Purpose: To show the functional, clinical, and biological significance of c-Jun-NH2-kinase (JNK)-1 in ovarian carcinoma.

Experimental Design: Analysis of the impact of JNK on 116 epithelial ovarian cancers was conducted. The role of JNK in vitro and in experimental models of ovarian cancer was assessed. We studied the role of N-5-[4-(4-methyl piperazine methyl)-benzoylamido]-2-methylphenyl-4-[3-(4-methyl)-pyridyl]-2-pyrimidine amine (WBZ_4), a novel JNK inhibitor redesigned from imatinib based on targeting wrapping defects, in cell lines and in experimental models of ovarian cancer.

Results: We found a significant association of pJNK with progression-free survival in the 116 epithelial ovarian cancers obtained at primary debulking therapy. WBZ_4 led to cell growth inhibition and increased apoptosis in a dose-dependent fashion in four ovarian cancer cell lines. In vivo, whereas imatinib had no effect on tumor growth, WBZ_4 inhibited tumor growth in orthotopic murine models of ovarian cancer. The antitumor effect was further increased in combination with docetaxel. Silencing of JNK-1 with systemically administered siRNA led to significantly reduced tumor weights compared with nonsilencing siRNA controls, indicating that indeed the antitumor effects observed were due to JNK-1 inhibition.

Conclusions: These studies identify JNK-1 as an attractive therapeutic target in ovarian carcinoma and that the redesigned WBZ_4 compound should be considered for further clinical development. Clin Cancer Res; 16(1); 184–94

Translational Relevance

We show here the functional, clinical, and biological significance of JNK-1 in ovarian carcinoma. Moreover, we provide evidence on the mechanisms of action and antitumor effects of a novel inhibitor of c-Jun-NH2-kinase that merits further clinical development.

c-Jun-NH2-kinases (JNK) or stress-activated protein kinases are members of the mitogen-activated protein kinase family. Of the three isoforms that have been described, JNK1 and JNK2 are ubiquitously expressed whereas JNK3 is only present in the brain, heart, and testes. Controversial effects have been reported for JNKs (13). JNKs have been shown to serve as mediators of apoptosis in response to cellular stress (17), have been show to sustain cell proliferation, and have been shown in survival in response to extracellular stimuli such as cytokines (1). JNKs are serine/threonine protein kinases that can be activated by a variety of stimuli including environmental stress (UV and ionizing radiation, heat shock, osmotic or redox shock), inflammatory cytokines, and growth factors. Induction by their own phosphorylation leads to the activation of different transcription factors including c-Jun and JunD. This induction, depending on the cellular context, can have effects at the level of apoptosis, differentiation, survival, and carcinogenesis. Based on gene disruption experiments, it was concluded that JNK1 and JNK2 have overlapping functions (810). However, Sabapathy et al. (11) showed that in fibroblasts, JNK1 has a prominent role in c-Jun activation leading to cell proliferation, whereas JNK2 inhibits cell proliferation by promoting c-Jun degradation in unstimulated cells.

Recent findings suggest that JNK mediates oncogenic functions in several cancer types including head and neck, gastric, and breast cancers, and melanoma (1219); thus, it may be an attractive therapeutic target. Gross and colleagues (15) showed that JNK inhibition, mediated by the competitive inhibitor SP600125 or by specific siRNA, inhibited growth of head and neck squamous cell carcinoma in vitro and in vivo (15). The antitumor effects were thought be mediated through effects on both tumor and endothelial cells. Specifically, JNK1 but not JNK2 has been shown to be important in promoting cell survival by controlling cell cycle arrest and apoptosis (12). Little is known about the role of JNK1 in ovarian cancer growth; therefore, we examined the potential of JNK1 as a therapeutic target in the current study. We used a novel small molecule inhibitor, N-5-[4-(4-methyl piperazine methyl)-benzoylamido]-2-methylphenyl-4-[3-(4-methyl)-pyridyl]-2-pyrimidine amine (WBZ_4), which was designed, synthesized, and characterized by our team to curb the potential for side effects of the original compound, imatinib (20). WBZ_4 was designed to target both c-Kit and JNK1 while avoiding the Abl-kinase. The redesign process was based on comparing the residence time of water molecules that solvate the interfacial aligned residues across known targets of imatinib. Nonconserved sites with low residence time that had a higher propensity to dehydrate (dewetting hotspots) were used to build blueprints for each target. By adding a methyl group to the original imatinib, we were able to attack a dewetting hotspot present in c-Kit and JNK1 but absent from BCR-Abl, thereby increasing the specificity of the drug toward the desired targets (20, 21). Here, we report on the biological effects of JNK1 inhibition, both in vitro and in vivo using experimental models of ovarian cancer.

