Human neuroblastoma (IMR-32), human cervical cancer (HeLa), and human lens epithelial (SRA01) cells were procured from American Type Culture Collection (ATCC) resource and grown in DMEM (Dulbecco’s Modified Essential Medium) containing 10% FBS (Fetal Bovine Serum) in the presence of penicillin (100 U/ml) and streptomycin (50 μg/ml) at 37˚C in a humidified incubator with 5% CO₂ supply. Cells (2 × 10⁵) were grown in a 6-well culture dish (TPP, Switzerland) on 22 × 22 mm cover glasses (Fisher Scientific, USA) were treated with different drugs. The stock solutions of GA, colcemid and MG132 were prepared in dimethyl sulfoxide (DMSO) and the stock solution of LiCl was prepared in sterile double distilled water.

Cells (1 × 10⁴) were grown in a 96-well plate (TPP, Switzerland), treated with colcemid or GA or their combination for 24 h. At the end of each time interval, 20 µl of MTT solution (5 mg/ml) was added to each well and incubated at culture conditions for 3 h. The MTT crystals in each well were solubilized with 200 µl of DMSO, and the absorbance was measured at 560 and 670 nm.

Cells scraped into phosphate buffer saline (PBS, pH 7.4) were collected by centrifugation at 4000 rpm for 5 min at room temperature and fixed using 70% ethanol for 2 h. Cells were stained with 50 μg/ml propidium iodide (PI) containing 5 μg RNase A for 30 min in dark at room temperature. Acquisition was performed using FACS (FACSCalibur, Becton Dickenson, USA) using FL2 parameter, and data analyzed using Cell-Quest Pro v5.2 software.

Nuclear staining was performed with DAPI (1 μg/ml) after fixing cells in 4% paraformaldehyde for 15 min and permeabilizing with 0.2% Triton X-100 in PBS (140 mM NaCl, 2.7 mM KCl, 10 mM Na₂HPO₄ and 1.8 mM K₂HPO₄, pH 7.3) for 10 min and observed under the microscope (Axiovert 200 microscope (20× magnification; Carl Zeiss; Germany) attached to a 35 mm CCD camera.

Total RNA was isolated by TRIzol (Sigma-Aldrich, USA) and the concentration was estimated using nanodrop 2000 (Thermo Scientific, USA) and RNA quality assessed by running on agarose gel. First strand cDNA was prepared from 1 μg total RNA in a reaction mixture provided in the cDNA synthesis kit (Invitrogen, USA). To check bubR1 expression using PCR, the following primers were used. Forward primer, 5’-TCTGGGGCTTGGGGAGTCA-3’ and reverse primer, 5’-GGGGTTGGCAGGCTTTTGG-3’. The PCR reaction was performed for GAPDH with the following primers, forward: 5’-TGAAGGTCGGTGTGAACGGATTTG-3’, reverse: 5’-TGATGGCATGGACTGTGGTCATGA-3’ (NCBI accession, NM_002046).

The siRNA for bubR1 (Accession No: NM_001211.5) was designed using BLOCK-iT^TM RNAi Designer software from Life Technologies, Invitrogen. The custom synthesized bubR1 siRNA (3 nos) are, Oligo-1: 5’-GCAGAGAAGAGAGCAGAAA-3’, Oligo-2:, 5’-GCCTCTGCAGAGTTGTGTA-3’, Oligo-3: 5’-GCTTGCAGCAAGAAATGAAT-3’. Each bubR1 siRNA (2 μg/well) in a 12-well Nunc plate was used for transfection into IMR-32 cells using lipofectamine LTX plus reagent (Invitrogen, USA) as per manufacturer’s instructions.

