The occurrence of chemotherapy-resistant tumors makes ovarian cancer (OC) the most lethal gynecological malignancy. While many factors may contribute to chemoresistance, the mechanisms responsible for regulating tumor vulnerability are under investigation. Our analysis of gene expression data revealed that Sab, a mitochondrial outer membrane (MOM) scaffold protein, was down-regulated in OC patients. Sab-mediated signaling induces cell death, suggesting that this apoptotic pathway is diminished in OC. We examined Sab expression in a panel of OC cell lines and found that the magnitude of Sab expression correlated to chemo-responsiveness; wherein, OC cells with low Sab levels were chemoresistant. The Sab levels were reflected by a corresponding amount of stress-induced c-Jun N-terminal kinase (JNK) on the MOM. BH3 profiling and examination of Bcl-2 and BH3-only protein concentrations revealed that cells with high Sab concentrations were primed for apoptosis, as determined by the decrease in pro-survival Bcl-2 proteins and an increase in pro-apoptotic BH3-only proteins on mitochondria. Furthermore, overexpression of Sab in chemoresistant cells enhanced apoptotic priming and restored cellular vulnerability to a combination treatment of cisplatin and paclitaxel. Contrariwise, inhibiting Sab-mediated signaling or silencing Sab expression in a chemosensitive cell line resulted in decreased apoptotic priming and increased resistance. The effects of silencing on Sab on the resistance to chemotherapeutic agents were emulated by the silencing or inhibition of JNK, which could be attributed to changes in Bcl-2 protein concentrations induced by sub-chronic JNK inhibition. We propose that Sab may be a prognostic biomarker to discern personalized treatments for OC patients.

Introduction

Ovarian cancer (OC) is the most lethal gynecological malignancy [1]. While the overall 5-year survival for OC is ∼45%, 60% of OC patients suffer from the metastatic disease, which has a survival rate of nearly 28%; whereas, locally detected OC cases have a 92% survival rate [2]. The high mortality associated with OC is attributed to a combination of poor detection and an extremely high recurrence rate (∼80%) [3]. Recurrent OC is typified by the presence of treatment-resistant tumor cells [4]. Specifically, OC patients are often characterized by their relative responsiveness to platinum-based drugs, which are commonly used chemotherapy agents for OC [4,5]. Patients are classified as platinum-sensitive (recurrence after 6 months), platinum-resistant (recurrence in less than 6 months) or platinum-refractory (recurrence during or upon completion of chemotherapy) [6], and this patient stratification is used, in part, to determine treatment options. In fact, the front-line treatment of paclitaxel and a platinum agent, such as cisplatin, only provides a modest improvement in survival among OC patients [7,8]. Thus, the molecular mechanisms driving chemoresistance need to be defined and targeted to enhance treatment responsiveness in OC patients. To date, many factors have been implicated and targeted in platinum-resistant/refractory OC to improve therapeutic efficacy including angiogenesis [9,10], perturbations in cellular signal transduction [1115], DNA maintenance and repair machinery [7,16], folate metabolism [17,18], and mitochondrial physiology [1921].

Recent research has demonstrated that changes in the apoptotic potential (ability to induce cell death) is altered in advanced OC patients [22,23]. Wherein, a decrease in the relative abundance of pro-apoptotic BH3-only proteins is linked to chemoresistance in OC. BH3-only proteins (Bad, Bid, Bik, Bim, Noxa, and Puma) are members of the Bcl-2 superfamily of proteins and facilitate apoptosis by sequestering pro-survival Bcl-2 proteins or enhancing mitochondrial outer membrane permeabilization (MOMP) by catalyzing the assembly of Bax–Bak pores [24,25]. Because of the association among BH3-only proteins, apoptosis, and chemo-responsiveness, BH3 profiling, a reproducible, high-throughput assessment of BH3-only protein levels, has been proposed to be an index for chemo-responsiveness in solid tumors, including OC tumors [2628]. Recent BH3 profiling efforts have found that indeed OC cells with high levels of BH3-only proteins were sensitive to chemotherapy [22,23]. Moreover, increasing BH3-only protein concentrations, emulating BH3-only functions with chemical mimetics (ABT-737), and inhibiting pro-survival Bcl-2 proteins have been found to enhance chemotherapeutic efficacy in resistant OC cells [23,24,29]. However, the precise molecular mechanisms controlling the relative levels of BH3-only proteins on mitochondria in OC are currently being investigated and may represent new therapeutic and prognostic targets.

The mitochondrial outer membrane (MOM) is the interface between mitochondria and the rest of the cell. The MOM is not only the primary site of signal integration for mitochondria [30], but the MOM is also the location of Bcl-2 superfamily proteins, including several BH3-only proteins [31,32]. Signaling complexes on the MOM may influence many aspects of mitochondrial physiology, including the recruitment, function, and retention of BH3-only proteins. Previously, we found that the MOM scaffold protein Sab organized pro-apoptotic signaling cascades in response to cytotoxic stress [3335]. Specifically, the c-Jun N-terminal kinase (JNK) translocated to the mitochondria and interacted with Sab in human cervical cancer (HeLa) cells treated with anisomycin [34]. Selective inhibition of the JNK–Sab interaction using a small, cell-permeable peptide (Tat-SabKIM1) prevented JNK-induced apoptotic events such as phosphorylation and emigration of Bcl-2 from the MOM [33]. Recently, we reported that sub-chronic treatment of HeLa cells with a chemo-sensitizer, LY294002, increased Sab expression and chemosensitivity towards cisplatin and paclitaxel, two drugs commonly used in combination to treat OC [35]. In general, enhanced Sab-mediated signaling may increase the sensitivity of cells to toxic stress [36], including OC tumors. However, we have yet to determine the precise molecular mechanisms by which Sab-mediated signaling may reduce chemoresistance.

In our current study, we examined the role of Sab-mediated signaling in OC chemo-responsiveness. Analysis of gene expression data from six, independent OC patient studies shows that Sab expression is reduced more than 5-fold on average across 970 OC samples when compared with healthy tissue. Profiling four commercially available OC cell lines revealed that cells with high Sab expression are sensitive to conventional chemotherapy agents and that Sab levels reflect the concentrations of BH3-only proteins on mitochondria. Additionally, overexpression of Sab in the chemoresistant SK-OV-3 cell line increased BH3-only protein levels and chemosensitivity to cisplatin and paclitaxel treatment. Conversely, inhibition of Sab-mediated signaling in a chemosensitive cell line, PA-1, was sufficient to enhance chemoresistance to paclitaxel and cisplatin. Our studies demonstrate that manipulation of MOM signaling, such as increasing Sab-mediated signaling events, may be a useful strategy to determine viable, personalized treatment options for OC patients.

Experimental procedures

Materials

Cell lines were purchased from American Type Culture Collection (ATCC, Manassas, VA). Human Ovarian Epithelium Cells were obtained from ScienCell Research Laboratories (Carlsbad, CA). Chemotherapeutic agents ABT-737, carboplatin, cisplatin, docetaxel, doxorubicin, etoposide, and paclitaxel were purchased from Sigma–Aldrich (St. Louis, MO). Antibodies were purchased from vendors as listed below. General laboratory supplies were purchased from Fisher Scientific (Pittsburgh, PA).

Gene expression analysis

Secondary data analysis was performed to determine the relative levels of Sab expression in normal and OC tissue samples. The expression of Sab (SH3-binding protein 5; SH3BP5) was examined using the Oncomine database (http://www.oncomine.orgn) in March 2017 [37]. By querying six datasets available for OC that included Sab as part of past studies [3843], we compared experimental data for control tissue and OC samples. The data were compiled from across six independent studies, which contained a total of 41 normal tissue samples and 970 OC samples (1011 samples in total). The data were presented as medians with error bars with limits of the 90th and 10th percentiles. All of the OC samples were from surgically removed primary site tumors (summarized in Table 1).

Table 1
Summary of ovarian cancer expression studies and observed changes in Sab expression

The data presented in the table was obtained from the Oncomine repository and statistically reexamined for Sab (SH3BP5) expression.

Study Total samples Cancer samples Normal samples Ovarian cancer types2 Normal tissue Genes analyzed Human genome array(s) log2 fold change Avg. fold change P Reference 
Abid 16 12 Serous carcinoma Ovary 8603 U95A-Av2 −2.331 −5.032 0.003 [38
Bonome 195 185 10 Ovarian carcinoma Ovarian Surface Epithelium 12 624 U133A −3.776 −13.699 7.18 × 10−9 [39
Hendrix 103 99 multiple2 Ovary 12 624 U133A −1.446 −2.725 1.43 × 10−5 [40
Lu 50 45 Multiple2 Ovarian Surface Epithelium 17 572 U95A-Av2 −1.285 −2.437 0.0672 [41
U95B 
U95C 
U95D 
U95E 
TCGA 594 586 Ovarian serous cystadenocarcinoma Ovary 12 624 U113A −1.684 −3.213 1.17 × 10−7 [42
Yoshihara 53 43 10 Ovarian serous adenocarcinoma Peritoneum 16 724 Agilent 1A Oligo v2 −2.619 −6.143 2.10 × 109 [43
Totals/Means1 1,011 970 41     2.190 5.542   
Study Total samples Cancer samples Normal samples Ovarian cancer types2 Normal tissue Genes analyzed Human genome array(s) log2 fold change Avg. fold change P Reference 
Abid 16 12 Serous carcinoma Ovary 8603 U95A-Av2 −2.331 −5.032 0.003 [38
Bonome 195 185 10 Ovarian carcinoma Ovarian Surface Epithelium 12 624 U133A −3.776 −13.699 7.18 × 10−9 [39
Hendrix 103 99 multiple2 Ovary 12 624 U133A −1.446 −2.725 1.43 × 10−5 [40
Lu 50 45 Multiple2 Ovarian Surface Epithelium 17 572 U95A-Av2 −1.285 −2.437 0.0672 [41
U95B 
U95C 
U95D 
U95E 
TCGA 594 586 Ovarian serous cystadenocarcinoma Ovary 12 624 U113A −1.684 −3.213 1.17 × 10−7 [42
Yoshihara 53 43 10 Ovarian serous adenocarcinoma Peritoneum 16 724 Agilent 1A Oligo v2 −2.619 −6.143 2.10 × 109 [43
Totals/Means1 1,011 970 41     2.190 5.542   
*

The number of samples from each study was added to arrive at the total numbers used for our evaluation. The mean fold changes were generated by averaging the fold changes from the studies;

The types of OC were provided by the studies; multiple types were used in the Hendrix and Lu studies, and we did not segregate the data based on type. In both studies, the types were clear cell adenocarcinoma, serous adenocarcinoma, endometrioid adenocarcinoma, and mucinous adenocarcinoma.

Cell culture

OC cell lines CaOV-3 (HTB-75), PA-1 (CRL-1572), SW-626 (HTB-78), and SK-OV-3 (HTB-77) OVCAR-3, OVCAR-8, ES-2, OV-21G, OV-90, and OV-112D cells were cultured according to the supplier's instructions in Dulbecco's Minimal Essential Medium (DMEM), Eagle's Minimal Essential Medium (EMEM), Lebovitz L15 medium, and McCoy's 5a medium, respectively, supplemented with 10% fetal bovine serum (FBS), 1000 U/ml penicillin, 100 mg/ml streptomycin, and 5 µg/ml plasmocin. For experiments, cells between passages 3 and 20 were used. To account for media-induced effects, cells were also adapted to DMEM and analyzed as described below. Human ovarian epithelium cells were cultured under standard cell culture conditions on poly-l-lysine-coated plates according to manufacturer's instructions using Ovarian Epithelial Cell Medium (ScienCell Research Laboratories) for no more than five passages.

Immunoblotting

To isolate proteins for analysis, cells were plated at 2.5 × 105 cells/well in six-well plates and 60-mm dishes. Following treatment, cells were lysed, and proteins were harvested as previously described. Briefly, cells were washed twice in phosphate-buffered saline (PBS; 137 mM NaCl, 2.7 mM KCl, 10 mM Na2HPO4, and 1.8 mM KH2PO4) and lysed in radioimmunoprecipitation assay buffer (RIPA; 50 mM Tris–HCl, pH 8.0, 150 mM NaCl, 1% Nonidet P-40, 0.5% deoxycholate, 0.1% SDS) supplemented with 1 mM phenylmethanesulfonyl fluoride (PMSF) and Halt Protease and Phosphatase Inhibitor Cocktails (Thermo Scientific). Cells were incubated while gently rocking at 4°C for 5 min, and then transferred to a sterile microcentrifuge tube. After 2 min on ice, cell disruption was completed using sonication. The lysate was cleared by centrifuging at 14 000×g for 15 min. Supernatant protein concentrations were measured using the Pierce BCA Assay kit. Proteins (25 µg) were resolved by SDS–PAGE and transferred onto low-fluorescence PVDF membranes. Membranes were placed in blocking buffer comprised of PBS with 5% bovine serum albumin (BSA) and incubated for 1 h at room temperature. The membranes were incubated in PBS containing 0.1% Tween 20 (PBST) and 5% BSA in the presence of primary antibodies overnight (4°C) while gently rocking. Primary antibodies specific for Sab (Novus Biologicals, H00009467-M01), Phospho-JNK [Thr183/Tyr185, Cell Signaling Technology (CST), 4668], JNK (Cell Signaling Technology, 9252), Phospho-c-Jun (Ser73, CST, 3270), c-Jun (CST, 9165), Phospho-Bcl-2 (Ser70) (CST, 2827), Bcl-2 (CST, 2870), Bcl-xL (CST, 2764), Mcl-1 (CST, 5453), Bad (CST, 9239), Bik (CST, 4592), Bim (CST, 2933), Bid (CST, 2002), PUMA (CST, 12450), Bax (CST, 5023), Bak (CST, 12105)Actin (CST, 4970), α-tubulin (CST, 2144), COX-IV (CST, 4850), GAPDH (CST, 5174), Calnexin (CST, 2679), Histone H3 (CST, 4499), TOM20 (Abcam, ab115746), TIMM23 (Abcam, ab116329), and PEX19 (Abcam, ab137072) were used at dilutions of 1:1000. Membranes were washed three times for 5 min in PBST and were incubated with secondary antibodies in the appropriate blocking buffer at a ratio of 1:20 000 for 1 h at RT gently rocking. The following secondary antibodies were used in the experiments below: IRDye 680RD goat anti-rabbit (926–32 211) and IRDye 800CW goat anti-mouse (926–68 070) (Licor Biosciences). Membranes were again washed three times for 5 min in PBST. Membranes were analyzed using fluorescence detection using the Odyssey CLx near infrared scanner (Licor Biosciences). The corresponding bands on the immunoblots were quantified and normalized using the Image Studio 2.0 software (Licor Biosciences). The fluorescence of specific bands of interest was divided by the fluorescence of the loading control band to equilibrate signal strength and loading. The resulting signal was then normalized by dividing the signals from treated samples by untreated controls for each experiment.

