Novel cancer related genes and methylation/expression markers associated with breast cancer.

B129

Purpose: to identify and characterize cancer related genes undergoing altered epigenetic regulation in breast cancer and to develop novel methylation/expression markers of carcinogenesis. Methods. We have elaborated a novel modification of arbitrary-primed methylation sensitive PCR to identify CpG islands differentially methylated in breast cancer and adjacent morphologically intact tissues. Genomic locations of the identified fragments were investigated by direct sequencing and BLAST search. Methylation frequencies in 108 breast cancer samples were evaluated by methylation sensitive PCR and direct sequencing of bisulphite-treated DNA. Fine methylation maps for the identified genes' promoter CpG islands were constructed by means of direct bisulphite sequencing. Expression patterns were evaluated for every of the genes under study by real-time RT-PCR. Results. CpG islands identified as differentially methylated in breast cancer and control samples in this study appeared to belong to SEMA6B, BIN1, VCPIP1, LAMC3, KCNH2, RPSA, CACNG4 and PSMF1 genes. The screening technique has proved its unbiased nature as we have demonstrated the methylation of a part of the BIN1 tumor suppressor gene promoter CpG island which had not been reported previously in spite of the attempts undertaken. The frequencies of abnormal methylation in breast cancer varied from 2% in RPSA to 38% in SEMA6B. Methylation mapping revealed nonrandom distribution of methylated cytosines across the CpG islands with certain regions within the same island being frequently and some rarely/never methylated in breast cancer. Abnotmal methylation was not necessarily restricted to the core regions of the CpG islands. Significant expression distortions were observed for all the genes under study with a pronounced shift towards underexpression (44%-94%). Most pronounced expression abnormalities were noted for LAMC3 (94% underexpression). We have estimated the rates of 10-fold expression decrease of the identified genes, as it might be a significant feature of a parameter as a routine diagnostic marker. A 10-fold cutoff resulted in LAMC3 and SEMA6B underexpression values as the most informative markers (59% and 38%, respectively). Conclusions. We have identified a number of candidate cancer related genes abnormal methylation of which has not been reported previously. Expression studies demonstrate their frequent downregulation in breast cancer. Further research will reveal their roles in carcinogenesis and may promote the development of novel cancer molecular markers. The study was partially supported by Applied Biosystems, USA.