TAK-901 | Mek signal

Biological Characterization of TAK-901, an Investigational, Novel, Multitargeted Aurora B Kinase Inhibitor

Pamela Farrell1, Lihong Shi1, Jennifer Matuszkiewicz1, Deepika Balakrishna1, Takashi Hoshino5, Lilly Zhang2, Sarah Elliott1, Robyn Fabrey1, Bumsup Lee1, Petro Halkowycz1, BiChing Sang3, Seigo Ishino5, Toshiyuki Nomura1, Mika Teratani6, Yoshikazu Ohta5, Charles Grimshaw1, Bheema Paraselli4, Takashi Satou6, and Ron de Jong1

Abstract

Protein kinases Aurora A, B, and C play essential roles during mitosis and cell division, are frequently

elevated in cancer, and represent attractive targets for therapeutic intervention. TAK-901 is an investigational, multitargeted Aurora B kinase inhibitor derived from a novel azacarboline kinase hinge-binder chemotype. TAK-901 exhibited time-dependent, tight-binding inhibition of Aurora B, but not Aurora A. Consistent with Aurora B inhibition, TAK-901 suppressed cellular histone H3 phosphorylation and induced polyploidy. In varioushuman cancer cell lines, TAK-901 inhibited cell proliferation with effective concentration values from 40 to 500 nmol/L. Examination of a broad panel of kinases in biochemical assays revealed inhibition of multiple kinases. However, TAK-901 potently inhibited only a few kinases other than Aurora B in intact cells, including FLT3 and FGFR2. In rodent xenografts, TAK-901 exhibited potent activity against multiple human solid tumor types, and complete regression was observed in the ovarian cancer A2780 model. TAK-901 also displayed potent activity against several leukemia models. In vivo biomarker studies showed that TAK-901 induced pharmacodynamic responses consistent with Aurora B inhibition and correlating with retention of TAK-901 in tumor tissue. These preclinical data highlight the therapeutic potential of TAK-901, which has entered phase I clinical trials in patients within a diverse range of cancers. Mol Cancer Ther; 12(4); 460–70. ©2013 AACR.

Introduction
Aurora kinases are key mitotic cell-cycle regulators, ensuring an orderly and accurate execution of mitosis and cell division (1–4). Aurora A localizes to centrosomes and spindle poles, and is required for mitotic spindle assembly and centrosome maturation (5), whereas Aurora B is a chromosome passenger protein essential for phos- phorylation of histone H3, chromosome segregation, and cytokinesis. Aurora B inactivation leads to abnormal chromosome alignment, silencing of the spindle check-

Authors’ Afﬁliations: 1Discovery Biology; 2Analytical Sciences and DMPK; 3Structural Biology; 4Chemistry, Takeda California, Inc., California; 5Pharmaceutical Research Laboratories; and 6Exploratory Translational

Medicine, Biology Research Laboratory, Takeda Pharmaceutical Company Ltd, Japan
Note: Supplementary data for this article are available at Molecular Cancer Therapeutics Online (http://mct.aacrjournals.org/).

Current address for B. Lee: Kolon Life Sciences, Inc., Gwacheon-Si, gyeonggi-do, Korea; current address for L. Shi, Quanticel Pharmaceuticals Inc., San Diego, California; and current address for B. Paraselli: Chemveda Life Sciences, San Diego, California.
Corresponding Author: Ron de Jong, Takeda San Diego, Inc. 10410 Science Center Drive, San Diego, CA 92121. Phone: 858-731-3719; Fax: 858-550-0526; E-mail: [email protected]
doi: 10.1158/1535-7163.MCT-12-0657
©2013 American Association for Cancer Research.

point function, and induces the cells to exit mitosis with- out completion of cell division (1–3, 6–11).
Aurora A and B are overexpressed in many human cancers and are associated with poor prognosis, progres- sion, and genetic instability (2, 4, 5, 11–13), making them attractive therapeutic targets. In vitro studies suggest that elevated Aurora A may promote oncogenesis (14, 15), and its overexpression is associated with chromosome instability and clinically aggressive disease (5, 15, 16). Elevated expression of Aurora B has also been seen in multiple human tumors and cancer cell lines and corre- lates with disease severity and a poor prognosis (11, 12, 15, 17–22).
Preclinical data have been reported for numerous Auro- ra kinase inhibitors that show either selective or pan- Aurora kinase inhibition. However, many pan-Aurora kinase inhibitors generate a similar polyploid cellular phenotype, which is a hallmark of Aurora B inactivation (11, 14). More than 10 small-molecule inhibitors have entered early clinical development, including selective inhibitors of Aurora A (23–25), B, or C (26–29) and pan- Aurora kinase inhibitors (30–35).
Here, we describe the biologic characterization of TAK- 901, a potent multitargeted Aurora B kinase inhibitor based on a novel chemical scaffold (Fig. 1A). TAK-901 shows tight-binding inhibition of Aurora B in biochemical studies and inhibits cellular Aurora B kinase as evidenc- ed by induction of polyploidy and dose-dependent

TAK-901, a Novel Multitargeted Aurora B Kinase Inhibitor

Figure 1. A, chemical structure of TAK-901. B, TAK-901 inhibition of Aurora A and B kinase/coactivator complexes and kinetic data of the Aurora B/INCENP enzyme complex. C, kinase inhibition proﬁle of TAK-901. D, enzyme reaction progression curves showing TAK-901 time-dependent binding to (D) and dissociation (E) from Aurora B/INCENP enzyme complex. Vehicle (*), 6.3 nmol/L (O), 25 nmol/L (&), 100 nmol/L (^), or 200 nmol/L (D) TAK-901, respectively.

suppression of phosphorylation of histone H3, a cellular substrate of Aurora B. While TAK-901 shows inhibition of multiple kinases in enzymatic assays, in intact cells it shows potent inhibition of a limited number of cellular kinases including FLT3 and FGFR2. Finally, we show the activity of TAK-901 against a range of human solid tumor models and various leukemia models and a dose-depen- dent inhibition of histone H3 phosphorylation in the A2780 xenograft model.

