Mitotic slippage: an old tale with a new twist

ABSTRACT

Targeting the mitotic machinery using anti-mitotic drugs for elimination of cancer cells is a century-old concept, which continues to be routinely used as a first line of treatment in the clinic. However, patient response remains unpredictable and drug resistance limits effectiveness of these drugs. Cancer cells exit from drug-induced mitotic arrest (mitotic slippage) to avoid subsequent cell death which is thought to be a major mechanism contributing to this resistance. The tumor cells that acquire resistance to anti-mitotic drugs have chromosomal instability (CIN) and are often aneuploid. In this review, we outline the key mechanisms involved in dictating the cell fate during perturbed mitosis and how these processes impede the efficacy of anti-mitotic therapies. Further, we emphasize the recent work from our laboratory, which highlights the functional role of CEP55 in protecting aneuploid cells from death. We also discuss the rationale of targeting CEP55 in vivo, which could prove to be a novel and effective therapeutic strategy for sensitizing cells to microtubule inhibitors and might offer significantly improved patient outcome.

Abbreviations: APC/C: Anaphase-Promoting Complex/Cyclosome; BAD: BCL2‐Associated agonist of cell Death; BAK1: BCL2 Antagonist Kinase1; BAX: BCL2‐Associated X; BCL2: B-cell Chronic Lymphocytic Leukaemia (CLL)/Lymphoma 2; BH: BCL2 Homology Domain; BID: BH3‐Interacting domain Death agonist; BIM: BCL2‐Interacting Mediator of cell death; BUB: Budding Uninhibited by Benzimidazoles; CDC: Cell Division Cycle; CDH1: Cadherin-1; CDK1: Cyclin-Dependent Kinase 1; CEP55: Centrosomal Protein (55 KDa): CIN: Chromosomal Instability; CTA: Cancer Testis Antigen; EGR1: Early Growth Response protein 1; ERK: Extracellular Signal-Regulated Kinase; ESCRT: Endosomal Sorting Complexes Required for Transport; GIN: Genomic Instability; MAD2: Mitotic Arrest Deficient 2; MCL1: Myeloid Cell Leukemia sequence 1; MPS1: Monopolar Spindle 1 Kinase; MYT1: MYelin Transcription factor 1; PLK1: Polo Like Kinase 1; PUMA: p53-Upregulated Mediator of Apoptosis; SAC: Spindle Assembly Checkpoint; TAA: Tumor-Associated Antigen

Introduction

For over a century, chemotherapy has been used to treat cancer. Today, it is a routine first-line treatment against multiple malignancies. Generally, anticancer chemotherapies induce genotoxic damage and activate molecular factors that modulate cell cycle checkpoints leading to cell death and tumor regression. More specifically, antimitotic chemotherapies affect cells in mitosis. This either involves chemotherapeutic agents that target microtubule dynamics – hence they are known as microtubule poisons or spindle poisons – or it involves small molecule inhibitors which induce an anti-proliferative effect as a consequence of perturbed microtubule dynamics, i.e., mitotic blockers or mitotic targeted therapies. Mechanistically, these inhibitors function by interfering with the regular mitotic processes, including spindle formation, faithful chromosome segregation and/or mitotic exit [1]. As such, they sustain activation of the spindle assembly checkpoint (SAC), a mechanism that in normal cells arrests mitotic progression at metaphase to ensure proper chromosome alignment at the metaphase plate prior to controlled chromosome segregation. In turn, inhibitor-sustained SAC activation leads to prolonged mitotic arrest and ultimately causes cell death, either in mitosis, a phenomenon referred to as mitotic catastrophe, or in G1 phase of the next cell cycle. Moreover, the cellular response to antimitotic drug is related to cell type and dosage of drug used and is not dictated by duration of mitotic arrest. In this review, we briefly highlight the key modulators influencing cell fate of cancer cells during perturbed mitosis imposed by antimitotic chemotherapies and then emphasize the latest strategies that our laboratory has implemented in order to overcome mitotic slippage for better clinical outcome.