Cell-based IC50s of chemotherapeutic agents

To determine the chemosensitivity of the OC cell lines, cells were treated with increasing concentrations (generally 0–100 µM) of drugs approved for the treatment of OC. Cells were plated at 1.5 × 104 cells per well in black-walled, clear bottom plates (PerkinElmer) and grown overnight. The cells were then treated with chemotherapeutic agents for 48 h. The cells were washed three times with PBS and fixed in 4% paraformaldehyde/PBS. The cells were then stained with 5 µM TO-PRO-3 for 45 min at RT [35,36]. The cells were then washed three times in Hank's Buffered Saline Solution (HBSS). The plate was imaged using the Odyssey CLx scanner (LI-COR Biosciences) and analyzed using the Image Studio 2.0 software (Licor Biosciences). The IC50s were then calculated using the GraphPad Prism7© software.

Mitochondrial isolation

To determine the relative abundance of proteins located on or within mitochondria, we isolated mitochondria as described in our previous work [33,36]. Mitochondrial preparations with greater than 80% purity were used for our studies. Mitochondria purity was determined by immunoblotting respective subcellular compartments for proteins characteristic of other organelles. For protein analyses, a total of 50 µg protein was loaded for each subcellular compartment sample. Additionally, the relative activities of cytosolic glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and lactate dehydrogenase (LDH) and mitochondrial citrate synthase (CS) were used to determine the relative levels of cytosolic and mitochondrial contamination in each subcellular fraction as described previously [33]. Again, 50 µg of protein was used for enzymatic assays from each subcellular fraction. Only preparations that had less than 20% of proteins anticipated to be associated with other compartments or cytosolic enzymatic activity were selected for analyses.

JNK activity assay

We have previously used a luciferase-based kinase activity assay (Promega's Kinase-Glo Assay) to determine cellular JNK activity [33,44]. In this study, mitochondria were isolated from each cell type, and the organelles were lysed using a native lysis buffer (25 mM Tris–HCl [pH 7.5], 150 mM NaCl, 1 mM EGTA, 1 mM EDTA, 1% Triton X-100) supplemented with protease and phosphatase inhibitors (Halt Cocktails — Thermo Fisher Scientific). Next, the mitochondrial lysate was placed in JNK activity buffer (25 mM HEPES, pH 7.4, 10 mM MgCl2, 2 mM dithiothreitol (DTT), 1 mg/ml BSA, and 1 µM ATP) containing 1 µM of peptide substrate: either c-Jun (1-79) peptide, recombinant Sab polypeptide (335–435), or mutant Sab polypeptide (335–435) with mutated JNK binding-sites to prevent kinase docking (SabKIM1/2L-A(335–435)). Sab peptides were expressed and purified as previously described [33,45]. As a control for JNK activity, specific lysates were pretreated 15 min with 10 µM SR-3306 (Tocris) to inhibit JNK. The assays were incubated for 1 h at 30°C, and the reaction was stopped by the addition of 50 mM EDTA. The reaction was combined with an equivalent volume of Kinase-Glo reagent and incubated at room temperature for 10 min. Luminescence was measured on the BioTek Synergy H1 microplate reader with 500 ms integration. ATP concentrations were determined by interpolating results onto an ATP standard curve acquired during each kinase reaction replicate. The data are reported as mean relative luminescence and normalized to protein concentration (mg) with error bars representing one standard deviation of the mean.

Apoptotic priming assay

To examine the potential for OC cell lines to induce apoptosis, we employed dynamic BH3-profiling, a high-throughput technique to predict the responsiveness of tumors to therapy [23,26,29,46]. 2.0 × 104 cells were used for the analysis of each cell line. The cells were trypsinized and dispersed into single cell suspensions, as determined by microscopy, in Trehalose Experiment Buffer [T-EB, 300 mM Trehalose, 10 mM HEPES-KOH (pH 7.7), 80 mM KCl, 1 mM EGTA, 1 mM EDTA, 0.1% BSA, and 5 mM succinate] at four times the final density. Bim BH3 peptide (MRPEIWIAQELRRIGDEFNA) was prepared as 10 distinct concentrations over a three log margin ranging in concentrations from 0.01 to 100 mM to determine the optimal conditions of the assay. A concentration of 500 nM was determined to be the optimal concentration of Bim BH3 assays in HeLa cells and later used for our profiling of OC cell lines. Next, 15 µl of Bim BH3 peptide was placed in a black 384-well plate (Nunc 262 260). One volume (7.5 µl) of the cell suspension was combined with one volume (7.5 µl) of JC-1/Digitonin Mastermix (4 mM JC-1, 40 µg/ml oligomycin, 20 mM β-mercaptoethanol, and 0.02% digitonin in T-EB). The cells were permeabilized for 10 min at room temperature. The 15 µl of cell/dye solution was then added to the Bim BH3 peptide in the respective well of the 384-well plate. When completed, the plate was placed inside a BioTek Synergy H1 plate reader and shaken for 15 s. The fluorescence was then monitored at 590 nM (excitation 545 nM; with 20 nM bandwidths) every 5 min for 120 min. As a positive control for complete depolarization, cells were treated with 3 µM carbonyl-cyanide-p-trifluoro-methoxy-phenylhydrazone (FCCP) and 250 nM valinomycin. For each biological replicate, experimental conditions were performed in replicates of five for each cell line; experiments with a standard deviation of less than 10% were used for our analysis. The percent priming (% priming) was found by subtracting the mean maximum depolarization (FCCP/Valinomycin) from the mean depolarization from Bim BH3 peptide treatment across replicates. The change in priming (Δ% priming) refers to the difference in priming between non-treated, and Bim BH3 peptide-treated cells [22].

Mitochondrial outer membrane permeabilization

To evaluate the integrity of the MOM, we chose to quantify the amount of cytosolic cytochrome c [36]. Extra-mitochondrial cytochrome c concentrations were assessed using immunoblotting of subcellular fractions (mitochondria and S100 (cytosol)). Briefly, 5 × 107 cells were grown in two 150-mm dishes, and the cells were resuspended in 800 µl of ice-cold homogenization buffer (250 mM sucrose, 20 mM HEPES, 10 mM KCl, 1.5 mM MgCl2, 1 mM EDTA, 1 mM EGTA, 1 mM DTT, pH 7.4 and Halt protease inhibitor cocktail). The cells were homogenized and centrifuged (750×g, 10 min, 4°C) to remove nuclei and intact cells. The supernatant was centrifuged (10 000×g, 10 min, 4°C). The resulting supernatant was transferred to a new tube, and the pellet (or mitochondria-containing fraction) was resuspended in homogenization buffer. The supernatant was centrifuged at 10 000×g (4°C) for 60 min yielding the cytosolic fraction. Finally, 50 µg of protein was resolved by SDS–PAGE, and cytochrome c was detected by immunoblotting as described above.

Manipulation of Sab-mediated signaling

To evaluate the role of Sab-mediated signaling in chemo-sensitivity, we transiently transfected SK-OV-3 (a low Sab expressing OC line) with plasmids designed to express Sab (pLOC:Sab), red fluorescent protein (RFP; pLOC:RFP), or a Sab mutant lacking MAPK-binding motifs (SabKIM-2L/A; pLOC: SabKIM-2L/A) [35,36]. Plasmid DNA and FugeneHD (Promega) were combined in Optimem (Invitrogen) at a ratio of 1:3 and incubated for 15 min at RT before addition to culture. Eight hours after the addition of the transfection complex to media, the media were exchanged. Protein levels were assessed at 72 h post-transfection. For drug-related studies, chemotherapeutic agents were added at 48 h post-transfection, and cell viability was measured at 96 h post-transfection (48 h after drug).

To reduce Sab-mediated signaling in a high Sab-expressing OC cell line (PA-1), we used peptide-based inhibition of protein–protein interactions on Sab as described in our prior studies. Cells were treated with up to 10 µM of Tat-SabKIM1 (a cell-permeable, inhibitory peptide) [33], scrambled control peptide (Tat-Scramble), or a variant of Tat-SabKIM1 lacking critical MAPK-binding residues (Tat-SabKIM1L/A). Cells were plated as described above and then treated with peptide for 30 min before drug treatment.

To decrease Sab expression, PA-1 cells were transiently transfected with pLKO.1 plasmids containing shRNAs either a luciferase-specific shRNA (control) or a shRNA for Sab that demonstrated over 85% silencing in 72 h (Supplementary Figure S3). Ectopic expression of a shRNA-resistant form of Sab (previously described in [35]) was used to rescue Sab levels in PA-1 cells silencing Sab. Briefly, 1.5 × 104 cells were plated in a 96-well plate, or 1.5 × 105 cells were plated a day before transfection in 35-mm dishes for experiments. Plasmids were mixed with FugeneHD (Promega) at a 3:1 ratio per manufacturer's recommendations and added to cells for up to 72 h. Protein levels were determined by immunoblotting as described above.

Manipulation of JNK expression and signaling

To reduce JNK signaling in OC cell lines, we either silenced JNK expression with shRNAs targeting JNK isoforms 1 and 2 or used the potent JNK-selective inhibitor (SR-3306). First, shRNAs for JNK1/2 were produced by cloning a conserved region of the cDNA encoding a portion of subdomain I of the kinase (N-LKPIGSGA-C; 5′-atttaaaa cctataggct caggagct-3′) with a loop sequence of 5′-uuauucaagauaa-3′ into the pLKO.1 vector. The plasmids were transiently transfected into either SK-OV-3 cells or PA-1 cells using Lipofectamine 2000 (Invitrogen) at a 4:1 ratio. Silencing was assessed at 120 h post-transfection by western blot analyses to assure knockdown. Relative levels of protein were compared endogenous proteins revealed by Ponceau S staining as described above (see Immunoblotting section).

For inhibition of JNK kinase activity, we introduced the pan-isoform, JNK selective inhibitor SR-3306 to prevent JNK phosphorylation of substrates throughout the cell. SR-3306 was prepared by dilution in DMSO, which serves as a vehicle loading control. OC cells were treated with 500 nM SR-3306 as described in the experiments below (or other doses as mentioned in the experiments). Incubation times ranged from 24 to 120 h depending on the particular experiment.

To selectively impair, the JNK–Sab interaction, we used the Tat-SabKIM1 peptide, which was described in our previous work [33]. Cells were treated with either 5 µM of the Tat-SabKIM1 peptide or a scrambled control (Tat-Scramble) for the indicated times in the experiments below. The effectiveness of the Tat-SabKIM1 peptide in PA-1 cells was verified by monitoring JNK translocation to mitochondria following treatment with staurosporine (STS) (Supplementary Figure S9).

Cell viability

To determine the extent of cell death, we utilized two methods: (1) Annexin-V with propidium iodide (PI) analyzed by fluorescent microscopy and (2) Caspase 3/7 fluorescent activity assay [4749]. For the Annexin-V/PI assays (Cayman Chemical), 1.5 × 104 cells were plated in a black-walled optically clear bottom 96-well plate (Nunc), and the cells were treated with the chemotherapy agents as described above. The plate was centrifuged at 400×g for 5 min, and the media were removed. The cells were then placed in 100 µl of Binding Buffer; the plate was centrifuged again (400×g, 5 min) and the buffer was removed. The cells were then incubated in 50 ml of Annexin-FITC/PI solution for 10 min at room temperature, and the solution was removed following centrifugation (400×g, 5 min). The cells were placed in 100 µl of HBSS, and the cells were imaged by fluorescent microscopy using the Applied Precision DeltaVision Elite Imaging System. Images were analyzed by counting cells stained only with unstained, Annexin-V-FITC-stained (early apoptosis), and double-labeled, Annexin-V/PI-stained cells (late apoptosis). As a positive control, cells were treated with 1 µM STS to induce significant cell death at 24 h. Images were assembled using Adobe Photoshop. For the cell-based caspase 3/7 activity assay, cells were grown in a clear 96-well plate as described for the Annexin-V/PI assay. The plate was centrifuged (400×g) for 5 min, and the supernatant was removed. The cells were placed in 150 µl of PBS and centrifuged (400×g for 5 min). After removing the buffer, the cells were placed in 100 µl of PBS supplemented with 1% Triton X-100 to lyse the cells. After 30 min of incubation on a room temperature orbital shaker, the plate was centrifuged at 800×g for 10 min, and 90 µl of the cleared lysate was transferred to a black 96-well plate. Next, either 10 µl of PBS or 10 µl of a caspase inhibitor (Ac-DEVD-CHO; 10 µM) was added to the lysate. Then, 100 µl of caspase substrate Ac-DEVD-AMC (100 µM) in PBS supplemented with 20 mM DTT was added to the well. The assay was incubated at 37°C for 30 min, and fluorescence was measured (excitation: 488 nm; emission: 535 nm) using the BioTek Synergy H1 plate reader. Each assay was performed in triplicate for each condition for a minimum of four biological replicates. Data are presented as mean relative fluorescent units.

In-cell western analysis

Black-walled optically clear bottomed 96-well plate (Thermo Scientific #H2861653050) were seeded with 200 µl of a mixture of trypsinized cells and media, with a count of 15 000 cells per well. Three hours after plating cells were treated with either 0.1% DMSO or a mixture of 20 nM cisplatin and 100 nM paclitaxel for 72 h under standard culture parameters. Cells were then fixed with 4% paraformaldehyde (Santa Cruz Biotechnology #SC-281692) in PBS for 20 min at RT. OC cells were next quenched with 100 mM glycine for 5 min at RT. Cells were permeabilized in PBS supplemented with 0.2% Triton X-100, gently rocking for 20 min at RT. After permeabilization, the cells were blocked with 150 µl of LI-COR blocking buffer (LI-COR Biosciences, 927–40 100), gently rocking the plate for 90 min at RT. Sab antibody was diluted in blocking buffer at a ratio of 1:1000 and 100 µl of this solution was added to each well. The plate was incubated on a rocker for 2.5 h at RT and washed with 100 µl of PBST (PBS + 0.1% Tween-20) for 5 min. This was repeated 5 times. A 1:800 ratio of anti-mouse secondary antibody (IRDye 800CW anti-mouse, LI-COR Biosciences, 926–32 210) and 1:1000 ratio of 1 mM of TO-PRO-3 (Invitrogen) were diluted in blocking buffer to make the secondary antibody solution. Cells were incubated with 100 µl of secondary antibody solution in each well for 45 min. The plate was washed five times with PBST. The wash buffer was removed, and the plate was scanned in Odyssey scanner, and the protein levels and TO-PRO-3 fluorescence were quantified using the LI-COR in-cell western (ICW) analysis in the Image Studio Software (LI-COR Biosciences).

Biological replicates and statistics

A minimum of five biological replicates were used for cell-based studies, while a minimum of six experimental replicates was evaluated for biochemical, fluorescence, and other measurements. The Mann–Whitney test was employed to determine differences between treatments. Statistical significance is indicated by an asterisk in figures in which the P-value is less than 0.05. Data are displayed as means with error bars representing plus and minus one standard deviation of the mean.

