Mitotic kinesin-like protein 2 (MKLP2/KIF20A) is a key mitotic regulator frequently overexpressed in human malignancies and its abundance is positively correlated with poor outcomes of the disease. Despite extensive research on MKLP2 as a potential target for oncology, the development of small-molecule inhibitors specific to MKLP2 remains limited. We have previously identified a benzoazipinone compound, HJ81 as a potent disruptor of Aurora kinase B (AURKB) localization during late mitosis. This study reveals that such disruption results from a failure of AURKB relocation at the onset of anaphase and this phenomenon can be specifically attributed to the disablement of MKLP2, a recognized facilitator of the relocation process. Further optimization of HJ81 leads to identifying compounds such as 12a as promising lead inhibitors of MKLP2-mediated processes, with improved pharmacokinetic properties. 12a inhibits the microtubule-stimulated ATPase activity of the recombinant MKLP2 in vitro. Significant suppression of tumor growth was observed in mice bearing the Calu-6 lung cancer cell line when treated with 12a at a well-tolerated dose. Overall, our findings suggest that benzoazipinone derivatives represent a novel chemical scaffold with the potential to be developed to mimic MKLP2 inhibition for cancer treatment.

Abstract The MYC oncogene, a pivotal regulator of various cellular processes, is deregulated in approximately 70 % of human malignancies. Mitotic Kinesin-like Protein 2 (MKLP2) plays a versatile role in both interphase and mitosis and has emerged as a significant prognostic indicator and therapeutic target in cancer. The undruggability of MYC and the scarcity of MKLP2 inhibitors, however, have impeded clinical translation. Our development of JMBI-001, a potent and orally bioavailable small-molecule compound, overcomes these barriers. JMBI-001 elicits loss of function phenotypes in MKLP2 and a synthetic lethal interaction with MYC overexpression. Extensive kinome and safety profiling have revealed no significant off-target effects, and the compound is well-tolerated in long-term animal studies. In preclinical models, JMBI-001 has demonstrated an average tumor growth inhibition rate of 75 % across more than 20 MYC-overexpressing tumor models, including those in the stomach, lung, colon, liver, breast, kidney, skin, and hematopoietic system. Its anti-tumor activity positively correlates with high MYC abundance, aligning with the selective eradication of cells with abundant MYC both in vitro and in vivo. Notably, JMBI-001 also robustly stimulates systemic anti-tumor immunity, enhancing NK and CD3+ T cell infiltration in tumors of syngeneic cancer models. Moreover, additive or synergistic effects have been observed when combined with anti-PD1 therapy even in tumors refractory to the immune checkpoint blockade (ICB) therapy. The dual therapeutic actions of JMBI-001 stem from its disruption of MKLP2 functionalities, leading to anomalies, such as Golgi fragmentation in interphase and multipolarity in pro-metaphase. These disruptions lead to two key outcomes: apoptosis and immunogenic cell death, marked by secreted ATP, released high mobility group protein B1 (HMGB1), and surface-exposed calreticulin. These abnormalities, primed and amplified by deregulated MYC, are not observed in non-transformed cells, suggesting their potential as pharmacodynamic markers for monitoring JMBI-001's activity in vivo. In conclusion, JMBI-001 represents a novel class of anticancer agents that simultaneously triggers MYC synthetic lethality and anticancer immunity by targeting MKLP2-mediated cellular processes. Its unique mechanisms of action, exceptional bioavailability, potency at low nanomolar concentrations, wide-spectrum efficacy, and favorable safety profile establish JMBI-001 as a promising clinical study-ready drug candidate for treating MYC-driven cancers. Citation Format: Ting Zhang, Qiong Shi, Julia Kalashova, Xumei Liu, Xiaohu Zhou, Chenglu Yang, Yan Long, Hongmei Li, Jinhua Li, Gang Lv, Duo Yu, Xuejiao Jiang, Shenqiu Zhang, Jing Zhang, Hong Liu, Dun Yang. An orally available small molecule JMBI-001 elicits MYC-synthetic lethality and anti-tumor immunity by disabling MKLP2-mediated cellular processes [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 5802.

