targeting prostate cancer subtype 1 by forkhead box m1 … · odyssey blocking buffer (li-cor...

Cancer Therapy: Preclinical

Targeting Prostate Cancer Subtype 1 by ForkheadBox M1 Pathway InhibitionKirsi Ketola1, Ravi S.N. Munuganti1, Alastair Davies1, Ka Mun Nip1,Jennifer L. Bishop1, and Amina Zoubeidi1,2

Abstract

Purpose: Prostate cancer was recently classified to three clinicallyrelevant subtypes (PCS) demarcated by unique pathway activationand clinical aggressiveness. In this preclinical study, we investigatedmolecular targets and therapeutics for PCS1, the most aggressiveand lethal subtype,with no treatment options available in the clinic.

Experimental Design: We utilized the PCS1 gene set and ourmodel of enzalutamide (ENZR) castration-resistant prostate can-cer (CRPC) to identify targetable pathways and inhibitors forPCS1. The findings were evaluated in vitro and in the ENZR CRPCxenograft model in vivo.

Results: The results revealed that ENZR CRPC cells are enrichedwith PCS1 signature and that Forkhead box M1 (FOXM1) path-

way is the central driver of this subtype. Notably, we identifiedMonensin as a novel FOXM1-binding agent that selectively targetsFOXM1 to reverse the PCS1 signature and its associated stem-likefeatures and reduces the growth of ENZRCRPC cells and xenografttumors.

Conclusions:Our preclinical data indicate FOXM1 pathway asa master regulator of PCS1 tumors, namely in ENZR CRPC,and targeting FOXM1 reduces cell growth and stemness in ENZR

CRPC in vitro and in vivo. These preclinical results may guideclinical evaluation of targeting FOXM1 to eradicate highly aggres-sive and lethal PCS1 prostate cancer tumors. Clin Cancer Res; 23(22);6923–33. �2017 AACR.

IntroductionProstate cancer is the second leading cause of cancer death

amongWesternmen (1). Androgen deprivation therapy targetingandrogen receptor (AR) and blocking its signaling is the corner-stone of therapies for prostate cancer patients withmetastatic andcastration-resistant disease (2, 3). These include second-line ther-apies, namely enzalutamide (ENZ) and abiraterone acetate, thatimprove survival of patients with castration-resistant prostatecancer (CRPC; refs. 4, 5). Nevertheless, the effects are not curative,and resistance to these therapies rapidly occurs (6, 7). Recentadvances in genotyping CRPC have underlined the role of het-erogeneity in reactivation of AR activity: the AR-driven resistancein CRPC remains dependent on AR signaling; for example, viaemergence of AR point mutations and splice variants (such asAR-V7) leading to acquired resistance to androgen deprivationtherapies (3, 6, 8–12). Moreover, cross-talk with other signalingpathways that drive AR activity has been described (6, 8). Incontrast, "AR-indifferent" disease, where the resistant cells lack ARexpression and/or signaling activity, has recently been reported tobe associated with cellular plasticity and neuroendocrine molec-ular features (13).

Due to acquired resistance and the significant biological het-erogeneity seen in prostate cancer tumors, there has long been aclinical need to identifymaster regulators that could be targeted totreat the most lethal, aggressive prostate cancer. In a recent study,prostate cancer was classified to three prostate cancer subtypes(PCS), PCS1, PCS2, and PCS3, by utilizing and integrating mul-tiple publically available prostate cancer gene expression data sets(n > 4,600; ref. 14). The luminal-like type 1 signature PCS1 wascharacterized as the most aggressive and lethal form of prostatecancer (14). Interestingly, analysis of the circulating tumor cellsfrom ENZ-resistant (ENZR) patient's revealed that most of theENZR patients belonged to the PCS1 subtype (14, 15). However,there are no molecular targets or therapeutic options for patientswith PCS1 tumors in the clinic.

In this study, we sought to identify novel targets and thera-peutics for the most lethal prostate cancer subtype, PCS1. Toachieve this goal, we utilized the PCS classifiers and our recentlygeneratedmodels of ENZR CRPC (14, 16) and found FOXM1 as atherapeutic target for tumors with a PCS1 subtype.

Materials and MethodsCells

The prostate carcinoma CRPC and ENZR CRPC cells weregenerated from LNCaP as previously reported (16, 17), testedand authenticated by whole-genome and whole-transcriptomesequencing (Illumina Genome Analyzer IIx, 2012), and tested asfree of mycoplasma contamination and grown in RPMI 1640/10% FBS/1% glutamate/1 % penicillin–streptomycin (Hyclone),and 10 mmol/L ENZ or DMSO.

CompoundsEnzalutamide (ENZ; Haoyuan Chemexpress) and Thiostrepton

(Sigma-Aldrich) were diluted in DMSO (Sigma-Aldrich), and

1Vancouver Prostate Centre, Vancouver, British Columbia, Canada. 2Departmentof Urology, Faculty of Medicine, University of British Columbia, Vancouver,British Columbia, Canada.

Note: Supplementary data for this article are available at Clinical CancerResearch Online (http://clincancerres.aacrjournals.org/).

Corresponding Author: Amina Zoubeidi, Vancouver Prostate Centre, 2660 OakStreet, Vancouver, British Columbia V6H3Z6, Canada. Phone: 604-809-4947;Fax: 604-875-5654; E-mail: [email protected]

doi: 10.1158/1078-0432.CCR-17-0901

�2017 American Association for Cancer Research.

ClinicalCancerResearch

www.aacrjournals.org 6923

on March 5, 2021. © 2017 American Association for Cancer Research. clincancerres.aacrjournals.org Downloaded from

Published OnlineFirst September 12, 2017; DOI: 10.1158/1078-0432.CCR-17-0901

http://crossmark.crossref.org/dialog/?doi=10.1158/1078-0432.CCR-17-0901&domain=pdf&date_stamp=2017-11-14

http://clincancerres.aacrjournals.org/

Monensin (Mon; Sigma-Aldrich) was diluted in EtOH. ENZconcentration of 10 mmol/L, Thiostrepton concentration of100 nmol/L, and Mon concentration of 10 nmol/L were used inall experiments unless otherwise noted.

Cell proliferation assayCell proliferation assay was performed on 96-/384-well plates

(Greiner) by plating 2,000/1,000 cells/well in 100/35mL ofmediaand left to attach overnight. Compound dilutionswere added andincubated for 72 hours. Cell viability was determined with Cell-Titer-Glo (CTG, Promega, Inc.) and confluence with live-cellimaging (Incucyte, Essen Bioscience Inc.). The luminescencesignal (700 nm) from CTG was quantified using Tecan 200 platereader (Tecan).

Gene expression analysis using bead arraysENZR CRPC cells were grown into approximately 70% conflu-

ence, and total RNA was extracted using TRIzol (Invitrogen).Integrity of the RNA was monitored prior to hybridization usinga Bioanalyzer 2100 (Agilent) according to the manufacturer'sinstructions. Note that 500 ng of purified RNA was amplifiedwith the TotalPrep Kit (Ambion), and the biotin-labeled cDNAwas hybridized to Agilent.

Analysis of gene expression dataDifferentially expressed genes from microarray were set to a

minimum fold change of > 1.5. The functional gene ontology,pathway annotations, and upstream regulator pathway analyses(z-score) were analyzed for the sets of differentially expressedgenes using Ingenuity Pathway Analysis Software (IngenuitySystems Inc.), and gene set enrichments were analyzed usingMSigDB. In order to identify drugs with similar or opposite effectson gene expression, Connectivity Map and LINCS Canvas data-base was used (18).

In silico transcriptomics analysesCancer Genome Browser, cBioPortal for Cancer Genomics

database, and Genesapiens database were utilized in FOXM1pathway gene expressions' analyses in prostate cancer patientsamples and survival plots (19, 20).

