Background
Drugs that disrupt mitotic progression are commonly referred to as 'anti-mitotics' and are extensively used for the treatment of cancer. Classical 'anti-mitotic' chemotherapeutics used in the clinic target microtubules and include the taxanes and vinca alkaloids [
1]. Despite success in the clinic, drug resistance and toxicity have limited their effectiveness, due to the broad role of tubulin in the cytoskeleton of normal and non-dividing cells [
1]. A new class of anti-mitotics have been developed that specifically target mitotic proteins such as Aurora kinase, polo-like kinase, kinesin spindle protein [
1,
2]. Such inhibitors are being characterised as potential chemotherapeutic agents since several induce mitotic failure leading to apoptotic cell death in cancer cells and xenograft mouse cancer models [
2,
3]. These mitotic proteins are either expressed only in dividing cells or apparently function exclusively during mitosis. In contrast to classical anti-mitotics, non-dividing differentiated cells should not be affected by such targeted inhibition, and thus they are predicted to be more efficacious. Many of these targeted inhibitors are currently in cancer clinical trials. Despite the differences in the protein being targeted, both classical and targeted anti-mitotics developed to date aim to disrupt the mitotic spindle or an early stage in mitosis.
We have recently reported a new class of targeted anti-mitotics that do not perturb the mitotic spindle but exclusively block cytokinesis [
4]. The targeted protein for inhibition is the endocytic protein, dynamin II (dynII). DynII is best known for its role in membrane trafficking processes, specifically in clathrin-mediated endocytosis [
5‐
7]. However, dynII also plays an essential role in the completion of the final stage of mitosis, cytokinesis [
4‐
6,
8‐
12]. We and others have developed several classes of dynamin inhibitors including dynasore [
13], dimeric tyrphostins (Bis-Ts), long chain amines and ammonium salts (MiTMABs (myristyl trimethyl ammonium bromides)), dynoles [
14‐
16], iminodyns [
17] and pthaladyns [
18]. Characterisation of the two most potent MiTMABs, MiTMAB and OcTMAB (collectively referred to as MiTMABs), revealed that they block the abscission phase of cytokinesis causing polyploidization, which is analogous to the dynII siRNA phenotype [
4,
8]. The MiTMAB dynamin inhibitors share many favourable characteristics with inhibitors of Aurora kinases, Plk and KSP: (i) they do not affect any other phase of the cell division cycle and (ii) possess anti-proliferative and cytotoxic properties that are selective for cancer cells [
4]. Thus, targeting cytokinesis with dynamin inhibitors may be a promising new approach for the treatment of cancer.
Apoptotic cell death is central to targeted anti-mitotic compounds being highly efficacious as chemotherapeutic agents and is thought to depend on their ability to cause mitotic failure and subsequent accumulation of polyploid cells [
3,
19‐
21]. The mechanism of apoptosis following mitosis failure is poorly understood. It is thought to be classical apoptosis, involving caspase activation and poly(ADP-ribose) polymerase 1 (PARP1) cleavage [
22]. However, cell death induced by caspase-independent mechanisms has been reported [
23,
24]. Apoptotic cell death does not always result following mitotic failure induced by an anti-mitotic. Various cellular responses, depending on the cell line and inhibitor analysed have been reported and include apoptosis, senescence and reversible mitotic arrest [
25]. An in-depth understanding of the mechanisms driving a particular cellular fate in response to targeted anti-mitotics is crucial for rational development and their potential application as chemotherapeutic agents.
In this study, we aimed to determine the fate of cells and the signalling mechanisms involved following treatment with MiTMABs, which exclusively block abscission during cytokinesis. We report that MiTMABs induce cell death following cytokinesis failure in several cancer cells and this was mediated by the intrinsic apoptotic pathway. The cellular response of cancer cells to MiTMABs appeared to correlate with expression of Bcl-2. Our results indicate that the anti-proliferative and cytotoxic properties of the MiTMAB dynamin inhibitors are due to their ability to induce apoptosis following cytokinesis failure. This provides the first evidence that targeting cytokinesis is a valid approach for the development of anti-cancer agents, and that dynII inhibitors are the first class of compounds in this new targeted anti-mitotic group.