Cells washed with PBS were scraped into HEPES lysis buffer (10 mM HEPES pH 7.4, 10 mM NaCl, 0.1 mM EDTA, 0.1 mM EGTA, 0.1% Triton X-100, 0.1 mM DTT and 1× protease inhibitor cocktail). After incubation at 4˚C for 1 h with gentle agitation, cell extracts were clarified by centrifugation at 13000 rpm for 20 min at 4˚C (Eppendorf Table Top Centrifuge 5417R). Protein concentration was determined using Bradford reagent (BioRad, USA), 20 μg total cell lysate was run on 8% SDS-PAGE and transferred onto nitrocellulose membrane (Amersham Biosciences, USA). Antibodies, Hsp90α/β (SC-7947), Hsp90a (SC-8262), Hsp90β (SC-1057), bubR1 (SC-16195), phosphorylated GSK-3β (SC-33039), GSK3β (SC-7291), cdk1 (SC-747), cdk2 (SC-163), cdk6 (SC- 7180), PCNA (SC-7907), p53 (SC-6243), cyclin B (SC- 752), ERK1 (SC-93), ERK2 (SC-154), phosphorylated ERK1/2 (SC-101760), grp94 (SC-11402), cyclin D1 (SC-8396), and GAPDH (SC-25778) were purchased from Santa Cruz Biotechnology Inc. (USA). Antibodies to β-catenin (Cat No. 610154) were purchased from BD Transduction Laboratories (USA). Immunodetection was performed using horseradish peroxidase (HRPO) conjugated goat antimouse (SC-2005), goat antirabbit (SC- 2030), rabbit antigoat (SC-2922) IgG antibodies (Santa Cruz, USA), and the luminescence was detected by Roche chemiluminescence detection kit (Roche, USA).

Cells were first fixed in 4% paraformaldehyde for 10 min, permeabilized with 0.2% Triton X-100 for 5 min, blocked with 2% BSA for 1 h. Cells were then incubated with primary antibodies, bubR1 (1:100 dilution in blocking buffer), aurk1, aurk2, plk1 and pericentrin (1:200 dilution in blocking buffer) for 1 h and subsequently incubated either with FITC conjugated IgG secondary antibodies. Nucleus was stained with DAPI (4’,6-diamidino- 2-phenylindole; 50 nM, VECTASHIELD, Vector Labs, USA) and cover glasses were sealed to glass slides and observed under the laser scanning confocal imaging microscope (100× magnification, Leica TCS SP5).

Data shown are mean ± SD of three independent experiments. Statistical significance was calculated by student’s t-test. Significance values represented are, *, p < 0.05; **, p < 0.01; ***, p < 0.001.

To examine the effect of colcemid and GA on IMR-32, HeLa, and SRA01, cells were treated with different concentrations of colcemid (2.0, 4.0, 8.0, 10.0, 15.0, and 20.0 μM) and GA (0.5, 1.0, 2.0, 4.0, 6.0, and 8.0 μM) for 24 h and analyzed for cytotoxicity. There was a concentration dependent increase in dead cells with colcemid treatment in all the three cell types examined. However, GA treatment induced cytotoxicity was observed in a dose-dependent manner in IMR-32 (Figure 1(a)) and HeLa (Figure 1(b)) cells compared to SRA01 (Figure 1(c)) cells. Since the objective of the study is to sensitize tumor cells to low concentrations of drugs in combination, colcemid and GA at concentrations 10.0 μM and 2.0 μM were chosen respectively.

First, we examined the combination effect of drugs on cell cycle in comparison to individual treatments against IMR-32, HeLa, and SRA01 cells in a 24 h treatment. In IMR-32, the combination treatment has resulted in the accumulation of 67.4% (p < 0.001) G2/M cells while colcemid and GA treatments showed 71% (p < 0.001) and 44.8 (p < 0.01) G2/M cells respectively (Figure 1(d)). In HeLa, the combination treatment has resulted in the accumulation of 70% (p < 0.001) G2/M cells while colcemid and GA treatments showed 46% (p < 0.001) and 34% (p < 0.001) G2/M cells respectively (Figure 1(e)). In SRA01, the combination treatment did not show any significant effect on cell cycle, especially in comparison with IMR-32 and HeLa cells (Figure 1(f)). Consistently, both IMR-32 and HeLa exhibited similar response to the combination drug treatment.