Quantitative real-time PCRTotal RNA was extracted, and 2 mg of total RNA was reverse

transcribed using the Transcriptor First Strand cDNA Synthesis

Kit (Roche Applied Science). Real-timemonitoring of PCR ampli-fication of cDNA was performed using DNA primers on ABIPRISM 7900 HT Sequence Detection System with SYBR PCRMaster Mix (Applied Biosystems). GAPDH levels were used asan internal standard, and each assay was performed in triplicate.

Molecular dockingEADock DSS engine–based Swissdock web server was uti-

lized to dock molecular structures of FOXM1 DNA–bindingdomain (pdb id: 3G73) and Mon (21). Mon chemical structurewas searched from Zinc database, dockings were performed 5times for each compound, and the results were analyzed usingUCSF Chimera. Molecular dynamics (MD) simulations of Monwas performed starting from their docking poses in DNA-binding domain of the FOXM1 protein as predicted by GlideSP program (22). Mutant forms of FOXM1 (R236A and Y241A;K278A and H287A) were created using MOE (23). All MDsimulations were performed with the CUDA-accelerated Amber14 program. FOXM1 force field parameters were obtained fromthe ff14SB force field and the ligand; Mon parameters camefrom generalized amber force field with charges derived froman RESP fit using an HF/6-31G electrostatic potential calculatedusing the Gaussian 09 program. MD simulations were carriedout within AMBER 14 on WestGrid facilities from ComputeCanada (https://www.westgrid.ca). The production of MD sim-ulation was conducted for 10 ns without any restraints underthe NPT ensemble condition at a temperature of 298 K andpressure of 1 atm.

Mutagenesis of FOXM1FOXM1 plasmid (DNASU) was double mutated on FOXM1-

Mon–binding site (R236A_Y241A or K278A_H287A) using theQ5 Site-Directed Mutagenesis Kit (New England Biosciences)using following primers: R236A and Y241A: For: TACTCTGC-CATGGCCATGATACAATTC and Rev: GGGTGGCGCCTCAGA-CACAGAGTTCTG; K278A and H287A: For: AACTCCATCCGCG-CGAACCTTTCCCTGCACGAC and Rev: CTTCCAGCCTGGCGC-GGCAATGTGCTTAAAGTAGG.

Chromatin immunoprecipitationCells treated with or without Mon (100 nmol/L) for 6 hours

were cross-linked with PFA (Sigma-Aldrich) and sonicated toshear DNA. Chromatin immunoprecipitation (ChIP) assay wasperformed using the ChIP Assay Kit (Agarose Beads) according tothe manufacturer's protocol (Millipore) and antibody againstFOXM1. The binding or FOXM1 to its target genes' promoters,PLK1, CDC25B, AURKB, and CCNB1, was addressed using qPCR.The primers for promoters were PLK1 promoter, FOR: CCA-GAGGGAGAAGATGTCCA and REV: GTCGTTGTCCTCGAAAA-AGC; CDC25B promoter, FOR: AAGAGCCCATCAGTTCCGCTTGand REV: CCCATTTTACAGACCTGGACGC; AURKB promoter,FOR: GGGGTCCAAGGCACTGCTAC and REV: GGGGCGGGA-GATTTGAAAAG; CCNB1 promoter, FOR: CGCGATCGCCCTG-GAAACGCA and REV: CCCAGCAGAAACCAACAGCCG; andActin control, FOR: AGCGCGGCTACAGCTTCA and REV: CG-TAGCACAGCTTCTCCTTAATGT.

Western blottingProtein lysates (5–50 mg) were run on SDS-PAGE and trans-

ferred to nitrocellulose membranes, which were blocked in

Translational Relevance

Androgen deprivation therapy including second-line treat-ment with enzalutamide is the mainstay of therapy for met-astatic and castration-resistant prostate cancer. Recently, threenovel prostate cancer subtypes were delineated from whichPCS1 was the most aggressive and lethal subtype. However,there are no specific therapeutic targets available for PCS1tumors. Here, we report Forkhead box M1 (FOXM1) pathwayas amaster regulator of the PCS1 subtype and ENZR CRPC andidentified Monensin as a novel FOXM1 inhibitor that selec-tively targets PCS1. These preclinical results may guide clinicalevaluation of targeting FOXM1 to target PCS1 in advancedprostate cancer including ENZR CRPC.

Ketola et al.

Clin Cancer Res; 23(22) November 15, 2017 Clinical Cancer Research6924



https://www.westgrid.ca


Odyssey Blocking Buffer (LI-COR Biosciences) at room temper-ature for 1 hour. Membranes were probed overnight at 4�C withprimary antibodies FOXM1 (1:1,000), GAPDH (1:5,000), andvinculin (1:5,000; Abcam), washed 3 times with PBS containing0.1% Tween for 5minutes, and incubated for 1 hour with 1:5,000diluted Alexa Fluor secondary antibodies (Invitrogen) at roomtemperature. Specific proteins were detected using ODYSSEY IRimaging system (LI-COR Biosciences).

Drug affinity responsive target stability assayCell lysates were incubated with vehicle or Mon, and the

proteins were degraded with different concentrations of pronaseaccording to the protocol as previously described (24). FOXM1protein level was obtained using Western blot. GAPDH was usedas a control.

Transfection and luciferase assay42DENZR and 42FENZR cells were reverse transfected on 96-

well plates (20,000 cells/well) with FOXM1 promoter regioncloned to pGL3-Basic Luciferase reporter plasmid (Promega)kindly provided by Dr. Pradip Raychaudhuri (University ofChicago; ref. 25) using Lipofectamine (0.5 mL/well; Invitro-gen). Renilla luciferase reporter construct was used as a trans-fection control. After 24 hours, concentration series of Mon(10 to 100 nmol/L) and control dilutions were added onto thecells for 18 hours. FOXM1–luciferase activity was measuredusing Dual-Luciferase Reporter Assay System (Promega) andmicroplate luminometer (Tecan) according to the manufac-turer's instructions. All experiments were carried out intriplicate.

Cell transfections42DENZR and 42FENZR cells were plated on 10 cm plates

(1 million cells/10 mL complete media; Corning Life Sciences)for 18 to 24 hours prior to transfection with 10 nmol/L FOXM1orcontrol siRNA (Santa Cruz Biotechnology) using Oligofectamine(Invitrogen) andOPTI-MEMmedia (Gibco). After 4 hours, OPTI-MEM media were replaced with complete media, and cells wereincubated for 18 hours prior second transfection. After 48 hours,cells were harvested. The same protocol as siRNA was used forshFOXM1 transfections using shFOXM1 or control shRNA (SantaCruz Biotechnology), and successfully transfected clones wereselected for and expanded in complete media containing 10 mg/mL puromycin. For transient FOXM1 overexpression, LNCaP and16DCRPC cells were seeded on 6-well plates, and FOXM1 plasmid(5 mg; DNASU) was transfected using Mirus T20/20 andOPTIMEM media (Invitrogen) according to the manufacturer'sinstructions. OPTI-MEMmedia were replaced after 24 hours withcomplete media� 10 mmol/L ENZ, and cells were harvested after48 hours.

Flow cytometryCells were exposed to Mon for 24, 48, and 72 hours (PI) or for

24 hours [aldehyde dehydrogenase (ALDH) activity and CD24þ/CD49b� staining], samples were stainedwith PI for 30minutes at4�C (subpopulation), CD24 and CD49b antibodies (1:20; Bio-legend) for 60 minutes, or ALDH reagent according to the man-ufacturer's instructions (Stem Cell Technologies). Live cells weregated using staining with viability dye eFluor 506 (eBioscience).Data were acquired from 10,000 events on a Canto II (BDBiosciences). The results were analyzed using FlowJo (TreeStar).

