Visfatin induces ovarian cancer resistance to anoikis by regulating mitochondrial activity

Purpose Ovarian cancer is characterized by recurrent peritoneal and distant metastasis. To survive in a non-adherent state, floating ovarian cancer spheroids develop mechanisms to resist anoikis. Moreover, ascitic fluid from ovarian cancer patients contains high levels of visfatin with anti-apoptotic properties. However, the mechanism by which visfatin induces anoikis resistance in ovarian cancer spheroids remains unknown. Here, we aimed to assess wheather visfatin which possess anti-apoptotic properties can induce resistance of anoikis in ovarian cancer spheroids. Methods Visfatin synthesis were examined using a commercial human visfatin ELISA Kit. Spheroid were exposed to visfatin and cell viability and caspase 3/7 activity were measured using CellTiter-Glo 3D cell viability assay and Caspase-Glo® 3/7 Assay System. mRNA and protein expression were analyzed by Real-time PCR and Western Blot analysis, respectively. Analysis of mitochondrial activity was estimated by JC-1 staining. Results First, our results suggested higher expression and secretion of visfatin by epithelial than by granulosa ovarian cells, and in non-cancer tissues versus cancer tissues. Interestingly, visfatin increased the proliferation/apoptosis ratio in ovarian cancer spheroids. Specifically, both the intrinsic and extrinsic pathways of anoikis were regulated by visfatin. Moreover, the effect of the visfatin inhibitor (FK866) was opposite to that of visfatin. Furthermore, both NAMPT and FK866 affected mitochondrial activity in ovarian cancer cells. Conclusion In conclusion, visfatin acts as an anti-apoptotic factor by regulating mitochondrial activity, leading to anoikis resistance in ovarian cancer spheroids. The finding suggest visfatin as a potential novel therapeutic target for the treatment of ovarian carcinoma with peritoneal dissemination.


Introduction
Ovarian cancer, the gynecological malignancy with the highest mortality rate, is characterized by recurrent peritoneal and distant metastasis. Clinical and case studies report that ascitic fluid contains ovarian cancer cells, either as single cells or as spheroid-like structures, which drive peritoneal dissemination [1,2]. As a barrier against metastases, normal cells undergo apoptosis after they lose contact with the extracellular matrix or neighboring cells. This cell death process is called "anoikis" [3]. Floating ovarian cancer spheroids have developed mechanisms to resist anoikis, and so have the ability to survive in a nonadherent state while traveling through the circulatory and lymphatic systems [4,5]. Anoikis involves two apoptotic pathways: the cell death receptor (extrinsic) and mitochondrial (intrinsic) pathways. Likewise, blockade of the mitochondrial pathway, for example, via over-expression of anti-apoptotic Bcl-2 family proteins, can also induce resistance to anoikis [6].
Interestingly, ascites from ovarian cancer patients contains significantly higher levels of visfatin (75.3 ± 28.1 ng/mL) than serum; these high levels are associated intraperitoneal dissemination of ovarian cancer. The source of visfatin in ascites fluid may be both ascites-derived ovarian cancer cells and/or peritoneal and omental adipocytes, which release soluble factors into the ascites and provide a pro-tumor microenvironment [7]. Visfatin is an adipocytokine and cytosolic enzyme with nicotinamide phosphoribosyltransferase (Nampt) activity in mammals. Furthermore, studies show that visfatin inhibits apoptosis of endometrial cancer [8] and breast cancer [9] cells. However, it is unclear whether visfatin triggers anoikis resistance in ovarian cancer spheroids.
The above data raise the intriguing hypothesis that visfatin present in ascites fluid of ovarian cancer patients possess antiapoptotic properties that drive anoikis resistance in ovarian cancer spheroids. Therefore, the main aim of this study was to examine involvement of visfatin in triggering anoikis resistance in ovarian cancer spheroids, taking into account the special role of mitochondria in anoikis.

