TAK-960, a Novel, Orally Available, Selective Inhibitor of Polo-Like Kinase 1, Shows Broad-Spectrum Preclinical Antitumor Activity in Multiple Dosing Regimens

Published OnlineFirst December 21, 2011; DOI: 10.1158/1535-7163.MCT-11-0762

Molecular Cancer Preclinical Development Therapeutics

TAK-960, a Novel, Orally Available, Selective Inhibitor of Polo-Like Kinase 1, Shows Broad-spectrum Preclinical Antitumor Activity in Multiple Dosing Regimens

Yuichi Hikichi1, Kohei Honda1, Kouki Hikami1, Hitoshi Miyashita1, Isao Kaieda1, Saomi Murai1, Noriko Uchiyama1, Maki Hasegawa1, Tomohiro Kawamoto1, Takashi Sato1, Takashi Ichikawa1, Sheldon Cao2, Zhe Nie2, Lilly Zhang2, Johnny Yang3, Keisuke Kuida3, and Erik Kupperman3

Abstract Polo-like kinase 1 (PLK1) is a serine/threonine protein kinase involved in key processes during mitosis. Human PLK1 has been shown to be overexpressed in various human cancers, and elevated levels of PLK1 have been associated with poor prognosis, making it an attractive target for anticancer therapy. TAK-960 [4-[(9- cyclopentyl-7,7-diﬂuoro-5-methyl-6-oxo-6,7,8,9-tetrahydro-5H-pyrimido[4,5-b][1,4]diazepin-2-yl)amino]-2- ﬂuoro-5-methoxy-N-(1-methylpiperidin-4-yl) benzamide] is a novel, investigational, orally bioavailable, potent, and selective PLK1 inhibitor that has shown activity in several tumor cell lines, including those that express multidrug-resistant protein 1 (MDR1). Consistent with PLK1 inhibition, TAK-960 treatment caused

accumulation of G2–M cells, aberrant polo mitosis morphology, and increased phosphorylation of histone H3

(pHH3) in vitro and in vivo. TAK-960 inhibited proliferation of multiple cancer cell lines, with mean EC50 values

ranging from 8.4 to 46.9 nmol/L, but not in nondividing normal cells (EC50 >1,000 nmol/L). The mutation status of TP53 or KRAS and MDR1 expression did not correlate with the potency of TAK-960 in the cell lines tested. In animal models, oral administration of TAK-960 increased pHH3 in a dose-dependent manner and significantly inhibited the growth of HT-29 colorectal cancer xenografts. Treatment with once daily TAK-960 exhibited significant efficacy against multiple tumor xenografts, including an adriamycin/paclitaxel-resistant xenograft model and a disseminated leukemia model. TAK-960 has entered clinical evaluation in patients with advanced cancers. Mol Cancer Ther; 11(3); 700–9. 2011 AACR.

Introduction addition to these general cellular functions, human PLK1 The polo-like kinases (PLK) are evolutionarily con- is overexpressed in various human cancers, including served serine/threonine kinases that were ﬁrst described head and neck squamous cell carcinoma, non-small cell in the so-called polo mutants of Drosophila melanogaster lung cancer, oropharyngeal carcinoma, melanoma, ovar- that failed to undergo normal mitotic progression (1). ian and endometrial carcinoma, and prostate cancer; Since then, many important functions of PLK1 have been elevated PLK1 has been associated with poor prognosis discovered, such as activation of cell division cycle 2, (2, 5, 6). Several reports have shown that PLK1 depletion chromosome segregation, centrosome maturation, bipo- by short interfering RNA or PLK1 inhibition by small- lar spindle formation, regulation of the anaphase-promot- molecule inhibitors resulted in cell-cycle arrest at mitosis ing complex, and execution of cytokinesis (2–4). In and induced apoptosis in cancer cell lines, but not in normal diploid cells or nondividing cells (7–10). Currently PLK1 inhibitors are being evaluated in registered clinical trials (4). To date, clinical responses have been reported

Authors' Afﬁliations: 1Takeda Pharmaceutical Company Ltd., Fujisawa, with intravenous volasertib (11). Japan; 2Takeda California, Inc., San Diego, California; and 3Millennium Antimitotic agents that target tubulin, such as taxanes, Pharmaceuticals, Inc., Cambridge, Massachusetts have been clinically validated; however, these drugs have Note: Supplementary data for this article are available at Molecular Cancer several drawbacks including development of peripheral Therapeutics Online (http://mct.aacrjournals.org/). neuropathy (12), the need for pretreatment to reduce The current address for S. Cao is Viva biotech Ltd., Shanghai, China. anaphylaxis (13), and the potential for tumors to become refractory or resistant due to multidrug resistance (6). Corresponding Author: Yuichi Hikichi, Takeda Pharmaceutical Company Ltd., 26-1 Muraoka Higashi 2-chome Fujisawa Kanagawa 251-8555, Consequently, there is a need for novel antimitotic drugs Japan. Phone: 81-466-32-1910; Fax: 81-466-29-4412; E-mail: that target nonmicrotubule proteins such as the mitotic [email protected] kinases (10). Here we report the preclinical evaluation of doi: 10.1158/1535-7163.MCT-11-0762 TAK-960 [4-[(9-cyclopentyl-7,7-diﬂuoro-5-methyl-6-oxo- 2011 American Association for Cancer Research. 6,7,8,9-tetrahydro-5H-pyrimido[4,5-b][1,4]diazepin-2-yl)

