MLN8054, a Small-Molecule Inhibitor of Aurora A, Causes Spindle Pole and Chromosome Congression Defects Leading to Aneuploidy^{▿ † (V体育官网入口)}

Cell culture and compound treatment. (VSports注册入口)

HCT-116, H460, and DLD1 human tumor cells were obtained from the American Type Culture Collection and maintained according to the distributor's recommendations. MLN8054 was diluted in distilled water and added to the cell culture medium at a final concentration of 0.25 μM. Bortezomib (Velcade; Millennium Pharmaceuticals, Inc.) was diluted in dimethyl sulfoxide (DMSO) and added to the cell culture medium at a final concentration of 100 nM.

Immunofluorescence staining. (V体育ios版)

HCT-116 cells grown on glass coverslips and treated as described above were fixed with 4% paraformaldehyde (Electron Microscopy Sciences) and then permeabilized with 0.5% Triton X-100 (Sigma) in phosphate-buffered saline (PBS). Blocking reagent (Roche) was added to cells for 30 min, followed by a 1-h incubation with the primary antibody in blocking reagent. The cells were washed in PBS and in PBS containing 0.05% Tween, followed by the addition of the secondary antibody in blocking reagent for 1 h. The cells were then washed with PBS and incubated with Hoechst 33321 in PBS (1:10,000; Molecular Probes) for 5 min. The cells were washed twice in PBS and mounted on glass slides with Fluoromount G (Electron Microscopy Sciences). The primary antibodies used in these studies included anti-α-tubulin mouse antibody (1:1,000; Sigma), anti-α-tubulin rabbit antibody (1:1,000; Abcam), anti-IAK1 (I_pI1 and Aurora-related kinase 1) mouse antibody (1:250; BD Transduction Laboratories), anti-pericentrin rabbit antibody (1:500; Abcam), anti-nuclear and mitotic apparatus protein (anti-NuMA) mouse antibody (1:250; Calbiochem), anti-γ-tubulin rabbit antibody (1:250; Sigma), and nuclear ANA (antinuclear antigen)-centromere autoantibody (CREST) (1:1,000; Cortex Biochem) to stain kinetochores. The secondary antibodies used in these studies included Alexa 488-conjugated goat anti-mouse immunoglobulin G (IgG) (1:250; Molecular Probes), Alexa 488-conjugated goat anti-rabbit IgG (1:250; Molecular Probes), Alexa 594-conjugated goat anti-mouse IgG (1:250; Molecular Probes) and Alexa 594-conjugated goat anti-rabbit IgG (1:250; Molecular Probes). Immunofluorescently labeled cells were visualized with 63× or 100× objectives and an LSM 5 Pa laser-scanning confocal microscope (Zeiss).

Abnormal mitotic spindle, centrosome, and spindle pole quantification.

The percentage of abnormal mitotic spindles was determined by evaluating mitotic spindle architecture from the immunofluorescently stained images. 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. Centrosomes and spindle poles were quantified by determining the number of distinct pericentrin- and NuMA-immunopositive spots, respectively, per mitotic cell.

Aneuploidy quantification.

HCT-116 cells were treated with DMSO or 0.25 μM MLN8054 for 24, 48, or 72 h. Aphidicholin (5 nM) was added for the final 16 h to arrest cells in the G₁ cell cycle phase. Kinetochores, α-tubulin, and DNA were immunofluorescently labeled as described above and imaged with an E800 microscope (Nikon Instruments) equipped with an automated XYZ stage (Prior Scientific), a filter wheel (Sutter Instruments), and a Cool Snap HQ camera (Roper Scientific) controlled by MetaMorph software (Molecular Devices). Z sections were acquired at 0.1-μm intervals with a 60× objective. Z-stack images were processed using MetaMorph software and compressed to single best-fit images. Grossly abnormal interphase nuclei were characterized as those that contained more than one distinct nucleus per cell or were dramatically misshapen. The number of kinetochores per cell was quantified by automated image processing using MetaMorph software.

"VSports" Video microscopy.

