BPR1K653, a Novel Aurora Kinase Inhibitor, Exhibits Potent Anti-Proliferative Activity in MDR1 (P-gp170)-Mediated Multidrug-Resistant Cancer Cells

Background Over-expression of Aurora kinases promotes the tumorigenesis of cells. The aim of this study was to determine the preclinical profile of a novel pan-Aurora kinase inhibitor, BPR1K653, as a candidate for anti-cancer therapy. Since expression of the drug efflux pump, MDR1, reduces the effectiveness of various chemotherapeutic compounds in human cancers, this study also aimed to determine whether the potency of BPR1K653 could be affected by the expression of MDR1 in cancer cells. Principal Findings BPR1K653 specifically inhibited the activity of Aurora-A and Aurora-B kinase at low nano-molar concentrations in vitro. Anti-proliferative activity of BPR1K653 was evaluated in various human cancer cell lines. Results of the clonogenic assay showed that BPR1K653 was potent in targeting a variety of cancer cell lines regardless of the tissue origin, p53 status, or expression of MDR1. At the cellular level, BPR1K653 induced endo-replication and subsequent apoptosis in both MDR1-negative and MDR1-positive cancer cells. Importantly, it showed potent activity against the growth of xenograft tumors of the human cervical carcinoma KB and KB-derived MDR1-positive KB-VIN10 cells in nude mice. Finally, BPR1K653 also exhibited favorable pharmacokinetic properties in rats. Conclusions and Significance BPR1K653 is a novel potent anti-cancer compound, and its potency is not affected by the expression of the multiple drug resistant protein, MDR1, in cancer cells. Therefore, BPR1K653 is a promising anti-cancer compound that has potential for the management of various malignancies, particularly for patients with MDR1-related drug resistance after prolonged chemotherapeutic treatments.


Introduction
Mitosis is a key step in cell cycle that is tightly regulated by many proteins. Abnormal expression or activation of these regulatory proteins could result in aberrant mitosis, leading to the development of cancers [1,2]. At the molecular level, Aurora kinases (Aurora-A, Aurora-B and Aurora-C) are serine/threonine kinases that function as key regulators of mitosis. Under normal physiological conditions, they are essential for spindle assembly, centrosome maturation, chromosomal segregation and cytokinesis [3,4]. Under pathological conditions, it has been demonstrated that Aurora kinases are over-expressed in various human cancers and also played important roles in the process of tumorigenesis [5,6,7,8]. For example, Aurora-A kinase is over-expressed in upper gastrointestinal adenocarcinomas [6]. In addition, a correlation between Aurora-A expression levels and tumor progression has been demonstrated in patients with head and neck squamous cell carcinoma [9]. On the other hand, Aurora-B kinase is frequently over-expressed in primary NSCLC and malignant gliomas, particularly glioblastomas [10,11]. Since over-expression of Aurora-A and Aurora-B is frequently associated with tumorigenesis, these molecules have been targeted for cancer therapy. The first proof-of-concept pan-Aurora kinase inhibitor, VX-680 (MK-0457, Tozasertib), was developed in 2004 by Vertex Pharmaceuticals (in collaboration with Merck) with an aim to target cancer cells. This specific inhibitor has been shown effective in targeting cancer cells both in vitro and in vivo, and has received approval from the US Food and Drug Administration (FDA) to enter clinical trials [12,13,14]. Since then, continuous efforts have been made by different pharmaceutical companies in search of potential Aurora kinase inhibitors that exhibit better therapeutic profile and specificity as compare to the first generation inhibitor, VX680 [15,16,17,18,19,20,21,22,23].
Despite early successes of the development of various Aurora kinase inhibitors, recent studies reveal that the effectiveness of many of these developed and clinically tested inhibitors, including VX680, PHA-739358 and AZD1152, can be affected by the expression of multidrug resistance protein MDR1 (P-gp170) in cancer cells [24,25]. In fact, over-expression of MDR1 also interferes with a broad range of different chemotherapeutic agents [2,26,27,28,29]. For examples, expression of the trans-membrane drug efflux pump, MDR1, reduces the sensitivity of cancer cells to paclitaxel, vincristine (anti-microtubule agents), doxorubicin (DNA intercalating agent), mitoxantrone, VP-16 (topoisomerase II inhibitors) and imatinib (tyrosine kinase inhibitor) [28,30,31,32,33,34]. Therefore, there has been great interest in identifying novel anti-cancer compounds that can overcome MDR1related resistance and also exhibit improved pharmacological profiles.
In this study, a novel pan-Aurora kinase inhibitor entitled BPR1K653 was developed and its potency against various MDR1negative and MDR1-positive cancer cells was evaluated. Results of the current study show that unlike the above mentioned chemotherapeutic agents, BPR1K653 is effective in targeting both MDR1negative and -positive cancer cells in vitro and in vivo. Furthermore, BPR1K653 exhibits favorable pharmacokinetic properties in vivo.

