Functional characterization of CFI-402257, a potent and selective Mps1/TTK … · Functional characterization of CFI-402257, a potent and selective Mps1/TTK kinase inhibitor, for

Functional characterization of CFI-402257, a potentand selective Mps1/TTK kinase inhibitor, for thetreatment of cancerJacqueline M. Masona, Xin Weia, Graham C. Fletchera, Reza Kiarasha, Richard Brokxa, Richard Hodgsona,Irina Beletskayaa, Mark R. Braya, and Tak W. Makb,1

aThe Campbell Family Institute Therapeutics Group at Princess Margaret Cancer Centre, University Health Network, Toronto, ON, Canada M5G 1L7; and bTheCampbell Family Institute for Breast Cancer Research at Princess Margaret Cancer Centre, University Health Network, Toronto, ON, Canada M5G 2M9

Contributed by Tak W. Mak, February 13, 2017 (sent for review January 6, 2017; reviewed by Sabine Elowe and Kari Wisinski)

Loss of cell-cycle control is a hallmark of human cancer. Cell-cyclecheckpoints are essential for maintaining genome integrity and bal-anced growth and division. They are specifically deregulated in cancercells and contain regulators that represent potential therapeutic targets.Monopolar spindle 1 (Mps1; also known as TTK protein kinase) is a corecomponent of the spindle assembly checkpoint (SAC), a genome-sur-veillance mechanism that is important for cell survival, and hasemerged as a candidate target for anticancer therapy. Here, we reportthe cellular and antitumor effects of CFI-402257, a potent (Mps1 Ki =0.09 ± 0.02 nM; cellular Mps1 EC50 = 6.5 ± 0.5 nM), highly selective,and orally active small-molecule inhibitor of Mps1 that was identifiedthrough a drug-discovery program. Human cancer cells treated withCFI-402257 exhibit effects consistent with Mps1 kinase inhibition, spe-cifically SAC inactivation, leading to chromosome missegregation, an-euploidy, and ultimately cell death. Oral administration of CFI-402257inmonotherapy or in combinationwith an anti-programmed cell death1 (PD-1) antibody in mouse models of human cancer results in inhibi-tion of tumor growth at doses that are well-tolerated. Our findingsprovide a rationale for the clinical evaluation of CFI-402257 in patientswith solid tumors.

Mps1/TTK | inhibitor | CFI-402257 | cancer

Mammalian cell division is tightly regulated by the activationand inactivation of proteins that regulate progression through

the phases of the cell cycle. To ensure that only healthy cells pro-liferate, cell-cycle checkpoints have evolved that play key roles ingenome maintenance under various stress conditions and evenduring an unperturbed cell cycle. These checkpoints are specificallyderegulated in cancer cells, and contain regulators that representpotential targets for the design of cancer cell-selective therapies (1).The spindle assembly checkpoint (SAC; also known as the mi-

totic checkpoint) is a signaling cascade that prevents chromosomemissegregation by arresting the cell cycle in mitosis until all chro-mosomes are properly attached to the mitotic spindle (2). This ar-rest is achieved by inhibiting the anaphase-promoting complex/cyclosome (APC/C), an E3 ubiquitin ligase that is essential formitotic progression. Normal cells have a robust SAC in which oneor more unattached chromosomes can produce a signal strongenough to inhibit all cellular APC/C activity and thereby blockprogression to anaphase. By contrast, many cancer cells have aweakened, but not absent, SAC response that may cause aneuploidyand drive malignancy. However, more severe disabling of SACcheckpoint signaling generates a level of chromosome instabilitythat exceeds the adaption capacity of cancer cells and is a potentialanticancer strategy. Of note, nontransformed cells appear to be lesssensitive than cancer cells to death caused by SAC inhibition (3, 4).The core SAC kinase, monopolar spindle 1 (Mps1; also known asTTK protein kinase), is a candidate target for this approach.Mps1 is an essential, dual-specificity kinase (5) that functions as an

important guardian of the fidelity of chromosome segregation. It iscritical for the recruitment of SAC proteins to unattached kineto-chores, mitotic checkpoint complex formation, and thus APC/C in-hibition. Mps1 is also required for chromosome alignment and error

correction (6–8). Inhibition of Mps1 activity causes cells toprematurely exit mitosis with unattached chromosomes, resultingin severe chromosome missegregation, aneuploidy, and ulti-mately cell death (4, 9–12). Mps1 is expressed during mitosis inproliferating cells. Overexpression of Mps1 is observed in severalhuman tumors, and correlates with worse prognosis (13–19).Further, Mps1 has been implicated in facilitating an aneuploidy-tolerant state in cancer cells (13). Together, these findingspointed to Mps1 as a rational drug target for anticancer therapy.We describe here the cellular and antitumor effects of CFI-

402257, a potent and selective Mps1 kinase inhibitor. CFI-402257shows strong antineoplastic activity on a broad panel of humancancer-derived cell lines, and causes effects consistent with de-pletion or inhibition of Mps1 (4, 9–12). Oral administration of CFI-402257 as a single agent to mice bearing human cancer xenograftsresults in inhibition of tumor growth at doses that are well-tolerated.Notably, the combination of CFI-402257 and an anti-programmedcell death 1 (PD-1) antibody causes tumor regression and increasessurvival in a murine colon cancer model that mimics the humandisease, suggesting that this might be a clinically viable approach.