Results

Sab expression is reduced in ovarian cancer

Our previous study in HeLa cells demonstrated that increasing the concentration of Sab on the MOM enhanced the sensitivity towards cisplatin and paclitaxel [35]. Because OC is typified by the presence of tumor cells resistant to platinum drugs, such as cisplatin, we wanted to determine if Sab expression may be altered in OC patients. To determine if Sab expression was changed in OC, we queried the Oncomine database [37] for Sab expression in OC samples and normal ovarian tissue (as identified in Table 1). Examination of six independent studies (Table 1) found that Sab levels were reduced on average ∼5.5-fold (2.19-fold in log2 scale) in primary site OC tumors when compared with normal tissue [3843]. We plotted the medians, 90th percentile values, and 10th percentile values for each study in Figure 1A. This analysis of 970 OC samples had a P-value less than 0.001 indicating a difference in Sab expression, at least at the mRNA level, between the healthy tissue and OC tumors (Table 1). For studies that had information on tumor grade and patient stage, we found that Sab expression was decreased beginning in lowest grade tumors and the earliest stage patients (Figure 1B,C). Analysis of Sab expression with respect to survival revealed that there was no difference between patients that survived or succumbed to disease over 5 years (Figure 1D). Because of the consistent lower expression of Sab, we examined the impact of Sab protein levels on OC cell physiology.

Summary of Sab expression levels in six OC gene expression studies.

Figure 1.
Summary of Sab expression levels in six OC gene expression studies.

(A) Six studies were identified in the Oncomine repository that matched our criteria of a 2-fold decrease in Sab expression and a P-value of less than 10−4. The critical data of these studies can be found in Table 1. We plotted the median value for Sab expression in healthy tissue (black circles) and ovarian cancer samples (gray squares) for each study, and the error bars represent the 90th and 10th percentiles for each study. An asterisk (*) is used to indicate studies with a P-value <0.005. (B) Data from the gene expression studies was used to illustrate Sab mRNA levels in tumors of different grades. (C) Sab expression levels were also examined in the context of FIGO Stage. The data in Figure 1B,C are presented as box plots with the maximum and minimum serving as the upper and lower limits, respectively, and the mean is represented by the horizontal line between the maximum and minimum. An asterisk (*) is used to indicate studies with a P-value <0.005 when compared with healthy tissue controls. (D) The relative abundance of Sab mRNA was also examined with respect to the extent of patient survival. Survival (white and blue bars) and deaths (gray and yellow bars) for years 1, 3, and 5 following treatment along with the absence (white and gray bars) or presence (blue and yellow bars) of disease are documented.

Figure 1.
Summary of Sab expression levels in six OC gene expression studies.

(A) Six studies were identified in the Oncomine repository that matched our criteria of a 2-fold decrease in Sab expression and a P-value of less than 10−4. The critical data of these studies can be found in Table 1. We plotted the median value for Sab expression in healthy tissue (black circles) and ovarian cancer samples (gray squares) for each study, and the error bars represent the 90th and 10th percentiles for each study. An asterisk (*) is used to indicate studies with a P-value <0.005. (B) Data from the gene expression studies was used to illustrate Sab mRNA levels in tumors of different grades. (C) Sab expression levels were also examined in the context of FIGO Stage. The data in Figure 1B,C are presented as box plots with the maximum and minimum serving as the upper and lower limits, respectively, and the mean is represented by the horizontal line between the maximum and minimum. An asterisk (*) is used to indicate studies with a P-value <0.005 when compared with healthy tissue controls. (D) The relative abundance of Sab mRNA was also examined with respect to the extent of patient survival. Survival (white and blue bars) and deaths (gray and yellow bars) for years 1, 3, and 5 following treatment along with the absence (white and gray bars) or presence (blue and yellow bars) of disease are documented.

Ovarian cancer cell lines have distinct levels of Sab expression and mitochondrial JNK signaling

To determine if the diminished Sab expression in the patient OC samples affects chemosensitivity, we used OC cell lines to examine Sab's role in chemo-responsiveness directly. First, we measured Sab protein levels in four commercially available OC cell lines, Caov-3, PA-1, SK-OV-3, and SW-626; human ovarian epithelial cells were used as a normal tissue control. Immunoblot analysis of Sab reveals that Caov-3 and SK-OV-3 lines have lower levels of Sab than that of ovary epithelial cells. Caov-3 and SK-OV-3 had 33% and 48% less Sab, respectively (Figure 2A,B). Meanwhile, the concentration of Sab was significantly elevated in PA-1 and SW-626 cells (Figure 2A, top panel); PA-1 had a mean increase in Sab levels of 6.7-fold that of ovarian epithelial cells, SW-626 was 4.2-fold higher (Figure 2A,B). α-Tubulin was used as a cellular loading control (Figure 2A, middle panel), while COX-IV was used as a control for mitochondrial density (Figure 2A, bottom panel). Ponceau S staining of membranes was used to normalize protein levels among cell types (Figure 2A). The results of four immunoblotting experiments were quantified and normalized to Ponceau S in Figure 2B.

The levels of Sab differ among OC cell lines and correspond to stress-induced mitochondrial JNK recruitment.

Figure 2.
The levels of Sab differ among OC cell lines and correspond to stress-induced mitochondrial JNK recruitment.

(A) OC cell lines were grown under standard cell culture conditions to ∼80% confluency. The cells were lysed, and the proteins were resolved by SDS–PAGE. Immunoblotting was performed to determine the relative abundance of Sab with a-Tubulin and COX-IV serving as cellular and mitochondrial loading controls. Ponceau staining of membranes was used to quantify the immunoblotting results. (B) Quantification of Sab abundance from whole cell lysates in Figure 2A. (C) Mitochondrial isolates were prepared from the four OC cell lines and assessed for the levels of Sab by western blot analysis. TOM20 was used as an outer mitochondrial loading control, while the relative contamination of other cellular compartments was assessed by monitoring the levels of Calnexin (ER), Histone H3 (nucleus), PEX-19 (peroxisomes), and LAMP2 (lysosomes). (D) The levels of Sab for each OC cell line were normalized to mitochondria abundance (TOM20 levels) and quantified for seven independent biological replicates. (E) The relative abundance of activated JNK on mitochondria was assessed in the presence and absence of stress. The top panel illustrates the relative levels of total JNK (JNK), activated JNK (P-JNK), with respect to Sab on mitochondria isolated from untreated OC cells. The bottom panel shows the same proteins following treatment with 1 mM STS for 45 min. TOM20 was used as a mitochondrial loading control for both panels. (F) In vitro kinase activity assays for JNK were performed on mitochondria using either a c-Jun (1-79) or a Sab (335–435) peptide containing JNK phosphorylation sites. Additionally, a Sab-peptide lacking JNK binding residues (SabKIM1/2L-A(335–435) was used as a negative control substrate. SR-3306 (500 nM), a potent JNK inhibitor, was used to impair JNK activity during cellular stress and prevent JNK translocation to mitochondria. An asterisk (*) is used to indicate studies with a P-value <0.01 with respect to controls; meanwhile a double asterisk (**) is used to demark differences between cell types and distinct treatments.

Figure 2.
The levels of Sab differ among OC cell lines and correspond to stress-induced mitochondrial JNK recruitment.

(A) OC cell lines were grown under standard cell culture conditions to ∼80% confluency. The cells were lysed, and the proteins were resolved by SDS–PAGE. Immunoblotting was performed to determine the relative abundance of Sab with a-Tubulin and COX-IV serving as cellular and mitochondrial loading controls. Ponceau staining of membranes was used to quantify the immunoblotting results. (B) Quantification of Sab abundance from whole cell lysates in Figure 2A. (C) Mitochondrial isolates were prepared from the four OC cell lines and assessed for the levels of Sab by western blot analysis. TOM20 was used as an outer mitochondrial loading control, while the relative contamination of other cellular compartments was assessed by monitoring the levels of Calnexin (ER), Histone H3 (nucleus), PEX-19 (peroxisomes), and LAMP2 (lysosomes). (D) The levels of Sab for each OC cell line were normalized to mitochondria abundance (TOM20 levels) and quantified for seven independent biological replicates. (E) The relative abundance of activated JNK on mitochondria was assessed in the presence and absence of stress. The top panel illustrates the relative levels of total JNK (JNK), activated JNK (P-JNK), with respect to Sab on mitochondria isolated from untreated OC cells. The bottom panel shows the same proteins following treatment with 1 mM STS for 45 min. TOM20 was used as a mitochondrial loading control for both panels. (F) In vitro kinase activity assays for JNK were performed on mitochondria using either a c-Jun (1-79) or a Sab (335–435) peptide containing JNK phosphorylation sites. Additionally, a Sab-peptide lacking JNK binding residues (SabKIM1/2L-A(335–435) was used as a negative control substrate. SR-3306 (500 nM), a potent JNK inhibitor, was used to impair JNK activity during cellular stress and prevent JNK translocation to mitochondria. An asterisk (*) is used to indicate studies with a P-value <0.01 with respect to controls; meanwhile a double asterisk (**) is used to demark differences between cell types and distinct treatments.

To evaluate if the differences in Sab protein levels among the four cell lines occurred on the MOM or elsewhere in the cell, we performed subcellular fractionation to isolate mitochondria and other subcellular compartments [3335]. Immunoblotting of subcellular fractions reveals that Sab is exclusively in the mitochondrial fraction for each of the four OC cell lines as compared with nuclear, cytosolic, and microsomal fractions (Figure 2C). We were unable to detect Sab at substantial levels in the ER, cytosolic or nuclear preparations (Supplementary Figure S1). Furthermore, the relative purity of each compartment fraction was assessed by monitoring the levels of nuclear (Histone H3), ER (Calnexin), peroxisomal (PEX19), and lysosomal (LAMP2) contamination in the mitochondrial isolates (Figure 2C). Only preparations with greater than 80% purity were used in our analysis according to protein analyses (Figure 2C and Supplementary Figure S1) and the relative activities of cytosolic glyceraldehyde-3-phosphate (GAPDH) and LDH and CS in discrete subcellular compartments (Supplementary Figure S1). The relative levels of Sab were quantified and presented in Figure 2D.

Extensive research demonstrates that mitochondrial translocation of JNK is highly dependent upon direct interactions with Sab. To examine if the relative Sab levels in the OC cells reflected mitochondrial JNK levels, we performed an immunoblot for JNK. JNK was not detected in any of the OC cell lines under standard culture conditions (Figure 2E). Equivalent mitochondrial loading and MOM loading in each sample were assessed by the signal intensity of COX-IV and TOM20, respectively, (Figure 2E). Because previous research indicates that mitochondrial translocation of JNK is dependent upon activation, we treated each of the four OC lines with 1 µM STS for 45 min and measured phosphorylated (active) JNK (Thr183/Tyr185) levels in mitochondrial extracts. Upon treatment with 1 µM STS, phospho-JNK levels on mitochondria increased in all four cell lines (Figure 2E); this was reflected by an increase in the amount of total JNK in the mitochondrial isolates (Figure 2E). The amount of JNK localized on mitochondria of SK-OV-3 and Caov-3 cells were 62% and 47% less of the amount of JNK on the mitochondria of PA-1 cells (Figure 2E; top panel). Also, SW-626 mitochondrial JNK levels were 21% less than JNK on PA-1 mitochondria (Figure 2E). Immunoblot analysis of whole cell extracts demonstrates that there was no significant difference in JNK abundance among the four OC cell types (Figure 2A), suggesting that the amount of JNK on mitochondria was dependent upon the concentrations of Sab. As observed in our previous studies only the higher molecular mass (54 kDa) JNK band was present on mitochondria (Figure 2E; top panel); accordingly, this band cross-reacted with the phospho-JNK-specific antibody in mitochondrial preps from each of the four OC cell lines (Figure 2E, top panel). Quantitation of western blot analyses can be found in Supplementary Figure S2.

To determine if the JNK present on mitochondria following STS treatment was indeed active, we lysed mitochondria isolated from each of the four OC cell lines in non-denaturing conditions and then measured JNK activity towards c-Jun and Sab peptides as well as a Sab mutant peptide (SabKIM1/2L-A(335–435)) incapable of binding JNK. Following exposure to 1 µM STS, JNK activity increased on mitochondria from all four cell lines (Figure 2F). However, JNK activity towards c-Jun and Sab peptides was increased over 10-fold in PA-1 and SW-626 cells, while only a 2.5- and 3.5-fold increase in activity was observed in SK-OV-3 and Caov-3 cells, respectively (Figure 2F). No kinase activity was observed with the SabKIM1/2L-A(335–435) peptide (Figure 2F), a peptide used previously for JNK activity assays. We have also found that JNK can phosphorylate the peptide primarily on Ser421 (unpublished data). The activity could be attributed to JNK as the introduction of 500 nM JNK specific inhibitor SR-3306 prevented kinase activity in the assay (Figure 2F). These studies reinforce that JNK activation drives translocation of the kinase to mitochondria.

To determine if JNK translocation was specific to STS, we treated cells with 25 µM anisomycin for 30 min and measured mitochondrial JNK levels and activity. JNK levels increased in all four OC cell lines, with higher amounts of JNK present on the mitochondria of PA-1 and SW-626 compared with SK-OV-3 and Caov-3 (Supplementary Figure S3). The JNK species on mitochondria cross-reacted with the phospho-specific JNK antibody (Supplementary Figure S3). Similarly, the activity of JNK towards c-Jun and Sab peptides in the mitochondrial fractions was greater in PA-1 and SW-626 cell compared with SK-OV-3 and Caov-3 (Supplementary Figure S3). SR-3306 was used to demonstrate that JNK was the kinase responsible for the activity. Taken together, these results demonstrate that Sab concentrations on the MOM of OC cells may dictate the local magnitude of JNK signaling.