Cells and culture conditions

The human ovarian epithelial cancer cells IGROV1, SKOV3ip1, SKOV3.TR, HeyA8, HEYA8.MDR, A2780PAR, and A2780CP20 cells have been described elsewhere (2224). All tumor cell lines were regularly screened for Mycoplasma using MycoAlert (Cambrex Bioscience) as described by the manufacturer. Cells were maintained in RPMI 1640 supplemented with 10% fetal bovine serum in 5% CO2/95% air at 37°C. In vitro assays were done at 70% to 85% cell density.

Drugs and reagents

The design and total synthesis of WBZ_4 has been described previously (21). For in vivo therapy, WBZ_4 and docetaxel (Aventis Pharma) were reconstituted in Ca2+- and Mg2+-free PBS. For in vitro testing, docetaxel, WBZ_4, and the JNK inhibitor SP600125 (Calbiochem) were reconstituted in 100% DMSO. When added to the cells, the final concentration of DMSO on the culture medium was 0.1% or less.

Western blot

Ovarian cancer cells were treated with 10 μmol/L WBZ_4 and collected at various time points. To activate JNK, A2780CP20 cells were seeded on 10-cm plates at 1 × 106 cells per plate. After attaching overnight, cells were treated with 10 μmol/L WBZ_4 or JNK inhibitor SP600125 (Calbiochem) and were incubated for 8 h. Next, medium was removed and saved, and the plates were washed twice with PBS. After the last wash, cells were exposed to 20, 40, and 80 J/m2 of UV light using an FB-UVXL-1000UV cross-linker (Fisher Scientific). After a 30-min recovery, cells were harvested and processed for Western blot analysis.

For p-JNK and p-c-JUN, protein lysates were separated on 10% SDS-PAGE gels and blotted onto nitrocellulose membranes (Life Science Research; Bio-Rad). Primary antibodies used were anti p-JNK (Thr183/Tyr185), anti-JNK, anti–p-c-JUN (Ser 61), anti c-JUN (Cell Signaling), and anti β-actin monoclonal antibody (Sigma-Aldrich). Secondary antibodies used were goat horseradish peroxidase (HRP) anti-rabbit and goat HRP anti-mouse (Cell Signaling).

In vitro cell viability assay

Cells were seeded on 96-well plates at 2 × 103 cells per well and incubated overnight. The next day, cells were treated with different concentrations of WBZ_4 (1, 5, 7.5, 10, and 20 μmol/L), docetaxel (0.005, 0.01, and 0.05 μmol/L), and/or the combination of both drugs and were incubated for 72 h. MTS (Promega) assay was done according to the manufacturers instructions. Plates were read at 490 nm in an ELISA plate reader (Kinetic Microplate Reader; Molecular Devices Corp.). The results are expressed in terms of percent of growth inhibition with respect to the untreated control.

Assessment of cell apoptosis

To determine cell death, cells were plated 24 h before treating them with 5, 10, and 20 μmol/L of WBZ_4. Cells were incubated further for 72 h and then collected. FITC Annexin-V and propidium iodide (PI) staining was detected by flow cytometry analysis (FACScan/CellQuest system, Becton Dickinson), measuring emission at 530 nm (FL1) and 575 nm (FL3).