700 Mol Cancer Ther; 11(3) March 2012

Antitumor Activity of TAK-960, a Selective PLK1 Inhibitor

amino]-2-fluoro-5-methoxy-N-(1-methylpiperidin-4-yl) minutes, blocked in StartingBlock TBS-T (Thermo Scien- benzamide], a potent and selective PLK1 inhibitor that tific) for 30 minutes, followed by incubation with anti- has oral bioavailability and shows single-agent antitumor bodies against a-tubulin and g-tubulin (Abcam) for 2 activity in a variety of tumor cell lines and xenograft hours. Alexa fluorophore–conjugated antibodies (Invitro- models of solid tumors and hematologic malignancies. gen Corporation) were used as secondary antibodies for detection, and 40,6-diamidino-2-phenylindole (DAPI) was Materials and Methods used for nuclei detection. Compounds Analysis of cell proliferation TAK-960 was synthesized at Takeda California, Inc., Cells were seeded into 96-well plates at 3,000 to 30,000 San Diego, CA. Paclitaxel was purchased from Bristol- cells per well in appropriate medium plus 10% fetal calf Myers Squibb. AraC was purchased from Sigma Aldrich. serum. After 24 hours, cells were treated with serial dilutions of TAK-960, and 72 hours later, the number of Cell culture viable cells was assessed using the CellTiter-Glo Assay Human tumor cell lines were obtained from the (Promega). Calculation of EC50 values and statistical anal- American Type Culture Collection with the exception ysis were done using GraphPad Prism software (Version of A2780 (ovarian cancer), SW620 (colorectal cancer), 5.01; GraphPad Software). and MRC5 (normal lung fibroblast; all 3 from Dainippon Sumitomo Pharma Co., Ltd.); KARPAS299 (lymphoma; Animal model studies Deutsche Sammlung von Mikroorganismen und Zell- Mice were housed and maintained within the faci- kulturen GmbH). K562ADR was generated at Takeda lity at TPC in accordance with the Takeda Experi- Pharmaceutical Company Ltd., (TPC) by exposing the mental Animal Care and Use Committee approved K562 parent cell line to doxorubicin at 10 nmol/L for protocol. Athymic nude mice (BALB/cAJc1-nu/nu) 8 days, 15 nmol/L for 11 days, 20 nmol/L for 4 days, or severe combined immunodeficiency (SCID) mice 30 nmol/L for 9 days, and 50 nmol/L for 10 days. (C.B17-Icr-scid/scid Jcl) of approximately 5 weeks of age MV4-11luc cells were established at TPC by transfecting were obtained from CLEA Japan, Inc. Nonobese dia- a firefly luciferase expression vector (Promega Corp.) betic (NOD)-SCID mice (NOD.CB17-Prkdcscid/J) were into MV4-11 cells. The tissue culture medium recom- obtained from Charles River Laboratories Japan, Inc. mended by suppliers was used for cultivation of each For subcutaneous-implanted tumor xenograft models, cell line. Commercially obtained cells were not authen- nude mice or SCID mice (for K562ADR leukemia xeno- ticated by the authors. grafts) were injected with 5 106 cells per mouse. When tumor volumes reached approximately 200 mm3,mice Biochemical kinase inhibition assays were randomly assigned to treatment groups. Tumor The inhibitory activity of TAK-960 was assessed by volumes and body weights were measured approxi- the TR-FRET (fluorescence resonance energy transfer) mately twice weekly throughout the study. Tumor assay, which measured the ATP-dependent phosphory- volume was calculated using calipers and the equation 2 lation of a biotinylated substrate peptide correspond- (L W )/2, in which L and W refer to the larger and smal- ing to residues 2,470 through 2,488 of the mTOR ler dimensions collected at each measurement, respec- protein (Biotin-AGAGTVPESIHSFIGDGLV). A total tively. Complete regression (CR) was defined as tumors of 288 kinases were screened for TAK-960 inhibition reduced to 13.5 mm3 or less and partial regression (PR) (1 mmol/L) using HotSpot technology (Reaction Bio- was defined as tumors reduced by more than 50% of their 3 logy Corporation) and IC50 values for the selected initial size but that were more than 13.5 mm . For the kinases were determined. disseminated leukemia study, NOD/SCID mice were pretreated intraperitoneally with TM-b1, an anti-CD122 Cell-cycle DNA analysis antibody (0.05 mg/mouse; BD Biosciences), to deplete HT-29 colorectal cancer cells were seeded in 12-well natural killer cells. The next day, mice were sublethally tissue culture plates, cultured overnight at 37 C, and irradiated with 3 Gray of X-ray, after which human AML incubated for 48 hours with dilutions of TAK-960. cells (MV4-11luc) were injected via tail vein. In vivo optical Cells were harvested by trypsin treatment and fixed images were obtained with an IVIS 200 optical imaging in 70% ethanol at 4 C. After treatment with RNase, system (Caliper Life Sciences) twice weekly. One hundred cells were stained with propidium iodide and DNA fifty mg/kg of D-luciferin (Promega) dissolved in PBS content was measured using a FACSCalibur System were administered to animals by intraperitoneal injection (BD Biosciences). 10 minutes before each image acquisition. Significant luminescence was detected around hind limb bone joints Immunocytochemistry on day 22 after implantation and vehicle, AraC, or TAK-960 HT-29 cells were treated with serial dilutions of TAK- was administered. AraC and paclitaxel were diluted in 960 for 24 hours and fixed with 4% paraformaldehyde for saline for intraperitoneal dosing and TAK-960 was sus- 10 minutes, permeabilized with 0.5% Triton X-100 for 5 pended in 0.5% methyl cellulose for oral dosing.

www.aacrjournals.org Mol Cancer Ther; 11(3) March 2012 701

Hikichi et al.