Activity-based drug screens have successfully led to the development of various inhibitors of the catalytic activity of aurora kinases (AURKs), major regulatory kinases of cell division. Disrupting the localization of AURKB, rather than its catalytic activity, represents a largely unexplored alternative approach to disabling AURKB-dependent processes. Localization disruptors could be just as specific as direct inhibitors of AURKB activity, may bypass their off-target and select on-target toxicities, and are likely less susceptible to drug resistance resulting from mutations of the AURKB catalytic site. In this study, we demonstrate that the pan-AURK inhibitor AMG900 works at a low concentration not by inhibiting the phosphorylation of H3 at Ser10, an AURKB substrate, but by disrupting the mitotic localization of AURKB. Structural deletion studies pinpoint this undescribed activity to the 2-phenoxy-3,4'-bipyridine moiety of AMG900. Guided by a mechanism-informed phenotypic screening (MIPS) assay, the drug fragment is optimized into a novel class of inhibitors that, at low nanomolar concentrations, can disable AURKB through disruption of its mitotic localization and have desirable oral PK properties. Hierarchical clustering of cell fitness profiles reveals that these compounds cluster with each other, rather than with known AURK inhibitors such as AMG900 and VX-680. Validation studies in mice demonstrate that compound 15a elicits mitotic arrest and apoptosis in NCI-H23 human lung adenocarcinoma xenografts, resulting in a pronounced suppression of tumor growth. The discovery and optimization of compounds that disrupt AURKB localization are successfully facilitated by MIPS. Our findings suggest that 2-phenoxy-3, 4'-bipyridine derivatives have the potential to be further developed as effective therapeutics for the treatment of malignancy by delocalizing AURKB.

We investigated a novel 4-phenoxy-quinoline-based scaffold that mislocalizes the essential mitotic kinase, Aurora kinase B (AURKB). Here, we evaluated the impact of halogen substitutions (F, Cl, Br, and I) on this scaffold with respect to various drug parameters. Br-substituted LXY18 was found to be a potent and orally bioavailable disruptor of cell division, at sub-nanomolar concentrations. LXY18 prevents cytokinesis by blocking AURKB relocalization in mitosis and exhibits broad-spectrum antimitotic activity in vitro. With a favorable pharmacokinetic profile, it shows widespread tissue distribution including the blood-brain barrier penetrance and effective accumulation in tumor tissues. More importantly, it markedly suppresses tumor growth. The novel mode of action of LXY18 may eliminate some drawbacks of direct catalytic inhibition of Aurora kinases. Successful development of LXY18 as a clinical candidate for cancer treatment could enable a new, less toxic means of antimitotic attack that avoids drug resistance mechanisms.

LG157 is a recently identified small-molecule inhibitor of mitotic kinesin-like protein 2 (MKLP2), an overlooked oncology target. This study aims to explore the drug developability of LG157 , by assessing its druglike properties, determining plasma drug exposure in various oral formulations, and exploring the self-emulsifying drug delivery system (SEDDS). Solubility of LG157 ranges from 175 to 228 μM across pH 1.0 to 13.0, with a Log D of 2.41 at pH 7.4. It showed a high protein binding rate of 92.58% in mouse plasma and 90.30% in human serum. The bioavailability radar plot aligns with experimental data (69-85%), indicating good bioavailability. In line with the computation prediction, preclinical formulation studies in mice reveal that all five formulations tested offer decent plasma LG157 exposure, with the highest level of LG157 exposure in the PEG300-based formulation. Subsequent tissue distribution studies in rats indicated that the compound is widely distributed with the highest concentration of LG157 in the liver and the lowest level in the brain. The optimal SEDDS formulation, SEDDS-F14, consists of 65% Oleic acid, 26.25% Tween 20, and 8.75% PEG400 as oil, surfactant, and co-surfactant, respectively. SEDDS optimization, based on the central composite design, has achieved the maximum loading of 188.7 mg/mL for LG157 . These findings support the developability of LG157 and encourage continued exploration and refinement of formulations for improved therapeutic efficacy. ![Figure][1] ### Competing Interest Statement The authors have declared no competing interest. [1]: pending:yes

