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

Experimental Design: We utilized the PCS1 gene set and our model of enzalutamide (ENZR) castration-resistant prostate cancer (CRPC) to identify targetable pathways and inhibitors for PCS1. The findings were evaluated in vitro and in the ENZR CRPC xenograft model in vivo.

Results: The results revealed that ENZR CRPC cells are enriched with PCS1 signature and that Forkhead box M1 (FOXM1) pathway is the central driver of this subtype. Notably, we identified Monensin as a novel FOXM1-binding agent that selectively targets FOXM1 to reverse the PCS1 signature and its associated stem-like features and reduces the growth of ENZR CRPC cells and xenograft tumors.

Conclusions: Our preclinical data indicate FOXM1 pathway as a 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 guide clinical evaluation of targeting FOXM1 to eradicate highly aggressive and lethal PCS1 prostate cancer tumors. Clin Cancer Res; 23(22); 6923–33. ©2017 AACR.

Translational Relevance

Androgen deprivation therapy including second-line treatment with enzalutamide is the mainstay of therapy for metastatic and castration-resistant prostate cancer. Recently, three novel prostate cancer subtypes were delineated from which PCS1 was the most aggressive and lethal subtype. However, there are no specific therapeutic targets available for PCS1 tumors. Here, we report Forkhead box M1 (FOXM1) pathway as a master regulator of the PCS1 subtype and ENZR CRPC and identified Monensin as a novel FOXM1 inhibitor that selectively targets PCS1. These preclinical results may guide clinical evaluation of targeting FOXM1 to target PCS1 in advanced prostate cancer including ENZR CRPC.

Prostate cancer is the second leading cause of cancer death among Western men (1). Androgen deprivation therapy targeting androgen receptor (AR) and blocking its signaling is the cornerstone of therapies for prostate cancer patients with metastatic and castration-resistant disease (2, 3). These include second-line therapies, namely enzalutamide (ENZ) and abiraterone acetate, that improve survival of patients with castration-resistant prostate cancer (CRPC; refs. 4, 5). Nevertheless, the effects are not curative, and resistance to these therapies rapidly occurs (6, 7). Recent advances in genotyping CRPC have underlined the role of heterogeneity in reactivation of AR activity: the AR-driven resistance in CRPC remains dependent on AR signaling; for example, via emergence of AR point mutations and splice variants (such as AR-V7) leading to acquired resistance to androgen deprivation therapies (3, 6, 8–12). Moreover, cross-talk with other signaling pathways that drive AR activity has been described (6, 8). In contrast, “AR-indifferent” disease, where the resistant cells lack AR expression and/or signaling activity, has recently been reported to be associated with cellular plasticity and neuroendocrine molecular features (13).

Due to acquired resistance and the significant biological heterogeneity seen in prostate cancer tumors, there has long been a clinical need to identify master regulators that could be targeted to treat 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 multiple publically available prostate cancer gene expression data sets (n > 4,600; ref. 14). The luminal-like type 1 signature PCS1 was characterized as the most aggressive and lethal form of prostate cancer (14). Interestingly, analysis of the circulating tumor cells from ENZ-resistant (ENZR) patient's revealed that most of the ENZR patients belonged to the PCS1 subtype (14, 15). However, there are no molecular targets or therapeutic options for patients with PCS1 tumors in the clinic.

In this study, we sought to identify novel targets and therapeutics for the most lethal prostate cancer subtype, PCS1. To achieve this goal, we utilized the PCS classifiers and our recently generated models of ENZR CRPC (14, 16) and found FOXM1 as a therapeutic target for tumors with a PCS1 subtype.

Cells

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

Compounds

Enzalutamide (ENZ; Haoyuan Chemexpress) and Thiostrepton (Sigma-Aldrich) were diluted in DMSO (Sigma-Aldrich), and Monensin (Mon; Sigma-Aldrich) was diluted in EtOH. ENZ concentration of 10 μmol/L, Thiostrepton concentration of 100 nmol/L, and Mon concentration of 10 nmol/L were used in all experiments unless otherwise noted.

Cell proliferation assay

Cell proliferation assay was performed on 96-/384-well plates (Greiner) by plating 2,000/1,000 cells/well in 100/35 μL of media and left to attach overnight. Compound dilutions were added and incubated for 72 hours. Cell viability was determined with CellTiter-Glo (CTG, Promega, Inc.) and confluence with live-cell imaging (Incucyte, Essen Bioscience Inc.). The luminescence signal (700 nm) from CTG was quantified using Tecan 200 plate reader (Tecan).

Gene expression analysis using bead arrays

ENZR CRPC cells were grown into approximately 70% confluence, and total RNA was extracted using TRIzol (Invitrogen). Integrity of the RNA was monitored prior to hybridization using a Bioanalyzer 2100 (Agilent) according to the manufacturer's instructions. Note that 500 ng of purified RNA was amplified with the TotalPrep Kit (Ambion), and the biotin-labeled cDNA was hybridized to Agilent.

Analysis of gene expression data

Differentially 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 expressed genes using Ingenuity Pathway Analysis Software (Ingenuity Systems Inc.), and gene set enrichments were analyzed using MSigDB. In order to identify drugs with similar or opposite effects on gene expression, Connectivity Map and LINCS Canvas database was used (18).