Orthotopic tumor implantation

Female athymic nude mice (NCr-nu, 8- to 12-wk-old.) were purchased from Taconic. To generate tumors, HeyA8 (3 × 105 cells/0.1 mL HBSS) and A2780-CP20 (1 × 106 cells/0.1 mL HBSS) were injected into the peritoneal cavity of nude mice. Mice (n = 10 per group for WBZ_4 experiment and 5 per group for the JNK-siRNA) were monitored for adverse effects of therapy and sacrificed on day 24 or when any of the mice began to appear moribund. Mouse weight, tumor weight, and tumor distribution were recorded.

Therapy experiments in nude mice

To determine the optimal dose of WBZ_4 needed to achieve antitumor effects, nude mice bearing HeyA8 tumors were randomly divided into four groups (n = 10 per group): PBS alone or 75, 150, or 300 mg/kg of WBZ_4 by gavage (p.o.). All therapeutic experiments were started 7 d after tumor cell injection. Once the mice in the control group became moribund (∼25 d), mice were sacrificed and the therapeutic effect of the different doses was evaluated in terms of tumor weight and number of nodules. To determine the therapeutic efficacy of WBZ_4 in combination with docetaxel, nude mice bearing HeyA8 or A2780CP20 tumors were randomly divided into four groups (n = 10 per group): (a) PBS (p.o.) daily; (b) docetaxel (50 μg/mouse for HeyA8-bearing mice and 35 μg/mouse for the A2780CP20-bearing mice) weekly i.p.; (c) WBZ_4 (p.o.) daily; and (d) WBZ_4 (p.o.) daily plus docetaxel i.p. weekly. For comparative studies between WBZ_4 and imatinib, nude mice bearing HeyA8 tumors were randomly split into six groups (n = 10 per group): (a) vehicle control (PBS); (b) imatinib (50 μg/mouse p.o. daily); (c) WBZ_4 (p.o.) daily; (d) 50 μg/mouse docetaxel (i.p) weekly; (e) 50 μg/mouse imatinib daily + 50 μg/mouse docetaxel weekly; and (f) WBZ_4 daily + 50 μg/mouse docetaxel weekly.

The therapeutic activity of JNK1 siRNA was evaluated using HeyA8 or SKOV3ip1 tumor cells injected i.p. Seven days later, the mice [HEYA8 (n = 5) and SKOV3ip1 (n = 10)] were randomly assigned to four treatment groups (all treatments were administered i.p.): (a) control siRNA-DOPC (5 μg/mouse) twice weekly; (b) control siRNA-DOPC (5 μg/mouse) twice weekly plus docetaxel (50 μg/mouse) weekly; (c) JNK1 siRNA-DOPC (5 μg/mouse) twice weekly; and (d) JNK1 siRNA-DOPC (5 μg/mouse) in combination with 50 μg/mouse docetaxel.

SiRNA incorporation into 1,2-dioleoyl-sn-glycero-3-phosphocholine liposomes

For in vivo delivery, the control and JNK1 siRNAs were incorporated into 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC) liposomes (Avanti Polar Lipids). SiRNA and DOPC were mixed in the presence of excess t-butanol at a ratio of 1:10 (w/w) as described previously (25). After Tween 20 was added, the mixture was frozen in an acetone-dry ice bath and was lyophilized. Before in vivo administration, the lyophilized powder was hydrated with Ca2+- and Mg2+-free PBS at a concentration of 25 μg/mL to achieve the desired dose in 200 μL per injection.

Immunohistochemistry

Expression of Ki67 and assessment of deoxynucleotidyl transferase-mediated dUTP nick end labeling (TUNEL)–positive cells were determined by immunohistochemical analysis using paraffin-embedded tumors as described previously (26). Ki67 primary antibody (BioCare Medical) and secondary goat HRP anti-rabbit (The Jackson Laboratory) were used. Immunohistochemistry for CD31 was done on freshly cut frozen tissue as described before (26). Primary antibody used was anti-CD31 (platelet/endothelial cell adhesion molecule-1, rat IgG; Pharmingen). Secondary antibody was goat HRP anti-rat (The Jackson Laboratory). Staining was analyzed with a ×10 objective on a Microphot-FX microscope (Nikon) equipped with a three-chip charge-coupled device color video camera (model DXC990, Sony).

