Therapeutic targeting of Aurora A kinase in Philadelphia chromosome-positive ABL tyrosine kinase inhibitor-resistant cells

Abelson murine leukemia viral oncogene homolog (ABL) tyrosine kinase inhibitors (TKIs) have been shown to be effective for treatment of chronic myeloid leukemia (CML) and Philadelphia chromosome-positive (Ph+) acute lymphoblastic leukemia patients. However, resistance to ABL TKIs can develop as a result of breakpoint cluster region-ABL point mutations. Aurora kinases regulate many processes associated with mitosis. In this study, we investigated whether inhibiting Aurora kinase can reduce the viability of Ph+ leukemia cells. Treatment with the Aurora kinase A inhibitor alisertib blocked Ph+ leukemia cell proliferation and Aurora kinase A phosphorylation; it also induced G2/M-phase arrest and increased the intracellular levels of reactive oxygen species. Combined treatment of Ph+ cells with ABL TKIs and alisertib was cytotoxic, with the fraction of senescent cells increasing in a time- and dose-dependent manner. Aurora A gene silencing suppressed cell proliferation and enhanced ABL TKI efficacy. In a mouse xenograft model, co-administration of ponatinib and alisertib enhanced survival and reduced tumor size; moreover, the treatments were well tolerated by the animals. These results indicate that inhibiting Aurora kinase can enhance the cytotoxic effects of ABL TKIs and is, therefore, an effective therapeutic strategy against ABL TKI-resistant cells, including those with the T315I mutation.


INTRODUCTION
Chronic myeloid leukemia (CML) is a myelo proliferative neoplasm characterized by cytogenetic aberration, namely, the translocation of the Philadelphia chromosome (Ph) [1], which generates a breakpoint cluster regionAbelson murine leukemia viral oncogene homolog 1 (BCRABL1) fusion oncogene that encodes the BCRABL oncoprotein [2]. Treatment options for CML include ABL tyrosine kinase inhibitors (TKIs) such as imatinib, which has been shown to be effective in patients [3]. However, resistance to imatinib can develop as a result of BCRABL point mutations [4,5], such as the T315I gatekeeper substitution mutation in the ABL kinase domain [6]. CML patients with T315I mutation exhibit a reduced response to ABL TKIs, including the second generation drugs nilotinib, dasatinib, and bosutinib [6].
Ponatinib (also known as AP24534) is an oral multi target thirdgeneration TKI [7] that can overcome imatinib resistance due to T315I mutation. Ponatinib is also effective against heavily pretreated resistant CML and in onethird of patients in the accelerated or blastic phases of the disease [8]. However, in a clinical trial, only 62% of refractory CML patients exhibited a cytogenetic response to ponatinib treatment 8 . Therefore, new therapeutic approaches are needed to overcome ABL TKI resistance and improve the outcome of Phpositive (Ph+) leukemia patients.
Mitosis is a normal part of the cell cycle [9] that is dysregulated in cancer. Aurora kinases are a family of serine/threonine kinases that are required for mitotic progression [10]. The mammalian genome contains three Aurora kinases (Aurora A, B, and C); Aurora A is essential for mitotic spindle assembly [11] and is overexpressed in cancer cells. Thus, Aurora A is an attractive target for Ph+ Research Paper leukemia treatment, including cases that are resistant to ABL TKIs [12,13].
In this study, we investigated whether inhibiting Aurora kinase A can reduce the viability of Ph+ leukemia cells that are resistant to ABL TKIs.

Aurora A expression in Ph+ leukemia cells
Aurora A is essential for normal mitotic spindle formation and centrosome maturation. Aurora A phosphorylation at Thr288 in the catalytic domain increases its kinase activity [14]. We, therefore, examined the phosphorylation level of Aurora A by immunoblotting to determine whether targeting Aurora A can affect mitosis in K562 and Ba/F3 T315I Ph+ leukemia cells. We found that Aurora A was phosphorylated in K562 and Ba/F3 T315I cells and that this was reduced in a dosedependent manner by treatment with alisertib for 48 h ( Figure 1A). It was previously reported that inhibiting Aurora kinase A leads to mitotic arrest [10]. We, therefore, examined cell cycle status by flow cytometry in alisertibtreated Ph+ leukemia cells. After treatment with 100 nM alisertib for 48 h, the G2/M fraction of K562 and Ba/F3 T315I cells was increased ( Figure 1B). To determine the effect of alisertib on cell growth, K562, Ba/F3 T315I, ABL TKI resistant K562 cells (K562 imatinib resistant: IR, nirotinib resistant: NR) were treated with various concentrations of alisertib for 72 h. We found that proliferation was decreased in a dosedependent manner ( Figure 1C, Supplementary Figure 1A) and also that the other Aurora A inhibitor, Aurora Inhibitor I, was inhibited in K562 cells (Supplementary Figure 1B). Moreover, cytotoxicity and caspase 3/7 activity were dosedependently enhanced in the presence of alisertib ( Figure 1D, 1E). Consistent with these observations, we observed that alisertib increased apoptosis in K562 and Ba/F3 T315I cells ( Figure 1F).

