Combined inhibition of NAD synthesis and C-terminal binding protein cooperatively induces cell death and inhibits growth of high grade serous ovarian carcinoma

Abstract

The transcriptional scaffolds C-terminal Binding Proteins (CtBP) 1 and 2 are overexpressed and act as oncogenic dependencies in multiple cancers but importantly encode a chemically targetable dehydrogenase domain. CtBP promotes survival of high grade serous ovarian carcinoma (HGSOC) cells by repressing expression of Death Receptors (DR) 4 and 5, which activate caspase 8-dependent apoptosis. We have previously developed a series of substrate competitive CtBP dehydrogenase inhibitors active in multiple cell and preclinical solid tumor models. In the current study, we validated CtBP1 and 2 overexpression in a longitudinal series of primary and metastatic/recurrent HGSOC cases. Our lead CtBP dehydrogenase inhibitor, JW-98, induced apoptosis and exhibited variable single agent IC₅₀ values in HGSOC cell lines. Importantly, depletion of nicotinamide adenine dinucleotide (NAD) species using the NAD synthesis inhibitor GMX1778 strikingly sensitized HGSOC cells to JW-98 treatment. Mechanistically, the JW-98/GMX1778 combination effectively abrogated CtBP dimerization that requires stoichiometric levels of intracellular NAD and is required for CtBP’s oncogenic transcriptional activities. Highlighting translational potential in late-stage HGSOC, combined JW-98/GMX1778 treatment of platinum-resistant OVCAR3 HGSOC mouse xenografts abrogated tumor growth without observable toxicity. Combined inhibition of CtBP and NAD synthesis represents a novel therapeutic strategy that could improve outcomes in chemoresistant HGSOC.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-30990-7.

Introduction

Epithelial Ovarian Cancer (EOC) afflicts nearly 20,000 women each year and accounts for more than 12,000 cancer deaths among US women annually¹. The most common EOC subtype, High Grade Serous Ovarian Carcinoma (HGSOC), is an especially lethal cancer due to a lack of early detection strategies resulting in diagnosis at an advanced stage, frequent development of chemoresistance to standard of care therapies, and lack of targetable driver oncogenes². Thus, novel therapeutic strategies for advanced/refractory disease have focused on targetable vulnerabilities and dependencies within oncogenic pathways that distinguish HGSOC cells from normal cells. We identified C-terminal binding protein (CtBP) family transcription factors, of which CtBP2 was already known to be overexpressed in the majority of EOC tumors³, as oncogenic drivers that promote survival of HGSOC cells by repressing expression of the proapoptotic Death Receptors 4/5⁴. Functionally, CtBP1 and 2 are paralogous dehydrogenases and transcriptional scaffolds/coregulators, and their overexpression in many solid tumors uniformly correlates with worse prognosis⁵. CtBP and its interactome contribute to malignant progression by repressing proapoptotic (DR4/5, Bik, Bax)⁵ and tumor-suppressor (E-cadherin, PTEN)⁵ gene expression, as well as transcriptionally activating oncogenes (Tiam1, c-Myc)^5,6. Furthermore, in both pancreatic and colon cancer mouse models, CtBP drives tumor progression and metastasis by promoting cancer stem cell activity^6,7. Additionally, a recently published study elucidated a non-transcriptional role of CtBP2 in promoting breast cancer metastasis via direct interaction with the anti-apoptotic protein Bcl-X_L⁸.

We have developed a library of CtBP dehydrogenase substrate competitive inhibitors (CtBPi) centered around the scaffold of the 1st generation CtBP inhibitor, hydroxyimino-3-phenylpropanoic acid (HIPP), that are active across a spectrum of solid tumor types, including colon and pancreatic cell lines and mouse models⁵. Our 2nd generation CtBPi, 4-chloro-HIPP (4-Cl-HIPP), phenocopied the hypomorphic effects of Ctbp2 allelic deletion in Apc min mice, attenuating intestinal polyposis and extending survival⁷. As nicotinamide adenine dinucleotide hydride (NADH) is embedded in CtBP’s ternary structure, we have further hypothesized that limiting the cellular concentration of total NAD species (NAD + /NADH) might enhance efficacy of CtBP inhibitors. Indeed, GMX1778, an inhibitor of the NAD synthetic enzyme nicotinamide phosphoribosyl transferase (NAMPT), synergized with 4-Cl-HIPP to kill pancreatic cancer cells in culture and strongly limit growth of pancreatic cancer xenograft tumors⁹.