To quantify microvessel density (MVD), 10 random fields at ×100 final magnification were examined for each tumor (one slide per mouse, five slides per group) and the number of microvessels per field was counted. To quantify Ki67 expression, the number of positive and negative cells (3,3′-diaminobenzidine staining) was counted in 10 random fields at ×100 magnification and the percent of Ki67-positive cells was calculated for each group. To quantify TUNEL-positive cells, the number of positive cells was counted in 10 random fields at ×100 magnification.

SiRNA transfection

Predesigned siRNA targeted against JNK1 [S 5′ GGAUGCAAAUCUUUGCCAA (dT) (dT); AS 5′ UUGGCAAAGAUUUGCAUCC (dT) (dT)] as well as the control siRNA (25) were purchased from Sigma-Aldrich). For transfection, cells were seeded on six-well plates at 8 × 104 cells per well and allowed to attach overnight. The next day, JNK1 and control siRNA were transfected using Lipofectamine 2000 (Invitrogen) following the manufacturer's instructions. Target downmodulation was confirmed by Western blot. To determine an effect on cell growth and apoptosis, cells were transfected and incubated for 72 h. Next, total cell number was counted and the results expressed as % of growth inhibition or percent of annexin V–positive cells (apoptosis) relative to the untransfected cells.

Analysis of ovarian cancer clinical specimens

One hundred and sixteen epithelial ovarian cancers were obtained at primary debulking surgery and protein was extracted from these tumors as previously described (2729). The median patient age was 61 y (range, 38-88). Of the 116 tumors, 92 were serous, 5 endometrioid, 3 clear cell, and 16 others. Of the tumors, 10 were stage 1, 16 were stage 2, 66 were stage 3, 18 were stage 4, and 6 were unknown for stage. The diagnosis stage was 1 in 10 cases, 2 in 11, 3 in 66, and 4 in 18 cases (remainder unknown), and the tumor was suboptimally debulked in 32 cases. The tumor lysates were subjected to reverse phase protein array with antibodies to JNK (Santa Cruz, Inc.) and phosphorylated JNK (Cell Signaling, Inc.) as previously described (2729). Note that these antibodies are not JNK isoform specific. The raw spot signal intensities from quantification of the reverse phase protein array slide images were processed by the R package SuperCurve (version 0.997; ref. 30). The protein concentrations for JNK and pJNK were then normalized by loading correction and the data were log2 centered as previously described (2729). Kaplan-Meier survival curves were used to determine the association between the concentrations of JNK and pJNK and survival end points [both progression-free (PFS) and overall survival times, both measured from the time of debulking surgery]. Cox proportional hazards models were applied to clinical and antibody variables that were significantly associated with outcome measures in univariate analysis.

Statistical analysis

For the in vivo experiments, differences in continuous variables (mean body weight, tumor weight, MVD, Ki-67, and TUNEL staining) were analyzed using the Student's t test for comparing two groups and by ANOVA for multiple group comparisons, with P < 0.05 considered statistically significant.

WBZ_4 inhibits JNK basal activity in ovarian cancer cells

We first used Western blot analysis to examine phospho-JNK-1 and 2 and JNK-1 and JNK-2 levels in different ovarian cancer cell lines (Supplementary Fig. S1). Then, we tested the efficacy of WBZ_4 for inhibiting JNK-1 expression in some of those cells. A2780CP20 cells were treated with WBZ_4 (10 μmol/L) and collected at different time points up to 24 hours after the treatment. Untreated cells and cells treated with 0.1% DMSO showed the presence of pJNK1/2 at low levels (Fig. 1A). Starved cells given serum shock were used as a positive control for pJNK activation, showing indeed an increase in the levels of pJNK1/2. Treatment with WBZ_4 resulted in a sustained decrease in pJNK1/2 levels noticeable as soon as 2 hours after the treatment and lasting for up to 12 hours. No changes were noted in total JNK1/2 levels.

