Compound C Inhibits Ovarian Cancer Progression via PI3K-AKT-mTOR-NFκB Pathway

Simple Summary Ovarian cancer is a deadly cancer due to its late diagnosis. Despite aggressive surgery and chemotherapy recurrence of a resistant aggressive disease is common. Thus, there is an unmet need to develop new therapeutics that target cancer cells and prevent recurrence and resistance. In the present study, we used multiple approaches to report and validate a novel therapeutic compound, compound C, that targets cancer cells and renders them more sensitive to standard of care therapy. Our study also reports novel mechanism of action of compound C and warrants its further development in the treatment of ovarian cancer patients. Abstract Epithelial Ovarian cancer (OvCa) is the leading cause of death from gynecologic malignancies in the United States, with most patients diagnosed at late stages. High-grade serous cancer (HGSC) is the most common and lethal subtype. Despite aggressive surgical debulking and chemotherapy, recurrence of chemo-resistant disease occurs in ~80% of patients. Thus, developing therapeutics that not only targets OvCa cell survival, but also target their interactions within their unique peritoneal tumor microenvironment (TME) is warranted. Herein, we report therapeutic efficacy of compound C (also known as dorsomorphin) with a novel mechanism of action in OvCa. We found that CC not only inhibited OvCa growth and invasiveness, but also blunted their reciprocal crosstalk with macrophages, and mesothelial cells. Mechanistic studies indicated that compound C exerts its effects on OvCa cells through inhibition of PI3K-AKT-NFκB pathways, whereas in macrophages and mesothelial cells, CC inhibited cancer-cell-induced canonical NFκB activation. We further validated the specificity of the PI3K-AKT-NFκB as targets of compound C by overexpression of constitutively active subunits as well as computational modeling. In addition, real-time monitoring of OvCa cellular bioenergetics revealed that compound C inhibits ATP production, mitochondrial respiration, and non-mitochondrial oxygen consumption. Importantly, compound C significantly decreased tumor burden of OvCa xenografts in nude mice and increased their sensitivity to cisplatin-treatment. Moreover, compound C re-sensitized patient-derived resistant cells to cisplatin. Together, our findings highlight compound C as a potent multi-faceted therapeutic in OvCa.


Introduction
Ovarian cancer (OvCa) is the leading cause of death from gynecologic malignancies in the United States with more than 75% of patients diagnosed at an advanced disease stage [1]. High-grade serous cancer (HGSC) is the most common pathological subtype and accounts for the highest lethality [2]. Recurrence of a chemo-resistant disease is very common due to suboptimal debulking of widespread inaccessible lesions in the peritoneal

In Vivo OvCa Cell Homing/Adhesion Assays
Confluent monolayer of ID8 cells were pre-treated with 5 µM of Compound C or DMSO for 18 h and were labelled with CellTracker™ Blue 7-amino-4-chloromethylcoumarin (CMAC) Dye (Invitrogen) for 30 min. C57BL6 mice were injected intaperitoneally with 1 × 10 6 CMAC-ID8 cells. After four hours, mice were euthanized, omenta were collected in 6 well cell plates and gently washed once with PBS (n = 4 mice per group). After which, specimens were preserved in 70% ethanol. Fluorescence imaging was immediately conducted using Olympus IX-70 at 460 nm (blue) and visualized in 5 fields/organ at 100× magnification and analyzed using Image J software.

RNA Extraction and Real-Time Quantitative Reverse Transcription Polymerase Chain Reaction (qRT-PCR)
Total RNA was isolated using RNEasy Kit (Qiagen, Germantown, MD, USA) [11,12,25]. Total RNA (1 µg) was reverse transcribed in a 20µl reaction using iScript cDNA synthesis kit (Bio-Rad). cDNA was amplified using forward and reverse primers (Table S2) with SsoAdvanced Universal SYBR Green Supermix (Bio-Rad). Reaction conditions were as follows: Polymerase activation and DNA denaturation for 30 s at 95 • C, followed by 35 cycles of 95 • C for 15 s and 60 • C for 30 s. PCR was performed in 96-well plates in CFX Connect Real-Time System (Bio-Rad). All experiments were performed in triplicates and were normalized to 18S mRNA as reference housekeeping gene.

Measurement of Mitochondrial Mass MitoTracker Staining
OvCa cells were seeded in 8-well LabTek slide chambers and were treated overnight with CC (5 µM). Cells were stained with MitoTracker Green (for mitochondrial mass) as previously described [25]. After fixing cells, slides were mounted in fluorogel containing DAPI and covered with coverslips. Images were acquired using Olympus FV1200 Confocal Imager (Tokyo, Japan). Average fluorescence intensity per cell was detected by PICO CellReporterXpress image acquisition and analysis software (Molecular Devices, San Jose, CA, USA), and measured by ImageJ Software.

In Vivo Tumor Xenografts
Luciferase-tagged SKOV3 (SKOV3-luc) cells were injected intraperitoneally (2 × 10 6 cells/100 µL PBS, using 27G syringe needles) in 6-8 weeks old female athymic nude mice (Charles River Laboratories, Wilmington, MA, USA) as earlier described [11,38]. Tumor growth was monitored bioluminescent imaging with IVIS Spectrum In Vivo Imaging System (PerkinElmer, Waltham, MA, USA). One week after tumor cell injection, mice were stratified into four groups: control/PBS, compound C (9 mg/kg/d), cisplatin (1 mg/kg/d), and a combination of CC (9 mg/kg/d) and cisplatin (1 mg/kg/d). Mice received treatment three times/ week for three weeks. Mice were weighed once a week and imaged with IVIS once every three weeks for eight weeks, after which mice were euthanized by isoflurane inhalation and cervical dislocation. Intraperitoneal tumors were dissected, weighed, and measured, and ascitic fluid was collected for further analysis [11][12][13]38].