In the current work, we demonstrate the cytotoxic efficacy in HGSOC cells and tumors of a novel 3rd generation CtBP inhibitor, JW-98, in which the carboxyl moiety of 4-Cl- HIPP has been esterified to improve cell penetration. Furthermore, limiting cellular NAD synthesis with NAMPT inhibitor GMX1778 significantly enhanced sensitivity of HGSOC cells to cell death induced by CtBP inhibitor⁹. Upon mechanistic investigation, we found that combined GMX1778/JW-98 treatment elicited stoichiometric disruption of transcriptionally active CtBP2 dimers, demonstrating on-target efficacy of this combination. When tested for therapeutic utility using a platinum-resistant HGSOC xenograft mouse model, the CtBP/NAMPT inhibitor combination abrogated tumor growth, demonstrating potential clinical utility in chemoresistant HGSOC. Overall, our new findings suggest combining NAD depleting agents with CtBP inhibitors could be an effective strategy to safely and effectively target late-stage refractory HGSOC.

Materials and methods

Cell lines and reagents

FT-282, OVCAR3, OVCAR4, OVCAR8, OVCA429, and Kuramochi HGSOC cells were obtained from the University of Pennsylvania Ovarian Cancer Research Center BioTrust^4,10 and were grown in DMEM supplemented with 10% Fetal Bovine Serum (Thermo Fisher Scientific, Cat No: A5670701) and 1% penicillin/ streptomycin (Thermo Fisher Scientific, Cat No: 15140122). All cell lines were routinely tested for mycoplasma (Abcam, Cat No: ab289834).

JW-98 synthesis

A Horner-Wadsworth-Emmons (HWE) reagent was synthesized (Fig. S1A) by first coupling dimethyl phosphite (1 eq) and ethyl glyoxylate (1 eq) with triethylamine (TEA, 0.2 eq) in dichloromethane (DCM) at -78 °C for 1 h to afford ethyl 2-(dimethoxyphosphoryl)-2-hydroxyacetate (Fig. S1A, 1.1). This was then TBS protected with tert-butyldimethylsilyl chloride (2 eq), imidazole (3 eq), and DMAP (0.15 eq) in DCM at room temperature overnight to yield the final HWE reagent, ethyl 2-((tert-butyldimethylsilyl)oxy)-2(dimethoxyphosphoryl) acetate (Fig. S1A, 1.2). 4-chlorobenzaldehyde scaffold was then subjected to an HWE reaction in the presence of LiHMDS (1.1 eq), where the aldehyde was used in excess (1.1 eq) relative to the HWE reagent (1 eq). The HWE reaction was carried out with LiHMDS as a base under reflux overnight in tetrahydrofuran (THF) following (Fig. S1B). After purification, the resulting silyl enol ether (2.1) underwent a two-step, one-pot reaction in which the silyl enol ether is TBS de-protected using triethylamine trihydroflouride (1.7 eq), and the resulting enol is converted to an oxime with hydroxylamine hydrochloride (1.7 eq) to yield the final oxime product in room temperature conditions overnight in two parts chloroform, and one part ethanol (Fig. S1B, 2.2). Average purity assessed by LC–MS was 95%.

Cell viability assays

3 X 10³ OVCAR3, OVCAR4, OVCAR8, OVCA429, or Kuramochi cells/well were seeded in 96 well plates, and after overnight incubation, treated with vehicle or JW-98 for 72 h, after which cell viability was determined by crystal violet staining. Absorbance was measured at 590 nm after dissolving the crystal violet stain in 10% glacial acetic acid and IC₅₀ values were calculated using GraphPad Prism version10.4.2 (Dotmatics).

Drug combination assays

2 X 10⁴ OVCAR3 cells/well were seeded in 6 well plates. After an overnight incubation, cells were pretreated with vehicle or NAMPT inhibitor GMX1778 for 24 h to deplete NAD levels, followed by treatment with vehicle or JW-98 for 6 d. Cell viability was assessed by crystal violet staining as described above.

Cross-linking assay

2.5 X 10⁵ OVCAR3 cells were treated with vehicle, GMX1778, JW-98, 4-Cl-HIPP or the combinations of JW-98 or 4-Cl-HIPP with GMX1778 for 24 h, followed by in vivo crosslinking with disuccimidylglutarate (DSG) and immunoblotting of crosslinked cell lysates for CtBP2 as previously described⁹.

