TRIGGERING ANAPHASE CATASTROPHE TO COMBAT ANEUPLOID CANCERS

Abstract

Cancer cells are genetically unstable and often have supernumerary centrosomes. When supernumerary centrosome clustering is inhibited at mitosis, multipolar cell division is forced, triggering apoptosis in daughter cells. This proapoptotic pathway is called anaphase catastrophe. Cyclin-dependent kinase 1 (CDK1) or CDK2 inhibitors can antagonize centrosome clustering and cause anaphase catastrophe to occur in lung cancer and other types of cancer. The centrosome protein CP110, a CDK1 and CDK2 phosphorylation substrate, engages anaphase catastrophe. Intriguingly, CP110 is downregulated by the KRAS oncoprotein, enhancing sensitivity of KRAS-driven cancers to CDK2 inhibitors. Anaphase catastrophe eradicates aneuploid cancer cells while relatively sparing normal diploid cells with two centrosomes. This therapeutic window discriminates between normal and neoplastic cells and can be exploited in the cancer clinic. The work reviewed here establishes that pharmacologically-induced anaphase catastrophe is useful to combat aneuploid cancers, especially when the KRAS oncoprotein is activated. This addresses an unmet medical need in oncology.

INTRODUCTION

Centrosome numbers are critical for proper spindle formation and chromosome segregation (1,2). In normal cells, centrosome numbers are tightly controlled. However, cancer cells are associated with genomic instability (3), a hallmark of cancer (4,5). Moreover, in cancer cells, centrosome numbers are frequently deregulated and supernumerary centrosomes exist (6-10). At mitosis, cancer cells cluster these supernumerary centrosomes into two spindle poles so that bipolar spindles are formed and chromosomes are properly segregated (11-15). When centrosome clustering is inhibited, cancer cells with supernumerary centrosomes are forced to undergo multipolar cell mitosis, which causes daughter cells to develop apoptosis after chromosome missegregation (7,16-18). This proapoptotic pathway is known as anaphase catastrophe (19,20).

Since normal cells do not typically possess supernumerary centrosomes with some exceptions like polyploid hepatocytes (21,22), anaphase catastrophe will affect preferentially cancer cells with supernumerary centrosomes while sparing normal cells from undergoing multipolar mitosis followed by apoptosis in daughter cells (13,19,23-26). This indicates that anaphase catastrophe induction can be therapeutically exploited to combat diverse cancers (19,27-29). Notably, it was found that cyclin-dependent kinase 1 (CDK1) or CDK2 inhibition hampers centrosome clustering and CDK1 or CDK2 inhibitors can trigger anaphase catastrophe and exert antineoplastic effects in lung cancers (24,30,31).

We review here the mechanisms engaged for anaphase catastrophe caused by CDK1 or CDK2 inhibition. Anaphase catastrophe is mediated through the centrosome protein, CP110, which is a direct substrate of CDK1 or CDK2, as reviewed (20). Intriguingly, CP110 expression is downregulated in KRAS-driven lung cancer cells and KRAS oncoprotein expressing cancer cells are particularly susceptible to anaphase catastrophe-inducing agents (20,24,31-33). Because treatment of KRAS-driven cancers is an unmet medical need, this finding has translational relevance. Furthermore, since anaphase catastrophe is a distinct pathway that is separate from those engaged by other chemotherapeutic drugs, combined regimens of agents that confer anaphase catastrophe with other antineoplastic drugs (especially those that disrupt chromosome stability) should be considered for treating many different cancer types.

ANAPHASE CATASTROPHE CAUSED BY CDK1 OR CDK2 INHIBITION

When lung cancer cells were treated with seliciclib (CYC202, Cyclacel), an orally bioavailable CDK2/7/9 inhibitor, multipolar spindles unexpectedly formed during mitosis (24). This was followed by multipolar cell division, and live cell imaging along with cytochrome C staining revealed that daughter cells after multipolar cell division exhibit apoptosis (24,32). Deregulation at anaphase can augment apoptosis, and this engaged a distinct death pathway designated as anaphase catastrophe (19,24).