Cell culture and chemicals
Three human ovarian cancer cell lines derived from ascitic fluid or the peritoneum were used as an in vitro model to investigate ovarian cancer cell tumorigenesis. These were the human epithelial ovarian carcinoma cell line OVCAR-3 (ATCC, Manassas, VA, USA); the human epithelial ovarian cancer cell line SKOV-3 (ATCC, Manassas, VI, USA); and the human AGCT-derived cell line KGN (Riken Cell Bank (RBRC-RCB1154), Ibaraki, Japan; approved by Drs. Yoshiro Nishi and Toshihiko Yanase). The human ovarian surface epithelial cell line HOSEpiC (ScienCell Research Laboratories, Carlsbad, CA), and a human non-luteinized granulosa cell line derived from antral follicles (HGrC1; a kind gift from Dr. Ikara Iwase (Nagoya University, Japan)) were used as non-cancer controls. OVCAR-3 cells were cultured in RPMI 1640 medium without phenol red, supplemented with 10% heat-inactivated, charcoal-stripped fetal bovine serum (FBS) (Biowest, Nuaillé, France

Visfatin cell synthesis analysis
To examine visfatin synthesis, non-cancerous ovarian cells and ovarian cancer cells were cultured for 48 h in appropriate medium. Basal visfatin levels in cell lysates was measured after 48 h of culture using a commercial human visfatin SimpleStep ELISA Kit (ab264623, Abcam). Absorbance was measured using a ELx808 microplate reader (Bio-Tek Instruments, Winooski, VT, USA). Data were analyzed using the KC Junior software (Bio-Tek) and normalized to total protein levels in cells (measured using the Bradford method; Bio-Rad Protein, Hercules, CA, USA).
Caspase -3, -8 and -9 activity assays KGN and SKOV-3 cells were exposed to visfatin for 24 h at concentrations 10, 100, or 1000 ng/ml in medium without FBS (Biowest). The medium was removed and the cells were lysed with the caspase assay buffer. The amount of protein in the lysates was measured using the reactive compound fluorescamine (MP Biomedicals, Illkirch Cedex, France). An equal amount of cytosolic extract (50 μg of protein) from each sample was analyzed. The assay was performed by addition of 100 μM of the substrate acetylo-DEVD-amino-4-methylcoumarin (Ac-DEVDAMC; Sigma-Aldrich), which is hydrolyzed by Caspase-3, 100 μM N-Acetyl-Ile-Glu-Thr-Asp-7-Amido-4-methylcoumarin (IETD-pNA) (Sigma-Aldrich), which is hydrolyzed by Caspase-8, or 100 μM N-Acetyl-Leu-Glu-His-Asp-7-amido-4-trifluoromethylcoumarin (Ac-LEHD-AFC) (Sigma-Aldrich), which is hydrolyzed by Caspase-9, followed by incubation at 37°C. The amount of fluorescent product was measured continuously for 120 min using a spectrofluorometer (FLx800; Bio-Tek Instruments, Winooski, VT, USA) at an excitation wavelength of 355 nm and an emission wavelength of 460 nm. Data were analyzed using KC JUNIOR software and normalized against the level of fluorescence in vehicle-treated cells.

Real-time PCR technique
Total RNA isolation and cDNA synthesis were carried out on cells at control and/or after treatment with visfatin (100 ng/mL) for 24 h, using the TaqMan Gene Expression Cells-to-CT kit (Applied Biosystems, ThermoFisher Scientific) according to the manufacturer's instructions. The resulting pre-amplified cDNA preparations were analyzed by real-time PCR using the StepOnePlus real-time PCR system (Applied Biosystems, ThermoFisher Scientific) and TaqMan gene expression assays in combination with Taq-Man Gene Expression Master Mix containing the ROX passive reference dye (Applied Biosystems, ThermoFisher Scientific. Real-time PCR using TaqMan Gene Expression Assays (Applied Biosystems, Thermo Fisher Scientific) was performed to measure basal expression of mRNA encoding NAMPT (Hs00237184_m1) in HOSEpiC, HGrC1, OVCAR-3, SKOV-3, and KGN cells and expression of mRNA encoding caspase-3 (Hs00234387_m1), Bax (Hs00180269_m1), Bcl-2 (Hs00236329_m1), and PARP1 (Hs00242302_m1) in KGN and SKOV-3 cells at 24 h posttreatment with visfatin (100 ng/mL), as described previously [10]. Expression was normalized to that of GAPDH (4310884E), and relative expression was calculated using the 2 −ΔΔCt method [11].

