NB compounds are potent and efficacious FOXM1 inhibitors in high-grade serous ovarian cancer cells

CAOV3 and OVCAR5 cell lines were a gift from Dr. Anirban Mitra (Indiana University). CAOV3, OVCAR5, OVCAR8 (National Cancer Institute Division of Cancer Treatment and Diagnosis Cell Line Repository, Bethesda, MD, USA), and COV318 (Sigma-Aldrich, St. Louis, MO, USA) cell lines were cultured in DMEM (Corning, Corning, NY, USA) supplemented with 10% fetal bovine serum (FBS) (Gibco, Fisher Scientific, Waltham, MA, USA) and 1% penicillin–streptomycin (pen-strep) (Gibco). OVCAR4 (National Cancer Institute Division of Cancer Treatment and Diagnosis Cell Line Repository) were cultured in RPMI-1640 (Gibco) supplemented with 10% FBS and 1% pen-strep. The human immortalized FTE cell line FT282, a gift from Dr. Ronny Drapkin (University of Pennsylvania) [33], was used to generate the clonal cell line FT282-C11 as previously described [34], and was cultured in DMEM and Ham’s F12, 50/50 mix (Corning) supplemented with 10% FBS and 1% pen-strep. All cell lines were maintained at 37 °C in a humidified incubator with 5% CO₂. Cell lines were authenticated using short tandem repeat (STR) analysis at the DNA Services Facility, University of Illinois at Chicago. Cell lines were confirmed mycoplasma-free by PCR analysis using Mycofind™ Mycoplasma PCR Detection Kit (Clongen Laboratories, LLC, Gaithersburg, MD, USA).

Details on the preparation and spectroscopic characterization of NB compounds were previously described [29]. NB-73, NB-115, Thiostrepton (Sigma-Aldrich #598226), FDI-6 (Sigma-Aldrich #SML1392), RCM-1 (R&D Systems, Minneapolis, MN, USA, #6845), N-phenylphenanthren-9-amine (Sigma-Aldrich, St. Louis, MO, USA, #761966), MG132 (Sigma-Aldrich #474790), Q-VD-OPh (R&D Systems #OPH001), and olaparib (Selleck Chemicals, Houston, TX, USA, #S1060) were dissolved in DMSO. NB-55 and monensin (R&D Systems #5223) were dissolved in ethanol. Carboplatin (Sigma-Aldrich #C2538) was dissolved in water. All compounds were stored at -20 °C.

Nuclear and cytoplasmic protein fractions were extracted using the NE-PER Nuclear and Cytoplasmic Extraction Reagents (ThermoFisher Scientific, Waltham, MA, USA). Protein was isolated following the manufacturer’s protocol including supplementation of the CER I and NER reagents with protease and phosphatase inhibitors (Sigma-Aldrich). Protein concentrations were determined as described above.

Equivalent amounts of protein per well, as determined by BCA assays, were loaded into Invitrogen™ NuPAGE™ 4–12%, Bis–Tris, 1.5 mm, mini protein gels (Fisher Scientific) or 7.5% Mini-Protein TGX precast protein gels (Bio-Rad, Hercules, CA, USA), and subsequently transferred to 0.45 µm polyvinylidene difluoride (PVDF) membranes (Millipore Sigma, St. Louis, MO, USA), using a wet transfer system. Membranes were stained with ThermoFisher Scientific™ Pierce™ Reversible Protein Stain Kit (Fisher Scientific) or Ponceau S (Acros Organics, Fisher Scientific, Waltham, MA, USA) to confirm equivalent protein loading. Membranes were blocked with 5% nonfat dry milk (Kroger, Cincinnati, OH, USA) in TBS-T at room temperature. Membranes were incubated in primary antibodies in 5% BSA (Sigma-Aldrich). Primary antibodies included anti- FOXM1 (Cell Signaling Technology, CST, Danvers, MA, USA) (CST #5436, 1:1,000), p-FOXM1 (Thr600) (CST #14655, 1:1,000–1:2,000), AURKB (Abcam, Waltham, MA, USA, #2254, 1:1,000), CCNB1 (CST #4138, 1:1,000–1:2,000), CDC25B (CST #9525, 1:1,000–1:2000), PLK1 (CST #4513, 1:1,000) FOXA1 (CST #53528, 1:1,000–1:2,000)), FOXK2 (CST #12008, 1:1,000), FOXO3a (CST #12829, 1:10,000), cleaved PARP (cl-PARP) (CST #5625, 1:1,000), ubiquitin (CST #14049, 1:20,000). alpha-tubulin (CST #2144, 1:5,000), β-actin (Santa Cruz Biotechnology, Dallas, TX, USA, # 47778, 1:10,000), and Lamin B1 (CST #12586, 1:1,000). Following primary antibody incubation, membranes were washed in TBS-T at room temperature and then incubated in goat anti-rabbit secondary antibody (CST #7074, 1:500–1:5,000) or horse anti-mouse secondary antibody (CST #7076, 1:500–1:10,000) in 5% non-fat dry milk (Kroger) in TBS-T for 1 h at room temperature. SuperSignal™ West Pico PLUS Chemiluminescent Substrate (Fisher Scientific) was used for protein detection. Ultra blue X-ray films (Light Labs, Aurora, CO, USA) were applied to the blots, and subsequently developed in a standard film processor. Quantification of protein expression was performed using Fiji software [35].