Xenograft experimentsAthymic nudemalemice (Harlan Sprague-Dawley), 5 weeks of

age,were injected s.c. in thebothflankswith 1�10642DENZR cellsin 200 mL of Matrigel without growth factors (BD Biosciences).Mice were castrated, and after 2 weeks of castration, mice weregiven ENZ (10mg/kg/d). When tumor size reached 200mm3, themice were divided into two groups: (a) vehicle only and (b) Mon(10 mg/kg/3 times a week). Mice were treated for 3.5 weeks.Tumor volumes were calculated by caliper measurements twice aweek to monitor tumor growth (tumor volume ¼ LW2 � 0.56).For ALDH activity experiments, Mon (10 mg/kg/3 times a week)and ENZ (10 mg/kg/d) treatments started the next day afterinjections of the cells, and the tumors were harvested at the sizeof 100 to 300 mm3.

Statistical analysesAll in vitro and in vivo data were assessed using the Student t test.

Disease-free survival was analyzed using Kaplan–Meier curves.Levels of statistical significance were set at �, P < 0.05; ��, P < 0.01;and ��, P < 0.001.

Results"AR-indifferent" ENZR CRPC cells are enriched with PCS1signature

Our gene expression profiling results showed that ENZR CRPC42DENZR and 42FENZR cells are enriched with PCS1 signature(gene expression fold changes were compared with 16DCRPC

control cells, Fig. 1A and B), whereas PCS2 signature is signifi-cantly downregulated in both cells (Fig. 1A). The results wereconfirmed by gene set enrichment analyses (GSEA; Fig. 1C andSupplementary Fig. S1). Taken together, the analysis of 42DENZR

and 42FENZR cells against the novel subtypes indicates that thesecells are enriched with the most aggressive and lethal PCS, PCS1.Thus, these results reveal that our ENZR cells could be furtherutilized to study novel targets and develop therapeutics for PCS1patient subtype.

FOXM1 is a master regulator pathway in PCS1To identify novel therapeutics for PCS1, we utilized the PCS1

gene set signature as well as the gene expression profiling of42DENZR and 42FENZR cells and performed Connectivity Map(cMap) combined with LINCS analyses (18, 26) to identifycompounds that could reverse all these three signatures(Fig. 2A). The results indicated that Thiostrepton, a Forkhead boxM1 (FOXM1) inhibitor, is themost enriched compound reversingall three signatures (connectivity scores of�98,�92, and�90, inPCS1, 42DENZR, and 42FENZR, respectively). To confirm that theFOXM1 pathway is amaster regulator in PCS1, we first confirmedthat FOXM1 was upregulated in 42DENZR and 42FENZR cellscompared with parental LNCaP and 16DCRPC cells (Supplemen-tary Fig. S2A). Second GSEA analyses were performed, and theanalysis revealed that the FOXM1 pathway is highly enriched andsignificantly activated in 42DENZR and 42FENZR (Fig. 2B; enrich-ment scores of 0.52 and 0.55 in 42DENZR and 42FENZR cells,respectively, P values < 0.01). Ingenuity upstream regulator path-way analysis also confirmed that the FOXM1 pathway wasenriched in PCS1, 42DENZR, and 42FENZR cells with z-scores of4.5, 3.8, and 3.3, respectively (Supplementary Fig. S2B). Thesedata were further confirmed using qRT-PCR (Fig. 2C), showingthat the FOXM1 pathway is upregulated in 42DENZR and 42FENZR.

Targeting PCS1 by FOXM1 Pathway Inhibition

www.aacrjournals.org Clin Cancer Res; 23(22) November 15, 2017 6925




In contrast, we found that AR (16), TP53, and RB1 pathways weredownregulated in these cells with z-scores of �5.16, �5.02, and�4.15 for AR, TP53, and RB1, respectively (P values < 0.001;Supplementary Table S1), further suggesting that these cells are"AR-indifferent." Interestingly, the FOXM1 pathway was found torepresent 24%of PCS1 signature (Fig. 2D), andPCS1was inducedwhen FOXM1was overexpressed in LNCaP and16DCRPC cells anddownregulated when FOXM1 was silenced in 42DENZR and42FENZR cells (Fig. 2E). Together, these data suggest that theFOXM1 pathway is a major regulator for PCS1.

Because PCS1 has been linked to high-risk prostate cancer, wenext addressed whether the observed FOXM1 pathway activationin PCS1 subtype is also seen in high-risk prostate cancer patientsand whether the pathway expression has any correlation withdisease-free survival. We found that high expression of FOXM1and its target genes' (AURKB, AURKA, BIRC5, PLK1, CCNB1,CCNA2, GTSE1, CCNE2, CDK1, CDKN3, and CENPA) correlatewith Gleason score (TCGA, N ¼ 550; blue, downregulation; red,upregulation; Fig. 2F). This was confirmed using Cox regressionanalysis showing a significant relationship between FOXM1 andhigh Gleason compared with normal prostate samples (correla-tion coefficient of 0.51, P value < 0.0001) as well as low PSA(correlation coefficient of �0.51, P value < 0.0001; Supplemen-tary Table S2). In addition, high FOXM1 pathway expression wasalso seen in metastatic prostate cancer patients compared withprimary patient samples analyzed using in Genesapiens database(Supplementary Fig. S2C; ref. 27). Finally, FOXM1, AURKB, PLK1,CCNB1, and SKP2mRNA expressions correlate with poor survivalin prostate cancer patients (Fig. 2G, analyzed from the data of

ref. 28). In summary, these results frombothunbiased compoundandpathway analyses reveal FOXM1as amajor pathway activatedinpatientswithPCS1 tumors aswell as 42DENZR and42FENZR cellsenriched with PCS1. FOXM1 pathway activity correlated withhigh risk and poor survival, which is in accordance with theprevious findings that PCS1 is the most aggressive and lethalsubtype of prostate cancer and that FOXM1 activity is associatedwith poor survival, metastasis, and resistance to therapy (29, 30).

Mon is a novel FOXM1 and PCS1 subtype–targeting agentAs the Connectivity Map results revealed the FOXM1 inhibitor

Thiostrepton as a potential agent targeting PCS1 tumors, we firstexplored its effect on 16DCRPC, 42DENZR, and 42FENZR cells. Theresults revealed that Thiostrepton reduces cell proliferation, andalthough differential effect was seen in 42DENZR and 42FENZR

compared with 16DCRPC (P value < 0.01), the EC50 values weresimilar between the cell lines (Fig. 3A). The antiproliferative effectof Thiostrepton was observed at lower concentration comparedwith 40 mmol/L of its IC50 for FOXM1 indicating possible toxicity,a concern that has been reported previously (31, 32). Thiostrep-ton is a natural compound that belongs to a group of naturalantibiotics produced by Streptomyces species. Notably, othernatural antibiotics were also among the most enriched com-pounds for PCS1, 42DENZR, and 42FENZR including bithionol,Mon,manumycin, oligomycin, selamectin, idarubicin, andCCCP(Table 1). Thus, we tested the binding of these compounds to theFOXM1 DNA–binding domain (pdb id 3G73) using in silicoSwissdock molecular docking (21). Thiostrepton–FOXM1-bind-ing affinity was used as a reference (31). We found that Mon has

42D

42F

0.0

0.20.1

0.40.3

0.50.6

Enri

chm

ent s

core

(ES)

P < 0.001

42D : PCS1

42D 16D ENZR CRPC

PCS1 PCS2 PCS3

MCM4CCNB1CDC6CDKN3EZH2TPX2FOXM1KIF11HMMRMKI67KNTC1RAB3BANK3GJB1SLC12A2CFDPTGDSLTBP4SPEGGABRPACTC1ASPACOL4A6SGCASLC2A5PAGE4ACOX2C16orf45

PCS1

PCS2

PCS3

B

A

0.0

0.20.1

0.40.3

0.50.6

Enri

chm

ent s

core

(ES)

42F : PCS1

42F 16D ENZR CRPC

ENZR

ENZR

ENZR

ENZR

P < 0.001

10

-10

0

ES = 0.64

ES = 0.66

42DENZR 42FENZR

5

-5

0

C

Figure 1.