Figure 1. TAK-960 is a potent and selective inhibitor of PLK1. A, chemical structure of TAK-960. B, inhibition curves of TAK-960 under the condition of 3 mmol/L ATP (*) and 1,000 mmol/L ATP (*). Data represent the mean of 22 samples. C, histogram of kinase selectivity. From a panel of 288 protein kinases, the number of kinases in which activity was inhibited by TAK-960 at 1,000 nmol/L is plotted by percent inhibition compared with controls. Percentages of more than 100% were considered to indicate complete inhibition.

pHH3 analysis pharmacodynamic (PD) marker using the tumor lysates The PathScan Phospho-Histone H3 (serine 10; pHH3) by the pHH3 ELISA as described above. sandwich ELISA was carried out as per the manufac- turer’s instructions (Cell Signaling Technology) using in vitro cell lysates or in vivo tumor tissue lysates. The Results increase in pHH3 was shown as the fold increase in pHH3 Kinase inhibition and selectivity of TAK-960 normalized to the value in a vehicle control. Tissue sam- TAK-960 was designed and synthesized as a novel ples from tumor xenografts were fixed in 10% formalin PLK1 inhibitor that binds to the ATP-binding pocket of before paraffin embedding. Sections were rehydrated and PLK1 (Fig. 1A). Data reported herein using the TR-FRET incubated with anti-pHH3 antibody. Peroxidase-conju- assay confirmed that TAK-960 is a potent PLK1 inhibitor gated antibody was used as a second antibody and with a mean IC50 of 1.5 nmol/L against the kinase domain detected by diaminobenzidine staining. Two representa- of PLK1 at 3 mmol/L ATP (Fig. 1B). The ATP concentration tive pictures were taken from each section, and pHH3- in living cells is much higher, in the millimolar range, positive cells were counted using an AxioVision system which significantly reduces the inhibitory activity of pure (Carl Zeiss). ATP-competitive inhibitors. In the presence of a high ATP concentration (1,000 mmol/L), the inhibitory activity of ¼ Pharmacokinetic/pharmacodynamic studies TAK-960 was reduced less than 5-fold (IC50 6.5 nmol/L; Female athymic nude mice (Hsd:Athymic Nude- Fig. 1B). Foxn1nu) from Harlan Labs, Inc. were housed and main- To determine the specificity of TAK-960 across the tained within the Comparative Medicine department at kinome, a panel of 288 kinases was screened by the the Sidney Kimmel Cancer Center (SKCC, San Diego, CA) HotSpot technology. Of the kinases tested, 243 (84%) were in accordance with SKCC Institutional Animal Care and inhibited by less than 20% and 6 (PLK1, PLK2, PLK3, FAK, Use Committee, and state and federal guidelines for the MLCK, and FES) were inhibited by more than 80% at 1,000 humane treatment and care of laboratory mice. Tumors nmol/L TAK-960 (Fig. 1C); PLK1 showed the greatest were collected at multiple time points after oral admin- inhibition (101.6%). To further examine the selectivity of istration of TAK-960 and stored on dry ice or at 80 C. TAK-960 for inhibition of PLK1, IC50 values for TAK-960 Homogenization was done using Beadbeater (Biospec inhibition of the 6 kinases that were inhibited by more Products, Inc.) with sample buffer containing 2% SDS. than 80% were determined in the presence of 10 mmol/L TAK-960 concentrations were determined by extraction ATP (Table 1). The IC50 value for inhibition of full-length from mouse plasma and tumor homogenates by protein PLK1 protein by TAK-960 was 0.8 nmol/L in this assay, precipitation, followed by liquid chromatography–tan- which was more than 20-fold lower than the next lowest dem mass spectrometry. pHH3 was determined as a IC50 value (PLK2: 16.9 nmol/L).

702 Mol Cancer Ther; 11(3) March 2012 Molecular Cancer Therapeutics

Antitumor Activity of TAK-960, a Selective PLK1 Inhibitor

Table 1. Enzyme selectivity proﬁle of TAK-960 of pHH3 at 24 hours after TAK-960 treatment clearly indicated a decrease in viability at 72 hours after TAK- on selected kinases 960 treatment (Fig. 2B). EC50 values for the increase of pHH3 and the loss of viability were 9.8 and 8.4 nmol/L, Kinase % inhibition at IC50 (nmol/L) respectively, similar to the IC50 value of 6.5 nmol/L for 1,000 nmol/L PLK1 kinase enzyme inhibition at 1,000 mmol/L ATP. PLK1 101.6 0.8 The effects of TAK-960 on chromosome dynamics were PLK2 88 16.9 examined by immunoﬂuorescent microscopy. The far-left PLK3 81.8 50.2 panel of Fig. 2C shows a representative control dimethyl FAK/PTK2 98.4 19.6 sulfoxide (DMSO)-treated HT-29 cell in G2–M phase. In MLCK/MYLK 90.3 25.6 contrast, accumulation of aberrant spindles was observed FES/FPS 81.5 58.2 following treatment with TAK-960 at 10 or 30 nmol/L. At 282 kinases <80 100 nmol/L or more concentration, the appearance and distribution of phenotypes consisted of monopolar spindles with characteristic polo spindle morphology (15). TAK-960 mechanism of action Mitotic arrest has been implicated as an outcome of Effect of TAK-960 on proliferation in human tumor PLK1 inhibition (14). Flow cytometry analysis showed cell lines, including those with TP53 or KRAS that treatment of HT-29 colorectal cancer cells with TAK- mutations or MDR1 overexpression 960 resulted in concentration-dependent accumulation of TAK-960 was tested for its ability to inhibit the prolif- cells in G2–M phase (Fig. 2A). A correlation between the eration of selected human cancer cells. The mean EC50 accumulation of G2–M–phase cells and the increase in values of proliferation inhibition in the 18 cell lines tested pHH3 was observed. In addition, the pattern of increase are shown in Fig. 3A, along with the mutation status of