BPR1K653 is a potent and selective pan-Aurora kinase inhibitor
In vitro kinase inhibition assay revealed that BPR1K653 ( Figure 1A) inhibited the activity of Aurora-A and -B kinase with an IC 50   and Table 1). The selectivity of BPR1K653 was then evaluated against different kinases. BPR1K653 exhibited less potency (i.e. IC 50 .10 mM) in inhibiting the activity of ALK, CHK1, cMET, EGFR, FLT3, VEGFR1 and VEGFR2 as compared to Aurora-A and Aurora-B kinase ( Table 1). The cellular activity of BPR1K653 was also examined. Activation of Aurora-A kinase requires an autophosphorylation on the Thr288 residue, whereas phosphorylation of the Thr232 residue is an essential regulatory mechanism for Aurora-B activation [35,36]. Here, Western blot analysis revealed that the amount of phosphor-Aurora-A, -B and -C kinase present in HCT116 cancer cells treated with a pan-Aurora kinase inhibitor, VX680 (positive control), was reduced in a concentrationdependent manner ( Figure 1C). Reduction of phosphor-Histone H3 (Ser10), a direct substrate of Aurora-B kinase, is widely used as an indicator of Aurora kinase inhibition in cells. Here, VX680 also reduced the amount of phosphor-Histone H3 (Ser10) present in cells as expect ( Figure 1C). Consistent with these findings, BPR1K653 induced a concentration-dependent decrease in phosphor-Aurora-A, -B and -C kinase in HCT116 cells. HCT116 cells treated with BPR1K653 also showed a concentration-dependent decrease in phosphor-Histone H3 ( Figure 1C).

BPR1K653 inhibits the proliferation of multiple human cancer cell lines regardless of their tissue origins and p53 status
To determine whether BPR1K653 could inhibit cell proliferation, a panel of 11 different cancer cell lines was treated with BPR1K653. For comparison, cells were also treated with two well-characterized Aurora kinase inhibitors, VX680, and PHA739358. It has been demonstrated that loss of p53 function induces multidrug resistance in some types of cancer [37]. Here, results of the clonogenic assay revealed that BPR1K653 was effective (i.e. IC 50 ,0.5 mM) against various types of cancer cells, including lung (A549), oral (HONE-1 and OECM-1) cervical (KB), colon (HT29), bladder (NTUB1) and leukemia/lymphoma (MV4-11 and IM9), regardless of their p53 status (Table 2). Moreover, the potency of BPR1K653 was shown to be higher than that of VX680 and PHA739358 in most of the tested cancer cell lines ( Table 2). The IC 50 values of VX680 and PHA739358 in various cancer cell lines (except in OECM-1 cells) were 2-10 folds higher than those of BPR1K653. The IC 50s of VX680 and BPR1K653 were equal in OECM-1 cells. Taken together, our results demonstrated that BPR1K653 is able to inhibit the proliferation of various types of cancer cell regardless of their tissue origins and p53 status.
BPR1K653 is equally potent in inhibiting the growth of the multiple-drug resistance protein (MDR1) -expressing cancer cells It has been widely demonstrated that over-expression of MDR1 (P-gp, drug efflux pump) induces drug resistance to various chemotherapeutic agents. To determine whether the potency of BPR1K653 is abrogated by MDR1 expression in cancer cells, three multidrug resistant MDR1-expressing cancer cell lines, KB-VIN10, KB-S15 and NTU0.017 [2,38,39,40], were treated with BPR1K653. As shown in Table 3, the IC 50 value of BPR1K653 to KB-VIN10 and KB-S15 was similar to those of the parental MDR1-negative KB cells. The IC 50 of BPR1K653 to KB-VIN10, KB-S15 and KB were 14 nM, 11 nM and 12 nM, respectively. In addition, the IC 50 value of BPR1K653 to the MDR1-expressing NTU0.017 cells was also similar to that of the parental MDR1negative NTUB1 cells (Table 3). Previous studies revealed that Aurora kinase inhibitors, VX680 and PHA739358, are substrates of MDR1 [24,25]. Consistently, all of our tested MDR1expressing cancer cell lines showed cross-resistant to VX680 and PHA739358 (Table 3). In addition, the level of MDR1 expression correlated with the level of VX680/PHA-739358 resistance in KB-VIN10 and KB-S15 cancer cells ( Figure 2). To further determine whether the potency of VX680 and PHA739358 in KB-VIN10, KB-S15 and NTU0.017 cells were actually affected by the expression of MDR1, cells were co-treated with the MDR1 modulator (negative regulator), verapamil, and cell viability was determined. Here, verapamil treatment (10 mM) was shown to be able to restore/enhance the sensitivity to both VX680 and PHA739358 in all of the tested MDR1-expressing cancer cells (Table 3). However, verapamil treatment could not further increase the sensitivity to BPR1K653 in both MDR-negative and MDR1-expressing cancer cells (data not shown). On the other hand, it has been demonstrated that a KB derived VP-16 resistant cancer cell line, KB-7D, over-expresses another type of the ATPdependent multi-drug efflux protein, MPR1 [41]. Interestingly, the IC 50 value of BPR1K653 to KB-7D was also similar to that of the parental MRP1-negative KB cells (Table 3).

