Faithful segregation of genetic material during mitosis is a vital cellular function that ensures the accurate passage of critical information from one generation to the next. Errors in this process can be highly deleterious, often times leading to aneuploidy- a hallmark of cancer and the leading genetic cause of birth defects and miscarriages. Given the potentially harmful nature of cells resulting from defects in mitosis, a majority of these cells are eliminated via apoptosis to prevent further propagation of this damage. However, this is not always the case as sometimes these cells evade cell death and continue to proliferate. Although the link between mitotic defects and activation of the cell death machinery is relatively well established, the underlying mechanism still remains unclear. The work presented here takes a closer look, in both the model system Caenorhabditis elegans and human tissue culture cells, at the direct effect of a variety of mitotic perturbations in an effort to better understand how mitotic defects can signal to cell death. In the adult C. elegans germline, which is composed of both mitotic and meiotic cells, nuclei that result from a failed mitosis acquire an excess amount of double-strand breaks, activate the DNA damage checkpoint and are unable to successfully complete meiotic recombination. These nuclei are preferentially eliminated by apoptosis, a response that is dependent on the initiation of recombination, suggesting that the deliberate double strand breaks that are required for recombination survey and mark for elimination nuclei with the improper complement of chromosomes. In human tissue culture cells, perturbations that prevent proper exit from mitosis while allowing for perfect, or near perfect, alignment of chromosomes result in a distinct phenotype called chromosome scattering. Cells with scattered chromosomes show uncoordinated loss of chromatid cohesion and subsequent spindle defects, yet still maintain the hallmarks of an active mitosis. Most notably, cells with scattered chromosomes show an increased susceptibility to apoptosis. Overall the results of these studies reveal two distinct mechanisms by which mitotic defects signal to cell death helping to shed light on the driving force behind this phenomenon

Centromeres are defined epigenetically in the majority of eukaryotes by the presence of chromatin containing the centromeric histone H3 variant CENP-A. Most species have a single gene encoding a centromeric histone variant whereas C. elegans has two: HCP-3 (also known as CeCENP-A) and CPAR-1. Prior RNAi replacement experiments showed that HCP-3 is the functionally dominant isoform, consistent with CPAR-1 not being detectable in embryos. GFP::CPAR-1 is loaded onto meiotic chromosomes in diakinesis and is enriched on bivalents until meiosis I. Here we show that GFP::CPAR-1 signal loss from chromosomes precisely coincides with homolog segregation during anaphase I. This loss of GFP::CPAR-1 signal reflects proteolytic cleavage between GFP and the histone fold of CPAR-1, as CPAR-1::GFP, in which GFP is fused to the C-terminus of CPAR-1, does not exhibit any loss of GFP signal. A focused candidate screen implicated separase, the protease that initiates anaphase by cleaving the kleisin subunit of cohesin, in this cleavage reaction. Examination of the N-terminal tail sequence of CPAR-1 revealed a putative separase cleavage site and mutation of the signature residues in this site eliminated the cleavage reaction, as visualized by retention of GFP::CPAR-1 signal on separating homologous chromosomes at the metaphase-anaphase transition of meiosis I. Neither cleaved nor uncleavable CPAR-1 were centromere-localized in mitosis and instead localized throughout chromatin, indicating that centromere activity has not been retained in CPAR-1. Although the functions of CPAR-1 and of its separase-dependent cleavage remain to be elucidated, this effort reveals a new substrate of separase and provides an in vivo biosensor to monitor separase activity at the onset of meiosis I anaphase.

We identified a synthetic lethality between PLK1 silencing and the expression of an oncogenic Epidermal Growth Factor Receptor, EGFRvIII. PLK1 promoted homologous recombination (HR), mitigating EGFRvIII induced oncogenic stress resulting from DNA damage accumulation. Accordingly, PLK1 inhibition enhanced the cytotoxic effects of the DNA damaging agent, temozolomide (TMZ). This effect was significantly more pronounced in an Ink4a/Arf(-/-) EGFRvIII glioblastoma model relative to an Ink4a/Arf(-/-) PDGF-β model. The tumoricidal and TMZ-sensitizing effects of BI2536 were uniformly observed across Ink4a/Arf(-/-) EGFRvIII glioblastoma clones that acquired independent resistance mechanisms to EGFR inhibitors, suggesting these resistant clones retain oncogenic stress that required PLK1 compensation. Although BI2536 significantly augmented the anti-neoplastic effect of EGFR inhibitors in the Ink4a/Arf(-/-) EGFRvIII model, durable response was not achieved until TMZ was added. Our results suggest that optimal therapeutic effect against glioblastomas requires a "multi-orthogonal" combination tailored to the molecular physiology associated with the target cancer genome.