Figure 2. TAK-960 induces accumulation of aberrant mitotic cells in HT-29 cells. A, DNA profiles of asynchronously cultured HT-29 cells treated with a variety of concentrations indicated in the figure for 48 hours were evaluated by flow cytometry. B, accumulation of pHH3 in HT-29 cells was determined by pHH3 ELISA at 24 hours (*). Growth inhibition at 72 hours (*) was measured by CellTiter-Glo as described in Materials and Methods. Data are shown as mean SD of 3 samples. C, HT-29 cells were treated for 24 hours with different concentrations of TAK-960 and analyzed by immunofluorescence for a-tubulin (green), g-tubulin (red), and chromatin (blue).

www.aacrjournals.org Mol Cancer Ther; 11(3) March 2012 703

Hikichi et al.

Figure 3. TAK-960 inhibits proliferation of human cancer cell lines regardless of TP53 and KRAS mutation and MDR1 expression status. A, B, and D, CellTiter-Glo assay was carried out on a panel of human cell lines treated with TAK-960 (concentration range ¼ 2–1,000 nmol/L) for 72 hours. A, the list of EC50 in cell lines with various TP53 and KRAS status, M, mutant, WT, wild-type, NA, not available. Shading indicates tumor origin. B, representative concentration– response curve of MRC5. MRC5 human normal lung ﬁbroblasts were seeded at 3 103 (proliferating, *) and 3 104 (quiescent, *) cells per well (in 100 mL of media) at 24 hours before TAK-960 treatment. Data represent the mean of 3 samples. C, MDR1 expression in selected cancer cell lines was determined by Western blotting. D, concentration– response curve of K562 leukemia cells (&) and K562ADR (&)for doxorubicin, paclitaxel, and TAK- 960. Data are shown as mean SD of 3 samples.

TP53 and KRAS. TAK-960 inhibited the proliferation of Because many chemotherapeutic agents, such as pac- these human cancer cell lines in a concentration-depen- litaxel and doxorubicin, are susceptible to the activity of dent manner, with mean EC50 values ranging from 8.4 to multidrug-resistant protein 1 (MDR1), the effect of 46.9 nmol/L (Fig. 3A). Recent studies have suggested a MDR1 expression on the potency of TAK-960 was eval- correlation between sensitivity to PLK1 inhibition and uated. Western blot analysis was carried out on selected TP53 or KRAS mutation status (9, 16, 17). However, no cell lines to conﬁrm expression levels of MDR1 (Fig. 3C). statistical differences were observed when EC50 values Although expression levels of MDR1 were high in HCT- were compared between cell lines with wild-type alleles 15 and COLO320DM colorectal cancer cells, EC50 values and those with mutated alleles of TP53 or KRAS (P > 0.05; were similar to those in cell lines that did not express 2-tailed Mann–Whitney test; Supplementary Fig. S1). MDR1 (Fig. 3A). The lack of effect of MDR1 expression Concentration–response curves for the representative was further shown with the doxorubicin-resistant normal human ﬁbroblast cells (MRC5) are shown K562ADR leukemia cell line. Interestingly, although in Fig. 3B. Although TAK-960 reduced the viability of this cell line was established as a doxorubicin-resistant proliferating MRC5 cells, viability at 1,000 nmol/L was cells, it showed resistance not only to doxorubicin but maintained at more than 50%. Furthermore, TAK-960 also to paclitaxel when compared with its parent cell did not affect the viability of quiescent MRC5 cells line K562 (Fig. 3D). Indeed, Western blot data showed (EC50 >1,000 nmol/L). that K562ADR expressed MDR1 at a level similar to that

704 Mol Cancer Ther; 11(3) March 2012 Molecular Cancer Therapeutics

Antitumor Activity of TAK-960, a Selective PLK1 Inhibitor

Figure 4. PK/PD and antitumor activity analysis for TAK-960 in mice bearing HT-29 colorectal cancer cells. A, TAK-960 was administered orally at 5, 10, and 30 mg/kg. The observed TAK-960 concentration for plasma (*) and tumor (*) of HT-29 tumor-bearing nude mice was plotted. The observed induction of pHH3 as a PD marker in HT-29 tumors was also plotted (&). Data represent the mean of 5 samples. The value of pHH3 is expressed as fold change (0 hour ¼ 1). B, tumor volume or C, body weight, nude mice bearing HT-29 tumors (n ¼ 5) were treated orally once daily for 21 days (*,QD21) or once daily for 3 days per week for 3 weeks (&,QD3/w3W) or twice weekly 2 days apart for 3 weeks (~,Q2D2/w3W) at the dosage indicated in figure. Arrows in the figure indicate the days of administration for each schedule. No regrowth was observed in 3 of 5 tumors in 10 mg/kg administered groups on once daily for 21 days schedule until the measurement on day 86. Data represent the mean SD of 5 samples. of HCT-15 and COLO320DM (Fig. 3C). TAK-960 inhib- (Fig. 4A, &) was observed at 24 hours post dose. Thus, ited proliferation of K562ADR cells with an EC50 value the peak in pHH3 levels occurred after the tumor concen- similar to that of the parent cell line, K562 (Fig. 3A tration of TAK-960 had peaked, which occurred between and D). 4 to 8 hours post dose. This result is consistent with the fact that increases in pHH3 are a consequence of mitotic Pharmacokinetic/pharmacodynamic and antitumor arrest induced by PLK1 inhibition by TAK-960 treatment. activity relationships in xenograft models Exposure to TAK-960 (AUC) in plasma and tumor was To confirm that TAK-960 inhibited PLK1 activity linearly correlated with dose. In addition, the peak of in vivo, mice bearing established HT-29 colorectal tumors pHH3 response (Emax) at 10 and 30 mg/kg was similar, were administered a single oral dose (5, 10, or 30 mg/kg, whereas total PD response (AUE) over the time was rel- Fig. 4A) of TAK-960. Plasma and tumor concentrations atively dose dependent (Supplementary Table S1). of TAK-960 were measured at various time points using To further evaluate relationship between PD response a high-performance liquid chromatography/tandem mass and antitumor activity, TAK-960 was evaluated in mice spectrometry assay. Plasma exposure to TAK-960 bearing established HT-29 colorectal tumor xenografts increased in an approximately dose-proportional manner using various regimens (Fig. 4B and C). Doses of 6.25 or and was partitioned preferentially into the tumor (Fig. 4A, 10 mg/kg were administered orally once daily for 21 *) compared with the circulating levels in plasma (Fig. 4A, days (QD21). Cyclic intermittent treatment schedules *). As shown previously, increased pHH3 is a marker were also tested in the same model. TAK-960 was of PLK1 inhibition/mitotic arrest. Levels of pHH3 were administered at 20 or 25 mg/kg orally twice weekly measured by ELISA in HT-29 colorectal tumor samples. 2daysapartfor3weeks(Q2D2/w3w). Doses of The maximum increase in tumor pHH3 over baseline 20 mg/kg were administered orally on a schedule of