In silico transcriptomics analyses

Cancer Genome Browser, cBioPortal for Cancer Genomics database, and Genesapiens database were utilized in FOXM1 pathway gene expressions' analyses in prostate cancer patient samples and survival plots (19, 20).

Quantitative real-time PCR

Total RNA was extracted, and 2 μg of total RNA was reverse transcribed using the Transcriptor First Strand cDNA Synthesis Kit (Roche Applied Science). Real-time monitoring of PCR amplification of cDNA was performed using DNA primers on ABI PRISM 7900 HT Sequence Detection System with SYBR PCR Master Mix (Applied Biosystems). GAPDH levels were used as an internal standard, and each assay was performed in triplicate.

Molecular docking

EADock DSS engine–based Swissdock web server was utilized to dock molecular structures of FOXM1 DNA–binding domain (pdb id: 3G73) and Mon (21). Mon chemical structure was searched from Zinc database, dockings were performed 5 times for each compound, and the results were analyzed using UCSF Chimera. Molecular dynamics (MD) simulations of Mon was performed starting from their docking poses in DNA-binding domain of the FOXM1 protein as predicted by Glide SP program (22). Mutant forms of FOXM1 (R236A and Y241A; K278A and H287A) were created using MOE (23). All MD simulations were performed with the CUDA-accelerated Amber 14 program. FOXM1 force field parameters were obtained from the ff14SB force field and the ligand; Mon parameters came from generalized amber force field with charges derived from an RESP fit using an HF/6-31G electrostatic potential calculated using the Gaussian 09 program. MD simulations were carried out within AMBER 14 on WestGrid facilities from Compute Canada (https://www.westgrid.ca/). The production of MD simulation was conducted for 10 ns without any restraints under the NPT ensemble condition at a temperature of 298 K and pressure of 1 atm.

Mutagenesis of FOXM1

FOXM1 plasmid (DNASU) was double mutated on FOXM1-Mon–binding site (R236A_Y241A or K278A_H287A) using the Q5 Site-Directed Mutagenesis Kit (New England Biosciences) using following primers: R236A and Y241A: For: TACTCTGCCATGGCCATGATACAATTC and Rev: GGGTGGCGCCTCAGACACAGAGTTCTG; K278A and H287A: For: AACTCCATCCGCGCGAACCTTTCCCTGCACGAC and Rev: CTTCCAGCCTGGCGCGGCAATGTGCTTAAAGTAGG.

Chromatin immunoprecipitation

Cells treated with or without Mon (100 nmol/L) for 6 hours were cross-linked with PFA (Sigma-Aldrich) and sonicated to shear DNA. Chromatin immunoprecipitation (ChIP) assay was performed using the ChIP Assay Kit (Agarose Beads) according to the manufacturer's protocol (Millipore) and antibody against FOXM1. 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: CCAGAGGGAGAAGATGTCCA and REV: GTCGTTGTCCTCGAAAAAGC; CDC25B promoter, FOR: AAGAGCCCATCAGTTCCGCTTG and REV: CCCATTTTACAGACCTGGACGC; AURKB promoter, FOR: GGGGTCCAAGGCACTGCTAC and REV: GGGGCGGGAGATTTGAAAAG; CCNB1 promoter, FOR: CGCGATCGCCCTGGAAACGCA and REV: CCCAGCAGAAACCAACAGCCG; and Actin control, FOR: AGCGCGGCTACAGCTTCA and REV: CGTAGCACAGCTTCTCCTTAATGT.

Western blotting

Protein lysates (5–50 μg) were run on SDS-PAGE and transferred to nitrocellulose membranes, which were blocked in Odyssey Blocking Buffer (LI-COR Biosciences) at room temperature for 1 hour. Membranes were probed overnight at 4°C with primary antibodies FOXM1 (1:1,000), GAPDH (1:5,000), and vinculin (1:5,000; Abcam), washed 3 times with PBS containing 0.1% Tween for 5 minutes, and incubated for 1 hour with 1:5,000 diluted Alexa Fluor secondary antibodies (Invitrogen) at room temperature. Specific proteins were detected using ODYSSEY IR imaging system (LI-COR Biosciences).

Drug affinity responsive target stability assay

Cell lysates were incubated with vehicle or Mon, and the proteins were degraded with different concentrations of pronase according to the protocol as previously described (24). FOXM1 protein level was obtained using Western blot. GAPDH was used as a control.

Transfection and luciferase assay

42DENZR and 42FENZR cells were reverse transfected on 96-well plates (20,000 cells/well) with FOXM1 promoter region cloned to pGL3-Basic Luciferase reporter plasmid (Promega) kindly provided by Dr. Pradip Raychaudhuri (University of Chicago; ref. 25) using Lipofectamine (0.5 μL/well; Invitrogen). Renilla luciferase reporter construct was used as a transfection control. After 24 hours, concentration series of Mon (10 to 100 nmol/L) and control dilutions were added onto the cells for 18 hours. FOXM1–luciferase activity was measured using Dual-Luciferase Reporter Assay System (Promega) and microplate luminometer (Tecan) according to the manufacturer's instructions. All experiments were carried out in triplicate.