Discussion
Dynamin inhibitors are a new class of targeted anti-mitotic compounds. In contrast to the classical (e.g. taxol) and known targeted (e.g. Aurora kinase and Plk inhibitors) anti-mitotic compounds which aim to disrupt the mitotic spindle, the MiTMAB dynamin inhibitors exclusively block cytokinesis without disrupting progression through any other stage of mitosis. Analogous to other anti-mitotic compounds, dynamin inhibitors also have putative anti-tumour activity [
4]. In this study, we show that two dynamin inhibitors called the MiTMABs induce cytokinesis failure and induce apoptosis in cancer cells and this appears to correlate with low expression of the anti-apoptotic proteins Bcl-2 and Mcl-1. Apoptosis occurred strictly following formation of a polyploid cell and was mediated via the intrinsic pathway. Over-expression of the anti-apoptotic protein, Bcl-2, blocked MiTMAB-induced apoptosis but not polyploidization. The induction of apoptosis exclusively following mitotic damage is analogous to the effect of targeted anti-mitotics, such as aurora kinase and Plk inhibitors [
1]. We also demonstrate that apoptosis is induced in cells that have failed cytokinesis due to treatment with the cytokinesis blocker, cytochalsin B. Therefore, this is the first study to demonstrate that cytokinesis blockers can specifically induce apoptotic cell death and thus represent a new class of anti-mitotics with potential anti-cancer activity. Our results indicate that dynamin II is the primary target in this new anti-mitotic action.
Cells exposed to MiTMAB undergo cell death via activation of the intrinsic apoptotic pathway. This was evident by the presence of cleaved caspase-3, -9, and PARP, an increase in DNA fragmentation (<2N DNA content), and membrane blebbing. We further demonstrate that this intrinsic apoptotic pathway involves a feedback caspase-8 amplification loop to drive the execution of apoptosis. MiTMAB-induced cell death exclusively occurred following cytokinesis failure and subsequent polyploidization. This was demonstrated by several findings. Independent single cell analysis using time-lapse microscopy revealed that those MiTMAB-treated cells that failed cytokinesis subsequently underwent apoptotic cell death. We observed an increase in polyploidization in MiTMAB-treated cells when apoptosis was blocked by ZVAD or Bcl-2 overexpression. Caspase-8, -9, -3 and PARP cleavage products were not observed in cells treated with MiTMABs that were not able to undergo a mitotic division (8 h treatment from G
1/S synchronization). Similar reports of cell death specifically following polyploidization in the presence of targeted inhibitors, such as aurora kinase, Plk and KSP inhibitors, have been reported [
1,
2,
40]. This indicates that inhibition of a specific target is not the trigger for apoptosis but rather that it is the phenotype or subsequent molecular alteration generated as a result of its disruption.
The ability of anti-mitotic compounds to induce apoptosis exclusively in dividing cells is the primary rationale that they may be efficacious chemotherapeutic compounds [
3,
19,
20,
41]. However, an increased level of polyploidization does not appear to translate into increased level of secondary apoptosis [
42]. Rather the resulting induction of apoptosis appears to be cell type specific. In line with this idea, the cellular response following exposure to a particular anti-mitotic varies and includes not only apoptosis, but also mitotic catastrophe, senescence and reversible mitotic arrest [
25]. One determinant thought to predict the cellular response to a particular anti-mitotic is the time spent blocked in mitosis [
43]. In the presence of the microtubule-stabilising drugs, ZM447439 (Aurora A/B inhibitor) and taxol, cells blocked in mitosis for >15 h undergo apoptosis shortly after mitotic exit, whereas those cells blocked in mitosis for <15 h showed variable fates with some cells living for days after mitotic exit [
43]. This analysis was carried out in HeLa cells, as done in the present study. In contrast to these findings, the MiTMABs, which block cytokinesis, did not trap cells at this mitotic stage for a long period of time, but only slightly delayed mitotic exit by approximately 30 mins [
4]. Nevertheless, time-lapse analysis indicated that every MiTMAB treated HeLa cell failing cytokinesis proceeded to apoptotic cell death approximately 7-10 hours after exiting mitosis. Conversely, we have previously shown that H460 cells spend a prolonged period of time trapped in cytokinesis in the presence of MiTMABs (up to 24 h) [
4] and these cells remained viable during the following 24 h time period of analysis. Thus, in the case of the MiTMAB-based dynamin inhibitors, the induction of apoptosis appears to correlate with a short (rather than long) period of time that cells spend trapped in cytokinesis. The significance of this correlation needs to be investigated in more detail. Rather, the difference in apoptotic response between these two cell lines likely represents the underlying difference in their molecular components, such as p53 status and Bcl-2 protein levels.