The challenge

Despite their efficacy, the long-term utility of antimitotic drugs in the clinic is problematic, because their use is accompanied by moderate to severe side effects. In addition, their usage frequently induces the malignant tumors’ becoming resistant to chemotherapeutic agents, leading to disease relapse. Cancer cells often acquire multiple resistance mechanisms that enable them to bypass mitotic arrest and prematurely exit mitosis, a phenomenon known as mitotic slippage [2,3]. This process, also known as endomitosis, may be followed by mitotic exit without undergoing cytokinesis, thus resulting in tetraploidy [4]. Alternatively, in cancer cells, SAC signaling may be impaired or hyper-activated, either of which results in unequal partitioning of sister chromatids and bipolar cell division. Yet another outcome can be the formation of multipolar mitotic spindles. Irrespective of the mechanism, these routes lead to genomic instability and genetically abnormal daughter cells. Notably, during perturbed mitosis, malignant cells pursue an alternate route. In contrast to normal cells, they overcome SAC-enforced arrest by slow degradation of cyclin B1. This enables premature mitotic exit and entry into interphase without undergoing correct segregation of chromosomes or cytokinesis resulting in the formation of tetraploid or multinucleated cells [5].

Interestingly, studies from the Taylor laboratory have shown a striking adaptive behavior of different cell types as a consequence of long-term mitotic arrest wherein the fate of these cells fluctuates between mitotic catastrophe and slippage [6,7]. The outcome is dictated by opposing activities of, respectively, the apoptotic machinery and the anaphase-promoting complex/cyclosome (APC/C), the latter of which leads to cyclin B1 degradation [6,7]. Termed the “competing networks-threshold model”, it underpins the critical role of cyclin B1 degradation, as a premature decline of cyclin B1 level below the mitotic exit threshold results in mitotic slippage. Conversely, if caspase activation arises before adequate cyclin B1 activation, it leads to breaching of the apoptosis threshold leading to mitotic catastrophe. Consistently, activities of the anti-apoptotic members of the BCL2 family, composed of BCL2, BCL-XL, MCL1, BCL-W, and BFL1, in association with their pro-apoptotic antagonists BIM, BID, PUMA, NOXA, BAD, BAX and BAK1, strongly influence patient outcome in the clinic and confer acquired resistance against anti-microtubule therapies [8]. Their activation is defined by “BH3-only proteins” mainly involving BIM and BID which facilitate dimerization of BAX and BAK1 in the mitochondrial outer membrane to form the toroidal pore. Consequently, apoptogenic factors such as Cytochrome-C, present in the inner mitochondrial membrane space, are released and trigger caspase machinery activation to initiate apoptosis [9,10].

The cell death machinery: a double-edged sword

Phos-tag gel analysis has revealed post-translational modifications of multiple phosphorylation sites located in an unstructured loop of BH3-only proteins (for example BCL2) between alpha 1 (BH4 domain) and alpha 2 (BH3 domain), which either inhibit or enhance the function of the protein and hence dictate cell fate during perturbed mitosis [11]. Importantly, these post-translational modifications dictate the fate of the proteins by either promoting or saving them from proteosomal degradation and hence influence whether cells breach the apoptotic threshold [12]. Notably, active CDK1 has been hypothesized to determine the outcome of mitotic exit or catastrophe, as independent studies have highlighted contrasting functions of CDK1 during perturbed mitosis. Allan et al. have shown that the CDK1/cyclin B1 complex inactivates caspase 9 by phosphorylation at Thr125 which leads to mitotic slippage [13]. In contrast, Harley et al. demonstrated that the complex induces MCL-1 phosphorylation at Thr92 causing its APC/C-mediated degradation to trigger mitotic arrest [14]. MCL1, a mitotic death timer and the most critical BCL2 family member, is prone to undergo intense phosphorylation and degradation during prolonged mitotic arrest, as it competes with cyclin B1 for binding to the proteolysis machinery; hence determining the cell fate of cancer cells during perturbed mitosis [15].