Cell transfections

42DENZR 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 FOXM1 or control siRNA (Santa Cruz Biotechnology) using Oligofectamine (Invitrogen) and OPTI-MEM media (Gibco). After 4 hours, OPTI-MEM media were replaced with complete media, and cells were incubated for 18 hours prior second transfection. After 48 hours, cells were harvested. The same protocol as siRNA was used for shFOXM1 transfections using shFOXM1 or control shRNA (Santa Cruz Biotechnology), and successfully transfected clones were selected for and expanded in complete media containing 10 μg/mL puromycin. For transient FOXM1 overexpression, LNCaP and 16DCRPC cells were seeded on 6-well plates, and FOXM1 plasmid (5 μg; DNASU) was transfected using Mirus T20/20 and OPTIMEM media (Invitrogen) according to the manufacturer's instructions. OPTI-MEM media were replaced after 24 hours with complete media ± 10 μmol/L ENZ, and cells were harvested after 48 hours.

Flow cytometry

Cells were exposed to Mon for 24, 48, and 72 hours (PI) or for 24 hours [aldehyde dehydrogenase (ALDH) activity and CD24+/CD49b staining], samples were stained with PI for 30 minutes at 4°C (subpopulation), CD24 and CD49b antibodies (1:20; Biolegend) for 60 minutes, or ALDH reagent according to the manufacturer's instructions (Stem Cell Technologies). Live cells were gated using staining with viability dye eFluor 506 (eBioscience). Data were acquired from 10,000 events on a Canto II (BD Biosciences). The results were analyzed using FlowJo (TreeStar).

Xenograft experiments

Athymic nude male mice (Harlan Sprague-Dawley), 5 weeks of age, were injected s.c. in the both flanks with 1 × 106 42DENZR cells in 200 μL of Matrigel without growth factors (BD Biosciences). Mice were castrated, and after 2 weeks of castration, mice were given ENZ (10 mg/kg/d). When tumor size reached 200 mm3, the mice 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 a week to monitor tumor growth (tumor volume = 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 after injections of the cells, and the tumors were harvested at the size of 100 to 300 mm3.

Statistical analyses

All 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.

“AR-indifferent” ENZR CRPC cells are enriched with PCS1 signature

Our gene expression profiling results showed that ENZR CRPC 42DENZR 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 significantly downregulated in both cells (Fig. 1A). The results were confirmed by gene set enrichment analyses (GSEA; Fig. 1C and Supplementary Fig. S1). Taken together, the analysis of 42DENZR and 42FENZR cells against the novel subtypes indicates that these cells are enriched with the most aggressive and lethal PCS, PCS1. Thus, these results reveal that our ENZR cells could be further utilized to study novel targets and develop therapeutics for PCS1 patient subtype.

Figure 1.

“AR-indifferent” ENZR CRPC cells are enriched with the most aggressive, lethal prostate cancer subtype, PCS1. A and B, Expression of PCS-specific genes, PCS1, PCS2, and PCS3 (14), in 42DENZR and 42FENZR cells (16). Fold changes were compared with 16DCRPC control cells. C, GSEA of PCS1 signature in 42DENZR and 42FENZR cells.

Figure 1.

“AR-indifferent” ENZR CRPC cells are enriched with the most aggressive, lethal prostate cancer subtype, PCS1. A and B, Expression of PCS-specific genes, PCS1, PCS2, and PCS3 (14), in 42DENZR and 42FENZR cells (16). Fold changes were compared with 16DCRPC control cells. C, GSEA of PCS1 signature in 42DENZR and 42FENZR cells.

Close modal

FOXM1 is a master regulator pathway in PCS1

To identify novel therapeutics for PCS1, we utilized the PCS1 gene set signature as well as the gene expression profiling of 42DENZR and 42FENZR cells and performed Connectivity Map (cMap) combined with LINCS analyses (18, 26) to identify compounds that could reverse all these three signatures (Fig. 2A). The results indicated that Thiostrepton, a Forkhead box M1 (FOXM1) inhibitor, is the most enriched compound reversing all three signatures (connectivity scores of −98, −92, and −90, in PCS1, 42DENZR, and 42FENZR, respectively). To confirm that the FOXM1 pathway is a master regulator in PCS1, we first confirmed that FOXM1 was upregulated in 42DENZR and 42FENZR cells compared with parental LNCaP and 16DCRPC cells (Supplementary Fig. S2A). Second GSEA analyses were performed, and the analysis revealed that the FOXM1 pathway is highly enriched and significantly activated in 42DENZR and 42FENZR (Fig. 2B; enrichment scores of 0.52 and 0.55 in 42DENZR and 42FENZR cells, respectively, P values < 0.01). Ingenuity upstream regulator pathway analysis also confirmed that the FOXM1 pathway was enriched in PCS1, 42DENZR, and 42FENZR cells with z-scores of 4.5, 3.8, and 3.3, respectively (Supplementary Fig. S2B). These data were further confirmed using qRT-PCR (Fig. 2C), showing that the FOXM1 pathway is upregulated in 42DENZR and 42FENZR. In contrast, we found that AR (16), TP53, and RB1 pathways were downregulated 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 to represent 24% of PCS1 signature (Fig. 2D), and PCS1 was induced when FOXM1 was overexpressed in LNCaP and 16DCRPC cells and downregulated when FOXM1 was silenced in 42DENZR and 42FENZR cells (Fig. 2E). Together, these data suggest that the FOXM1 pathway is a major regulator for PCS1.