Materials and Methods
Reagents
TAK-901 was synthesized at Takeda San Diego. Its discovery and synthesis will be described elsewhere.

Cell culture

All cell lines were obtained from the American Type Culture Collection (ATCC) except for A2780 (Sigma). All

Farrell etell lines were authenticated by ATCC or Sigma (mor- phology, growth curve analysis, isoenzyme species anal- ysis, DNA fingerprinting, and mycoplasma dectection) and used within 6 months of receipt or resuscitation. Cell lines were cultured according to the manufacturer’s specifications.

Aurora A and B kinase assays
Enzyme activities of Aurora A/TPX2 and Aurora B/INCENP complexes were assayed at room tempera- ture in 50 mmol/L HEPES, pH 7.3, 10 mmol/L NaCl, 10 mmol/L MgCl2, 0.01% Brij35, and 1 mmol/L dithiothrei- tol containing serially diluted TAK-901, and the product was quantified using IMAP detection reagents (Molecular Devices). Aurora A/TPX2 (2 nmol/L) was assayed with 100 nmol/L FL-Kemptide and 1 mmol/L ATP. Aurora B/INCENP (0.8 nmol/L) was assayed with 100 nmol/L 5-carboxy-fluorescein-GRTGRRNSI-NH2 (FL-PKAtide) and 10 mmol/L ATP. For time-dependent inhibition, Aurora B/INCENP was incubated with TAK-901 for
1 hour at room temperature followed by addition of 150 mmol/L ATP to initiate the reaction.

Aurora B/INCENP enzyme kinetic studies
A LabChip, off-chip mobility shift assay (Caliper Life Sciences) was used to measure the slow progress curve of TAK-901 binding to Aurora B/INCENP. To measure the on-rate, the enzyme reaction was initiated by addition of
0.1 nmol/L Aurora B/INCENP to a reaction mixture containing 1 mmol/L ATP, 2 mmol/L FL-PKAtide, and serial dilutions of TAK-901. Enzyme kinetics studies mea- suring TAK-901 binding to Aurora B/INCENP were conducted according to Morrison and Walsh (36), and the kinetic parameters were derived from the equations described therein.

Assays of multiple kinases
Twenty protein kinases were assayed at Takeda San Diego to determine inhibitory concentration (IC50) values using different assay technologies according to the man- ufacturer’s specifications. All kinase assays were con- ducted at room temperature and at a concentration of ATP equal to or below the Km for ATP for that particular kinase. Inhibitory activities of TAK-901 were measured against approximately 150 additional protein kinases by screening at 0.1 and 1 mmol/L using the kinase profiling services at Millipore and Invitrogen. Assay details are described by these providers. IC50 values for enzyme inhibition were estimated from these 2 data points and 0% and 100% inhibition values using XLfit4 MicroSoft Excel curve-fitting software. For kinases exhibiting esti- mated IC50 values of 300 nmol/L or less, IC50 values from full dose–response curves were determined.

Cell proliferation and viability assays
Cells were plated in 96-well microtiter plates and incu- bated with serial dilutions of TAK-901 for 72 hours. Cell proliferation was determined by ELISA analysis of bro-

modeoxyuridine (BrdUrd) incorporation into DNA (Roche Diagnostics). IMR-90 immortalized lung fibro- blasts were seeded in 96-well microtiter plates and cul- tured for 3 to 4 days until confluent. Cells were then incubated with serial dilutions of TAK-901 for 72 hours. The MTS assay was conducted as per the manufacturer’s instructions (Promega). The concentration of TAK-901 required to inhibit half of the maximal effective concen- tration (EC50) was determined from the dose–response curves.

Aurora B kinase assays in cells
Histone H3 phosphorylation assay. PC3 cells were incubated with TAK-901 for 4 hours. Proteins were sep- arated by SDS-PAGE, transferred to polyvinylidene difluoride membrane, and incubated with phospho-his- tone H3 (Millipore) and proliferating cell nuclear antigen (PCNA; Santa Cruz Biotechnology) and secondary IRDye- conjugated antibodies (Rockland). Blot images were collected on the Odyssey scanner (Li-Cor), and signals corresponding to PCNA and phosphorylated H3 were quantified using Li-Cor software.
Polyploidy analysis by FACS and immunocytochem- istry. HL60 cells were incubated with TAK-901 for 48 hours. Cells were harvested, fixed, stained with propi- dium iodide, and analyzed by flow cytometry. PC3 cells were treated for 48 hours, formalin-fixed, washed, and permeabilized. After blocking with 10% normal goat serum, actin was stained with FITC-phalloidin (Sigma- Aldrich), and nuclei were stained by washing in Hoechst dye (Calbiochem).

Inhibition of cellular tyrosine kinases
Cells were treated with TAK-901 for 4 hours before lysis. Western blot analyses were prepared with various antibodies: actin (Sigma), phospho-FGFR, FLT3, phos- pho-418-SRC, SRC, pY705-Stat3, and Stat3 (Cell Signal- ing), FGFR2 (Santa Cruz Biotechnology), phospho-Y412- ABL (Invitrogen), 4G10 pTyr monoclonal (Upstate), and ABL monoclonal (BD Pharmingen). Immunoblotting and quantification of detected (phospho) proteins were con- ducted as described above. Inhibition of MEK1/2 signal- ing was determined in A375 cells treated for 1 hour, lysed, and incubated with SureFire Phospho-ERK1/2 cocktail
for 4 hours at 25◦C. Proximity luminescent signals were read on a Perkin Elmer Envision reader. FAK phosphor-
ylation was measured in PC3 cells incubated for 30 min- utes followed by adhesion to the fibronectin-coated surface for 2 hours. FAK-Y397 phosphorylation was deter- mined using an ELISA kit according to the manufacturer’s specifications (Calbiochem). Inhibition of AXL autopho- sphorylation was determined in HEK293 cells stably expressing full-length AXL fused to GFP using an ELISA kit that detects phosphorylated AXL (Cell Signaling).