In a recent study, Haschka et al. have systematically determined the impact of individual BCL2 family proteins and have mechanistically shown that the phosphorylation of BCL2 and BCLX during mitosis triggers NOXA-dependent MCL1 degradation, which primes mitotic cell death in a BIM-dependent manner [11]. The study demonstrates a comprehensive model that involves requirement of the BH3-only proteins to initiate the molecular process of apoptosis in cells undergoing perturbed mitosis. On the other hand, during mitotic arrest, BIM is known to be heavily phosphorylated at S69 by Aurora A kinase, which leads to its ubiquitin-dependent degradation and results in mitotic slippage [16]. Furthermore, ERK has been shown to confer taxol resistance, as ERK inhibition in the presence of low dosages of mitotic poisons increased baseline levels of BIM and led to BIM-dependent mitotic cell death [17]. In addition, multiple studies have shown the contribution of BH3-only proteins, such as activating BAX and BAK1, in association with post-translational modification or proteasomal degradation of pro-apoptotic BCL2 protein, which modulates apoptosis under mitotic stress [18,19]. BAX and BAK1, though functionally redundant, are critical for mitochondrial apoptosis, as their deficiency or depletion provides significant advantage in mitotic slippage during aberrant mitosis [20]. Interestingly, using a genome-wide RNAi screen, the Taylor group again have demonstrated evidence that MYC induces upregulation of certain BH3-only proteins, namely BIM, BAD and NOXA, while it represses BCL-xL levels leading to p53-independent cell death by cooperating with EGR1 during abrogated mitosis [21]. Further, they have shown in vitro that chemical inhibition of BCL-xL leads to accelerated mitotic entry and cell death of MYC-deficient cells. This is consistent with work previously reported by the Mitchison group wherein, following RNAi silencing and the pan-BCL2 inhibitor navitoclax, also known as ABT-263, BCL-xL was identified to critically block cell death in response to antimitotic drugs [22].

Chromosomal instability: an impenetrable defense

Genomic instability (GIN), is a well-defined hallmark of cancer facilitating procurement of evolutionary somatic mutations leading to tumor plasticity and heterogeneity [23,24]. It is a phenomenon characterized by accumulation of genetic alterations due to improper chromosome segregation, failure of cell cycle checkpoint control and perturbed mitosis, defects in telomere maintenance or failure to repair damaged DNA and genome doubling [25,26]. Chromosomal instability (CIN), refers to the aberrant gain or loss of chromosomes during mitosis as a result of chromosome missegregation and leads to the formation of aneuploid daughter cells, i.e., cells with an abnormal chromosome number [27]. Both GIN and CIN are well established to provide phenotypic variation to accommodate clonal expansion of cancer cells, resistance to chemotherapy and tumor relapse [28,29].

Notably, there exists a direct link between CIN and mitotis-associated cell death regulation, which requires careful evaluation to improve clinical outcome. As such, there exists variation in the level of numerical (nCIN) and structural (sCIN) chromosomal abnormalities between cell populations [30]. Features of aneuploidy include abnormalities in chromosome numbers, which arise from persistent chromosome segregation errors during mitosis and contributes to oncogenic clonal evolution [31]. A number of cellular mechanisms causing aneuploidy have now been delineated and it is evident that aneuploidy can result from the perturbation of a variety of pathways that normally ensure faithful segregation of chromosomes during mitosis [29]. CIN is common in solid cancers and mostly arises due to a weakened or hyper-active mitotic checkpoint [27,32,33]. Aberrant expression of a wide variety of mitotic genes such as PLK1, AURK-A and -B, WEE1, MPS1, MAD2, FOXM1 have individually been shown to cause CIN in tumor cells [34]. Many of these are also part of CIN gene expression signatures, which predict poor patient outcome for a range of solid cancers [35–37]. Mitotic slippage is a prerequisite for aneuploid cell survival. The altered genomic make-up of aneuploid cells may result in these cells’ acquiring drug resistance. Indeed, CIN has been shown to directly promote drug resistance in cell lines [38–40]. In addition, clinical data strongly support the intricate link between CIN and drug resistance, as members of drug resistance gene expression signatures frequently overlap with those of CIN signatures and both are strong predictors of poor cancer patient survival [35–37,41,42].