Fig. 1.

WBZ_4 inhibits JNK activity in ovarian cancer cells. A, basal p-JNK levels in cells treated with WBZ_4. Protein extracts from A2780CP20 cells treated with 10 μmol/L WBZ_4 and collected at different time points were probed with pJNK, total JNK, and β-actin. U, untreated sample; D, vehicle treated sample; S, serum-shocked positive control. B, UV-dependent JNK activation and c-Jun phosphorylation. Protein extracts from cells treated with 10 μmol/L of WBZ_4 or SP600125 and exposed to UV lights were probed with antibodies against the phosphorylated and total form of JNK and c-Jun. To show equal loading, blots were also probed with β-actin. UV, level of energy in J/m2; WBZ, WBZ_4; SP, SP600125. C and D, densitometric analysis of blots shown in B.

Fig. 1.

WBZ_4 inhibits JNK activity in ovarian cancer cells. A, basal p-JNK levels in cells treated with WBZ_4. Protein extracts from A2780CP20 cells treated with 10 μmol/L WBZ_4 and collected at different time points were probed with pJNK, total JNK, and β-actin. U, untreated sample; D, vehicle treated sample; S, serum-shocked positive control. B, UV-dependent JNK activation and c-Jun phosphorylation. Protein extracts from cells treated with 10 μmol/L of WBZ_4 or SP600125 and exposed to UV lights were probed with antibodies against the phosphorylated and total form of JNK and c-Jun. To show equal loading, blots were also probed with β-actin. UV, level of energy in J/m2; WBZ, WBZ_4; SP, SP600125. C and D, densitometric analysis of blots shown in B.

Close modal

UV induction has been shown to activate JNK phosphorylation as well as that of its downstream targets (1, 4, 31). To determine whether WBZ_4 can block JNK activation, A2780CP20 cells were treated with 10 μmol/L of either WBZ_4 or SP600125 for 8 hours followed by exposure to UV light (20, 40, and 80 J/m2). Western blot analysis was done on the cell extracts to assess the levels of both p-JNK and p-c-Jun as indicators of JNK activity (Fig. 1B). JNK was readily activated in untreated A2780CP20 cells as shown by increased levels of p-JNK and p-c-Jun. However, at UV levels of 20 and 40 J/m2, WBZ_4 and SP600125 prevented phosphorylation of both JNK and c-Jun. Particularly, when the UV levels were 40 J/m2, the addition of WBZ_4 or SP600125 reduced p-JNK levels by 5-fold (P < 0.001). Under the same UV levels, the addition of SP600125 or WBZ_4 reduced p-c-Jun levels by 2.5- and 3-fold (P < 0.01), respectively (Fig. 1C-D). At 80 J/m2, neither drug was capable of inhibiting JNK activity, suggesting that this high level of UV overrides any type of inhibition occurring in the cells.

WBZ_4 inhibit growth of ovarian cancer cell lines

On the basis of the JNK1 inhibitory activity of WBZ_4, we next asked whether this inhibitor would affect in vitro growth of ovarian cancer cells. Following treatment for 72 hours, dose-dependent growth inhibition was observed in all the cell lines tested with IC50 values ranging between 7.5 and 15 μmol/L (Fig. 2A). To assess whether the effects on cell viability were related to apoptosis, fluorescence-activated cell sorting (FACS) analysis after FITC-Annexin V staining was performed. This analysis showed a dose-dependent increase in the percentage of cells labeled positively with Annexin V, indicating that the growth inhibition observed was likely a result of cellular apoptosis (Fig. 2B).

Fig. 2.