www.aacrjournals.org Mol Cancer Ther; 11(3) March 2012 705

Hikichi et al.

once daily for 3 consecutive days per week for 3 weeks ules, tumor recurrence was observed by 48 days after (QD3/w3w). All TAK-960 treatment schedules the ﬁrst dosing day (Supplementary Fig. S2). inhibited tumor growth in a dose-dependent manner (Fig. 4B) without obvious body weight loss (Fig. 4C); TAK-960 antitumor activity in various models of however, the degree of inhibition was schedule depen- human cancer dent. Tumor regression was greater in the continuous Nine additional tumor xenograft models, HCT116 dosing group (10 mg/kg, QD 21; CR: 3 of 5 mice at day (colorectal cancer), PC-3 (prostate cancer), BT474 (breast 86, Fig. 4B). In the intermittent dose group, 1 CR and cancer), A549 (lung cancer), NCI-H1299 (lung cancer), 1 PR were observed at the 20 mg/kg TAK-960 dose NCI-H1975 (lung cancer), A2780 (ovary cancer), and (QD3/w3w, Fig. 4B). An additional xenograft study MV4-11 (myeloid leukemia) were tested using the opti- was carried out to evaluate another intermittent dosing mal dose of 10 mg/kg once daily for 2 weeks selected on schedule (Q2D2/w3w), in which the total amount the basis of experiments with HT-29 colorectal cancer of TAK-960 administrated was similar to that achieved xenografts. In each of these models (Fig. 5A–H), sub- with daily dosing; however, in all intermittent sched- stantial antitumor activity and good tolerability was

Figure 5. Single-agent efﬁcacy of TAK-960 in various tumor xenograft models using human cancer cell lines. A, HCT116 colorectal cancer; B, PC-3 prostate cancer; C, BT474 breast cancer; D, A549 lung cancer; E, NCI-H1299 lung cancer; F, NCI-H1975 lung cancer; G, A2780 ovary cancer; and H, MV4-11 leukemia cells were inoculated subcutaneously in nude mice or SCID mice. Vehicle (*, 0.5% MC) or TAK-960 (*, 10 mg/kg) administered orally once daily for 14 days. Data are shown as mean SD of 5 samples. I, survival of mice inoculated with MV4-11luc was assessed following treatment with vehicle, AraC, and TAK-960. AraC 5 mg/kg was administrated intraperitoneally once daily for 5 days per week for 2 weeks (QD5/w2w). TAK-960 (7.5 mg/kg) was administered orally once daily for 9 consecutive days (QD9).

706 Mol Cancer Ther; 11(3) March 2012 Molecular Cancer Therapeutics

Antitumor Activity of TAK-960, a Selective PLK1 Inhibitor

observed (no treatment-related deaths occurred in this TAK-960 in vivo efficacy in MDR1-expressing regimen). tumors On the basis of the significant tumor growth inhibition K562ADR was established as an adriamycin-resistant observed in the subcutaneously implanted MV4-11 leukemia cell line and also showed resistance to paclitaxel human leukemia model (Fig. 5H), TAK-960 efficacy was (Fig. 3D). Paclitaxel induces G2–M arrest and increases also explored in a model of disseminated leukemia using pHH3, as observed with PLK1 inhibition. Therefore, we MV4-11luc cells, which may more closely reflect the path- compared changes in PD markers between TAK-960 and ogenesis of the human disease. In this model, treatment paclitaxel in K562 or K562ADR cells that were inoculated with TAK-960 at 7.5 mg/kg orally once daily for 9 days subcutaneously in SCID mice. After a single oral admin- showed a significant increase in median survival com- istration of TAK-960 or intraperitoneally administered pared with vehicle (39.5 vs. 25 days, respectively; P ¼ paclitaxel, pHH3-positive cells in tumor tissues were 0.0055, Log-rank (Mantel–Cox) test; Fig. 5I). Disseminated detected and counted immunohistochemically. In K562 tumor growth was also evaluated by in vivo whole body tumors, the maximum increase in pHH3 was observed at fluorescence imaging 14 days after the first day of dosing; 24 hours post dose at 30 mg/kg TAK-960, consistent with inhibition of tumor growth was observed in animals HT-29 colorectal cancer cells (Fig. 6A) and 20 mg/kg of treated with TAK-960 but not in animals treated with intraperitoneal paclitaxel. TAK-960 also showed a similar AraC (Supplementary Fig. S3A). Animals treated with pattern of pHH3 response in K562ADR tumors, but pac- vehicle or AraC showed increasing total body flux (i.e., litaxel did not induce pHH3 in MDR1-expressing tumors tumor growth) during 2 weeks of treatment; flux was (Fig. 6A–C). Indeed, TAK-960 at 10 mg/kg (once daily for lower in animals treated with TAK-960 and reached 6 days per week for 2 weeks, orally) inhibited tumor statistical significance on day 14 (P ¼ 0.0017; Supplemen- growth in the MDR1-expressing K562ADR-bearing leu- tary Fig. S3B). kemia xenograft model, whereas paclitaxel at 10 mg/kg