BPR1K653 induces endo-replication in both MDR1negative and -positive cancer cells
Further experiments were performed to reconfirm the above findings that the effectiveness of BPR1K653 is not affected by the MDR1 expression in cells. Inhibition of Aurora kinases induces endoreduplication of cells, indicating by the formation of polyploidy [14]. Here, results of immunofluorescence microscopy and flow cytometric analysis clearly showed that BPR1K653 induced the formation of polyploidy (populations .4N) in KB cells ( Figure 3A and B, and Figure S1A). The MDR1-expressing KB-VIN10 cells treated with the same concentrations of BPR1K653 as had been applied to KB cells also induced the formation of polyploidy ( Figure 3A and C, and Figure S1A). In contrast, VX680 only induced the formation of polyploidy in KB cells but not in KB-VIN10 cells under the same treatment concentrations ( Figure 3A, B and C). However, formation of the polyploidy population was shown in KB-VIN10 cells co-treated with 10 mM of the MDR-inhibitor, verapamil, and VX680 ( Figure 3C). These results are consistent with the findings of the above clonogenic assay that expression of MDR1 in cancer cells affects the effectiveness of VX680 but not of BPR1K653.
To determine whether BPR1K653 also induces endo-replication in cancer cell lines other than KB and its derivative, HONE-1 cells were treated with BPR1K653 and cellular contents were analyzed by microcopy and flow cytometry. Both immunofluorescence microscopy and flow cytometric analysis clearly showed that BPR1K653 promoted the formation of polyploidy (populations .4N) in HONE-1 cells in a concentration-dependent manner ( Figure 3D and E).
BPR1K653 reduces Histone H3 phosphorylation and cyclin B1 expression in both MDR1-negative and -positive cancer cells Western blot analysis was performed to reconfirm that the effectiveness of BPR1K653 is not affected by the MDR1 expression in cancer cells. Histone H3 is a direct substrate of Aurora-B kinase, and endo-replicating cells usually show reduction of the expression of cyclin B1. In this experiment, inhibition of Histone H3 phosphorylation and down-regulation of cyclin B1 expression were shown in both KB and KB-VIN10 cells treated with the same concentrations, 12 (IC 50 ), 24 (26 IC 50 ) and 36 nM (36 IC 50 ) of BPR1K653 in a concentration-dependent manner ( Figure 4A and B). Consistent with these findings, VX680 treatment (i.e. 170 nM and 255 nM) also inhibited the phosphorylation of Histone H3 and the expression of cyclin B1 in KB cells ( Figure 4A). However, same VX680 treatment could not induce the above molecular changes in the MDR1-expressing KB-VIN10 cells. Verapamil treatment (10 mM) was shown to restore the sensitivity to VX680 in KB-VIN10 cells, as indicated by a reduction in the Histone H3 phosphorylation and cyclin B1 expression ( Figure 4B).
To determine whether BPR1K653 also reduces Histone H3 phosphorylation and cyclin B1 expression in cancer cell lines other than KB and its derivative, HONE-1 cells was treated with BPR1K653 and intracellular proteins were analyzed by Western blotting. Western blot analysis clearly demonstrated that both the phosphorylation of Histone H3 and expression of cyclin B1 were decreased in BPR1K653-treated HONE-1 cells ( Figure 4C).