Therapeutic targeting of MYC directly has proven difficult, but several means to target MYC indirectly using a synthetic lethal drug approach have been proposed. Synthetic lethal approaches for MYC have sought to take advantage of vulnerabilities MYC imposes related to either metabolic reprogramming, apoptotic signaling or the cycling of cancer cells. Here, we describe in detail the cell division phenotypes induced by a MYC synthetic lethal compound, dimethylfasudil (diMF). DiMF is a known ROCK inhibitor, but structurally related ROCK inhibitors are not synthetic lethal with MYC, so the activity of diMF is not related to blockade of this family of kinases. Instead, this compound induced multiple cell cycle-related liabilities. These included the early mitotic arrest of cycling cells followed by mitotic catastrophe-induced death and the induction of polyploidy in cells that do manage to pass through mitosis. As early as prometaphase, we noted diminished staining for BUB1 kinase, which binds to kinetochores and regulates the mitotic spindle checkpoint and chromosome congression. Kinetochore proteins, such as CENP-F, failed to localize at the metaphase plate, confirming a deficit in centromere assembly. This, presumably, contributed to the development of segregation anomalies in diMF-treated cells. In anaphase cells, the protein regulator of cytokinesis 1 (PRC1), failed to be recruited to the midzone, leading to a cascade of defects that included failed recruitment of the chromosomal passenger protein complex, the centralspindlin complex and polo-like-kinase 1 (PLK1). These observations correlate well with the cell death phenotypes induced by diMF, which may serve as a prototype MYC synthetic lethal compound to explore synthetic lethal therapy or as a scaffold upon which to build superior compounds. The phenotypes described here serve as examples of MYC synthetic lethal drug effects that can be used to explore and maximize drug discovery programs.

We combined a mechanism-informed phenotypic screening (MIPS) assay with a structural simplification strategy to guide the discovery of compounds that disrupt the localization of the mitotic regulator, Aurora kinase B (AURKB), rather than inhibiting its catalytic activity. An initial hit 4-(4-methylthiophen-2-yl)-N-(4-(quinolin-4-yloxy)phenyl)phthalazin-1-amine was identified after screening an in-house library of small molecules and phenocopied the loss of function mutations in AURKB without inhibiting its catalytic activity. We isolated this hit compound activity to its 4-phenoxy-quinoline moiety. The fragment was further optimized into a class of new chemical entities that potently disrupt the mitotic localization of AURKB at low nanomolar concentrations and consequently elicit severe growth inhibition in diverse human cancer cell lines. A lead compound, N-(3-methoxy-5-(6-methoxyquinolin-4-yl)oxy)phenyl)acetamide possessed desirable pharmacokinetic properties such as AUC0-infinity: 227.15 [ng center dot h/mL/(mg/kg)]; Cmax: 3378.52 ng/mL T1/2: 3.52 h; and F%: 42 % and produced the AURKB-inhibitory phenotypes in a mouse xenograft model. A lead compound is a powerful tool for interrogating the regulation of AURKB and has the potential to be further developed as a first-in-class oncology therapeutic.

Abstract The potential for synthetic lethality occurs when the same oncogenic events that promote carcinogenesis, also create vulnerabilities for cancer cells. Targeting cancers with these vulnerabilities, but not resident normal tissues, is the promise of synthetic lethal therapies. The compound LC30 is a potent and orally bioavailable compound that is synthetic lethal with deregulated MYC. No off-target liabilities have been demonstrated by kinome and safety profiling and long-term treatment is well tolerated by rodents and canines. LC30 is active against a wide spectrum of cell line xenografts derived from both liquid and solid cancers, including B and T cell malignancies, lung, colon, gastric, mammary and liver cancer cell lines, with efficacy that correlates with levels of MYC. Experiments aimed at unlocking the key to LC30’s synthetic lethality have revealed that LC30 is active in both resting cells and during cell division, so it is not simply an anti-mitotic compound. In resting cells, LC30 disrupts golgi structure, a phenotype that possibly relates to MYC control of centrosomal gene expression. Anti-angiogenic effects are observed in xenografts and have been confirmed using in vitro assays of endothelial cell migration and tubule formation. We speculate that the golgi phenotype and resulting faulty trafficking of secretory vesicles containing pro-angiogenic factors may underly the vascular phenotype. In dividing cells, LC30 potently deregulates the positioning of the chromosomal passenger protein complex (CPPC), a mitotic complex comprised of a catalytic subunit, Aurora B kinase (AURKB), the scaffolding protein inner centromere protein (INCENP) as well as survivin and borealin. The CPPC is dynamically relocated during mitotic progression so that AURKB activity can be localized to specific locations where activity is required. The phenotypic effects of LC30 treatment are consistent with this compound functioning as an inhibitor of the localization, but not catalytic activity of the CPPC. This includes the induction of multipolar spindles, micronuclei and polyploidy, leading to mitotic arrest and both mitotic and post-mitotic apoptosis. This unique mode-of-action, to disrupt CPPC passage in the mitotic cell, is a key component of the MYC synthetic lethality as cells without MYC overexpression do not develop multipolar spindles, lethal polyploidy and apoptose with treatment. LC30 represents a new class of anticancer compound that is a non-kinase, synthetic lethal inhibitor that unlocks MYC-induced vulnerabilities in cancer cells. Citation Format: Qiong Shi, Ting Zhang, Julia Kalashova, Jinhua Li, Chenglu Yang, Hongmei Li, Xiaohu Zhou, Yan Long, Yidan Xia Abuliezi, Gang Lv, Duo Yu, Shenqiu Zhang, Jing Zhang, Thaddeus D Allen, Hong Liu, Dun Yang. An orally available small molecule inhibitor for synthetic lethal targeting of MYC expressing tumors [abstract]. In: Proceedings of the AACR-NCI-EORTC Virtual International Conference on Molecular Targets and Cancer Therapeutics; 2023 Oct 11-15; Boston, MA. Philadelphia (PA): AACR; Mol Cancer Ther 2023;22(12 Suppl):Abstract nr LB_C10.