ResultsCFI-402257 Is a Selective Inhibitor of Mps1.Mps1 was identified as apotential therapeutic target through a systematic approach thatcombined RNAi screening with gene-expression analysis in hu-man breast cancers and cell lines, which previously yielded polo-like kinase 4 as an anticancer drug target (20). Reduction of

Significance

At present, microtubule-targeting agents are the most importantantimitotic drugs used in the clinic. However, there is an urgentneed for the discovery of new approaches to more effectively tar-get tumor cells with less toxicity. Emerging strategies for anticancertherapy include exploiting cell-cycle checkpoint vulnerabilities andgenomic instability in cancer cells. The spindle assembly checkpoint(SAC) is important for cell survival, and its inactivation generateslethal genomic instability in cancer cells. Inhibition of SAC signalingthrough targeting of monopolar spindle 1 (Mps1) has provided anindication of the feasibility of such an approach. We report here thecellular and antitumor effects of CFI-402257, a potent and specificsmall-molecule inhibitor ofMps1. CFI-402257 is currently in a phase Iclinical trial (ClinicalTrials.gov ID: NCT02792465).

Author contributions: J.M.M., X.W., G.C.F., R.B., R.H., M.R.B., and T.W.M. designed research;J.M.M., X.W., G.C.F., R.K., R.B., R.H., and I.B. performed research; J.M.M., X.W., G.C.F., R.K., R.B.,R.H., I.B., M.R.B., and T.W.M. analyzed data; and J.M.M., M.R.B., and T.W.M. wrote the paper.

Reviewers: S.E., Université Laval; and K.W., University of Wisconsin School of Medicineand Public Health.

The authors declare no conflict of interest.

Freely available online through the PNAS open access option.1To whom correspondence should be addressed. Email: [email protected].

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1700234114/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1700234114 PNAS Early Edition | 1 of 6

CELL

BIOLO

GY

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

7, 2

020

Mps1 resulted in a loss of cell viability in 8 of 14 cell lines (Fig. S1),andMps1mRNA was found to be overexpressed in 38% of all breasttumors and in 85% of triple-negative/basal-like (TNBC) tumors (Fig.S2A) (21), the subtype considered to be the most genomically un-stable class of breast cancer (22). Of note, reduction of Mps1inhibited the cell viability of all four basal breast cancer cell lines andboth immortalized basal breast cell lines examined. Immunoblotanalysis was used to compare Mps1 levels in breast cancer cell lines tothose in primary human mammary epithelial cells derived fromnormal breast tissue samples. Mps1 levels were found to be signifi-cantly higher in breast cancer cells relative to their normal tissuecounterparts and in basal breast cancer cell lines relative to luminalbreast cancer cell lines (Fig. S2B).On the basis of these studies, we initiated a drug discovery

program that culminated in the discovery of CFI-402257 as apotent and selective small-molecule inhibitor of Mps1 (Fig. 1A)(23). In an assay using recombinant human Mps1, CFI-402257inhibited Mps1 with an IC50 value of 1.2 ± 0.4 nM, and was ATPcompetitive with a Ki value of 0.09 ± 0.02 nM. Consistent withthese results, CFI-402257 inhibited autophosphorylation of Mps1at threonine 12/serine 15 with an EC50 value of 6.5 ± 0.5 nM incells exogenously expressing human Mps1 (Fig. 1B). Mps1 thre-onine 12/serine 15 autophosphorylation occurs as a consequenceof kinase activation, and thus enables identification of active Mps1in cells (24). Immunoblotting of endogenous phosphorylatedMps1 in cell lysates was technically challenging; therefore, cellsexogenously expressing Mps1 were used to determine the EC50value for inhibition of Mps1 autophosphorylation by CFI-402257.To evaluate selectivity, CFI-402257 was profiled against a panel

of 265 recombinant protein and lipid kinases and a panel of 98nonkinase targets, including G protein-coupled receptors, nuclearreceptors, membrane channels, and enzymes. Of the 265 kinasestested, none showed >50% inhibition when incubated with 1 μMCFI-402257, and no significant inhibitory activity was observedagainst Aurora kinase, Polo-like kinase, or Cyclin-dependent kinasefamily members. Only 3 of 98 nonkinase targets were inhibited≥50% with 10 μM CFI-402257 (Table S1). Overall, these resultsindicate that CFI-402257 selectively and potently inhibits Mps1 ki-nase activity in cells with minimal off-target effects.

Cellular Effects of CFI-402257. Mps1 contributes to the alignment ofchromosomes on the metaphase plate, and is a core component ofthe SAC, a surveillance mechanism that ensures mitotic fidelity andgenome stability. Depletion or inhibition of Mps1 causes cells to exitmitosis with unattached chromosomes, resulting in severe chromo-some missegregation, aneuploidy, and cell death (4, 9–12). CFI-402257 inhibited the activation of the SAC with an IC50 value of 64 ±5 nM as monitored in an assay in which the disappearance of histoneH3 serine 10 phosphorylation was assessed in HCT116 human coloncarcinoma cells treated with the spindle poison nocodazole (Fig. 1C).This phosphorylation event in histone H3 is a posttranslational modi-fication that identifies mitotic cells (prophase to anaphase), and is not aresult of Mps1 activity. To examine the effect of CFI-402257 on thefidelity of chromosome segregation, HCT116 cells were treated withthe inhibitor, and the number of chromosome segregation errorspresent in mitotic cells was determined by immunostaining for cen-tromeres, α-tubulin to visualize the mitotic spindle, and DAPI. Treat-ment with 200 nM CFI-402257 caused a massive increase inchromosomemissegregations relative to DMSO control cells (Fig. 2A).SAC inactivation and gross chromosome segregation errors caused