Cells were seeded into 96-well plates and treated with serial dilutions with the designated compounds for 72 h. Medium was then removed from the wells, and the cells were frozen at -80 °C for storage. After thawing, cell viability was analyzed using the CyQUANT™ Cell Proliferation Assay for cells in culture (Invitrogen, ThermoFisher Scientific, Waltham, MA, USA), according to manufacturer’s instructions. The CyQuant assay utilizes a DNA binding fluorescent dye, and the signal is proportional to the number of live cells, which thus reflects the effect of drug treatment on both cell proliferation and cytotoxicity. To integrate these two measures (proliferation and cytotoxicity) into one simple term, we refer to CyQuant data as a measure of cell viability. Fluorescence intensity was measured using POLARstar OPTIMA microplate multi-detection plate reader (BMG LabTech, Cary, NC, USA) with settings specified by the manufacturer’s instructions.

RNA was extracted using TRIzol Reagent (Invitrogen) and purified with the Direct-zol™ RNA MiniPrep Kit (Zymo Research, Irvine, CA, USA) following the manufacturer’s protocol with DNase treatment. RNA integrity was confirmed by running denatured samples on agarose gels in MOPS buffer containing formaldehyde. RNA yield and purity was assessed using a Nanodrop 2000 instrument (ThermoFisher) and by determining 260/280 and 260/230 nm sample absorbance. cDNA was generated using 200–1,000 ng RNA with the High-Capacity cDNA Reverse Transcription kit (Applied Biosystems, ThermoFisher Scientific, Waltham, MA, USA) following the manufacturer’s protocol. cDNA was diluted 1:5 in PCR-grade water (Sigma) and 1.0 μl sample was added to a mix of iTaq Universal SYBR® Green Supermix and primers. PCR primer sequences are listed in Table 1. Reaction mixtures were run using the CFX Connect Real-Time System (Bio-Rad) with an annealing temperature of 60^OC for 40 cycles. Standard curves were generated using products from endpoint RT-PCR purified by gel-purification (QIAquick PCR purification kit, Qiagen, Germantown, MD, USA). mRNA measurement for genes of interest were normalized to 18S rRNA.

CAOV3 cells were seeded into 6-well or 35 mm plates and treated with DMSO or NB-73. RNA and protein were isolated in parallel from untreated samples at the time of treatment (0 h) and at time points after treatment. RNA was isolated and used for RT-qPCR analysis. Whole cell protein extracts were isolated and used for western blotting.

To assess FOXM1 protein expression, cells were seeded into 6-well plates and treated with DMSO or NB compounds the next day. At 3- or 6-h post-treatment, media was aspirated and cells were washed with PBS and incubated with drug-free media until reaching 24, 48, or 72 h post the initial treatment. At time points post-treatment, protein was isolated from washout and no washout (i.e., continuous treatment) treatment conditions and used for western blotting. To assess the effects of drug washout on cell viability, cells were seeded in 96-well plates, treated as described above, and cell viability was assessed at 72 h post the initial treatment using CyQuant assays.