The MYC oncoprotein represents an intriguing target for cancer treatment, but its therapeutic potential has been hindered by the absence of specific pharmacological inhibitors. In this study, we demonstrate that the phenoxy quinoline compound LXY18 selectively targets and eliminates cells overexpressing MYC, leaving non-transformed cells unharmed. This synthetic lethality arises from an acute induction of multipolarity, resulting in a persistent arrest in early mitosis followed by cell death in mitosis or after mitotic slippage. Distinctively, LXY18’s action contrasts with other antimitotic compounds, as they either fail to induce mitotic arrest or elicit mitotic arrest irrespective of MYC abundance. Furthermore, the MYC abundance in a panel of 98 tumor cell lines correlates with their sensitivity to LXY18. Collectively, our findings uncover LXY18 as an MYC- enabled mitotic blocker and open a new avenue to selectively target MYC-overexpressing tumor cells without affecting normal cells.

This study investigated the metabolism of LXY18, a quinolone-based compound that suppresses tumorigenesis by blocking AURKB localization. Metabolite profiling of LXY18 in liver microsomes from six species and human S9 fractions revealed that LXY18 undergoes various conserved metabolic reactions, such as N-hydroxylation, N-oxygenation, O-dealkylation, and hydrolysis, resulting in ten metabolites. These metabolites were produced through a combination of CYP450 enzymes, and non-CYP450 enzymes including CES1, and AO. Two metabolites, M1 and M2 were authenticated by chemically synthesized standards. M1 was the hydrolyzed product catalyzed by CES1 whereas M2 was a mono-N-oxidative derivative catalyzed by a CYP450 enzyme. AO was identified as the enzyme responsible for the formation of M3 with the help of AO-specific inhibitors and LXY18 analogs, 5b and 5c. M1 was the intermediate of LXY18 to produce M7, M8, M9, and M10. LXY18 potently inhibited 2C19 with an IC50 of 290 nM but had a negligible impact on the other CYP450s, indicating a low risk of drug-drug interaction. Altogether, the study provides valuable insights into the metabolic process of LXY18 and its suitability as a drug candidate. The data generated serves as a significant reference point for conducting further safety assessments and optimizing drug development.