Xenograft models
Nude mice or rats were inoculated subcutaneously in the flank with suspensions of human cancer cells. When

TAK-901, a Novel Multitargeted Aurora B Kinase Inhibitor

Tumors in implanted mice reached 100 to 150 mm3 (200 mm3 in rats), animals were randomized into treatment groups of 6 (rats) or 8 to 10 (mice) before treatment initiation. Tumor volume was measured twice weekly and was calculated using the formula: tumor volume (mm3) ¼ (a × b2)/2, in which “a” is the largest diameter and “b” is the smallest diameter of the measured tumor.
Animal health was closely monitored, and distressed animals were immediately euthanized. TAK-901 or vehicle was administered intravenously twice daily on dosing days in 12% Captisol, 25 mmol/L citrate, pH 3.0, at 5 mL/gm body weight. Antitumor efficacy was expressed as mean relative growth of treated versus control tumors (%T/C) using the formula: [(T — T0)/(C — C0)] × 100, in
which C and T are mean control and drug-treated tumor
volumes, and C0 and T0 are initial tumor volumes, respec- tively. Regression percentage was calculated using the formula: [1-(T/T0)] × 100, in which T and T0 are treated and are initial tumor volumes, respectively.

In vivo pharmacodynamic studies
Nude rats bearing A2780 tumors averaging 250 to 500 mg received an intravenous dose of TAK-901 as described above. Plasma samples were collected by terminal cardiac puncture under CO2 anesthesia. Tumors were dissected and snap-frozen at —80◦C. Whole tumor tissue lysates were prepared in 62.5 mmol/L Tris-HCl, pH 7, 1% SDS, and 10% glycerol. Immunoblotting was conducted as described above. Plasma levels of TAK-901 were deter- mined after protein precipitation using acetonitrile.
Tumor samples were homogenized in 4 volumes of water using a mechanical homogenizer. Determination of TAK- 901 concentrations in plasma and tumor tissues was con- ducted by liquid chromatography/mass spectrometry. For polyploidy detection, dissected tumors were forma- lin-fixed, sectioned, and stained with hematoxylin and eosin using standard histologic procedures.

Results
Inhibition of Aurora A and B kinases
TAK-901 (Fig. 1A) is derived from the novel azacarbo- line kinase hinge binder. By optimization through struc- ture-based drug design using structural information from Aurora A kinase cocomplexes, TAK-901 was discovered as a potent inhibitor of Aurora kinases in biochemical assays. TAK-901 inhibited Aurora A/TXP2 and Aurora B/INCENP with IC50 values of 0.021 and 0.015 mmol/L, respectively, (Fig. 1B) and also inhibited Aurora A, B, and C kinases with comparable potencies without their coac- tivators (Fig. 1C). Following preincubation with Aurora B/INCENP in the presence of high ATP concentration, TAK-901 showed a 10-fold decrease in IC50 from 0.015 to 0.0017 mmol/L (Fig. 1B), which was dependent on the presence of the INCENP coactivator protein (data not shown). In enzyme kinetic studies, TAK-901 displayed time-dependent tight binding of Aurora B/INCENP with
a measured on-rate constant value of 5.6 × 105 M—1sec—1

(Fig. 1B). Enzyme reaction progression curves following substrate phosphorylation illustrate the slow binding of TAK-901 to Aurora B/INCENP through gradual loss of enzyme activity (Fig. 1D). Dissociation of TAK-901 from Aurora B/INCENP was shown by the slow recovery of enzymatic activity in an enzyme reaction progression curve (Fig. 1E) and was extremely slow with an off-rate of 1.3 × 10—5 sec—1 and an inhibitor enzyme complex half- life of 920 minutes at room temperature (Fig. 1B). The
affinity constant for TAK-901 binding to Aurora B/ INCENP was 0.02 nmol/L. Similar experiments with Aurora A/TPX2 failed to show time-dependent binding of TAK-901 to Aurora A (data not shown).

In vitro kinase inhibition profile
A total of 169 kinases were examined in 2-point con- centrations and dose–response screens. Of these, 80 kinases had IC50 values more than 1,000 nmol/L, whereas 89 and 56 kinases had IC50 values less than 1,000 nmol/L and less than 100 nmol/L, respectively (Fig. 1C). In con- trast to Aurora B/INCENP, TAK-901 did not exhibit time- dependent inhibition of several other kinases (LCK, c- SRC, and Aurora A/TPX2; data not shown). Several kinases potently inhibited by TAK-901 were identified by a cellular target-profiling method using affinity chro- matography of TAK-901 that was coupled to resin (Sup- plementary Fig. S1A). Notably, Aurora A and B were among the kinase targets exhibiting the strongest affinity for TAK-901. Clk1, Clk4, LRRK2, and SNF1LK were also high-affinity targets, and enzymatic inhibition by TAK- 901 of these and other identified kinases were confirmed by in vitro kinase assays.

Cellular biomarkers of Aurora B and inhibition of cell proliferation
In PC3 cells, TAK-901 suppressed histone H3 phos- phorylation in a dose-dependent manner with an EC50 value of 0.16 mmol/L (Fig. 2A). TAK-901 treatment caused polyploidy in PC3 cells visualized by multinuclear cells detected by immunofluorescence microscopy (Fig. 2B). Multiple concentrations of TAK-901 induced polyploidy in HL60 cells as shown by cell-cycle analysis using flow cytometry (Fig. 2C). Octaploid cells (8 n), which failed 2 cell divisions, were detected starting at 200 nmol/L. At concentrations above 3.2 mmol/L, TAK-901 caused accu- mulation of sub-G0 cell populations. These studies show that Aurora B kinase is a main target of TAK-901 in intact cells.
The effect of TAK-901 on cell proliferation was deter- mined by BrdUrd incorporation in cultured cell lines. For most cell lines, EC50 values were 50 to 200 nmol/L (Fig. 2D), which correlates well with cellular Aurora B inhibi- tion potencies. TAK-901 is a substrate of the P-glycopro- tein (PgP) drug efflux pump, as shown by the decrease in EC50 from 38 nmol/L in the uterine sarcoma MES-SA cells to more than 50 mmol/L in drug-resistant MES-SA/Dx5 expressing high levels of PgP. Colorectal cancer lines HCT15 and DLD1 also express PgP and were notably

Farrell et al.