WBZ_4 and the JNK inhibitor SP600125 inhibit growth of ovarian cancer cells. A, percentage growth inhibition of ovarian cancer cells versus different concentrations of WBZ_4. Cells were treated with the drug for 72 h followed by MTS assay. B, percentage of FITC-Annexin V–positive cells. A2780CP20 cells were treated with WBZ_4 and 72 h later, Annexin V labeling was assayed by FACS: 5 μmol/L (*, P < 0.05), 10 μmol/L (***, P < 0.001), and 20 μmol/L (***, P < 0.001). C, percentage growth inhibition versus WBZ_4 and SP600125. Cell viability was assessed by MTS assay 72 h after treatment. A2780CP20: 5 μmol/L WBZ_4 (*, P < 0.05), 10 μmol/L WBZ_4 (***, P < 0.001), and 20 μmol/L WBZ_4 (***, P < 0.001); 5 μmol/L SP600125 (*, P < 0.05), 10 μmol/L SP600125 (**, P < 0.01), and 20 μmol/L SP600125 (**, P < 0.01). HEYA8: 5 μmol/L WBZ_4 (P > 0.05), 10 μmol/L WBZ_4 (*, P < 0.05), and 20 μmol/L WBZ_4 (***, P < 0.001); 5 μmol/L SP600125 (*, P < 0.05), 10 μmol/L SP600125 (**, P < 0.01), and 20 μmol/L SP600125 (***, P < 0.001). D, cell cycle distribution of HeyA8 and A2780CP20 cells treated with WBZ_4 and SP600125. Cell cycle was evaluated by FACS based on the propidium iodide staining of cells treated with WBZ_4 and SP600125 for 72 h. Columns, mean; bars, SEM.

Fig. 2.

WBZ_4 and the JNK inhibitor SP600125 inhibit growth of ovarian cancer cells. A, percentage growth inhibition of ovarian cancer cells versus different concentrations of WBZ_4. Cells were treated with the drug for 72 h followed by MTS assay. B, percentage of FITC-Annexin V–positive cells. A2780CP20 cells were treated with WBZ_4 and 72 h later, Annexin V labeling was assayed by FACS: 5 μmol/L (*, P < 0.05), 10 μmol/L (***, P < 0.001), and 20 μmol/L (***, P < 0.001). C, percentage growth inhibition versus WBZ_4 and SP600125. Cell viability was assessed by MTS assay 72 h after treatment. A2780CP20: 5 μmol/L WBZ_4 (*, P < 0.05), 10 μmol/L WBZ_4 (***, P < 0.001), and 20 μmol/L WBZ_4 (***, P < 0.001); 5 μmol/L SP600125 (*, P < 0.05), 10 μmol/L SP600125 (**, P < 0.01), and 20 μmol/L SP600125 (**, P < 0.01). HEYA8: 5 μmol/L WBZ_4 (P > 0.05), 10 μmol/L WBZ_4 (*, P < 0.05), and 20 μmol/L WBZ_4 (***, P < 0.001); 5 μmol/L SP600125 (*, P < 0.05), 10 μmol/L SP600125 (**, P < 0.01), and 20 μmol/L SP600125 (***, P < 0.001). D, cell cycle distribution of HeyA8 and A2780CP20 cells treated with WBZ_4 and SP600125. Cell cycle was evaluated by FACS based on the propidium iodide staining of cells treated with WBZ_4 and SP600125 for 72 h. Columns, mean; bars, SEM.

Close modal

WBZ_4 was originally designed to specifically inhibit c-Kit and JNK1 to avoid potential cardiotoxicity (21). Given that ovarian cancer cells do not express detectable levels of c-Kit (data not shown), we reasoned that WBZ_4 was most likely functioning through JNK-1 inhibition. The JNK inhibitor SP600125 was used as a positive control. As shown in Fig. 2C, ovarian cancer cells treated with increasing concentrations of SP600125 for 72 hours showed dose-dependent growth inhibition (IC50 levels between 25 and 30 μmol/L). The IC50s observed with SP600125 were similar to those published for other tumor cell lines (12, 15, 19, 32).

