In recent years, many groups have shown that centrosomes are amplified in the majority of breast cancers and other solid tumors (Lingle et al., PNAS, 1998; Pihan et al., Can. Res., 1998). In an effort to test the hypothesis that centrosome amplification in breast cancer leads to aneuploidy through the formation of multipolar mitoses, we analyzed aneuploidy and centrosome amplification in 20 breast tumors as compared to 7 normal breast tissues from women who underwent cosmetic breast reduction surgery. We performed fluorescence in situ hybridization (FISH) using centromeric probes to chromosomes 3, 7, and 17 to measure both ploidy and chromosomal instability, and performed immunofluorescence microscopy to measure centrosome size and number on all 27 tissues. We determined that the extent of centrosome amplification is linearly correlated with the level of chromosomal instability. Furthermore, we showed that centrosomes in the 3 diploid tumors present in the cohort were not amplified in size or number relative to the normal control tissues. We have initiated additional studies of centrosome amplification and mammary carcinogenesis in a rat model. In this model, aneuploid mammary tumors are induced by constant exposure to estrogen. The level of estrogen required to induce tumors in 100% of the rats within 4 months is approximately the same as the physiological level of estrogen associated with pregnancy. Upon exposure to estrogen, the mammary tissue goes through changes in histology that are similar to changes thought to be precursors to human breast cancer. An increase in aurora A protein and activity and the presence of amplified centrosomes are found in the earliest histologically distinct changes in the rat mammary tissues. Aurora A is a centrosome-associated kinase that is over-expressed in many human breast cancers. The centrosome-associated proteins centrin and gamma-tubulin are also overexpressed in the rat mammary tumors. Immunofluorescence of the rat tissues using gamma-tubulin as a centrosome marker demonstrated that centrosome were amplified in 30% of the cell in dysplastic lesions in just 3 months of estrogen exposure. Similarly, ductal carcinoma in situ lesions after 4-5 months of exposure also had amplified centrosomes in approximately 35% the cells. Approximately 90% of the cells in invasive tumors present after 6 months of estrogen exposure had centrosome amplification. These tumors are aneuploid and display chromosomal instability. These results indicate that estrogen exposure in the ACI rat model is sufficient to initiate mammary tumor formation and progression, and that this may occur through increased levels of aurora A expression, which in turn promote centrosome amplification and chromosomal instability. In addition to increased exposure to estrogen, other possible initiators of centrosome amplification include transient overexpression of cyclin D1 and non-lethal levels of gamma-irradiation. We propose that it is possible for amplified centrosomes to cause aneuploidy to arise in a diploid tissue or population of cells. This can happen when a tripolar mitosis results in 2, rather than 3, daughter cells. We have observed instances in which one daughter cell contains one spindle pole and an associated nuclear mass, and the other daughter cell results from the fusion of nuclear masses associated with the remaining 2 spindle poles. If the chromosomes were segregated evenly between the three spindle poles, then the first daughter cell would contain approximately 32 chromosomes and the fused daughter cell would contain 64 chromosomes. The first daughter cell would very like not survive, due to the lack of essential genes. However, the fused daughter cell, with 64 chromosomes, may possibly survive. The surviving cell would be aneuploid, with extra copies of some chromosomes, and would also have an extra centrosome. During the next cell division, this cell could yield aneuploid daughter cells, some of which may have a proliferative advantage. When centrosome amplification is accompanied by permissive lapses in cell cycle checkpoints, the potential for malignant growth is present. These lapses could result from specific genetic mutations and amplifications, epigenetic gene silencing, or from massive chromosomal instability caused by the centrosome amplification. Centrosome amplification, therefore, can serve to exacerbate and/or generate genetic instabilities associated with cancers.

[Proc Amer Assoc Cancer Res, Volume 46, 2005]