MLN8054, a small-molecule inhibitor of Aurora A, causes spindle pole and chromosome congression defects leading to aneuploidy

Abstract

Aurora A kinase plays an essential role in the proper assembly and function of the mitotic spindle, as its perturbation causes defects in centrosome separation, spindle pole organization, and chromosome congression. Moreover, Aurora A disruption leads to cell death via a mechanism that involves aneuploidy generation. However, the link between the immediate functional consequences of Aurora A inhibition and the development of aneuploidy is not clearly defined. In this study, we delineate the sequence of events that lead to aneuploidy following Aurora A inhibition using MLN8054, a selective Aurora A small-molecule inhibitor. Human tumor cells treated with MLN8054 show a high incidence of abnormal mitotic spindles, often with unseparated centrosomes. Although these spindle defects result in mitotic delays, cells ultimately divide at a frequency near that of untreated cells. We show that many of the spindles in the dividing cells are bipolar, although they lack centrosomes at one or more spindle poles. MLN8054-treated cells frequently show alignment defects during metaphase, lagging chromosomes in anaphase, and chromatin bridges during telophase. Consistent with the chromosome segregation defects, cells treated with MLN8054 develop aneuploidy over time. Taken together, these results suggest that Aurora A inhibition kills tumor cells through the development of deleterious aneuploidy.

Figures

FIG. 1.

Aurora A inhibition or ablation results in the formation of abnormal mitotic spindles. Abnormal spindles were defined as those that did not display canonical bipolar spindle formation, as defined by the existence of a clearly visible metaphase plate straddled by undisrupted radial arrays of microtubules emanating from opposite poles. (A and B) HCT-116 cells were treated with DMSO or 0.25 μM MLN8054 for 5, 24, and 48 h. A final concentration of 100 nM bortezomib was added for the final 2 h to arrest mitotic cells at the metaphase-to-anaphase transition. (A) Representative immunofluorescent images of mitotic spindles obtained from cells stained with anti-α-tubulin rabbit antibody and Hoechst stain (DNA). Four example MLN8054-treated mitotic cells are shown. The scale bars represent 10 μm. (B) The percentages of abnormal spindles were quantified 5, 24, and 48 h after treatment with DMSO or MLN8054. The P values demonstrate a statistical difference between control- and Aurora MLN8054-treated cells in the incidence of abnormal spindles at all time points. (C and D) HCT-116 cells were transfected with control (luciferase [Luc]) or Aurora A siRNAs for 5, 24, or 48 h. A 100 nM concentration of bortezomib was added for the final 2 h to arrest mitotic cells at the metaphase-to-anaphase transition. (C) Western blot probed for Aurora A and β-actin, which served as a protein-loading control. (D) The percentages of abnormal spindles were quantified 5, 24, and 48 h after siRNA transfection. The P values demonstrate a statistical difference between control- and Aurora A siRNA-transfected cells in the incidence of abnormal spindles at all time points.

FIG. 2.

Aurora A inhibition delays mitotic progression, but cells ultimately divide at a high frequency. HCT-116 cells constitutively expressing EGFP-α-tubulin and mmRed-histone H2B were treated with DMSO (A) or 0.25 μM MLN8054 (B). Images were captured every 5 min using a time-lapse imaging system. Two cells were tracked for each treatment, A and B. Daughter cells from the first division are labeled A1, A2, B1, or B2, and daughter cells from the second division are labeled A1a, A1b, B1a, or B1b. Yellow letters represent the first division, and red letters represent the second division. The number in the lower corner of each image indicates elapsed time (h:min). The times to progress from prophase to telophase over the first (C) and second (D) divisions and from prophase to prophase (E) as well as the percentages of cells that divided (F) were quantified in control- and MLN8054-treated cells.

FIG. 3.

Aurora A inhibition results in centrosome and spindle pole defects. HCT-116 cells were treated with DMSO (control) or 0.25 μM MLN8054 for 5 h. Two example MLN8054-treated cells are shown in each panel. A 100 nM concentration of bortezomib was added for the final 2 h to arrest mitotic cells at the metaphase-to-anaphase transition. (A) Representative immunofluorescent images of centrosomes and spindle poles from cells stained with anti-pericentrin rabbit antibody, anti-NuMA mouse antibody, and Hoechst stain (DNA). (B) Representative immunofluorescent images of centrosomes and spindle poles from cells stained with anti-γ-tubulin rabbit antibody, anti-NuMA mouse antibody, and Hoechst stain (DNA). (C) Representative immunofluorescent images of mitotic spindles and spindle poles from cells stained with anti-α-tubulin rabbit antibody, anti-NuMA mouse antibody, and Hoechst stain (DNA). (D) Representative immunofluorescent images of mitotic spindles and centrosomes stained with anti-α-tubulin mouse antibody, anti-pericentrin rabbit antibody, and Hoechst stain (DNA).