CAOV3 cells were seeded into 96-well plates in DMEM media containing Incucyte Nuclight Rapid Red Dye (Sartorius, Bohemia, NY, USA, Cat #4717) 1:1,000 (for live cell nuclear labeling) and Incucyte Cytotox Green Dye (Sartorius Cat #4633) 250 μM (for counting of dead cells). The following day, cells were treated with DMSO or NB compounds at 0.1, 0.5 or 1.0 μM concentrations. After treatment, cells were incubated in a Incucyte S3 time lapse imager (Sartorius). We obtained 5 pictures per well every 2 h, for a total of 72 h. Data were analyzed and plotted using Incucyte Live-cell Imaging and Analysis software.

Cells were seeded into 6-well plates and treated with the designated compound(s) for 24 h. For cell collection, the adherent cells were trypsinized, centrifuged at 500 × g to form a pellet, and washed with PBS. For cell fixation, the washed cells were centrifuged at 500 × g to form a pellet, PBS was aspirated from the cell pellet, and 1.0 mL 70% ice-cold ethanol was slowly added dropwise while the cell pellet was gently vortexed. Cells were stored in -20 °C for at least 3 h. For cell staining, the fixed cells were washed with 4.5 mL PBS thrice, centrifuging at 2,000 × g to form a pellet each time, and the final cell pellet was resuspended in 200 µL FxCycle™ PI/RNase Staining Solution (ThermoFisher Scientific). The cells were incubated for 30 min, protected from light. Cells were transferred to a 1.5 mL microcentrifuge tube for analysis. Cell cycle was analyzed by flow cytometry using the Cell Cycle program on the Guava® Muse® Cell Analyzer. A stained sample was used to adjust instrument settings, determine gating strategies, and define the DNA profile content histogram. Experimental samples were thoroughly resuspended before loading onto the Guava® Muse® Cell Analyzer, and 10,000 events were acquired for each experimental sample.

For 2D anchorage-dependent colony formation assays, 2500 cells in single-cell suspension were seeded into triplicate wells of 6-well plates, incubated for 24 h, then treated with the designated compounds and allowed to form colonies for 7 days (OVCAR4, OVCAR5, OVCAR8) or 14 days (CAOV3). At the end of incubation period, cells were fixed in methanol, stained with crystal violet, washed with water, and air-dried overnight. Pictures of the wells were captured and colonies were counted using Count and Plot Histograms of Colony Size (countPHICS) software [36], a macro written for ImageJ. The threshold for colony size was automatically determined by countPHICS.

For 3D anchorage-independent colony formation assays, 2500 cells in single-cell suspension were seeded into a 0.4% SeaPlaque™ Agarose (Lonza Bioscience, Rockland, ME, USA) liquid solution and placed on top of a solidified layer of 0.8% SeaPlaque™ Agarose solution in triplicate wells of 6-well plates. The top layer with cells was allowed to solidify at room temperature for at least 30 min. After 24 h, cells were treated with the designated compounds in the media above the two agarose layers and allowed to form colonies for 14 days. Following incubation, cells were stained with crystal violet and washed with water. Pictures of the wells were taken, and colonies were counted using countPHICS software [36]. The threshold for colony size was automatically determined by countPHICS.

CyQuant assays were performed as described above. Concentrations of each compound were decided based on the diagonal method to measure drug synergy and are presented in Table 2 [37]. Drug interactions and synergy assessment was determined using CompuSyn software (ComboSyn, Inc., Paramus, NJ, USA).

For simultaneous drug treatment, cells were seeded into 96 well plates at 2500 cells per well and 24 h later the media was removed and treated with NB compounds and carboplatin. 72 h later, media was removed, and the plates were transferred to -80° C. Later, plates were thawed and processed for CyQuant assays. For sequential drug treatment, cells were seeded into 96 well plates at 2500 cells per well, and 24 h later the media was removed and treated with NB compounds. 24 h later, the media was removed, and cells were treated with carboplatin. 48 h later, the media was removed, plates were transferred to -80° C. Later, plates were thawed and processed for CyQuant assays.

Experiments were performed using three biological replicates, typically each with technical triplicate measurements. In some cases, representative data from one replicate are presented (e.g., western blot images). Statistical analyses were performed using GraphPad Prism and specific tests used are provided in Figure Legends. If the numerical p-values are not specifically indicated, the significance notation used is * for p < 0.05, ** for p < 0.01, *** for p < 0.001, and **** for p < 0.0001.