by CFI-402257 had a major impact on the cell-cycle progression andsurvival of human cancer cells. CFI-402257 induced a dose-dependentdysregulation of the cell cycle, resulting in an increase in the frequencyof cells exhibiting an aneuploid DNA content (Fig. 2B). Massive an-euploidy was observed with 100 nM CFI-402257, and significant poly-ploidy was not observed with CFI-402257 concentrations as high as3,000 nM as a result of the potency and selectivity of the compound.The ability of CFI-402257 to cause aneuploidy in cancer cells wasconfirmed by metaphase chromosome spread assay (Fig. 2C). CFI-402257–induced aneuploidy was accompanied by a progressive accu-mulation of apoptotic cells that were detectable as early as 16 hfollowing treatment (Fig. 2D). The effects described here forCFI-402257 are in accordance with cellular depletion of Mps1protein or inhibition of Mps1 with other compounds (4, 9–12).Thus, CFI-402257 inhibits Mps1 in human cancer cells andmodulates the physiological role of the kinase.We evaluated the effect of CFI-402257 on growth inhibition (i.e.,

IC50) in a panel of human cancer-derived cell lines covering variouscancer histologies after 5 d. CFI-402257 exerted potent growth in-hibitory activity across the vast majority of cell lines with a medianIC50 value of 15 nM (range, 1 to >1,500 nM; Fig. 3 and Table S2).Notably, no single gene mutation in several proteins relevant tooncogenesis (e.g., APC, BRAF, CDKN2A, PIK3CA, RAS andTP53; Sanger Institute Catalog Of Somatic Mutations In Cancerdatabase, cancer.sanger.ac.uk/cosmic/) or cell doubling time wascorrelated with statistical significance with the resistance/sensitivityof the cell lines to CFI-402257.

Toxicity Studies of CFI-402257. The tolerability and pharmacoki-netic properties of CFI-402257 were initially evaluated in mice.To determine a tolerable dose range for in vivo efficacy studies,mice were given CFI-402257 orally for 21 d, with doses rangingfrom 4.5 to 6.5 mg/kg. The upper dose level [i.e., maximumtolerated dose (MTD)] for once-daily (QD) administration ofCFI-402257 was between 6 and 6.5 mg/kg, defined as the highest dosethat did not cause >10% loss in body weight or induce overt toxicity(i.e., signs of ill health or abnormal behaviors, including hunchedposture, poor grooming, urine stains, diarrhea, or dehydration) inaccordance with institutional humane endpoint guidelines.Comprehensive toxicologic evaluation of CFI-402257 adminis-

tered to rats and dogs has been completed. In summary, dose-lim-iting toxicities were found to be reversible gastrointestinal effects andmyelosuppression, which are consistent with the proposed mecha-nism of action of the molecule (Table S3). The toxicological profileof CFI-402257 supports its clinical development to determine itspotential as an effective treatment in oncology.

In Vivo Characterization of the Anticancer Potential of CFI-402257. Phar-macokinetic parameters were determined by treating mice with singleoral doses of CFI-402257 (Fig. S3). Plasma was obtained at specifiedtime points, and concentrations of CFI-402257 were determined by

C

Pho

spho

-his

tone

H3

(S10

)/hi

ston

e H

3

EC50

-10 -9 -8 -7 -6log [CFI-402257]

1.5

1.0

0.5

0.0

B

-10 -9 -8 -7 -6

1.0

0.8

0.6

0.4

0.2

0.0

log [CFI-402257]Pho

spho

-Mps

1 (T

12/S

15)/M

ps1

EC50

nM CFI-402257DM

SO

1 10 25 50 100

200

400

Mps1

moc

k

phospho-Mps1(T12/S15)

Mps1

A

Fig. 1. CFI-402257 is a selective inhibitor of Mps1. (A) The chemical structure ofCFI-402257. (B) HCT116 cells transfectedwith full-length humanMps1were treatedwith nocodazole for 17 h. Cells were pretreated with MG132 for 30 min beforetreatmentwith CFI-402257 or DMSO for 4 h. Lysates were analyzed by immunoblotanalysis with antibodies against phospho-Mps1 (T12/S15) and Mps1. (C) HCT116cells were treated with nocodazole for 17 h before treatment with CFI-402257 orDMSO for 4 h. Lysates were analyzed by immunoblot analysis with antibodiesagainst phospho-histone H3 (S10) and histoneH3. Band intensities were quantified,and EC50 values were calculated by using GraphPad Prism software.

2 of 6 | www.pnas.org/cgi/doi/10.1073/pnas.1700234114 Mason et al.

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

7, 2

020

liquid chromatography (LC)-tandem MS (LC/MS/MS). CFI-402257was absorbed rapidly after oral administration, and exposure [interms of area under the curve (AUC) and maximum plasmaconcentrations (Cmax)] was dose-dependent and approximatelydose-linear to 55 mg/kg. At the repeat-dose MTD of 6.5 mg/kg,the Cmax was 1,070 ng/mL and the AUC0–24h was 3,170 ng·h/mL.The elimination t1/2 ranged between 2 and 3 h. These data dem-onstrate that CFI-402257 has a favorable pharmacokinetic profile.CFI-402257 has demonstrated efficacy as a monotherapy in

cancer cell line and patient-derived xenograft (PDX) models acrossvarious indications, and representative studies are described here.CFI-402257 given orally QD showed dose-dependent activity inmice with established tumors from xenografted MDA-MB-231human TNBC cells [5 mg/kg CFI-402257, tumor growth inhibition(TGI) = 74%, P = 0.02; 6 mg/kg CFI-402257, TGI = 89%, P =0.004; Fig. 4A]. Similar results were obtained in tumors from xen-ografted MDA-MB-468 human TNBC cells in mice (5 mg/kg CFI-402257, TGI = 75%, P = 0.003; 6 mg/kg CFI-402257, TGI = 94%,P = 0.001; Fig. 4B). CFI-402257 also demonstrated antitumor ac-tivity in a platinum-resistant PDX model of high-grade serousovarian cancer (6.5 mg/kg CFI-402257, TGI = 61%, P = 0.04;75 mg/kg carboplatin, TGI = 97%, P = 0.03; Fig. 4C). Further, CFI-402257 was found to be similarly efficacious in a platinum-sensitive