FIG. 4.

Quantification of centrosomes and spindle poles in control- and MLN8054-treated cells. HCT-116 cells were treated with DMSO (control) or 0.25 μM MLN8054 for 5 h. A 100 nM concentration of bortezomib was added for the final 2 h to arrest mitotic cells at the metaphase-to-anaphase transition. (A) The percentage of mitotic cells containing 1, 2, 3, or 4 centrosomes was determined by counting the number of pericentrin-immunopositive spots in mitotic cells. (B) The percentage of mitotic cells containing 1, 2, 3, or 4 spindle poles was determined by counting the number of NuMA-immunopositive spots in mitotic cells. Average values ± standard deviations and median values are shown. The P values demonstrate a statistical difference between control- and MLN8054-treated cells in centrosomes and spindle poles.

FIG. 5.

Centrosome amplification due to Aurora A inhibition increases over time. HCT-116 cells were treated with DMSO (control) or 0.25 μM MLN8054 for 5, 24, and 48 h. A 100 nM concentration of bortezomib was added for the final 2 h to arrest mitotic cells at the metaphase-to-anaphase transition. (A) Representative immunofluorescent images of mitotic spindles and centrosomes stained with anti-α-tubulin mouse antibody, anti-pericentrin rabbit antibody, and Hoechst stain (DNA). Two example MLN8054-treated cells are shown. The scale bars represent 5 μm. (B) Percentage of mitotic cells containing three or more centrosomes. The P values demonstrate a statistical difference between control- and MLN8054-treated cells in the percentages of cells with amplified centrosomes 24 and 48 h after treatment.

FIG. 6.

Aurora A inhibition disrupts normal chromosome segregation during mitosis. HCT-116 cells were treated with DMSO (control) or 0.25 μM MLN8054 for 5 and 24 h. Representative immunofluorescent images of metaphase, anaphase, and telophase cells stained with anti-α-tubulin rabbit antibody and Hoechst stain (DNA) are shown. The yellow arrows indicate misaligned or missegregated chromosomes at each mitotic phase. The scale bars represent 5 μm.

FIG. 7.

Aurora A inhibition leads to aneuploidy. HCT-116 cells were treated with DMSO (control) or 0.25 μM MLN8054 for 24, 48, or 72 h. A 5 μM concentration of aphidicholin was added for the final 16 h to arrest cells in the G1 phase of the cell cycle. (A) Representative immunofluorescent images of interphase cells stained with anti-α-tubulin rabbit antibody (green) and Hoechst stain (DNA, blue). Three example MLN8054-treated cells are shown. The yellow arrows indicate binucleated nuclei of various sizes. The scale bars represent 10 μm. (B) Percentage of interphase cells containing grossly abnormal nuclei. The P values demonstrate statistical differences between control- and MLN8054-treated cells in the percentages of cells with abnormal nuclei 24, 48, and 72 h after treatment. (C) Representative immunofluorescent images of control- and MLN8054-treated cells stained with anti-α-tubulin rabbit antibody (green), nuclear ANA-centromere autoantibody (kinetochores, red), and Hoechst stain (DNA, blue). (D) Number of kinetochores per cell in control- and MLN8054-treated cells 24, 48, and 72 h after treatment.

FIG. 8.

Proposed model of the mechanism of action for MLN8054, a small-molecule inhibitor of Aurora A. Chromosomes are represented in blue, kinetochores in red, mitotic spindles in green, and centrosomes in yellow. The inhibition of Aurora A leads to the formation of monopolar, bipolar, and multipolar spindles, often with unseparated centrosomes. Monopolar and multipolar spindles may resolve to form bipolar spindles or may die directly from defective mitoses. Although mitotic spindle defects lead to mitotic delays, many cells divide, at an incidence near that of untreated cells. Aurora A inhibition leads to chromosome congression defects at metaphase, lagging chromosomes in anaphase, and telophase chromatin bridges. Cells treated with MLN8054 ultimately become aneuploid, which leads to death. The solid arrows within this model represent steps supported by evidence presented in this work.