We investigated the effect of eight previously reported FOXM1 inhibitors on HGSOC cell viability. We utilized CAOV3 and OVCAR4 cells, which are validated to model HGSOC both in vitro and in vivo [38, 39], and measured the effect on cell viability using CyQuant assays. The data are summarized in Table 3 (drugs ordered alphabetically). Of the eight tested compounds, two had low potency (FDI-6, NP9) and one had no observable activity (RCM-1) on HGSOC cells. Monensin had high potency but low efficacy in one cell line (OVCAR4). Thiostrepton showed high potency and efficacy. The three diarylethene compounds (NB compounds) also showed high potency and efficacy, with NB-73 and NB-115 outperforming NB-55.

To investigate the cancer cell selectivity of FOXM1 inhibitors, we utilized the immortalized (FTE) cell line FT282-C11 as non-malignant, cell of origin matched control [33, 34]. FT282-C11 cells have low expression of FOXM1, FOXM1-P (activated FOXM1), CCNB1 (canonical FOXM1 target), FOXA1, and FOXK2 (all oncoproteins), and elevated FOXO3a (a tumor suppressor), as compared to HGSOC cells (Fig. 1A), validating their utility as a normal control. We tested a sub-set of FOXM1 inhibitors, including the three NB compounds and the two most widely used FOXM1 inhibitors, thiostrepton and FDI-6. The data summarized in Table 4 (drugs ordered alphabetically) revealed that NB-73 and NB-115 have higher selectivity (approximately tenfold) for HGSOC cells compared to the other three compounds, which exhibit little selectivity. Representative CyQuant data for NB-73 and NB-115 are shown in Fig. 1B and C. We further validated the activity of NB-73 and NB-115 on two additional HGSOC cell lines, OVCAR5 and OVCAR8 (Fig S1). Thus, based on our analysis of FOXM1 inhibitor potency, efficacy, and selectivity, we chose NB-73 and NB-115 for further study.

We measured FOXM1 protein expression as the primary molecular pharmacodynamic (PD) endpoint, as NB compounds are known to directly bind to FOXM1, promoting its degradation [29]. In addition, we measured a transcriptionally active form of FOXM1, FOXM1-pThr600 (i.e., FOXM1-P), to infer the effect of drug treatment on the FOXM1 pathway [40, 41]. In both cell lines, NB compounds, at low micromolar concentrations, suppressed FOXM1 and FOXM1-P expression at 24-, 48, and 72-h post-treatment. We observed the most robust FOXM1 suppression at the 48-h time point in CAOV3 cells and the 72-h time point in OVCAR4 cells (Fig. 2A-D).

Next, we examined whether NB compounds suppress nuclear FOXM1 expression, which is likely the biologically active fraction of FOXM1, by quantifying protein expression in the nuclear and cytosolic compartments of HGSOC cells. Notably, in both HGSOC cell lines, NB compounds suppressed FOXM1 in both the cytosolic and nuclear compartments (Fig. 2E-F). FOXM1-P was enriched in the nuclear compartment, particularly in CAOV3 cells, and was suppressed by NB compound treatment (Fig. 2E-F).

Next, we measured the effect of NB compound treatment on the protein expression of FOXM1 target genes. We analyzed four canonical FOXM1 targets: Cyclin B1 (CCNB1), PLK1, AURKB, and CDC25B [42‐44]. As seen for FOXM1 and FOXM1-P, low micromolar treatments with NB-73 or NB-115 suppressed FOXM1 targets at the protein level in both cell lines (Fig. 3A-D). Because FOXM1 promotes the gene expression of its targets, we also utilized RT-qPCR to measure the expression of four well characterized FOXM1 mRNA targets following NB compound treatment (FOXM1, CCNB1, SKP2, CDC25B). FOXM1 was included in these genes as FOXM1 is reported to regulate its own gene expression [15]. The data showed that NB compound treatment causes suppressed of FOXM1 target genes in both cell lines (Fig. 3E-F).