PDX model of high-grade serous ovarian cancer (Fig. 4D). CFI-402257 was well-tolerated at the doses examined as measured bylittle to no decrease in body weight (<10% change) and normalbehavior. Together, these results indicate that CFI-402257 can in-hibit the growth of a range of tumor types and may be effective ina clinical setting.To determine the pharmacodynamics of CFI-402257 in vivo,

phospho-histone H3 serine 10-positive cells were counted in theMDA-MB-231 breast tumor xenografts treated with the daily doseMTD of 6 mg/kg for 3 d or a large acute dose of 35 mg/kg twice daily(BID) for five doses (Fig. 5). Relative to vehicle controls, a decreasein phospho-histone H3 serine 10-positive cells per square millimeterof tumor tissue was measured in CFI-402257–treated tumors (40phospho-histone H3-positive cells per square millimeter with 6 mg/kgCFI-402257 QD × 3 treatment, and 29 phospho-histone H3-positivecells per square millimeter with 35 mg/kg CFI-402257 BID × 5treatment vs. 70 phospho-histone H3-positive cells per squaremillimeter with vehicle control treatment). Thus, CFI-402257reduces the mitotic index in vivo, consistent with inhibition ofMps1 in vivo.CFI-402257 induces genomic instability and apoptotic cell death,

and therefore could promote tumor immunity. To explore the po-tential to combine Mps1 inhibitors with immune checkpoint

A B

C

D

Fig. 2. Mechanism of action studies for CFI-402257. (A) CFI-402257 causes aberrant chromosomal segregation in cancer cells. HCT116 cells were synchronized by usinga double thymidine block, released after 6 h, and treated with 200 nM CFI-402257 or DMSO for 2 h. (Left) Representative mitotic figures of CFI-402257– or DMSO-treated cells. Staining of centromeres is in green, α-tubulin in red, and DAPI in blue. (Scale bar: 5 μm.) (Right) Graph of the mean percentage of anaphase andtelophase cells with and without lagging chromosomes from three independent experiments, with a total of 100 cells examined per treatment condition. Meanpercentage of lagging chromosomes ± SEM: DMSO, 11 ± 3%, CFI-402257, 79 ± 6% (P = 0.0002, Student’s t test). (B) CFI-402257 induces aneuploidy with no significantpolyploidy in cancer cells. HCT116 cells were treated with CFI-402257 or DMSO for 2 d before staining with propidium iodide and analysis of DNA content by flowcytometry. Percentages of 8N and 16N cells are normalized to the DMSO control. Aneuploidy is defined as having an abnormal numbers of chromosomes. Polyploidy isdefined as having an 8N DNA population greater than or equal to 15% after normalization to the DMSO control. Data are representative of three independentexperiments. (C) CFI-402257 causes aneuploidy in cancer cells. (Top) Graph of the chromosome counts on metaphase spreads from HCT116 cells treated with 200 nMCFI-402257 or DMSO for 28 h. Data are representative of three independent experiments, with a total of 98 spreads examined per treatment condition. (Bottom)Representative images of metaphase spreads from CFI-402257– or DMSO-treated cells. Chromosomes were stained with DAPI. (D) HCT116 cells were treated for theindicated amounts of time with 0, 50, or 100 nM CFI-402257. Lysates were analyzed by immunoblot analysis with antibodies against cleaved PARP and GAPDH.

Mason et al. PNAS Early Edition | 3 of 6

CELL

BIOLO

GY

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

7, 2

020

inhibitors, immunocompetent BALB/cJ mice were inoculatedwith syngeneic CT26 mouse colon carcinoma cells and then treatedwith CFI-402257 alone and in combination with an anti–PD-1 an-tibody (Fig. 6). Tumors in the vehicle-treated control arm grewrapidly, and the average size was >1,500 mm3 by day 11 of treat-ment. Although there was tumor growth delay in the anti–PD-1antibody- and the CFI-402257–treated single-agent arms, therewere no instances in which complete regression was observed. In thecombination anti–PD-1 antibody and CFI-402257–treated arm,however, two of the eight tumors completely regressed. Very similarresults were also seen in a duplicate experiment (Fig. S4), again withcomplete regression (two of eight tumors) only seen in the combi-nation arm. In the former experiment, the two animals inwhich complete regression had occurred were rechallenged byinoculation with CT26 cells on day 31. Tumors did not grow ineither mouse, indicating that immunity to the CT26 cells hadbeen generated.