Figure 2. TAK-901 inactivates cellular Aurora B kinase and inhibits cell proliferation. A, immunoblot analysis of histone H3 phosphorylation in PC3 cells treated with TAK-901. B, PC3 cells were incubated for 48 hours with dimethyl sulfoxide or 0.2 mmol/L TAK-901 and stained for actin (green) and DNA (blue). C, DNA content histograms of HL60 cells after incubation for 48 hours with a concentration dilution series of TAK-901. D, TAK-901 EC50 values for cell proliferation (DNA synthesis) inhibition in cells. HCT15, DLD1, and MES-SA/Dx5 cells express PgP.

more resistant. Proliferation of preconfluent normal lung fibroblast IMR-90 cells was inhibited with an EC50 of 88 nmol/L, whereas the EC50 for loss of viability in nondi- viding confluent cultures of IMR-90 was 2.4 mmol/L (Fig. 2D), thereby showing low cytotoxicity in nonproliferating cells that lack Aurora B expression.

Cellular kinase inhibition profiles
TAK-901 showed strong inhibition of phosphorylation of tyrosine kinase FGFR2 in a concentration-dependent manner in KATO-III cells and FLT3 in MV4–11 cells with EC50 values of 0.22 mmol/L and 0.25 mmol/L, respectively (Fig. 3A–C). This was equal to or 3-fold lower than the inhibition of cellular Aurora B kinase activity measured in the same cell line (Fig. 3C). In other cell models, TAK-901 exhibited less potent inhibition of cellular c-SRC, FAK, ABL, JAK2, B-Raf-V600E, and AXL kinases displaying EC50

values in the micromolar range despite the fact that many of these kinases are low nanomolar inhibitors in biochem- ical kinase assays (Fig. 3C and Supplementary Fig. S2).

Tumor growth inhibition in multiple xenograft tumors
The in vivo antitumor efficacy of TAK-901 was evalu-
ated in subcutaneous solid tumor and leukemia xenograft models using repeated and intermittent regimens of twice daily intravenous dosing. TAK-901 was usually admin- istered in cycles of 2 consecutive days per week. Tumor growth inhibition (TGI) by TAK-901 was statistically significant in many of these in vivo efficacy models in a dose-dependent manner (Table 1 and Fig. 4).
Tumor regressions were observed at the highest dose levels of TAK-901 in several xenograft models, including 95% at 45 mg/kg twice daily in (nude rat) ovarian A2780,

TAK-901, a Novel Multitargeted Aurora B Kinase Inhibitor

Figure 3. Proﬁle of various TAK-901 kinase targets. A, MV4-11 cells, 2 mmol/L FLT3/MTK inhibitor. B, KATO-III cells, 10 mmol/L Aurora B inhibitor. C, proﬁle of TAK-901 cellular activity against Aurora B kinase and several cross-reacting kinases. EC50 values derived from dose–response immunoblotting or cellular reporter.39% at 30 mg/kg twice daily in colorectal HCT116, and 58% at 30 mg/kg twice in acute myeloid leukemia (AML) HL60 models, respectively (Table 1). Complete regres- sions (95%) were observed post-dose in all A2780 xeno- graft animals, with partial responses (PR) in HCT116 and HL60 models.

In A2780 and AML MV4–11 nude mice xenograft mod- els, TAK-901 caused tumor stasis for 30 days at 30 and 40 mg/kg/injection, respectively (Table 1 and Fig. 4D). This was accompanied by significant increases in lifespan in the MV4–11 xenograft model (data not shown). Several models exhibited more modest TGI in response to TAK- 901, such as the H460 and H1299 lung cancer xenografts. Overall, TAK-901 was well tolerated at efficacious dose levels and caused little or no weight loss (Table 1).
Intensifying the dosing schedule from once weekly to 2 consecutive days per week showed marked improvement in antitumor activity in the nude rat A2780 xenograft model, as evidenced by tumor regression rather than TGI. The nude mouse A2780 xenograft model showed no difference in TGI between dosing 2 consecutive days per week or spaced days apart (Table 1).
Single-agent treatment with intraperitoneal 2 mg/kg irinotecan or i.v. 10 mg/kg twice daily TAK-901 resulted in TGI of approximately 50% in the HCT116 xenograft model (Fig. 4E). Combining both agents substantially

enhanced the TGI showing complete stasis until onset of outgrowth (Fig. 4E). The TAK-901/irinotecan combina- tion was associated with minimal weight loss. Similarly, nude mice with HL60 leukemia xenografts were treated intraperitoneally with 2 mg/kg daunorubicin and intra- venously with 15 mg/kg/injection TAK-901 (Fig. 4F). Single-agent activity of TAK-901 resulted in 75% TGI and irinotecan alone resulted in 20% TGI but was accompa- nied by 15% decrease in body weight. A notable increase in efficacy was observed when TAK-901 and daunorubi- cin were coadministered, leading to complete tumor stasis.

Aurora B kinase inhibition in A2780 xenograft tumors
Following a single intravenous dose of TAK-901, A2780 tumors in nude rats showed a strong suppression of histone H3 phosphorylation at 20 mg/kg, which returned to normal levels after 8 to 12 hours. At 40 mg/kg, histone H3 phosphorylation was more durably suppressed and did not return to control levels, remaining 72% inhibited at
12 hours post-dose (Fig. 5A). The pharmacodynamic response was associated with retention of TAK-901 in the tumor tissue. Although TAK-901 concentrations remained relatively constant in tumor tissues throughout the time course, its plasma levels declined to undetectable

Farrell et al.