Efficacy of ABL TKIs and alisertib against Ph+ cells
ABL TKIs are a standard treatment for Ph+ leukemia patients. To investigate the efficacy of ABL TKIs and Aurora kinase inhibitor, Ph+ cells were treated with the ABL TKIs imatinib, nilotinib or ponatinib alone or in combination with alisertib. Cotreatment with imatinib, nilotinib, or ponatinib with alisertib had a synergistic effect that was more potent than the treatment with a single drug (Supplementary Figure 2A-2D). Cytotoxicity and caspase 3/7 activity were also increased by ABL TKI and alisertib treatment (Figure 2A, 2B). Immunoblot analysis revealed that imatinib or ponatinib and alisertib treatment increased caspase 3 and PARP activity and reduced CrkL phosphorylation in K562 and Ba/F3 T315I cells ( Figure 2C). These results indicate that the combination of ABL TKIs and Aurora kinase inhibitor is effective against Ph+ leukemia cells, including those with the T315I mutation.

Alisertib induces cellular senescence in Ph+ cells
Senescence is a terminal cellular outcome that has a cytostatic effect [15]. To determine whether alisertib induces cellular senescence, we examined SA β-gal activity in K562 and Ba/F3 T315I cells. SA-β-gal staining was increased by alisertib starting on day 1; after 72 h of treatment, the number of SA-β-gal-positive cells was increased in a dosedependent manner ( Figure 3A, Supplementary Figure 3A), an effect that was attenuated in the presence of Nacetyllcysteine (NAC) ( Figure 3B, Supplementary Figure 3B), a nonspecific ROS scavenger [16]. ROS can cause premature senescence and induce apoptosis; we found that intracellular ROS levels in Ph+ cells were dosedependently increased by alisertib treatment ( Figure 3C). NAC and synthetic antioxidants abrogated this effect ( Figure 3D).

Aurora A silencing increases ABL TKI activity against Ph+ cells
To evaluate the effect of inhibiting of Aurora A kinase on the leukemia cell response to ABL TKIs, we used a siRNA to knock down Aurora A expression ( Figure 4A). Aurora A knockdown enhanced imatinib induced cell death relative to control siRNAtransfected cells, as evidenced by the increase in cytotoxicity, caspase 3/7 activity, and apoptosis ( Figure 4B-4D). Aurora kinase A forms a protein complex with cMyc in liver cancer [17]. Immunoblot analysis revealed that cMyc expression was reduced upon Aurora A siRNA transfection ( Figure 4E). cMyc and Aurora A expression were found to be correlated in clinical CML samples, as determined by quantitative reverse transcription PCR ( Figure 4F). These results indicate that downregulation of Aurora A is associated with leukemia cell death.

Efficacy of ABL TKI combined with alisertib in a mouse model
To evaluate the efficacy of alisertib in vivo, nude mice or NODSCID mice were subcutaneously or intraperitoneally inoculated with Ba/F3 T315I cells and the average tumor volume was evaluated every 3 days. We also examined patientderived xenograft (PDX) models as they are functional to study cancer biology and for drug screening [18]. Ponatinib (20 mg/kg, 5 days/ week) or alisertib (30 mg/kg, 5 days/week) administered orally inhibited the growth of Ba/F3 T315I cell tumors in vivo to a greater extent than the vehicle control (PBS) (P < 0.05) ( Figure 5A), whereas cotreatment with alisertib and ponatinib decreased tumor volume ( Figure 5B, www.oncotarget.com  Figure 4B) to a greater extent than either drug alone. The Kaplan-Meier analysis revealed that co treatment with ponatinib and alisertib improved survival ( Figure 5D). All doses were well tolerated and body weight loss was not observed in the treatment groups ( Figure 5E).
We also found that CrkL phosphorylation was reduced and PARP activity was increased after ponatinib and alisertib treatment in mouse tumor samples ( Figure 5F). The readout of the flow cytometric analysis of human CD45 positive cells was decreased in the spleen and in the peripheral blood samples by ponatinib and alisertib combination  Figure 4C). These results indicate that alisertib increases the efficacy of ABL TKIs and does not have noticeable side effects.