DiscussionThe primary function of the SAC is to prevent chromosome mis-segregation and the generation of aneuploid or polyploid cells. Thus,the SAC ensures healthy proliferation and precise division in cells, andtargeting this checkpoint could provide new therapeutic opportunitiesfor killing cancer cells. Differently from antimitotic/antitubulin ther-apy, which arrests cell proliferation, inactivation of the SAC wouldlead to an accelerated mitosis and the consequent generation of lethalgenomic instability in cancer cells. The strong requirement for Mps1for the SAC has made it a candidate for this therapeutic approach,and has generated intense interest in developing small-molecule in-hibitors against this kinase (4, 9–12, 25, 26).We report here the functional characterization of CFI-402257,

a potent and orally active inhibitor of Mps1 that has entered aphase I clinical trial (ClinicalTrials.gov ID: NCT02792465). CFI-402257 is highly selective toward Mps1 relative to other proteinand lipid kinases, and did not inhibit several kinases that areknown for their role in mitosis. CFI-402257 pharmacologically re-capitulates previous findings obtained upon Mps1 deletion or withother Mps1-specific inhibitors (4, 9–12). Specifically, CFI-402257treatment potently inhibited Mps1 kinase activity in cancer cells, andcauses cancer cells to exit mitosis with missegregated chromosomes,which results in severe aneuploidy and finally cell death. Cancer cellline profiling demonstrated that CFI-402257 has broad and strongantiproliferative activity. Although cell lines with high and low sen-sitivity toward CFI-402257 treatment were identified, no single geneimplicated via mutation in human cancer could be clearly correlated.This suggests that multiple genes and/or kinetochore-specific factorsmay determine the cellular sensitivity to Mps1 inhibitors. Furtherinvestigation is ongoing to identify candidate predictive biomarkersfor CFI-402257 and to understand which particular class of malig-nancies might benefit from Mps1-targeted agents.

Mouse plasma concentration–time profiles indicate that, follow-ing oral administration of efficacious doses of CFI-402257 in mice,plasma levels of CFI-402257 were sustained and remained abovethe EC50 value for half-maximal inhibition of cellular Mps1 auto-phosphorylation and the median growth-inhibition IC50 value forgreater than 12 h. Moreover, CFI-402257 monotherapy demon-strated dose-dependent antitumor activity. Analysis of xenografttumors from mice treated with an efficacious dose of CFI-402257showed pharmacodynamic effects consistent with inhibition ofMps1. Together, these results provide evidence that the antitumoractivity of CFI-402257 results from its in vivo inhibition of Mps1.Mps1 inhibitors BAY 1161909 (ClinicalTrials.gov ID:

NCT02138812), BAY 1217389 (ClinicalTrials.gov ID:NCT02366949), and S 81694 (EudraCT no.: 2014–002023-10) arecurrently in phase I clinical trials. As has been presented for CFI-402257, these inhibitors have cellular on-target activity and havesignificant tumor growth-inhibitory activity in preclinical models (27,28). However, comparison of the reported pharmacological prop-erties and preclinical data for these inhibitors reveals differences.CFI-402257 shows higher kinase selectivity compared with BAY1161909, BAY 1217389, or S 81694. Although additional activitiescould provide enhanced efficacy in particular settings, they may alsoresult in reduced tolerance. Specifically, the toxicities observed forthis class of drug targets (i.e., myelosuppression and gastrointestinaldisturbances) suggest that the balance between efficacy and toler-ability may not be one that allows additional activities to contributeto an antitumor effect. These inhibitors have been evaluated byusing different dosing schedules and routes of delivery. Daily oral

Gro

wth

inhi

bitio

n IC

50 ( n

M)

10000

1000

100

10

1

SK

Br-

3 B

xPC

-3

CA

L51

T-47

D

MD

A-M

B-4

36

A54

9 D

U-1

54

MD

A-M

B-3

61

HC

C20

2 S

W48

BT

-474

P

C-3

H

CT1

16

HT2

9 M

DA

-MB

-468

M

DA

-MB

-231

H

CC

1937

M

DA

-MB

-330

O

VC

AR

-3

NC

I-H23

P

AN

C-1

H

CC

1419

S

K-O

V-3

Fig. 3. CFI-402257 inhibits human cancer cell growth. After 5 d of treatment, cellgrowth was determined by measuring total protein content by sulforhodamineB (SRB) assay, and growth-inhibition IC50 values were calculated by using GraphPadPrism software.

Day

Tum

or v

olum

e (m

m3 )

VehicleCFI-402257, 5 mg/kg PO, QDCFI-402257, 6 mg/kg PO, QD

0

100

200

300

400

500

0 2 4 6 8 10 12 14 16 18 20 22Day

Tum

or v

olum

e (m

m3 )

VehicleCFI-402257, 5 mg/kg PO, QDCFI-402257, 6 mg/kg PO, QD

0

150

200

250

300

350

0 2 4 6 8 10 12 14 16 18 20

100

50

Tum

or v

olum

e (m

m3 )

VehicleCFI-402257, 6.5 mg/kg PO, QDCarboplatin, 75 mg/kg IP, QWX2

0

250

500

750

1000

Day0 2 4 6 8 10 12 14 16 18 20

Tum

or v

olum

e (m

m3 )

