide. Both Alu elements belong to the AluSx subfamily and were 269 bp and 310 bp, respectively. Homology analysis of the AluSx sequences included in the deletion was performed using BLAST analysis ( Figure 4). e entire MSH2 gene was also analysed by RepeatMasker program, as already described in the literature [25], to verify the presence of repeat sequences. In this study, a total of 190 repeat sequences, including 106 Alu-type SINE sequences, 19 L1-type LINE sequences, 12 simple repeat sequences, and 12 LTR sequences were identi�ed, and their positions on the gene de�ned. Of these, 32 Alu-type SINE sequences, one L1-type LINE sequence, one LTR, and three simple repeat sequences were located in the genomic region between exons 5 and 7.

Discussion
e 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 identi�ed one large rearrangement in the MSH2 gene and none in the MLH1 gene. erefore, large rearrangements in the MLH1 and MSH2 genes occur at a low frequency in our patient cohort (1.6%).
e rearrangement in MSH2 identi�ed 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. e 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 ( Figure  2). DNA extracted from the tumour tissues of the two patients showed an MSI-H status, with instability at all markers analysed. e 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. e exact breakpoints could not be ascertained because of the presence of identical 11bp sequences at both ends; in fact using the RepeatMasker program, the breakpoints of this deletion were found to lie Ctrl2  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. erefore, 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.
e complete deletion of exon 6 has been previously reported to cause Lynch syndrome in a Dutch family [26], however the deletion was classi�ed as resulting from nonhomologous recombination, as the breakpoints did not fall in Alu sequences. e breakpoint characterised in this study therefore demonstrates that we have identi�ed a novel deletion.
e MSH2 and MLH1 genes are known to have a high density of Alu sequences, 34% and 21%, respectively, several large rearrangements in this gene have been reported [16,27]. 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 identi�ed. erefore, 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 and MSH2 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 [28,29], 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 [30,31]. erefore, 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 clari�ed.
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 unclassi�ed 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 genes [7,32]. Furthermore, it is also possible that the existence of other as yet undiscovered genes may confer susceptibility to colon cancer in Lynch syndrome families. e 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 identi�ed a number of loci that appear confer more increases in colon cancer risk [35,36]. Further studies are needed to better identify the underlying genetic risk factors associated with disease in these families.

Conclusions
is paper is the �rst signi�cant study on contribution of large MLH1 and MSH2 genomic rearrangements in Southern Italian Lynch syndrome patients, negative for point mutation in MMR genes. Our results enlarge the spectrum of large rearrangements in MSH2 genes and at the same time indicate that these genomic rearrangements seem to be a less frequent mutational event in our population. Nonetheless, we believe that the detection of large rearrangements in the MLH1 and MSH2 genes should be included in the routine testing for Lynch syndrome, especially considering the simplicity of the MLPA assay.