Table 1. Antitumor activity of TAK-901 in xenograft models

in nude rats 40 mg/kg, b.i.d. i.v., qwk × 4 97c – ND 20 mg/kg, b.i.d. i.v., 2d on/5d off, wk × 3 93c –d 3 (3) 45 mg/kg, b.i.d. i.v., 2d on/5d off, wk × 3 >100c 95d 4 (3)
H460 SCLC in nude rats 20 mg/kg, b.i.d. i.v., 2d on/5d off, wk × 3 48 – 6 (19) 45 mg/kg, b.i.d. i.v., 2d on/5d off, wk 3 67 – 8 (19)
A2780 ovarian carcinoma 30 mg/kg, b.i.d. i.v., days 1,4,8,11 99c – 5 (6) in nude mice 30 mg/kg, b.i.d. i.v., 2d on/5d off, wk × 2 98c – 5 (4)

HCT116 colorectal carcinoma in nude mice

30 mg/kg, b.i.d. i.v., days 1,3,6,8,10,13,15 >100c 58e No BW loss

H1299 NSCLC in nude mice 15 mg/kg, b.i.d. i.v., 2d on/5d off, wk × 2 23 – No BW loss 30 mg/kg, b.i.d. i.v., 2d on/5d off, wk × 2 24 – No BW loss
HL60 AML in nude mice 15 mg/kg, b.i.d. i.v., 2d on/5d off, wk × 2 58a – No BW loss 30 mg/kg, b.i.d. i.v., 2d on/5d off, wk × 2 >100c 39f No BW loss
MV4–11 AML in nude mice 20 mg/kg, b.i.d. i.v., 2d on/5d off, wk × 3 74b – 3 (3) 40 mg/kg, b.i.d. i.v., 2d on/5d off, wk × 3 93c – 4 (3)
K562 CML in SCID mice 20 mg/kg, b.i.d. i.v., 2d on/5d off, wk × 3 63c – 14 (28) 40 mg/kg, b.i.d. i.v., 2d on/5d off, wk × 3 69c – 12 (28)
NOTE: TGI was calculated as the difference between control and treated tumor volumes at the end of treatment period or vehicle group discontinuation. Statistical signiﬁcance was calculated by one-way ANOVA. % Regression is deﬁned as percent reduction in treated versus initial tumor volume at the end of treatment period. % BW loss is deﬁned as the maximum percent BW change at nadir.
aP < 0.05; bP < 0.01; cP < 0.001.
dAt 20 mg/kg, 1 complete regression out of 6 rats, at 45 mg/kg, all animals achieved complete regression during post-dosing period. e4 of 8 showed PR and f1 of 5 showed PR.
Abbreviations: 2d on/5d off, 2 consecutive days of dosing per week; AML, acute myeloid leukemia; b.i.d, twice daily; BW, body weight; CML, chronic myeloid leukemia; ND, not determined; NSCLC, non–small cell lung carcinoma; SCID, severe combined immunodeﬁcient mice; SCLC, small cell lung carcinoma.

levels. In addition, TAK-901 caused polyploidy in nude mice A2780 xenograft tumors after repeated dosing (Fig. 5B).
Discussion
We describe here the biochemical and pharmacologic characterization of TAK-901, an investigational, novel, ATP-competitive, and multitargeted Aurora B kinase inhibitor. TAK-901 contains the azacarboline chemical scaffold that interacts with the Aurora kinase hinge region via 2 hydrogen bonds (data not shown) and represents the first potent azacarboline-based kinase inhibitor.
In biochemical assays, TAK-901 inhibited all 3 Aurora isoforms, A, B, and C, without their accessory proteins at low nanomolar concentrations in enzymatic assays. Similar potencies were observed when TAK-901 was tested against Aurora A and B in complex with their respective coactivating partners, TPX2 and INCENP. These enzyme complexes may more accurately resemble physiologic functions as Aurora A and B interact with their accessory proteins in the cellular environment, and these interactions affect their subcellular location, kinase activity, and protein conformations (37–40). We further showed with enzyme kinetic studies that TAK-901 bound

Aurora B/INCENP tightly in a time-dependent man- ner. Although TAK-901 exhibited potent inhibition of Aurora A/TPX2, this was not time-dependent. In con- trast, TAK-901 displayed a slow on-rate to inhibit Aurora B/INCENP and once formed, the TAK-901-bound Aurora B/INCENP complex dissociated slowly with a half-life of 920 minutes. This favorable biochemical property of extremely long residence time on Aurora B/INCENP was also reported recently for GSK1070916, an Aurora B/C- selective kinase inhibitor (26). Although the biochemical methods used were different, the TAK-901 dissociation rate seems of comparable magnitude with GSK1070916, whereas their kinetic analysis indicated that Aurora kinase inhibitors AZD1152 and VX-680/MK-0457 exhibit much faster dissociation rates (26). Therefore, the Aurora B tight-binding property may be less common among reported Aurora B kinase inhibitors.
Our results showed an apparent decreased competition with ATP substrate resulting from the slow relaxation of the tightly bound TAK-901–Aurora B/INCENP complex to a more weakly bound state from which TAK-901 can dissociate and ATP can effectively compete for binding to the active site. Perhaps the Aurora B tight binding causes the observed potent cellular activity. This notion was supported by compounds in the azacarboline series that

TAK-901, a Novel Multitargeted Aurora B Kinase Inhibitor

Figure 4. In vivo antitumor activity of TAK-901 in human tumor and leukemia xenograft models. Tumor-bearing animals were treated intravenously twice daily (b.i.d.) with either vehicle or TAK-901 on 2 consecutive days per week (A, C, and D) or every other day (B) for 2 or 3 cycles. E and F, combined effects of TAK-901 and therapeutic agents on tumor growth.

either lacked or possessed this tight-binding property (data not shown). Our cell proliferation studies showed that TAK-901 yielded a strong growth inhibitory effect across a range of human cancer cell lines. These cell proliferation effects were consistently associated with potency of Aurora B inactivation as measured by histone H3 phosphorylation inhibition in multiple cell lines and polyploidy induction in HL60 cells. Moreover, cell-cycle

analysis in several other cell lines also showed that the strength of cell proliferation inhibition was associated with the appearance of polyploidy cells (data not shown). Overall, Aurora B kinase inhibition was the consistent phenotype across cell lines suggesting that it may be the predominant cellular activity of TAK-901 despite its potent inhibitory activity against a range of kinases in biochemical assays.

Farrell et al.