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 and 42DENZR 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 of PCS1, 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, The mRNA expression of FOXM1, AURKB, AURKA, BIRC5, PLK1, CCNB1, CCNA2, GTSE1, CDK1, CDK3, CENPA, and KLK3 in prostate cancer patients (TCGA, n = 550). G, Survival plots of FOXM1, AURKB, PLK1, SKP2, and CCNB1 in prostate cancer patient data analyzed using Taylor data set in cBioportal (28).

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 and 42DENZR 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 of PCS1, 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, The mRNA expression of FOXM1, AURKB, AURKA, BIRC5, PLK1, CCNB1, CCNA2, GTSE1, CDK1, CDK3, CENPA, and KLK3 in prostate cancer patients (TCGA, n = 550). G, Survival plots of FOXM1, AURKB, PLK1, SKP2, and CCNB1 in prostate cancer patient data analyzed using Taylor data set in cBioportal (28).

Close modal

Because PCS1 has been linked to high-risk prostate cancer, we next addressed whether the observed FOXM1 pathway activation in PCS1 subtype is also seen in high-risk prostate cancer patients and whether the pathway expression has any correlation with disease-free survival. We found that high expression of FOXM1 and its target genes' (AURKB, AURKA, BIRC5, PLK1, CCNB1, CCNA2, GTSE1, CCNE2, CDK1, CDKN3, and CENPA) correlate with Gleason score (TCGA, N = 550; blue, downregulation; red, upregulation; Fig. 2F). This was confirmed using Cox regression analysis showing a significant relationship between FOXM1 and high Gleason compared with normal prostate samples (correlation coefficient of 0.51, P value < 0.0001) as well as low PSA (correlation coefficient of −0.51, P value < 0.0001; Supplementary Table S2). In addition, high FOXM1 pathway expression was also seen in metastatic prostate cancer patients compared with primary patient samples analyzed using in Genesapiens database (Supplementary Fig. S2C; ref. 27). Finally, FOXM1, AURKB, PLK1, CCNB1, and SKP2 mRNA expressions correlate with poor survival in prostate cancer patients (Fig. 2G, analyzed from the data of ref. 28). In summary, these results from both unbiased compound and pathway analyses reveal FOXM1 as a major pathway activated in patients with PCS1 tumors as well as 42DENZR and 42FENZR cells enriched with PCS1. FOXM1 pathway activity correlated with high risk and poor survival, which is in accordance with the previous findings that PCS1 is the most aggressive and lethal subtype of prostate cancer and that FOXM1 activity is associated with poor survival, metastasis, and resistance to therapy (29, 30).

Mon is a novel FOXM1 and PCS1 subtype–targeting agent

As the Connectivity Map results revealed the FOXM1 inhibitor Thiostrepton as a potential agent targeting PCS1 tumors, we first explored its effect on 16DCRPC, 42DENZR, and 42FENZR cells. The results revealed that Thiostrepton reduces cell proliferation, and although differential effect was seen in 42DENZR and 42FENZR compared with 16DCRPC (P value < 0.01), the EC50 values were similar between the cell lines (Fig. 3A). The antiproliferative effect of Thiostrepton was observed at lower concentration compared with 40 μmol/L of its IC50 for FOXM1 indicating possible toxicity, a concern that has been reported previously (31, 32). Thiostrepton is a natural compound that belongs to a group of natural antibiotics produced by Streptomyces species. Notably, other natural antibiotics were also among the most enriched compounds for PCS1, 42DENZR, and 42FENZR including bithionol, Mon, manumycin, oligomycin, selamectin, idarubicin, and CCCP (Table 1). Thus, we tested the binding of these compounds to the FOXM1 DNA–binding domain (pdb id 3G73) using in silico Swissdock molecular docking (21). Thiostrepton–FOXM1-binding affinity was used as a reference (31). We found that Mon has highest binding affinity to FOXM1 with ΔG of −12.94 kcal/mol and FullFitness of −1556.22 kcal/mol compared to Thiostrepton (31) or other hit compounds (Table 1). This effect was translated on the ability of Mon to exert a superior inhibitory effect on FOXM1 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-expressing 42DENZR cells by reducing cell proliferation (Fig. 3C) and inducing apoptosis (Supplementary Fig. S3C).

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 of Thiostrepton 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 various concentrations of Mon. Graph represents pooled data from three independent experiments. D, Frequent contact maps of FOXM1–Mon binding reveal that Arg236, Tyr241, Lys278, and His287 interact with Mon over 80% of the total MD simulation time. E, Effect of Mon on mutated FOXM1 compared with WT as measured by transcriptional activity (FOXM1 activity P values of <0.01 and <0.001 for R236A_Y241A or K278A_H287A, respectively). F, Effect of Mon on mutated FOXM1 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 without Mon (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 the manufacturer'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 independent experiments. K, Effect of Mon on prostate cancer subtype signatures PCS1, PCS2, and PCS3 in 42DENZR cells analyzed by gene expression profiling and GSEA. *, P < 0.05; **, P < 0.01; and ***, P < 0.001.