Western blot analysis
After treating with visfatin (100 ng/mL) for 48 h, the cells were lysed in lysis buffer. Proteins were separated in 4-20% Mini-Protean TGX Precast Protein Gels (Bio-Rad, Hercules, CA, USA) and transferred to Trans-Blot Turbo Mini PVDF transfer packs (Bio-Rad) using the Trans-Blot Turbo transfer system (Bio-Rad). The blots were blocked for 1 h with 0.02 M Tris-buffered saline containing 5% BSA and 0.1% Tween 20, and then incubated overnight at 4°C with primary antibodies ( Table 1). The membranes were then washed three times in TBST (Tris-buffered saline, 0.1% Tween 20) and incubated for 1 h at room temperature with horseradish peroxidase (HRP)-conjugated secondary antibodies (Table 1) as described previously [10]. β-actin was used as a loading control (Table 1). Immunopositive bands were visualized using WesternBright Sirius Western blotting HRP substrate (Advansta, Menlo Park, CA, USA). Quantification of protein bands at three independent experiments was measured by densitometry using Vision-Works LS Acquisition and Analysis software (UVP, Upland, CA, USA).

Quantification of mitochondrial activity
To measure the amount of ATP in OVCAR-3 and KGN cells treated with NAMPT (50 ng/mL), cells were treated with FK866 (10 nM), or co-treated with NAMPT (50 ng/ mL) and FK866 (10 nM), and used in the Cell Titer-Glo Assay (Promega, Charbonnieresles-Bains, France) to assess the presence of metabolically active cells. OVCAR-3 and KGN cells were lysed according to the manufacturer's instructions and luminescence was measured using a SpectraMax L luminometer (Molecular Devices, San Jose CA, USA). Furthermore, JC-1 staining was performed as an indicator of the electrochemical potential of the inner mitochondrial membrane. This fluorescent dye can be used to distinguish active and inactive mitochondria in the same cell. After entering the mitochondria, fluorescence changes from green (JC1 monomers, indicative of inactive mitochondria) to red as the mitochondrial membrane becomes polarized and aggregates of JC-1 form (indicative of active mitochondria). The ratio of this green/red fluorescence is independent of mitochondrial shape, density, or size, but depends only on the membrane potential (for details see Chazotte [12] and Krawczyk et al., [13]). KGN cells treated with visfatin and/or FK866 as described above were incubated with 10 μg/mL JC-1 (Sigma-Aldrich) at 37°C in serum-free medium for 10 min. Thereafter, the medium was replaced by fresh medium and fluorescence was detected using an Axiocam 503 bright field/fluorescence microscope (Zeiss). Both monomeric (excitation wavelength, 490 nm; emission wavelength, 500-550 nm) and aggregated (excitation wavelength, 555 nm; emission wavelength, 575-620 nm) forms of JC-1 were detected. The obtained images were merged using ImageJ. A predominance of red fluorescence was indicative of active mitochondria, while a predominance of green fluorescence was indicative of inactive mitochondria.

Statistical analysis
Statistical data analysis was performed using tools within Prism software (GraphPad Software Inc., San Diego, USA). The mean ± SEM or SD of datasets generated in triplicate (at least three independent experiments per condition) were compared using one-way or two-way analysis of variance, followed by Tukey's test, or using a parametric Student's t test (*P < 0.05, **P < 0.01, ***P < 0.001).