FOXM1 is a member of the forkhead (FOX) transcription factor family, which has ~ 50 members unified by a conserved DBD but consists of proteins of diverse function [45]. Thus, an important characteristic of any FOXM1 inhibitor is its selectivity within the FOX protein family. To test this in the context of NB compound treatment of HGSOC cells, we used western blotting to determine the expression of three additional FOX family members: FOXA1, FOXK2, and FOXO3a. FOXA1 and FOXK2 function as oncoproteins in ovarian cancer [46‐48], while, in contrast, FOXO3a is a tumor suppressor [49]. Remarkably, NB compound treatment suppressed FOXA1 (significantly in OVCAR4, trend apparent in CAOV3) and FOXK2, with no significant effect on FOXO3a (Fig. 4). It should be noted that NB compound suppression of FOXA1 and FOXK2 was less consistent and occurred at higher drug concentrations as compared to their effect on FOXM1 (Figs. 2 and 4). Nevertheless, the data indicate, for the first time, that NB compounds can suppress other oncogenic FOX family members.

The observation that NB compounds suppressed FOXM1 at both the protein and mRNA levels (Figs. 2 and 3E-F) led us to further investigate the mechanism of action (MOA) of NB compounds in HGSOC cells. We hypothesized that, following treatment with NB compounds, FOXM1 protein suppression would precede FOXM1 mRNA suppression. To test this, we conducted a kinetic study in which we measured FOXM1 mRNA and protein at time points up to 24 h in CAOV3 cells. As hypothesized, FOXM1 protein was suppressed prior to FOXM1 mRNA and was apparent as early as 3 h post-treatment (Fig. 5). These data are in agreement with studies showing that NB compounds target FOXM1 protein in breast cancer cells [29]. Consistently, co-treatment of NB compounds with the proteasome inhibitor MG132 rescued FOXM1 protein suppression in HGSOC cell lines (Fig. 6).

An important PD parameter is the length of time of a drug’s effect after drug removal. We sought to determine whether NB compound treatment has a sustained effect on FOXM1 in HGSOC cells by conducting drug washout studies [50]. We treated CAOV3 cells with NB compounds using three different regimens: 1) no drug washout (i.e., treatment followed by continuous drug incubation), 2) drug washout after 6 h of incubation, or 3) drug washout after 3 h of incubation (Fig. 7A). Notably, the suppressive effects of NB compounds on FOXM1 were largely maintained for 72 h after drug washout (Fig. 7B-D). To further assess the durability of NB compound effects on HGSOC cells, we determined NB compound IC50 values in CAOV3 cells treated with the three different treatment regimens (Fig. 7A). We observed a 2–threefold decrease of NB compound potency with drug washout, supporting that the effects of NB compounds are largely sustained in HGSOC cells for up to 72 h after drug removal (Fig. 7E-F). Additionally, we confirmed the sustained effect of NB compounds on FOXM1 suppression and cell viability using OVCAR4 cells (Fig S2).

To investigate the cell physiological effects of NB compounds in HGSOC cells over time, we used the Incucyte platform to perform simultaneous quantitative kinetic analysis of live cells (i.e., cell proliferation) and dead cells (i.e., cell death) following NB compound treatment of CAOV3 cells. We selected three drug concentrations for study (0.1, 0.5, and 1.0 μM) based on the drug potency observed in CyQuant cell viability assays (Table 3). CAOV3 cells treated with NB-73 showed a dose-dependent suppression of cell proliferation, which began within 2–4 h post-treatment in cells treated with 0.5 μM or 1.0 μM NB-73 (Fig. 8A). Notably, cells treated with 1.0 μM NB-73, but not the lower drug concentrations, exhibited a marked increase in cell death by ~ 20 h post-treatment (Fig. 8B). Similar results were obtained for NB-115 treatment (Fig. 8C-D). Surprisingly, in this assay, 0.1 μM NB-115 showed increased proliferation compared to the DMSO control (Fig. 8C). This might reflect hormesis, an adaptive response of cells to moderate stress, which is widely described in pharmacology [51]. In summary, at doses below their IC50 value, NB compounds appear to mainly inhibit HGSOC cell proliferation, while at doses above their IC50 value, the major effect is cell death. This finding agrees with observations in breast cancer cells, in which cell death was the principal outcome of NB compound treatment [29].