Small molecule inhibitors of aurora kinases are currently being investigated in oncology clinical trials. The long-term effects of these inhibitors on proliferating euploid cells have not been adequately studied. We examined the effect of the reversible pan-aurora kinase inhibitor VX-680 on p53-competent human euploid cells. Circumscribed treatment with VX-680 blocked cytokinesis and arrested cells in G1 or a G1-like status. Approximately 70% of proliferatively arrested cells had 4N DNA content and abnormal nuclei. The remaining 30% of cells possessed 2N DNA content and normal nuclei. The proliferative arrest was not due to the activation of the tumor suppressor Rb and was instead associated with rapid induction of the p53–p21 pathway and p16. The induction was particularly evident in cells with nuclear abnormalities but was independent of activation of the DNA damage response. All of these effects were correlated with the potent inhibition of aurora kinase B. After release from VX-680, the cells with normal nuclei robustly resumed proliferation whereas the cells with abnormal nuclei underwent senescence. Irrespective of their nuclear morphology or DNA content, cells pre-treated with VX-680 failed to grow in soft agar or form tumors in mice. Our findings indicate that an intermittent treatment strategy might minimize the on-target side effects of Aurora Kinase B (AURKB) inhibitory therapies. The strategy allows a significant fraction of dividing normal cells to resume proliferation.

Plants are a rich source for bioactive compounds. However, plant extracts can harbor a mixture of bioactive molecules that promote divergent phenotypes and potentially have confounding effects in bioassays. Even with further purification and identification, target deconvolution can be challenging. Corynoline and acetylcorynoline, are phytochemicals that were previously isolated through a screen for compounds able to induce mitotic arrest and polyploidy in oncogene expressing retinal pigment epithelial (RPE) cells. Here, we shed light on the mechanism by which these phytochemicals can attack human cancer cells. Mitotic arrest was coincident to the induction of centrosome amplification and declustering, causing multi-polar spindle formation. Corynoline was demonstrated to have true centrosome declustering activity in a model where A549 cells were chemically induced to have more than a regular complement of centrosomes. Corynoline could inhibit the centrosome clustering required for pseudo-bipolar spindle formation in these cells. The activity of AURKB, but not AURKA or polo-like kinase 4, was diminished by corynoline. It only partially inhibited AURKB, so it may be a partial antagonist or corynoline may work upstream on an unknown regulator of AURKB activity or localization. Nonetheless, corynoline and acetylcorynoline inhibited the viability of a variety of human cancer derived cell lines. These phytochemicals could serve as prototypes for a next-generation analog with improved potency, selectivity or in vivo bioavailability. Such an analog could be useful as a non-toxic component of combination therapies where inhibiting the chromosomal passenger protein complex is desired.

We used mechanism-informed phenotypic screening to identify and optimize compounds that phenocopy the genetic depletion of the mitotic aurora kinase B (AURKB) kinase. After assaying nine aryl fused seven-membered lactam compounds, we identified a hit compound 6a that was subsequently optimized to five lead compounds with low nanomolar activity, represented by the lead compound 6v (19 nM). With excellent drug-like properties, these compounds reproduced the loss of function in phenotypes of AURKB and exhibited potent cytotoxic activities in various cancer cell lines. Collectively, these data support that seven-membered lactam-based analogs might be valuable for further development as a new type of antimitotic agents for the treatment of cancer.

To identify novel bioactive compounds, an image-based, cell culture screening of natural product extracts was conducted. Specifically, our screen was designed to identify phytochemicals that might phenocopy inhibition of the chromosomal passenger protein complex in eliciting mitotic and cytokinetic defects. A known alkaloid, scoulerine, was identified from the rhizomes of the plant Corydalis decumbens as being able to elicit a transient mitotic arrest followed by either apoptosis induction or polyploidy. In examining the mitotic abnormality further, we observed that scoulerine could elicit supernumerary centrosomes during mitosis, but not earlier in the cell cycle. The localization of NUMA1 at spindle poles was also inhibited, suggesting diminished potential for microtubule recruitment and spindle-pole focusing. Polyploid cells emerged subsequent to cytokinetic failure. The concentration required for scoulerine to elicit all its cell division phenotypes was similar, and an examination of related compounds highlighted the requirement for proper positioning of a hydroxyl and a methoxy group about an aromatic ring for activity. Mechanistically, scoulerine inhibited AURKB activity at concentrations that elicited supernumerary centrosomes and polyploidy. AURKA was only inhibited at higher concentrations, so AURKB inhibition is the likely mechanism by which scoulerine elicited division defects. AURKB inhibition was never complete, so scoulerine may be a suboptimal AURK inhibitor or work upstream of the chromosomal passenger protein complex to reduce AURKB activity. Scoulerine inhibited the viability of a variety of human cancer cell lines. Collectively, these findings uncover a previously unknown activity of scoulerine that could facilitate targeting human cancers. Scoulerine, or a next-generation analogue, may be useful as a nontoxic component of combination therapies where inhibiting the chromosomal passenger protein complex is desired.

