Background
Eukaryotic cells have developed stringent cell cycle controls to ensure mitosis occurs consistently error free. Cell cycle checkpoints have evolved to ensure the inheritance of undamaged DNA, and that each daughter cell receives the correct complement of chromosomes. Aberrant expression and function of proteins that regulate the mitotic spindle, and other cell cycle checkpoints can lead to aneuploidy and contribute to cancer progression [
1]. The Aurora family of evolutionarily conserved serine/threonine kinases regulates entry into mitosis, centrosome maturation and the mitotic spindle checkpoint [
2]. Mammalian genomes contain three members of this kinase family, Aurora-A, B and C. Aurora-A was first characterized in
Drosophila melanogaster where mutants exhibited defects in centrosome separation [
3]. Aurora-B is a chromosomal passenger protein that begins mitosis localized to the centromeres but at the onset of anaphase relocates to the spindle equator [
4]. Aurora-B kinase is known to regulate processes such as kinetochore and microtubule interactions [
5‐
8] and cytokinesis [
9,
10]. Aurora-C is expressed specifically in the male testis [
11] and has meiotic functions [
12].
Aurora-A is critical for mitotic entry, as well as the mitotic spindle checkpoint involving chromosome maturation and segregation [
13‐
15]. Two proteins known to bind and initiate activation of Aurora-A are TPX2 [
16,
17] and Ajuba [
13]. Upon binding, TPX2 or Ajuba stimulate Aurora-A to undergo autophosphorylation and subsequent activation. Once activated, Aurora-A phosphorylates downstream targets such as TPX2, thus regulating the attachment of microtubules to the kinetochore during spindle assembly [
18‐
20]. Aurora-A also phosphorylates the tumor suppressor protein p53, resulting in MDM2 dependent degradation and cell cycle progression [
21]. Aurora-A is overexpressed in ovarian [
22‐
24], breast [
25], colorectal [
26] and metastatic prostate cancer [
27] and is upregulated in response to simian virus 40 (SV40) small tumor (ST) antigen [
28]. In addition, amplification of human chromosome 20q13.2, which contains Aurora-A, frequently occurs in ovarian cancer [
29]. Overexpression of Aurora-A causes transformation in rodent fibroblasts [
30] and tumors in nude mice [
31], consistent with the possibility that Aurora-A is an oncogene.
The current standard of care for advanced ovarian cancer is debulking surgery followed by combination chemotherapy of carboplatin and paclitaxel [
32]. Unfortunately, the majority of patients relapse within 18 months of first-line therapy, and 24–59% of relapse patients treated with paclitaxel progress to resistant disease [
33]. Paclitaxel causes cell death by stabilization of microtubule dynamics resulting in activation of the spindle assembly checkpoint and apoptosis [
34]. Previous studies have investigated the link between Aurora-A levels and sensitivity or resistance to paclitaxel. One study demonstrated that overexpression of Aurora-A in HeLa cells induces resistance to paclitaxel [
35] while another study reported sensitization of pancreatic cancer cells to paclitaxel by siRNA knockdown of Aurora-A [
36]. Interestingly, a recent study in ovarian cancer cells reported that overexpression of Aurora-A could increase cell survival in the presence of paclitaxel [
37].
Through microarray profiling of ovarian cancer samples, we have observed that Aurora-A was significantly overexpressed in ovarian carcinomas compared to adenomas. We confirmed Aurora-A expression at the protein level by staining tissue microarrays from the same patients. Recently, Aurora kinases have been exploited as novel drug targets with the development of a handful of small molecule inhibitors, all of which have been or are in clinical trials (Reviewed in [
38]). To determine if the Aurora kinase family is an effective therapeutic target for ovarian tumors that have acquired resistance to paclitaxel, we tested the ability of VE-465, an Aurora kinase family inhibitor (gift of Merck & Co. and Vertex Pharmaceuticals), to induce apoptosis in the presence and absence of paclitaxel in taxol-sensitive 1A9 and taxol-resistant PTX10 ovarian cancer cells [
39]. VE-465 potently induced apoptosis in both paclitaxel resistant and sensitive ovarian cancer cells. In addition, VE-465 synergistically enhanced apoptosis in combination with paclitaxel in taxol-sensitive cells at low doses (1–10 nM). Our data indicate that VE-465 is effective at inducing apoptosis in both taxol-sensitive and taxol-resistant ovarian cancer cell lines, and thus may be an effective therapy for patients with ovarian cancer, including those patients with taxol-resistant disease.