BPR1K653 induces apoptosis in both MDR1-negative and -positive cancer cells
Previous studies revealed that targeting Aurora kinases induces cell endo-replication and subsequent cell apoptosis [14]. To Table 3. BPR1K653 exhibits anti-proliferative activity against various MDR1/MRP1-positive cancer cells.   determine whether BPR1K653 is able to induce apoptosis in both MDR1-positive and -negative cancer cells, KB and KB-VIN10 cells were treated with BPR1K653 and apoptotic properties were analyzed by Annexin-V, real-time caspase-3/-7 activity imaging and TUNEL assays. Here, both cytoplasmic volume and the size of nucleus were increased in the BPR1K653-treated KB and KB-VIN10 cells, indicating that BPR1K653 induced cell endoreplication as expected ( Figure 5A and C, and Figure S1A). Translocation of the phosphatidylserine molecule from the innerleaflet of cell membrane to the outer membrane indicates the occurrence of early apoptosis. Results of the Annexin-V assay showed that BPR1K653 induced the translocation of the phosphatidylserine molecule in both KB and KB-VIN10 cells, as indicating by the green fluorescent label ( Figure 5A). BPR1K653 also induced the caspase-3/-7 activity and DNA fragmentation in both KB and KB-VIN10 cells under the same treatment conditions ( Figure 5B, C, D, and Figure S1A). In contrast, VX680 only induced the translocation of the phosphatidylserine molecule, caspase-3/-7 activity and DNA fragmentation in KB cells and not in the MDR1-expressing KB-VIN10 cells ( Figure 5A, B, C and D). Moreover, cleavage of PARP was only shown in the MDR1-expressing KB-VIN10 cells treated with either BPR1K653 or VX680/verapamil (co-treatment), and not with VX680 alone, as revealed by the Western blot analysis ( Figure 5E). BPR1K653 also induced apoptosis in HONE-1 cells, as indicated by the induction of caspae-3/-7 activity in vitro ( Figure  S1B).

BPR1K653 suppresses the growth of both human MDR1negative and -positive cancer xenografts in vivo
Although the above results showed that BPR1K653 exhibits potent anti-cancer effect in vitro, experiments were performed to determine whether BPR1K653 is also able to inhibit the activity of Aurora kinases and the growth of both MDR1-negative/positive tumors in vivo. KB cells were grown as s.c. tumors in nude mice. When well-established KB xenografts were palpable with tumor size of ,75 mm 3 , mice were randomized into vehicle control and treatment groups of five animals each. The treated mice received either 15 mg/kg of BPR1K653 or 30 mg/kg of VX680 i.p. for 5 days/week for 2 consecutive weeks. Results of the immunohistochemical analysis of the tumor tissue sections showed that administration of BPR1K653 reduced the amount of phosphor-Histone H3 positive cells present in tumor tissues as compared to the control (10% vs 60%) ( Figure 6A). A decrease in the rate of tumor growth in mice treated with either BPR1K653 or VX680 5 days/week for 2 consecutive weeks was also observed. There was a ,73% decrease in tumor volume on day 30 in the animals treated with BPR1K653 (P,0.05). In addition, there was a ,68% decrease in tumor volume on Day 30 in the animals treated with VX680 (P,0.05; Figure 6B). BPR1K653 was well-tolerated at the dosage of 15 mg/kg with no signs of toxicity in the KB xenograft tumor model as the loss of body weight after treatment was less than 10% in the treatment group as compare to the control group ( Figure 6C). To determine whether the inhibition of tumor growth in BPR1K653-treated mice was related to the increases of apoptotic cancer cell populations, tumors were surgically removed from the mice 12 days post-treatment and tissue sections were analyzed by TUNEL assay. Results of the TUNEL assay showed that the amount of apoptotic cells present in the tumor tissue of BPR1K653-treated mice was significantly higher than those in the control mice (55% vs 7%) ( Figure 6D). This is consistent with the result of the above in vitro experiment that BPR1K652 is able to induce cancer cells apoptosis.
Notably, BPR1K653 is also as effective toward MDR1expressing tumor xenograft as it is in cultured MDR1-expressing cells. Here, KB-VIN10 tumor xenograft was used to evaluate the efficacy of BPR1K653 against MDR1-expressing tumor in vivo. Due to the slow growing properties of KB-VIN10, the treated mice received either 15 mg/kg of BPR1K653 or 30 mg/kg of VX680 i.p. for 5 days/week for 3 consecutive weeks instead of 2 weeks as in KB-implanted mice. In comparison to the control mice, growth of KB-VIN10 tumor was significantly inhibited in mice treated with 15 mg/kg of BPR1K653. There was a ,50% decrease in tumor volume on Day 42 in the animals treated with BPR1K653 (P,0.05). In contrast, VX680 did not exhibit significant tumor growth inhibitory effect in mice transplanted with KB-VIN10 cells ( Figure 6E). Moreover, BPR1K653 was welltolerated at the dosage of 15 mg/kg (5 days/week for 3 consecutive weeks) with no signs of toxicity in the KB-VIN10 xenograft tumor model as the loss of body weight after treatment was less than 10% in the treatment group as compare to the control group ( Figure 6F). Thus, BPR1K653 exerts potent antitumoral efficacy toward both MDR-negative and MDR-expressing tumor xenografts.

