Revista médica de Chile
versão impressa ISSN 0034-9887
ZAVALA G, Luis et al. Hierarchical clustering analysis to detect associations between clinical and pathological features of gastric tumors and hypermethylation of suppressor genes. Rev. méd. Chile [online]. 2007, vol.135, n.1, pp. 17-25. ISSN 0034-9887. doi: 10.4067/S0034-98872007000100003.
Background:Methylation is an inactivation mechanism for tumor suppressor genes, that can have important clinical implications. Aim: To analyze the methylation status of 11 tumor suppressor genes in pathological samples of diffuse gastric cancer. Material and methods: Eighty three patients with diffuse gastric cancer with information about survival and infection with Epstein Barr virus, were studied. DNA was extracted from pathological slides and the methylation status of genes p14, p15, p16, APC, p73, FHIT, E-caderin, SEMA3B, BRCA-1, MINT-2 y MGMT, was studied using sodium bisulphite modification and polymerase chain reaction. Results were grouped according to the methylation index or Hierarchical clustering (TIGR MultiExperiment Viewer). Results: Three genes had a high frequency of methylation (FHIT, BRCA1, APC), four had an intermediate frequency (p15, MGMT, p14, MINT2) and four had a low frequency (p16, p73, E-cadherin, SEMA3B). The methylation index had no association with clinical or pathological features of tumors or patients survival. Hierarchical clustering generated two clusters. One grouped clinical and pathological features with FHIT, BRCA1, and APC and the other grouped the other eight genes and Epstein Barr virus infection. Two significant associations were found, between APC and survival and p16/p14 and Epstein Barr virus infection. Conclusions: Hierarchical clustering is a tool that identifies associations between clinical and pathological features of tumors and methylation of tumor suppressor genes
Palavras-chave : Cluster analysis; Genes, supressor; Stomach neoplasms.