DISCUSSION
Alisertib is an oral and selective Aurora kinase A inhibitor [19] that functions as an oncoprotein. We showed here that Aurora A is expressed in Ph+ leukemia cells and that inhibiting Aurora A rendered the cells more sensitive to the cytotoxic and proapoptotic effects of ABL TKIs. The positive correlation between Aurora A and cMyc expression in primary CML samples suggests that blocking Aurora kinase activity in combination with ABL TKI treatment can decrease the proliferation and survival of Ph+ cells as a result of cMyc downregulation.
Cellular senescence is considered a tumor suppressing mechanism. Our results demonstrated that induction of senescence blocked Ph+ cell growth and increased ROS levels. Alisertib exposure has been linked to an increase in ROSmediated apoptosis signaling. Thus, ROS play an essential role in alisertibinduced apoptosis in Ph+ cells.
Interestingly, in the mouse xenograft tumor model, cotreatment with ABL TKI, ponatinib, and alisertib prolonged survival and reduced spleen size. Moreover, a combination of 20 mg/kg/day ponatinib and 30 mg/kg/ day alisertib suppressed tumor growth as compared to administration of each drug alone. The treatments were well tolerated, suggesting that alisertib combination strategies are clinically feasible. In fact, alisertib and other Aurora kinase inhibitors are now in clinical trial for the treatment of various types of tumors [19][20][21][22]. The most common side effects in phase 1 studies were thrombocytopenia, neutropenia, and gastrointestinal toxicity [19]. However, in animal models, alisertib was shown to decrease tumor burden and increase overall survival without toxicity [23]. In the mouse study, ponatinib was more active in the T315I group. The pharmacokinetics of ponatinib, by a phase 1 Japanese study, shows that C max is 86.4 ng/ml in patients treated with 45 mg. These results indicated that ponatinib was more effective in the mouse study possibly because   the serum concentration reached a higher value than the in vitro treatment dose [24].
In this study, we investigated the efficacy of combination therapy with an ABL TKI and alisertib against Ph+ cells with T315I mutation and ABL TKI resistant cells. Although the underlying mechanisms have yet to be fully defined, our data indicate that Aurora A is an important regulator of tumorigenesis in Ph+ cells. Alisertib induced senescence in Ph+ cells by stimulating ROS production. Thus, Aurora A inhibition is an effective treatment strategy for Ph+ leukemia patients. Moreover, combination therapy with ABL TKI and alisertib exerts a synergistic effect; this combination may improve the clinical outcome of patients with acquired imatinib resistance.

Ethics statement
The present investigation has been conducted in accordance with the ethical standards, according to the Declaration of Helsinki and to national and international guidelines and has been approved by the Tokyo Medical University's review board.

Chemicals
Alisertib (MLN8237), Aurora A Inhibitor Ⅰ and ponatinib were purchased from Selleck Chemicals (Houston, TX, USA) and MedKoo Biosciences (Chapel Hill, NC, USA); imatinib was provided by Novartis Pharma AG (Basel, Switzerland). Stock solutions of alisertib and ponatinib were prepared in dimethyl sulfoxide, and imatinib was dissolved in distilled water, aliquoted, and stored at −20° C. Other reagents were obtained from SigmaAldrich (St. Louis, MO, USA).

Cell culture
The K562 Ph+ leukemia cell line was obtained from the American Type Culture Collection (Manassas, VA, USA). T315Imutant Ba/F3 BCRABL cells and ABL TKI resistant K562 cells (K562 IR, K562 NR) were previously established [25,26]. All cell lines were cultured in Roswell Park Memorial Institute 1640 medium containing 10% fetal bovine serum with or without 500 nM imatinib or nilotinib at 37° C in a humidified atmosphere with 5% CO 2 . Fresh peripheral CML blood samples were collected from patients, and mononuclear cells were isolated using LymphoSepare (ImmunoBiological Laboratories, Okayama, Japan) and either used immediately or cryopreserved in liquid nitrogen until use. The study protocol was approved by the Institutional Review Board of Tokyo Medical University (No. 1974), and written informed consent was obtained from all patients in accordance with the Declaration of Helsinki.

Cell viability assays
Ph+ leukemia cells were treated with alisertib alone or in combination with imatinib or ponatinib and viability was evaluated by Trypan Blue exclusion or with Cell Counting Kit (Dojindo Laboratories, Kumamoto, Japan) followed by measurement of absorbance at 450 nm. The experiments were performed with triplicate samples.

Cytotoxicity assay
Cells were treated with various concentration of alisertib and/or imatinib or ponatinib. Cytotoxicity was evaluated based on lactose dehydrogenase (LDH) release using the Cytotoxicity LDH Assay kit with watersoluble tetrazolium [WST] salt (Dojindo Laboratories) according to the manufacturer's protocol. The amount of LDH released form dead cells was measured using an EnSpire Multimode Plate Reader (PerkinElmer, Waltham, MA, USA).

Caspase activity
Caspase activity in leukemia cells was examined using the Caspase Glo 3/7 assay kit (Promega, Madison, WI, USA) according to the manufacturer's protocol. Absorbance was measured on a microplate reader.