"AR-indifferent" ENZR CRPC cells areenriched with the most aggressive,lethal prostate cancer subtype, PCS1. AandB, Expression of PCS-specific genes,PCS1, PCS2, and PCS3 (14), in 42DENZR

and 42FENZR cells (16). Fold changeswere compared with 16DCRPC controlcells. C, GSEA of PCS1 signature in42DENZR and 42FENZR cells.

Ketola et al.





A

42D

42FENZR

ENZR

Dise

ase-

free

mRNA Upregula�on No change

P < 0.001 100

50

0 0 40 80 120 160

Dise

ase-

free

100

50

0

P < 0.05

Dise

ase-

free

100

50

0

Dise

ase-

free

100

50

0

Dise

ase-

free

100

50

0

CancerNormalHigh Gleason

Months disease-free Months disease-free Months disease-free Months disease-free Months disease-free 0 40 80 120 160 0 40 80 120 160 0 40 80 120 160 0 40 80 120

PCS1

1000-100

GTSE1CCNA2CCNB1

PLK1BIRC5

AURKAAURKBFOXM1

KLK3CENPACDNK3

CDK1CCNE2

FOXM1 AURKB PLK1 SKP2 CCNB1P < 0.001 P < 0.1P < 0.1

Pa�ents (n = 550), blue low, red high m

RNA expression

G

E

PCa Pa�entsPCS1 signature

Compounds (connec�vity score)

Comparison

Common hits (incl. FOXM1 inhibitor

Thiostrepton)

PCa Cell lines

42D42FENZR

ENZR

42D 16D Enric

hmen

t sco

re (E

S)

0.0

0.20.1

0.40.3

0.5

42F 16D Enric

hmen

t sco

re (E

S)

0.0

0.20.1

0.40.3

0.5FOXM1 PathwayFOXM1 Pathway

ENZR ENZRCRPC CRPC

B

D

0

2

4

6

12

Rela

vite

mRN

A

FOXM1AURKBCCNB1PLK1SKP2

***

***

***

**

***

**

**

**

42D ENZR 42F ENZR16D CRPC

**

C

FOXM1AURKB

PLK1CCNB1AURKA

BUB1MELK0

10

20

30 PCS1PCS2PCS3

% o

f Sub

type

gen

es

FOXM1 Pathway

PCS1

Moc

k

FOXM

1 oe

Moc

k

FOXM

1 oe

shCt

rl

shFO

XM1

shCt

rl

shFO

XM1

16D CRPC 42D ENZR 42F ENZRLNCaP F

- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

- - - - - - - -

1 13 0.3

ES: 0.55P < 0.05

ES: 0.52P < 0.05

Figure 2.

FOXM1 is a master regulator pathway in PCS1. A, Common compounds reversing the gene expression signature of prostate cancer (PCa) patient PCS1 subtype and42DENZR and 42FENZR cells analyzed using cMap and LINCS (18, 26). B, GSEA of the FOXM1 pathway in 42DENZR and 42FENZR cells. C, Expression of FOXM1,AURKB, PLK1, CCNB1, and SKP2 at mRNA levels by qRT-PCR in 16DCRPC, 42DENZR, and 42FENZR cells. � , P < 0.05; �� , P < 0.01; and �� , P < 0.001. D, The percentage ofPCS1, PCS2, and PCS3 subtype–specific genes in the FOXM1 pathway. E, The effect of FOXM1 overexpression in LNCaP and 16DCRPC cells and silencing in 42DENZR

and 42FENZR cells on PCS1 signature genes. F, ThemRNAexpression of FOXM1, AURKB, AURKA, BIRC5, PLK1, CCNB1, CCNA2, GTSE1, CDK1, CDK3, CENPA, andKLK3 inprostate cancer patients (TCGA, n¼ 550).G, Survival plots of FOXM1, AURKB, PLK1, SKP2, andCCNB1 in prostate cancer patient data analyzed using Taylor data set incBioportal (28).






-0.5

-0.3-0.4

-0.1-0.2

0.00.1

-0.6 -0.10

0.00-0.05

0.050.100.15

-0.050.050.150.25

-0.15-0.25

0.200.25

Enri

chm

ent s

core

(ES)

Enri

chm

ent s

core

(ES)

Enri

chm

ent s

core

(ES)Mon: PCS1 Mon: PCS2 Mon: PCS3

P < 0.001

Mon Mon Mon

A

K

00.20.40.60.8

11.2

0 5 10 15 20

Thiostrepton

Concentra�on (µmol/L) Concentra�on (µmol/L)

Rela

�ve

prol

ifera

�on

42DENZR

42FENZR

16DCRPC

P = 0.3 P = 0.4

00.2

0.4

0.60.8

11.2

0 0.001 0.010 0.1 1Rela

�ve

prol

ifera

�on

H

ES = -0.66

ES = 0.23 ES = 0.25

PCS1 PCS2 PCS3

- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

-5 5

Mon

00.010.020.030.040.050.060.07

PLK1 CDC25B AURKB CCNB1

% o

f Inp

ut

ctrlMon

Monensin

***

***** **

E

0

0.5

1.0

1.5

0 10 20 40 60 80 100

Luci

fera

se a

c�vi

ty

0

0.2

0.4

0.6

0.8

1.0

1.2

Ctrl Mon0

0.2

0.4

0.6

0.8

1.0

1.2

Ctrl Mon

Mon (nmol/L)

Rela

�ve

mRN

A le

vels FOXM1

AURKBCCNB1PLK1SKP2

J

Rela

�ve

mRN

A le

vels

** **

***

***

*

******

**

**

******

*** **

****

FOXM1AURKBCCNB1PLK1SKP2

GEn

richm

ent s

core

(ES)

-0.5-0.6

-0.3-0.4

-0.1-0.2

0.0FOXM1 pathway

Mon

P < 0.001

Ctrl

I42DENZR 42FENZR

42D ENZR

42F ENZR

FOXM1 ac�vity

** **** **

**

ES = -0.63

B

42DENZR

42FENZR

16DCRPCLNCaP

0.0

0.2

0.4

0.6

0.8

1.0

1.2

Ctrl Mon

Rela

�ve

luci

fera

se a

c�vi

tyWT

+ R236A_Y241A

+ K278A_H287A

**

***

0.0

0.2

0.4

0.6

0.8

1.0

1.2

Ctrl Mon

Rela

�ve

cell p

rolif

era�

on

WT

+ R236A_Y241A

+ K278A_H287A

**

***

H287 Y241

R236

K278

C

FOXM1 ac�vity Cell prolifera�onF

0

20

40

60

80

100

232 252 272 292 312

Freq

uenc

y %

Residue ID

R236Y241

K278H287D

H287 Y241

R236

K278

Ctrl

Ctrl Ctrl Ctrl

Figure 3.

