Mutation spectrum in South American Lynch syndrome families

Background Genetic counselling and testing for Lynch syndrome have recently been introduced in several South American countries, though yet not available in the public health care system. Methods We compiled data from publications and hereditary cancer registries to characterize the Lynch syndrome mutation spectrum in South America. In total, data from 267 families that fulfilled the Amsterdam criteria and/or the Bethesda guidelines from Argentina, Brazil, Chile, Colombia and Uruguay were included. Results Disease-predisposing mutations were identified in 37% of the families and affected MLH1 in 60% and MSH2 in 40%. Half of the mutations have not previously been reported and potential founder effects were identified in Brazil and in Colombia. Conclusion The South American Lynch syndrome mutation spectrum includes multiple new mutations, identifies potential founder effects and is useful for future development of genetic testing in this continent.


Background
Since the initial reports on disease-predisposing mutations in the mismatch-repair (MMR) genes MLH1 [MIM:120436], MSH2 [MIM:609309] and MSH6 [MIM:600678] in the early 1990'ies, a large number of studies have contributed to the establishment of the molecular map of Lynch syndrome with over 3,072 unique genetic MMR gene variants identified. These data are predominantly based on studies from North America, Europe and Asia. The mutations affect MLH1 in 42%, MSH2 in 33%, MSH6 in 18% and PMS2 in 8% [1]. Nonsense mutations, frameshift mutations and missense mutations predominate, whereas large genomic rearrangements and splice-site variants constitute <10% of the alterations [1].
The South American population is ethnically mixed from American Indian and European ancestors. In Uruguay and Argentina, European ancestry predominates. In Brazil, significant African and American Indians roots apply. In Chile, Colombia, Peru and Bolivia, Spanish colonist and American Indian ancestry influence the populations [2,3]. Mutation screening in South American families suspected of Lynch syndrome has identified disease-predisposing germline mutations in MLH1 and MSH2 in 16-45% of families that fulfill the Amsterdam criteria and/or the Bethesda guidelines [2][3][4][5][6][7]. Hereditary colorectal cancer registries have been established in Argentina, Brazil, Uruguay and Chile with the aim to collect and share data on the MMR gene mutation spectrum, identify potential founder mutations, interpret the role of unclassified genetic variants and to study cancer risks in the South American Lynch syndrome population. We used published data and unpublished register data to describe the mutation spectrum in South American Lynch syndrome families.

Ethics statement
All patients provided an informed consent for inclusion into the South American registers during genetic counseling sessions and is in compliance with the Helsinki Declaration.

Patient selection
Families that fulfilled the Amsterdam criteria [8,9] and/ or the Bethesda guidelines [10] were selected from the hereditary cancer registries at the Hospital Italiano (Buenos Aires, Argentina), the Hospital de las Fuerzas Armadas (Montevideo, Uruguay), the Clinica Los Condes (Santiago, Chile), the Barretos Cancer Hospital (Barretos, Brazil) and from two databases in Colombia and in Southeastern Brazil ( Figure 1, Table 1) [2,3,5,11]. Patients were informed about their inclusion into the registries, which generally contained data on family history, age at onset and results of genetic testing.

Disease-predisposing mutations
Methods to assess MMR status, e.g. microsatellite instability analysis and MMR protein staining, varied between the countries and were excluded from the present study since these data were incomplete. Molecular diagnosis was generally based on direct sequencing of MLH1 and MSH2. Chilean and Brazilian families were also analyzed for large genomic rearrangements using the multiplex ligation-dependent probe amplification (MLPA) method (performed using the SALSA kit P003, MRC-Holland, Amsterdam, Netherlands).

Mutation nomenclature
Mutation nomenclature was in accordance with the Human Genome Variation Society (HGVS) guidelines [12]. Mutations in the MLH1 or MSH2 genes were considered deleterious if they: a) were classified as pathogenic in LOVD database; b) introduced a premature stop codon in the protein sequence (nonsense or frameshift mutation); c) occurred at donor or acceptor splice sites; or d) represented whole-exon deletions or duplications. All identified mutations were correlated to the MMR Gene Unclassified Variants Database (www.mmruv.info), the Mismatch Repair Genes Variant Database (http://www.med.mun.

Variants of uncertain significance
To establish the pathogenicity of variants of uncertain significance, web-based programs, i.e. Polyphen, MAPP-MMR, SIFT, P-mut and PON-MMR were applied to predict the effect of an amino acid substitution based on protein structural change and/or evolutionary conservation [13][14][15][16][17].

Statistical analysis
The statistical analyses were performed using the statistical software package IBM SPSS Statistics 20 (SPSS, Chicago, IL, USA).