Figure 5. In vivo effects of TAK-901 on Aurora B pharmacodynamic markers. A, histone H3 phosphorylation in nude rat A2780 xenograft tumors. A single dose of TAK-901 was administered intravenously and tumor tissue and plasma samples were harvested at various time points. Quantiﬁed and normalized histone H3 phosphorylation and total plasma and tumor concentrations of TAK-901. B, polyploidy in nude mice A2780 xenograft tumors. 30 mg/kg
TAK-901 was administered intravenously twice daily in 2 cycles of 2 consecutive days per week. Five days after last dose, tumors were dissected and stained with hematoxylin and eosin.
Our studies provide a comprehensive characterization of TAK-901 in vitro inhibitory properties through bio- chemical screening against a large panel of kinases and cellular target profiling. Multiple kinases other than Auro- ra kinases were identified that were potently inhibited by TAK-901. Besides Aurora A and B, CLK1 and CLK4 were the highest affinity kinase targets found in the cellular target profiling study. CLK kinases are part of signal transduction pathways that regulate pre-mRNA splicing. A specific inhibitor of CLK1 and CLK4 affects alternative splicing but has no effect on cancer cell growth (41). Many of the additional TAK-901 kinase targets are mediators of signaling pathways implicated in malignancy and can potentially contribute to the anticancer activity of TAK- 901. Because in vitro enzyme activity does not always correlate with potency in intact cells, TAK-901 was eval- uated in cellular models to quantify its inhibitory effects on specific kinases. In contrast to potent enzymatic inhi- bition, TAK-901 showed considerably less activity against AXL, ABL, c-SRC, FAK, and JAK2 in corresponding cellular signaling pathway models. FLT3 and FGFR2 receptor tyrosine kinases were potently inhibited by

TAK-901 in cells at levels close or comparable with Aurora B and may represent significant off-target activity. How- ever, the combined data showed a substantial window of cellular selectivity of TAK-901 for Aurora B versus other kinases. This selectivity for Aurora B in intact cells com- pared with other kinases potently inhibited in biochem- ical assays is presumably due to the Aurora B/INCENP tight-binding kinetics, as discussed above, whereas other kinases remain susceptible to more readily reversible competition with ATP substrate. Therefore, we speculate that the actual selectivity for Aurora B may be greater under physiologic conditions, such as high cellular ATP concentrations, as compared with selectivity measured in biochemical assays.
Activated cellular FLT3 kinase was also strongly inhib- ited by TAK-901. Mutations in FLT3 are associated with AML (42), suggesting that TAK-901 may have therapeutic potential in Flt-mutated AML (43). Tumors may only be partially sensitive to single-agent therapies, requiring interdiction of multiple protein kinases and other protein signaling targets for optimal anticancer therapy. Similar to TAK-901, many of the Aurora kinase inhibitors that have

TAK-901, a Novel Multitargeted Aurora B Kinase Inhibitor

progressed to clinical trials also inhibit multiple other kinases in vitro and these off-target activities may con- tribute to antitumor activities.
TAK-901 showed marked in vivo activity against a range
of leukemia and solid tumor xenograft models using intravenous administration and a well-tolerated schedule of repeated, intermittent dosing. TAK-901 showed high levels of growth inhibition in AML models, HL60, and MV4–11, and in ovarian (A2780) and colorectal cancer (HCT116) models, including prolonged tumor regres- sions. Similar to the observations in cell models, the antitumor effects of TAK-901 in the A2780 xenograft model are consistent with dose-responsive inhibition of Aurora B kinase. The Aurora B pharmacodynamic effects in these xenograft tumors correlated with the level of antitumor activity. Moreover, TAK-901 was effectively retained in tumor tissue correlating with the sustained pharmacodynamic effect and suggests that, for the treat- ment of solid tumors, TAK-901 will reach its intended target tissue. TAK-901 may have the potential to be com- bined with standard-of-care chemotherapy agents, irino- tecan (colorectal cancer) and daunorubicin (AML), based on preliminary results in 2 xenograft models. In vivo combination with TAK-901 resulted in decreased tumor growth compared with chemotherapy alone, offering the potential for enhanced antitumor activity. The ability to combine with Aurora B kinase inhibitors may be broad as several other Aurora kinase inhibitors have also shown potentiation with different antimitotic agents in cancer cells and xenograft models (34, 44, 45). Because many cancers are sustained by abnormalities in multiple pathways, combination of TAK-901 with stan- dard-of-care and/or other targeted agents with additive or synergistic mechanisms of action may increase treat- ment efficacy.
In summary, our results show that TAK-901 is a novel, potent, tight-binding Aurora B kinase inhibitor with mul-

titargeted inhibitory activity against other kinases, show- ing a typical Aurora B phenotype in cells and in vivo activity across multiple cancer cell lines and xenograft models. TAK-901 has entered phase I clinical trials in patients with a diverse range of cancers.

Disclosure of Potential Conﬂicts of Interest
No potential conflicts of interest were disclosed.

Authors’ Contributions
Conception and design: P. Farrell, L. Shi, J. Matuszkiewicz, Y. Ohta, B. Paraselli, T. Satou, R. de Jong
Development of methodology: P. Farrell, L. Shi, J. Matuszkiewicz, D. Balakrishna, P. Halkowycz, C. Grimshaw, R. de Jong Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): P. Farrell, L. Shi, J. Matuszkiewicz, T. Hoshino,
S. Elliott, R. Fabrey, B. Lee, S. Ishino, T. Nomura, M. Teratani, Y. Ohta,
C. Grimshaw, T. Satou, R. de Jong
Analysis and interpretation of data (e.g., statistical analysis, biostatis- tics, computational analysis): P. Farrell, L. Shi, J. Matuszkiewicz,
T. Hoshino, L. Zhang, S. Elliott, B. Lee, Y. Ohta, C. Grimshaw, T. Satou,
R. de Jong
Writing, review, and/or revision of the manuscript: P. Farrell, L. Shi,
J. Matuszkiewicz, T. Hoshino, L. Zhang, B. Lee, S. Ishino, T. Nomura,
Y. Ohta, C. Grimshaw, T. Satou, R. de Jong
Administrative, technical, or material support (i.e., reporting or orga- nizing data, constructing databases): P. Farrell, L. Shi, B. Lee, B. Sang,
R. de Jong
Study supervision: P. Farrell, B. Lee, Y. Ohta, C. Grimshaw, R. de Jong

Acknowledgments
The authors thank Shawn O’Connell, Misa Iwatani Cheryl Napier, Cheri Routt, Randall Jackson, and Beth Hollister for technical assistance. The authors also thank Catriona Scott and Stephen Mosley of FireKite for the writing assistance during the development of this publication.