Annexin V/Propidium Iodide (PI) staining

2 X 10⁶ OVCAR3 cells were seeded in 15 cm dishes, and after an overnight incubation, cells were treated with vehicle or indicated JW-98 drug concentrations for 5 d. Following drug treatment, attached and floating cells were collected and counted using trypan blue exclusion assay, and 1 X 10⁶ treated cells were stained using Alexa Fluor 488 Annexin V/Dead Cell Apoptosis Kit as per manufacturer instructions (Thermo Fisher Scientific, Cat No: V13242). AnnexinV/PI-stained cells were analyzed using BD X Fortessa 20 (BD Biosciences) or Agilent NovoCyte Quanteon (Agilent) flow cytometers, and data analyzed using BD FACSDiva Software (BD Biosciences) or NovoExpress (Agilent). The percentage of apoptotic cells (Annexin V/PI positive cells) was quantified after recording 20,000 events.

Western blotting

Approximately 3 X 10⁵ cells (OVCAR 3, 4 and OVCA429) were seeded in 6 cm cell culture dishes. After an overnight incubation, cells were treated with increasing concentrations of JW-98 (0, 50, 100 and 200 µM) for 72 h, followed by scraping into the media and centrifugation at 8000 rpm for 5 min at 4 °C to pellet both live and floating dead cells. Pelleted cells were then lysed using RIPA Lysis and Extraction Buffer (Thermo Fisher Scientific Cat No: 89900) for 30 min on ice and lysates cleared by centrifugation at 15,000 rpm for 20 min at 4 °C. Next, lysate protein concentrations were determined (Pierce BCA Protein Assay Kit, Thermo Fisher Scientific, Cat No: 23227) before mixing equal amounts of lysate protein with 4X Laemmli buffer containing 10% beta mercaptoethanol and boiling at 95 °C for 5 min. Lysates were separated on 4–15% Mini-PROTEAN TGX Precast Protein Gels (Biorad, Cat No: 4561083) using 1X MOPS-SDS running buffer and transferred to 0.2 µM nitrocellulose (Thermo Fisher Scientific, Cat No: 88024) for immunoblotting. After transfer, the nitrocellulose membranes were cut for blotting to sizes based on the approximate molecular weights of the proteins of interest. Nitrocellulose membranes were blocked using 5% BSA in 20 mM tris pH 7.4, 150 mM NaCl, 0.05% Triton X-100 (TBST) for 1 h at room temperature and then incubated with the following primary antibodies: Vinculin (Cell Signaling Technology, 1:1000 Cat No: 13901), Cleaved-PARP1 (Cell Signaling Technology, 1:1000 Cat No: 9541), Caspase-8 (Cell Signaling Technology, 1:500 Cat No: 9746), Cleaved Caspase-3 (Cell Signaling Technology, 1:500 Cat No: 9661). After overnight incubation in primary antibody, membranes were washed in TBST 3X for 5 min each. The membranes were then incubated in HRP-conjugated secondary antibodies (goat anti-mouse IgG (H + L)-HRP Conjugate; Bio-Rad, Cat No: 1706516; goat anti-rabbit IgG (H + L)-HRP Conjugate, Bio-Rad, Cat No: 1706515; both diluted 1:10,000) for 1 h at room temperature. After washing 3X for 5 min each in TBST, ECL reagent (SuperSignal West Atto Ultimate Sensitivity Substrate, Cat No: A38554) was added to the membrane and chemiluminescent signal detected using a Bio-Rad Chemidoc MP Imaging system. Densitometry was performed to quantify the signal intensity using ImageJ software (Image J).

NAD+ /NADH measurement

3 X 10³ OVCAR3 or OVCA429 cells/well were seeded in a 96 well plate. After overnight attachment of cells, the cells were treated with vehicle, GMX1778, or combination of GMX1778 and nicotinic acid (NA) for 48 h, following which total combined NAD+ /NADH levels were assessed using Promega-glo NAD/NADH measurement kit (Promega, Cat No: G9071) as per manufacturer instructions.

Xenograft study

2 X 10⁶ OVCAR3 cells in PBS mixed with Matrigel basement membrane matrix (Corning, Cat No: CLS354234) in a 1:1 ratio were subcutaneously injected into the right flank of 2-month-old NOD-scidIL2Rg^null (NSG) female mice (JAX). 10 days after tumor cell injection, mice were randomly assigned to each of 4 treatment groups and received by oral gavage 3x/week for 2 weeks: vehicle (10% DMSO +  90% corn oil; N = 5), JW-98 (100 mg/Kg; N = 5), GMX1778 (30 mg/Kg; N = 5), or JW-98/GMX1778 (doses as per single agents; N = 4) combination. Tumor volumes were measured weekly by calipers and calculated using the formula 0.5 X Length X Width². Mice were euthanized at the end of the study by exposure to 5% isoflurane as delivered by a calibrated vaporizer, followed by cervical dislocation. Statistical analysis was performed using paired student’s t-test and GraphPad Prism software. Animal studies were carried out in accordance with relevant guidelines and regulations and approved by the USC Institutional Animal Care and Use and Committee. In addition, animal studies were conducted and are reported in accordance with ARRIVE guidelines (https://arriveguidelines.org).