Basal level of visfatin in ovarian cancer cells from the databases
Visfatin expression in normal ovarian and ovarian cancer cells was checked with the help of CSIOVDB, which is a microarray gene expression database of ovarian cancer subtypes (http://csiovdb.mc.ntu.edu.tw/CSIOVDB.html) [14]. Analysis using CSIOVDB identified expression of NAMPT in each analyzed sample. The level of visfatin is significantly higher in OSE than in normal stroma. However, there were no differences in the level of NAMPT according to the type of ovarian tumor. Interestingly, that database also revealed expression of mRNA encoding NAMPT in peritoneal metastases (Fig. 1).

Basal levels of visfatin in non-cancer and cancer ovarian cell lines
Based on the data obtained from the databases, we investigated the role of visfatin in ovarian cancer tumors. To do this we measured basal expression of visfatin (NAMPT) mRNA in human ovarian non-cancer (HOSEpiC and HGrC1) and human ovarian cancer (OVCAR-3, SKOV-3, and KGN) cell lines by qRT-PCR. The relative quantity

Effect of visfatin on the proliferation/apoptosis ratio in ovarian cancer cell lines
Pro-cancerogenic activity is related to resistance to anoikis, which in terns enables cancer cells to survive. Therefore, we analyzed the effect of visfatin on the proliferation/apoptosis (P/A) ratio of OVCAR-3 (Fig. 3a), SKOV-3 (Fig. 3b), and KGN (Fig. 3c) spheroids in non-adherent cell cultures. Proliferation was evaluated by counting the number of viable cells using the CellTiter-Glo 3D cell viability assay after  (Fig. 3b, c; *P < 0.05, **P < 0.01, and ***P < 0.001). Visfatin had no effect on the P/A ratio of epithelial ovarian cancer (OVCAR-3) cells (Fig. 3a).

FK866 reduces the proliferation/apoptosis ratio and triggers mitochondrial dysfunction, leading to decreased levels of ATP in KGN cells
To evaluate the effect of NAD + depletion on ovarian cancer survival, KGN cells were treated with FK866 (10 or 100 nM). As shown in Fig. 6a, KGN cells exposed to FK866 (100 nM)

Visfatin increases mitochondrial activity, but FK866 decreases mitochondrial activity
Because the ATP level in KGN cells increased markedly after visfatin treatment, we decided to check whether these changes are associated with mitochondrial activity and the chemical potential of an inner membrane of these organelles. Therefore, cells pre-treated with visfatin (NAMPT) and/or its inhibitor (FK866) were stained with JC1. Use of this chemical compound enables precise distinction between active and inactive mitochondria (for further clarification see Material and Methods, and Chazotte [12]). Inactive mitochondria with a low membrane potential emitted green fluorescence (Fig. 7A-D), while active mitochondria with a high electrochemical potential emitted red fluorescence (Fig. 7A'-D'). Our analyses showed that treatment with NAMPT resulted in a significant increase in mitochondrial with visfatin (100 ng/mL) for 24 h, and levels of BID/tBID protein (k) in KGN cells. Expression of mRNA in vehicle-treated controls was set to RQ = 1.0. Each bar represents the mean ± SEM of three independent experiments. *P < 0.05, **P < 0.01, and ***P < 0.001 activity (Fig. 7B'), which can be clearly seen in merged images with predominant red fluorescence (representing active mitochondria with a high membrane potential) (Fig. 7B"). By contrast, treatment with FK866 led to a marked reduction in mitochondrial activity (Fig. 7D), resulting in predominant green fluorescence in the merged image (Fig. 7D"). It is interesting to note that when cells were treated simultaneously with both visfatin and FK866, no increase in activity was observed (not shown).