Apoptosis assay
Apoptosis was assessed with the Annexin V detection kit (BD Biosciences, Franklin Lakes, NJ, USA) according to the manufacturer's protocol. Cells were analyzed by flow cytometry on a FACS Verse instrument (BD Biosciences).

Detection of reactive oxygen species (ROS)
Cells were treated with indicated concentration of alisertib and/or ABL TKIs for 72 h. Intracellular ROS levels were analyzed with the 2',7'dichlorofluorescein diacetate (DCFDA)/H2DCFDA Cellular Reactive Oxygen Species Detection Assay kit (Abcam, Cambridge, UK) according to the manufacturer's protocol. Fluorescence was measured on a microplate reader.

β-Galactosidase (β-gal) staining
Senescence-associated (SA)-β-gal staining was used to evaluate senescence in Ph+ cells. SA-β-gal activity was qualitatively assessed with a kit (Cell Signaling Technology, Danvers, MA, USA) according to the manufacturer's instructions along with examination on an inverted microscope (Olympus, Tokyo, Japan). The percentage of senescent cells was counted as the number of SA-β-gal-positive cells per total number of cells.

Cell cycle analysis
Cell cycle status was analyzed with the BD Cycletest Plus DNA Reagent kit (BD Biosciences) according to the manufacturer's protocol. After 48 h of treatment with alisertib, K562 and Ba/F3 T315I cell cycle distribution was analyzed by flow cytometry.

Immunoblot analysis
Immunoblotting was performed according to a previously described method [27,28]. Briefly, after incubation in indicated concentrations of inhibitor, cells were washed with icecold phosphatebuffered saline (PBS) and lysed in radioimmunoprecipitation lysis buffer. The protein content of lysates was determined with a kit (BioRad, Hercules, CA, USA) and proteins were resolved by polyacrylamide gel electrophoresis and transferred to a polyvinylidene difluoride membrane (Millipore, Billerica, MA, USA) that was incubated with primary antibodies at the appropriate dilution for 1 h. The blots were then probed with the secondary antibodies and developed with an enhanced chemiluminescence system (Amersham Pharmacia Biotech, Little Chalfont, UK). Antibodies against the following proteins were used: phosphorylated (p)ABL, pCrkL, cleaved caspase 3, cleaved poly (ADP ribose) polymerase (PARP), and pAurora, pAurora A (Thr288)/Aurora B (Thr232)/Aurora C (Thr198) (all from Cell Signaling Technology); CrkL (Millipore); Aurora A (GeneTex, Irvine, CA, USA); and ABL, c-Myc, and β-actin (Santa Cruz Biotechnology, Santa Cruz, CA, USA). Three independent experiments were performed. Protein band intensity was evaluated using ImageJ software (National Institutes of Health, Bethesda, MD, USA).

Small interfering (si)RNA transfection
siRNA targeting Aurora A was purchased from SigmaAldrich. K562 cells were transfected with Aurora A or control siRNA by electroporation as previously described [29].

Real-time PCR
Total RNA was extracted from cultured cells using the RNeasy Mini kit (Qiagen, Venlo, Netherlands) and subjected to DNase digestion with the RNaseFree DNase Set (Qiagen). Reverse transcription was performed using the First Strand cDNA Synthesis kit (OriGene Technologies, Rockville, MD, USA). Realtime reverse transcription PCR was performed using FastStart Essential DNA Green Master Mix according to the recommended protocol on a LightCycler 96 system (Roche, Mannheim, Germany). Primers were purchased from Takara Bio, Otsu, Japan), and β-actin or glyceraldehyde 3-phosphate dehydrogenase served as an internal control.

In vivo assay
Mice were maintained under specific pathogenfree conditions. Experiments were performed with the approval of the Institutional Animal Care and Use Committee of Tokyo Medical University. Female mice BALB/c nu/nu mice or NODSCID mice (6 weeks old) were subcutaneously or intraperitoneally injected with 1 × 10 7 Ba/F3 T315I cells or T315I positive human samples, and then orally administered 20 mg/kg ponatinib, 30 mg/kg alisertib, or both for 5 days/week. Control mice were administered PBS by the same route. At predetermined time points, tumor size and mouse survival were recorded. The average tumor weight and spleen size per mouse were calculated and used to determine mean tumor volume or weight ± SEM (n = 5) for each group. Mouse spleen, tumor cells, and peripheral blood samples were collected at indicated time points, and the samples were also analyzed by flow cytometry.

Statistical analysis
A Student's ttest and twoway analysis of variance were used to compare the means of drug treatment and control groups. P < 0.05 was considered statistically significant.

Author contributions
SO performed the experimental procedures; TT, YT, and KO designed and coordinated the study and interpreted data. All authors have read and approved the final manuscript.

ACKNOWLEDGMENTS
We thank the Tokyo Medical University Research Center for providing technical assistance.