Mon is a novel FOXM1 and PCS1 subtype–targeting agent.A,Relative cell proliferation of 16DCRPC, 42DENZR, and 42FENZR cells in response to various concentrations ofThiostrepton using CTG cell proliferation assay. Graph represents pooled data from three independent experiments. B, Effect of Mon (10 to 100 nmol/L)on FOXM1 transcriptional activity assessed by luciferase activity assay. C, Relative cell proliferation of 16DCRPC, 42DENZR, and 42FENZR cells in response to variousconcentrations of Mon. Graph represents pooled data from three independent experiments. D, Frequent contact maps of FOXM1–Mon binding reveal thatArg236, Tyr241, Lys278, and His287 interact with Mon over 80% of the total MD simulation time. E, Effect of Mon onmutated FOXM1 compared withWT as measuredby transcriptional activity (FOXM1 activity P values of <0.01 and <0.001 for R236A_Y241A or K278A_H287A, respectively). F, Effect of Mon on mutatedFOXM1 compared with WT as measured by cell proliferation (FOXM1 activity P values of <0.001 and <0.01 for R236A_Y241A or K278A_H287A, respectively).G,MD simulation showing that Mon could not form critical H-bond interactions with Tyr241 and His 287 as they were mutated to Ala. H, Cells treated with or withoutMon (100 nmol/L) for 6 hours were cross-linked with PFA (Sigma-Aldrich) and sonicated to shear DNA. ChIP assay was performed using the ChIP Assay Kit(Agarose Beads) according to themanufacturer's protocol (Millipore) and antibody against FOXM1. The FOXM1 binding to its target genes' promoters, PLK1, CDC25B,AURKB, and CCNB1 was evaluated using qPCR. I, Effect of Mon on the FOXM1 pathway in 42DENZR cells assessed by gene expression profiling and GSEA.J, mRNA expression of FOXM1 and its targets, AURKB, PLK1, SKP2, and CCNB1, assessed by qRT-PCR. Graph represents pooled data from three independentexperiments. K, Effect of Mon on prostate cancer subtype signatures PCS1, PCS2, and PCS3 in 42DENZR cells analyzed by gene expression profilingand GSEA. � , P < 0.05; �� , P < 0.01; and �� , P < 0.001.

Ketola et al.





highest binding affinity to FOXM1 with DG of �12.94 kcal/moland FullFitness of�1556.22 kcal/mol compared to Thiostrepton(31) or other hit compounds (Table 1). This effect was translatedon the ability of Mon to exert a superior inhibitory effect onFOXM1 transcriptional activity compared with Thiostrepton(Supplementary Fig. S3A). Importantly, Mon showed a dose-dependent inhibition on FOXM1 transcriptional activity (Fig.3B) and FOXM1 expression at protein levels (Supplementary Fig.S3B). Mon has predominant selectivity for high FOXM1-expres-sing 42DENZR cells by reducing cell proliferation (Fig. 3C) andinducing apoptosis (Supplementary Fig. S3C).

To further gain insight into the molecular interactions betweenMon and FOXM1, we conducted explicit solventMD simulations.During 10 ns MD simulations, we observed that Mon was tightlybound to the DNA-binding region of FOXM1 throughout thesimulation period, and Arg236, Tyr241, Lys278, and His287residues of FOXM1 interact with Mon over 80% of the total MDsimulation time using a frequent contact map (Fig. 3D). Impor-tantly, the compound forms strong H-bond interactions with His287 and Tyr241, whereas Leu291, His287, and Trp281 residuesmake strong hydrophobic contacts with the chemical core (Fig.3D). Using drug DARTS assay (24), we confirmed that Mon bindsto FOXM1 in cell assay (Supplementary Fig. S3D). To furtherevaluate if Mon binds to FOXM1 DNA-binding region, in silico–guided site-directed mutagenesis was performed. Based on thecontact frequency map as shown in Fig. 3D, two double mutants(R236A and Y241A; K278A and H287A) were generated thatcontribute to Mon binding to FOXM1 and were overexpressedLNCaP cells. We found that Mon exhibited weaker effect onFOXM1 transcriptional activity compared with WT (P values of<0.01 and <0.001 for R236A_Y241A or K278A_H287A,respectively; Fig. 3E) as well as on cell proliferation (P values of<0.001 and <0.01 for R236A_Y241A or K278A_H287A,respectively; Fig. 3F). We next performed MD simulations onR236A and Y241A, K278A, andH287A forms of FOXM1mutantsto study the binding to Mon. We found that Mon could not formcritical H-bond interactions with Tyr241 andHis 287 as they weremutated to Ala (Fig. 3G). These data support that Mon binds toFOXM1 DNA–binding domain. Finally, we confirmed that Monaffects FOXM1 binding to DNA using ChIP and confirmed thatMon reduced FOXM1 binding to promoters of its target genes(PLK1, CDC25B, AURKB, and CCNB1; Fig. 3H).

A genome-wide analysis of 42DENZR cells treated with Mon(Supplementary Table S3) revealed FOXM1 as the most signifi-cant master regulator pathway inhibited by Mon (z-score �3.8, Pvalue < 0.001; Supplementary Fig. S3E), which was confirmedusing GSEA (Fig. 3I, enrichment score, �0.63, P value < 0.001).Moreover, these data were further confirmedwith qRT-PCR show-ing that key FOXM1 target genes were downregulated after Montreatment (Fig. 3J). To identify whether FOXM1 inhibition by

Mon affects PCS-specific genes and pathways, we compared Mongene expression data with PCS1, PCS2, and PCS3 signatures usingGSEA. The results showed thatMon significantly reduces the PCS1signature (enrichment score,�0.66, P value < 0.001) but does nothave a significant effect on PCS2 or PCS3, indicating that Monselectively targets PCS1 (Fig. 3K). Taken together, these dataconfirm that Mon directly binds to and targets FOXM1, resultingin blunting of the FOXM1 pathway and, in turn, PCS1 signature.

PCS1 and ENZR CRPC cells display cancer stem–like featuresthat are targeted by FOXM1 inhibition

In particular, PCS1-specific gene signature was also reported tobe enriched with stem-like phenotype (14). We compared thesubtype-specific genes of PCS1, PCS2, and PCS3 with embryonicstem cell core signatures and found that approximately one-thirdof the PCS1 genes belong to these signatures (33, 34), whereasPCS2 andPCS3 signatures did not display these genes (<1%of thegenes listed; Fig. 4A). Our ENZR CRPC cells enriched with PCS1were also significantly enriched with ECS (ES score of 0.44, Pvalues < 0.01), whereas Mon significantly downregulates ECS (ESscore,�0.40, P value < 0.05; Fig. 4B). Based on the above finding,we explored theMon effect on the population of CD24�/CD49bþ

cells, ALDH activity, and tumorspheres, as readouts of stem-likephenotype. First, we found that 42DENZR and 42FENZR wereenriched with CD24�/CD49bþ population (Supplementary Fig.S4A) and exhibit high ALDH (Supplementary Fig. S5A) comparedwith 16DCRPC. Targeting FOXM1 using Mon or siRNA reducesCD24�/CD49bþ population (Fig. 4C; Supplementary Fig. S4Band S4C) or ALDH activity (Fig. 4D; Supplementary Fig. S5B andS5C). Interestingly, ALDHHigh cells which display higher expres-sion of FOXM1 (cells sorted from 42DENZR and 42FENZR cellsusing FACS)weremore sensitive toMoncomparedwithALDHLow

cells (Fig. 4E; Supplementary Fig. S5D and S5E). In addition,Monsignificantly reduced the number and size of 42DENZR and42FENZR cell tumorspheres (Fig. 4F). These data suggest that Montargets not only FOXM1 but also stem cell-like phenotypes.