Immunohistochemistry (IHC)
IHC was performed on formalin-fixed paraffin-embedded sections as earlier described [11,25,38]. After de-paraffinization, antigen-retrieval (by boiling in 0.01% citric acid for 15 min). Sections were incubated with the indicated primary antibodies (Table S1). After washing, sections were developed with secondary antibodies in Vectastain ABC ELITE (Vector Laboratories, Inc., Newark, CA, USA) according to manufacturer's instructions. Slides were developed by diaminobenzidine (DAB) as a chromogen and hematoxylin as the nuclear counterstain. Negative controls were included omitting the primary antibody. Slides were scanned using Olympus VS120 Automated Slide Scanner (Olympus). Digital image analysis was carried out as earlier described [11,25,38]. The frequency of positive staining was determined by the percentage of positive cells counted in whole tumor section with three tumor sections/experimental condition examined.

Docking
Coordinates for dorsomorphin were obtained from Pubchem (CID 11524144) [46] and prepared using Babel [47] and AutoDockTools [48] for conversion to a Protein Data Bank (pdb) and Protein Data Bank, Partial Charge (Q), & Atom Type (T) (pdbqt file) respectively. The relevant protein structure files were obtained from the Research Collaboratory for Structural Bioinformatics (RCSB) PDB IDs 1E7V, 4JSV, and 3GUT [49][50][51]. Non-protein atoms were removed manually, and the files prepared as pdbqt files also using AutoDock Tools [52]. The docking was performed using AutoDock Vina [48], using the default exhaustiveness. The 40 Angstrom cubic grid was centered at the relevant middle of the relevant binding pocket for each structure. For p65relA a grid of 80 Angstroms was also tested to see if an alternative binding mode could be found; one was not.

Statistical Analysis
Data were analyzed by two-tailed unpaired Student's t-test, multiple t-test, and oneand two-way analysis of variance (ANOVA) with Sidak-Holm test. Differences were deemed significant at p < 0.05. GraphPad Prism 7.0 (San Diego, CA, USA).

Compound C Inhibits OvCa Proliferation and Clonogenic Survival
To determine the effects of CC on OvCa cell malignant phenotype, we treated SKOV3, OVCA3, IGROV1, and CAOV3, and murine ID8 OvCa cell lines with increasing concentrations of CC and determined the effect on proliferation and clonogenic survival. We found that CC exerted an inhibitory effect on OvCa proliferation in the five OvCa cell lines in a time and dose dependent manner ( Figure 1A-E). Consistently, CC inhibited clonogenic survival in a dose-dependent manner in the five OvCa cell lines ( quency of positive staining was determined by the percentage of positive cells counted in whole tumor section with three tumor sections/experimental condition examined.

Docking
Coordinates for dorsomorphin were obtained from Pubchem (CID 11524144) [46] and prepared using Babel [47] and AutoDockTools [48] for conversion to a Protein Data Bank (pdb) and Protein Data Bank, Partial Charge (Q), & Atom Type (T)(pdbqt file) respectively. The relevant protein structure files were obtained from the Research Collaboratory for Structural Bioinformatics (RCSB) PDB IDs 1E7V, 4JSV, and 3GUT [49][50][51]. Nonprotein atoms were removed manually, and the files prepared as pdbqt files also using AutoDock Tools [52]. The docking was performed using AutoDock Vina [48], using the default exhaustiveness. The 40 Angstrom cubic grid was centered at the relevant middle of the relevant binding pocket for each structure. For p65relA a grid of 80 Angstroms was also tested to see if an alternative binding mode could be found; one was not.

Statistical Analysis
Data were analyzed by two-tailed unpaired Student's t-test, multiple t-test, and oneand two-way analysis of variance (ANOVA) with Sidak-Holm test. Differences were deemed significant at p < 0.05. GraphPad Prism 7.0 (San Diego, CA, USA).

Compound C Inhibits OvCa Proliferation and Clonogenic Survival
To determine the effects of CC on OvCa cell malignant phenotype, we treated SKOV3, OVCA3, IGROV1, and CAOV3, and murine ID8 OvCa cell lines with increasing concentrations of CC and determined the effect on proliferation and clonogenic survival. We found that CC exerted an inhibitory effect on OvCa proliferation in the five OvCa cell lines in a time and dose dependent manner ( Figure 1A

Compound C Inhibits OvCa Cell Migration and Invasiveness
Gene set enrichment derived from GSE60135 of SKOV3 cells treated with CC [40] showed significant inhibition of the epithelial mesenchymal transition (EMT) signature in SKOV3 cells treated with 5 µM of CC ( Figure 3A). Thus, we determined the effect of CC on migration and matrix invasiveness of OvCa cells. Consistently, we found that CC significantly inhibited OvCa cell migration and matrix invasion in all OvCa cell lines ( Figure 3B,C).

Compound C Inhibits OvCa Cell Migration and Invasiveness
Gene set enrichment derived from GSE60135 of SKOV3 cells treated with CC [40] showed significant inhibition of the epithelial mesenchymal transition (EMT) signature in SKOV3 cells treated with 5 µM of CC ( Figure 3A). Thus, we determined the effect of CC on migration and matrix invasiveness of OvCa cells. Consistently, we found that CC significantly inhibited OvCa cell migration and matrix invasion in all OvCa cell lines ( Figure  3B,C).