Pharmacokinetics of BPR1K653 in rats
Finally, pharmacokinetic studies of BPR1K653 were accessed over a 24 h period to examine plasma concentrations of BPR1K653 after a single intravenous administration (Table 4). After a single administration of BPR1K653 at a dosage of 5 mg/ kg to rats, BPR1K653 achieved a maximum plasma concentration of 10 mM (5463 ng/mL) at 2 min after dosing, and the estimated BPR1K653 plasma concentration remained at a concentration of 3.9 nM (2.1 ng/mL) 24 h after dosing. The plasma half-life, total body clearance, and volume of distribution at the steady state (Vss) were 3.960.7 h, 49.3610.6 mL/min/kg and 10.665.1 L/kg, respectively.

Discussion
Aurora kinases have emerged as key regulators of mitosis and evidence indicates abnormalities in their expression and activity are closely related to the development and progression of various cancers. In this study, we have developed a novel pan-Aurora kinase inhibitor BPR1K653 and further demonstrated its efficacy in targeting various types of cancers in vitro. Our pervious x-ray cocrystallography studies had demonstrated the physical interactions between the precursor compound of BPR1K653 and Aurora kinases [42], and the current in vitro kinase inhibition study has confirmed the target specificity of BPR1K653. Consistent with the molecular changes observed in cells treated with Aurora-B kinase specific siRNA oligos and with different pan-Aurora kinase inhibitors such as VX680 and SNS-314 [14,43,44], BPR1K653 treatment also induces endo-replication of cells and reduces amount of phosphorylated Histone H3 present in cells. In addition, BPR1K653 is able to induce cancer cell apoptosis but not autophagy ( Figure S2), which is the common result in cells treated with Aurora kinase inhibitors [43]. Interestingly, BPR1K653 is active in all of the tested p53-wildtype/-negative/-mutant cancer cell lines at low nano-molar concentrations (IC 50 ,20 nM), despite limited ability of another pan-Aurora kinase inhibitor VX680 to induce endo-replication and subsequent apoptosis has been shown in cancer cells with normal p53-dependent post-mitotic checkpoint function in other study [14]. Taken together, BPR1K653 is selectively inhibiting Aurora kinases, and unlike VX680, it is able to target various types of cancer cells regardless of their p53 status.
Drug resistance is a common problem in the management of neoplastic diseases, and the effectiveness of many chemotherapeutic drugs is limited by the fact that they are substrates for the efflux pump MDR1 (P-gp170). For example, the Aurora kinase inhibitor AZD1152/AZD1152-HQPA (Barasertib) was shown to be the substrate of MDR1 [24]. Moreover, our reference Aurora kinase inhibitors, VX680 (Tozasertib) and PHA-739358 (Danusertib), were previously shown ineffective in targeting the MDR1expressing SA-Dx5 (doxorubicin resistant), MB-231-PTX and H460-PTX (both paclitaxel resistant) cancer cells by other investigators [25]. In this study, BPR1K653 was shown to be equally effective against two KB-derived MDR1-positive cancer cell lines (KB-VIN10 and KB-S15) and one NTUB1-dervided MDR1-positive cancer cell line (NTU0.017) in vitro. This feature is distinct from those of the well-characterized Aurora kinase inhibitors, VX680 and PHA-739358, because our tested MDR1positive cancer cells are more resistant to these chemotherapeutic agents than their