HCT-116 cells constitutively expressing enhanced green fluorescent protein (EGFP)-conjugated α-tubulin and mmRed-conjugated histone H2B were treated with DMSO or 0.25 μM MLN8054. Upon compound treatment, cells were added to a humidified chamber at 37°C with 5% CO₂ attached to an Eclipse TE2000-U microscope (Nikon Instruments) equipped with an automated XYZ stage (Prior Scientific), a filter wheel (Sutter Instruments), and an Orca-ER camera (Hamamatsu) controlled by MetaMorph software. Cells were imaged every 5 min for over 24 h. Time-lapse images were processed using MetaMorph software. Prophase onset was defined by the point of centrosome separation, which we took to occur at the time of separation of clustered EGFP-conjugated α-tubulin (centrosomes) that occurred subsequent to mmRed-conjugated histone H2B condensation. Telophase onset was defined by the appearance of a midbody, as measured by tracking EGFP-conjugated α-tubulin.

RNAi.

Suspended HCT-116 tumor cells (2 × 10⁵) were transfected with 50 nM of GL2 (sense, 5′ CGUACGCGGAAUACUUCGA 3′) or Aurora A (sense, 5′ AUGCCCUGUCUUACUGUCA 3′) siRNAs (Dharmacon) using 2 μl Lipofectamine 2000 reagent (Invitrogen) and 98 μl Opti-mem I (Invitrogen) on 6-well BioCoat cell culture plates (BD Biosciences) containing poly-d-lysine-coated coverslips. The cells were harvested 5, 24, and 48 h after transfection and processed for immunofluorescence and Western blot analysis.

Western blots.

The cells were lysed in radioimmunoprecipitation assay buffer (150 mM NaCl, 1% NP-40, 0.5% sodium deoxycholate, 0.1% sodium dodecyl sulfate [SDS], 50 mM Tris-HCl [pH 7.5]) containing complete protease inhibitor cocktail III (Boehringer-Ingelheim). Lysates were denatured by heating to 100°C in NuPAGE SDS sample buffer (Invitrogen) containing 0.1% β-mercaptoethanol. Proteins were resolved by SDS-polyacrylamide gel electrophoresis, Western blotted, and detected with an ECL Western blotting detection system (Amersham Biosciences). The primary antibodies used for Western blotting included anti-IAK1/Aurora A kinase mouse antibody (1:250) and anti-β-actin mouse antibody (1:10,000; Abcam). The secondary antibody used was goat anti-mouse IgG conjugated to horseradish peroxidase (1:10,000; Southern Biotechnologies).

Statistical analysis.

The statistical significance (P value) for various experiments was assessed either by using a one-sided P value for the Poisson distribution or by using the two-sided t test.

MLN8054 and Aurora A RNAi lead to defects in mitotic spindle formation. (VSports app下载)

MLN8054 is an ATP-competitive selective Aurora A inhibitor that completely inhibits Aurora A in cells at 0.25 μM without affecting Aurora B activity (35). The effect of Aurora A inhibition via MLN8054 on spindle organization and chromosome alignment was examined by using immunofluorescence staining (Fig. 1). HCT-116 cells were exposed to 0.25 μM MLN8054 for 5, 24, or 48 h, and the proteasome inhibitor bortezomib was added 2 h before fixation to trap mitotic cells immediately prior to anaphase onset so that prometaphase cells, with their accompanying incomplete spindle formation and chromosome alignment, would not be construed as being abnormal. MLN8054-treated cells showed a variety of spindle organization defects after 24 h, such as monopolarity, multipolarity, and severe chromosome alignment defects (Fig. 1A). Similar observations were made for MLN8054 HCT-116 cells in the absence of bortezomib (35). The percentages of cells with abnormally formed mitotic spindles were quantified at 5, 24, and 48 h following MLN8054 treatment (Fig. 1B). MLN8054 caused the formation of abnormal spindles at very high frequencies, ranging from 77 to 95% over the different exposure times. Control samples at each time displayed only 3% abnormal mitotic spindles.

Formation of abnormal mitotic spindles was also quantified in cells depleted of Aurora A by RNAi. Aurora A depletion was partial 5 h after siRNA transfection and complete 24 and 48 h after siRNA transfection (Fig. 1C). The percentages of abnormal spindles formed in HCT-116 cells upon transfection with Aurora A siRNA for 24 and 48 h were similar to those observed with MLN8054 treatment (Fig. 1D). Far fewer abnormal spindles were observed at 5 h, consistent with the partial Aurora A depletion at this time point. These results demonstrate that the spindle defects observed in MLN8054-treated cells were similar to those obtained upon RNAi-mediated Aurora A depletion.