Ovarian cancer cell lines with low Sab levels are resistant to chemotherapy agents

Because mitochondrial JNK signaling on Sab is linked to apoptosis [33,5055], we measured the chemosensitivity of the four OC cell lines to determine if Sab levels correlated to chemo-responsiveness. First, we examined the potency of paclitaxel/cisplatin treatment, an approach employed for resistant and refractory tumors [5]. SK-OV-3 cells and Caov-3 cells had considerably greater cellular viability as assessed by TO-PRO-3 fluorescence following cisplatin/paclitaxel treatment when compared with PA-1 and SW-626 cells (Figure 3A,B). Next, using conventional and clinically relevant chemotherapeutic agents, we assessed the potency of each with TO-PRO-3 staining and near-infrared imaging (demonstrated in Supplementary Figure S4). The data presented in Table 2 summarizes the drugs and IC50 values for each of the cells lines. SK-OV-3 and Caov-3 consistently had higher IC50 values than PA-1 and SW-626 cells. Specifically, SK-OV-3 and Caov-3 had IC50 values greater than 10-fold higher than PA-1 and SW-626 when treated with ABT-737, ABL-263, TW-37, and Obatoclax mesylate (Table 2). To determine if Sab levels correlated to the cellular sensitivities towards chemotherapeutic agents, we derived the Pearson's coefficient for Sab levels (from Figure 2A) in the cell lines (Supplementary Figure S5). Sab levels were strongly correlated to chemosensitivity across all four cell lines. The correlation curves and one-sided Pearson's coefficients can be found in Supplementary Figure S5 for each drug presented in Table 2. To determine if the trends were a result of the small sample size, we incorporated six additional cell lines (OVCAR3, OVCAR8, ES-2, OV-21G, OV-90, and OV-112D) into our study. In Figure 3C, we examined the relative abundance of Sab for each cell line using western blot analysis, and we found that the cells had comparable levels of Sab without noticeable differences in the levels of COX-IV (mitochondria) or tubulin. We then examined the cellular responses to a 72-hour treatment with 25 nM cisplatin and 100 nM paclitaxel using an ICW. The ICW analysis revealed that all six cells had similar levels of Sab and were sensitive to the combination of cisplatin and paclitaxel (Figure 3D). Therefore, we calculated the levels of Sab for each cell line and normalized the relative fluorescence to human ovarian epithelial cell Sab levels (Figure 3E). Again, the six additional cell lines had similar levels of Sab compared with PA-1 and SW-626 cells. Furthermore, the OVCAR3, OVCAR8, ES-2, OV-21G, OV-90, and OV-112D cells had comparable IC50 values for cisplatin and paclitaxel (Figure 3E). Finally, the six additional cell lines were all similarly susceptible to treatment with 100 nM cisplatin and 100 nM paclitaxel after 48 h, and the six cell lines had comparable responses to the platinum and taxol treatment. Given the differences among cells with varying Sab levels, we examined the potential impact of Sab concentrations in low and high Sab expressing cell lines and monitored OC chemovulnerability.

Cisplatin/Paclitaxel treatment affects OC cell lines differently.

Figure 3.
Cisplatin/Paclitaxel treatment affects OC cell lines differently.

(A) TO-PRO-3 staining was used to assess the viability of OC cell types following treatment with vehicle (0.1% DMSO) or a combination of 25 nM cisplatin and 100 nM paclitaxel for 72 h. A representative experiment with duplicates of each treatment is shown to illustrate the technique. (B) The quantified results of the cell viability are shown for each OC cell line. The vehicle is the left bar, while the right bar represents the relative fluorescence of cells treated with cisplatin and paclitaxel. An asterisk (*) is used to indicate studies with a P-value <0.01. (C) Six additional OC cell lines were analyzed for relative Sab levels using western blot analysis. COX-IV was used as a mitochondrial control, and tubulin was used as a cellular loading control. (D) In-Cell western analysis was performed on the six cells lines to detect the relative changes in Sab levels and cell viability (TO-PRO-3) in the presence and absence of 25 nM cisplatin and 100 nM Paclitaxel (Pt/Taxol) or a vehicle (DMSO) control. (E) The relative concentrations of Sab was determined for each cell line based on Sab concentrations in human ovarian epithelial cells, IC50 values for cisplatin and paclitaxel were calculated for each of the remaining six cell lines, and the relative viability for each cell line was determined by comparing cells treated with cisplatin/paclitaxel for 72 h to cells treated with DMSO. The values presented in the table are means plus/minus one standard deviation.

Figure 3.
Cisplatin/Paclitaxel treatment affects OC cell lines differently.

(A) TO-PRO-3 staining was used to assess the viability of OC cell types following treatment with vehicle (0.1% DMSO) or a combination of 25 nM cisplatin and 100 nM paclitaxel for 72 h. A representative experiment with duplicates of each treatment is shown to illustrate the technique. (B) The quantified results of the cell viability are shown for each OC cell line. The vehicle is the left bar, while the right bar represents the relative fluorescence of cells treated with cisplatin and paclitaxel. An asterisk (*) is used to indicate studies with a P-value <0.01. (C) Six additional OC cell lines were analyzed for relative Sab levels using western blot analysis. COX-IV was used as a mitochondrial control, and tubulin was used as a cellular loading control. (D) In-Cell western analysis was performed on the six cells lines to detect the relative changes in Sab levels and cell viability (TO-PRO-3) in the presence and absence of 25 nM cisplatin and 100 nM Paclitaxel (Pt/Taxol) or a vehicle (DMSO) control. (E) The relative concentrations of Sab was determined for each cell line based on Sab concentrations in human ovarian epithelial cells, IC50 values for cisplatin and paclitaxel were calculated for each of the remaining six cell lines, and the relative viability for each cell line was determined by comparing cells treated with cisplatin/paclitaxel for 72 h to cells treated with DMSO. The values presented in the table are means plus/minus one standard deviation.

Table 2
Cell-based IC50 values (µM) for chemotherapeutic agents in OC cell lines

IC50 values are presented as means plus or minus one standard deviation of the mean. The IC50s were determined using TO-PRO-3 staining and were confirmed by cell death assays. For each assay, each dose over the range of 10 nM to 1 mM was performed in quadruplet including the 0.01% DMF control. The means and standard deviations were based on a minimum of three biological replicates.

Chemotherapy Caov-3 PA-1 SK-OV-3 SW-626 
Paclitaxel 9.7 ± 3.1 0.11 ± 0.05 13.7 ± 4.6 0.57 ± 0.13 
Docetaxel 0.45 ± 0.19 <0.01 1.22 ± 0.38 0.01 ± 0.004 
Doxorubicin 0.41 ± 0.13 0.04 ± 0.008 0.5 ± 0.11 0.11 ± 0.007 
Etoposide 28.9 ± 8.17 0.13 ± 0.08 10.9 ± 3.33 8.1 ± 2.67 
Carboplatin 48 ± 12.17 1.9 ± 0.93 65 ± 17.6 5.8 ± 2.33 
Cisplatin 34.5 ± 63.3 0.79 ± 0.16 60.2 ± 72.7 3.2 ± 1.61 
ABT-263 189 ± 49.1 0.96 ± 0.33 215 ± 67.6 3.2 ± 0.86 
ABT-727 8.9 ± 2.67 0.09 ± 0.03 15 ± 4.33 0.47 ± 0.16 
Obatoclax 1.4 ± 0.67 0.07 ± 0.01 0.6 ± 0.17 0.09 ± 0.02 
TW-37 1.69 ± 0.81 0.1 ± 0.05 0.8 ± 0.23 0.4 ± 0.13 
Chemotherapy Caov-3 PA-1 SK-OV-3 SW-626 
Paclitaxel 9.7 ± 3.1 0.11 ± 0.05 13.7 ± 4.6 0.57 ± 0.13 
Docetaxel 0.45 ± 0.19 <0.01 1.22 ± 0.38 0.01 ± 0.004 
Doxorubicin 0.41 ± 0.13 0.04 ± 0.008 0.5 ± 0.11 0.11 ± 0.007 
Etoposide 28.9 ± 8.17 0.13 ± 0.08 10.9 ± 3.33 8.1 ± 2.67 
Carboplatin 48 ± 12.17 1.9 ± 0.93 65 ± 17.6 5.8 ± 2.33 
Cisplatin 34.5 ± 63.3 0.79 ± 0.16 60.2 ± 72.7 3.2 ± 1.61 
ABT-263 189 ± 49.1 0.96 ± 0.33 215 ± 67.6 3.2 ± 0.86 
ABT-727 8.9 ± 2.67 0.09 ± 0.03 15 ± 4.33 0.47 ± 0.16 
Obatoclax 1.4 ± 0.67 0.07 ± 0.01 0.6 ± 0.17 0.09 ± 0.02 
TW-37 1.69 ± 0.81 0.1 ± 0.05 0.8 ± 0.23 0.4 ± 0.13 

Ovarian cancer cell mitochondria with high Sab concentrations are primed for apoptosis

Because of the Sab-mediated JNK signaling is linked to Bcl-2 levels on mitochondria [33,51,55], we measured the apoptotic potential in the four OC cell lines. To examine the extent of apoptotic priming in each of the cancer cell lines, we first performed dynamic BH3 profiling [22,27,29] to assess the relative extent of priming. The addition of Bim BH3 peptide caused a rapid depolarization of mitochondrial membrane potential in all four OC cell lines; however, the loss of JC-1 fluorescence was slower in SK-OV-3 cells than PA-1 (Figure 4A). Additionally, the PA-1 and SW-626 cells with greater Sab abundance were found to have a greater percent depolarization than Caov-3 and SK-OV-3 cells that have relatively low Sab levels (Figure 4B). Consequently, the PA-1 and SW-626 cells had a more significant change in priming (Δ%Priming) compared with Caov-3 and SK-OV-3 cells (Figure 4C). Next, we verified the results of the dynamic BH3 profiling by measuring the cytosolic levels of cytochrome c by immunodetection within cytoplasmic subcellular fractions in the presence and absence of 500 nM Bim-BH3 peptide (Figure 4D; quantified in 4E). The extent of cytochrome c present in the cytosol of OC cell lines was modest compared with cells treated with 1 µM STS for 2 h (Figure 4D,E). OC cells with low Sab levels (Caov-3 and SK-OV-3) had decreased cytochrome c release into the cytosol compared with cells with higher Sab concentrations, PA-1, and SW-626. To examine if the loss of mitochondrial membrane integrity was indeed due to apoptotic priming, the relative levels of Bcl-2 proteins (Bcl-2, Bcl-xL, and Mcl-1), BH3-only proteins (Bid, Bik, Bim, Bad, and Puma), and pro-apoptotic proteins (Bax and Bak) were measured by immunoblotting mitochondria from OC cell lines (Figure 4F). The quantified data presented in Supplementary Figure S6 indicate that PA-1 and SW-626 cells have higher levels of pro-apoptotic BH3-only proteins than Caov-3 and SK-OV-3 cells, which have higher levels of pro-survival Bcl-2 proteins. This may indicate that ovarian cancer cells with elevated Sab levels may be primed for apoptosis.

OC cells with high Sab concentrations are primed for apoptosis and chemosensitive.

Figure 4.
OC cells with high Sab concentrations are primed for apoptosis and chemosensitive.

(A) Dynamic BH3 profiling was used to evaluate the extent of mitochondrial priming between PA-1 cells with high Sab levels and SK-OV-3 cells with relatively low Sab abundance. Cells were treated with vehicle (0.1% DMSO) or 500 nM BIM BH3-only peptide, and JC-1 fluorescence was monitored. FCCP/Valinomycin was used to depolarize mitochondria in this assay. (B) The percent depolarization and (2) the change relative priming (Δ%Priming) were calculated based on the dynamic BH3 profiling for each of the OC cell lines. An asterisk (*) is used to indicate studies with a P-value <0.01. (D) The abundance of cytosolic cytochrome c was measured using western blot analysis of cytosolic fractions taken from OC cells in the absence (top panel) and presence of 500 nM BIM peptide (bottom panel). GAPDH was used as a cytosolic loading control. The relative fluorescence of cytosolic cytochrome c was quantified (right panel). An asterisk (*) is used to indicate studies with a P-value <0.01. (E) Pro-survival Bcl-2 protein levels were monitored on isolated mitochondria from each of the four OC cells lines; individually, Bcl-2, Bcl-xL, and Mcl-1 levels were measured, while TOM20 served as a mitochondrial loading control. (F) Pro-apoptotic protein levels on mitochondrial were measured by western blot analysis. Bim, Bad, Bak, and Bax were detected on isolated mitochondria from the OC cell lines, and COX-IV was used as a mitochondrial loading control.

Figure 4.
OC cells with high Sab concentrations are primed for apoptosis and chemosensitive.

(A) Dynamic BH3 profiling was used to evaluate the extent of mitochondrial priming between PA-1 cells with high Sab levels and SK-OV-3 cells with relatively low Sab abundance. Cells were treated with vehicle (0.1% DMSO) or 500 nM BIM BH3-only peptide, and JC-1 fluorescence was monitored. FCCP/Valinomycin was used to depolarize mitochondria in this assay. (B) The percent depolarization and (2) the change relative priming (Δ%Priming) were calculated based on the dynamic BH3 profiling for each of the OC cell lines. An asterisk (*) is used to indicate studies with a P-value <0.01. (D) The abundance of cytosolic cytochrome c was measured using western blot analysis of cytosolic fractions taken from OC cells in the absence (top panel) and presence of 500 nM BIM peptide (bottom panel). GAPDH was used as a cytosolic loading control. The relative fluorescence of cytosolic cytochrome c was quantified (right panel). An asterisk (*) is used to indicate studies with a P-value <0.01. (E) Pro-survival Bcl-2 protein levels were monitored on isolated mitochondria from each of the four OC cells lines; individually, Bcl-2, Bcl-xL, and Mcl-1 levels were measured, while TOM20 served as a mitochondrial loading control. (F) Pro-apoptotic protein levels on mitochondrial were measured by western blot analysis. Bim, Bad, Bak, and Bax were detected on isolated mitochondria from the OC cell lines, and COX-IV was used as a mitochondrial loading control.

Ectopic expression of Sab induces apoptotic priming

To directly determine if Sab-mediated signaling contributes to apoptotic priming in ovarian cancer cell lines, we ectopically expressed either RFP, Sab, or a Sab mutant incapable of binding JNK (SabKIM1L-A) [35] (Figure 5A) in chemoresistant SK-OV-3 cells over the course of 72 h (Figure 5B). By 72 h, Sab was expressed 5- to 7-fold higher than SK-OV-3 cells that are expressing RFP or mock-transfected cells (Figure 5B; quantified in Figure 5C). Dynamic BH3 profiling reveals that SK-OV-3 cells expressing Sab have a higher percent depolarization and change in depolarization compared with mock-transfected and RFP-expressing cells (Figure 5D), an effect that was not observed in cells expressing SabKIM1L-A (Figure 5D). Furthermore, analysis of cytosolic cytochrome c reveals that compared with mock transfected and RFP-expressing SK-OV-3 cells, Sab expressing SK-OV-3 cells have increased cytosolic cytochrome c (Figure 5E). This increase is not observed in SabKIM1L-A-expressing SK-OV-3 cells (Figure 5E). Additionally, the mitochondrial levels of Bcl-2 family proteins were assessed (Figure 5F). Ectopic expression of Sab in SK-OV-3 cells changed the relative abundance of Bcl-2 and BH3-only proteins on mitochondria when compared with mock transfected, RFP-expressing, and SabKIM1L-A-expressing cells (Figure 5G; quantified in Supplementary Figure S6B). These results suggest that altering Sab levels on the MOM may change mitochondrial context and apoptotic potential.

Artificially elevating Sab expression in resistant OC cells increases vulnerability to chemotherapy.

Figure 5.
Artificially elevating Sab expression in resistant OC cells increases vulnerability to chemotherapy.