Given that the pharmacologic target of seliciclib is CDK2 (IC50s for seliciclib are 100 nmol/L for CDK2, 160 nmol/L for CDK5, 540 nmol/L for CDK7, and 950 nmol/L for CDK9) (34-36), CDK2 inhibition was assessed as potentially responsible for inducing anaphase catastrophe. When CDK1, CDK2, CDK5, and CDK9 were individually downregulated using independent siRNAs, it was found that knock-down of CDK1 or CDK2 each induced anaphase catastrophe in lung cancer cells, independently confirming the prior findings obtained using a pan-CDK inhibitor (30). In addition to seliciclib, other selective CDK2 or pan-CDK inhibitors including dinaciclib (SCH727965, Merck) (30), CCT68127 (Cyclacel) (31), and CYC065 (Cyclacel) (31) also triggered anaphase catastrophe. Notably, anaphase catastrophe was augmented after a CDK2 inhibitor treatment not only in lung cancer cells, but also in other cancer cells (personal communication, Dr. Masanori Kawakami), implicating anaphase catastrophe as a broadly active antineoplastic mechanism in cancer therapy.

Centrosomes play a pivotal role in spindle formation. Given this, studies were undertaken to determine whether centrosome numbers or biology were affected by CDK2 antagonism. It was found that lung cancer cells frequently possess supernumerary centrosomes and supernumerary centrosomes were not clustered at mitosis after CDK2 inhibitor treatment (24,30,31). Supernumerary centrosomes are known to cluster into two spindle poles so that chromosomes are properly segregated with bipolar spindle formation (11-15). In contrast, CDK2 inhibitors antagonize this clustering of supernumerary centrosomes, resulting in multipolar spindle formation followed by multipolar mitosis. This confers anaphase catastrophe, as summarized in Figure 1.

Fig. 1.

Summary of anaphase catastrophe. Mitosis in either normal cells with two centrosomes or cancer cells with supernumerary centrosomes is displayed. CDK1 or CDK2 inhibition blocks clustering of supernumerary centrosomes into two poles, forcing cells to undergo multipolar cell division. This results in apoptotic death of daughter cells, known as anaphase catastrophe. The sizes of the arrows indicate the relative impact of the highlighted pathway. The red rectangle indicates that CDK1 or CDK2 inhibition blocks centrosome clustering.

Because supernumerary centrosomes are required for anaphase catastrophe induction, anaphase catastrophe after CDK2 inhibition preferentially occurs in polyploid cancer cells that have supernumerary centrosomes while relatively sparing normal cells that do not possess supernumerary centrosomes. Indeed, CDK2 inhibitors did not substantially augment anaphase catastrophe in immortalized pulmonary epithelial cells that are not as chromosomally unstable as are lung cancer cells (24,31). This finding identified the anaphase catastrophe pathway as worth exploiting as an antineoplastic mechanism to engage in the cancer clinic.

CP110 AS A MEDIATOR OF ANAPHASE CATASTROPHE

Based on the fact that CDK2 inhibition causes anaphase catastrophe, known substrates of CDK2 phosphorylation were comprehensively screened to uncover potential protein targets affected by CDK2 inhibition that would cause anaphase catastrophe (32). Knock-down of CP110, a centrosomal protein that is directly phosphorylated by cyclin E/CDK2, cyclin A/CDK2, and cyclin B/CDK1 (37), was found to trigger anaphase catastrophe. This means CP110 is involved in the augmentation of anaphase catastrophe after CDK2 inhibition (32).

CP110 is involved in centrosome duplication (37), maturation (38), inhibition of premature centrosome separation (37), and cytokinesis (39,40). Depending on the cell cycle phase, CP110 expression levels and its localization are regulated (37). CP110 expression, which is low at the early G1 cell cycle phase, is increased during the G1/S transition (37). Subsequently, CP110 expression decreases during the G2/M phase and diminishes after mitosis (37). In the G1/S cell cycle phase, CP110 regulates centrosome duplication and maturation with CP110 peak expression during the S phase (37,38). In the G2/M phase, CP110 protein levels decline, leading to centrosome separation and cytokinesis (39,40). Both CP110 knock-down and engineered mutation of its CDK2 phosphorylation sites result in the unscheduled separation of centrosomes (37).