Cells treated with MLN8054 divide despite spindle abnormalities.

To visualize the cell cycle progression of cells treated with 0.25 μM MLN8054, we used HCT-116 cells constitutively expressing EGFP-α-tubulin and mmRed-histone H2B. Cells treated with DMSO or 0.25 μM MLN8054 were imaged every five minutes for over 24 h using time-lapse microscopy (see Videos S1 and S2 in the supplemental material). Examples of control- and MLN8054-treated cells progressing through mitosis are shown in Fig. 2A and B, respectively. All of the tracked control-treated cells underwent an apparently normal division (Fig. 2A). However, some cells treated with MLN8054 undergo an abnormal division, displaying missegregated chromosomes at division (Fig. 2B; 3:05 h, cell B1) and micronuclei formation (Fig. 2B; 25:25 h, cell B2a).

MLN8054 treatment prolonged mitosis, increasing the average time from prophase to telophase from 67 to 131 min for the first mitotic division following treatment (Fig. 2C). A similar delay was observed at the second mitotic division (Fig. 2E). There was also an increase in overall cell cycle time, from 23 to 31 h, although this difference did not achieve statistical significance. A large percentage (65.5%) of MLN8054-treated cells completed cytokinesis; this was slightly lower than the percentage of cells in the control sample that divided in the same time frame (75.5%). From the observation that approximately 82% of the spindles in an MLN8054-treated sample at 24 h are abnormal (Fig. 1B), we may infer that a minimum of 58% of cells presenting with abnormal spindles underwent division within this time frame. This result was further corroborated using two other cell types. DLD1 and H460 cells also divide in the presence of MLN8054, although some cells appear to die during or shortly after division (see Fig. S1 and S2 in the supplemental material). Collectively, these results suggest that cells lacking functional Aurora A are capable of dividing despite the presence of spindle organization defects.

Cells treated with MLN8054 are able to establish bipolar spindles in the absence of centrosome separation. (V体育2025版)

Aurora A plays an important role in centrosome maturation and separation (16, 21), and many of the MLN8054-treated cells observed by us at early time points have monopolar spindles. Despite this, MLN8054-treated cells were found to undergo division at a rate near that of control-treated cells. To understand this apparent paradox, we examined the centrosomes and spindle poles in HCT-116 cells treated with 0.25 μM MLN8054 for 5 h and bortezomib for the final 2 h. Cells were fixed and stained for centrosomes by using pericentrin and γ-tubulin and for spindle poles by using NuMA (Fig. 3). Control-treated cells at the metaphase-to-anaphase transition presented primarily with two centrosomes and two spindle poles per cell (Fig. 3A and B and 4). The MLN8054-treated cells frequently presented with only one centrosome per mitotic cell (Fig. 3A and B), consistent with the observations that Aurora A participates in centrosome maturation and separation. However, these cells frequently formed acentrosomal spindle poles (NuMA positive), forming two or more spindle poles per cell, where centrosomal markers were detected at only one of the poles (Fig. 3A and B). Moreover, radial arrays of microtubules emanated from all spindle poles, whether or not centrosomes were present (Fig. 3C and D). We note further that many of these bipolar spindles in MLN8054-treated cells show chromosome alignment defects. Given that these cells were treated with MLN8054 for only 5 h, the centrosome and spindle pole defects observed occurred within a single mitotic episode, and not subsequent to previous abnormal divisions.

To further establish the relationship between unseparated centrosomes and spindle poles in MLN8054-treated cells, centrosomes and spindle poles were quantified in mitotic cells from the above experiment (Fig. 4) as the number of distinct pericentrin- and NuMA-positive spots, respectively. As expected, nearly all control-treated cells presented with two centrosomes and two spindle poles per cell. Of the MLN8054-treated cells, 66% presented with only one centrosome, whereas only 11% contained only one spindle pole. The majority of MLN8054-treated cells formed more than one spindle pole. Of these, the majority were bipolar, but tri- and tetrapolar spindles were also observed. When centrosomes are present, they act as the dominant sites for microtubule nucleation. Thus, a cell with two separated centrosomes must contain at least two functional spindle poles, as defined by tubulin staining. Based on this premise, we may infer that 83% of the cells containing a single centrosome or unseparated centrosomes are capable of organizing one or more acentrosomal spindle poles.

