A salient feature of EZH2 overexpressing human invasive breast carcinomas is their high histological grade and poorly-differentiated cells with pleomorphic nuclei (8
). EZH2 overexpressing invasive carcinomas are largely ER negative and exhibit BRCA1 downregulation (8
). We discovered that EZH2 regulates the intracellular distribution of BRCA1 protein in benign breast cells and in ER negative breast cancer cells. To draw these conclusions we investigated the effect of EZH2 on the intracellular localization of BRCA1 protein utilizing independent and complementary gain- and loss-of function approaches. By measuring the nuclear and cytoplasmic expression of BRCA1 protein at different time points after release from G1/S cell cycle block, it was concluded that EZH2 overexpression in MCF10A induced nuclear export with cytoplasmic retention of BRCA1 protein. Consistent with this observation, while BRCA1 was mainly localized to the cytoplasm of CAL51 breast cancer cells, it was translocated to the nucleus upon lentiviral-mediated EZH2 KD.
The mechanisms governing the nuclear-cytoplasmic shuttling of BRCA1 protein are not fully elucidated but recent studies implicate the membrane serine/threonine protein kinase B, Akt (23
). A tumorigenic mechanism of Akt upon its phosphorylation is the induction of cytoplasmic localization of tumor suppressor proteins including p21 Cip1/WAF1 and FOXO3a (37
). The functional relationship between Akt and BRCA1 is complex and contextual (23
). The PI3K/Akt pathway promoted nuclear translocation of BRCA1 and reciprocally, BRCA1 deficiency was able to activate the PI3K/Akt signaling (26
). Akt–1 activation was shown to induce cytoplasmic retention of BRCA1 protein in breast cancer cells (24
). By using pharmacologic pathway inhibition and transient specific siRNA interference of Akt isoforms, we provide direct evidence that the effect of EZH2 on BRCA1 intracellular localization necessitates the activation of Akt-1, while Akt-2 and Akt-3 are dispensable for this function. Immunostaining of surgical samples highlights the relevance of our mechanistic studies to human breast cancer as EZH2 overexpression is significantly associated with increased pAkt-1 and with decreased pBRCA1 nuclear protein.
The stepwise progression from an aypical lesion to full-blown malignancy with metastatic capacity is associated with increases in genomic instability (40
). BRCA1 deficiency can cause tetraploidy and aneuploidy (41
). However, whether EZH2 regulates genomic stability is not known. Conditional EZH2 upregulation induced numerical chromosomal alterations in MCF10A cells as early as 72 hours after addition of doxycycline. Of note, over 50% of polyploid cells were near-tetraploid. These results are intriguing as several lines of evidence show that tetraploidy can be an initiator of chromosomal instability and tumorigenesis in vivo,
and has been detected in human tissues before aneuploidy occurs (42
). Our data on CAL51 breast cancer cells support the possible therapeutic role of EZH2 blockade in breast cancer, as EZH2 KD was sufficient to significantly decrease the percentage of tetraploid breast cancer cells. Thus, preventing or reverting tetraploidization through EZH2 inhibition may halt breast cancer development.
Although multiple mechanisms can lead to aneuploidy (40
), alterations in mitosis play an important role. Overexpression of Aurora kinases A and B are required for centrosome maturation, bipolar spindle assembly and mitotic entry, and their overexpression in human cells results in abnormal mitosis and aneuploidy (40
). We demonstrate that transient EZH2 overexpression in benign breast cells was sufficient to induce aberrant mitosis with extra centrosomes. The effect of EZH2 on mitosis was also evident in CAL51 breast cancer cells. While CAL51 controls exhibited aberrant mitosis with supernumerary centrosomes and multiple mitotic spindles, EZH2 KD abrogated these abnormalities. Mechanistically, EZH2 overexpression increased the messenger RNA and protein levels of Aurora kinase A and B and enhanced their kinase activity. These data implicate EZH2 in mitosis and in the regulation of Aurora kinase function in benign and in breast cancer cells.
Although Akt has been reported to play a role in mitosis and aneuploidy, the specific mechanisms have not been fully defined. Likewise, the specific role of each Akt isoform in the maintenance of genomic stability is unknown. Akt was shown to mediate abnormal checkpoint control and aneuploidy in PTEN-deficient cells by impairing CHK1 through phosphorylation, ubiquitination, and reduced nuclear localization (47
). Especially intriguing in light of our data are results from a recent study demonstrating that Akt-1 activation induced supernumerary centrosomes and genomic instability through cytoplasmic retention of BRCA1 in a hamster ovary cell line (25
). Here, we demonstrate that the effects of EZH2 overexpression on mitosis and genomic instability require specific activation of Akt-1. Interestingly, our data suggest a novel role for Akt-2 during mitosis unrelated to EZH2 expression. We observed that Akt-2 siRNA inhibition caused a 3-fold decrease in the number of cells undergoing mitosis in an EZH2-independent manner. Based on our data, we hypothesize that the blunting of mitoses may explain the absence of mitotic defects in Akt-2 KD cells after induction of EZH2 overexpression, as was observed with Akt-3 KD. The function of Akt-2 in mitosis warrants further study.
In conclusion, these data show a novel function of EZH2 in breast tumorigenesis: its ability to promote the nuclear export of BRCA1, induce aberrant mitosis and genomic instability. Our results enable us to pinpoint one mechanism by which EZH2 controls BRCA1 intracellular localization and genomic stability by activating Akt-1. In breast cancer cells, EZH2 downregulation induces nuclear localization of BRCA1, decreased mitotic aberrations and reverses tetraploidy. We propose that modulation of EZH2 expression may be a valid strategy to prevent or halt neoplastic progression in the breast.