Grant Support
This study was funded by the Millennium Pharmaceuticals, Inc.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Received July 19, 2012; revised December 11, 2012; accepted January 7,
2013; published OnlineFirst January 28, 2013.

References
1. Carmena M, Earnshaw WC. The cellular geography of Aurora kinases. Nat Rev Mol Cell Biol 2003;4:842–54.
2. Fu J, Bian M, Jiang Q, Zhang C. Roles of Aurora kinases in mitosis and
tumorigenesis. Mol Cancer Res 2007;5:1–10.
3. Carmena M, Ruchaud S, Earnshaw WC. Making the Auroras glow:
regulation of Aurora A and B kinase function by interacting proteins. Curr Opin Cell Biol 2009;21:796–805.
4. Meraldi P, Honda R, Nigg EA. Aurora kinases link chromosome
segregation and cell division to cancer susceptibility. Curr Opin Genet Dev 2004;14:29–36.
5. Karthigeyan D, Prasad SB, Shandilya J, Agrawal S, Kundu TK. Biology
of Aurora A kinase: implications in cancer manifestation and therapy. Med Res Rev 2011;31:757–93.
6. Becker M, Stolz A, Ertych N, Bastians H. Centromere localization of
INCENP-Aurora B is sufﬁcient to support spindle checkpoint function. Cell Cycle 2010;9:1360–72.
7. Ditchﬁeld C, Johnson VL, Tighe A, Ellston R, Haworth C, Johnson T, et al. Aurora B couples chromosome alignment with anaphase by
targeting BubR1, Mad2, and Cenp-E to kinetochores. J Cell Biol 2003;161:267–80.

8. Hauf S, Cole RW, LaTerra S, Zimmer C, Schnapp G, Walter R, et al. The small molecule Hesperadin reveals a role for Aurora B in correcting kinetochore-microtubule attachment and in maintaining the spindle assembly checkpoint. J Cell Biol 2003;161:281–94.
9. Liu D, Lampson MA. Regulation of kinetochore-microtubule attach-
ments by Aurora B kinase. Biochem Soc Trans 2009;37:976–80.
10. Girdler F, Gascoigne KE, Eyers PA, Hartmuth S, Crafter C, Foote KM,
et al. Validating Aurora B as an anti-cancer drug target. J Cell Sci 2006;119:3664–75.
11. Pollard JR, Mortimore M. Discovery and development of aurora kinase
inhibitors as anticancer agents. J Med Chem 2009;52:2629–51.
12. Lucena-Araujo AR, de Oliveira FM, Leite-Cueva SD, Dos Santos GA,
Falcao RP, Rego EM. High expression of AURKA and AURKB is associated with unfavorable cytogenetic abnormalities and high white blood cell count in patients with acute myeloid leukemia. Leuk Res 2011;35:260–4.
13. Pohl A, Azuma M, Zhang W, Yang D, Ning Y, Winder T, et al. Pharma-
cogenetic proﬁling of Aurora kinase B is associated with overall survival in metastatic colorectal cancer. Pharmacogenomics J 2011;11:93–9.

Farrell et al.

14. Cheung CH, Coumar MS, Hsieh HP, Chang JY. Aurora kinase inhibi- tors in preclinical and clinical testing. Expert Opin Investig Drugs 2009;18:379–98.
15. Katayama H, Brinkley WR, Sen S. The Aurora kinases: role in cell
transformation and tumorigenesis. Cancer Metastasis Rev 2003;22: 451–64.
16. Barton VN, Foreman NK, Donson AM, Birks DK, Handler MH, Vibhakar
R. Aurora kinase A as a rational target for therapy in glioblastoma. J Neurosurg Pediatr 2010;6:98–105.
17. Chen YJ, Chen CM, Twu NF, Yen MS, Lai CR, Wu HH, et al. Over-
expression of Aurora B is associated with poor prognosis in epithelial ovarian cancer patients. Virchows Arch 2009;455:431–40.
18. Chiefﬁ P, Cozzolino L, Kisslinger A, Libertini S, Staibano S, Mansueto
G, et al. Aurora B expression directly correlates with prostate cancer malignancy and inﬂuence prostate cell proliferation. Prostate 2006; 66:326–33.
19. Kurai M, Shiozawa T, Shih HC, Miyamoto T, Feng YZ, Kashima H, et al.
Expression of Aurora kinases A and B in normal, hyperplastic, and malignant human endometrium: Aurora B as a predictor for poor prognosis in endometrial carcinoma. Hum Pathol 2005;36:1281–8.
20. Sorrentino R, Libertini S, Pallante PL, Troncone G, Palombini L,
Bavetsias V, et al. Aurora B overexpression associates with the thyroid carcinoma undifferentiated phenotype and is required for thyroid carcinoma cell proliferation. J Clin Endocrinol Metab 2005;90:928–35.
21. Tatsuka M, Katayama H, Ota T, Tanaka T, Odashima S, Suzuki F, et al.
Multinuclearity and increased ploidy caused by overexpression of the aurora- and Ipl1-like midbody-associated protein mitotic kinase in human cancer cells. Cancer Res 1998;58:4811–6.
22. Zeng WF, Navaratne K, Prayson RA, Weil RJ. Aurora B expression
correlates with aggressive behaviour in glioblastoma multiforme. J Clin Pathol 2007;60:218–21.
23. Dees EC, Infante JR, Burris HA, Astsaturov IA, Stinchcombe T, Liu H,
et al. Phase I study of the investigational drug MLN8237, an Aurora A kinase (AAK) inhibitor, in patients (pts) with solid tumors. J Clin Oncol. 2010;28(Suppl.):15s: abstr 3010.
24. Macarulla T, Cervantes A, Elez E, Rodriguez-Braun E, Baselga J, Rosello S, et al. Phase I study of the selective Aurora A kinase inhibitor MLN8054 in patients with advanced solid tumors: safety, pharmaco- kinetics, and pharmacodynamics. Mol Cancer Ther 2010;9:2844–52.
25. Manfredi MG, Ecsedy JA, Meetze KA, Balani SK, Burenkova O, Chen
W, et al. Antitumor activity of MLN8054, an orally active small-molecule inhibitor of Aurora A kinase. Proc Natl Acad Sci U S A 2007;104: 4106–11.
26. Anderson K, Lai Z, McDonald OB, Stuart JD, Nartey EN, Hardwicke
MA, et al. Biochemical characterization of GSK1070916, a potent and selective inhibitor of Aurora B and Aurora C kinases with an extremely long residence time. Biochem J 2009;420:259–65.
27. Boss DS, Witteveen PO, van der Sar J, Lolkema MP, Voest EE,
Stockman PK, et al. Clinical evaluation of AZD1152, an i.v. inhibitor of Aurora B kinase, in patients with solid malignant tumors. Ann Oncol 2011;22:431–7.
28. Hardwicke MA, Oleykowski CA, Plant R, Wang J, Liao Q, Moss K, et al.
GSK1070916, a potent Aurora B/C kinase inhibitor with broad antitu- mor activity in tissue culture cells and human tumor xenograft models. Mol Cancer Ther 2009;8:1808–17.
29. Wilkinson RW, Odedra R, Heaton SP, Wedge SR, Keen NJ, Crafter C,
et al. AZD1152, a selective inhibitor of Aurora B kinase, inhibits human