Compound C Inhibits PI3K-AKT-mTOR and NFκB in OvCa
Recent reports indicated PI3K-AKT-mTOR and NFκB axis are of the most amplified and hyperactivated pathways in OvCa [8,[53][54][55][56]. We have recently shown that the transcripts of the key molecules involved in these pathways are not only associated with poor patients' survival in TCGA data, but the expression of these transcripts positively correlated with each other as well, suggesting positive feedforward activation [8,[53][54][55][56]. In addition, Gene Set Enrichment Analysis (GSEA) of GSE60135 study revealed inhibition of

Compound C Inhibits PI3K-AKT-mTOR and NFκB in OvCa
Recent reports indicated PI3K-AKT-mTOR and NFκB axis are of the most amplified and hyperactivated pathways in OvCa [8,[53][54][55][56]. We have recently shown that the transcripts of the key molecules involved in these pathways are not only associated with poor patients' survival in TCGA data, but the expression of these transcripts positively correlated with each other as well, suggesting positive feedforward activation [8,[53][54][55][56]. In addition, Gene Set Enrichment Analysis (GSEA) of GSE60135 study revealed inhibition of PI3K-AKT-mTOR as well as inflammatory signatures in CC-treated SKOV3 cells ( Figure 4A,B). Given that this axis represents the central hub of oncogenic signaling linking cancer cell proliferation, invasiveness as well as metabolic reprogramming of cancer cells, we sought to determine whether CC exerts its inhibitory effect on OvCa cells through inhibition of this pathway. Compound C decreased the expression as well as the activation and phosphorylation of key regulators in PI3K-AKT-mTOR and NFkB signaling pathways in two OvCa cell lines, SKOV3 and OVCAR3 ( Figures 4C and S2). In addition, we found that CC inhibited basal and LPA-induced activation and nuclear translocation of p65RelA subunit of NFκB (Figures 4D,E and S3).  To confirm the specificity of the inhibitory effect of compound C on PI3K-AKT and NFκB, we used two independent approaches. First, we performed in silico simulation and modeling of the structure of CC and the structures of the key nodes in this pathway. We found that CC ( Figure 5A) bound to PI3K-p110α subunit at ASP 950, LYS 890, and ALA 805 ( Figure 5B). Computational modeling also revealed that CC binds to mTOR at 130nM. To confirm the specificity of the inhibitory effect of compound C on PI3K-AKT and NFκB, we used two independent approaches. First, we performed in silico simulation and modeling of the structure of CC and the structures of the key nodes in this pathway. We found that CC ( Figure 5A) bound to PI3K-p110α subunit at ASP 950, LYS 890, and ALA 805 ( Figure 5B). Computational modeling also revealed that CC binds to mTOR at 130 nM. While the binding is not at the catalytic site, it is nearby in a pocket/cleft in mTOR molecule at residues PRO 1940, PRO 1975, TYR 2144 ( Figure 5C). Moreover, in silico modeling also revealed CC binds p65RelA subunit of NFkB near the DNA binding site and three nearest interacting residues, ASP 80, ARG 84 and ASN 190 ( Figure 5D). Secondly, we stably overexpressed constitutively active myristoylated PI3K-p110α catalytic subunit, as well as AKT1 and IKKα, with pBABE as a vector control ( Figures 6A  and S4). Overexpression of myristoylated PI3Kp110α in SKOV3 cells led to modest change in total and phospho-AKT (ser473) but markedly increased phosphorylation of downstream mTOR at ser2448, IKKα as well as total and phosphorylated (ser536) p65RelA subunit of NFκB ( Figures 6A and S4). Overexpression of constitutively active AKT1 increased the expression of PI3Kp110α, activation and phosphorylation of mTOR (ser2448), IKKα as well as total and phosphorylated (ser536) p65RelA subunit of NFκB. Consistently, overexpression of IKKα increased the expression of PI3K-p110α, activation and phosphorylation of mTOR (ser2448), as well as total and phosphorylated (ser536) p65RelA subunit of NFκB but had no effect on AKT expression or activation ( Figures 6A and S4). These data further confirm the interconnected feedforward activation loop of PI3K-AKT-mTOR-IKKα-NFκB in OvCa cells. Phenotypically, overexpression of myristoylated PI3K-p110α, AKT1, and IKKα significantly increased proliferation of SKOV3 at 72-96 h for PI3K-p110α, and 48-96h for AKT1 and IKKα, respectively ( Figure 6B-D). Importantly, they mitigated the inhibitory effect of CC on cell proliferation ( Figure 6B-D). Similarly, overex- Secondly, we stably overexpressed constitutively active myristoylated PI3K-p110α catalytic subunit, as well as AKT1 and IKKα, with pBABE as a vector control ( Figures 6A and S4). Overexpression of myristoylated PI3Kp110α in SKOV3 cells led to modest change in total and phospho-AKT (ser473) but markedly increased phosphorylation of downstream mTOR at ser2448, IKKα as well as total and phosphorylated (ser536) p65RelA subunit of NFκB ( Figures 6A and S4). Overexpression of constitutively active AKT1 increased the expression of PI3Kp110α, activation and phosphorylation of mTOR (ser2448), IKKα as well as total and phosphorylated (ser536) p65RelA subunit of NFκB. Consistently, overexpression of IKKα increased the expression of PI3K-p110α, activation and phosphorylation of mTOR (ser2448), as well as total and phosphorylated (ser536) p65RelA subunit of NFκB but had no effect on AKT expression or activation ( Figures 6A and S4). These data further confirm the interconnected feedforward activation loop of PI3K-AKT-mTOR-IKKα-NFκB in OvCa cells. Phenotypically, overexpression of myristoylated PI3K-p110α, AKT1, and IKKα significantly increased proliferation of SKOV3 at 72-96 h for PI3K-p110α, and 48-96 h for AKT1 and IKKα, respectively ( Figure 6B-D). Importantly, they mitigated the inhibitory effect of CC on cell proliferation ( Figure 6B-D). Similarly, overexpression of myristoylated PI3Kp110α, AKT1, and IKKα significantly increased SKOV3 migration and matrix-invasion and mitigated the inhibitory effect of CC ( Figure 6E,F). Together, these data further support the specificity of the inhibitory effect of CC on the key nodes PI3K-AKT-NFκB axis. Values were compared to untreated cells pBABE vector controls (n = 3/experimental condition; repeated twice). * p < 0.05 compared to pBABE vector control treated with DMSO, ** p < 0.05 comparing CC-treatment to corresponding DMSO control, Student's t-test. **** p < 0.0001.

