Identification and validation of novel amplification target genes in human breast cancer

Gene amplification is a major mechanism for overexpression of potential oncogenes in cancer. Several amplifications have already been described in breast cancer. In order to find new amplified regions, we screened 30 human breast tumors for gene amplifications using the Affymetrix SNP 10k 2.0 microarrays. For this purpose, we developed a new analysis procedure leading to an increased signalto-noise ratio which allowed us the discovery of new small single gene amplicons (<Mb). Two of them, the ESR1 gene on 6q25 and the NFIB gene on 9p24, were further investigated. Fluorescence insitu hybridization of these two genes was performed on a TMA comprising more than 2200 breast cancer samples. NFIB amplification was found in 5% of all breast cancers analyzed, but with an increased amplification rate in medullary carcinoma (19%). NFIB amplified breast cancers showed a higher Ki67 proliferation index. Functional analysis with RNA interference of the NFIB gene in three tumor cell lines suggested a proliferation supporting role of the NFIB protein in breast cancer. The ESR1 gene was amplified in 21% of the breast cancer samples analyzed. The ESR1 amplified patients defined a subgroup of ER positive breast cancer patients with prolonged survival, suggesting that patients with ESR1 gene amplification optimally benefit from hormonal therapy. Since this amplification was also found in histological benign and precancerous breast lesion, we suggest that ESR1 gene amplification is an early mechanism in breast cancer development. Furthermore, the use of potentially heterogeneous markers, like the Ki67 proliferation index on a breast cancer TMA was investigated. All expected associations between Ki67 and other previously analyzed molecular markers could be reproduced with high statistical significance using a breast cancer TMA containing only one tissue sample per tumor. This leads to the conclusion that associations with cell proliferation can be reliably analyzed in a TMA format.