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 of Thiostrepton 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 various concentrations of Mon. Graph represents pooled data from three independent experiments. D, Frequent contact maps of FOXM1–Mon binding reveal that Arg236, Tyr241, Lys278, and His287 interact with Mon over 80% of the total MD simulation time. E, Effect of Mon on mutated FOXM1 compared with WT as measured by transcriptional activity (FOXM1 activity P values of <0.01 and <0.001 for R236A_Y241A or K278A_H287A, respectively). F, Effect of Mon on mutated FOXM1 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 without Mon (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 the manufacturer'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 independent experiments. K, Effect of Mon on prostate cancer subtype signatures PCS1, PCS2, and PCS3 in 42DENZR cells analyzed by gene expression profiling and GSEA. *, P < 0.05; **, P < 0.01; and ***, P < 0.001.

Close modal
Table 1.

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

DrugTargetFOXM1, estimated ΔG (kcal/mol)FOXM1, FullFitness (kcal/mol)
Monensin Antibiotic –12.94 –1556.22 
Idarubicin Antibiotic, topoisomerase II inhibitor –7.96 –807.24 
Bithionol Antibacterial and anthelmintic –6.49 –788.62 
Manumycin Antibiotic, Ras inhibitor –6.46 –601.04 
CCCP Antibiotic, inhibitor of oxidative phosphorylation –6.16 –776.52 
Thiostrepton FOXM1 inhibitor –9.4 kcal/mol (31)  
Oligomycin Antibiotic, mitochondrial ATP synthase inhibitor 3D structure not available in the ZINC database  
Selamectin Anthelmintic 3D structure not available in the ZINC database 
DrugTargetFOXM1, estimated ΔG (kcal/mol)FOXM1, FullFitness (kcal/mol)
Monensin Antibiotic –12.94 –1556.22 
Idarubicin Antibiotic, topoisomerase II inhibitor –7.96 –807.24 
Bithionol Antibacterial and anthelmintic –6.49 –788.62 
Manumycin Antibiotic, Ras inhibitor –6.46 –601.04 
CCCP Antibiotic, inhibitor of oxidative phosphorylation –6.16 –776.52 
Thiostrepton FOXM1 inhibitor –9.4 kcal/mol (31)  
Oligomycin Antibiotic, mitochondrial ATP synthase inhibitor 3D structure not available in the ZINC database  
Selamectin Anthelmintic 3D structure not available in the ZINC database 

To further gain insight into the molecular interactions between Mon and FOXM1, we conducted explicit solvent MD simulations. During 10 ns MD simulations, we observed that Mon was tightly bound to the DNA-binding region of FOXM1 throughout the simulation period, and Arg236, Tyr241, Lys278, and His287 residues of FOXM1 interact with Mon over 80% of the total MD simulation time using a frequent contact map (Fig. 3D). Importantly, the compound forms strong H-bond interactions with His 287 and Tyr241, whereas Leu291, His287, and Trp281 residues make strong hydrophobic contacts with the chemical core (Fig. 3D). Using drug DARTS assay (24), we confirmed that Mon binds to FOXM1 in cell assay (Supplementary Fig. S3D). To further evaluate if Mon binds to FOXM1 DNA-binding region, in silico–guided site-directed mutagenesis was performed. Based on the contact frequency map as shown in Fig. 3D, two double mutants (R236A and Y241A; K278A and H287A) were generated that contribute to Mon binding to FOXM1 and were overexpressed LNCaP cells. We found that Mon exhibited weaker effect on FOXM1 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 on R236A and Y241A, K278A, and H287A forms of FOXM1 mutants to study the binding to Mon. We found that Mon could not form critical H-bond interactions with Tyr241 and His 287 as they were mutated to Ala (Fig. 3G). These data support that Mon binds to FOXM1 DNA–binding domain. Finally, we confirmed that Mon affects FOXM1 binding to DNA using ChIP and confirmed that Mon 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 significant master regulator pathway inhibited by Mon (z-score −3.8, P value < 0.001; Supplementary Fig. S3E), which was confirmed using GSEA (Fig. 3I, enrichment score, −0.63, P value < 0.001). Moreover, these data were further confirmed with qRT-PCR showing that key FOXM1 target genes were downregulated after Mon treatment (Fig. 3J). To identify whether FOXM1 inhibition by Mon affects PCS-specific genes and pathways, we compared Mon gene expression data with PCS1, PCS2, and PCS3 signatures using GSEA. The results showed that Mon significantly reduces the PCS1 signature (enrichment score, −0.66, P value < 0.001) but does not have a significant effect on PCS2 or PCS3, indicating that Mon selectively targets PCS1 (Fig. 3K). Taken together, these data confirm that Mon directly binds to and targets FOXM1, resulting in blunting of the FOXM1 pathway and, in turn, PCS1 signature.