(A) A schematic representation of Sab showing the two leucine residues in the KIM1 motif necessary for JNK binding. The bottom panel shows a Sab-variant (SabKIM1(L-A)) generated by site-directed mutagenesis that lacks the leucine residues required for JNK binding. (B) Ectopic expression of Sab and SabKIM1(L-A) in chemoresistant SK-OV-3 cells was assessed using western blot analysis with TOM20 and Tubulin as loading controls for mitochondria and whole cell respectively. (C) The relative levels of activated JNK (P-JNK) and total JNK (JNK) on mitochondria (top panel) and in whole cell lysates (bottom panel) were assessed in SK-OV-3 cells following 45 min of 1 µM STS treatment. (D) Dynamic BH3 profiling of SK-OV-3 cells with and without ectopic expression of Sab was performed in the presence of vehicle or 500 nM BIM peptide, and treatment with FCCP/Valinomycin was used as a depolarization control. (E) The %Depolarization and (F) Δ%Priming were calculated from the BH3 profiling for SK-OV-3 cells expressing either RFP, Sab, or SabKIM1(L-A). An asterisk (*) is used to indicate studies with a P-value <0.01. (G) Western blot analysis was used to assess the relative abundance of pro-survival (left panel) and pro-apoptotic (right panel) in SK-OV-3 cells expressing either RFP, Sab, or SabKIM1(L-A).

Figure 5.
Artificially elevating Sab expression in resistant OC cells increases vulnerability to chemotherapy.

(A) A schematic representation of Sab showing the two leucine residues in the KIM1 motif necessary for JNK binding. The bottom panel shows a Sab-variant (SabKIM1(L-A)) generated by site-directed mutagenesis that lacks the leucine residues required for JNK binding. (B) Ectopic expression of Sab and SabKIM1(L-A) in chemoresistant SK-OV-3 cells was assessed using western blot analysis with TOM20 and Tubulin as loading controls for mitochondria and whole cell respectively. (C) The relative levels of activated JNK (P-JNK) and total JNK (JNK) on mitochondria (top panel) and in whole cell lysates (bottom panel) were assessed in SK-OV-3 cells following 45 min of 1 µM STS treatment. (D) Dynamic BH3 profiling of SK-OV-3 cells with and without ectopic expression of Sab was performed in the presence of vehicle or 500 nM BIM peptide, and treatment with FCCP/Valinomycin was used as a depolarization control. (E) The %Depolarization and (F) Δ%Priming were calculated from the BH3 profiling for SK-OV-3 cells expressing either RFP, Sab, or SabKIM1(L-A). An asterisk (*) is used to indicate studies with a P-value <0.01. (G) Western blot analysis was used to assess the relative abundance of pro-survival (left panel) and pro-apoptotic (right panel) in SK-OV-3 cells expressing either RFP, Sab, or SabKIM1(L-A).

Increasing Sab abundance sensitizes SK-OV-3 cells to chemotherapies

Because Sab levels may affect apoptotic priming, and we previously demonstrated that increasing Sab sensitizes HeLa cells to toxic agents [35], we examined whether ectopic expression of Sab enhanced the chemo-responsiveness of resistant SK-OV-3 cells towards cisplatin/paclitaxel treatment. To assess cell death in mock-transfected SK-OV-3 cells, cells transfected with an empty vector, or cells expressing either Sab or SabKIM1L-A, we utilized microscopic analysis of Annexin V and PI staining (Figure 6A). Compared with mock-transfected cells and empty vector cells, Sab-expressing SK-OV-3 cells had a considerable increase in the number of apoptotic cells in the presence of 25 nM cisplatin and 100 nM paclitaxel after 72 h (Figure 6B). Specifically, these cells had a marked rise in the number of late apoptotic cells (those cells stained with both Annexin V and PI – Figure 5B). SK-OV-3 cells expressing the MAPK-binding deficient mutant, SabKIM1L-A, did not differ from mock-transfected and RFP expressing SK-OV-3 cells (Figure 6B). To complement the microscopic analysis, we performed an assay to detect caspase 3 and 7 activity in the cells exposed to 25 nM cisplatin and 100 nM paclitaxel. Mock-transfected and empty vector transfected cells had little caspase activation following treatment with cisplatin and paclitaxel (Figure 6C); conversely, Sab-expressing cells had a substantial increase in caspase activity in the presence of chemotherapeutic agents, which was comparable to SK-OV-3 cells treated with 1 µM of STS for 24 h (Figure 6C). Again, expression of SabKIM1L-A in SK-OV-3 cells did not alter caspase activity levels in response to 25 nM cisplatin and 100 nM paclitaxel treatment (Figure 6C). To determine if Sab-mediated signaling was involved in cisplatin and paclitaxel-induced death in sensitive cell lines, such as PA-1, we incubated PA-1 with increasing amounts of either Tat-Scramble or Tat-SabKIM1 and measured the extent of cell death following treatment with 0.25 µM cisplatin and 0.10 µM paclitaxel (Figure 6D). Inhibition of Sab-mediated signaling with the Tat-SabKIM1 peptide prevented PA-1 cell death following cisplatin and paclitaxel administration when compared with untreated and cells treated with Tat-Scramble peptide (Figure 6D). PA-1 cells treated with a mutant Sab peptide, Tat-SabKIM1L-A, which cannot bind MAPKs did not differ from cells treated with the Tat-Scramble peptide (Figure 6D). Again, we verified the induction of apoptosis in PA-1 cells following 25 nM cisplatin and 100 nM paclitaxel treatment by using a caspase activity (Figure 6E). PA-1 cells treated with 1 µM Tat-SabKIM1 peptide had decreased caspase activity following treatment with chemotherapy agents as compared with cells treated with either 1 µM Tat-Scramble or Tat-SabKIM1L-A (Figure 6E). To confirm that the inhibition of apoptosis observed with the Tat-SabKIM1 peptide was due to loss of Sab-mediated signaling, we silenced Sab expression in PA-1 cells using shRNAs specific for Sab (Figure 6F; quantified in Figure 6G). Analysis of cell death using Annexin V and PI revealed that PA-1 cells with deficient Sab-expression were resistant to cisplatin and paclitaxel treatment when compared with mock-transfected cells or cells transfected with a luciferase-specific shRNA (Figure 6H). Cell death induction was restored in Sab-deficient PA-1 cells by ectopic expression of a shRNA-resistant Sab construct (Figure 6F and 6H) [35]. Additionally, silencing Sab expression in PA-1 cells impaired caspase activation when compared with cells expressing a control shRNA (Figure 6I), and expression of a shRNA-resistant Sab restore caspase activity following 25 nM cisplatin and 100 nM paclitaxel treatment. There was no changes in cellular viability observed by the introduction of the overexpression and silencing constructs (Supplementary Figure S7) suggesting that the induced effects were related to the changes in Sab levels. Taken together, these results suggest that Sab facilitates signaling necessary to induce apoptosis in OC cells.

Inhibiting Sab-mediated signaling in sensitive OC cells enhances resistance.

Figure 6.
Inhibiting Sab-mediated signaling in sensitive OC cells enhances resistance.

(A) Fluorescent microscopy of a TUNEL assay performed on SK-OV-3 cells treated with vehicle (0.1% DMSO) or 1 µM STS for 2 h. Cells were stained with DAPI for total cell number, while Annexin V and PI staining were used to detect early and late apoptosis respectively. (B) The TUNEL assays were scored by counting the fluorescent cells in a field of view and based on fluorescence SK-OV-3 cells expressing either RFP, Sab, or SabKIM1(L-A) were assessed for apoptosis levels in the presence and absence of 25 µM cisplatin and 10 µM paclitaxel after 72 h. (C) Caspase assays were performed using cell lysates obtain from SK-OV-3 cells expressing RFP, Sab, or SabKIM1(L-A) that were exposed to 1% DMSO, Cisplatin/Paclitaxel for 72 h, or 1 µM STS for 24 h. (D) Chemosensitive PA-1 cells with high Sab levels were surveyed for apoptosis using TUNEL assays in the presence and absence of 25 nM cisplatin and 100 nM paclitaxel after 72 h in the presence of increasing concentrations of Tat-Scramble, Tat-SabKIM1, or Tat-SabKIM1(L-A) peptides. Both early and late apoptotic cells were scored in these experiments. (E) Caspase assays were used to assess the extent of apoptosis induced in PA-1 cells treated with cisplatin/paclitaxel and 1 µM of either Tat-Scramble, Tat-SabKIM1, or Tat-SabKIM1(L-A) peptides. STS was used as a positive control for apoptosis. (F) To selectively impair Sab-mediated signaling in chemosensitive PA-1 cells, we used shRNA-mediated gene silencing of Sab. Luciferase-specific shRNAs were used as a negative control, while two Sab-specific shRNAs were used. Additionally, a version of Sab resistant to shRNA #1 was expressed in PA-1 cells treated with Sab shRNA #1 as a rescue control. (G) The levels of Sab were assessed by western blot analysis using COX-IV as a mitochondrial loading control and tubulin as a cellular loading control. (H) A TUNEL assay was performed to assess the levels of apoptosis in cells treated with Luc or Sab shRNAs and compared with a Sab rescue in PA-1 cells. (I) Caspase assays were employed to indicate the level of apoptosis in the PA-1 cells treated with shRNAs. An asterisk (*) is used to indicate studies with a P-value <0.01. A double asterisk (**) is used to demark differences between cell types and distinct treatments.

Figure 6.
Inhibiting Sab-mediated signaling in sensitive OC cells enhances resistance.

(A) Fluorescent microscopy of a TUNEL assay performed on SK-OV-3 cells treated with vehicle (0.1% DMSO) or 1 µM STS for 2 h. Cells were stained with DAPI for total cell number, while Annexin V and PI staining were used to detect early and late apoptosis respectively. (B) The TUNEL assays were scored by counting the fluorescent cells in a field of view and based on fluorescence SK-OV-3 cells expressing either RFP, Sab, or SabKIM1(L-A) were assessed for apoptosis levels in the presence and absence of 25 µM cisplatin and 10 µM paclitaxel after 72 h. (C) Caspase assays were performed using cell lysates obtain from SK-OV-3 cells expressing RFP, Sab, or SabKIM1(L-A) that were exposed to 1% DMSO, Cisplatin/Paclitaxel for 72 h, or 1 µM STS for 24 h. (D) Chemosensitive PA-1 cells with high Sab levels were surveyed for apoptosis using TUNEL assays in the presence and absence of 25 nM cisplatin and 100 nM paclitaxel after 72 h in the presence of increasing concentrations of Tat-Scramble, Tat-SabKIM1, or Tat-SabKIM1(L-A) peptides. Both early and late apoptotic cells were scored in these experiments. (E) Caspase assays were used to assess the extent of apoptosis induced in PA-1 cells treated with cisplatin/paclitaxel and 1 µM of either Tat-Scramble, Tat-SabKIM1, or Tat-SabKIM1(L-A) peptides. STS was used as a positive control for apoptosis. (F) To selectively impair Sab-mediated signaling in chemosensitive PA-1 cells, we used shRNA-mediated gene silencing of Sab. Luciferase-specific shRNAs were used as a negative control, while two Sab-specific shRNAs were used. Additionally, a version of Sab resistant to shRNA #1 was expressed in PA-1 cells treated with Sab shRNA #1 as a rescue control. (G) The levels of Sab were assessed by western blot analysis using COX-IV as a mitochondrial loading control and tubulin as a cellular loading control. (H) A TUNEL assay was performed to assess the levels of apoptosis in cells treated with Luc or Sab shRNAs and compared with a Sab rescue in PA-1 cells. (I) Caspase assays were employed to indicate the level of apoptosis in the PA-1 cells treated with shRNAs. An asterisk (*) is used to indicate studies with a P-value <0.01. A double asterisk (**) is used to demark differences between cell types and distinct treatments.

JNK activity is a component of Sab-mediated effects on OC cell vulnerability

Because of the relationship between JNK, Sab, and events contributing to apoptosis, we used JNK selective inhibitors and JNK1/2-selective shRNAs to determine if the phenotypes associated with Sab overexpression in chemoresistant OC cells could be attributed to JNK kinase activity. To control for off-target consequences of small molecule JNK inhibitors, we silenced the expression of JNK isoforms 1 and 2 (JNK1/2) in SK-OV-3 cells 48 h before overexpressing Sab, and we confirmed the silencing was sustained through 72 h of Sab overexpression (120 h of silencing) in Figure 7A. At the time of transfection in SK-OV-3 cells with the pLOC plasmids, we introduced increasing amounts (0, 5 nM, 50 nM, 500 nM, and 5 µM) of SR-3306 to impair JNK activity during Sab overexpression. It should be noted that the presence of the JNK inhibitor or JNK shRNAs had no impact on Sab overexpression (Supplementary Figure S8). However, silencing JNKs or the presence of SR-3306 reduced the percent depolarization in a dose-dependent manner compared with cells treated with a Luc shRNA or DMSO (Figure 7B). This was accompanied by a corresponding decreased in the percent of priming in Sab-overexpressing SK-OV-3 cells (Figure 7B). To assess if the changes in the BH3-only responsiveness were due to changes in the levels of Bcl-2 and BH3-only proteins, we performed western blots of SK-OV-3 cells overexpressing Sab that have been treated with either 0.1% DMSO or 500 nM of SR-3306 for 72 h. In Figure 7C, we found that the levels of anti-apoptotic Bcl-2 proteins, Bcl-2 and Bcl-xL increased on mitochondria in the presence of SR-3306; meanwhile, pro-apoptotic BH3-only proteins, Bim and Bad, decreased in SK-OV-3 cells exposed to 500 nM SR-3306. To determine if the changes induced by the JNK inhibitor impacted MOMP, we quantified the levels of cytochrome c in the cytosol. Again, the presence of 500 nM SR-3306 diminished the release of cytochrome c from mitochondria in Sab overexpressing SK-OV-3 cells treated with 500 nM BIM BH3-peptide when compared with those treated with DMSO (Figure 7D). We next measured the vulnerability of SK-OV-3 cells overexpressing Sab to cisplatin/paclitaxel treatment in the presence and absence of JNK activity. In Figure 7E, inhibition of JNK activity either through silencing or the presence of SR-3306 reduce the extent of cell death in the presence of 25 nM cisplatin and 100 nM paclitaxel for 72 h. Finally, to determine if impaired JNK activity could prevent Sab-mediated events under endogenous conditions, we silenced JNK in PA-1 cells (Figure 7A) and examined the impact on chemovulnerability. Silencing JNK was able to reduce apoptotic priming (Figure 7F), mitochondrial Bcl-2 and Bim levels (Figure 7G) and prevented cisplatin/paclitaxel-induced cell death (Figure 7H). These events were comparable to the treatment of PA-1 with 500 nm SR-3306. Additionally, we examined the impact of acute (24 h) administration of Sab inhibition (5 µM Tat-SabKIM1 peptide) or the JNK inhibitor (500 nM SR-3306), and we found that there was no significant shift in anti-apoptotic or pro-apoptotic Bcl-2 proteins in this shorter time frame (Supplementary Figure S9). These data illustrate that mitochondrial JNK activity in part responsible for the changes in cellular sensitivity to chemotherapeutic agents.