Compound C Inhibits OvCa Cells-Mesothelial Interactions In Vitro and In Vivo
The mesothelial cell monolayer is the first barrier that OvCa cells from the primary tumor encounter for peritoneal colonization and spread [12][13][14]57,58]. To investigate the effect of CC on the ability of OvCa cells to adhere to the mesothelial layer, human OvCa cells were treated with either CC (5µM) or vehicle (DMSO) for 30 min and were allowed to adhere to uncoated-, and matrigel-coated plates, or mesothelial monolayers on 96 well plates for 2hrs [12][13][14]. We found that CC significantly inhibited OvCa cells adhesion to the uncoated-, and matrigel-coated wells as well as mesothelial monolayers ( Figure 7A,B). To further verify the effect of CC on OvCa cell chemotaxis and adhesion to the mesothelial surface in vivo, fluorescent-labelled ID8 cells were injected intraperitoneally into C57B6 mice as earlier described [11]. Two hours later, mice were euthanized, omenta and mesentery dissected, fluorescent OvCa cells that homed to and adhered to mesothelial cells covering mesentery and omenta were visualized under a fluorescent microscope, and fluorescent signal quantified. We found that CC significantly inhibited the homing and adhesion of OvCa cells to the mesothelial cells coving the omentum ( Figure 7C). To determine the mechanism of the inhibitory effect of CC on OvCa cell-mesothelial cell interactions, we treated OvCa cells and mesothelial cells in mono and co-cultures with CC and determined the effect on both cell types. Consistent with our earlier reports [10], co-cultures of OvCa cells with mesothelial cells increased phosphorylation and activation of

Compound C Inhibits OvCa Cells-Mesothelial Interactions In Vitro and In Vivo
The mesothelial cell monolayer is the first barrier that OvCa cells from the primary tumor encounter for peritoneal colonization and spread [12][13][14]57,58]. To investigate the effect of CC on the ability of OvCa cells to adhere to the mesothelial layer, human OvCa cells were treated with either CC (5 µM) or vehicle (DMSO) for 30 min and were allowed to adhere to uncoated-, and matrigel-coated plates, or mesothelial monolayers on 96 well plates for 2 h [12][13][14]. We found that CC significantly inhibited OvCa cells adhesion to the uncoated-, and matrigel-coated wells as well as mesothelial monolayers ( Figure 7A,B). To further verify the effect of CC on OvCa cell chemotaxis and adhesion to the mesothelial surface in vivo, fluorescent-labelled ID8 cells were injected intraperitoneally into C57B6 mice as earlier described [11]. Two hours later, mice were euthanized, omenta and mesentery dissected, fluorescent OvCa cells that homed to and adhered to mesothelial cells covering mesentery and omenta were visualized under a fluorescent microscope, and fluorescent signal quantified. We found that CC significantly inhibited the homing and adhesion of OvCa cells to the mesothelial cells coving the omentum ( Figure 7C). To determine the mechanism of the inhibitory effect of CC on OvCa cell-mesothelial cell interactions, we treated OvCa cells and mesothelial cells in mono and co-cultures with CC and determined the effect on both cell types. Consistent with our earlier reports [10], co-cultures of OvCa cells with mesothelial cells increased phosphorylation and activation of p65RelA subunit of NFkB. Compound C decreased phosphorylation and activation of p65RelA subunit of NFκB in mesothelial and OvCa cells in mono and cocultures (Figures 7D and S5) and significantly decreased the transcript levels of target pro-inflammatory markers as IL6, IL8, CCL2, VEGF, IL1β and TNFα in both MESO301 and SKOV3 cells ( Figure S6).

Compound C Inhibits OvCa Cell-Macrophage Crosstalk
We next determined the effect of CC on another key player in the OvCa TME, namely tumor associated macrophages (TAMs). The crosstalk between OvCa cells and TAMs is instigated by OvCa cells secretome attracting macrophages to the peritoneal TME and their phenotypic switch to proinflammatory tumor associated phenotype. In turn, the secretome of TAMs induces OvCa cell migration and invasion [13,59]. To elucidate the effect of CC on the OvCa cells-TAMs crosstalk, we determined the effect of CC on OvCa cellsinduced macrophage chemotaxis, and macrophage-induced OvCa cell matrix invasiveness [12,13,38]. We found that CC inhibited macrophage migration towards OvCa cells, i.e., chemotaxis ( Figure 8A), as well as macrophage induced OvCa cell invasiveness (Figure 8B). To further elucidate the mechanism of inhibition of U937-OvCa crosstalk, we