Several reports suggest that p53 status is critical for determining the cellular response following polyploidization [
21,
44,
45]. It is possible that MiTMAB-induced cell death is influenced by p53 status since its expression or mutation status also correlated with sensitivity (HeLa: p53
wt but almost undetectable levels due to HPV, HT29: p53
mut and SW480: p53
mut) and resistance (MCF-7 and H460 contain p53
wt) to apoptosis. Given that this gene is frequently lost or mutated in cancers [
46], the ability of dynamin inhibitors to induce apoptosis following polyploidization in cells lacking functional p53 could be a favourable characteristic as a potential chemotherapeutic agent. It could be particularly relevant to those drug resistant cancers that often develop following p53 mutation. However, the contribution of p53 in determining the cellular response following polyploidization is under debate and is complicated by its multiple roles. For example, in response to aurora kinase inhibitors, p53
wt is required for G
1 arrest of tetraploid cells [
21] and for inducing apoptosis following tetraploid formation [
45]. Therefore, p53 status alone is not the sole predictor of the cellular response following polyploidization.
The expression of Bcl-2 and Mcl-1, but not Bcl-XL, appears to correlate with the ability of cells to undergo apoptosis following exposure to MiTMABs. There are six anti-apoptotic Bcl-2 family members identified and several of these appear to contribute to drug resistance in cancer cells [
37,
38], suggesting that inhibition of multiple Bcl-2 family members will be necessary to achieve an optimal therapeutic effect. The development of antagonists toward Bcl-2 [
47] and Mcl-1 [
48] provide an attractive hypothesis that MiTMABs may synergise with these antagonists to sensitise resistant cell lines to undergo apoptosis. In line with this idea, the Bcl-2 antagonists, ABT-737 or ABT-263, have been shown to synergise with Plk and aurora kinase inhibitors [
49] as well as conventional chemotherapeutic drugs, such as vincristine,
in vitro and
in vivo[
50].
Conclusions
Overall, our findings demonstrate that the MiTMAB family of dynamin inhibitors induce apoptosis in a concentration-dependent manner following polyploidization. More specifically, these are the first reported targeted anti-mitotic compounds which induce polyploidy by exclusively blocking cytokinesis. Thus, dynamin inhibitors are a new class of anti-mitotic compounds with potential anti-cancer action. MiTMAB-induced apoptosis is not only dependent on cytokinesis failure and polyploidization but also on specific molecular components of the apoptotic machinery, such as Bcl-2. Thus, inhibitors of these anti-apoptotic proteins, such as the Bcl-2 inhibitor ABT-737, may act synergistically with the MiTMAB dynamin inhibitors, broadening their therapeutic potential for the treatment of cancer.
Authors' contributions
SJ designed and conducted experiments and carried out data analysis. AB and PJR participated in intellectual discussion of the data and manuscript writing. MC contributed to experimental design, co-ordination of the project, data analysis and manuscript writing. All authors read and approved the manuscript.
Competing interests
The authors declare that they have no competing interests.