Mon reduces ENZR CRPC xenograft growth, FOXM1 pathway,and ALDH activity in vivo

Next, we investigatedMoneffect on the growthof ENZ-resistantCRPC in vivo.Mice bearing 42DENZR xenograft tumorswere treatedwith vehicle or Mon (10 mg/kg) and monitored for tumor sizes.The mice were observed for 30 days for signs of weight loss,toxicity, and behavioral changes. No in vivo toxicity was seen inresponse to Mon (Supplementary Fig. S6A). The results indicatedthatMon reduced tumor growth comparedwith vehicle treatment(Fig. 5A). Moreover, Mon significantly reduced the expression ofFOXM1 and its pathwaymembers, CCNB1 and SKP2 (Fig. 5B), aswell as ALDH activity in the tumors (Fig. 5C; Supplementary Fig.S6B). Together, these results convey that Mon reduces the growth,

Table 1. In silico binding of antibiotic compounds reversing PCS1, 42DENZR, and 42FENZR signatures to FOXM1 DNA–binding domain analyzed using SwissDock

Drug Target FOXM1, estimated DG (kcal/mol) FOXM1, FullFitness (kcal/mol)

Monensin Antibiotic –12.94 –1556.22Idarubicin Antibiotic, topoisomerase II inhibitor –7.96 –807.24Bithionol Antibacterial and anthelmintic –6.49 –788.62Manumycin Antibiotic, Ras inhibitor –6.46 –601.04CCCP Antibiotic, inhibitor of oxidative phosphorylation –6.16 –776.52Thiostrepton FOXM1 inhibitor –9.4 kcal/mol (31)Oligomycin Antibiotic, mitochondrial ATP synthase inhibitor 3D structure not available in the ZINC databaseSelamectin Anthelmintic 3D structure not available in the ZINC database






Time (h)0 20 40 60 80

0.51.01.52.02.53.03.54.04.5

Rela

�ve

confl

uenc

e

0.81.01.21.41.61.82.02.22.42.6

CtrlMon

1.01.52.02.53.03.54.0

Ctrl Mon

0.01.02.03.04.05.06.0

Ctrl Mon

A

42D ALDHENZR

Ctrl 10 nmol/L 100 nmol/L

42D ALDHENZR 42F ALDHENZR 42F ALDHENZR

42DENZR

42FENZR

0

2

4

6

8

0

2

4

6

810

12

0

5

10

15

0

5

10

15

% A

LDH-

Posi�

ve c

ells

CD49

-Hig

h ce

lls (%

)

CD49

-Hig

h ce

lls (%

)

Time (h)0 20 40 60 80

Time (h)0 20 40 6080

Time (h)0 20 40 60 80

C

D

E

42DENZR 42FENZR16DCRPC 42DENZR 42FENZR

42DENZR 42FENZR42DENZR 42FENZR16DCRPC

0.0

CtrlMon

Ctrl Mon

***

***

*

***

** ***

*** ***

010

20

3040

5060

0 10 1000

5

1015

2025

30

Tum

orsp

here

s, n

0 10 100Mon (nmol/L)Mon (nmol/L)

**

**

***

42DENZR 42FENZR

0

5

10Ctrl siRNAFOXM1 siRNA

42DENZR 42FENZR

0

5

10

15

CD49

-Hig

h ce

lls (%

)

42DENZR 42FENZR

***

**

**

**

Enric

hmen

t sco

re (E

S)

0.000.10

0.300.20

0.40

42D

Wong - Embyonic Stem Cell Core

Enric

hmen

t sco

re (E

S)

-0.40-0.30-0.20

Wong - Embyonic Stem Cell Core

-0.10-0.000.10

ENZR

P < 0.01

P < 0.05

0

10

20

30

40

Ben-Porath -

PCS1PCS2PCS3

B

% o

f Sub

type

gen

es

FOXM1 siRNACtrl siRNACtrl

Mon

High HighLow Low

F

ES = 0.44

ES = -0.40

Tum

orsp

here

s, n

% A

LDH-

Posi�

ve c

ells

% A

LDH-

Posi�

ve c

ells

Rela

�ve

confl

uenc

e

Rela

�ve

confl

uenc

e

Rela

�ve

confl

uenc

e

Wong - EmbryonicStem Cell Core

EmbryonicStem 1

16DCRPC Mon

100 μm

100 μm100 μm

100 μm

100 μm

100 μm

Ctrl

Figure 4.

PCS1 and ENZR CRPC cells display stem-like features that are targeted by FOXM1 inhibition. A, The percentage of PCS1, PCS2, and PCS3 subtype–specific genes inembryonic stem cell signatures (33, 34). B, Mon effect on Wong and colleagues' embryonic stem cell signature assessed using GSEA on ctrl- and Mon-treated42DENZR cells. C, CD49high population. Percentage of high CD49b-expressing cells in 42DENZR and 42FENZR cells in comparison with 16DENZR cells using flowcytometry (left). Effect of Mon on the percentage CD49high population in 42DENZR and 42FENZR cells using flow cytometry (middle). Effect of FOXM1 siRNA onCD49high population in 42DENZR and 42FENZR cells using flow cytometry (right). D, ALDH activity. Percentage of ALDH activity in 42DENZR and 42FENZR

cells in comparison with 16DENZR cells assessed by Aldefluor ALDH activity using flow cytometry (left). Effect of Mon on ALDH activity in 42DENZR and42FENZR cells in comparison with 16DENZR cells assessed by Aldefluor ALDH activity using flow cytometry (middle). Effect of FOXM1 siRNA ALDH activity in 42DENZR

and 42FENZR cells in comparison with 16DENZR cells assessed by Aldefluor ALDH activity using flow cytometry (right). E, Relative confluence of 42DENZR and42FENZR cells with high and low ALDH activity in response to Mon treatment assessed by IncuCyte (Essen Biosciences). Graphs represent pooled data from threeindependent experiments. F, Effect of Mon on the size and number of 42DENZR and 42FENZR tumorspheres. �, P < 0.05; �� , P < 0.01; and ��, P < 0.001.

Ketola et al.





FOXM1 pathway activity, and ALDH activity in 42DENZR tumorsin vivo.

DiscussionThis preclinical study addresses the major clinical challenge

of targeting the most aggressive subtype of prostate cancerPCS1 derived from 4,600 patient's data (14). FOXM1 wasfound to be the most enriched master regulator pathway in"AR-indifferent" ENZR CRPC cells that we identified as PCS1.Interestingly, we found that PCS1 signature is upregulated in"AR-indifferent" CRPC-neuroendocrine (NEPC) patients inthe 2016 Beltran cohort (compared with CRPC-Adeno; Sup-plementary Fig. S7A; ref. 13) and in NPp53 abiraterone-exceptional nonresponder mice (compared with NPp53 vehi-cle) that transdifferentiate to neuroendocrine (SupplementaryFig. S7B; ref. 35) as well as in RB1, TP53, and PTEN doubleand triple knockout compared with wild type in the 2017Ku data set (Supplementary Fig. S7C; ref. 36). These datasuggest that PCS1 signature is found in a broad phenotypeof aggressive prostate cancer. In agreement, our ENZR cellsexhibit a luminal B-like subtype by PAM50 classification(Pearson correlation 0.47, P values < 0.01), which is associ-ated with the poorest clinical prognoses in prostate cancer

(37). In addition, FOXM1 and its target genes were upregu-lated in high-risk prostate cancer and correlated with poorsurvival and was found to be increased in TP53Alt/RB1Altphe-notype characterized by aggressive NEPC features (36). Impor-tantly, mitotic kinase AURKA and N-Myc that previously werelinked to NEPC phenotype (13) are FOXM1 target genes (38,39). Together, these findings further link the PCS1 signatureto FOXM1 pathway, AR indifferent and/or neuroendocrineprostate cancer.

We discovered Mon as a novel FOXM1-binding agent withhigher binding affinity to previously established FOXM1 inhib-itor Thiostrepton (40). Mon targets and reduces FOXM1 path-way activity and reduces PCS1 signature. Mon selectively inhi-bits cell proliferation in high FOXM1-expressing cells in vitroand in vivo without any toxicity, induces apoptosis, and reducesself-renewal. These data are in accordance with previous reportsindicating that FOXM1 regulates major hallmarks of cancersuch as proliferation/cell cycle, metastasis, genomic instability,stem cell renewal, DNA damage repair, and drug resistance (29,41–43). Together, these findings suggest that Mon is as apromising drug candidate to inhibit master regulator FOXM1and PCS1 tumors.