Methods
Tumor samples, RNA isolation, Microarray Hybridization and Normalization
Cell Culture and Drug Treatment
PTX10 and 1A9 cells were cultured in RPMI media (Mediatech, Herndon, VA) supplemented with 10% fetal bovine serum and grown in 5% CO2 at 37°C. Two days before treatment 1.5 × 105 cells were seeded in each well of a 6-well plate (Corning, Corning, NY). On day one of treatment combinations of 15 ng/mL paclitaxel (Sigma-Aldrich, St. Louis, MO) and either Dimethyl Sulfoxide (DMSO) control or the indicated concentration of of VE-465 (Vertex Pharmaceuticals, Abingdon, United Kingdom) were added to 2 mL of fresh RPMI and incubated for 96 hours prior to FACS analysis or caspase 3/7 activity assays.
Fluorescence Activated Cell Sorting (FACS) Analysis
Following drug treatment, cells were washed from the plate in media, centrifuged at 3000 rpm to pellet and washed once with cold PBS. Pellets were resuspended and fixed in 70% Ethanol/PBS at -20°C overnight. On the day of analysis, pellets were washed once with PBS and digested with 500 μl of 0.1 mg/mL PBS/RNaseA (Sigma-Aldrich, St. Louis, MO) by incubating at 37°C for 15 minutes. DNA content was assessed by staining with 500 μl of 25 μg/mL PBS/Propidium Iodide (Sigma-Aldrich, St. Louis, MO). Cell suspensions were transferred to 5 mL collection tubes for FACS analysis. Samples were processed using a Becton Dickson FACSCalibur analyzer (Becton Dickson, San Jose, CA) and data analyzed using the FlowJo software package (Tree Star, Ashland, OR).
Drug Treatment and Caspase Assay
One day before drug treatment, each well of a white-walled, 96 well luminometer plate (Nalge Nunc International, Rochester, NY) was coated with a 1:4 dilution of BD matrigel matrix (BD biosciences, Bedford, MA) and RPMI media. The plates were incubated at room temperature for one hour and excess matrigel was removed before 4800 cells were seeded in each well in triplicate. On day one of treatment, cells were treated with or without 15 ng/mL paclitaxel (Sigma-Aldrich, St. Louis, MO) plus varying concentrations and combinations of VE-465 (Vertex Pharmaceuticals, Abingdon, United Kingdom), or with 50 μM z-vad (EMD Chemicals, San Diego, CA). Z-vad is a general caspase inhibitor and was used as a negative control to block caspase activity and apoptosis. Control cells were left untreated. Three independent biological replicates were performed, luminescence measured and data analyzed.
The Caspase-Glo™ 3/7 Assay (Promega, Madison, WI) lyophilized substrate (DEVD-aminoluciferin powder) was resuspended in Caspase- Glo™ 3/7 lysis buffer and equilibrized to room temperature. Forty-eight or 72 hours after cell treatment, the Caspase- Glo™ 3/7 reagent was added in a 1:1 volume ratio to each well of the 96 well luminometer plate. Immediately following the addition of the reagent, the contents of the wells were gently mixed with a plate shaker at 500 rpm for 30 seconds. After one hour incubation, the luminescence was measured with a Synergy HT plate reader (BioTek Instruments, Winooski, VT). Culture medium was used as a blank and "no-cell background" values were determined.