The random nature of CIN results in the generation of cells with a wide variety of reshuffled karyotypes. At the cellular level, this may cause lethality or provide cells with malignant advantages, such as drug resistance. At the tumor level, this induces tumor heterogeneity, which in turn paves the way for clonal expansion of cancer cells. Treatment with chemotherapeutic drugs may eradicate non-resistant cells, but resistant clones may survive and expand, resulting in tumor relapse and poor patient outcome [28,29]. Studies have highlighted the involvement of intra-tumor heterogeneity in facilitating the adaptation of tumors to environmental stress and induce network rewiring causing intrinsic chemo-resistance leading to worse overall and relapse free survival [39]. Mechanistically, cancer cells may increase their likelihood to become drug-resistant by modulating the expression of key apoptotic regulators, such as BCL-xL, referred to above. This would sustain tolerance to ongoing CIN by shifting the “competing networks” towards mitotic slippage, rather than mitotic cell death. In any case, it is evident that CIN promotes intrinsic multi-drug resistance both in vitro and in cancer patients [36,39].

Mitotic kinase inhibitors: a flawed rationale

Mitotic kinase inhibitors have been shown to facilitate CIN in tumor cells. Hence, they have been widely explored on the rationale that the clinical efficacy of mitotic kinase inhibitors could be endorsed for stimulation of mitotic arrest, which would selectively eliminate dividing cells with minimal side effects. Mitosis, a critical phase of the cell cycle, requires careful orchestration of several molecular factors, such as cyclins and CDKs, and coordination of numerous checkpoints for monitoring the order, integrity and fidelity of major events in the cell cycle [43,44]. At G2 phase, CDK1 levels remain low, though an increased activity of CDK1 in conjunction with B type cyclins is required for mitotic entry. Therefore, activation of CDK1 is the focal point of many signaling pathways that control the commitment of a cell to mitosis [45,46]. The full activation of CDK1 and the entry of M phase are restrained by WEE1 and MYT1, the endogenous inhibitors of CDK1 [47,48]. These kinases phosphorylate CDK1 on a tyrosine residue (Y15) and the adjacent threonine residue (T14) upon binding to cyclin B1 disabling a conformational change required for actively binding to downstream substrates [48–50]. Thus, upon binding of CDK1 with cyclin B and phosphorylation by CAK and WEE1, the CDK1 is primed but requires dephosphorization of Y15 and T14 by CDC25 phosphatase to initiate entry into mitosis [51]. CDC25, a dual-specificity protein phosphatase, enables the dephosphorylation of CDK1 at Y15 and T14 allowing its translocation into the nucleus enabling G2/M transition and initiation of mitosis [52–54].

The SAC, also called mitotic checkpoint, is known for arresting mitotic progression during cell cycle dysfunction and ensures proper chromosome alignment at the metaphase plate prior to controlled segregation of the chromosomes. The SAC’s main function is to delay anaphase until accurate kinetochore-microtubule attachment of each chromosome at the metaphase plate [51,55,56]. Upon improper attachment, an inhibitory signal arises from the kinetochore to induce recruitment of a protein machinery involving MAD2, BUBR1, BUB3 and MPS1 (also called TTK) caused due to tension at the kinetochore sensed by AURKB, survivin and INCENP [57–59]. The APC/C complex is the major mitotic ubiquitin ligase that controls timely degradation of numerous regulators of mitosis including Cyclin B, AURORA kinase and PLKs [60]. APC/C depends on CDC20 and CDH1 for substrate specificity and its activity is controlled heavily by the SAC. The APC/C-CDH1 complex is activated in late mitosis after all chromosomes contact bipolar spindle. It then degrades B-type Cyclins and PTTG1 (securin) [61]. A significant number of genetic alterations, mainly gene amplification, of these mitotic regulators have been demonstrated both in vitro and in vivo to contribute to tumorigenesis. Thus, these mitotic kinase proteins have been widely explored as potential targets for cancer therapeutics by rationalizing the fact that their inhibition might induce mitotic aberrations and induce mitotic cell death [34].