To examine whether combination treatment promote cytotoxicity on a prolonged treatment, tumor cells were continuously treated for 72 h and the percent cytotoxicity was measured. In IMR-32, prolonged treatment has increased the cytotoxicity both with colcemid (44.9%, p < 0.01) and GA (53.5%, p < 0.01), which is 34.9% and 48.5% increase respectively compared to 24 h treatment. In HeLa also prolonged treatment increased the cytotoxicity with colcemid (30%, p < 0.01) and GA (49%, p < 0.01), which is 20% and 29% increase respectively compared to 24 h treatment. Interestingly, the combination treatment showed enhanced cytotoxicity in IMR-32 (85%, p < 0.001) and HeLa (67.7%, p < 0.001) cells, which is 75% and 57% respectively compared to their 24 h treatments (Figure 2(a)). Since some of the chemotherapeutic drugs show reversible effects after their withdrawal, we examined the cytotoxic effects of drugs after 24 h treatments followed by 24 h recovery without any treatment. In IMR-32, we observed increased cytotoxicity with colcemid (24%, NS) and with GA (32.7%, p < 0.05), which is 19% and 22.7% increase respectively compared to 24 h treatment. In HeLa also we observed increased cytotoxicity

with colcemid (26%, p < 0.01) and with GA (27%, p < 0.05), which is 16% and 22% increase respectively compared to 24 h treatment. To our surprise, we observed significant increase in cytotoxicity in the combination treatment for IMR-32 (59.33%, p < 0.001) and HeLa (66.3%, p < 0.001) cells, which is 49% and 56% increase respectively in comparison with 24 treatment (Figure 2(b)). These results suggest that tumor cells exhibit atypical response to the combination treatment. Although both IMR-32 and HeLa cells showed similar response to the combination treatments, after observing enhanced cytotoxicity in IMR-32 (85%) compared to HeLa (67.7%) on prolonged treatments, subsequent experiments were performed using IMR-32 cells.

To examine whether the combination treatment interferes with biochemical pathways, cell lysates from IMR-32 cells after 24 h treatment were subjected to immunoblot analysis with antibodies to proliferative enzymes and cell cycle regulators. The combination treatment resulted in decreasing the levels of PCNA and cyclin E. Increased p53 levels correlating with decreased cyclin E and cdk1, cdk2, and cdk6, and their upstream kinase ERK1/2 activation suggest severed cell cycle machinery, especially the mitotic machinery. Concurrently, we observed a significant increase in the APC/C inhibitor kinase, bubR1. Individual drugs although showed varied biochemical response, the effect induced by the combination treatment was apparently significant suggesting severed cell mitosis. Further, decreased procaspase-3 levels in the combination treatment suggest mitotic damage lead to cell death activation (Figure 3(a)). These results were in agreement with G2/M cell cycle arrest in 24 treatment (Figure 1) and enhanced cytotoxicity in a 72 h treatment (Figure 2). Next, we examined for bubR1 expression in SRA01 cells and compared with HeLa cells. While a moderate, but significant increase in bubR1 levels was observed in HeLa cells in response to the combination treatment, SRA01 cells did not show any bubR1 expression (Figure 3(b)) suggesting tumor selective effects of combination treatment. After observing increased levels of p53 protein in our immunoblot experiments, we want to examine whether p53 increase relate to increased transcription. Also to examine whether increased p53 levels correlate with p21 transcription, RT-PCR analysis was performed in control and drugs treated cells. Although there is a complete decrease in p53 transcription in the GA and colcemid treatments, but not on recovery, combination treatment showed significant increase in its induced transcription. The increased p53 transcription, however, did not correlate with increased p21 levels suggesting its unconventional role in the combination treatment (Figrue 3(c)).

Since bubR1 nuclear location decides the metaphase progression [22], its localization in 24 h combination treated IMR-32 cells were examined by immunofluorescence analysis. While individual drug treatments, colcemid and GA showed lack of bubR1 nuclear localization, the combination treatment showed significant increase in its nuclear localization (Figure 4(a)). The increase in bubR1 nuclear localization therefore suggests its possible interference in mitosis progression. Analysis of DAPI stained IMR-32 cell nuclei after 48 h colcemid, GA, and combination treatments showed defects in mitotic progression compared to control cells. The colcemid treatment showed unsuccessful metaphase to anaphase transition, which may be due to its anti-tubulin effects affecting the spindle assembly. Colcemid treated cells prevailed over mitotic arrest however, resulted in unequal segregation of sister chromatids in the anaphase and telophase transition of mitosis. GA treatment showed prolonged metaphase arrest suggesting delayed or inhibited metaphase to anaphase transition. Interestingly, the combination treatment showed loss of chromosomal architecture, which may be due to prolonged metaphase block by the GA, and thus, relate to the enhanced nuclear bubR1 levels (Figure 4(b)).