Figure 6. TAK-960 shows PD activity in paclitaxel-resistant model. A, pHH3, as a PD marker for TAK-960 and paclitaxel in mice bearing K562 or K562ADR leukemia cells, was analyzed by immunohistochemistry. TAK-960 30 mg/kg was administered orally and 20 mg/kg of paclitaxel was administered intraperitoneally. The time course of the PD response was assessed up to 48 hours after the single dose of TAK- 960 (B) or paclitaxel (C) in mice bearing K562 (*) and K562ADR (*) tumors. Data are shown as the mean of 3 samples. D, the single-agent efﬁcacy study was carried out using the K562ADR-bearing xenograft model. Vehicle (*, 0.5% methyl cellulose) or TAK-960 (*, 10 mg/kg) was administered orally, once daily for 6 days per week for 2 weeks (QD6/w2w). Paclitaxel (~, 10 mg/kg) was administered intraperitoneally once daily for 5 days per week for 2 weeks (QD5/w2w). Data are shown as the mean SD of 3 samples.

www.aacrjournals.org Mol Cancer Ther; 11(3) March 2012 707

Hikichi et al.

(once daily for 5 days per week for 2 weeks, intraperito- found that mutation of Arg136 to Gly in the PLK1 ATP- neally), which is the maximum tolerated dose of this binding pocket could alter the sensitivity to PLK1 inhibi- schedule in our preliminary study, did not show signif- tors in a compound-specific manner. Further efforts will icant tumor growth inhibition (Fig. 6D). be required to define potential patient stratification factors to help identify the tumors most vulnerable to TAK-960 and to determine whether combinatorial approaches may Discussion lead to durable tumor regression. Targeting the mitotic machinery is a proven strategy in Our results showed that TAK-960 inhibited intracellu- the treatment of cancer. In this article, we have described lar PLK1 and led to dysregulation of the cell cycle. The the preclinical activities of TAK-960, a novel, potent, and mechanism of action of TAK-960 was indicated in HT-29 selective PLK1 inhibitor. The inhibitory activity of pure colorectal cancer cells, in which concentrations that inhib- ATP-competitive inhibitors is significantly reduced in the ited growth at 72 hours were able to induce pHH3 at presence of high ATP concentrations. However, the inhib- 24 hours and cause aberrant mitosis in a dose-dependent itory activity of TAK-960 in the presence of a high ATP manner, suggesting that pHH3 would be a predictive concentration (1,000 mmol/L) was reduced by less than 5- marker of antitumor activity. Consistent with in vitro ¼ fold (IC50 6.5 nmol/L), suggesting that TAK-960 could observations, TAK-960 treatment in mice bearing HT-29 maintain potent PLK1 inhibition in living cells. Further- colorectal cancer or K562ADR leukemia xenografts more, dissociation of TAK-960 after binding to PLK1 induced pHH3 in tumors and showed significant anti- showed a slow off-rate for TAK-960 (data not shown), tumor efficacy in vivo. Recently, a number of PLK1 resulting in a longer residence time at the target and a substrates have been reported (25, 26). Although the avail- prolonged pharmacologic effect in cells or tissues. TAK- ability of an appropriate antibody against PLK1 substrates 960 inhibited proliferation in all cancer cell lines tested but for immunohistochemistry or ELISA is still limited, the did not inhibit nondividing normal fibroblasts, indicating reliability of PLK1 substrates as potential predictive or that TAK-960 has few off-target effects. diagnostic biomarkers should be compared with pHH3 Our data are consistent with others suggesting that in future studies. sensitivity to PLK1 inhibition did not necessarily correlate Early preclinical studies of taxanes showed poor with the doubling time of cancer cells. Degenhardt and correlations between efficacy and peak mitotic index colleagues (17) suggested that sensitivity to the PLK1 (27, 28). Although there seemed to be a poor correlation inhibitor GSK461364 was associated with loss of p53 between doses and peak pHH3 levels (Emax), we found a function. Luo and colleagues (16) reported that K-ras good correlation between doses of TAK-960 and total mutant DLD1 cells were more sensitive to PLK1 knock- PD responses over the time for pHH3 (AUE) in the HT- down or to the PLK1 inhibitor BI2536 compared with 29 xenograft model. Given that the peak level of PD K-ras wild-type DLD1 cells. However, TAK-960 at nano- response was the same for 10 and 30 mg/kg of TAK-960 molar concentrations exhibited strong antiproliferative at 24 hours post dose (Supplementary Table S1), the activity in all tumor cell lines tested, regardless of the Emax PD response could be saturated at 10 mg/kg. mutation status of TP53 or KRAS; no statistically signif- Considering the relationship between pharmacokinetic icant correlation was observed for sensitivity and muta- (PK),PD,andTAK-960dose,itislikelythatconsecutive tion status in vitro. Following mitotic arrest by an antimi- daily 10 mg/kg administration would result in an totic agent, cells can either die or exit from mitosis in a effective PD response without toxic drug accumulation. tetraploid state, known as mitotic slippage. Because the QD dosing at 10 mg/kg showed superior antitumor genetic status of TP53 mainly affects the sensitivity of the activity compared with intermittent dosing in the HT- cells that have exited from mitosis (17), examining wheth- 29 colorectal cancer xenograft model without obvious er cells are likely to undergo mitotic slippage might body weight loss. Using the QD schedule, TAK-960 improve the accuracy of TP53 as a biomarker. exhibited significant antitumor activity as a single agent Elucidating the mechanism of drug sensitivity should in several solid tumor xenograft and hematologic malig- take drug efflux into account. Although the debate about nancy models, including MDR1-expressing tumors. In the clinical relevance of efflux-mediated drug resistance is the disseminated leukemia model using MV4-11luc currently ongoing, HCT-15 and COLO320DM cells express cells, QD treatment with TAK-960 significantly extend- MDR1 and are resistant to several chemotherapy drugs ed survival, whereas treatment with AraC did not. In (18–20). A number of chemotherapy-resistant cell lines K562ADR-bearing tumors, which overexpress MDR1, have been developed using K562 leukemia cells (21–23). thelackofPDresponseorefficacywaslikelydueto In this study, we also produced adriamycin-resistant K562 drug efflux. As mentioned above, the ability to exit from cells and confirmed that these cells overexpressed MDR1 mitosis may be another influencing factor for drug and were resistant to paclitaxel. We found that TAK-960 sensitivity. However, such resistance could also be retained its effectiveness, irrespective of MRD1 status, in observed by a decrease in pHH3. Taken together, PD HCT-15, COLO320DM, and K562ADR tumor cells. Resis- response for pHH3, especially AUE, could be a valu- tance to PLK1 inhibition might also be due to mutations able biomarker not only for prognosis but also for in the ATP-binding site of PLK1. Scutt and colleagues (24) determining a suitable dose and treatment schedule.