A synthetic lethal effect arises when a cancer-associated change introduces a unique vulnerability to cancer cells that makes them unusually susceptible to a drug's inhibitory activity. The synthetic lethal approach is attractive because it enables targeting of cancers harboring specific genomic or epigenomic alterations, the products of which may have proven refractory to direct targeting. An example is cancer driven by overexpression of MYC. Here, we conducted a high-content screen for compounds that are synthetic lethal to elevated MYC using a small-molecule library to identify compounds that are closely related to, or are themselves, regulatory-approved drugs. The screen identified dimethylfasudil, a potent and reversible inhibitor of Rho-associated kinases, ROCK1 and ROCK2. Close analogs of dimethylfasudil are used clinically to treat neurologic and cardiovascular disorders. The synthetic lethal interaction was conserved in rodent and human cell lines and could be observed with activation of either MYC or its paralog MYCN. The synthetic lethality seems specific to MYC overexpressing cells as it could not be substituted by a variety of oncogenic manipulations and synthetic lethality was diminished by RNAi-mediated depletion of MYC in human cancer cell lines. Collectively, these data support investigation of the use of dimethylfasudil as a drug that is synthetic lethal for malignancies that specifically overexpress MYC.

in the CNS and therapeutic resistance which are repetitively seen at the time of microtubule-targeting, the use of these drugs is limited [1-4]. The new generation of mitotic drugs aims for the mitotic regulatory machinery which involves the motor proteins, mitotic kinesins, or the Aurora and polo-like kinases and complexes which are expressed only at the time of cell division [2]. Research efforts are intended towards developing superior antimitotic drugs that would not be only more specific in their action but would also lessen the burden of side effects on patients. Also, because cancer cells demonstrate vast phenotypic miscellany they are characteristically responsive to phenotypic screening which would assist in translating the molecular mechanism as a therapeutic approach in treating cancer with familiar cellular phenotypes following the theory of mechanism-informed phenotypic screening.

Natural sources are a valuable reserve of chemical diversity for drug development. However, creative screening assays are required to unlock this potential. A high-content screen was developed to isolate anti-mitotic and polyploidy-inducing activity in mammalian cancer cells from plant extracts. The assay was utilized in an analysis of more than 2000 medicinal plants and activity was identified in a rhizome extract from Corydalis longicalcarata. Two bioactive benzophenanthridine alkaloids, corynoline and its close analog acetylcorynoline, were purified. These compounds display pleiotropic effects on cell division, including prevention of chromosome congression, compromise of the spindle checkpoint response, and blockade of cytokinesis. Corynoline and acetylcorynoline are known entities but the potential for development as anti-mitotic drugs has never been ascribed to these two phytochemicals. This study highlights the requirement and ability of novel bioassays to unlock novel function for known phytochemicals.

Inhibition of Aurora-B kinase is a synthetic lethal therapy for tumors that overexpress the MYC oncoprotein. It is currently unclear whether co-occurring oncogenic alterations might influence this synthetic lethality by conferring more or less potency in the killing of tumor cells. To identify such modifiers, isogenic cell lines were utilized to test a variety of cancer genes that have been previously demonstrated to promote survival under conditions of cellular stress, contribute to chemoresistance and/or suppress MYC-primed apoptosis. It was found that Bcl-2 and Bcl-xL, two antiapoptotic members of the Bcl-2 family, can partially suppress the synthetic lethality, but not multinucleation, elicited by a pan-aurora kinase inhibitor, VX-680. Suppression was show to stem from the inhibition of autophagy, specifically in multinucleated cells, rather than a general inhibition of apoptosis. The anti-autophagic activity of Bcl-2 also impacted polyploid cell recovery in colony-forming assays, suggesting a route of escape from MYC-VX-680 synthetic lethality that may have clinical consequences. These findings expand on previous conclusions that autophagic death of VX-680-induced polyploid cells is mediated by Atg6. Bcl-2 and Bcl-xL negatively modulate MYC-VX-680 synthetic lethality and it is the anti-autophagic activity of these two Bcl-2 family proteins, specifically in multinucleate cells, that contributes to resistance to Aurora kinase-targeting drugs.