Compound C Inhibits OvCa Cell-Macrophage Crosstalk
We next determined the effect of CC on another key player in the OvCa TME, namely tumor associated macrophages (TAMs). The crosstalk between OvCa cells and TAMs is instigated by OvCa cells secretome attracting macrophages to the peritoneal TME and their phenotypic switch to proinflammatory tumor associated phenotype. In turn, the secretome of TAMs induces OvCa cell migration and invasion [13,59]. To elucidate the effect of CC on the OvCa cells-TAMs crosstalk, we determined the effect of CC on OvCa cells-induced macrophage chemotaxis, and macrophage-induced OvCa cell matrix invasiveness [12,13,38]. We found that CC inhibited macrophage migration towards OvCa cells, i.e., chemotaxis ( Figure 8A), as well as macrophage induced OvCa cell invasiveness ( Figure 8B). To further elucidate the mechanism of inhibition of U937-OvCa crosstalk, we treated mono-and co-cultures with CC for 24 h and found that CC decreased the activation and phosphorylation of p65RelA subunit of NFκB in both mono and cocultures ( Figures 8C and S7). Consistently, CC significantly decreased the expression of NFkB target genes in both cell types ( Figure S8). 8C and S7). Consistently, CC significantly decreased the expression of NFkB target genes in both cell types ( Figure S8).

Compound C Suppresses Cellular Bioenergetics
Our findings of the inhibitory effect of CC on PI3K-AKT-mTOR pathway prompted us to determine whether CC also inhibits metabolic programming. Realtime monitoring of cellular bioenergetics using Seahorse mito-stress assay, revealed that CC significantly inhibited basal and maximal respiration, ATP production, and non-mitochondrial O2 consumption in SKOV3, OVCAR3 and IGROV1 cell lines ( Figure 9A,B). We next determined whether CC mitigates the effect of LPA, the bone a fide activator of PI3K-AKT-mTOR pathway on cellular bioenergetics. We found that LPA treatment of OvCa cell lines SKOV3 and OVCAR3 significantly induced basal and maximal respiration, non-mitochondrial O2 consumption, and ATP production ( Figure 10). CC significantly mitigated LPA-induced basal and maximal respiration, non-mitochondrial O2 consumption, and ATP production in SKOV3 and OVCAR3. These data indicate that CC exerts its inhibitory effect through inhibition of mitochondrial respiration and ATP production as measures of oxidative phosphorylation (OXPHOS). We also determined the effect of CC on glycolytic rate by measuring extracellular acidification rate (ECAR) and cellular ATP production from glycolysis. We found CC did not exert a significant effect on ECAR or ATP production from glycolysis ( Figure S9A,B). We further confirmed the preferential inhibitory effect of CC on mitochondrial function as an ATP source in OvCa cells by calculating the ratio of mitochondrial respiration, represented as OCR, to that of glycolysis, represented as the proton

Compound C Suppresses Cellular Bioenergetics
Our findings of the inhibitory effect of CC on PI3K-AKT-mTOR pathway prompted us to determine whether CC also inhibits metabolic programming. Realtime monitoring of cellular bioenergetics using Seahorse mito-stress assay, revealed that CC significantly inhibited basal and maximal respiration, ATP production, and non-mitochondrial O 2 consumption in SKOV3, OVCAR3 and IGROV1 cell lines ( Figure 9A,B). We next determined whether CC mitigates the effect of LPA, the bone a fide activator of PI3K-AKT-mTOR pathway on cellular bioenergetics. We found that LPA treatment of OvCa cell lines SKOV3 and OVCAR3 significantly induced basal and maximal respiration, non-mitochondrial O 2 consumption, and ATP production ( Figure 10). CC significantly mitigated LPA-induced basal and maximal respiration, non-mitochondrial O 2 consumption, and ATP production in SKOV3 and OVCAR3. These data indicate that CC exerts its inhibitory effect through inhibition of mitochondrial respiration and ATP production as measures of oxidative phosphorylation (OXPHOS). We also determined the effect of CC on glycolytic rate by measuring extracellular acidification rate (ECAR) and cellular ATP production from glycolysis. We found CC did not exert a significant effect on ECAR or ATP production from glycolysis ( Figure S9A,B). We further confirmed the preferential inhibitory effect of CC on mitochondrial function as an ATP source in OvCa cells by calculating the ratio of mitochondrial respiration, represented as OCR, to that of glycolysis, represented as the proton efflux rate (PER) (Figure S10A,B). We next determined the effect of CC on mitochondrial mass using MitoTracker green-fluorescent dye [25] and found that CC did not exert significant effect on mitochondrial mass in OVCAR3 and IGROV1 cells with a trend though insignificant decrease in SKOV3 cells ( Figure S10).
Cancers 2021, 13, x FOR PEER REVIEW 15 of 24 mass using MitoTracker green-fluorescent dye [25] and found that CC did not exert significant effect on mitochondrial mass in OVCAR3 and IGROV1 cells with a trend though insignificant decrease in SKOV3 cells ( Figure S10).