We next investigated the effect of NB compound treatment on HGSOC cell cycle profiles using flow cytometry analysis of propidium iodide-stained cells. We performed the analyses 24 h after treatment with NB-73 or NB-115, using both CAOV3 and OVCAR4 cells. As shown in Fig S3A, in CAOV3 cells treatment with NB compounds did not lead to significant changes in cell cycle distribution, although a trend was apparent towards reduction of cells in S phases with small increases in G1 and G2 cells. In contrast, OVCAR4 showed significant increases in G1 populations along with some decreases in S phase cells (Fig S3B). However, the modest nature of the observed changes in both cell lines suggests that altered cell cycle is not the major cellular outcome of NB compound treatment in HGSOC cells.

To investigate the mechanism by which NB compounds cause HGSOC cell death, we measured the activity of caspases 3 and 7, which are activated during apoptosis [52]. In both CAOV3 and OVCAR4 cells treated with NB-73 and NB-115, but not in FT282-C11 cells, there was a robust increase in caspase 3/7 activity, supporting an apoptotic mechanism (Fig. 9A). In agreement, NB compounds caused the induction of cleaved PARP (cl-PARP), another biomarker of apoptosis [53] (Fig. 9B). Most notably, co-treatment of CAOV3 cells with NB compounds and the pan-caspase inhibitor Q-VD-OPh [54] inhibited caspase 3/7 activity (as expected) and rescued the cells from drug-induced cytotoxicity (Fig. 9C and D). Together, these data implicate apoptosis as a major effect of NB compounds on HGSOC cells.

As NB compound treatment promotes apoptosis and apoptosis can lead to the degradation of cell signaling proteins [55, 56], we investigated whether FOXM1 suppression is a non-specific consequence of apoptosis in NB compound treated cells. For this, we measured FOXM1, FOXM1-P, and FOXM1 target expression after co-treatment of CAOV3 cells with NB compounds and the pan-caspase inhibitor Q-VD-OpH. Notably, FOXM1 and FOXM1-P remained suppressed after co-treatment (Fig. 10A and B). Furthermore, canonical FOXM1 targets exhibited the same behavior (Fig. 10C and D), and similar data were obtained for OVCAR4 cells (data not shown). Thus, the data support that FOXM1 pathway suppression is a primary outcome of NB compound treatment and is not an indirect consequence of apoptosis.

Two-dimensional (2D) anchorage-dependent colony formation is considered an in vitro assessment of self-renewing cells [57]. Based on data linking FOXM1 to tumor stem cell phenotypes [58], we examined the impact of NB compounds on this phenotype in four HGSOC cell lines: CAOV3, OVCAR4, OVCAR5, and OVCAR8. Notably, NB compounds restricted HSGOC 2D colony formation at sub-micromolar concentrations (Fig. 11).

Three-dimensional (3D) colony formation, an in vitro measure of anchorage-independent cell growth, is a classical measure of cellular transformation [59]. To test the effect of NB compounds on this phenotype, we utilized OVCAR5 and OVCAR8 HGSOC cells, which exhibited strong 3D growth compared to other HGSOC cell lines examined (data not shown). Similar to the outcome of the 2D colony formation assays, treatment of HGSOC cells with NB compounds inhibited 3D colony formation at sub-micromolar concentrations (Fig. 12).

To investigate the potential use of the NB compounds in combination with important HGSOC chemotherapies, we investigated combination treatments of NB-73 or NB-115 with carboplatin or the PARP inhibitor olaparib. Combinations with these agents are supported by the function of FOXM1 in stimulating the expression of DNA repair genes [7], as well as studies using genetic depletions of FOXM1 [60]. Notably, except for the carboplatin + NB-115 combination CAOV3 cells, carboplatin synergized with both NB compounds in both cell lines (Table 5). In contrast, the combination of NB compounds with olaparib was not synergistic (Table S1). To further examine the carboplatin combination, we used two different treatment regimens, the first in which NB compounds and carboplatin were administered to cells simultaneously, the second in which the NB compound was used first, and carboplatin was administered 24 h later. In both cases, we measured cell viability by CyQuant assay 72 h after the first treatment. Although the data were similar in both cases, the sequential regimen resulted in slightly greater synergy (Table 5).