VehicleCFI-402257, 6.5 mg/kg PO, QDCarboplatin, 75 mg/kg IP, QWX2

0

250

500

750

Day0 2 4 6 8 10 12 14 16 18 20

A B

DC

Fig. 4. In vivo characterization of the anticancer potential of CFI-402257 onhuman xenograft tumors. (A) CD-1 Nude mice with established MDA-MB-231xenografts were treated for 22 d (n = 7). CFI-402257 5 mg/kg orally QD vs. vehicle,TGI= 74%, P = 0.02; CFI-402257 6mg/kg orally (PO) QD vs. vehicle, TGI= 89%, P =0.004. (B) CD-1 nude mice with establishedMDA-MB-468 xenografts were treatedfor 21 d (n = 7). CFI-402257 5 mg/kg orally QD vs. vehicle, TGI = 75%, P = 0.003;CFI-402257 6 mg/kg orally QD vs. vehicle, TGI = 94%, P = 0.001. (C) NOD scidgamma (NSG) mice with established human platinum-resistant ovarian tumorxenografts were treated for 21 d (n = 6). CFI-402257 6.5 mg/kg orally QD vs. ve-hicle, TGI= 61%, P= 0.04; carboplatin 75mg/kg i.p. weekly ×2 (QWX2) vs. vehicle,TGI = 97%, P = 0.03. (D) NSG mice with established human platinum-sensi-tive ovarian tumor xenografts were treated for 21 d (n = 6). CFI-402257 6.5mg/kg orally QD vs. vehicle, TGI = 66%, P = 0.11; carboplatin 75 mg/kg i.p.QWX2 vs. vehicle, TGI = 124%, P = 0.02. Data are represented as mean ± SEM(for figure clarity, only positive error bars are shown). P values were calculated byusing Student’s t test.

4 of 6 | www.pnas.org/cgi/doi/10.1073/pnas.1700234114 Mason et al.

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

7, 2

020

dosing has been used for CFI-402257, intermittent oral dosing forBAY 1161909 and BAY 1217389, and intermittent i.v. dosing for S81694. Notably, the preclinical studies for these inhibitors haveshown that both daily and intermittent dosing schedules are effi-cacious and generally well-tolerated, giving confidence that anefficacious and tolerated schedule can be found in patients.Preclinical efficacy results for BAY 1161909 and BAY 1217389showed moderate activity as monotherapies, but strong coopera-tivity with taxanes, and this combination has advanced to the clinic(discussed later). CFI-402257 and S 81694 showed single-agentactivity in preclinical studies, and have entered clinical testing inmonotherapy. The relative importance of these factors and othersubtle differences between these drug candidates, and others thatmay emerge, as well as the genetic makeup of the tumor, will beestablished by ongoing clinical studies.Mechanism-based combination strategies are being explored

for Mps1 inhibitors in an effort to expand their therapeuticutility and delay the onset of resistance. Mps1 inhibitors aremitosis-promoting agents that, when used with antitubulin agents,which perturb proper chromosome alignment, are expected toincrease chromosome segregation errors above the threshold re-quired to kill cancer cells. In support of this hypothesis, Mps1inhibitors have been shown to sensitize cancer cells to taxanetreatment (29), and this approach is being tested clinically (27).Furthermore, the combination of Mps1 and cyclin-dependentkinase 4/6 inhibition has been investigated as a strategy to in-crease the therapeutic window for Mps1 inhibitors (30). Cdk4/6inhibitors trigger G1 cell cycle arrest in Rb1-competent cells, andthus might improve the tolerability and efficacy of Mps1 inhibi-tors. Early results support this hypothesis, and suggest that thecombination may prove particularly effective in Rb1-dysregu-lated cancers. The rationale for another combination approachfor Mps1 inhibitors may lie in the proteotoxic or metabolic stresscaused by the chromosome imbalance in aneuploid cancer cells(31). Here, treatment of cancer cells with Mps1 inhibitors thatincrease aneuploidy could synergize with ones that target theproteotoxic or metabolic stress induced by aneuploidy. Whetheraneuploidy-induced stress pathway inhibitors might synergizewith Mps1 inhibitors remains to be determined. Chemotherapy,targeted therapies, and radiotherapy can promote tumor im-munity by inducing immunogenic cell death (32, 33). Combi-nations of these cancer therapies with immune-checkpointinhibitors are being explored to achieve additive or synergistic

clinical activity. Here, we show that the combination of CFI-402257 and an anti–PD-1 antibody causes tumor regression andincreases the survival in a murine colon cancer model. Thereason for this response is likely complex and under investigation.One possible explanation may lie in the genomic instability andapoptosis in tumor cells caused by CFI-402257–induced Mps1inhibition. These events might lead to the creation of new immu-nogenic epitopes, which are able to induce cytotoxic T lymphocyteactivity against tumor cells. Further studies of immune-checkpointinhibitors are merited to support their use in combination treat-ment with Mps1 inhibitors.In summary, CFI-402257 is a potential small-molecule in-

hibitor for the treatment of cancer with a mechanism of actionthat is consistent with potent and selective inhibition of Mps1.CFI-402257 has been approved for human use by Health Canada,and a phase I clinical trial has been initiated.

MethodsCompounds and Plasmids. CFI-402257 was synthesized as described previously(23). The hydrochloride or bisphosphate hemihydrate salt forms of CFI-402257 were used in all studies. Thymidine, nocodazole, and MG132 wereobtained from Sigma. Rat IgG2a anti–PD-1 antibody (clone RMP1-14; cat. no.BE0146) and rat IgG2a isotype control (clone 2A3; cat. no. BE0089) wereobtained from Bio X Cell. Carboplatin was obtained from Teva Canada.The expression construct encoding full-length human Mps1 was obtainedfrom Origene.

Animals. C.B.-17 SCID and NSG mice were obtained from the animal colony atthe Ontario Cancer Institute of the University Health Network (Toronto). CD-1 Nude mice were obtained from Charles River Laboratories. BALB/cJ micewere obtained from The Jackson Laboratory. Six- to eight-week-old femaleanimals were used for the studies and were allowed unrestricted access tofood and water. All animal procedures were approved by the institutionalanimal care and use committee of the University Health Network (Toronto).