Combination Therapy of Compound C and Cisplatin, Reduced Tumor Burden in SKOV3 Xenografts in Athymic Nude Mice
A major challenge in OvCa treatment is chemo-resistance, specifically to platinumderived compounds (reviewed in [60]). This prompted us to study the ability of compound C in combination with cisplatin to reduce tumor burden in vivo in OvCa cell xenografts. We injected OvCa cells intraperitoneally (ip) in athymic nude mice and ip tumor growth by IVIS bioluminescent imaging and quantification of photons flux were used for non-invasive monitoring of tumor burden in live mice. One week after tumor cell injection, mice were stratified into treatment groups that received CC, cisplatin, in mono and combinatorial therapy as well as vehicle (PBS) control ( Figure 11A). Monotherapy with CC resulted in a significant decrease in tumor burden as determined the number and size of peritoneal tumor nodules ( Figure 11B) by IVIS imaging and quantification of photon flux ( Figure S11). Cisplatin alone significantly decreased tumor burden as determined by nodule count and size, and photon flux (Figures 11B and S11). Combinatorial treatment with CC and cisplatin significantly reduced tumor burden as determined by nodule count and size and photon flux, when compared with either treatment alone (Figures 11B and  S11). Immunostaining of tumor excised from mice treated with CC exhibited significant decrease in proliferation index (nuclear ki67), mean vascular density (CD31), and tumor associated macrophages infiltration (CD68) immunostaining ( Figure 11C,D). Furthermore, tumors from CC-treated mice exhibited significant decrease in the protein expression of PI3Kp110, pAKT, mTOR as well as nuclear p65RelA subunit of NFκB ( Figure  11C,D).  SKOV3 and (B). OVCAR3 stimulated with LPA, in presence or absence of CC for 6 h. Bar graphs represent the means ± SEM of the basal and maximal respiration, spare respiratory capacity, ATP production, non-mitochondrial respiration, in (C). SKOV3 and (D). OVCAR3 cells treated with LPA ± CC (n = 5/experimental condition, repeated twice). * p < 0.05, compared DMSO, ** p < 0.05 compared to LPA, Student's t-test.

Combination Therapy of Compound C and Cisplatin, Reduced Tumor Burden in SKOV3 Xenografts in Athymic Nude Mice
A major challenge in OvCa treatment is chemo-resistance, specifically to platinumderived compounds (reviewed in [60]). This prompted us to study the ability of compound C in combination with cisplatin to reduce tumor burden in vivo in OvCa cell xenografts. We injected OvCa cells intraperitoneally (ip) in athymic nude mice and ip tumor growth by IVIS bioluminescent imaging and quantification of photons flux were used for non-invasive monitoring of tumor burden in live mice. One week after tumor cell injection, mice were stratified into treatment groups that received CC, cisplatin, in mono and combinatorial therapy as well as vehicle (PBS) control ( Figure 11A). Monotherapy with CC resulted in a significant decrease in tumor burden as determined the number and size of peritoneal tumor nodules ( Figure 11B) by IVIS imaging and quantification of photon flux ( Figure S11). Cisplatin alone significantly decreased tumor burden as determined by nodule count and size, and photon flux (Figures 11B and S11). Combinatorial treatment with CC and cisplatin significantly reduced tumor burden as determined by nodule count and size and photon flux, when compared with either treatment alone (Figures 11B and S11). Immunostaining of tumor excised from mice treated with CC exhibited significant decrease in proliferation index (nuclear ki67), mean vascular density (CD31), and tumor associated macrophages infiltration (CD68) immunostaining ( Figure 11C,D). Furthermore, tumors from CC-treated mice exhibited significant decrease in the protein expression of PI3Kp110, pAKT, mTOR as well as nuclear p65RelA subunit of NFκB ( Figure 11C,D).

Compound C Synergized with Cisplatin in Platinum-Resistant Patient Derived OvCa Cells
To further validate the therapeutic efficacy of CC on resistant and recurrent OvCa, we isolated OvCa cells from ascitic fluid and OvCa omental tumor of a patient with OvCa after two cycles of neoadjuvant platinum therapy (AF1, OM1: platinum-sensitive), and another patient who underwent six cycles of neoadjuvant platinum therapy (AF2, OM2: platinum resistant, Figure 12A). We found that CC significantly reduced proliferation of patient derived primary cells in a dose-dependent manner ( Figure 12B). Moreover, cisplatin-sensitive AF1 was more susceptible to the cytotoxic effect of cisplatin than the cisplatin-resistant AF2 as determined by IC50 of 11.3µM and 25.3µM for AF1 and AF2, respectively; an effect that was not observed in their matching omental tumor cells OM1 and OM2 ( Figure 12B). Interestingly, OM1 and OM2 were more sensitive to the cytotoxic effect of cisplatin as compared with their matching AF1 and AF2, with IC50 of 6.787µM and 7.958µM for OM1 and OM2, compared to 11.3µM and 25.3µM for AF1 and AF2, respectively. Interestingly, treatment of the four cell types with their respective IC50 concentration of CC and cisplatin exerted a synergistic effect in all four cell populations with combination indices < 1 ( Figure 12B,C), strongly suggesting the potential efficacy of CC in

Compound C Synergized with Cisplatin in Platinum-Resistant Patient Derived OvCa Cells
To further validate the therapeutic efficacy of CC on resistant and recurrent OvCa, we isolated OvCa cells from ascitic fluid and OvCa omental tumor of a patient with OvCa after two cycles of neoadjuvant platinum therapy (AF1, OM1: platinum-sensitive), and another patient who underwent six cycles of neoadjuvant platinum therapy (AF2, OM2: platinum resistant, Figure 12A). We found that CC significantly reduced proliferation of patient derived primary cells in a dose-dependent manner ( Figure 12B). Moreover, cisplatin-sensitive AF1 was more susceptible to the cytotoxic effect of cisplatin than the cisplatin-resistant AF2 as determined by IC 50 of 11.3 µM and 25.3 µM for AF1 and AF2, respectively; an effect that was not observed in their matching omental tumor cells OM1 and OM2 ( Figure 12B). Interestingly, OM1 and OM2 were more sensitive to the cytotoxic effect of cisplatin as compared with their matching AF1 and AF2, with IC 50 of 6.787 µM and 7.958 µM for OM1 and OM2, compared to 11.3 µM and 25.3 µM for AF1 and AF2, respectively. Interestingly, treatment of the four cell types with their respective IC 50 concentration of CC and cisplatin exerted a synergistic effect in all four cell populations with combination indices < 1 ( Figure 12B,C), strongly suggesting the potential efficacy of CC in conjunction with current chemotherapy to overcome resistance in high-grade ovarian tumors. conjunction with current chemotherapy to overcome resistance in high-grade ovarian tumors.