Gain of CP110 expression antagonizes the induction of anaphase catastrophe caused by either genetic or pharmacological antagonism of CDK2 (30-32). When the potential CDK2 phosphorylation sites of CP110 were replaced by alanine residues, anaphase catastrophe after CDK2 inhibition was not antagonized by engineered gain of expression of this mutant CP110 species (32). These findings indicated that inhibition of supernumerary centrosomes clustering and induction of anaphase catastrophe after CDK2 antagonism were mediated by opposing CP110 phosphorylation. Specifically, individually mutating 10 different candidate CDK2 phosphorylation sites within CP110 uncovered serine 170 and threonine 194 that are located between the CP110 coiled-coil and destruction box (D-box) domains (37) were responsible for this regulation of centrosome clustering and the subsequent appearance of anaphase catastrophe after CDK2 inhibition (33). These phosphorylation sites are displayed in Figure 2.

Fig. 2.

The protein structure of CP110 and its phosphorylation sites. The 10 potential CDK2 phosphorylation sites in the CP110 amino acid sequence are shown. In black boxes are the two residues, which, once mutated, can alter CDK2-driven induction of anaphase catastrophe.

CP110 protein expression changes during cell cycle progression are regulated by ubiquitination-dependent mechanisms (41,42). CP110 is associated with the F-box protein cyclin F and ubiquitinated by the SCF (Skp1-Cul1-F-box protein)^{cyclin F} ubiquitin ligase complex that leads to degradation of CP110 during the G2 phase of the cell cycle (41,42). Knock-down of cyclin F, a key component of SCF (Skp1-Cul1-F-box protein)^{cyclin F} ubiquitin ligase complex, by siRNA attenuated the destabilization of CP110 in the G2 cell cycle phase and this led to centrosome and mitotic abnormalities (41). This indicated that SCF^cyclinF-mediated degradation of CP110 was necessary for proper mitosis and genomic integrity (41).

Deubiquitination is also critical for regulation of CP110 protein (42). CP110 is deubiquitinated by the ubiquitin-specific protease 33 (USP33), a deubiquitinating enzyme, that counteracts SCF^cyclinF-mediated ubiquitination (43). Depletion of USP33 leads to destabilization of CP110 protein, which antagonizes centrosome amplification (43). When USP33 is co-depleted with cyclin F, the mitotic defects that are caused by knock-down of cyclin F (SCF^cyclinF) are attenuated (43). Likewise, USP33 knock-down that destabilizes CP110 protein enhanced anaphase catastrophe caused by CDK2 inhibition, thereby confirming a direct role for CP110 in causing anaphase catastrophe (33).

CP110 AND KRAS ONCOPROTEIN

Unexpectedly, the basal cyclin F protein expression was elevated in KRAS mutant-expressing lung cancers as compared with KRAS wild-type lung cancer cells (33). As noted, cyclin F is a key negative regulator of CP110 protein expression (41). When the KRAS oncoprotein was introduced, CP110 protein levels declined; when KRAS was repressed in lung cancer cells, CP110 expression increased (33). This occurred through the ubiquitin ligase SCF (cyclin F) that can target CP110 protein for destabilization (33). CP110 expression was also substantially lowered in KRAS mutant lung cancers as compared with KRAS wild-type ones in murine lung cancer cellular, genetically-engineered mouse lung cancer models, and in human lung cancer cases, as individually assessed by immunohistochemical assays (32).

High throughput pharmacogenomic analyses revealed that lung cancer cells harboring KRAS mutations were especially sensitive to CDK2 inhibitors, such as seliciclib (24) and CCT68127 (31), as compared to KRAS wild-type-expressing lung cancer cells. CDK2 inhibitors were found to inhibit CP110 activity more readily in KRAS mutant lung cancer cells where CP110 expression was basically reduced (32). Since KRAS mutations are detected in about 30% of lung cancers (44) and the treatment of lung cancer cases driven by the KRAS oncoprotein is an unmet medical need with poor clinical prognosis (44-48), the observed preclinical response of KRAS mutant lung cancers to CDK2 inhibitors has substantial translational implications (19,24,31). Furthermore, KRAS mutations are also detected frequently in other cancers such as pancreatic and colon cancers (49). Thus, engagement of anaphase catastrophe could be pursued as an antineoplastic approach beyond lung cancer cases. Preliminary studies indicate that CYC065-triggered anaphase catastrophe is found not only in lung cancer but also in other KRAS-driven cancer cells (personal communication, Dr. Masanori Kawakami). Because bipolar cells do not readily exhibit anaphase catastrophe, a strategy to exploit the anaphase catastrophe pathway in the cancer clinic is promising, given the therapeutic window that exists for this induced death program.