Centrosome amplification in the presence of MLN8054 increases over time.

The overexpression of both wild-type and kinase-dead Aurora A constructs in Chinese hamster ovary cells induced centrosome amplification through defects in cell division (38). We thus sought to test the effect of Aurora A inhibition by MLN8054 on centrosome number in mitotic cells over time. The number of centrosomes was assessed by using pericentrin in HCT-116 cells treated with DMSO or 0.25 μM MLN8054 for 5, 24, or 48 h and bortezomib for the final 2 h. Control-treated cells primarily contained two centrosomes, whereas cells treated with MLN8054 for 24 h occasionally displayed three centrosomes per mitotic cell (Fig. 5A). The number of cells with more than two centrosomes rose steadily, from 0% at 5 h to 9% at 24 h and 14% at 48 h (Fig. 5B). The increase in centrosome amplification with time upon Aurora A inhibition using MLN8054 is consistent with a model in which centrosome amplification occurs due to defects in previous cell divisions.

MLN8054 induces chromosomal defects in mitosis, leading to aneuploidy.

Cells treated with MLN8054 for 5 and 24 h display chromosomal defects in all stages of mitosis, including chromosome alignment defects in metaphase, lagging chromosomes in anaphase, and chromatin bridges in telophase (Fig. 6). Given these defects, we explored the manifestation of aneuploidy in MLN8054-treated cells over time. HCT-116 cells were treated with 0.25 μM MLN8054 for 24, 48, or 72 h, and 5 μM aphidicholin was added 16 h before fixation to trap cells in the G₁ cell cycle phase. Synchronizing cells in G₁ provided a standardized means to evaluate nuclear integrity and DNA content after cells progressed through mitosis in the presence of MLN8054. Interphase nuclei from cells treated with MLN8054 displayed a variety of defects in form, including micronucleation, binucleation, and multinucleation (Fig. 7A). The frequency of these defects, collectively characterized as grossly abnormal interphase nuclei, increased with time, reaching a maximum at approximately 35% of total cells.

As a second, independent measure of aneuploidy, the number of kinetochores per cell was quantified in G₁ cells prepared as described above. In G₁, cells have one kinetochore per chromosome; therefore, measuring kinetochores is reflective of chromosome content. We deemed this approach superior to flow cytometry for the purpose of quantifying aneuploidy, as interpretation of DNA profiles generated by flow cytometry can be complicated by fragmented DNA derived from apoptotic cells. Cells were stained for kinetochores, α-tubulin, and DNA, and the numbers of kinetochores per cell were determined using serial-section confocal immunofluorescent microscopy and image processing (Fig. 7C). Histograms plotting the distribution of kinetochores in control- or MLN8054-treated cells were generated (Fig. 7D). The distribution of kinetochores in control-treated cells remained essentially unchanged over time. The distribution of kinetochores in MLN8054-treated cells at 24 h overlapped that of control-treated cells. However, at 48 and 72 h, there were large populations of cells that had dramatic increases in kinetochores per cell. In fact, the increased distribution in kinetochores per cell at 48 h was similar to previous findings demonstrating an increased distribution in the DNA content determined by flow cytometry 48 h after the addition of MLN8054 (35). Interestingly, there was a peak of cells with a complement of kinetochores approximately twofold more that of the untreated cells, suggesting that these cells may have failed to complete cytokinesis prior to exiting mitosis and doubled their DNA content in the subsequent S phase. This is consistent with the results of previous reports demonstrating that perturbation of Aurora A can lead to a low incidence of cytokinesis failures (36). There were not a significant number of cells with a kinetochore-per-cell distribution below the range for kinetochores per cell of the control-treated samples. This suggests that cells with a suboptimal complement of DNA cannot survive and is consistent with previous reports describing massive chromosomal loss leading directly to cell death (31).

Overall, the findings described in this study demonstrate that inhibition of Aurora A by using MLN8054 leads to chromosome segregation defects that, in turn, cause severe aneuploidy over time.