tumor xenograft growth by inducing apoptosis. Clin Cancer Res 2007;13:3682–8.
30. Arbitrario JP, Belmont BJ, Evanchik MJ, Flanagan WM, Fucini RV,
Hansen SK, et al. SNS-314, a pan-Aurora kinase inhibitor, shows potent anti-tumor activity and dosing ﬂexibility in vivo. Cancer Che- mother Pharmacol 2010;65:707–17.
31. Carpinelli P, Ceruti R, Giorgini ML, Cappella P, Gianellini L, Croci V,
et al. PHA-739358, a potent inhibitor of Aurora kinases with a selective target inhibition proﬁle relevant to cancer. Mol Cancer Ther 2007; 6:3158–68.
32. Curry J, Angove H, Fazal L, Lyons J, Reule M, Thompson N, et al.
Aurora B kinase inhibition in mitosis: strategies for optimising the use of aurora kinase inhibitors such as AT9283. Cell Cycle 2009;8:1921–9.
33. Harrington EA, Bebbington D, Moore J, Rasmussen RK, Ajose-Adeo-
gun AO, Nakayama T, et al. VX-680, a potent and selective small- molecule inhibitor of the Aurora kinases, suppresses tumor growth in vivo. Nat Med 2004;10:262–7.
34. Jani JP, Arcari J, Bernardo V, Bhattacharya SK, Briere D, Cohen BD,
et al. PF-03814735, an orally bioavailable small molecule aurora kinase inhibitor for cancer therapy. Mol Cancer Ther 2010;9:883–94.
35. McLaughlin J, Markovtsov V, Li H, Wong S, Gelman M, Zhu Y, et al.
Preclinical characterization of Aurora kinase inhibitor R763/AS703569 identiﬁed through an image-based phenotypic screen. J Cancer Res Clin Oncol 2010;136:99–113.
36. Morrison JF, Walsh CT. The behavior and signiﬁcance of slow-binding
enzyme inhibitors. Adv Enzymol Relat Areas Mol Biol 1988;61: 201–301.
37. Anderson K, Yang J, Koretke K, Nurse K, Calamari A, Kirkpatrick RB,
et al. Binding of TPX2 to Aurora A alters substrate and inhibitor interactions. Biochemistry 2007;46:10287–95.
38. Bayliss R, Sardon T, Vernos I, Conti E. Structural basis of Aurora-A
activation by TPX2 at the mitotic spindle. Mol Cell 2003;12:851–62.
39. Honda R, Korner R, Nigg EA. Exploring the functional interactions
between Aurora B, INCENP, and survivin in mitosis. Mol Biol Cell 2003;14:3325–41.
40. Sessa F, Mapelli M, Ciferri C, Tarricone C, Areces LB, Schneider TR,
et al. Mechanism of Aurora B activation by INCENP and inhibition by hesperadin. Mol Cell 2005;18:379–91.
41. Muraki M, Ohkawara B, Hosoya T, Onogi H, Koizumi J, Koizumi T, et al.
Manipulation of alternative splicing by a newly developed inhibitor of Clks. J Biol Chem 2004;279:24246–54.
42. Reilly JT. FLT3 and its role in the pathogenesis of acute myeloid
leukaemia. Leuk Lymphoma 2003;44:1–7.
43. Moore AS, Blagg J, Linardopoulos S, Pearson AD. Aurora kinase
inhibitors: novel small molecules with promising activity in acute myeloid and Philadelphia-positive leukemias. Leukemia 2010;24: 671–8.
44. VanderPorten EC, Taverna P, Hogan JN, Ballinger MD, Flanagan WM,
Fucini RV. The Aurora kinase inhibitor SNS-314 shows broad thera- peutic potential with chemotherapeutics and synergy with microtu- bule-targeted agents in a colon carcinoma model. Mol Cancer Ther 2009;8:930–9.
45. Yang J, Ikezoe T, Nishioka C, Tasaka T, Taniguchi A, Kuwayama Y,
et al. AZD1152, a novel and TAK-901 selective aurora B kinase inhibitor, induces growth arrest, apoptosis, and sensitization for tubulin depolymerizing agent or topoisomerase II inhibitor in human acute leukemia cells in vitro and in vivo. Blood 2007;110:2034–40.