Immunohistochemistry

A tissue microarray of formalin-fixed paraffin-embedded (FFPE) tissue sections from a longitudinal set of human HGSOC specimens¹¹ were deparaffinized and dehydrated, following which antigen unmasking was performed using Retriever 2100 (Aptum Biologics). Slides were blocked in 5% Goat serum (Thermo Fisher Scientific, Cat No: 31872) for 1 h at room temperature and incubated with primary antibodies to CtBP1 (mouse anti-CtBP1, 1:25; BD Biosciences, Cat No: 612042) or CtBP2 (mouse anti-CtBP2, 1:25; BD Biosciences, Cat No: 612044) overnight at 4 °C. Following three 5 min washes in 1X PBST, the slides were incubated in secondary antibody (Goat Anti-Mouse IgG (H + L)-HRP, 1:200; Bio-Rad, Cat No: 1706516) for 1 h at room temperature and stained using DAB substrate (DAKO Chromogen, Agilent) per manufacturer instructions. Nuclei were counter-stained using Mayer’s Hematoxylin (Electron Microscopy Sciences, Cat No: 2617303) for 3 min. Slides were then dehydrated and coverslipped as previously described¹².The intensity of CtBP1/2 nuclear staining was scored on a scale 0–3; 0 for no staining, 1 + for weak staining, 2 + for moderate staining and 3 + for strong staining⁶.

Statistical analysis

All the results presented represent an average of at least three independent experiments. All the pairwise comparisons between groups were performed using paired student t test using GraphPad Prism software. Chi-square/equivalence test was used to compare the statistical differences between the groups and test for equivalence when a categorical variable was the outcome¹³.

Results

Expression of CtBP1/2 across the HGSOC disease spectrum

Multiple studies have shown linkage of overexpression of either or both CtBP1 and CtBP2 in various solid tumors to poor outcomes⁵. A prior study analyzed CtBP2 expression in an EOC case series that included a limited number of HGSOC cases, but CtBP2 expression specifically in HGSOC was not reported³. In addition, that series only analyzed primary tumors, and not tumors at recurrence or metastasis, when novel agents such as CtBP inhibitors might be deployed. To assess whether expression of CtBP1/2 was maintained across the continuum of HGSOC from primary resection through recurrence/progression, we analyzed CtBP1/2 expression by immunohistochemistry (IHC) in a longitudinal patient-matched series of 42 HGSOC cases (21 unique cases each for CtBP1 or CtBP2). This unique collection includes primary tumors and concurrent metastases obtained at primary debulking surgery, recurrent tumors obtained at second-look surgeries after chemotherapy (usually platinum/taxane combination), as well as normal fallopian tube, representing the normal tissue from which HGSOC arises (representative images shown in Fig. 1A, B, C)¹¹. We then scored IHC staining intensities for CtBP1/2 in each case using a 0–3 + scale and aggregated the data as the ratio of positively (1+ -3+) staining specimens for each specimen category (Fig. 1D). Notably, CtBP1 was uniformly expressed across all categories (80–90% positively staining), with no statistically significant differences in the ratio of positive specimens between primary and recurrent/metastatic categories (Fig. 1B, D). Moreover, applying a statistical equivalence test revealed that within a biologically realistic variation of +/− 20% in the proportion of specimens staining positively for CtBP1, the expression of CtBP1 was statistically equivalent across all categories (Fig. 1D). The ratio of CtBP2-positive specimens was lower in primary tumors (~ 55%) but demonstrated a trend toward more universal positive expression (~ 75%) in recurrent/metastatic tumors that also demonstrated statistical non-equivalence (Fig. 1D). Thus, HGSOC tumors might initially select for expression of CtBP1 that does not increase with recurrence or metastasis, whereas the frequency of CtBP2 expression may increase with recurrence/metastasis, though confirmation of this hypothesis will require additional specimens to attain adequate statistical power.

Fig. 1.