However, inhibitors such as those targeting PLK1, MPS1, WEE1 and AURORA and AURORB, either alone or in combination with adjuvant and neoadjuvant chemotherapy, have not been successful in the clinic [62]. One of the major challenges faced by these inhibitors is to tackle the tumor heterogeneity imposed by CIN, leading to mitotic slippage, cell survival and in turn poor survival. However, we must not disregard the importance of targeting the mitotic spindle and cleavage apparatus in the future. Instead, application of better strategies for druggable vulnerabilities in mitosis may provide considerable clinical benefit. Collectively, these evidences emphasize the fact that CIN triggers an aggressive tumor phenotype and imposes a vital clinical challenge against the efficacy of classical chemotherapy and small molecule mitotic kinase inhibitors.

CEP55: a rising star of tumorigenesis

The current need in clinic is to explore the mitotic candidates that dictate the fate of aneuploid cells during perturbed mitosis which could potentially broaden our mechanistic understanding of how these cells resist mitotic poisons. Our laboratory has recently demonstrated that CEP55, a centrosomal protein, plays a pivotal role in dictating the cell fate of aneuploid cells during perturbed mitosis [63]. We initially identified CEP55 (also known as c10orf3 and FLJ10540) as a critical component of cell abscission that is regulated post-translationally in a phosphorylation-dependent manner by CDK1, ERK2 and PLK1 for timely recruitment to the midbody, independent of the cell cycle [64]. Similar to other centrosomal proteins, CEP55 is a highly coiled-coil protein and localizes primarily to the pericentriolar material (PCM) but is also allied with the mother centriole [64]. Various independent studies over the past decade have highlighted the critical role of CEP55 in recruiting the ESCRT machinery at the midbody and facilitating equal segregation of cytoplasmic contents between the daughter cells (elaborately discussed in our previous review [65]).

Alongside its role in the regulation of cytokinesis in mitotic cells, the gene signature of CEP55 has been reported to be associated with poor prognosis in multiple tumors [65]. Its expression has been significantly correlated with aggressiveness, tumor stage, metastasis and poor prognosis across multiple tumor types. Carter et al. have shown that CEP55 is part of the top 70 genes identified as responsible for CIN (CIN70 signature) from analyses involving 12 different cancer types [37]. In addition, CEP55 is part of a CIN signature of 10 genes whose high expression is associated with drug resistance, CIN and cell proliferation [67]. CEP55, also a potent cancer testes antigen and a key regulator of spermatogenesis [68], has been linked mainly to regulation of the PI3K/AKT pathway and its overexpression is associated with proliferation, invasion, migration and metastasis [66]. CEP55 binds to and stabilizes the catalytic subunit, p110 of PIK3CA and increases AKT signaling, in vitro and in vivo [68,69]. Independently, studies have shown the interplay between CEP55 and FOXM1 [70,71] in the cancer context, which is negatively regulated by TP53 in a PLK1-dependent manner [72]. However, the mechanisms underlying this are not well defined. Independent studies in pancreatic and hepatocellular cancer have demonstrated the involvement of CEP55 in the regulation of the NF-ĸB and JAK2- STAT3 signaling pathway respectively [73,74]. CEP55 has been identified as a cancer testis antigen (CTA) and tumor-associated antigen (TAA) [75]. This makes CEP55 an ideal candidate for vaccine therapy. Preliminary studies have shown efficacy of CEP55-based immunotherapy vaccines with higher potential in the treatment of chemotherapy-resistant colon cancer stem cells and tumor-initiating cells [75,76]. Taken together, these evidences highlight the rise of CEP55’s role in promoting tumorigenesis and indicate its importance as a potent therapeutic target.