Immunofluorescence
PTX10 and 1A9 cells were grown on cover slips (Fisher Scientific, Hampton, NH) in 6-well culture dishes (Corning, Corning, NY). Cells were washed 3 times with cold PBS and fixed in 4% paraformaldehyde for 15 minutes at room temperature, permeablized on ice for 2 minutes in 0.5% Tween-20/PBS and blocked in 5% nonfat dry milk (NFDM) for 30 minutes at room temperature. Mitotic cells were stained with anti-phospho-Histone H3 Serine 10 (Upstate, Charlottesville, VA) with 5% NFDM at a 1:200 dilution for 2 hours at 4°C. Secondary antibody of anti-Rabbit AlexaFluor 488 (Molecular Probes, Eugene, OR) was applied at a 1:400 dilution for 45 minutes at room temperature. Cells were washed 3 times in PBS and stained with TOPro (Molecular Probes, Eugene, OR) at a concentration of 3 μg/μl for 15 minutes to reveal the nucleus. Cover slips were mounted on slides and visualized using a Zeiss Axiovert 35 fluorescence microscope.
Western Blot
60% conflutent cells were lysed in lysis buffer (0.137 M NaCl, 0.02 M TRIS pH 8.0, 10% Glycerol, and 1% NP-40), 50 μg total lysate separated by SDS-PAGE electrophoresis and transferred to nitrocellulose for immunoblotting. Immunoblots were probed with an antibody to Aurora-A (Abcam Inc., Cambridge, MA), Aurora-B (GenScript, Piscitaway, NJ), phosphoAurora-A and -B (Cell Signaling, Danvers, MA), p53 (Santa Cruz Biotechnology, Santa Cruz, CA) and phospho(S315)p53 (Cell Signaling, Danvers, MA). To ensure equal loading blots were then probed with a monoclonal antibody to PP2A, catalytic subunit (BD Biosciences, San Jose, CA).
Tissue Microarray Analysis
TMA sections were stained at the WCI Tissue and Pathology Core Facility
http://www.pathology.emory.edu/WCIPathCore/ with H&E and with Aurora A antibody (1:300 dilution, Abcam, Cambridge, MA). Staining was scored on a four level scale (0 = no staining, 1 = weak staining, 2 = moderate staining, 3 = intense staining) by a GU pathologist.
Discussion
Recently, we identified Aurora-A kinase to be significantly overexpressed in carcinoma patients compared to adenomas [
22]. Our data suggested that reduced p53 activity can lead to improved clinical outcome for ovarian cancer patients undergoing chemotherapy [
22]. One mechanism that might contribute to this phenomenon is that Aurora-A renders cells resistant to paclitaxel-induced apoptosis and stimulates Akt1 and Akt2 activity in wild-type p53 but not p53-null ovarian cancer cells [
37]. Thus, p53-null tumors would be more responsive to chemotherapy regimens. Here, we have shown that the mitotic kinase Aurora-A is overexpressed in ovarian carcinomas compared to adenomas. Furthermore, we have demonstrated that the pan-Aurora inhibitor VE-465 can synergize with paclitaxel to induce apoptosis and is a potent killer of taxane-sensitive and resistant ovarian cancer cells.
Although other Aurora family members were not overexpressed, other genes known to interact with Aurora-A kinase were significantly increased. One of the most significantly overexpressed was TPX2, an activator and substrate of Aurora-A [
16,
17]. Recently, a link between another Aurora-A substrate, BRCA1, and TPX2 has been demonstrated [
57]. Juokov et al. showed that loss of BRCA1 expression leads to mislocalization of TPX2 along microtubules instead of at the aster poles, suggesting a mechanism by which BRCA1 mutation could lead to chromosomal instability [
57]. TPX2 was overexpressed 15-fold in carcinomas and provides a possible mechanism for increased activation of Aurora-A kinase. These observations have implications for ovarian cancer because overexpression of Aurora-A can induce resistance to the chemotherapeutic paclitaxel [
35]. We predicted that ovarian cancer patients who overexpress Aurora-A would have a higher chance of becoming resistant to taxanes and possibly benefit from a different treatment strategy targeted at Aurora-A and other Aurora family members. To test this prediction, we evaluated the compound VE-465 as a pan-Aurora kinase inhibitor and inducer of apoptosis in ovarian cancer cell lines. Although VE-465 is not specific to Aurora-A, it is highly selective and effective at inhibiting Aurora family kinases and offered a unique opportunity to evaluate the entire family of kinases as a therapeutic target. Our results indicate that VE-465 is able to induce apoptosis in the paclitaxel resistant, ovarian cancer cell line PTX10 in a dose dependent manner and synergize with paclitaxel in the 1A9 paclitaxel-sensitive cell line.