Expression of CtBP across the spectrum of HGSOC. (A–C) IHC using CtBP1 or CtBP2 antibodies performed on sections of resected normal fallopian tube (FT; A) or primary HGSOC (B, C). (D) Quantitation of percent of patients positively stained for CtBP1/2 using IHC from a series of HGSOC cases representing intra-patient primary, metastatic, and recurrent tumor. Scale bar = 50 µM. “Positive” indicates 1+ to 3+ staining. ns = p > 0.05; *statistically equivalent within allowed variation of + /–20%; ^$ not statistically equivalent within allowed variation of + /–20%. Error bars indicate +/– 1.0 SD.

Development of a novel 3rd generation small molecule CtBP inhibitor

We have iteratively improved the potency of small molecule CtBPi’s¹⁴. The 1st generation CtBPi, HIPP, was rationally designed from substrate/catalytic domain interactions observed in the co-crystal structure of CtBP with its substrate 2-keto-4-methylthio-2-oxo butyric acid (MTOB) and based on the dehydrogenase enzymatic mechanism¹⁴. We then identified the potent 2^nd generation CtBPi 4-Cl-HIPP (enzymatic IC₅₀ = 173 nM) based on a structure–activity relationship study of substituents on the HIPP phenyl ring⁷. In developing the 3rd generation of CtBPi’s, we sought to improve cell-permeability and in vivo activity through esterification of the 2nd generation CtBPi 4-Cl-HIPP to form our lead CtBPi JW-98 (Fig. 2). However, it should be noted that active intracellular cytoplasmic esterases readily catalyze hydrolysis of ester moieties such as incorporated into JW-98, facilitating the intracellular conversion of the prodrug to the active free acid, 4-Cl-HIPP¹⁵.

Fig. 2.

Chemical structures of 1st–3rd generation HIPP-class CtBP substrate-competitive dehydrogenase inhibitors. JW-98 is the ethyl ester of 4-Cl-HIPP.

JW-98 induces cell death by apoptosis in HGSOC cells

Our previous work in HGSOC cells demonstrated repression of pro-apoptotic genes DR4/5 by CtBP in HGSOC cells⁴. Furthermore, transient or stable knockdown of CtBP1/2 led to cell death by caspase 8-dependent apoptosis, unveiling the CtBP dependency of HGSOC cell survival. To determine if a CtBPi could phenocopy genetic CtBP loss in HGSOC, we assessed the cytotoxic efficacy of JW-98 in a panel of genetically validated HGSOC cell lines (OVCAR3, OVCAR4, OVCAR8, OVCA429 and Kuramochi), of which OVCAR3, OVCAR4, and OVCAR8 were derived from platinum-exposed/resistant tumors, and OVCAR8 also exhibits in vitro platinum-resistance¹⁶. Upon exposure of the HGSOC cell lines to increasing JW-98 concentrations over a period of 72 h followed by assessment of cell viability, we observed measurable JW-98 IC₅₀’s in the range of ~ 50 to 150 µM in OVCAR3, OVCAR4, and Kuramochi cells, while OVCAR8 and OVCA429 cells were resistant to JW-98 (Fig. 3A). Notably, the non-transformed fallopian tube epithelial cell line hTERT-FT-282 cell¹⁷ was also resistant to JW-98 (IC₅₀ > 200 µM; Fig. 3A) consistent with cancer cell selective cytotoxic effects of JW-98.

Fig. 3.

Cytotoxicity of JW-98 in HGSOC cells. (A) The indicated cell lines were treated with vehicle or increasing doses of JW-98 for 72 h, and cell viability was assessed by crystal violet staining, with absorbance of solubilized dye recorded at 590 nm. IC₅₀ values were calculated using GraphPad Prism (N = 3). (B) OVCAR3 cells were treated with vehicle or JW-98 for 5 d and analyzed via Annexin V-488/Propidium Iodide (PI) staining. The percentage of Annexin V positive apoptotic cells was quantified using flow cytometry. (C) OVCAR3 cells treated with JW-98 for 72 h were lysed and immunoblotted for indicated apoptotic markers. The original uncropped blots are shown in Fig. S6. Significance was established by comparison between groups using paired student t-test. **p < 0.01. N = 3. Error bars indicate + / − 1.0 SD.

Thus, despite esterification of 4-Cl-HIPP to improve cell penetration, the IC₅₀ values of JW-98 in HGSOC cells as a single agent were higher than optimal in HGSOC cells exhibiting measurable IC₅₀’s, suggesting that further therapeutic development of JW-98 in HGSOC might require additional strategies to enhance efficacy at lower doses. In addition, to understand whether CtBP abundance predicts the sensitivity of HGSOC cell lines to JW-98 treatment, we immunoblotted CtBP1/2 protein levels in each cell line (Fig. S2A). However, upon performing linear regression analysis of normalized CtBP1/2 protein abundance to JW-98 IC_50, we found no statistically significant correlations (Fig. S2B).