COMBINATIONS WITH OTHER CHEMOTHERAPEUTIC AGENTS

Since anaphase catastrophe is a distinct antineoplastic mechanism, combination therapy with chemotherapeutic agents that act on cooperating pathways is an attractive anticancer strategy to explore. For example, the microtubule-targeting agent taxane causes improper chromosome alignment that enhances chromosome missegregation (50) and potentially has cooperative effects with anaphase catastrophe induction that is caused by enhanced chromosome missegregation. Consistent with this view, different CDK2 inhibitors (seliciclib, dinaciclib, and CCT68127) that confer multipolar anaphase catastrophe were shown to have synergistic or additive antineoplastic effects when combined with taxanes in treated lung cancer cells (24,30,31).

Another promising candidate that potentially has synergistic or additive effects with induced anaphase catastrophe is a polo-like kinase 4 (PLK4) inhibitor (50). PLK4 is a polo-like kinase family member of the serine/threonine kinases that has a critical role in centriole duplication (51,52). Depletion of PLK4 arrests centriole duplication, and gain of PLK4 expression over-duplicates centrosomes (51-54). CFI-400945 is a PLK4 inhibitor that has a dual effect on centriole number, based on its concentration (55-57). At high concentrations, CFI-400945 represses PLK4 activity and inhibits centriole duplication. At lower concentrations, CFI-400945 incompletely represses PLK4 activity and inhibits autophosphorylation of PLK4 that is needed for its degradation, leading to increased PLK4 levels and centriole over-duplication (55). In either case, centrosome number becomes aberrant, which leads to chromosome missegregation after CFI-400945 treatment (55).

CFI-400945 was shown to cause multipolar mitosis after inducing aberrant centrosome numbers, and this agent is reported to have antineoplastic effects in diverse cancers, including breast cancer (54) and lung cancer (58). As summarized in Figure 3, it is rational to hypothesize that this PLK4 inhibitor has cooperative anticancer effects with CDK2 inhibitors because both inhibitors induce multipolar mitosis through distinct mechanisms (induced aberrant centrosome number and anaphase catastrophe). Consistent with this hypothesis, the combination of CDK2 (seliciclib) and PLK4 (CFI-400945) inhibitors had synergistic antineoplastic effects in lung cancer cells (58).

Fig. 3.

Cooperative effects of a PLK4 inhibitor and a CDK2 inhibitor. A PLK4 inhibitor generates supernumerary centrosomes and with a CDK2 antagonist can hamper the clustering of supernumerary centrosomes. Since both mechanisms lead to multipolar mitosis and chromosome missegregation, they are hypothesized to have cooperative antineoplastic effects.

Combinations of anaphase catastrophe mediators with immune-checkpoint inhibitors are also worth exploring given that anaphase catastrophe causes a distinct type of cell death after multipolar mitosis that could augment immunogenicity by exposing neoantigens (59).

CONCLUSIONS

Anaphase catastrophe is a previously unrecognized proapoptotic death program that is caused by inhibition of centrosome clustering. It preferentially eradicates cancer cells with supernumerary centrosomes while relatively sparing bipolar cells. The centrosomal protein CP110 is a direct CDK1 and CDK2 target and has a key role in centrosome clustering (32,37). CDK2 inhibitors alter the phosphorylation of CP110 and inhibit clustering of supernumerary centrosomes, thereby triggering anaphase catastrophe in cancer cells (32). Notably, CP110 protein expression is downregulated by the KRAS oncoprotein (33). Therefore, anaphase catastrophe induction by CDK2 (or CDK1) inhibitors has translational relevance. Activating this pathway holds the promise of combating KRAS-driven cancers where successful treatment is currently an unmet medical need. Furthermore, combinations of anaphase catastrophe-inducing agents with those enhancing other chromosome-destabilizing mechanisms are promising strategies to consider for cancer therapy or prevention. In summary, anaphase catastrophe is a novel antineoplastic mechanism that is worth studying in the laboratory and cancer clinic.

DISCUSSION

Due to technical problems with the Grand Hotel audiovisual equipment, the questions by Dr. Licht associated with this paper and the response by Dr. Dmitrovsky could not be transcribed.