parental MDR1-negative cells. Indeed, coincubation of the MDR1 inhibitor, verapamil, was shown to be effective in re-sensitizing the MDR1-expressing cancer cells to both VX680 and PHA-739358, whereas the same treatment could not enhance the sensitivity to BPR1K653 in neither MDR1negative (KB and NTUB1) nor MDR1-expressing cells (KB-VIN10, KB-S15 and NTU0.017). Importantly, BPR1K653 is also effective in inhibiting the growth of both MDR1-negative KB and MDR1-expressing KB-VIN10 cancer cells in vivo, further supporting the hypothesis that over-expression of the common drug efflux pump MDR1 could not interfere with the efficacy of BPR1K653 in targeting cancer cells. Since chemotherapeutic compounds such as paclitaxel, vincristine (anti-microtubule agents), doxorubicin (DNA intercalating agent), tretinoin (all-trans retinoic acid), mitoxantrone, VP-16 (topoisomerase II inhibitors) and imatinib (tyrosine kinase inhibitor) are all substrates of the drug efflux pump MDR1, the use of BPR1K653 may be beneficial in patients that are resistant to the above compounds after prolonged therapeutic treatments [30,31,32].
It has been known that most newly-developed anti-cancer compounds that perform well in vitro, do not progress to the clinical stage due various factors such as unfavorable pharmacokinetic properties and reduced potency in vivo. In this study, we have shown that BPR1K653 exhibits favorable pharmacokinetic properties in vivo. The maximum achievable plasma concentration of BPR1K653 (10 mM, 5463 ng/mL) after a single administration at a dosage of 5 mg/kg to rat is more than 80-fold and 200-fold above the in vitro kinase inhibition IC 50 of Aurora-A and -B kinase respectively. Even though at 24 h after dosing, the plasma levels of BPR1K653 (2 ng/mL) was still high enough to inhibit the activity of both Aurora-A and Aurora-B kinase. In addition, the high Volume of distribution at the steady state (Vss) value (10.6 l/kg) indicates that the distribution of BPR1K653 into deep compartments, including tumor and tissues is expected. Taken together, these favorable pharmacokinetic properties suggest that BPR1K653 dosing once a day is sufficient for continuous inhibition of the activity of both Aurora-A and Aurora-B kinase.
In conclusion, BPR1K653 is a potent pan-Aurora kinase inhibitor that is able to target cancer cells regardless of their tissue origins, MDR1 or p53 status. These key features distinguish this compound from other previously developed Aurora kinase inhibitors and anti-cancer compounds. At the molecular level, results of this study suggest that BPR1K653 can be used as a tool to study the molecular functions of Aurora kinases in the MDR1induced drug resistant cancer cells in the future. As BPR1K653 exhibits favorable pharmacokinetic properties in animal models, further evaluations are warranted to determine whether BPR1K653 is also effective in clinical situations.

Ethics statement
The animals used in this study were housed and the experiments were carried out at an International Association for Assessment and Accreditation of Laboratory Animal Care-accredited animal facility at the National Health Research Institutes, Tainan, Taiwan R.O.C.. The Institutional Animal Care and Use Committees for Biotechnology and the National Health Research Institutes approved uses of animals in these studies (approval number: NHRI-IACUC-099070).