Discussion
Our current study demonstrates the potent inhibitory effects of CC in OvCa. First, we show that CC inhibits malignant phenotypes of OvCa including cell survival, proliferation, migration, and matrix invasion. We further showed that CC inhibited OvCa-stromal interactions, namely with mesothelial cells and macrophages, and demonstrated the inhibitory effect of CC on the expression and activation of the most amplified and hyperactivated oncogenic pathways in OvCa patients namely, PI3K-AKT-mTOR-NFκB. The inhibitory effect of CC on OvCa cells interactions with mesothelial cells and macrophages is mediated through inhibition of the expression and activation of p65RelA subunit of NFκB in both cell types. Importantly, CC inhibited mitochondrial bioenergetics, with modest effect on mitochondrial mass in OvCa cells. In vivo, CC in combination with cisplatin, decreased tumor burden, and corroborated our in vitro studies demonstrating its strong synergy with cisplatin in both platinum sensitive and resistant patient derived OvCa ascitic fluid and solid omental tumors.

Discussion
Our current study demonstrates the potent inhibitory effects of CC in OvCa. First, we show that CC inhibits malignant phenotypes of OvCa including cell survival, proliferation, migration, and matrix invasion. We further showed that CC inhibited OvCa-stromal interactions, namely with mesothelial cells and macrophages, and demonstrated the inhibitory effect of CC on the expression and activation of the most amplified and hyperactivated oncogenic pathways in OvCa patients namely, PI3K-AKT-mTOR-NFκB. The inhibitory effect of CC on OvCa cells interactions with mesothelial cells and macrophages is mediated through inhibition of the expression and activation of p65RelA subunit of NFκB in both cell types. Importantly, CC inhibited mitochondrial bioenergetics, with modest effect on mitochondrial mass in OvCa cells. In vivo, CC in combination with cisplatin, decreased tumor burden, and corroborated our in vitro studies demonstrating its strong synergy with cisplatin in both platinum sensitive and resistant patient derived OvCa ascitic fluid and solid omental tumors.
The anti-tumorigenic effects of CC have been studied in several cancers including colon, ovarian, and breast cancers, glioma and B-cell lymphoblastic leukemia [32,[61][62][63][64], and have historically been attributed to inhibition of AMPK and BMP pathways [64].
However, AMPK-and BMP-independent tumoricidal activity were reported [29,64,65]. Comparing the anti-tumor activity of CC to those of AMPK putative activators, AICAR and the potent biguanide phenformin, we found that CC exerted a more potent tumoricidal effect on OvCa cell lines in micromolar concentrations compared to AICAR and phenformin that exerted their effects in millimolar concentrations. In addition, re-analysis of published transcriptomic data of OvCa cell line SKOV3 treated with CC using GSEA, identified inhibitory effect of CC on PI3K-AKT-mTOR, the most commonly hyperactivated pathway with activating mutations, and amplification in many cancers including OvCa [7,8]. The anti-tumorigenic effect of CC has been also attributed to its inhibitory effect on angiogenesis through an effect on endothelial cells in in vitro assays and inhibition of tumor angiogenesis in a B16 melanoma mouse model [33,34]. This anti-angiogenic effect of CC was further supported by inhibition of the angiogenesis signature in SKOV3 treated with CC GSE60135 [40] ( Figure S11C). Our data in the present study further highlighted the antiangiogenic effect of CC not only through a direct effect on OvCa cells but through an effect on the interactions of OvCa cells with macrophages and mesothelial cells in co-cultures that significantly induced VEGF transcripts in these cells. The reciprocal feedback loop between angiogenesis and PI3K-AKT-mTOR and NFκB pathways has long been established [66][67][68][69][70][71]. Thus, our findings herein that CC inhibits VEGF transcript expression in vitro as well as angiogenesis of OvCa cell xenografts in vivo further highlight the therapeutic utility of CC in OvCa.
The role of CC as an activator or inhibitor of PI3K-AKT-mTOR-NFκB pathways is contextual not only disease-specific but are also model system-specific with substantial technical variabilities in the studies reporting such role [65,[72][73][74]. Furthermore, known chemical modifications and/or homologs of CC were designed to enhance its effects on BMP/BMP receptors as an anti-fibrotic agent in pulmonary fibrosis and diabetic nephropathy [75][76][77]. However, the therapeutic efficacy of these molecules on OvCa cells was understudied with only two studies reporting the therapeutic potential of CC in OvCa [32,40]. Our report in the present study highlights a novel compelling and multi-faceted anti-tumorigenic effect of CC that warrants further development of CC for the treatment of OvCa as an inhibitor of PI3K-AKT-mTOR-NFkB axis. Importantly, we showed that the significant suppression of proliferation, angiogenesis, TAM infiltration as well as the expression and activity of PI3K and nuclear p65RelA in SKOV3 xenografts from mice treated with CC. Moreover, we also found that CC exerted a significant cytotoxic effect on platinum-sensitive and platinumresistant omental and ascitic fluid patient-derived primary OvCa cells (OM1, AF1, OM2, and AF2, respectively). Hence, our findings suggest that CC works synergistically with cisplatin to inhibit OvCa progression in vivo and in vitro. In the literature, only one earlier study demonstrated that CC increased animal survival of A2780-s xenografts; however, this study focused on the inhibitory effect on CC on BMP-SMAD5 pathway [32].
The complex OvCa milieu is dictated by the bioenergetic alterations of the OvCa cells themselves and thus we sought to determine the effect of CC on OvCa cells bioenergetics. We found that CC inhibited oxidative phosphorylation, in particular suppressing basal and maximal respiration, spare respiratory capacity, ATP production, non-mitochondrial respiration, and proton leak in three OvCa cell lines, while having modest though insignificant effects on the glycolysis and fatty acid oxidation. Importantly, CC mitigated the stimulatory effect of LPA, initially reported as OvCa promoting factor and a putative activator of PI3K-AKT-mTOR and NFκB [78,79] on mitochondrial bioenergetics and ATP production. Given our earlier reports of the reliance of OvCa stem cells on oxidative phosphorylation [80], our results further suggest that CC could be used to target OvCa stem cells and prevent relapse.
Several clinical trials are currently underway that target PI3K pathway either as single agents (NCT01068483, NCT01833169, NCT04836663, NCT02307240, NCT04586335, NCT01936363, NCT01708161), or in combination with standard of care platinum, taxane compounds and the more recent poly-ADP ribose polymerase inhibitors, PARPi (NCT05295589, NCT00216112, NCT03586661). However, as earlier summarized [55,56,81], these agents did not go beyond phase 1 or early phase 2 clinical trials with reported adverse effects as hyperglycemia, myopathies, mucositis, and neuropathies that halted progression to advanced phases of trials [81]. Our study shows that while CC exhibited a significant anti-tumor effect in vitro, its effect as a single agent was not as potent as cisplatin. This is also consistent with the use of PI3K inhibitors in clinical trials in combination with standard of care therapeutics. Our finding that CC not only exerted a synergistic effect with cisplatin in vitro and in vivo, but it also sensitized cisplatin resistant OvCa cells to cisplatin, would justify its use in combination with cisplatin to improve the disease outcome, reduce the dose of cisplatin and the adverse events of cisplatin therapy.