708 Mol Cancer Ther; 11(3) March 2012 Molecular Cancer Therapeutics

Antitumor Activity of TAK-960, a Selective PLK1 Inhibitor

Through PK/PD and efficacy studies, we established Acknowledgments the optimal dosing regimen for TAK-960 in the mouse The authors thank Stephen Mosley and Natalie Beavan of FireKite, Ltd. for their editorial assistance during the development of this manuscript, xenograft model; however, an optimal human dose and which was funded by Millennium Pharmaceuticals, Inc; Dr. Akira Shibuya schedule will need to be determined in clinical studies. and Dr. Shuichi Furuya for the scientific discussion and advice; the entire TAK-960 team members for their support and many contributions to this In summary, TAK-960 has high oral bioavailability, project. thus providing the potential for convenient, controlled, and flexible dosing schedules compared with intrave- Grant Support nously administered drugs. TAK-960 is also a potent, This research was supported by Takeda Pharmaceutical Company, selective PLK1 inhibitor with activity in various tumor Ltd., Takeda California, Inc., and Millennium Pharmaceuticals, Inc. The costs of publication of this article were defrayed in part by the cell lines in vitro and in vivo, including cells that over- payment of page charges. This article must therefore be hereby marked express MDR1. TAK-960 is currently undergoing phase I advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate evaluation in adult patients with advanced cancers. this fact.

Disclosure of Potential Conﬂicts of Interest All authors are employees of Takeda Pharmaceutical Company, Ltd., Received September 23, 2011; revised December 2, 2011; accepted or Takeda California, Inc., or Millennium Pharmaceuticals, Inc. December 9, 2011; published OnlineFirst December 21, 2011.

