Caracterización de nuevos perfiles moleculares en carcinoma de endometrio esporádico

Abstract

[ES] Endometrial carcinoma is the most frequent gynaecological tumour in developed countries. The origin is sporadic in 95% of them and its main risk factor is hormone exposure. It has been proposed several classifications depending on the followed criteria to characterize sporadic endometrial carcinomas. Bearing in mind clinical-pathological characteristics, Bokhman et al., in 1983, proposed a dualistic model in which endometrial carcinomas were classified in two groups: type 1 or endometrioid carcinomas (with a good prognosis and estrogens-dependent) and type 2 or non-endometrioid carcinomas (with a poor prognosis and non-estrogens-dependent). Later, Vogelstein in 1988, developed a theory for colorectal cancer, embraced and adapted for endometrial carcinoma, which said that tumours are a consequence of an accumulation of genetic and epigenetic alterations. According to this fact, it has been observed that type 1 and type 2 endometrial carcinomas have specific molecular profiles. Type 1 carcinomas are mainly associated with MSI and mutations in PTEN, PIK3CA, KRAS, CTNNB1, FGFR2 and MMR genes and type 2 carcinomas are associated with mutations in TP53, CDKN2A, and CDH1 genes, loss of expression of HER2/ERBB2 and EGFR/ERBB1 proteins and loss of heterozigosity (LOH) in different chromosomes. Currently, whole exome sequencing studies are being carried out in order to look for new target genes in different types of tumours. Two of them are ARID1A and PPP2R1A genes that are highly mutated in endometrioid and non-endometrioid carcinomas respectively. On the other hand, telomere length variations have been associated with a susceptibility of developing different kind of tumours. However, it has not been yet clarified if the telomere shortening is a cause or a consequence of tumour development and the exact role of telomerase. Moreover, new criteria are being established with the aim of elaborate a more accurate classification that would help to better identify, characterize and treat the different types of endometrial carcinomas. The aim of the present work has been the analysis of 14 genes implicated in tumorigenesis (including PPP2R1A and ARID1A genes); MSI; LOH; as well as the methylation of the MMR genes promoter and HDACs expression; the telomere length and TERT-1327C>T and TERC-63G>A polymorphisms analysis in a cohort of 86 sporadic endometrial carcinomas and 23 blood samples. We have used a wide variety of techniques for the realization of our study. We have extracted DNA from tumour and blood samples, and RNA and total proteins from tumour samples. The analysis of different genes has been carried out by PCR, CSGE-Heteroduplex and Sanger sequencing using several databases in order to check if the mutations found had been previously described. For the characterization of the new mutations, we have performed RT-PCR and Western blot assays (as well as for HDAC protein expression analysis). Several prediction programs have helped us to determine the mutations pathogenicity. We also have studied LOH by qPCR and RFLP; gross alterations in MMR genes by MLPA; MMR promotermethylation by MS-MLPA; and telomere length and telomerase polymorphisms by real time PCR. Statistical analysis has been carried out with SPSS v18.0, GenEx 5.3.6 Enterprise, and MULTBiplot programs. We have found 213 mutations: 99 described mutations and other 114 non-previously described mutations that have been characterized in our work. The most mutated genes have been PTEN and ARID1A. MSI has been carried out when we have obtained DNA from both tumour and blood samples of the same patient observing that it is a frequent event in our patients. In contrast with Lynch syndrome, D17S250 was the most altered marked in our population. On the other hand, we have observed that the gross alterations and the point mutations do not explain the MSI and loss of expression of MMR genes. The most methylated gene has been hMLH1, but methylation of MMR genes has not been consistent with either the lack of MMR proteins expression and MSI. The HDAC2 protein has been the histone deacetylase which expression has been absent in a highest number of studied tumours. Both methylation and HDAC2 protein expression patterns have differed depending on the type of tumour. The telomere length has not shown any relation with neither type nor grade of tumour and C and G alleles of TERT-1327C>T and TERC-63G>A polymorphisms were only slightly related to mixed and grade 3 endometrioid carcinomas respectively (the groups with shortest telomeres). In our work, PPP2R1A gene has shown mutations mainly associated with serous carcinomas. We have described a new probably pathogenic mutation in the exon 2 of the gene, presented in a grade 1 endometrioid carcinoma, and a pole of polymorphisms situated in the promoter region implicated in the binding of several transcription factors. When we have analyzed PPP2R1A protein expression, we have observed a new pattern of protein expression at 110 kDa, maybe related to posttranslational modifications. On the other hand, we have observed that ARID1A gene appears commonly mutated in endometrioid and non-endometrioid carcinomas with a high number of nosense and frameshift mutations. We have described 32 new mutations including an inframe alteration that causes a change in the mRNA splicing. According to our results, the most adequate classification among those that have been proposed until now, is the classification that bears in mind the characteristics of the different histological subtypes. The clinical-pathological criteria are not accurate and they share a high identity with the classification by histological grades. We have observed that grade 1 and grade 2 endometrioid carcinomas could be grouped in a low-grade endometrioid cluster showing very similar molecular characteristics. At the same time the molecular profile of low-grade endometrioid carcinomas differs from grade 3 endometrioid carcinomas genetic profile. Grade 3 endometrioid carcinomas share characteristics with low-grade endometrioid and serous carcinomas and because of that, it is not possible to group them with any other histological subtype. Moreover, mixed carcinomas and carcinosarcomas show a molecular profile that depends on the type of their components suggesting the importance of making histological studies of this kind of tumours before their molecular analysis.