Conclusions
The challenge in developing novel treatment options for OvCa lies in the fact that there is significant upregulation of oncogenic feedback mechanisms, and only a few of the current therapeutics target the interactions of OvCa in the TME that drive recurrence and resistance. Thus, our findings that CC inhibits the interactions of CC with mesothelial cells and macrophages further highlight CC as a promising therapeutic agent especially given that it was well tolerated in the in vivo preclinical models. Our findings that computational modeling showed docking of CC into PI3K, mTOR and p65RelA warrant further development of CC (and perhaps analogs) to improve potency as a multi-valent therapeutic.
Supplementary Materials: The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/cancers14205099/s1, Figure S1: Effects of phenformin and AICAR on colony formation of OvCa cell lines. A-C. SKOV3, OVCAR3, and IGROV1 were treated with AICAR (0-2 mM), and D-E. phenformin (0-1 mM). All experiments were performed in triplicates/ experimental condition and were repeated at least twice; **** p < 0.0001, one way Analysis of Variance (ANOVA); Figure S2: Full WBs of Figure 4C; Figure S3: Full WBs of Figure 4D; Figure S4: Full WBs of Figure 6A; Figure S5: Full WBs of Figure 7D; Figure S6: Effect of CC on the expression of NFkB target genes in SKOV3 and mesothelial cells in mono-and co-cultures. Bras represent the means ± SEM of the relative expression of the transcripts of the indicated genes determined by qRT-PCR. * p < 0.05, comparing CC-to DMSO-treated cells. ** p < 0.05, comparing cells in mono-to cocultures, multiple t-tests; Figure S7: Full WBs of Figure 8C; Figure S8: Effect of CC on the expression of NFkB target genes in SKOV3 and U937 macrophages in mono-and co-cultures. Bras represent the means ± SEM of the relative expression of the transcripts of the indicated genes determined by qRT-PCR. * p < 0.05, comparing CC-to DMSO-treated cells. ** p < 0.05, comparing cells in mono-to cocultures, multiple t-tests; Figure S9: The Effect of CC on glycolytic rate. A. Seahorse tracing of the proton efflux rate (PER) in SKOV3 and OVCAR3 treated with 5 µM of CC for 18 h, as described in material and methods. B. Bars representing means ± SEM of the basal and compensatory glycolysis, post 2DG acidification, basal proton efflux rate,%PER from glycolysis and mitochondrial oxygen consumption rate/glycolysis proton efflux rate (mitoOCR/glycoPER) ratio in SKOV3 and OVCAR3 cells treated with CC. A-B are representatives of 3 experiments (n = 3/experimental condition) * p < 0.05, Student's t-test; Figure S10: The effect of CC on mitochondrial mass. A. Confocal microscopy images of MitoTracker and merged images of SKOV3, OVCAR3 and IGROV1 cells treated with CC or DMSO control for 18 h (scale bar, 10 µm) B. Bar graphs represent means ± SEM of quantifies fluorescent pixel intensity of using ImageJ. p-values are determined by Student's t-test, comparing CC-to DMSOtreated groups; Figure S11: (A) IVIS imaging of tumor-bearing mice 8-weeks after ip injection of SKOV3-luc, and (B) box plots of the quantification of the photon flux/second in the experimental in the experimental cohorts. * p < 0.05, compared to the vehicle control. # p < 0.05 compared to monotherapy with either cisplatin (cis) and CC, Mann-Whitney's test. (C) GSEA analysis of SKOV3 cells treated with CC showing enrichment of a signature of angiogenesis; Table S1: List of Antibodies Used; Table S2: List of primers used.  Informed Consent Statement: Informed consent was obtained from all subjects involved in the study. Specimens were handled de-identified and cannot be linked to patients' information.

Data Availability Statement:
The data presented in this study are available in the article and the Supplementary Materials.