Given the effect established for the SP600125 JNK inhibitor in arresting cells at the G2-M phase of the cell cycle (33), we studied the effects of both SP600125 and WBZ_4 on cell cycle distribution of the ovarian cancer cells. Treatment of either A2780 or HeyA8 cells with 15 μmol/L of SP600125 led to a decrease in the percentage of cells in the G1 phase and an increase in the cells at the S and G2-M phases of the cell cycle (Fig. 2D) in accordance with the published data (20). However, in the HeyA8 cells, WBZ_4 induced a clear increase in the percentage of cells in the sub-G1 population (dead cells) at the two concentrations tested (10 and 15 μmol/L), whereas in the A2780CP20 cells, 10 μmol/L of WBZ_4 led to arrest of the cells at the G2-M. Higher concentrations (15 μmol/L) resulted in an ∼20% increase in the sub-G1 population with a concomitant decrease in the percentage of cells at the G1 phase of the cell cycle.

Therapeutic efficacy of WBZ_4

To evaluate in vivo the therapeutic effect of the combination of WBZ_4 and docetaxel (3439), we first performed preliminary dose-response experiments for WBZ_4. Nude mice bearing HeyA8 tumors were treated daily with one of three different doses of WBZ_4 (75, 150, and 300 mg/kg) for ∼3 weeks. Assessment of tumor weight and number of nodules at the end of the experiment revealed that the 150 mg/kg dose gave the most consistent results, which was selected for subsequent experiments (Supplementary Fig. S2). The in vivo antitumor activity of WBZ_4 was determined in nude mice implanted with either HeyA8 or A2780CP20 cells (Fig. 3). Seven days later, mice were randomly distributed into four treatment groups: (a) control (PBS), (b) WBZ_4 (gavage), (c) docetaxel (i.p.), and 4) WBZ_4 (gavage) plus docetaxel (i.p.). The in vitro effect of docetaxel in these cell lines alone or in combination with WBZ_4 is shown in Supplementary Fig. S3. In the HeyA8-tumor bearing mice, comparison of the mean tumor weights across different groups revealed that WBZ_4 and docetaxel resulted in a statistically significant reduction in mean tumor weight when compared with the PBS-treated mice (P < 0.01), and this reduction was increased even further when both drugs were combined (P < 0.001). Each drug alone also induced a statistically significant decrease in the number of nodules (P < 0.05), which were further decreased when both drugs were combined (P < 0.01; Fig. 3A). In the A2780CP20 model (Fig. 3B), WBZ_4 and the combination of WBZ_4 plus docetaxel were effective in reducing tumor weight compared with controls (P < 0.05). In this model, each treatment induced a reduction in the number of nodules compared with controls (P < 0.05 for the combination group).

Fig. 3.

Therapeutic efficacy of WBZ_4 in combination with docetaxel. HeyA8 (A) or A2780CP20 (B) cells were implanted i.p. as described in Materials and Methods. Mice were randomly allocated to one of the following groups, with therapy beginning 1 wk after tumor cell inoculation: PBS, PBS+ docetaxel (DOC), WBZ_4, and WBZ_4 + docetaxel. Left columns, number of nodules; center columns, mean tumor weights (bars, SD). Right columns, individual weights. C, WBZ_4 therapeutic activity was compared with that of imatinib in HeyA8 tumor–bearing mice. Nude mice were injected i.p. with HeyA8 cells and randomly allocated to one of the following groups, with therapy beginning 1 wk after tumor cell inoculation: vehicle, imatinib, WBZ_4, PBS+ docetaxel, WBZ_4 + docetaxel, and WBZ_4 + imatinib. D, pJNK levels in tumor samples treated with vehicle control, WBZ_4, or imatinib (three mice per group are shown). *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Fig. 3.