CEP55: a vital target of aneuploid cells

Our laboratory has extensively characterized the physiological role of CEP55 in normal homeostasis and its molecular function promoting CIN and tumorigenesis in vivo by using breast cancer (BC) and transgenic mouse models. Notably, in BC, we have first reported that overexpression of CEP55 protects aneuploid cells, as loss of CEP55 causes copy number changes suggesting that CEP55 expression determines the survival of genomically unstable cells among heterogeneous cell populations. Moreover, in BC patients we observed that higher CEP55 expression is associated with gain of chromosome 20 (20q arm) and leads to chemotherapeutic resistance, particularly against docetaxel, as CEP55 knockdown resulted in reduced proliferation and cell death. In addition, high level of CEP55 was found to be inversely associated with relapse-free survival of chemotherapy-treated BC patients and in vitro facilitated substantial mitotic slippage alongside increased polyploidy when cells were challenged with mitotic poisons [63]. Mechanistically, we have shown that loss of CEP55 leads to significantly better mitotic arrest and induced mitotic cell death during perturbed mitosis, led by premature mitotic entry as indicated by premature activation of CDK1. We also observed suppression of baseline BCL2, BCL-xL and MCL1 levels, in contrast to higher BIM level, which led to increased Caspase-2 and finally PARP and Caspase-3 cleavage in CEP55 knock-down cells during aberrant mitosis. Notably, we have also demonstrated that the MAPK pathway transcriptionally controls CEP55 in a MYC-dependent manner and ERK1/2 inhibition mimicked CEP55 depletion in vivo. Therefore, using this model, we rationalized that dual targeting of MEK1/2 and PLK1 using specific small molecule inhibitors could serve as synergistic therapeutic strategies to target CIN or aneuploid BC cells in patients [63]. Thus, we put forth a model in which overexpressed CEP55 protects heterogeneous aneuploid cell populations and its depletion induces premature CDK1 activation, hence leading to cell death by facilitating breach of the apoptosis threshold and preventing mitotic slippage (Figure 1). Thus, the link between CEP55 and aneuploidy represents a new twist in the tale of mitotic slippage. We expect that further exploring this may well improve therapeutic approaches.

Figure 1.

Cell fate during perturbed mitosis influenced by CEP55 expression. When cancer cells (heterogeneous cell population) are exposed to anti-mitotic chemotherapy drugs, they undergo arrest in mitosis due to chronic activation of the Spindle Assembly Checkpoint (SAC). Cell fate is determined by two competing networks, one that allows exit of mitotically arrested cells and the other that activates cell death. Higher CEP55 expression facilitates exit from mitotic arrest (mitotic slippage) resulting in polyploidy and resistance to anti-mitotic drug-induced death. However, when CEP55 levels are reduced by MEK1/2 inhibition followed by treatment with anti-mitotic drugs (PLK1 inhibition) cells initiate premature entry into mitosis and activate the cell death pathway.

Independent studies which involve MEK1/2 inhibition in association with docetaxel [77] and PLK1 [78] showed reduced BC and melanoma xenograft growth in vivo respectively. This supports our model and could significantly impact clinical practice, in particular the treatment of aggressive, heterogeneous and genomically unstable malignant tumors. Collectively, strategies involving elimination of CIN/aneuploid cell populations in combination with targeting mitotic proteins could potentially overcome mitotic slippage and induce mitosis-associated cell death. Hence, by imposing synthetic lethality, these tactics could prove to be vital in the development of novel and effective targeted therapies for better clinical outcome and should be encouraged to be translated from bench to bedside.

Funding Statement

This work was supported by National Health & Medical Research Council (NH&MRC) Program Grant [ID 1017028].

Acknowledgments

We thank the members of the Duijf and Khanna laboratories for helpful discussions. We would like to thank QIMR Berghofer Medical Research Institute for providing the necessary facilities and support.

Disclosure statement

No potential conflict of interest was reported by the authors.