Kinase inhibition assay
Aurora-A and Aurora-B kinase -The recombinant GST-tagged Aurora-A (residues S123-S401) containing kinase domain was expressed in Sf9 insect cells. The recombinant full length Histagged Aurora-B (residues M1,A344) was purchased from Invitrogen (PV3970). The kinase assay were carried out in 96well plates with the tested compound at either 37uC (Aurora-A) for 90 min or 30uC (Aurora-B) for 120 min.
ALK -The recombinant His-tagged ALK (residues V1058-P1620) containing kinase domain was expressed in Sf9 insect cells. The kinase assay was carried out in 96-well plates with the tested compound at 30uC for 120 min.
CHK1/2 -The recombinant His-tagged CHK1 (residues M1-T476) or CHK2 (residues M1-L543) containing kinase domain were expressed in Sf9 insect cells. The kinase assay was carried out in 96-well plates with the tested compound at 30uC for 120 min.
c-Met -The recombinant GST-tagged c-Met (residues K956-S1390) containing kinase domain was expressed in Sf9 insect cells. The kinase assay was carried out in 96-well plates with the tested compound at 30uC for 120 min.
EGFR -The recombinant GST-tagged EGFR (residues G696-G1022) containing kinase domain was expressed in Sf9 insect cells. The kinase assay was carried out in 96-well plates with the tested compound at 37uC for 60 min.
FLT3 -GST-tagged FLT3-KDWT containing the FLT3 kinase catalytic domain (residues Y567,S993) were expressed in Sf9 insect cells. The FLT3WT Kinase-Glo assays were carried out in 96-well plates at 30uC for 4 h with the tested compound.
VEGFR1/2 -The recombinant GST-tagged VEGFR1 (residues R781-I1338) or VEGFR2 (residues V789-V1356) containing kinase domain were expressed in Sf9 insect cells. The kinase assay was carried out in 96-well plates with the tested compound at 30uC for 120 min.
Composition of the reaction buffers used in different kinase inhibitory assays is described in Figure S3.

Clonogenic assay
Two hundred cells in logarithmic growth phase were seeded in a 6-well plate. The cells were exposed to various concentrations of the test drugs for a three-generation period. At the end of the incubation period, cells were fixed and stained with 50% ethanol containing 0.5% methylene blue for 30 min. The plates were washed five times with water and allowed to air-dry. Colonies were countered manually. The IC 50 value resulting from 50% inhibition of cell growth was calculated graphically as a comparison with the growth of the control group. Each value represents the average of at least three independent experiments run in triplicates.

Cell cycle analysis
Cell cycle progression was monitored using flow cytometry. After drug treatment, cells were trypsinized, washed with PBS and fixed in 80% ethanol at 220uC for 1 h. The fixed cells were stained with propidium iodide (containing RNase) at room temperature in the dark for 20 min. The DNA content was determined by a fluorescence-activated cell sorting IV flow cytometer (BD Biosciences, Franklin Lakes, NJ). For each analysis, 10,000 cells were counted and the percentage of cells in each phase was calculated using the ModFit LT software (Verity Software House, Topsham, ME). Experiments were repeated independently at least three times.

Real-time Caspase-3/-7 activity imaging
Caspase-3/-7 activity was analyzed with the MagicRed TM DEVD real-time caspase activity detection kit (catalog number #935, Immunochemistry Technologies LLC, Bloomington, MN). Briefly, cells were cultured in chamber-slides and incubated with test agents for various durations. Cells were then incubated with the Caspase-3/-7 substrate MR-(DEVD2) in culture medium for 1 hour, and then co-incubated with Hoechst 33342 for 15 min. Cells were viewed with a UV-enabled inverted-microscope at an excitation wavelength of 540 nm-560 nm and emission at 610 nm. Experiments were repeated independently at least two times.