Discussion
Obesity-related cancers, including ovarian cancer, secrete abnormal levels of adipokines such as visfatin [15,16]. Furthermore, visfatin is present in ascites fluid from ovarian cancer patients, suggesting a connection with intraperitoneal dissemination [7]. Data obtained from databases indicate that both normal and various types of cancer ovarian cells express NAMPT. Databases show higher levels of NAMPT  transcripts in OSE cells than in stroma cells, which is consistent with our results showing higher mRNA expression and secretion of visfatin by epithelial ovarian HOSE-piC cells (which are OSE cells) than ovarian granulosa HGrC1 cells (which are stroma cells). Moreover, NAMPT is expressed by cells taken from peritoneal metastases. However, no study has examined NAMPT expression in ovarian granulosa cell tumors, which are rare but can metastasize aggressively to the peritoneal cavity. Our results showed that visfatin is synthesized and secreted by ovarian cancer cell lines from ascitic fluid, and by adult granulosa cell tumor cells (KGN cells). These findings agree with reports showing that visfatin is produced by ovarian cancer cells [7]. In addition, the expression level varied between non-cancer cells and cancer cells: higher basal levels of NAMPT transcripts and higher secretion were noted in noncancer epithelial ovarian cells (HOSEpiC) than in epithelial ovarian cancers (OVCAR-3) and (SKOV-3), and in noncancer ovarian granulosa cells (HGrC1) than in adult granulosa (KGN) cells. The results of a previous metaanalysis revealed significantly higher circulating visfatin levels in the serum in patients with various cancers than in controls [17]. However, our results focusing on local visfatin sources indicate an inverse correlation. This suggests that local levels of visfatin may be different to circulating levels.
Ascites fluid contains single ovarian cancer cells and spheroid-like structures; the latter are thought to favor peritoneal dissemination [2]. Floating ovarian cancer spheroids acquire the ability to survive in a non-adherent state because they are resistant to anoikis. Because visfatin possesses proliferative, anti-apoptotic, pro-angiogenic, and enzymatic activity properties [18], we examined its effect on the P/A ratio in spheroids of three ovarian cancer cell lines grown under non-adhesion conditions to reflect in vivo conditions. We observed that visfatin increased the P/A ratio in SKOV-3 and KGN cells. However, there was no effect in the OVCAR-3 cell line, which expressed the highest levels of visfatin. After observing the anti-apoptotic effects of the visfatin, we evaluated its effect on caspase activation. We found downregulated expression of caspase-3 mRNA, accompanied by decreased caspase-3 activity, in KGN and SKOV-3 cells. In addition, we showed decreasing cleavage of caspase-3 following visfatin treatment. Wang et al. [8] reported that visfatin stimulates proliferation, and inhibits apoptosis of, endometrial carcinoma in both Ishikawa and KLE cells. Furthermore, visfatin may promote proliferation and inhibit apoptosis of colon and breast cancer cells [9,19]. PARP is one of the best known substrates for caspase activity, and PARP cleavage to yield fragments of 89 and 24 kDa is a universally accepted hallmark of apoptosis [20]. Our analysis of PARP1 expression in KGN and SKOV-3 cells showed no changes in NAMPT gene expression, or the cleaved form PARP1 protein, after visfatin treatment. These observations are partly in line with those of Gholinejad et al. [9], who reported that visfatin improves cell viability, and prevents TNF-α-induced apoptosis and PARP cleavage, in breast cancer cells.