Mechanism of cell death induced by JW-98

Next, we assessed the mechanism of cell death induced by JW-98 in cells that did exhibit measurable single agent sensitivity. Our prior work in ovarian and other solid tumor cells (pancreatic and colon) demonstrated that chemical inhibition or genomic depletion of CtBP1/2 leads to apoptosis⁴. Hence, we analyzed JW-98 treated OVCAR3 cells with Annexin V and propidium iodide staining followed by flow cytometry to assess the role of apoptosis in JW-98 cytotoxicity. Notably, JW-98 treatment caused a significant increase in apoptotic cells (~ 30%) in a dose dependent manner compared to vehicle treatment (Fig. 3B). Furthermore, we have previously shown that CtBP1/2 depletion in HGSOC cells leads to apoptosis specifically via the caspase 8-dependent extrinsic apoptotic pathway (EAP)⁴. To investigate this possibility in JW-98 treated cells, we immunoblotted JW-98 treated OVCAR3 cell lysates for EAP markers¹⁸ and observed a dose-dependent decrease in levels of procaspase 8 along with increased levels of cleaved caspases 3, 8, and PARP1 (Fig. 3C). Furthermore, JW-98 treatment of OVCAR4 and 8 cells also revealed increased levels of cleaved PARP1 (Fig. S3) consistent with a universal apoptotic mechanism of cytotoxic action for JW-98 in HGSOC cells. Thus, CtBP chemical inhibition induces EAP-mediated apoptosis in HGSOC cells in a manner similar to CtBP1/2 depletion via RNA interference⁴.

Combined inhibition of CtBP and NAD synthesis disrupts oligomeric CtBP complexes

We and others have shown that CtBP activity is dependent on the availability of cellular NAD for oligomerization, which drives formation of transcriptional complexes¹⁹. The active CtBP oligomer is found in tetrameric form as a dimer of dimers⁵. In cancer cells, NAD is synthesized using either the Salvage or Preiss-Handler pathways, which are dependent on the NAMPT and Nicotinate phosphoribosyl transferase (NAPRT) enzymes, respectively, though only NAMPT chemical inhibitors are currently commercially available. Moreover, CtBP oligomerization status in a cell can serve as a biomarker for activated CtBP-dependent oncogenic transcriptional programs²⁰. Previously, we demonstrated efficient disruption of cellular CtBP dimerization by the 2^nd generation CtBPi 4-Cl-HIPP in pancreatic cancer cells in the setting of NAD depletion induced by the NAD synthesis inhibitor GMX1778, a potent inhibitor of NAMPT⁹.

To establish if JW-98 similarly disrupts CtBP dimerization in HGSOC, we first established whether GMX1778 treatment could deplete NAD levels in OVCAR3 cells. Indeed, treatment of OVCAR3 cells with a non-toxic sub-IC₅₀ dose of GMX1778 (Fig. S4A) strongly depleted total intracellular NAD levels (NAD + /NADH species), which was rescued by addition of the NAD precursor nicotinic acid, demonstrating that NAD depletion by GMX1778 was due to on-target activity (Fig. 4A). We next assayed CtBP dimerization in OVCAR3 cells treated with vehicle, JW-98, GMX1778, or the JW-98/GMX1778 combination, by treating drug-exposed live cells with DSG crosslinker followed by immunoblotting of lysates for CtBP2 (Figs. 4B, S4B). Strikingly, neither drug alone reduced dimerization, but the combination potently reduced and even eliminated dimerization in a dose dependent manner for JW-98 (Fig. 4C).

Fig. 4.