Xenograft Tumor Growth. The PDXs were generated from platinum-resistant orplatinum-sensitive high-grade serous ovarian cancers that were obtained from theUniversity Health Network Tissue Bank with patient consent and research ethicsboard approval (protocol 06–0903T) (34). Xenografts were established by injectionof 1 × 106 H2K-depleted tumor cells in 1:1 HBSS:growth factor-reduced Matrigel(BD Biosciences) into mammary fat pad of female NSG mice. Animals were ran-domized, and treatments were initiated when tumor volumes reached100−170 mm3. Cancer cell line xenografts were established by injectings.c. 1 × 107 tumor cells into the flank of female mice. Animals were randomized,

Pho

spho

-his

tone

H3

(S10

)po

sitiv

e nu

clei

per

mm

2

0

10

20

30

40

50

60

70

80p=0.009

p=0.002

Vehicle 6 mg/kgCFI-402257

QDx3

35 mg/kgCFI-402257

BIDx5

Fig. 5. In vivo effect of CFI-402257 on human xenograft tumors. C.B.-17 severecombined immunodeficiency (SCID) mice with established MDA-MB-231 xe-nografts were treated with CFI-402257 6 mg/kg orally QD or vehicle for 3 d (n =3) or CFI-402257 35mg/kg orally BID for 5 d (n = 3); mice were killed and tumortissue was removed 4 h after the final dose. Graph shows the mean ± SEM ofthe number of phospho-histone H3 (S10)-positive nuclei per square millimeterof tumor tissue, with 14.4 mm2 of tumor tissue analyzed for the vehicle controltumors, 14.5 mm2 of tumor tissue analyzed for the 6 mg/kg CFI-402257–treatedtumors, and 32.7 mm2 of tumor tissue analyzed for the 35 mg/kg CFI-402257–treated tumors. P values were calculated by using Student’s t test.

Control

)

2000 Anti-PD-1, 150 µg, Day 0, 3, 6, 10 IP

)

2000

Day10 20 30 40 500

Tum

or v

olum

e (m

m3 )

0

500

1000

1500

Day10 20 30 40 500

Tum

or v

olum

e (m

m3 )

0

500

1000

1500

CFI-402257, 6 mg/kg PO, QD

Tum

or v

olum

e (m

m3 )

500

1000

1500

2000 CFI-402257 + Anti-PD-1

Tum

or v

olum

e (m

m3 )

500

1000

1500

2000

Day10 20 30 40 500

0

Day10 20 30 40 500

0

Fig. 6. CFI-402257 in combination with anti–PD-1 antibodies induces completeregressions in the syngeneic CT26 model. When CT26 tumors reached an averagetarget size of ∼60 mm3, Balb/cJ mice were treated with four doses of anti–PD-1antibody (150 μg on days 0, 3, 6, and 10) or 21 doses of CFI-402257 6 mg/kg orally(PO) QD. The size of each individual tumor within each treatment arm is plotted.

Mason et al. PNAS Early Edition | 5 of 6

CELL

BIOLO

GY

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

7, 2

020

and treatments were initiated when tumor volumes reached 80−100 mm3. Totreat an established tumor, CFI-402257 and the vehicle [10% (vol/vol) NMP/40%(vol/vol) PEG-300/50% (vol/vol) water for the hydrochloride form, or 80–90% (vol/vol) PEG-300 or PEG-400/20–10% (vol/vol) water for the bisphosphate hemi-hydrate form] were administered by oral gavage, and carboplatin, anti–PD-1antibody, and isotype control were administered by i.p. injection to mice asdescribed earlier. Animal weights were monitored daily, and tumor volume wasmeasured three times per week. Tumor volume (in cubic millimeters) was de-fined as length × width2/2. Percentage of TGI was defined as 100 × [1 −(TVf,treated − TVi,treated)/(TVf,control − TVi,control)], where TVf is the average tumorvolume at the end of study and TVi is the average tumor volume at the

initiation of treatment. In cases in which tumor regression occurred, percentageof tumor regression was defined as 100 × [1 − (TVf,treated/TVi,treated)]. At thecompletion of the study, the mice were killed by an anesthetic overdose, andtumor tissue was removed for further studies.

ACKNOWLEDGMENTS. The authors thank Benjamin Neel for providing theovarian cancer PDX models and members of The Campbell Family Institutefor Breast Cancer Research for helpful discussions. This work was supportedby the Princess Margaret Cancer Foundation, the Canadian Institutes ofHealth Research, Genome Canada, and the California Institute forRegenerative Medicine.

1. Dominguez-Brauer C, et al. (2015) Targeting mitosis in cancer: Emerging strategies.Mol Cell 60(4):524–536.

2. Lara-Gonzalez P, Westhorpe FG, Taylor SS (2012) The spindle assembly checkpoint.Curr Biol 22(22):R966–R980.

3. Stolz A, et al. (2009) Pharmacologic abrogation of the mitotic spindle checkpoint byan indolocarbazole discovered by cellular screening efficiently kills cancer cells.Cancer Res 69(9):3874–3883.

4. Kwiatkowski N, et al. (2010) Small-molecule kinase inhibitors provide insight intoMps1 cell cycle function. Nat Chem Biol 6(5):359–368.

5. Lauzé E, et al. (1995) Yeast spindle pole body duplication gene MPS1 encodes anessential dual specificity protein kinase. EMBO J 14(8):1655–1663.

6. Abrieu A, et al. (2001) Mps1 is a kinetochore-associated kinase essential for the ver-tebrate mitotic checkpoint. Cell 106(1):83–93.