References 1. Sunkel CE, Glover DM. Polo, a mitotic mutant of Drosophila displaying 16. Luo J, Emanuele MJ, Li D, Creighton CJ, Schlabach MR, West- abnormal spindle poles. J Cell Sci 1988;89:25–38. brook TF, et al. A genome-wide RNAi screen identifies multiple 2. Strebhardt K, Ullrich A. Targeting polo-like kinase 1 for cancer therapy. synthetic lethal interactions with the Ras oncogene. Cell 2009;137: Nat Rev Cancer 2006;6:321–30. 835–48. 3. Archambault V, Glover DM. Polo-like kinases: conservation and diver- 17. Degenhardt Y, Greshock J, Laquerre S, Gilmartin AG, Jing J, Richter M, gence in their functions and regulation. Nat Rev Mol Cell Biol et al. Sensitivity of cancer cells to PLK1 inhibitor GSK461364A is 2009;10:265–75. associated with loss of p53 function and chromosome instability. Mol 4. Lens SM, Voest EE, Medema RH. Shared and separate functions of Cancer Ther 2010;9:2079–89. polo-like kinases and aurora kinases in cancer. Nat Rev Cancer 18. Naito M, Matsuba Y, Sato S, Hirata H, Tsuruo T. MS-209, a quinoline- 2010;10:825–41. type reversal agent, potentiates antitumor efficacy of docetaxel in 5. Eckerdt F, Yuan J, Strebhardt K. Polo-like kinases and oncogenesis. multidrug-resistant solid tumor xenograft models. Clin Cancer Res Oncogene 2005;24:267–76. 2002;8:582–8. 6. Szakacs G, Paterson JK, Ludwig JA, Booth-Genthe C, Gottesman 19. van Rensburg CE, Joone GK, O'Sullivan JF. Clofazimine and B4121 MM. Targeting multidrug resistance in cancer. Nat Rev Drug Discov sensitize an intrinsically resistant human colon cancer cell line to 2006;5:219–34. P-glycoprotein substrates. Oncol Rep 2000;7:193–5. 7. Steegmaier M, Hoffmann M, Baum A, Len art P, Petronczki M, Krssak 20. Xia Z, Zhu Z, Zhang L, Royal C, Liu Z, Chen Q, et al. Specific reversal of M, et al. BI 2536, a potent and selective inhibitor of Polo-like kinase 1, MDR1/P-gp-dependent multidrug resistance by RNA interference in inhibits tumor growth in vivo. Curr Biol 2007;17:316–22. colon cancer cells. Oncol Rep 2008;20:1433–9. 8. Liu X, Lei M, Erikson RL. Normal cells, but not cancer cells, survive 21. Assef Y, Rubio F, Colo G, del Monaco S, Costas MA, Kotsias BA. severe Plk1 depletion. Mol Cell Biol 2006;26:2093–108. Imatinib resistance in multidrug-resistant K562 human leukemic cells. 9. Guan R, Tapang P, Leverson JD, Albert D, Giranda VL, Luo Y. Small Leuk Res 2009;33:710–6. interfering RNA-mediated Polo-like kinase 1 depletion preferentially 22. Dumontet C, Jaffrezou JP, Tsuchiya E, Duran GE, Chen G, Derry WB, reduces the survival of p53-defective, oncogenic transformed cells et al. Resistance to microtubule-targeted cytotoxins in a K562 leuke- and inhibits tumor growth in animals. Cancer Res 2005;65:2698–704. mia cell variant associated with altered tubulin expression and poly- 10. Jackson JR, Patrick DR, Dar MM, Huang PS. Targeted anti-mitotic merization. Bull Cancer 2004;91:E81–112. therapies: can we improve on tubulin agents? Nat Rev Cancer 2007;7: 23. Peng X, Gong F, Xie G, Zhao Y, Tang M, Yu L, et al. A proteomic 107–17. investigation into adriamycin chemo-resistance of human leukemia 11. Bug G, Schlenk RF, Muller-Tidow C, Lubbert M, Kramer A, Fleischer F, K562 cells. Mol Cell Biochem 2011;351:233–41. et al. Phase I/II study of BI 6727 (volasertib), an intravenous Polo-like 24. Scutt PJ, Chu MLH, Sloane DA, Cherry M, Bignell CR, Williams DH, kinase-1 (PLK1) inhibitor, in patients with acute myeloid leukemia Eyers PA. Discovery and exploitation of inhibitor-resistant aurora and (AML): Results of the dose finding for BI 6727 in combination with polo kinase mutants for the analysis of mitotic networks. J Biol Chem low-dose cytarabine. Blood 2010;116:abstr 3316. 2009;284:15880–93. 12. Wolf S, Barton D, Kottschade L, Grothey A, Loprinzi C. Chemotherapy- 25. Santamaria A, Wang B, Elowe S, Malik R, Zhang F, Bauer M, et al. The induced peripheral neuropathy: prevention and treatment strategies. Plk1-dependent phosphoproteome of the early mitotic spindle. Mol Eur J Cancer 2008;44:1507–15. Cell Proteomics 2010;10:M110 004457. 13. Hennenfent KL, Govindan R. Novel formulations of taxanes: a review. 26. Liu XS, Song B, Liu X. The substrates of Plk1, beyond the functions in Old wine in a new bottle? Ann Oncol 2006;17:735–49. mitosis. Protein Cell 2010;1:999–1010. 14. Lane HA, Nigg EA. Antibody microinjection reveals an essential role for 27. Schimming R, Mason KA, Hunter N, Weil M, Kishi K, Milas L. Lack human polo-like kinase 1 (Plk1) in the functional maturation of mitotic of correlation between mitotic arrest or apoptosis and antitumor centrosomes. J Cell Biol 1996;135:1701–13. effect of docetaxel. Cancer Chemother Pharmacol 1999;43: 15. Gilmartin AG, Bleam MR, Richter MC, Erskine SG, Kruger RG, Madden 165–72. L, et al. Distinct concentration-dependent effects of the polo-like 28. Milross CG, Mason KA, Hunter NR, Chung WK, Peters LJ, Milas L. kinase 1-specific inhibitor GSK461364A, including differential effects Relationship of mitotic arrest and apoptosis to antitumor effect of on apoptosis. Cancer Res 2009;69:6969–77. paclitaxel. J Natl Cancer Inst 1996;88:1308–14.

www.aacrjournals.org Mol Cancer Ther; 11(3) March 2012 709

TAK-960, a Novel, Orally Available, Selective Inhibitor of Polo-Like Kinase 1, Shows Broad-spectrum Preclinical Antitumor Activity in Multiple Dosing Regimens

Yuichi Hikichi, Kohei Honda, Kouki Hikami, et al.

Mol Cancer Ther 2012;11:700-709. Published OnlineFirst December 21, 2011.

Updated version Access the most recent version of this article at: doi:10.1158/1535-7163.MCT-11-0762

Supplementary Access the most recent supplemental material at: Material http://mct.aacrjournals.org/content/suppl/2011/12/20/1535-7163.MCT-11-0762.DC1

Cited articles This article cites 27 articles, 8 of which you can access for free at: http://mct.aacrjournals.org/content/11/3/700.full#ref-list-1

Citing articles This article has been cited by 2 HighWire-hosted articles. Access the articles at: http://mct.aacrjournals.org/content/11/3/700.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at Subscriptions [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://mct.aacrjournals.org/content/11/3/700. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.