NAMPT inhibitor/JW-98 combination inhibits CtBP oligomerization and induces cytotoxicity in HGSOC cell lines. (A) OVCAR3 or OVCA429 cells were treated for 48 h with vehicle, GMX1778 (1 nM), or GMX1778 (1 nM) + nicotinic acid (NA) (10 µM), and NAD + /NADH levels were measured in cell lysates and normalized to levels in vehicle-only treated cells. (B) OVCAR3 cells were treated with vehicle (1st 3 lanes) or GMX1778 (3 nM) for 24 h, followed by addition of vehicle (0) or indicated concentrations of JW-98 or 4-Cl-HIPP (250 µM; Pos.) for 48 h prior to crosslinking live cells with DSG. Un-X indicates an un-crosslinked control. Cross-linked cell lysates were immunoblotted using CtBP2 antibody. Results of a representative experiment from three independent experiments are shown. The original uncropped blot is shown in Fig. S7. (C) Densitometric analysis of B was performed using ImageJ software and percentage of CtBP2 dimer abundance relative to that in vehicle-only treated cells was determined. (D) OVCAR3 cells were pretreated with GMX1778 (1 nM) for 24 h, followed by addition of increasing concentrations of JW-98 for 6 d. Cell viability was assessed via crystal violet staining and absorbance of solubilized dye was recorded at 590 nm. N = 3. All pairwise comparisons were made using student’s t-test. *p < 0.05; **p < 0.01; ***p < 0.001. Error bars indicate + /- 1.0 SD.

NAD depletion sensitizes HGSOC cells to CtBP inhibitor induced cell death

To determine if disruption of CtBP oligomers by the JW-98/GMX1778 combination translated into cellular cytotoxic effects, we pre-treated OVCAR3 HGSOC cells with a sub-IC₅₀ dose of GMX1778 or vehicle for 24 h, followed by treatment with vehicle or increasing doses of JW-98 for 6 days, followed by assessment of cell viability (Fig. 4D). Strikingly, GMX1778, which had little effect by itself at 1 nM, strongly sensitized the cytotoxic effect of JW-98 at JW-98 concentrations well below its single agent IC₅₀ (Fig. 4D). Indeed, in the presence of a non-cytotoxic concentration of GMX1778, there was a ~ threefold loss of viability (75% to 25%) at a JW-98 dose of 25 µM, which is well below the single agent IC₅₀ (Fig. 4D). Strikingly, pre-treatment of JW-98 resistant OVCAR8 and OVCA429 cells with GMX1778 also strongly accentuated JW-98 cytotoxicity with 95% and 50% loss of viability, respectively, at a JW-98 dose of 100 µM, which alone, was non-toxic in these cells (Fig. S3A). These results suggest the NAD depletion strongly potentiates JW-98 induced cytotoxicity in HGSOC cells.

NAMPT inhibitor GMX1778 potentiates anti-cancer effect of JW-98 in vivo

To establish the in vivo efficacy of the CtBP/NAD synthesis dual inhibition strategy, we subcutaneously xenografted immunodeficient NSG mice with OVCAR3 cells and treated the mice 3X/week for 2 weeks by oral gavage with vehicle, GMX1778 (30 mg/kg), JW-98 (100 mg/kg), or the combination of JW-98 and GMX1778, and tumor volume was measured on Days 0, 7, and 14 (Fig. 5A). OVCAR3 cells were chosen, in part, due to their derivation from a clinically platinum-resistant tumor¹⁶, thus mirroring the clinical challenge of treating late-stage chemoresistant HGSOC. The mice suffered no visible ill effects over the 14-day course of treatment, as their weights remained stable, and gross examination at necropsy revealed no morphologic abnormalities in vital organs (heart, lungs, liver, digestive tract; Fig. S5). By Day 14, GMX1778 treatment alone only modestly reduced tumor volume ~ 35%, as did JW-98 alone, though these effects did not achieve statistical significance (Fig. 5B). However, combination therapy with GMX1778/JW-98 abrogated any growth of tumor from baseline over the 14-day experiment, and relative to mean tumor size in vehicle-treated mice, caused a significant 70% reduction in mean tumor volume at Day 14 (p < 0.01; Fig. 5B). The strikingly efficacious, yet, safe, activity of the JW-98/GMX1778 combination in OVCAR3-xenografted mice is consistent with the highly cooperative effects of this drug combination observed in cultured HGSOC cells (Fig. 4), pointing to a novel therapeutic option for the enormous unmet need for novel therapies in refractory HGSOC.

Fig. 5.

Combined CtBP/NAD synthesis inhibition abrogates growth of HGSOC xenografts. (A) OVCAR3 cells were injected subcutaneously into the right flank of NSG mice and tumors were allowed to grow for 10 days, after which animals were treated by oral gavage 3x/week for 2 weeks (indicated by arrows) with vehicle (corn oil, N = 5), GMX1778 (30 mg/kg, N = 5), JW-98 (100 mg/kg, N = 5) or combination of GMX1778 and JW-98 (N = 4). (A) Tumor volumes were measured weekly. (B) % change in mean treatment group tumor volume normalized to vehicle-treated group tumor volume. Paired student t-test was used for pairwise group comparisons. **p < 0.01. Error bars indicate + / − 1.0 SD.

