Nearly a century ago, cell biologists postulated that the chromosomal aberrations blighting cancer cells might be caused by a mysterious organelle-the centrosome-that had only just been discovered. For years, however, this enigmatic structure was neglected in oncologic investigations and has only recently reemerged as a key suspect in tumorigenesis. A majority of cancer cells, unlike healthy cells, possess an amplified centrosome complement, which they manage to coalesce neatly at two spindle poles during mitosis. This clustering mechanism permits the cell to form a pseudo-bipolar mitotic spindle for segregation of sister chromatids. On rare occasions this mechanism fails, resulting in declustered centrosomes and the assembly of a multipolar spindle. Spindle multipolarity consigns the cell to an almost certain fate of mitotic arrest or death. The catastrophic nature of multipolarity has attracted efforts to develop drugs that can induce declustering in cancer cells. Such chemotherapeutics would theoretically spare healthy cells, whose normal centrosome complement should preclude multipolar spindle formation. In search of the 'Holy Grail' of nontoxic, cancer cell-selective, and superiorly efficacious chemotherapy, research is underway to elucidate the underpinnings of centrosome clustering mechanisms. Here, we detail the progress made towards that end, highlighting seminal work and suggesting directions for future research, aimed at demystifying this riddling cellular tactic and exploiting it for chemotherapeutic purposes. We also propose a model to highlight the integral role of microtubule dynamicity and the delicate balance of forces on which cancer cells rely for effective centrosome clustering. Finally, we provide insights regarding how perturbation of this balance may pave an inroad for inducing lethal centrosome dispersal and death selectively in cancer cells.

We have previously shown that a non-toxic noscapinoid, EM011 binds tubulin without altering its monomer/polymer ratio. EM011 is more active than the parent molecule, noscapine, in inducing G2/M arrest, inhibiting cellular proliferation and tumor growth in various human xenograft models. However, the mechanisms of mitotic-block and subsequent cell death have remained elusive. Here, we show that EM011-induced attenuation of microtubule dynamics was associated with impaired association of microtubule plus-end tracking proteins, such as EB1 and CLIP-170. EM011 treatment then led to the formation of multipolar spindles containing ‘real’ centrioles indicating drug-induced centrosome amplification and persistent centrosome declustering. Centrosome amplification was accompanied by an upregulation of Aurora A and Plk4 protein levels, as well as a surge in the kinase activity of Aurora A, suggesting a deregulation of the centrosome duplication cycle. Cell-cycle phase-specific experiments showed that the ‘cytotoxicity-window’ of the drug encompasses the late S-G2 period. Drug-treatment, excluding S-phase, not only resulted in lower sub-G1 population but also attenuated centrosome amplification and spindle multipolarity, suggesting that drug-induced centrosome amplification is essential for maximal cell death. Subsequent to a robust mitotic arrest, EM011-treated cells displayed diverse cellular fates suggesting a high degree of intraline variation. Some ‘apoptosis-evasive’ cells underwent aberrant cytokinesis to generate rampant aneuploidy that perhaps contributed to drug-induced cell death. These data indicate that spindle multipolarity induction by means of centrosome amplification has an exciting chemotherapeutic potential that merits further investigation.

Classical anti-mitotic drugs have failed to translate their preclinical efficacy into clinical response in human trials. Their clinical failure has challenged the notion that tumor cells divide frequently at rates comparable to those of cancer cells in vitro and in xenograft models. Given the preponderance of interphase cells in clinical tumors, we asked whether targeting amplified centrosomes, which cancer cells carefully preserve in a tightly clustered conformation throughout interphase, presents a superior chemotherapeutic strategy that sabotages interphase-specific cellular activities, such as migration. Herein we have utilized supercentrosomal N1E-115 murine neuroblastoma cells as a test-bed to study interphase centrosome declustering induced by putative declustering agents, such as Reduced-9-bromonoscapine (RedBr-Nos), Griseofulvin and PJ-34. We found tight ‘supercentrosomal' clusters in the interphase and mitosis of ~80% of patients' tumor cells with excess centrosomes. RedBr-Nos was the strongest declustering agent with a declustering index of 0.36 and completely dispersed interphase centrosome clusters in N1E-115 cells. Interphase centrosome declustering caused inhibition of neurite formation, impairment of cell polarization and Golgi organization, disrupted cellular protrusions and focal adhesion contacts—factors that are crucial prerequisites for directional migration. Thus our data illustrate an interphase-specific potential anti-migratory role of centrosome-declustering agents in addition to their previously acknowledged ability to induce spindle multipolarity and mitotic catastrophe. Centrosome-declustering agents counter centrosome clustering to inhibit directional cell migration in interphase cells and set up multipolar mitotic catastrophe, suggesting that disbanding the nuclear–centrosome–Golgi axis is a potential anti-metastasis strategy.