In conclusion, our results reveal FOXM1 as a major pathwayactivated in PCS1 and ENZR CRPC and Mon as a novel FOXM1

C

0

2

4

6

% A

LDH

-Pos

i�ve

cel

ls

Vehicle Mon

A B

mRN

A Ex

pres

sion

rela

�ve

to c

ontr

ol

FOXM1

SKP2CCNB1

0

2

4

6

8

10

12

0 0.5 1 1.5 2 2.5 3 3.5

***

*

Tum

or g

row

th (f

old

chan

ge)

Weeks

*

00.2

0.4

0.6

0.8

1

1.2

Vehicle Mon

**

** *

VehicleMon

0 10 3 10 4 10 50

50K

100K

150K

200K

250K

SSCA

0 10 3 10 4 10 50

50K

100K

150K

200K

250K

SSCA

MonVehicle

5.7%+/- 1.3

1.5%+/- 0.5

ALDH (FITC) ALDH (FITC)

Figure 5.

Mon reduces ENZR CRPC xenograft growth, FOXM1 pathway, and ALDH activity in vivo. A, Effect of vehicle or Mon on tumor growth of 42DENZR xenografts. Graphrepresents pooled data from 6 vehicle- and 6 Mon-treated tumors. B, Relative mRNA expression assessed by qRT-PCR of FOXM1, CCNB1, and SKP2 invehicle- vs. Mon-treated xenografts.C,Percentage of cellswith highALDHactivity in 42DENZR tumor xenografts treatedwith vehicle orMon. Graph represents pooleddata from 4 vehicle- and 4 Mon-treated tumors. � , P < 0.05; �� , P < 0.01.






and PCS1 subtype–targeting agent that also reduces stem-likephenotype in vitro and in vivo. High FOXM1 pathway expressionalso correlated with aggressive prostate cancer phenotype inprostate cancer patients including high Gleason score and poorsurvival. Because there is a lack of third-line treatment options forENZR CRPC in the clinic and there are no therapies available forPCS1, our results indicate that targeting the FOXM1 pathwaymayprovide a novel therapeutic strategy for this aggressive subset ofprostate cancer associated with treatment resistance.

Disclosure of Potential Conflicts of InterestNo potential conflicts of interest were disclosed.

Authors' ContributionsConception and design: K. Ketola, A. ZoubeidiDevelopment of methodology: K. Ketola, R.S.N. Munuganti, A. Davies,A. ZoubeidiAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): K. Ketola, K.M. Nip, J.L. BishopAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): K. Ketola, R.S.N. Munuganti, K.M. Nip, A. ZoubeidiWriting, review, and/or revision of the manuscript: K. Ketola, A. Davies,K.M. Nip, J.L. Bishop, A. Zoubeidi

Administrative, technical, or material support (i.e., reporting or organizingdata, constructing databases): K. Ketola, K.M. NipStudy supervision: K. Ketola, A. Zoubeidi

AcknowledgmentsThe authors thank Dr. Pradip Raychaudhuri, University of Chicago, for the

FOXM1 luciferase plasmid.

Grant SupportThiswork is supported byProstateCancerCanada andproudly fundedby the

Movember Foundation (T2013-01). A. Zoubeidi is supported byMichael SmithFoundation for Health Research, and K. Ketola is supported by Prostate CancerCanada and US Department of Defense (PC141530). R.S.N. Munuganti issupported by Prostate Cancer Canada and Michael Smith Foundation forHealth Research, A. Davies is supported by Canadian Institute for HealthResearch and Prostate Cancer Foundation, and J.L. Bishop is supported byProstate Cancer Foundation.

The costs of publication of this article were defrayed in part by thepayment of page charges. This article must therefore be hereby markedadvertisement in accordance with 18 U.S.C. Section 1734 solely to indicatethis fact.

Received March 27, 2017; revised July 27, 2017; accepted August 31, 2017;published OnlineFirst September 12, 2017.

References1. Siegel RL, Miller KD, Jemal A. Cancer statistics, 2015. CA Cancer J Clin

2015;65:5–29.2. Huggins C, Hodges CV. Studies on prostatic cancer: I. The effect of

castration, of estrogen and of androgen injection on serum phosphatasesin metastatic carcinoma of the prostate. 1941. J Urol 2002;168:9–12.

3. Chen CD, Welsbie DS, Tran C, Baek SH, Chen R, Vessella R, et al.Molecular determinants of resistance to antiandrogen therapy. Nat Med2004;10:33–9.

4. ScherHI, Fizazi K, Saad F, TaplinME, SternbergCN,Miller K, et al. Increasedsurvival with enzalutamide in prostate cancer after chemotherapy. N Engl JMed 2012;367:1187–97.

5. de Bono JS, Logothetis CJ, Molina A, Fizazi K, North S, Chu L, et al.Abiraterone and increased survival in metastatic prostate cancer. N Engl JMed 2011;364:1995–2005.

6. Watson PA, Arora VK, Sawyers CL. Emerging mechanisms of resistance toandrogen receptor inhibitors in prostate cancer. Nat Rev Cancer 2015;15:701–11.

7. Scher HI, Beer TM, Higano CS, Anand A, Taplin ME, Efstathiou E, et al.Antitumour activity of MDV3100 in castration-resistant prostate cancer: aphase 1-2 study. Lancet 2010;375:1437–46.

8. Ferraldeschi R, Welti J, Luo J, Attard G, de Bono JS. Targeting the androgenreceptor pathway in castration-resistant prostate cancer: progresses andprospects. Oncogene 2015;34:1745–57.

9. Joseph JD, LuN,Qian J, Sensintaffar J, ShaoG, BrighamD, et al. A clinicallyrelevant androgen receptor mutation confers resistance to second-gener-ation antiandrogens enzalutamide and ARN-509. Cancer Discov 2013;3:1020–9.

10. Korpal M, Korn JM, Gao X, Rakiec DP, Ruddy DA, Doshi S, et al. AnF876L mutation in androgen receptor confers genetic and pheno-typic resistance to MDV3100 (enzalutamide). Cancer Discov 2013;3:1030–43.

11. Li Y, Chan SC, Brand LJ, Hwang TH, Silverstein KA, Dehm SM. Androgenreceptor splice variants mediate enzalutamide resistance in castration-resistant prostate cancer cell lines. Cancer Res 2013;73:483–9.

12. Antonarakis ES, Lu C, Wang H, Luber B, Nakazawa M, Roeser JC, et al. AR-V7 and resistance to enzalutamide and abiraterone in prostate cancer. NEngl J Med 2014;371:1028–38.

13. Beltran H, Prandi D, Mosquera JM, Benelli M, Puca L, Cyrta J, et al.Divergent clonal evolution of castration-resistant neuroendocrine prostatecancer. Nat Med 2016;22:298–305.

14. You S, KnudsenBS, ErhoN, AlshalalfaM, TakharM, Al-DeenAshabH, et al.Integrated classification of prostate cancer reveals a novel luminal subtypewith poor outcome. Cancer Res 2016;76:4948–58.

15. MiyamotoDT, ZhengY,Wittner BS, Lee RJ, ZhuH, Broderick KT, et al. RNA-Seq of single prostate CTCs implicates noncanonical Wnt signaling inantiandrogen resistance. Science 2015;349:1351–6.

16. Bishop JL, Thaper D, Vahid S, Davies A, Ketola K, Kuruma H, et al. Themaster neural transcription factor BRN2 is an androgen receptor-sup-pressed driver of neuroendocrine differentiation in prostate cancer. CancerDiscov 2017;7:54–71.

17. Kuruma H, Matsumoto H, Shiota M, Bishop J, Lamoureux F, Thomas C,et al. A novel antiandrogen, Compound 30, suppresses castration-resistantand MDV3100-resistant prostate cancer growth in vitro and in vivo. MolCancer Ther 2013;12:567–76.