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

In particular, PCS1-specific gene signature was also reported to be enriched with stem-like phenotype (14). We compared the subtype-specific genes of PCS1, PCS2, and PCS3 with embryonic stem cell core signatures and found that approximately one-third of the PCS1 genes belong to these signatures (33, 34), whereas PCS2 and PCS3 signatures did not display these genes (<1% of the genes listed; Fig. 4A). Our ENZR CRPC cells enriched with PCS1 were also significantly enriched with ECS (ES score of 0.44, P values < 0.01), whereas Mon significantly downregulates ECS (ES score, −0.40, P value < 0.05; Fig. 4B). Based on the above finding, we explored the Mon effect on the population of CD24/CD49b+ cells, ALDH activity, and tumorspheres, as readouts of stem-like phenotype. First, we found that 42DENZR and 42FENZR were enriched with CD24/CD49b+ population (Supplementary Fig. S4A) and exhibit high ALDH (Supplementary Fig. S5A) compared with 16DCRPC. Targeting FOXM1 using Mon or siRNA reduces CD24/CD49b+ population (Fig. 4C; Supplementary Fig. S4B and S4C) or ALDH activity (Fig. 4D; Supplementary Fig. S5B and S5C). Interestingly, ALDHHigh cells which display higher expression of FOXM1 (cells sorted from 42DENZR and 42FENZR cells using FACS) were more sensitive to Mon compared with ALDHLow cells (Fig. 4E; Supplementary Fig. S5D and S5E). In addition, Mon significantly reduced the number and size of 42DENZR and 42FENZR cell tumorspheres (Fig. 4F). These data suggest that Mon targets not only FOXM1 but also stem cell-like phenotypes.

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 in embryonic stem cell signatures (33, 34). B, Mon effect on Wong and colleagues' embryonic stem cell signature assessed using GSEA on ctrl- and Mon-treated 42DENZR cells. C, CD49high population. Percentage of high CD49b-expressing cells in 42DENZR and 42FENZR cells in comparison with 16DENZR cells using flow cytometry (left). Effect of Mon on the percentage CD49high population in 42DENZR and 42FENZR cells using flow cytometry (middle). Effect of FOXM1 siRNA on CD49high 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 and 42FENZR 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 and 42FENZR cells with high and low ALDH activity in response to Mon treatment assessed by IncuCyte (Essen Biosciences). Graphs represent pooled data from three independent 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.

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 in embryonic stem cell signatures (33, 34). B, Mon effect on Wong and colleagues' embryonic stem cell signature assessed using GSEA on ctrl- and Mon-treated 42DENZR cells. C, CD49high population. Percentage of high CD49b-expressing cells in 42DENZR and 42FENZR cells in comparison with 16DENZR cells using flow cytometry (left). Effect of Mon on the percentage CD49high population in 42DENZR and 42FENZR cells using flow cytometry (middle). Effect of FOXM1 siRNA on CD49high 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 and 42FENZR 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 and 42FENZR cells with high and low ALDH activity in response to Mon treatment assessed by IncuCyte (Essen Biosciences). Graphs represent pooled data from three independent 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.

Close modal

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

Next, we investigated Mon effect on the growth of ENZ-resistant CRPC in vivo. Mice bearing 42DENZR xenograft tumors were treated with 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 in response to Mon (Supplementary Fig. S6A). The results indicated that Mon reduced tumor growth compared with vehicle treatment (Fig. 5A). Moreover, Mon significantly reduced the expression of FOXM1 and its pathway members, CCNB1 and SKP2 (Fig. 5B), as well as ALDH activity in the tumors (Fig. 5C; Supplementary Fig. S6B). Together, these results convey that Mon reduces the growth, FOXM1 pathway activity, and ALDH activity in 42DENZR tumors in vivo.

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. Graph represents pooled data from 6 vehicle- and 6 Mon-treated tumors. B, Relative mRNA expression assessed by qRT-PCR of FOXM1, CCNB1, and SKP2 in vehicle- vs. Mon-treated xenografts. C, Percentage of cells with high ALDH activity in 42DENZR tumor xenografts treated with vehicle or Mon. Graph represents pooled data from 4 vehicle- and 4 Mon-treated tumors. *, P < 0.05; **, P < 0.01.

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. Graph represents pooled data from 6 vehicle- and 6 Mon-treated tumors. B, Relative mRNA expression assessed by qRT-PCR of FOXM1, CCNB1, and SKP2 in vehicle- vs. Mon-treated xenografts. C, Percentage of cells with high ALDH activity in 42DENZR tumor xenografts treated with vehicle or Mon. Graph represents pooled data from 4 vehicle- and 4 Mon-treated tumors. *, P < 0.05; **, P < 0.01.

Close modal

This preclinical study addresses the major clinical challenge of targeting the most aggressive subtype of prostate cancer PCS1 derived from 4,600 patient's data (14). FOXM1 was found 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 in the 2016 Beltran cohort (compared with CRPC-Adeno; Supplementary Fig. S7A; ref. 13) and in NPp53 abiraterone-exceptional nonresponder mice (compared with NPp53 vehicle) that transdifferentiate to neuroendocrine (Supplementary Fig. S7B; ref. 35) as well as in RB1, TP53, and PTEN double and triple knockout compared with wild type in the 2017 Ku data set (Supplementary Fig. S7C; ref. 36). These data suggest that PCS1 signature is found in a broad phenotype of aggressive prostate cancer. In agreement, our ENZR cells exhibit a luminal B-like subtype by PAM50 classification (Pearson correlation 0.47, P values < 0.01), which is associated with the poorest clinical prognoses in prostate cancer (37). In addition, FOXM1 and its target genes were upregulated in high-risk prostate cancer and correlated with poor survival and was found to be increased in TP53Alt/RB1Altphenotype characterized by aggressive NEPC features (36). Importantly, mitotic kinase AURKA and N-Myc that previously were linked to NEPC phenotype (13) are FOXM1 target genes (38, 39). Together, these findings further link the PCS1 signature to FOXM1 pathway, AR indifferent and/or neuroendocrine prostate cancer.

