Genetic and epigenetic alterations of genes/loci on human chromosome 3 were determined in four types of epithelial cancers, namely ovarian, colon, prostate and breast, using NotI-microarrays that contained 180 clones associated with genes/loci. We demonstrated inactivation of clusters of tumor suppressor genes (and putative tumor suppressor genes) on the 3p- and 3q-arms. We found that genes, such as ITGA9, LRRC3B, THRB, RBSP3 and SEMA3B was inactivated due to methylation of the gene promoters. Other genes, for example, NKIRAS1, PPM1M, PRICKLE2 and GPX1 were inactivated in tumors due to heterozygous and homozygous deletions. Genes GORASP1, GNAI2, NKIRAS1, GPX1, GPX3, PPM1M, PRICKLE2, SEMA3B, BHLHE40, BCL6 and ITGA9 showed both, genetic and/or epigenetic aberrations; their expression were downregulated in the studied epithelial tumors, as was shown by a quantitative PCR. For the putative tumor suppressor gene SEMA3B, we detected methylation of the promoter and intron CpG-islet in renal and lung tumors and also decreased gene expression in these tumor types. Moreover, we shown the tumor suppressor activity of SEMA3B both, in vitro and in vivo.
In result of an analysis of the NotI- microarray data, concerning 7 types of epithelial tumors we found, that 74 genes/loci showed the significant genetic and/or epigenetic aberrations. Twenty of them are common for several tumor types; changes in expression in 23 genes/loci are tumor-specific. The majority of them are found in samples of prostate cancer tissue.
Based on the NotI-microarray data analysis, we proposed the panels of markers for diagnostics and sub-division of ovarian and prostate cancers.
Also, we identified some features of tumor-stromal specific gene expression in more in prostate cancer, using tumor samples (~57 genes) and model cell lines (~140 genes). A comparative analysis of the relative expression of 65 genes in tumor cell lines LNCaP, DU145 and PC3 and in the conventionally normal cell line PNT2 revealed 35 differentially expressed genes, encoding proteins involved in the TP53, NF-kB, and WNT signaling pathways and in cell adhesion, invasiveness and metastasizing. Genes, such as IL1B, TAGLN, EFNA5, IL8, CXCL1 and CCNB2 were among them. Moreover, our study on the relative expression levels of 84 genes in comparison between LNCaP and PC3 cell lines allowed us to identify 36 differentially expressed genes, associated with apoptosis, adhesion, invasiveness and metastasizing, including MET, MMP1, MTA2, NME4, PLAU, TGFB1, SERPINB5 and SERPINE1. Expression of the seven differentially expressed genes, namely TAGLN, EFNA5, IL1B, PLAU, TGFB1, EPDR1 and FOS, correlated with the stage and a Gleason score in prostate adenocarcinomas, especially, when expression levels were compared with such in the conditionally-normal tissues and adenomas. The data, obtained on clinical samples, only partially coincided with the results on cell lines, which can be explained by heterogeneity of individual tumors and admixture of other cell types.
First, we monitored the presence of a TMPRSS2-ERG (EF194202.1) fusion transcript in adenocarcinomas of the different stage and Gleason score, in a group of Ukrainian patients. We assessed levels of relative expression of more than 50 genes from different functional groups in prostate tumors. More than 30 genes, associated with epithelial-to-mesenchymal cell transition (EMT), prostate cancer, tumor-associated fibroblasts, macrophages and immune-associated genes showed differential expression. Based on the cluster analysis of these gene groups, we could identify three molecular subtypes of prostate adenocarcinomas, that showed a high degree of correlation between groups of tumor-associated genes and stromal genes.
The obtained data on the relative expression allowed us to predict potential ranges of the sensitivity of tumor cells to the specific inhibitor drugs, using expression pattern of several genes, such as AR, PTEN, COX2, LDLR, HMGCR, FASN and CTLA4.
We developed an approach for the analysis of relative gene expression data, to create biomarker panels, based on the machine learning method and MDR analysis. Using this approach, we proposed the 12 biomarker panel for diagnostics and sub-typing of prostate adenocarcinomas. This panel, including genes, such as CDH2, CXCL12, CCL17, ESR1, FN1, IL1R1, HIF1A, HOTAIR, KRT18, PCA3, S100A4 and VDR, can be used to differentiate adenocarcinomas by the stage and Gleason score with the high statistical probability.