18. Duan QN, Flynn C, Niepel M, Hafner M, Muhlich JL, Fernandez NF, et al.LINCS Canvas Browser: interactive web app to query, browse and inter-rogate LINCS L1000 gene expression signatures. Nucleic Acids Res 2014;42:W449–W60.

19. Gao J, Aksoy BA, Dogrusoz U, Dresdner G, Gross B, Sumer SO, et al.Integrative analysis of complex cancer genomics and clinical profiles usingthe cBioPortal. Sci Sig 2013;6:pl1.

20. Kilpinen S, Autio R, Ojala K, Iljin K, Bucher E, Sara H, et al. Systematicbioinformatic analysis of expression levels of 17,330 human genes across9,783 samples from175 types of healthy and pathological tissues. GenomeBiol 2008;9:R139.

21. Grosdidier A, Zoete V, Michielin O. SwissDock, a protein-small moleculedocking web service based on EADock DSS. Nucleic Acids Res 2011;39:W270–7.

22. Friesner RA, Banks JL, Murphy RB, Halgren TA, Klicic JJ, Mainz DT, et al.Glide: a new approach for rapid, accurate docking and scoring. 1.Method and assessment of docking accuracy. J Med Chem 2004;47:1739–49.

23. Chemical Computing Group. Molecular Operating Environment. Quebec:Chemical ComputingGroup; 2008. Available from:www. chemcomp.com.

24. Lomenick B, Jung G, Wohlschlegel JA, Huang J. Target identification usingdrug affinity responsive target stability (DARTS). Curr Protoc Chem Biol2011;3:163–80.

25. Petrovic V, Costa RH, Lau LF, Raychaudhuri P, Tyner AL. Negative regu-lation of the oncogenic transcription factor FoxM1 by thiazolidinedionesand mithramycin. Cancer Biol Ther 2010;9:1008–16.


Ketola et al.



www. chemcomp.com


26. Lamb J, Crawford ED, Peck D, Modell JW, Blat IC, Wrobel MJ, et al. Theconnectivity map: Using gene-expression signatures to connect smallmolecules, genes, and disease. Science 2006;313:1929–35.

27. Kilpinen S, Autio R, Ojala K, Iljin K, Bucher E, Sara H, et al. Systematicbioinformatic analysis of expression levels of 17,330 human genes across9,783 samples from175 types of healthy and pathological tissues. GenomeBiol 2008;9:R139.

28. Taylor BS, Schultz N, Hieronymus H, Gopalan A, Xiao Y, Carver BS, et al.Integrative genomic profiling of human prostate cancer. Cancer Cell2010;18:11–22.

29. Aytes A,Mitrofanova A, LefebvreC, AlvarezMJ, Castillo-MartinM, Zheng T,et al. Cross-species regulatory network analysis identifies a synergisticinteraction between FOXM1 and CENPF that drives prostate cancer malig-nancy. Cancer Cell 2014;25:638–51.

30. Kalin TV, Wang IC, Ackerson TJ, Major ML, Detrisac CJ, Kalinichenko VV,et al. Increased levels of the FoxM1 transcription factor accelerate devel-opment and progression of prostate carcinomas in both TRAMP and LADYtransgenic mice. Cancer Res 2006;66:1712–20.

31. Chen Y, Ruben EA, Rajadas J, Teng NN. In silico investigation of FOXM1binding and novel inhibitors in epithelial ovarian cancer. Bioorg MedChem 2015;23:4576–82.

32. Gormally MV, Dexheimer TS, Marsico G, Sanders DA, Lowe C, Matak-Vinkovic D, et al. Suppression of the FOXM1 transcriptional programmevia novel small molecule inhibition. Nat Commun 2014;5:5165.

33. Wong JC, Jack MM, Li Y, O'Neill C. The epigenetic bivalency of corepancreatic beta-cell transcription factor genes within mouse pluripotentembryonic stem cells is not affected by knockdown of the polycombrepressive complex 2, SUZ12. PLoS One 2014;9:e97820.

34. Ben-Porath I, Thomson MW, Carey VJ, Ge R, Bell GW, Regev A, et al. Anembryonic stem cell-like gene expression signature in poorly differentiatedaggressive human tumors. Nat Genet 2008;40:499–507.

35. Zou M, Toivanen R, Mitrofanova A, Floch N, Hayati S, Sun Y, et al.Transdifferentiation as a mechanism of treatment resistance in a mousemodel of castration-resistant prostate cancer. Cancer Discov 2017;7:736–49.

36. Ku SY, Rosario S, Wang Y, Mu P, Seshadri M, Goodrich ZW, et al. Rb1 andTrp53 cooperate to suppress prostate cancer lineage plasticity, metastasis,and antiandrogen resistance. Science 2017;355:78–83.

37. Zhao SG, Chang SL, ErhoN, YuM, Lehrer J, AlshalalfaM, et al. Associationsof luminal and basal subtyping of prostate cancer with prognosis andresponse to androgen deprivation therapy. JAMA Oncol 2017.

38. Lefebvre C, Rajbhandari P, AlvarezMJ, Bandaru P, LimWK, SatoM, et al. Ahuman B-cell interactome identifies MYB and FOXM1 asmaster regulatorsof proliferation in germinal centers. Mol Syst Biol 2010;6:377.

39. Wang IC, Ustiyan V, Zhang Y, Cai Y, Kalin TV, Kalinichenko VV. Foxm1transcription factor is required for the initiation of lung tumorigenesis byoncogenic Kras(G12D.). Oncogene 2014;33:5391–6.

40. HegdeNS, SandersDA, Rodriguez R, Balasubramanian S. The transcriptionfactor FOXM1 is a cellular target of the natural product thiostrepton. NatChem. 2011;3:725–31.

41. Bella L, Zona S, Nestal de Moraes G, Lam EW. FOXM1: a key oncofoetaltranscription factor in health and disease. Sem Cancer Biol 2014;29:32–9.

42. Halasi M, Gartel AL. FOX(M1) news–it is cancer. Mol Cancer Ther 2013;12:245–54.

43. Koo CY,Muir KW, Lam EW. FOXM1: From cancer initiation to progressionand treatment. Biochim Biophys Acta 2012;1819:28–37.






2017;23:6923-6933. Published OnlineFirst September 12, 2017.Clin Cancer Res Kirsi Ketola, Ravi S.N. Munuganti, Alastair Davies, et al. InhibitionTargeting Prostate Cancer Subtype 1 by Forkhead Box M1 Pathway

Updated version

10.1158/1078-0432.CCR-17-0901doi:

Access the most recent version of this article at:

Material

Supplementary

http://clincancerres.aacrjournals.org/content/suppl/2017/09/12/1078-0432.CCR-17-0901.DC1

Access the most recent supplemental material at:

Cited articles

http://clincancerres.aacrjournals.org/content/23/22/6923.full#ref-list-1

This article cites 41 articles, 13 of which you can access for free at:

Citing articles

http://clincancerres.aacrjournals.org/content/23/22/6923.full#related-urls

This article has been cited by 5 HighWire-hosted articles. Access the articles at:

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

Subscriptions

Reprints and

[email protected]

To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://clincancerres.aacrjournals.org/content/23/22/6923To request permission to re-use all or part of this article, use this link



http://clincancerres.aacrjournals.org/lookup/doi/10.1158/1078-0432.CCR-17-0901

http://clincancerres.aacrjournals.org/content/suppl/2017/09/12/1078-0432.CCR-17-0901.DC1

http://clincancerres.aacrjournals.org/content/23/22/6923.full#ref-list-1

http://clincancerres.aacrjournals.org/content/23/22/6923.full#related-urls

http://clincancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]

http://clincancerres.aacrjournals.org/content/23/22/6923


targeting prostate cancer subtype 1 by forkhead box m1 … · odyssey blocking buffer (li-cor...

Documents