JNK activity plays a role in the Sab-mediated effects in OC cell lines.

Figure 7.
JNK activity plays a role in the Sab-mediated effects in OC cell lines.

(A) Western blot analysis of SK-OV-3 and PA-1 cells transfected with plasmids expressing either a luciferase-specific (Luc) shRNA or one of two JNK-selective shRNAs (JNK shRNA 1 or 2) for 120 h (top panel). Tubulin was used to represent sample loading. The relative fluorescence of JNK bands were normalized to protein content (Ponceau staining) and plotted in the bottom panel. An asterisk (*) represents a difference (P < 0.05) between control and Luc shRNA expressing cells, while double asterisks (**) illustrate a significant difference between the two JNK shRNAs. (B) SK-OV-3 cells were treated with DMSO or increasing concentrations of JNK-selective inhibitor SR-3306 (5, 50, 500, and 5000 nM) 24-hours after transfection for Sab overexpression (pLOC:Sab), additionally, SK-OV-3 cells expressing either the Luc shRNA or JNK shRNA 2 48 h prior to transfection with pLOC:Sab. After 72 h of Sab overexpression, the cells were subjected to dynamic BH3 profiling with the Bim BH3-peptide. The percent depolarization (%Depolarization — white bars) and change in percent priming (Δ%Priming – blue bars) were calculated, and the means of four experiments were plotted with error bars representing the standard deviation. An asterisk is used to indicate differences (P < 0.05) between untreated or vehicle-treated controls and experimental conditions. (C) The relative abundance of Bcl-2, Bcl-xL, and Bim were assessed in SK-OV-3 cells that had been treated with Luc shRNA, JNK shRNA 2 or 500 nM SR-3306 as described in panel B. COX-IV was used to represent the relative levels of mitochondrial proteins per sample. (D) The relative levels of cytochrome c were measured using an ELISA of cytoplasmic extracts from SK-OV-3 cells that were exposed to either Luc shRNA, JNK shRNA2, or 500 nM SR-3306 prior to Sab overexpression and treated with 500 nM BIM BH3-peptide for 45 min. The relative fluorescence of cytochrome c in cytoplasmic preparations was normalized to protein concentrations in individual wells. Data shown are the means for at least three replicate studies, and the error bars represent one standard deviation from the mean. An asterisk (*) indicates differences (P < 0.05) between controls and treatments, while double asterisks indicate differences among BIM BH3-peptide exposed samples. (E) Cell viability was assessed for SK-OV-3 cells that were exposed to Luc shRNA, JNK shRNA2, or 500 nM SR-3306 before Sab overexpression followed by treatment with either DMSO or 25 nM cisplatin and 100 nM paclitaxel for 72 h. The percent cells viable was obtained by normalizing to untreated controls; data were plotted as means of four replicates with standard deviation used for error bars. An asterisk is used to indicate differences (P < 0.05) between untreated or vehicle-treated controls and experimental conditions. (F) PA-1 cells (have high levels of Sab) were treated with DMSO, increasing concentrations of JNK-selective inhibitor SR-3306 (5, 50, 500, and 5000 nM), or shRNAs targeting Luc or JNK. Dynamic BH3-profiling was performed with the Bim BH3-peptide, and the percent depolarization (%Depolarization – white bars) and change in percent priming (Δ%Priming – blue bars) are shown. An asterisk is used to indicate differences (P < 0.05) between untreated or vehicle-treated controls and experimental conditions. (G) Western blot analyses of Bcl-2, Bcl-xL, and Bim were performed for PA-1 cells that had been treated with Luc shRNA, JNK shRNA 2 or 500 nM SR-3306 as described above. COX-IV is used to represent the relative amounts of mitochondria protein in each sample. (H) Cell viability was assessed for PA-1 cells exposed to either Luc shRNA, JNK shRNA2, or 500 nM SR-3306 followed by treatment with either DMSO or 25 nM cisplatin and 100 nM paclitaxel for 72 h. Data analysis was handled as described for SK-OV-3 cells above.

Figure 7.
JNK activity plays a role in the Sab-mediated effects in OC cell lines.

(A) Western blot analysis of SK-OV-3 and PA-1 cells transfected with plasmids expressing either a luciferase-specific (Luc) shRNA or one of two JNK-selective shRNAs (JNK shRNA 1 or 2) for 120 h (top panel). Tubulin was used to represent sample loading. The relative fluorescence of JNK bands were normalized to protein content (Ponceau staining) and plotted in the bottom panel. An asterisk (*) represents a difference (P < 0.05) between control and Luc shRNA expressing cells, while double asterisks (**) illustrate a significant difference between the two JNK shRNAs. (B) SK-OV-3 cells were treated with DMSO or increasing concentrations of JNK-selective inhibitor SR-3306 (5, 50, 500, and 5000 nM) 24-hours after transfection for Sab overexpression (pLOC:Sab), additionally, SK-OV-3 cells expressing either the Luc shRNA or JNK shRNA 2 48 h prior to transfection with pLOC:Sab. After 72 h of Sab overexpression, the cells were subjected to dynamic BH3 profiling with the Bim BH3-peptide. The percent depolarization (%Depolarization — white bars) and change in percent priming (Δ%Priming – blue bars) were calculated, and the means of four experiments were plotted with error bars representing the standard deviation. An asterisk is used to indicate differences (P < 0.05) between untreated or vehicle-treated controls and experimental conditions. (C) The relative abundance of Bcl-2, Bcl-xL, and Bim were assessed in SK-OV-3 cells that had been treated with Luc shRNA, JNK shRNA 2 or 500 nM SR-3306 as described in panel B. COX-IV was used to represent the relative levels of mitochondrial proteins per sample. (D) The relative levels of cytochrome c were measured using an ELISA of cytoplasmic extracts from SK-OV-3 cells that were exposed to either Luc shRNA, JNK shRNA2, or 500 nM SR-3306 prior to Sab overexpression and treated with 500 nM BIM BH3-peptide for 45 min. The relative fluorescence of cytochrome c in cytoplasmic preparations was normalized to protein concentrations in individual wells. Data shown are the means for at least three replicate studies, and the error bars represent one standard deviation from the mean. An asterisk (*) indicates differences (P < 0.05) between controls and treatments, while double asterisks indicate differences among BIM BH3-peptide exposed samples. (E) Cell viability was assessed for SK-OV-3 cells that were exposed to Luc shRNA, JNK shRNA2, or 500 nM SR-3306 before Sab overexpression followed by treatment with either DMSO or 25 nM cisplatin and 100 nM paclitaxel for 72 h. The percent cells viable was obtained by normalizing to untreated controls; data were plotted as means of four replicates with standard deviation used for error bars. An asterisk is used to indicate differences (P < 0.05) between untreated or vehicle-treated controls and experimental conditions. (F) PA-1 cells (have high levels of Sab) were treated with DMSO, increasing concentrations of JNK-selective inhibitor SR-3306 (5, 50, 500, and 5000 nM), or shRNAs targeting Luc or JNK. Dynamic BH3-profiling was performed with the Bim BH3-peptide, and the percent depolarization (%Depolarization – white bars) and change in percent priming (Δ%Priming – blue bars) are shown. An asterisk is used to indicate differences (P < 0.05) between untreated or vehicle-treated controls and experimental conditions. (G) Western blot analyses of Bcl-2, Bcl-xL, and Bim were performed for PA-1 cells that had been treated with Luc shRNA, JNK shRNA 2 or 500 nM SR-3306 as described above. COX-IV is used to represent the relative amounts of mitochondria protein in each sample. (H) Cell viability was assessed for PA-1 cells exposed to either Luc shRNA, JNK shRNA2, or 500 nM SR-3306 followed by treatment with either DMSO or 25 nM cisplatin and 100 nM paclitaxel for 72 h. Data analysis was handled as described for SK-OV-3 cells above.

Discussion

Mitochondrial JNK signaling on the MOM scaffold protein Sab has been shown to be an instrumental event in the induction of mitochondria-induced cell death; furthermore, altering the concentration of Sab-mediated JNK signaling has been shown to affect chemosensitivity in tissues (brain and heart) and cancer cells. In our current study, we extend these observations by demonstrating that chemoresistant cancer cells have diminished Sab expression compared with healthy tissue and restoring or enhancing Sab levels on the MOM is critical to reestablishing proper apoptotic signaling in OC cell lines. Our analysis of six independent OC transcriptome analyses [3843] found that Sab mRNA levels were reduced over 5-fold when compared with normal tissues (Figure 1, and Table 1). Previous studies have demonstrated in various tissues and cell lines that diminished Sab-mediated signaling prevents apoptosis [33,36,53,54,56,57]. To date, there are no published reports regarding the transcriptional regulation of Sab. It stands to reason that the decreased reliance on mitochondrial health status and the suppression of apoptotic mechanisms in cancer cells [19] may be partly responsible for the reduction in Sab expression. Indeed, we recently described that JNK-deficient murine embryonic fibroblasts (MEFs) had a significant decrease in Sab on the MOM compared with wild-type MEFs, which was accompanied by a corresponding resistance to conventional chemotherapeutic and chemosensitizing agents [35]. In our opinion, this would suggest that perturbations in signal transduction pathways, specifically those responsible for apoptosis, may be instrumental to the diminished Sab levels in OC.

The diminished concentrations of Sab mRNA in the earliest stages of OC suggests that the change in Sab expression may be an early event in oncogenesis, and the sustained suppression of Sab mRNA levels implies that its expression may be detrimental to tumor progression. It is essential to keep in mind that Sab protein levels were not assessed in the patient studies, and our studies explore how Sab protein concentrations impact OC cell biology, which could mean that post-translational mechanisms are in place in OC cells that suppress Sab protein concentrations on the MOM that are distinct from the transcriptional programs driving Sab expression. Without a discernable relationship to patient survival, it may also be that silencing Sab expression may have an impact on tumor cell biology beyond cell death. For example, mitochondrial JNK signaling has been shown to impair pyruvate dehydrogenase and respiratory complex I [34,58,59]. So, perhaps, silencing Sab expression is a means to mitigate bioenergetic or undiscovered consequences of mitochondrial JNK signaling.

Hendrix and colleagues included a brief survey of patient genetic data into their study and diminished Sab expression was observed in patients harboring hallmark mutations in catenin beta 1 (CTNNB1), Phosphatase and tensin homolog (PTEN), and phosphatidylinositol-4,5-bisphosphate 3-kinase alpha catalytic subunit (PI3KCA). While patients with mutations in K-Ras, p53, and adenomatous polyposis coli (APC). The interaction between PTEN-PI3K will require further investigation as previous research links PTEN loss to JNK-induced cell survival [6062] and PI3K signaling to JNK-mediated cancer progression [63]. PTEN/PI3K-related silencing of Sab may be a means to limit JNK signaling to the nucleus during oncogenesis.

Intriguingly, one OC patient from The Cancer Genome Atlas (TCGA) repository [42] had a Sab-overexpression phenotype; however, the genomic sequence of Sab from this patient possessed early termination site, which truncated the protein at residue 379 (Sab usually is 455 amino acids in length). The loss of the C-terminal portion removes the two kinase interaction motifs (KIMs) responsible for mitochondrial MAPK signaling. Without the KIMs, it is unlikely that this Sab variant would have any impact on apoptotic induction, as previous studies have demonstrated that mutation of the KIMs that disrupt MAPK binding prevent the induction of cell death.

We demonstrated that the relative levels of Sab correlate to chemo-responsiveness in OC cells. Over-expression of Sab improved the chemosensitivity of resistant SK-OV-3 cells; meanwhile, inhibiting Sab-mediated signaling with the Tat-SabKIM1 peptide or silencing Sab expression increased chemoresistance in vulnerable PA-1 cells (Figures 2 and 5). These effects were mediated by the relative levels of JNK on mitochondria. JNK signaling has long been implicated in apoptotic responses [64]. JNK signaling can contribute to apoptosis through the combination of nuclear and cytosolic/mitochondrial mechanisms. An interesting observation was that the six additional cell lines analyzed (Figure 3C–E) were found to have comparable levels of Sab and similar responses to cisplatin and paclitaxel. It may be interested to determine if prolonged cell culture adaptations may influence Sab concentrations; this means that direct analysis of patient tumors will be imperative when continuing this line of research.

The primary mitochondrial alteration facilitated by Sab-mediated JNK signaling in OC cells was the change in Bcl-2/BH3-only proteins at mitochondria (Figures 3 and 4). Previous research establishes this possibility, as JNK can directly phosphorylate Bcl-2 and BH3-only proteins affecting their functions and localizations [65]. It is likely that the post-translational modification of Bcl-2 and BH3-only proteins ascribed to JNK activity would likely alter the stability of these proteins; whereby, anti-apoptotic proteins may be targeted for proteolysis similar to other JNK substrates, and pro-apoptotic proteins, like Bim, may be preserved by altering their protein–protein interactions. For example, we observed a marked decrease in mitochondrial Bcl-2 levels, an event previously noted in breast cancer cells treated with paclitaxel [66,67]. Following stress, such as chemotherapy, JNK can phosphorylate Bcl-2 on Ser70 inducing its migration from the MOM [33,68]. Furthermore, JNK can impair Bcl-xL activity by phosphorylating Thr47 and Thr115 [51]. Enhancing the amount of Sab on the MOM would increase the concentration of activated JNK that could modify Bcl-2 proteins and hasten the impairment of anti-apoptotic mechanisms on mitochondria.

JNK can directly phosphorylate BH3-only proteins and increase their pro-apoptotic potential. During UV-induced apoptosis JNK has been shown to phosphorylate Bim and Bmf [24,68,69]. This phosphorylation causes Bim and Bmf to dissociate from dynein and myosin V motor complexes and activate Bax and Bak to initiate apoptosis. Furthermore, phosphorylation of Bim by JNK can complement JNK's phosphorylation of Bcl-2 and Bcl-xL because Bim can bind and neutralize Bcl-2 and Bcl-xL to promote apoptosis as well. Additionally, JNK can phosphorylate Bad on Ser128, which promotes apoptosis [70]. Bad, similar to Bim, can promote apoptosis by inhibiting Bcl-2 proteins. JNK can improve Bad activity by phosphorylating 14-3-3ξ, which causes the release of Bad from the scaffold [71]. JNK can also affect the BH3-only proteins of the extrinsic pathway of apoptosis at the mitochondrial level. JNK can induce the cleavage of Bid, thereby elevating the levels of t-Bid, which will, in turn, activate Bax and Bak [72,73]. During TNF-α-evoked apoptosis in HeLa cells, JNK activity induces a caspase 8-dependent cleavage of Bid that produces the product jBid that promotes the release of Smac/DIABLO [74]. In addition to the direct modulation of Bcl-2 function, mitochondrial JNK signaling facilitated by Sab would also enhance the activity of BH3-only proteins contributing to apoptosis in OC cells treated with chemotherapy. Therefore, the relative abundance of JNK on mitochondria could tip the balance of anti- and pro-apoptotic events by altering the relative abundance of Bcl-2 and BH3-only proteins, respectively.

