Abstract
Polyploidy is a common outcome of chemotherapies, but there is conflicting evidence as to whether polyploidy is an adverse, benign or even favourable outcome. We show Aurora B kinase inhibitors efficiently promote polyploidy in many cell types, resulting in the cell cycle exit in RB and p53 functional cells, but hyper-polyploidy in cells with loss of RB and p53 function. These hyper-polyploid cells (>8n DNA content) are viable but have lost long-term proliferative potential in vitro and fail to form tumours in vivo. Investigation of mitosis in these cells revealed high numbers of centrosomes that were capable of supporting functional mitotic spindle poles, but these failed to progress to anaphase/telophase structures even when AURKB inhibitor was removed after 2-3 days. However, when AURKB inhibitor was removed after 1 day and cells had failed a single cytokinesis to become tetraploid, they retained colony forming ability and long-term proliferative potential. Mathematical modelling of the potential for polyploid cells to produce viable daughter cells demonstrated that cells with >8n DNA and >4 functional spindle poles approach zero probability of a viable daughter, supporting our experimental observations. These findings demonstrate that tetraploidy is tolerated by tumour cells, but higher ploidy states are incompatible with long-term proliferative potential. Model for AURKBi driven hyper-polyploid cells formation and fate. Aurora B inhibitor (AURKBi) treatment of RB+p53 defective cells efficiently promotes failed cell division. One failed cell division produces three possible outcomes, continued proliferation of the tetraploid daughter, cell death, or if AURKBi is continued, high polyploid states. Once cell have failed cell division >twice and have >8n DNA content they will continue to undergo rounds of endomitosis even in the absence of AURKBi to either become viable hyper-polyploid or die. The hyper-polyploid cells have no long-term proliferative potential.