Discussion

The efficacy of HGSOC treatment is often limited by chemoresistance. In addition, HGSOC is a complex disease driven by a combination of tumor suppressor loss and numerous non-mutated oncogenic dependencies, such as overexpression of CtBP⁴. We now show, for the first time, continued maintained expression of both CtBP paralogs in primary, recurrent, and metastatic HGSOC. While CtBP1 expression was statistically equivalent among all the categories, CtBP2 exhibited a trend toward higher expression in recurrent/metastatic disease, though a larger sample size would be needed to determine if the trend to higher expression achieved statistical significance. Nevertheless, having confirmed ubiquitous expression of CtBP1/2 in advanced disease, we tested a 3^rd generation CtBP substrate-competitive dehydrogenase inhibitor, JW-98, in a panel of genetically validated HGSOC cell lines—including cell lines derived from clinically platinum resistant tumors- OVCAR3, OVCAR4, and OVCAR 8¹⁶. Unfortunately, the single agent activity of JW-98 in HGSOC cells was disappointing, leading to consideration as to whether high stoichiometry of NAD species, which co-occupy the CtBP active site⁵, might be interfering with JW-98’s ability to inhibit CtBP functions.

The CtBP transcriptional co-regulators are reported to dimerize and/or tetramerize in the presence of NAD + or NADH to form transcriptional complexes regulating their oncogenic activities²¹. NAD is a key metabolite required for normal cellular growth and proliferation; however, tumor cells are in greater need of NAD than normal cells to leverage the ever-growing demand for cellular ATP needed for proliferation and growth²². We strategized that limiting NAD synthesis in HGSOC cells using the NAMPT inhibitor GMX1778 would cooperatively enhance JW-98’s CtBP inhibitory activity and lead to efficient disruption of CtBP oligomeric complexes that drive oncogenic transcriptional programs (such as repression of DR4/5⁴), ultimately causing cell death. Indeed, our prior work in pancreatic cancer has shown that limiting NAD synthesis improved the cytotoxic efficacy of an earlier generation CtBPi, by promoting disruption of CtBP oligomeric complexes⁹. When we limited NAD synthesis in HGSOC cells using GMX1778, we indeed observed that JW-98 efficiently disrupted CtBP dimerization as assessed by in vivo crosslinking in living cells. Importantly, NAD depletion was also synthetically lethal with JW-98, leading to enhanced cell death in vitro, both in clinically platinum-resistant HGSOC cell lines (OVCAR3, OVCAR4, OVCAR8) and cell lines resistant to single agent JW-98 activity (OVCAR8, OVCA429).

Evaluation of JW-98 safety and efficacy in vivo revealed that combined treatment of mice with GMX1778 and JW-98 safely and completely abrogated HGSOC OVCAR3 xenograft tumor growth. Indeed, mouse weights remained stable throughout the experiment for all cohorts, indicating overall good tolerance to the regimen, including the combination treated mice (Fig. S5). Moreover, the GMX1778 dose used was 70% lower than that used in single agent xenograft studies²³, which helped to limit any toxicity in mice treated with the combination regimen. Though mouse internal organs remained morphologically normal, they were not analyzed histologically for toxicity in this xenograft study. However, a prior study testing the related drug 4-Cl-HIPP (the non-esterified version of JW-98) revealed no evidence of toxicity by morphologic/histologic evaluation of vital organs at similar doses (100 mg/kg 3x/week)⁶.

HGSOC is a lethal cancer in women due to the lack of effective systemic therapies in advanced, especially chemo-resistant, disease. Overall, our findings strongly support the combined inhibition of NAD synthesis and CtBP as a safe and innovative strategy to target HGSOC. The ease of performing in vivo crosslinking in tumor cells may point to development of a future predictive pharmacodynamic biomarker for CtBP efficacy. Furthermore, based on our exciting results combining CtBP and NAD synthesis inhibition in HGSOC xenografts, this strategy should be explored in additional pre-clinical HGSOC models, including patient-derived xenografts, to determine the broad applicability of this therapeutic strategy across both chemosensitive and chemoresistant HGSOC cells and tumors.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

KKC, JW, NSM, DD, IK, NK, MMD, and AEW conducted the experiments. RD, JWC, SO, JW, KCE, KKC, and SRG conceived the experiments and analyzed the results. All authors reviewed the manuscript.

Data availability

All data generated and/or analyzed during the current study will be made available from the corresponding author on reasonable request.

Declarations

Competing interests

The authors declare that they have no competing interests.