The impact of mitochondrial JNK signaling on Bcl-2 and BH3-only proteins has been well documented [6268]. This relationship balances between transcriptional regulation of expression and direct phsophoregulation, which calls into question the roles of Sab abundance in these distinct responses. One may surmise that modest levels of Sab on mitochondria permit low-grade JNK activation to adjust mitochondrial functions and health; whereas robust levels of JNK activity on mitochondria induce apoptosis through the post-translational activities of Bcl-2 and BH3-only proteins. Thus, it may be likely that lower doses of retrograde JNK signaling from mitochondria over time lead to small adjustments in transcription of Bcl-2 and BH3-only proteins. For this reason, acute JNK inhibition did not affect the levels of Bcl-2 and Bim in our studies, while sub-chronic exposures resulted in a change in the concentrations of these two proteins. Thus, Sab levels and those of Bcl-2 and BH3-only protein may depend upon the level of chronic nuclear JNK signaling in cancer cells. Consequently, one may need to consider the relative levels of Sab and Bcl-2/BH3-only profiles while regarding total JNK activities in the cell.

We have previously demonstrated that increasing Sab levels in mammalian cells can increase the vulnerability towards toxic agents. First, we demonstrated that the sub-chronic exposure to the chemosensitizer LY294002 increased Sab levels on the MOM, which led to an increase in mitochondrial JNK signaling and induction of apoptosis [35]. Again, we demonstrate that acute inhibition (24 h or less) of Sab-mediated signaling or JNK activity do not alter the Bcl-2/BH3-only content of mitochondria (Supplementary Figure S9). It is likely that the presence of elevated mitochondrial JNK signaling on Sab induces modest amounts of mitochondrial dysfunction which amplify JNK activity leading to the accumulation of pro-apoptotic Bcl-2 proteins on mitochondria [34,35]. This previous research combined with our current study would suggest that increasing Sab levels would be a useful technique to improve chemo-responsiveness in cancer cells. However, we have recently reported that elevated Sab levels in the heart can increase the risk for cardiotoxicity following chemotherapy [36]. We show that artificially enhancing Sab in cardiomyocyte cell lines sensitized the cells to imatinib [14]. One may anticipate that targeted methods to selectively up-regulate Sab expression in OC tumors may be a safer approach to sensitize cancer cells than that of general Sab increase throughout the body.

We propose that Sab levels may be utilized better as a prognostic biomarker to guide the selection and magnitude of therapy for OC patients. Specifically, low Sab concentrations in a tumor biopsy or resection may suggest a more rigorous therapeutic regimen; whereby, high Sab levels may allow for a less ambitious and toxic approach to treating patients. Nonetheless, cancer, especially OC, are complex conditions and Sab concentrations may need to be considered in the cellular context of the tumor for choosing the most efficacious therapeutic approach. Because little is known regarding the regulation of Sab expression, future studies in our group will be focused on understanding the molecular mechanisms responsible for controlling Sab levels on the MOM.

Our current study reinforces the vital role that scaffold proteins play in the regulating cell signaling responses, especially how scaffold proteins, such as Sab, can influence the magnitude and outcomes at specific subcellular sites within cells. We demonstrate that increasing the concentration of a specific MOM scaffold protein can enhance signaling pathways on mitochondria and alter organelle and cellular physiology. Specifically, we have demonstrated that increasing Sab levels in OC cells increases mitochondrial JNK signaling and apoptotic priming, which enhances the vulnerability of OC cells to chemotherapy agents. Moreover, we uncovered that Sab levels are decreased in OC patient samples, and restoring Sab levels in OC cells with low Sab expression enhances chemo-responsiveness. Thus, assessing Sab concentrations in the context of genetic and histological testing, including dynamic BH3 profiling, may be a useful prognostic tool for OC.

Abbreviations

     
  • APC

    adenomatous polyposis coli

  •  
  • BSA

    bovine serum albumin

  •  
  • CS

    citrate synthase

  •  
  • CST

    cell signaling technology

  •  
  • DMEM

    Dulbecco's Minimal Essential Medium

  •  
  • EMEM

    Eagle's Minimal Essential Medium

  •  
  • FBS

    fetal bovine serum

  •  
  • HBSS

    Hank's Buffered Saline Solution

  •  
  • ICW

    in-cell western

  •  
  • JNK

    N-terminal kinase

  •  
  • KIMs

    kinase interaction motifs

  •  
  • LDH

    lactate dehydrogenase

  •  
  • MEFs

    murine embryonic fibroblasts

  •  
  • MOM

    mitochondrial outer membrane

  •  
  • MOMP

    mitochondrial outer membrane permeabilization

  •  
  • PBS

    phosphate-buffered saline

  •  
  • PMSF

    phenylmethanesulfonyl fluoride

  •  
  • PTEN

    Phosphatase and tensin homolog

  •  
  • RFP

    red fluorescent protein

  •  
  • STS

    staurosporine

Author Contribution

The contribution of the authors is as follows: J.W.C. designed the experiments and wrote the manuscript. I.P., S.M.H., G.M.P., and T.P.C. conducted the experiments.

Acknowledgments

The authors thank Drs Carolyn Runowicz, Fenfei Leng, and Yuk-Ching Tse Dinh for their helpful comments during the preparation of this manuscript. The authors acknowledge generous funding from the Hearing the Ovarian Cancer Whisper Foundation, Florida International University, and the FIU Division of Research.

Competing Interests

The Authors declare that there are no competing interests associated with the manuscript.