Results
In total, 110 families harbored MMR gene variants, of which 99 were classified as Lynch syndrome predisposing and 11 were regarded as variants of uncertain significance. Mutations in MLH1 and MSH2 were identified in 37% (range 27-65% in the different countries/registries) of the families that fulfilled the Amsterdam criteria and/ or Bethesda guidelines (Table 1). When the Amsterdam criteria were considered, the mutation detection rate was 55% (81/147), whereas 15% families that fulfilled the Bethesda guidelines had disease-predisposing mutations.
The mean age at diagnosis was 35-44 years for colorectal cancer and 41-49 years for endometrial cancer in the different registries (Table 1). Pedigree information was available from 54 families and showed that among the Lynch syndrome-associated tumors, 65% were colorectal cancers (of which 43% were located in the right side of the colon), 22% endometrial cancers and 13% constituted other Lynch syndrome-associated cancer types.
Of the 99 disease-predisposing MMR gene mutations, 60% affected MLH1 and 40% affected MSH2 (Table 2). Frameshift and nonsense mutations were the most common alterations (36% and 31%, respectively), followed by splice site mutations (13%), missense mutations (12%) and large deletions (8%) (Figure 2a). Though the mutations were spread over the genes, hot-spot regions included exons 16 and 18 in MLH1 (13% of the mutations  each) and exon 13 in MSH2 (24% of the mutations each) (Figure 2b). In total, 10 mutations identified in at least two South American families were classified as recurrent. Among these, the MSH2 c.2152C > T identified in Brazil represents an internationally hot-spot. Three founder mutations were identified in five South American families. The MLH1 c.545 + 3A > G and the MSH2 c.942 + 3A > T have been identified as founder mutations in Italy and in Newfoundland and were also identified in Brazilian families [3]. The MLH1 c.1039-8T_1558 + 896Tdup has been suggested to represent a founder mutation in Colombia [2,11]. Mutations that were unique and herein first reported in more than one family included the MLH1 c.1853delAinsTTCTT in Brazil, the MLH1 c.1856delG in Colombia and the MSH2 c.25252_2526delAG in Brazil (Table 2) (Figure 3).

Discussion
In South America, disease-predisposing mutations linked to Lynch syndrome have been identified in 99 families, which corresponds to 37% of the families that fulfilled the Amsterdam criteria and/or Bethesda guidelines and underwent genetic testing. The mutation rate is high compared to prevalence rates of 28% for MLH1 and 18% for MSH2 in the Asian population, 31% and 20% in a multi-ethnic American population and 26% and 19% in European/Australian populations [18]. The mutation spectrum is predominated by MLH1 (60%) and MSH2 (40%) mutations [3,[19][20][21][22], but the seemingly larger contribution than the 42% and 33% reported in the InSIGHT database could reflect failure to test for MSH6 and PMS2 mutations in most South American studies [1]. Referral bias in populations that have more recently been screened for mutations represents a potential limitation, but the strong contribution from MLH1 and MSH2 could also reflect population structure [2,4,5,7]. Frameshift mutations and nonsense mutations were the most common types of mutations, which are in agreement with findings from other populations [1,[23][24][25][26], with hotspots in exons 16 and 18 of MLH1 and in exon 13 of MSH2 (Figure 2b). Exon 16 and 18 in MLH1 has been identified as a genetic hot spot also in other populations with 26% of the MLH1 mutations reported herein [3,18]. The frequent mutations in MSH2 exon 13 may be linked to the c.2152C>, which was first identified in Portuguese Lynch syndrome families. This alteration accounted for 35% (6/17) of the MSH2 mutations in the Brazilian population, which is in line with the Portuguese migration to Brazil [3,27].
Founder mutations have been identified in several populations where they significantly contribute to disease predisposition and thereby allow for directed genetic testing [28]. Two of the mutations identified in South American Lynch syndrome families have been suggested to constitute potential founder mutations in other populations, e.g. the Italian MLH1 c.545 + 3A > G and the Newfoundland MSH2 c.942 + 3A > T [3]. The Spanish founder mutations MLH1 c.306 + 5G > A and c.1865 T > A and MSH2 c.2635-3 T > C; c2635-5C > T; c.2063 T > G were, however, not observed in South American Lynch syndrome families [27][28][29][30]. In Colombia, the MSH2 c.1039-8T_1558 + 896Tdup was suggested to represent a founder mutation [2,11]. The Colombian population has a mixed ancestry with a strong influence from Spanish colonists and thereby genetically differs from previously studied populations [2,6].  .05 then the aminoacid (AA) substitution is predicted to affect protein function, if PolyPhen score >0.5 then the AA substitution is predicted to affect protein function, if MAPP-MMR score >4.55 then the AA substitution is predicted to affect protein function, If P-mut score > 0.5, the AA substitution is classified as pathological, if PON-MMR score >0.7615, the AA substitution is classified as pathogenic.

Conclusions
In conclusion, disease-predisposing mutations in MLH1 and MSH2 have been identified in a relatively large proportion of the South American families suspected of Lynch syndrome that have been tested. Genetic hot-spot regions, internationally recognized founder mutations and potential South American founder mutation have been recognized, which is of relevance for genetic counseling and testing that are increasingly available in South America.