We discovered Mon as a novel FOXM1-binding agent with higher binding affinity to previously established FOXM1 inhibitor Thiostrepton (40). Mon targets and reduces FOXM1 pathway activity and reduces PCS1 signature. Mon selectively inhibits cell proliferation in high FOXM1-expressing cells in vitro and in vivo without any toxicity, induces apoptosis, and reduces self-renewal. These data are in accordance with previous reports indicating that FOXM1 regulates major hallmarks of cancer such 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 a promising drug candidate to inhibit master regulator FOXM1 and PCS1 tumors.

In conclusion, our results reveal FOXM1 as a major pathway activated in PCS1 and ENZR CRPC and Mon as a novel FOXM1 and PCS1 subtype–targeting agent that also reduces stem-like phenotype in vitro and in vivo. High FOXM1 pathway expression also correlated with aggressive prostate cancer phenotype in prostate cancer patients including high Gleason score and poor survival. Because there is a lack of third-line treatment options for ENZR CRPC in the clinic and there are no therapies available for PCS1, our results indicate that targeting the FOXM1 pathway may provide a novel therapeutic strategy for this aggressive subset of prostate cancer associated with treatment resistance.

No potential conflicts of interest were disclosed.

Conception and design: K. Ketola, A. Zoubeidi

Development of methodology: K. Ketola, R.S.N. Munuganti, A. Davies, A. Zoubeidi

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): K. Ketola, K.M. Nip, J.L. Bishop

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): K. Ketola, R.S.N. Munuganti, K.M. Nip, A. Zoubeidi

Writing, 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 organizing data, constructing databases): K. Ketola, K.M. Nip

Study supervision: K. Ketola, A. Zoubeidi

The authors thank Dr. Pradip Raychaudhuri, University of Chicago, for the FOXM1 luciferase plasmid.