7. Maure JF, Kitamura E, Tanaka TU (2007) Mps1 kinase promotes sister-kinetochore bi-orientation by a tension-dependent mechanism. Curr Biol 17(24):2175–2182.

8. Tighe A, Staples O, Taylor S (2008) Mps1 kinase activity restrains anaphase during anunperturbed mitosis and targets Mad2 to kinetochores. J Cell Biol 181(6):893–901.

9. Colombo R, et al. (2010) Targeting the mitotic checkpoint for cancer therapy withNMS-P715, an inhibitor of MPS1 kinase. Cancer Res 70(24):10255–10264.

10. Hewitt L, et al. (2010) Sustained Mps1 activity is required in mitosis to recruit O-Mad2to the Mad1-C-Mad2 core complex. J Cell Biol 190(1):25–34.

11. Jemaà M, et al. (2013) Characterization of novel MPS1 inhibitors with preclinicalanticancer activity. Cell Death Differ 20(11):1532–1545.

12. Tardif KD, et al. (2011) Characterization of the cellular and antitumor effects of MPI-0479605,a small-molecule inhibitor of the mitotic kinase Mps1. Mol Cancer Ther 10(12):2267–2275.

13. Daniel J, Coulter J, Woo JH, Wilsbach K, Gabrielson E (2011) High levels of the Mps1checkpoint protein are protective of aneuploidy in breast cancer cells. Proc Natl AcadSci USA 108(13):5384–5389.

14. Landi MT, et al. (2008) Gene expression signature of cigarette smoking and its role inlung adenocarcinoma development and survival. PLoS One 3(2):e1651.

15. Salvatore G, et al. (2007) A cell proliferation and chromosomal instability signature inanaplastic thyroid carcinoma. Cancer Res 67(21):10148–10158.

16. Tannous BA, et al. (2013) Effects of the selective MPS1 inhibitor MPS1-IN-3 on glioblastomasensitivity to antimitotic drugs. J Natl Cancer Inst 105(17):1322–1331.

17. Thykjaer T, et al. (2001) Identification of gene expression patterns in superficial andinvasive human bladder cancer. Cancer Res 61(6):2492–2499.

18. Yuan B, et al. (2006) Increased expression of mitotic checkpoint genes in breast cancercells with chromosomal instability. Clin Cancer Res 12(2):405–410.

19. Slee RB, et al. (2014) Selective inhibition of pancreatic ductal adenocarcinoma cellgrowth by the mitotic MPS1 kinase inhibitor NMS-P715. Mol Cancer Ther 13(2):307–315.

20. Mason JM, et al. (2014) Functional characterization of CFI-400945, a Polo-like kinase4 inhibitor, as a potential anticancer agent. Cancer Cell 26(2):163–176.

21. Hu Z, et al. (2006) The molecular portraits of breast tumors are conserved acrossmicroarray platforms. BMC Genomics 7:96.

22. Shah SP, et al. (2012) The clonal and mutational evolution spectrum of primary triple-negative breast cancers. Nature 486(7403):395–399.

23. Liu Y, et al. (2016) Discovery of pyrazolo[1,5-a]pyrididine TTK inhibitors: CFI-402257is a potent, selective, bioavailable anticancer agent. ACS Med Chem Lett 7(7):671–675.

24. Xu Q, et al. (2009) Regulation of kinetochore recruitment of two essential mi-totic spindle checkpoint proteins by Mps1 phosphorylation. Mol Biol Cell 20(1):10–20.

25. Innocenti P, et al. (2016) Rapid discovery of pyrido[3,4-d]pyrimidine inhibitors ofmonopolar spindle kinase 1 (MPS1) using a structure-based hybridization approach.J Med Chem 59(8):3671–3688.

26. Liu Y, et al. (2015) The discovery of orally bioavailable tyrosine threonine kinase (TTK)inhibitors: 3-(4-(heterocyclyl)phenyl)-1H-indalone-5-carboxamides as anticancer agents.J Med Chem 58(8):3366–3392.

27. Wengner AM, et al. (2016) Novel Mps1 kinase inhibitors with potent antitumor ac-tivity. Mol Cancer Ther 15(4):583–592.

28. Colombo R, et al. (2015) Abstract 1638: Preclinical characterization of the novel TTKkinase inhibitor S81694 for the treatment of triple negative breast cancer. Cancer Res75(15 suppl):1638. doi: 10.1158/1538-7445.AM2015-1638.

29. Janssen A, Kops GJ, Medema RH (2009) Elevating the frequency of chromosome mis-segregation as a strategy to kill tumor cells. Proc Natl Acad Sci USA 106(45):19108–19113.

30. Martinez R, et al. (2015) Mitotic checkpoint kinase Mps1 has a role in normal physi-ology which impacts clinical utility. PLoS One 10(9):e0138616.

31. Torres EM, et al. (2007) Effects of aneuploidy on cellular physiology and cell division inhaploid yeast. Science 317(5840):916–924.

32. Emens LA, Middleton G (2015) The interplay of immunotherapy and chemotherapy:Harnessing potential synergies. Cancer Immunol Res 3(5):436–443.

33. Swart M, Vergrugge I, Beltman JB (2016) Combination approaches with immune-checkpoint blockade in cancer therapy. Front Oncol 6:223.

34. Stewart JM, et al. (2011) Phenotypic heterogeneity and instability of human ovariantumor-initiating cells. Proc Natl Acad Sci USA 108(16):6468–6473.

6 of 6 | www.pnas.org/cgi/doi/10.1073/pnas.1700234114 Mason et al.

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

7, 2

020