VE-465 and paclitaxel are both drugs that function by targeting mitotic cells, but induce apoptosis by different mechanisms. Paclitaxel alters microtubule dynamics and induces the spindle checkpoint resulting in mitotic arrest and eventual apoptosis. VE-465, on the other hand, inhibits the activity of the Aurora kinase family and subsequent mitotic entry. We found that many PTX10 cells treated with VE-465 bypass the spindle checkpoint resulting in missegregation of chromosomes and aneuploidy, possibly due to the inhibition of other family members such as Aurora-B. Thus, in addition to inhibiting mitotic entry, VE-465 appears to induce apoptosis by causing catastrophic chromosomal abnormalities due to the absence of an intact spindle assembly checkpoint in cells that do proceed through mitosis.
Intriguingly, 1A9 cells were more sensitive to VE-465 than PTX10 cells and this correlates with the roughly two fold higher expression of Aurora-A in the 1A9 cell line. Significant cell death was observed at low concentrations in 1A9 cells such as 1–25 nM relative to 50–75 nM for PTX10 cells, suggesting that at low doses VE-465 synergizes with paclitaxel in taxol-sensitive ovarian cancer cells. Interestingly, at low concentrations VE-465 has a K
i more specific to Aurora-A (1 nM) than Aurora-B (26 nM) or -C (9 nM). This suggests the synergistic effects are due to the specific inhibition of Aurora-A and not other family members. However, at higher concentrations, we found no evidence that paclitaxel and VE-465 synergized to induce apoptosis in PTX10 cells. This could be because a very high percentage of cells are undergoing apoptosis at high doses, or possibly due to the inherent nature of the resistance of PTX0 cells. PTX10 cells harbor a point mutation in the M40 β-tubulin isotype resulting in a phenylalanine to valine mutation [
39] which may alter the binding of paclitaxel to microtubules. It is possible that this particular form of resistance does not coincide with the function of Aurora kinases and therefore no synergism is seen when treating with a combination of both drugs. Tumors that exhibit other forms of taxane resistance such as Aurora-A overexpression, alternate point mutations, modulations in tubulin isotypes, decreased tubulin expression and changes in post-translational modifications may respond synergistically when treated with VE-465 and paclitaxel. Alternatively, a synergistic effect may be observed prior to the acquisition of taxol resistance, or in combination with other drugs that target different cellular pathways such as tyrosine kinase receptor signals or apoptosis resistance pathways. Aurora kinase inhibitors represent a promising alternative to taxane therapy, especially for patients who overexpress the mitotic kinase Aurora-A, or other family members, or whose disease continues to progress during taxane therapy [
58].
Treatment of patients with different drugs in a serial fashion allows for clones that are resistant to one therapy to arise by drug-resistance selection. However, combinatorial therapies may be more effective, as has been shown using cocktail therapies for the treatment of the rapidly evolving human immunodeficiency virus [
59]. Thus, initial combinatorial chemotherapy using Aurora-inhibitors, paclitaxel, and other chemotherapeutic agents could be an effective approach to prevent the development of chemo-resistant ovarian cancers.
Acknowledgements
The authors would like to thank Mohamed Ali-Seyed for assistance with propidium iodide staining, Lilya Matyunina for RNA preparation, Dr. Evi Giannakakou for 1A9 and PTX10 cell lines, and Merck & Co. and Vertex Pharmaceuticals for VE-465. CSM was supported in part by R01-CA106826. CDS was supported by DOD predoctoral fellowship PC060145. Tissue Microarrays were stained in the Winship Cancer Institute Research Pathology Core Laboratory. The authors thank Dianne Alexis for technical assistance with TMA staining.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
CDS performed the flow cytometry, immunofluorescence, drug treatments, and immunoblotting experiments and wrote the initial draft. NL performed the caspase 3/7 and qPCR assays. AOO read and scored the TMA. SL generated the tissue microarray. JFM generated the microarray expression data. BBB provided the ovarian patient tissue samples. CSM directed the research, analyzed the microarray data, and co-wrote the manuscript. All authors read and approved the manuscript.