Visualization of apoptosis by the TUNEL assay
Under in vitro conditions, cells were seeded and cultured in 8well chamber-slides, and treated with various compounds. The treated cells were washed with PBS, fixed with 4% paraformal-dehyde for 30 min on ice, and permeabilized with PBST at room temperature. Apoptotic cells were stained by the TUNEL agent using the TMR (red) In-Situ Apoptosis Detection Kit (catalog number #12156792910, Roche Diagnostic, Mannheim, Germany). Cells were counterstained with DAPI to detect the nucleus, and examined by fluorescence microscopy. Amount of red fluorescence labeled (DNA fragmented) cells were counted and percentage of apoptotic cells were calculated as follows: Amount of the red fluorescence labeled cells 4 Total cells available6100. Experiments were repeated independently at least two times.
Under in vivo conditions, tumors were dissected from the euthanized mice and instantly stored under 280uC. Tumor tissue sections were prepared from the use of cryostats (Leica Microsystems, Buffalo Grove, IL), and subsequently fixed with ice-cold methanol. Tissue sections were stained by the TUNEL reagent using Fluorescent (green) In-Situ Apoptosis Detection Kit (catalog number #11684795910, Roche Diagnostic, Mannheim, Germany). Cells were counterstained with DAPI to detect nucleus, and examined by fluorescence microscopy. Amount of green fluorescence labeled (DNA fragmented) cells were counted and percentage of apoptotic cells were calculated as follows: Amount of green fluorescence labeled cells 4 Total cells available6100. Experiments were repeated independently at least two times. Drug treatments and monitoring of the in vivo antitumor activity BPR1K653 was dissolved completely in a vehicle mixture of DMSO/cremophor/saline (1:2:7). Selected dose of BPR1K653 was decided base on the following conditions: 1/2 of the dosage that caused noticeable body weight loss (.10%) in the treated mice during toxicity study. In the KB xenograft study, when the size of a growing tumor reached $75 mm 3 , the xenograft tumorbearing nude mice were treated with either BPR1K653 or VX680 i.p. (5 mice per treatment group) at a dosage of 15 mg/kg or 30 mg/kg, respectively, for 5 days/week for 2 consecutive weeks.

Immunohistochemistry
Tumors were harvested and instantly stored at 280uC. Frozen cryostat sections were fixed with ice-cold methanol for 10 min. After washing with PBS, endogenous peroxidase was blocked using 3% hydrogen peroxide in TBS for 5 min. Immunostaining process was carried out according to the user's manual of the ABC Peroxidase Staining Kit (Pierce Biotechnology, Rockford, IL). Briefly, the tissues were incubated with a protein-blocking solution for 20 min, and subsequently stained with an anti-phosphorylated Histone H3 (Ser10) polyclonal antibody for 1 hour at room temperature. Then, the samples were incubated with the ABC reagent for 30 min, and subsequently incubated with the metal enhanced DAB substrate. The sections were counterstained with hematoxylin.

Pharmacokinetic studies of BPR1K653 in rats
Male Sprague-Dawley rats weighing 300-400 g each (8-12 weeks old) were obtained from BioLASCO, Taiwan Co., Ltd., Ilan, Taiwan. Animals were surgically prepared with a jugularvein cannula one day prior to dosing and fasted overnight (,18-20 h) prior to dosing. Water was available ad libitum throughout the experiment. Single 5 mg/kg dose of BPR1K653, as a DMA/ PEG (20/80, v/v) solution, was separately administered to groups of 3 rats each intravenously by a bolus injection via the jugularvein cannula. At 0 (prior to dosing), 2, 5, 15 and 30 min, and at 1, 2, 4, 6, 8 and 24 h after dosing, a blood sample (0.15 mL) was collected from each animal via the jugular-vein cannula and stored in ice (0-4uC). Plasma was separated from the blood by centrifugation (14,000 g for 15 min at 4uC in a Beckman Model Allegra TM 6R centrifuge) and stored in a freezer (220uC). All samples were analyzed for the parent drug by LC-MS/MS. LC/ MS/MS conditions: The chromatographic system consisted of an Agilent 1200 series LC system and an Agilent ZORBAX Eclipse XDB-C8 column (5 mm, 3.06150 mm) was connected to a MDS Sciex API3000 tandem mass spectrometer, which was equipped with a Turbo V TM ESI in the positive scanning mode at 600uC. Data was acquired via the multiple reactions monitoring (MRM) system. The MS/MS ion transitions were monitored at m/z of 541.4/106.4 for BPR1K653. The collision energy of 58.0 V was used for the analyst, BPR1K653. A gradient HPLC method was employed for the separation. Mobile phase A consisted of water containing 0.1% formic acid, and mobile phase B consisted of acetonitrile. The gradient profile was shown as follows (min/%B): 0.0-1.2/5, 1.3-3.9/95, 4.0-5.0/5. The flow rate was set to be 1.5 mL/min. The auto-sampler was programmed to inject 15 mL sample aliquots in every 5 min. The retention time of BPR1K653 was 2.39 min. Plasma concentration data were analyzed with noncompartmental method.

Statistical analysis
For all statistical analysis, values were expressed as mean 6 SD. Values were compared using Student's t-test. P,0.05 was considered significant.