Induction of anoikis occurs through interplay between two apoptosis pathways: the intrinsic pathway and the extrinsic pathway. The extrinsic pathway is activated by upregulated FAS and FasL (and inhibited by FLIP), leading to activation of caspase-8, followed by activation of caspase-7 and caspase-3. Loss of cell adhesion also activates the intrinsic pathway by increasing activation of proapoptotic proteins such as Bax, which inactivate antiapoptotic Bcl-2 proteins and release cytochrome c from the mitochondria, which activates caspase 9, followed by caspase 3 [21]. Our results showed that visfatin reduces both caspase 8 and caspase 9 activity in KGN and SKOV-3 cells. Additionally, we found downregulated Bax expression and upregulated Bcl-2 expression at the gene and protein levels. This consistent with experiments showing that visfatin decreases expression of pro-apoptotic proteins (Bax and cleaved caspase) and simultaneously increases expression of anti-apoptotic proteins (Bcl-2), culminating in neuroprotective effects against ischemic injury [22]. Similarly, Xiang et al. [23] showed that visfatin protects rat pancreatic β-cells against IFN-γ-induced apoptosis. The link between the two apoptosis pathways is the BID protein, which is activated after cleavage of caspase-8 [24,25]. Interestingly, we noted increased expression of cleaved BID, suggesting that both the intrinsic and extrinsic pathways are involved in the action of visfatin. Thus, visfatin plays a role in anoikis resistance by inhibiting both the intrinsic and extrinsic pathways.
To confirm our observations, we used the nicotinamide analog FK866 (also known as WK175 and APO866) which acts as a competitive inhibitor of NAMPT [26]. FK866 was the first nanomolar-effective NAMPT inhibitor, and is currently the most widely used in the clinic. FK866 has shown positive results when used to treat cancer [27]. We found that FK866 (100 nM) significantly inhibited proliferation, and increased apoptosis, of KGN cells. Moreover, the P/A ratio decreased after treatment with FK866. In vitro studies show that FK866 induces apoptosis of liver cancer cells (HepG2) via highly specific, non-competitive inhibition of nicotinamide phosphoribosyltransferase (NAPRT) [28]. Furthermore, FK866 inhibits growth of a wide variety of cancer cell lines, and of tumors in in vivo models [29,30].
The role of mitochondria in caspase activation during apoptosis is fairly well characterized. However, disruption of mitochondrial function during apoptosis mediated by caspase cleavage has also been demonstrated [31]. Waterhouse et al. [32] reported mitochondrial damage and a rapid fall in ATP levels in apoptotic cells upon caspase activation. Therefore, we next analyzed the effect of NAMPT and FK866 on mitochondrial function. We observed increased the ATP content after treatment of KGN cells with visfatin, whereas FK866 decreased ATP levels. Gherke et al. [33] reported that FK866 triggers death of chronic lymphocytic leukemia cells by reducing cellular NAD and ATP levels in a time-and concentration-dependent manner. Therefore, to detect changes in mitochondrial activity, we performed JC-1 staining. Our experiments showed that visfatin increased, and FK866 decreased, mitochondrial activity. These results are in line with a previous observation that FK866 inhibits mitochondrial respiratory activity [28]. Interestingly, a previous study showed that a NAMPT inhibitor induced apoptosis accompanied by activation of caspases, DNA fragmentation, and disruption of mitochondrial transmembrane potential, in primary adult T-cell leukemia/lymphoma (which shows high expression of NAMPT) [34]. Consistent with this, Liu et al., [30] demonstrated that targeting of NAD + by FK866 reduced mitochondrial membrane potential, which ultimately increased apoptosis, and inhibited proliferation, of gastric cancer cells. In addition, NAMPT silencing reduced intracellular NAD and ATP levels [30].
To sum up, the observations that ATP levels and mitochondrial activity increase in parallel with downregulation of caspase expression after treatment with visfatin support the hypothesis that visfatin is an ani-apoptotic factor that triggers anoikis resistance in ovarian cancer spheroids. The visfatin inhibitor FK866 exerted pro-apoptotic properties by decreasing ATP levels and mitochondrial activity in parallel with upregulating caspase activity, thereby stimulating anoikis in ovarian cancer spheroids. Thus, targeting/inhibiting visfatin may be a potential novel therapeutic approach to inhibiting peritoneal dissemination of ovarian carcinoma.