References

References
1
Parkin
,
D.M.
,
Bray
,
F.
,
Ferlay
,
J.
and
Pisani
,
P.
(
2005
)
Global cancer statistics, 2002
.
CA Cancer J. Clin.
55
,
74
108
2
Jemal
,
A.
,
Siegel
,
R.
,
Ward
,
E.
,
Hao
,
Y.
,
Xu
,
J.
,
Murray
,
T.
et al. 
(
2008
)
Cancer statistics, 2008
.
CA Cancer J. Clin.
58
,
71
96
3
Ledermann
,
J.
,
Harter
,
P.
,
Gourley
,
C.
,
Friedlander
,
M.
,
Vergote
,
I.
,
Rustin
,
G.
et al. 
(
2012
)
Olaparib maintenance therapy in platinum-sensitive relapsed ovarian cancer
.
N. Engl. J. Med.
366
,
1382
1392
4
Gore
,
M.E.
,
Fryatt
,
I.
,
Wiltshaw
,
E.
and
Dawson
,
T.
(
1990
)
Treatment of relapsed carcinoma of the ovary with cisplatin or carboplatin following initial treatment with these compounds
.
Gynecol. Oncol.
36
,
207
211
5
Sandercock
,
J.
,
Parmar
,
M.K.B.
,
Torri
,
V.
and
Qian
,
W.
(
2002
)
First-line treatment for advanced ovarian cancer: paclitaxel, platinum and the evidence
.
Br. J. Cancer
87
,
815
824
6
Yap
,
T.A.
,
Carden
,
C.P.
and
Kaye
,
S.B.
(
2009
)
Beyond chemotherapy: targeted therapies in ovarian cancer
.
Nat. Rev. Cancer
9
,
167
181
7
Placido
,
S.D.
,
Scambia
,
G.
,
Vagno
,
G.D.
,
Naglieri
,
E.
,
Lombardi
,
A.V.
,
Biamonte
,
R.
et al. 
(
2004
)
Topotecan compared with no therapy after response to surgery and carboplatin/paclitaxel in patients with ovarian cancer: multicenter Italian trials in ovarian cancer (MITO-1) randomized study
.
J. Clin. Oncol.
22
,
2635
2642
8
Bookman
,
M.A.
,
Greer
,
B.E.
and
Ozols
,
R.F.
(
2003
)
Optimal therapy of advanced ovarian cancer: carboplatin and paclitaxel vs. cisplatin and paclitaxel (GOG 158) and an update on GOG0 182-ICON5
.
Int. J. Gynecol. Cancer
13
,
735
740
9
Spannuth
,
W.A.
,
Sood
,
A.K.
and
Coleman
,
R.L.
(
2008
)
Angiogenesis as a strategic target for ovarian cancer therapy
.
Nat. Clin. Pract. Oncol.
5
,
194
204
10
Mesiano
,
S.
,
Ferrara
,
N.
and
Jaffe
,
R.B.
(
1998
)
Role of vascular endothelial growth factor in ovarian cancer
.
Am. J. Pathol.
153
,
1249
1256
11
Matei
,
D.
,
Emerson
,
R.E.
,
Lai
,
Y.C.
,
Baldridge
,
L.A.
,
Rao
,
J.
,
Yiannoutsos
,
C.
et al. 
(
2005
)
Autocrine activation of PDGFR[alpha] promotes the progression of ovarian cancer
.
Oncogene
25
,
2060
2069
12
Matei
,
D.
,
Chang
,
D.D.
and
Jeng
,
M.-H.
(
2004
)
Imatinib mesylate (Gleevec) inhibits ovarian cancer cell growth through a mechanism dependent on platelet-derived growth factor receptor α and Akt inactivation
.
Clin. Cancer Res.
10
,
681
690
13
Levine
,
D.A.
,
Bogomolniy
,
F.
,
Yee
,
C.J.
,
Lash
,
A.
,
Barakat
,
R.R.
,
Borgen
,
P.I.
et al. 
(
2005
)
Frequent mutation of the PIK3CA gene in ovarian and breast cancers
.
Clin. Cancer Res.
11
,
2875
2878
14
Obata
,
K.
,
Morland
,
S.J.
,
Watson
,
R.H.
,
Hitchcock
,
A.
,
Chenevix-Trench
,
G.
,
Thomas
,
E.J.
et al. 
(
1998
)
Frequent PTEN/MMAC mutations in endometrioid but not serous or mucinous epithelial ovarian tumors
.
Cancer Res.
58
,
2095
2097
PMID:
[PubMed]
15
Wiener
,
J.R.
,
Windham
,
T.C.
,
Estrella
,
V.C.
,
Parikh
,
N.U.
,
Thall
,
P.F.
,
Deavers
,
M.T.
et al. 
(
2003
)
Activated Src protein tyrosine kinase is overexpressed in late-stage human ovarian cancers
.
Gynecol. Oncol.
88
,
73
79
16
Bryant
,
H.E.
,
Schultz
,
N.
,
Thomas
,
H.D.
,
Parker
,
K.M.
,
Flower
,
D.
,
Lopez
,
E.
et al. 
(
2005
)
Specific killing of BRCA2-deficient tumours with inhibitors of poly(ADP-ribose) polymerase
.
Nature
434
,
913
917
17
Kalli
,
K.R.
,
Oberg
,
A.L.
,
Keeney
,
G.L.
,
Christianson
,
T.J.H.
,
Low
,
P.S.
,
Knutson
,
K.L.
et al. 
(
2008
)
Folate receptor alpha as a tumor target in epithelial ovarian cancer
.
Gynecol. Oncol.
108
,
619
626
18
Gibbs
,
D.D.
,
Theti
,
D.S.
,
Wood
,
N.
,
Green
,
M.
,
Raynaud
,
F.
,
Valenti
,
M.
et al. 
(
2005
)
BGC 945, a novel tumor-selective thymidylate synthase inhibitor targeted to α-folate receptor–overexpressing tumors
.
Cancer Res.
65
,
11721
11728
19
Hanahan
,
D.
and
Weinberg
,
R.A.
(
2011
)
Hallmarks of cancer: the next generation
.
Cell
144
,
646
674
20
Chappell
,
N.P.
,
Teng
,
P.N.
,
Hood
,
B.L.
,
Wang
,
G.
,
Darcy
,
K.M.
,
Hamilton
,
C.A.
et al. 
(
2012
)
Mitochondrial proteomic analysis of cisplatin resistance in ovarian cancer
.
J. Proteome Res.
11
,
4605
4614
21
Ara
,
G.
,
Kusumoto
,
T.
,
Korbut
,
T.T.
,
Cullere-Luengo
,
F.
and
Teicher
,
B.A.
(
1994
)
cis-diamminedichloroplatinum(II) resistant human tumor cell lines are collaterally sensitive to PtCl evidence for mitochondrial involvement
.
Cancer Res.
54
,
1497
1502
PMID:
[PubMed]
22
Montero
,
J.
,
Sarosiek
,
K.A.
,
DeAngelo
,
J.D.
,
Maertens
,
O.
,
Ryan
,
J.
,
Ercan
,
D.
et al. 
(
2015
)
Drug-induced death signaling strategy rapidly predicts cancer response to chemotherapy
.
Cell
160
,
977
989
23
Chonghaile
,
T.N.
,
Sarosiek
,
K.A.
,
Vo
,
T.-T.
,
Ryan
,
J.A.
,
Tammareddi
,
A.
,
Moore
,
V.D.G.
et al. 
(
2011
)
Pretreatment mitochondrial priming correlates with clinical response to cytotoxic chemotherapy
.
Science
334
,
1129
1133
24
Letai
,
A.
,
Bassik
,
M.C.
,
Walensky
,
L.D.
,
Sorcinelli
,
M.D.
,
Weiler
,
S.
and
Korsmeyer
,
S.J.
(
2002
)
Distinct BH3 domains either sensitize or activate mitochondrial apoptosis, serving as prototype cancer therapeutics
.
Cancer Cell
2
,
183
192
25
Shamas-Din
,
A.
,
Brahmbhatt
,
H.
,
Leber
,
B.
and
Andrews
,
D.W.
(
2011
)
BH3-only proteins: orchestrators of apoptosis
.
Biochim. Biophys. Acta, Mol. Cell Res.
1813
,
508
520
26
Ryan
,
J.A.
,
Brunelle
,
J.K.
and
Letai
,
A.
(
2010
)
Heightened mitochondrial priming is the basis for apoptotic hypersensitivity of CD4+ CD8+ thymocytes
.
Proc. Natl Acad. Sci. U.S.A.
107
,
12895
12900
27
Ryan
,
J.
and
Letai
,
A.
(
2013
)
BH3 profiling in whole cells by fluorimeter or FACS
.
Methods
61
,
156
164
28
Letai
,
A.G.
(
2008
)
Diagnosing and exploiting cancer's addiction to blocks in apoptosis
.
Nat. Rev. Cancer
8
,
121
132
29
Certo
,
M.
,
Del Gaizo Moore
,
V.
,
Nishino
,
M.
,
Wei
,
G.
,
Korsmeyer
,
S.
,
Armstrong
,
S.A.
et al. 
(
2006
)
Mitochondria primed by death signals determine cellular addiction to antiapoptotic BCL-2 family members
.
Cancer Cell
9
,
351
365
30
Chandel
,
N.S.
(
2014
)
Mitochondria as signaling organelles
.
BMC Biol.
12
,
34
31
Bender
,
T.
and
Martinou
,
J.-C.
(
2013
)
Where killers meet—permeabilization of the outer mitochondrial membrane during apoptosis
.
Cold Spring Harb. Perspect. Biol.
5
,
a011106
32
Brunelle
,
J.K.
and
Letai
,
A.
(
2009
)
Control of mitochondrial apoptosis by the Bcl-2 family
.
J. Cell Sci.
122
,
437
441
33
Chambers
,
J.W.
,
Cherry
,
L.
,
Laughlin
,
J.D.
,
Figuera-Losada
,
M.
and
Lograsso
,
P.V.
(
2011
)
Selective inhibition of mitochondrial JNK signaling achieved using peptide mimicry of the Sab kinase interacting motif-1 (KIM1)
.
ACS Chem. Biol.
6
,
808
818
34
Chambers
,
J.W.
and
LoGrasso
,
P.V.
(
2011
)
Mitochondrial c-Jun N-terminal kinase (JNK) signaling initiates physiological changes resulting in amplification of reactive oxygen species generation
.
J. Biol. Chem.
286
,
16052
16062
35
Chambers
,
T.P.
,
Portalatin
,
G.M.
,
Paudel
,
I.
,
Robbins
,
C.J.
and
Chambers
,
J.W.
(
2015
)
Sub-chronic administration of LY294002 sensitizes cervical cancer cells to chemotherapy by enhancing mitochondrial JNK signaling
.
Biochem. Biophys. Res. Commun.
463
,
538
544
36
Chambers
,
T.P.
,
Santiesteban
,
L.
,
Gomez
,
D.
and
Chambers
,
J.W.
(
2017
)
Sab mediates mitochondrial dysfunction involved in imatinib mesylate-induced cardiotoxicity
.
Toxicology
382
,
24
35
37
Rhodes
,
D.R.
,
Kalyana-Sundaram
,
S.
,
Mahavisno
,
V.
,
Varambally
,
R.
,
Yu
,
J.
,
Briggs
,
B.B.
et al. 
(
2007
)
Oncomine 3.0: genes, pathways, and networks in a collection of 18,000 cancer gene expression profiles
.
Neoplasia
9
,
166
180
38
Adib
,
T.R.
,
Henderson
,
S.
,
Perrett
,
C.
,
Hewitt
,
D.
,
Bourmpoulia
,
D.
,
Ledermann
,
J.
et al. 
(
2004
)
Predicting biomarkers for ovarian cancer using gene-expression microarrays
.
Br. J. Cancer
90
,
686
692
39
Bonome
,
T.
,
Levine
,
D.A.
,
Shih
,
J.
,
Randonovich
,
M.
,
Pise-Masison
,
C.A.
,
Bogomolniy
,
F.
et al. 
(
2008
)
A gene signature predicting for survival in suboptimally debulked patients with ovarian cancer
.
Cancer Res.
68
,
5478
5486
40
Hendrix
,
N.D.
,
Wu
,
R.
,
Kuick
,
R.
,
Schwartz
,
D.R.
,
Fearon
,
E.R.
and
Cho
,
K.R.
(
2006
)
Fibroblast growth factor 9 has oncogenic activity and is a downstream target of Wnt signaling in ovarian endometrioid adenocarcinomas
.
Cancer Res.
66
,
1354
1362
41
Lu
,
K.H.
,
Patterson
,
A.P.
,
Wang
,
L.
,
Marquez
,
R.T.
,
Atkinson
,
E.N.
,
Baggerly
,
K.A.
et al. 
(
2004
)
Selection of potential markers for epithelial ovarian cancer with gene expression arrays and recursive descent partition analysis
.
Clin. Cancer Res.
10
,
3291
3300
42
Cancer Genome Atlas Research Network
. (
2011
)
Integrated genomic analyses of ovarian carcinoma
.
Nature
474
,
609
615
43
Yoshihara
,
K.
,
Tajima
,
A.
,
Komata
,
D.
,
Yamamoto
,
T.
,
Kodama
,
S.
,
Fujiwara
,
H.
et al. 
(
2009
)
Gene expression profiling of advanced-stage serous ovarian cancers distinguishes novel subclasses and implicates ZEB2 in tumor progression and prognosis
.
Cancer Sci.
100
,
1421
1428
44
Chambers
,
J.W.
,
Pachori
,
A.
,
Howard
,
S.
,
Ganno
,
M.
,
Hansen
, Jr,
D.
,
Kamenecka
,
T.
et al. 
(
2011
)
Small molecule c-jun-N-terminal kinase (JNK) inhibitors protect dopaminergic neurons in a model of Parkinson's disease
.
ACS Chem. Neurosci.
2
,
198
206
45
Laughlin
,
J.D.
,
Nwachukwu
,
J.C.
,
Figuera-Losada
,
M.
,
Cherry
,
L.
,
Nettles
,
K.W.
and
LoGrasso
,
P.V.
(
2012
)
Structural mechanisms of allostery and autoinhibition in JNK family kinases
.
Structure
20
,
2174
2184
46
Vo
,
T.T.
,
Ryan
,
J.
,
Carrasco
,
R.
,
Neuberg
,
D.
,
Rossi
,
D.J.
,
Stone
,
R.M.
et al. 
(
2012
)
Relative mitochondrial priming of myeloblasts and normal HSCs determines chemotherapeutic success in AML
.
Cell
151
,
344
355
47
Allen
,
R.T.
,
Hunter Iii
,
W.J.
and
Agrawal
,
D.K.
(
1997
)
Morphological and biochemical characterization and analysis of apoptosis
.
J. Pharmacol. Toxicol. Methods.
37
,
215
228
48
Anuradha
,
C.D.
,
Kanno
,
S.
and
Hirano
,
S.
(
2001
)
Oxidative damage to mitochondria is a preliminary step to caspase-3 activation in fluoride-induced apoptosis in HL-60 cells
.
Free Radic. Biol. Med.
31
,
367
373
49
Kluck
,
R.M.
,
Bossy-Wetzel
,
E.
,
Green
,
D.R.
and
Newmeyer
,
D.D.
(
1997
)
The release of cytochrome c from mitochondria: a primary site for Bcl-2 regulation of apoptosis
.
Science
275
,
1132
1136
50
Aoki
,
H.
,
Kang
,
P.M.
,
Hampe
,
J.
,
Yoshimura
,
K.
,
Noma
,
T.
,
Matsuzaki
,
M.
et al. 
(
2002
)
Direct activation of mitochondrial apoptosis machinery by c-Jun N-terminal kinase in adult cardiac myocytes
.
J. Biol. Chem.
277
,
10244
10250
51
Kharbanda
,
S.
,
Saxena
,
S.
,
Yoshida
,
K.
,
Pandey
,
P.
,
Kaneki
,
M.
,
Wang
,
Q.
et al. 
(
2000
)
Translocation of SAPK/JNK to mitochondria and interaction with Bcl-x(L) in response to DNA damage
.
J. Biol. Chem.
275
,
322
327
52
Chauhan
,
D.
,
Li
,
G.
,
Hideshima
,
T.
,
Podar
,
K.
,
Mitsiades
,
C.
,
Mitsiades
,
N.
et al. 
(
2003
)
JNK-dependent release of mitochondrial protein, smac, during apoptosis in multiple myeloma (MM) cells
.
J. Biol. Chem.
278
,
17593
17596
53
Nijboer
,
C.H.
,
Bonestroo
,
H.J.
,
Zijlstra
,
J.
,
Kavelaars
,
A.
and
Heijnen
,
C.J.
(
2013
)
Mitochondrial JNK phosphorylation as a novel therapeutic target to inhibit neuroinflammation and apoptosis after neonatal ischemic brain damage
.
Neurobiol. Dis.
54
,
432
444
54
Win
,
S.
,
Than
,
T.A.
,
Min
,
R.W.M.
,
Aghajan
,
M.
and
Kaplowitz
,
N.
(
2016
)
c-Jun N-terminal kinase mediates mouse liver injury through a novel Sab (SH3BP5)-dependent pathway leading to inactivation of intramitochondrial Src
.
Hepatology
63
,
1987
2003
55
Schroeter
,
H.
,
Boyd
,
C.S.
,
Ahmed
,
R.
,
Spencer
,
J.P.
,
Duncan
,
R.F.
,
Rice-Evans
,
C.
et al. 
(
2003
)
c-Jun N-terminal kinase (JNK)-mediated modulation of brain mitochondria function: new target proteins for JNK signalling in mitochondrion-dependent apoptosis
.
Biochem. J.
372
,
359
369
56
Win
,
S.
,
Than
,
T.A.
,
Fernandez-Checa
,
J.C.
and
Kaplowitz
,
N.
(
2014
)
JNK interaction with Sab mediates ER stress induced inhibition of mitochondrial respiration and cell death
.
Cell Death Dis.
5
,
e989
57
Win
,
S.
,
Than
,
T.A.
,
Han
,
D.
,
Petrovic
,
L.M.
and
Kaplowitz
,
N.
(
2011
)
c-Jun N-terminal kinase (JNK)-dependent acute liver injury from acetaminophen or tumor necrosis factor (TNF) requires mitochondrial Sab protein expression in mice
.
J. Biol. Chem.
286
,
35071
35078
58
Zhou
,
Q.
,
Lam
,
P.Y.
,
Han
,
D.
and
Cadenas
,
E.
(
2008
)
c-Jun N-terminal kinase regulates mitochondrial bioenergetics by modulating pyruvate dehydrogenase activity in primary cortical neurons
.
J. Neurochem.
104
,
325
335
59
Zhou
,
Q.
,
Lam
,
P.Y.
,
Han
,
D.
and
Cadenas
,
E.
(
2009
)
Activation of c-Jun-N-terminal kinase and decline of mitochondrial pyruvate dehydrogenase activity during brain aging
.
FEBS Lett.
583
,
1132
1140
60
Yuan
,
Z.Q.
,
Feldman
,
R.I.
,
Sussman
,
G.E.
,
Coppola
,
D.
,
Nicosia
,
S.V.
and
Cheng
,
J.Q.
(
2003
)
AKT2 inhibition of cisplatin-induced JNK/p38 and Bax activation by phosphorylation of ASK1: implication of AKT2 in chemoresistance
.
J. Biol. Chem.
278
,
23432
23440
61
Zhang
,
R.
,
Luo
,
D.
,
Miao
,
R.
,
Bai
,
L.
,
Ge
,
Q.
,
Sessa
,
W.C.
et al. 
(
2005
)
Hsp90–Akt phosphorylates ASK1 and inhibits ASK1-mediated apoptosis
.
Oncogene
24
,
3954
62
Wu
,
D.-N.
,
Pei
,
D.-S.
,
Wang
,
Q.
and
Zhang
,
G.-Y.
(
2006
)
Down-regulation of PTEN by sodium orthovanadate inhibits ASK1 activation via PI3-K/Akt during cerebral ischemia in rat hippocampus
.
Neurosci. Lett.
404
,
98
102
63
Vivanco
,
I.
,
Palaskas
,
N.
,
Tran
,
C.
,
Finn
,
S.P.
,
Getz
,
G.
,
Kennedy
,
N.J.
et al. 
(
2007
)
Identification of the JNK signaling pathway as a functional target of the tumor suppressor PTEN
.
Cancer Cell
11
,
555
569
64
Tournier
,
C.
,
Hess
,
P.
,
Yang
,
D.D.
,
Xu
,
J.
,
Turner
,
T.K.
,
Nimnual
,
A.
et al. 
(
2000
)
Requirement of JNK for stress- induced activation of the cytochrome c-mediated death pathway
.
Science
288
,
870
874
65
Dhanasekaran
,
D.N.
and
Reddy
,
E.P.
(
2008
)
JNK signaling in apoptosis
.
Oncogene
27
,
6245
6251
66
Srivastava
,
R.K.
,
Mi
,
Q.-S.
,
Hardwick
,
J.M.
and
Longo
,
D.L.
(
1999
)
Deletion of the loop region of Bcl-2 completely blocks paclitaxel-induced apoptosis
.
Proc. Natl Acad. Sci. U.S.A.
96
,
3775
3780
67
Yamamoto
,
K.
,
Ichijo
,
H.
and
Korsmeyer
,
S.J.
(
1999
)
BCL-2 is phosphorylated and inactivated by an ASK1/Jun N-terminal protein kinase pathway normally activated at G2/M
.
Mol. Cell. Biol.
19
,
8469
8478
68
Lei
,
K.
,
Nimnual
,
A.
,
Zong
,
W.X.
,
Kennedy
,
N.J.
,
Flavell
,
R.A.
,
Thompson
,
C.B.
et al. 
(
2002
)
The Bax subfamily of Bcl2-related proteins is essential for apoptotic signal transduction by c-Jun NH2-terminal kinase
.
Mol. Cell. Biol.
22
,
4929
4942
69
Marani
,
M.
,
Tenev
,
T.
,
Hancock
,
D.
,
Downward
,
J.
and
Lemoine
,
N.R.
(
2002
)
Identification of novel isoforms of the BH3 domain protein bim which directly activate Bax to trigger apoptosis
.
Mol. Cell. Biol.
22
,
3577
3589
70
Donovan
,
N.
,
Becker
,
E.B.E.
,
Konishi
,
Y.
and
Bonni
,
A.
(
2002
)
JNK phosphorylation and activation of BAD couples the stress-activated signaling pathway to the cell death machinery
.
J. Biol. Chem.
277
,
40944
40949
71
Tsuruta
,
F.
,
Sunayama
,
J.
,
Mori
,
Y.
,
Hattori
,
S.
,
Shimizu
,
S.
,
Tsujimoto
,
Y.
et al. 
(
2004
)
JNK promotes Bax translocation to mitochondria through phosphorylation of 14-3-3 proteins
.
EMBO J.
23
,
1889
1899
72
Sarosiek
,
K.A.
,
Chi
,
X.
,
Bachman
,
J.A.
,
Sims
,
J.J.
,
Montero
,
J.
,
Patel
,
L.
et al. 
(
2013
)
BID preferentially activates BAK while BIM preferentially activates BAX, affecting chemotherapy response
.
Mol. Cell
51
,
751
765
73
Madesh
,
M.
,
Antonsson
,
B.
,
Srinivasula
,
S.M.
,
Alnemri
,
E.S.
and
Hajnóczky
,
G.
(
2002
)
Rapid kinetics of tBid-induced cytochrome c and Smac/DIABLO release and mitochondrial depolarization
.
J. Biol. Chem.
277
,
5651
5659
74
Deng
,
Y.
,
Ren
,
X.
,
Yang
,
L.
,
Lin
,
Y.
and
Wu
,
X.
(
2003
)
A JNK-dependent pathway is required for TNFα-induced apoptosis
.
Cell
115
,
61
70

Author notes

*

These authors contributed equally to work described in this manuscript.

Supplementary data