Centrosome amplification (CA) and resultant chromosomal instability have long been associated with tumorigenesis. However, exacerbation of CA and relentless centrosome declustering engender robust spindle multipolarity (SM) during mitosis and may induce cell death. Recently, we demonstrated that a noscapinoid member, reduced bromonoscapine, (S)-3-(R)-9-bromo-5-(4,5-dimethoxy-1,3-dihydroisobenzofuran-1-yl)-4-methoxy-6-methyl-5,6,7,8-tetrahydro-[1,3]dioxolo-[4,5-g]isoquinoline (Red-Br-nos), induces reactive oxygen species (ROS)-mediated autophagy and caspase-independent death in prostate cancer PC-3 cells. Herein, we show that Red-Br-nos induces ROS-dependent DNA damage that resulted in high-grade CA and SM in PC-3 cells. Unlike doxorubicin, which causes double-stranded DNA breaks and chronic G2 arrest accompanied by ‘templated' CA, Red-Br-nos-mediated DNA damage elicits de novo CA during a transient S/G2 stall, followed by checkpoint abrogation and mitotic entry to form aberrant mitotic figures with supernumerary spindle poles. Attenuation of multipolar phenotype in the presence of tiron, a ROS inhibitor, indicated that ROS-mediated DNA damage was partly responsible for driving CA and SM. Although a few cells (∼5%) yielded to aberrant cytokinesis following an ‘anaphase catastrophe', most mitotically arrested cells (∼70%) succumbed to ‘metaphase catastrophe,' which was caspase-independent. This report is the first documentation of rapid de novo centrosome formation in the presence of parent centrosome by a noscapinoid family member, which triggers death-inducing SM via a unique mechanism that distinguishes it from other ROS-inducers, conventional DNA-damaging agents, as well as other microtubule-binding drugs.

Background: Centrosome amplification (CA) which refers to presence of supernumerary or abnormally large centrosomes drives tumor progression by promoting chromosomal instability and the generation of aggressive tumor clones. Although the role of CA in cancer progression is well-defined, no studies have yet discussed how CA is induced in tumor cells. We report here that intra-tumoral hypoxia, which is considered one of the major contributors to intratumor heterogeneity, induces CA via HIF-1α. Methods: We first immunohistochemically labeled 24 breast carcinoma and uninvolved adjacent normal tissue samples for HIF-1α and calculated weighted indices (WIs) for nuclear HIF-1α. Adjacent serial sections from the same tumors were also immunofluorescently labeled for γ-tubulin and CA was calculated. Using public microarray datasets (Kao dataset, n=327), we investigated whether centrosomal gene expression is enriched in breast tumors characterized by a hypoxia gene expression signature. Finally, to determine the role of hypoxia in CA induction we exposed cultured TNBC cells (MDA-MB-231 and MDA-MB-468) to hypoxia and overexpressed (OE) and knocked out (KO) HIF-1α in TNBC cells and quantitated CA. Additionally, to discern the biological pathway through which HIF-1α induces CA we performed ChIP assay and in silico analyses to identify the possible targets of HIF-1α. Results: A strong positive correlation between nuclear HIF-1α WI and CA was found in breast tumor samples (Spearman's rho p=0.722, p<0.001). In addition, we found that higher nuclear HIF-1α was associated with worse overall survival (p=0.041; HR=1.03). Our in silico findings suggest that breast tumors with high expression of hypoxia-associated genes exhibited higher expression of centrosomal genes than breast tumors with low expression of hypoxia-associated genes. In addition, cells cultured in hypoxic conditions exhibited ˜1.5 fold higher (p<0.05) CA when compared to the cells cultured in normoxic conditions. Interestingly level of CA decreased when HIF-1α KO TNBC cells were exposed to hypoxia and it increased when HIF-1α OE TNBC cells were culture in normoxic conditions. Furthermore, we discovered that HIF-1α induced CA by directly regulating the expression of Plk4 which was confirmed by performing ChIP assay. Our results indicated HIF-1α interaction with the motif in the PLK4 promoter from genomic DNA of MDA-MB 231 cells under hypoxic conditions, was significantly (p=0.04) higher when compared with the cells cultured under normoxic conditions. Plk4 mRNA expression was assessed using the online BC gene expression data sets (n=25). We found significantly higher expression of Plk4 in TNBC (n=374) when compared with non-TNBC (n=4098) and it was associated with poor overall survival (HR=1.76; p=0.054) in TNBC. Conclusion: Collectively our findings suggest that hypoxia drives CA in TNBC via HIF-1α and contribute to poor outcomes. Thus, determination of CA and HIF-1α can help risk stratification in TNBC patients for more personalized treatments. Citation Format: Mittal K, Choi DH, Maganti N, Ogden A, Melton BD, Kaur J, Gupta MV, Jonsdottir K, Janseen EAM, Aleskandarany MA, Rakha EA, Rida PCG, Aneja R. Hypoxia induced centrosome amplification via HIF-1α/Plk4 signaling axis associates with poorer overall survival in TNBC [abstract]. In: Proceedings of the 2017 San Antonio Breast Cancer Symposium; 2017 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2018;78(4 Suppl):Abstract nr P1-01-23.