If bubR1 accumulation is considered to be the major cause for defect in metaphase to anaphase transition, siRNA knockdown of bubR1 should decrease G2/M cells. To investigate this, IMR-32 cells after post bubR1 siRNA transient transfection (24 h) were subjected to 24 h combination treatment and examined for bubR1 transcript levels as well as protein levels in RT-PCR and immunoblot analyses respectively. A decrease in bubR1 mRNA transcript has correlated with decreased bubR1 protein levels compared to Figure 3(a) suggest functional siRNA. When siRNA transfected cells were challenged with GA for 24 h, and analyzed in FACS, we observed a 50% decrease in G2/M cells compared to non-transfected controls (Figure 5(b)). These experiments established a correlation between bubR1 stability with functional Hsp90 and suggest Hsp90 involvement in bubR1 quality control. To investigate this correlation, IMR-32 cells were treated with combination drugs with and without MG132 for 24 h and the cell lysates were immunoblotted for bubR1 expression. An increase in bubR1 levels (1.4 folds) upon MG132 co-treatment in comparison to combination treatment suggests Hsp90-dependent proteasome-mediated degradation of bubR1.

Malignant cells may activate the canonical pathway through gain-of-function mutations in β-catenin [23] or by noncanonical pathway through Wnt/receptor signaling [24]. Since β-catenin stable expression induces G2/M arrest that can stimulate cells to apoptosis [25], we have examined for β-catenin expression and its upstream regulator GSK3β activities and correlated with the cell proliferative marker cyclin D1. Further, β-catenin is implicated in cancer due to its ability to induce the transcription of c-myc and cyclin D1 [26]. Since β-catenin is present in mitotic cells and localized to centrosomes [27], we examined whether β-catenin expression is influenced by bubR1-mediated mitotic arrest. Primarily, we observed increased β-catenin levels in GA treatment compared to colcemid and its combination treatments (Figure 6(a)). Akt happens to be the primary regulator of GSK3β phosphorylation, where Hsp90 controls Akt activity in tumor cells [28]. As shown in Figure 6(a), LiCl inhibited GSK3β phosphorylation and thereby decreased the accumulation of β-catenin however, only in the combination treatment. Since β-catenin expression correlates with cyclin D1, we examined for cyclin D1 levels by immunoblot analysis and found a correlation between decreased β-catenin levels with decreased cyclin D1 levels. The inability of MG132 treatment in retaining β-catenin or cyclin D1 in the combination treatment suggested proteasome-independent, but Akt-GSK3β phosphorylation-dependent regulation (Figure 6(b)). These findings suggested that the combination treatment in addition to targeting mitosis, also targets the non-canonical wnt signaling.

Considering that chromosomal defects can affect the downstream mitotic kinases, we examined the distribution of aurora kinases, aurk1 [29] and aurk2, and pololike kinase, plk1 by cytoimmunofluorescence [30]. The untreated cells showed polar distribution of aurk1 and its association with spindle poles. However, in the treatment, first there is a decrease in aurk1 expression levels, which is associated defective bipolar distribution (Figure 7(a)). In untreated cells, we observed aurk2 localizing to chromosome arms and as the mitosis progresses, it localized to the centromere and midbody. In case of combination treatment, along with a significant decrease in its expression, aurk2 was unable to localize to the centromere and midbody (Figure 7(b)). The plk1 is implicated in mitotic entry and exit, and also in cytokinesis, however, in association with aurk1 and aurk2 [30]. In untreated cells, plk1 is distributed to centromere during mitosis. Its inappropriate distribution in the treatment has correlated with lack of interaction with aurks and defective mitosis (Figure 7(c)). Centrosome is composed of centrioles and pericentriolar material (PCM), where pericentrin is a key component of PCM that acts as a microtubule organization center [31]. Untreated cells showed pericentrin enrichment and alignment with centrosome. In the treatment, both its enrichment and alignment with the centrosome were absent (Figure 7(d)). These experiments suggested that the early defects in metaphase are associated with a decrease, defective distribution, and functions of mitotic kinases.