This work is supported by Prostate Cancer Canada and proudly funded by the Movember Foundation (T2013-01). A. Zoubeidi is supported by Michael Smith Foundation for Health Research, and K. Ketola is supported by Prostate Cancer Canada and US Department of Defense (PC141530). R.S.N. Munuganti is supported by Prostate Cancer Canada and Michael Smith Foundation for Health Research, A. Davies is supported by Canadian Institute for Health Research and Prostate Cancer Foundation, and J.L. Bishop is supported by Prostate Cancer Foundation.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
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 phosphatases in 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 Med
2004
;
10
:
33
9
.
4.
Scher
HI
,
Fizazi
K
,
Saad
F
,
Taplin
ME
,
Sternberg
CN
,
Miller
K
, et al
Increased survival with enzalutamide in prostate cancer after chemotherapy
.
N Engl J Med
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 J Med
2011
;
364
:
1995
2005
.
6.
Watson
PA
,
Arora
VK
,
Sawyers
CL
. 
Emerging mechanisms of resistance to androgen 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: a phase 1-2 study
.
Lancet
2010
;
375
:
1437
46
.
8.
Ferraldeschi
R
,
Welti
J
,
Luo
J
,
Attard
G
,
de Bono
JS
. 
Targeting the androgen receptor pathway in castration-resistant prostate cancer: progresses and prospects
.
Oncogene
2015
;
34
:
1745
57
.
9.
Joseph
JD
,
Lu
N
,
Qian
J
,
Sensintaffar
J
,
Shao
G
,
Brigham
D
, et al
A clinically relevant androgen receptor mutation confers resistance to second-generation 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
An F876L mutation in androgen receptor confers genetic and phenotypic resistance to MDV3100 (enzalutamide)
.
Cancer Discov
2013
;
3
:
1030
43
.
11.
Li
Y
,
Chan
SC
,
Brand
LJ
,
Hwang
TH
,
Silverstein
KA
,
Dehm
SM
. 
Androgen receptor 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
.
N Engl 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 prostate cancer
.
Nat Med
2016
;
22
:
298
305
.
14.
You
S
,
Knudsen
BS
,
Erho
N
,
Alshalalfa
M
,
Takhar
M
,
Al-Deen Ashab
H
, et al
Integrated classification of prostate cancer reveals a novel luminal subtype with poor outcome
.
Cancer Res
2016
;
76
:
4948
58
.
15.
Miyamoto
DT
,
Zheng
Y
,
Wittner
BS
,
Lee
RJ
,
Zhu
H
,
Broderick
KT
, et al
RNA-Seq of single prostate CTCs implicates noncanonical Wnt signaling in antiandrogen resistance
.
Science
2015
;
349
:
1351
6
.
16.
Bishop
JL
,
Thaper
D
,
Vahid
S
,
Davies
A
,
Ketola
K
,
Kuruma
H
, et al
The master neural transcription factor BRN2 is an androgen receptor-suppressed driver of neuroendocrine differentiation in prostate cancer
.
Cancer Discov
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-resistant and MDV3100-resistant prostate cancer growth in vitro and in vivo
.
Mol Cancer 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 interrogate 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 using the cBioPortal
.
Sci Sig
2013
;
6
:
pl1
.
20.
Kilpinen
S
,
Autio
R
,
Ojala
K
,
Iljin
K
,
Bucher
E
,
Sara
H
, et al
Systematic bioinformatic analysis of expression levels of 17,330 human genes across 9,783 samples from 175 types of healthy and pathological tissues
.
Genome Biol
2008
;
9
:
R139
.
21.
Grosdidier
A
,
Zoete
V
,
Michielin
O
. 
SwissDock, a protein-small molecule docking 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 Computing Group
; 
2008
.
Available from
: www.chemcomp.com.
24.
Lomenick
B
,
Jung
G
,
Wohlschlegel
JA
,
Huang
J
. 
Target identification using drug affinity responsive target stability (DARTS)
.
Curr Protoc Chem Biol
2011
;
3
:
163
80
.
25.
Petrovic
V
,
Costa
RH
,
Lau
LF
,
Raychaudhuri
P
,
Tyner
AL
. 
Negative regulation of the oncogenic transcription factor FoxM1 by thiazolidinediones and mithramycin
.
Cancer Biol Ther
2010
;
9
:
1008
16
.
26.
Lamb
J
,
Crawford
ED
,
Peck
D
,
Modell
JW
,
Blat
IC
,
Wrobel
MJ
, et al
The connectivity map: Using gene-expression signatures to connect small molecules, genes, and disease
.
Science
2006
;
313
:
1929
35
.
27.
Kilpinen
S
,
Autio
R
,
Ojala
K
,
Iljin
K
,
Bucher
E
,
Sara
H
, et al
Systematic bioinformatic analysis of expression levels of 17,330 human genes across 9,783 samples from 175 types of healthy and pathological tissues
.
Genome Biol
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 Cell
2010
;
18
:
11
22
.
29.
Aytes
A
,
Mitrofanova
A
,
Lefebvre
C
,
Alvarez
MJ
,
Castillo-Martin
M
,
Zheng
T
, et al
Cross-species regulatory network analysis identifies a synergistic interaction between FOXM1 and CENPF that drives prostate cancer malignancy
.
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 development and progression of prostate carcinomas in both TRAMP and LADY transgenic mice
.
Cancer Res
2006
;
66
:
1712
20
.
31.
Chen
Y
,
Ruben
EA
,
Rajadas
J
,
Teng
NN
. 
In silico investigation of FOXM1 binding and novel inhibitors in epithelial ovarian cancer
.
Bioorg Med Chem
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 programme via novel small molecule inhibition
.
Nat Commun
2014
;
5
:
5165
.
33.
Wong
JC
,
Jack
MM
,
Li
Y
,
O'Neill
C
. 
The epigenetic bivalency of core pancreatic beta-cell transcription factor genes within mouse pluripotent embryonic stem cells is not affected by knockdown of the polycomb repressive complex 2, SUZ12
.
PLoS One
2014
;
9
:
e97820
.
34.
Ben-Porath
I
,
Thomson
MW
,
Carey
VJ
,
Ge
R
,
Bell
GW
,
Regev
A
, et al
An embryonic stem cell-like gene expression signature in poorly differentiated aggressive 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 mouse model 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 and Trp53 cooperate to suppress prostate cancer lineage plasticity, metastasis, and antiandrogen resistance
.
Science
2017
;
355
:
78
83
.
37.
Zhao
SG
,
Chang
SL
,
Erho
N
,
Yu
M
,
Lehrer
J
,
Alshalalfa
M
, et al
Associations of luminal and basal subtyping of prostate cancer with prognosis and response to androgen deprivation therapy
.
JAMA Oncol
2017
.
38.
Lefebvre
C
,
Rajbhandari
P
,
Alvarez
MJ
,
Bandaru
P
,
Lim
WK
,
Sato
M
, et al
A human B-cell interactome identifies MYB and FOXM1 as master regulators of proliferation in germinal centers
.
Mol Syst Biol
2010
;
6
:
377
.
39.
Wang
IC
,
Ustiyan
V
,
Zhang
Y
,
Cai
Y
,
Kalin
TV
,
Kalinichenko
VV
. 
Foxm1 transcription factor is required for the initiation of lung tumorigenesis by oncogenic Kras(G12D.)
.
Oncogene
2014
;
33
:
5391
6
.
40.
Hegde
NS
,
Sanders
DA
,
Rodriguez
R
,
Balasubramanian
S
. 
The transcription factor FOXM1 is a cellular target of the natural product thiostrepton
.
Nat Chem
. 
2011
;
3
:
725
31
.
41.
Bella
L
,
Zona
S
,
Nestal de Moraes
G
,
Lam
EW
. 
FOXM1: a key oncofoetal transcription 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 progression and treatment
.
Biochim Biophys Acta
2012
;
1819
:
28
37
.

Supplementary data