Therapeutic efficacy of WBZ_4 in combination with docetaxel. HeyA8 (A) or A2780CP20 (B) cells were implanted i.p. as described in Materials and Methods. Mice were randomly allocated to one of the following groups, with therapy beginning 1 wk after tumor cell inoculation: PBS, PBS+ docetaxel (DOC), WBZ_4, and WBZ_4 + docetaxel. Left columns, number of nodules; center columns, mean tumor weights (bars, SD). Right columns, individual weights. C, WBZ_4 therapeutic activity was compared with that of imatinib in HeyA8 tumor–bearing mice. Nude mice were injected i.p. with HeyA8 cells and randomly allocated to one of the following groups, with therapy beginning 1 wk after tumor cell inoculation: vehicle, imatinib, WBZ_4, PBS+ docetaxel, WBZ_4 + docetaxel, and WBZ_4 + imatinib. D, pJNK levels in tumor samples treated with vehicle control, WBZ_4, or imatinib (three mice per group are shown). *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Close modal

Because WBZ_4 was redesigned from imatinib, we also asked whether the activity of WBZ_4 is distinct from imatinib against ovarian carcinoma. In the HeyA8 model, treatment with imatinib did not affect tumor growth (Fig. 3C). However, WBZ_4 monotherapy reduced tumor growth by 50% (P = 0.03) and the combination of WBZ_4 and docetaxel reduced tumor growth by 90% (P < 0.001; Fig. 3C). Tumor samples harvested from this experiment confirmed that WBZ_4 reduced the levels of pJNK whereas imatinib had no effect on p-JNK (Fig. 3D).

Effect of WBZ_4 therapy on angiogenesis, cell proliferation, and apoptosis

Next, we examined the effects of WBZ_4 therapy on cell proliferation, apoptosis, and MVD. WBZ_4 or docetaxel treatment resulted in decreased MVD when compared with controls (Fig. 4A), which was further decreased by combination treatment. Cell proliferation was assessed by staining for the nuclear marker Ki-67. WBZ_4 induced a significant decrease in the Ki-67 index (P < 0.05; Fig. 4B), which was decreased further in the combination group (P < 0.001). TUNEL staining showed similar effects on apoptosis in response to therapy (Fig. 4C).

Fig. 4.

Effect of WBZ_4 with or without docetaxel on cell proliferation, apoptosis, and MVD. A, MVD was determined after immunohistochemical peroxidase staining for CD31. B, Ki67 immunohistochemistry. Original magnification, ×100. Columns, mean percentage of Ki67-positive cells; bars, SD. Five animals per group and at least five fields per animal were counted. C, TUNEL immunohistochemistry. Tumor sections from each group were stained for TUNEL. Representative slides from each group are shown. The number of apoptotic cells was counted. Columns, mean number of TUNEL-positive cells; bars, SD. Ten fields per slide and at least five slides per group (all from different animals) were counted.

Fig. 4.

Effect of WBZ_4 with or without docetaxel on cell proliferation, apoptosis, and MVD. A, MVD was determined after immunohistochemical peroxidase staining for CD31. B, Ki67 immunohistochemistry. Original magnification, ×100. Columns, mean percentage of Ki67-positive cells; bars, SD. Five animals per group and at least five fields per animal were counted. C, TUNEL immunohistochemistry. Tumor sections from each group were stained for TUNEL. Representative slides from each group are shown. The number of apoptotic cells was counted. Columns, mean number of TUNEL-positive cells; bars, SD. Ten fields per slide and at least five slides per group (all from different animals) were counted.

Close modal

JNK1 silencing inhibits cell growth

To confirm that JNK inhibition leads to growth arrest in ovarian cancer cells, we used JNK1 siRNA to specifically target JNK1 in the HEYA8, A2780CP20, SKOV31ip1, and SKOV3.TR ovarian cancer cells (Fig. 5). Cells treated with siRNA showed a 60% to 90% decrease in JNK1 expression compared with the untreated or the control siRNA-treated cells (Fig. 5A). Compared with control siRNA, a 30% to 40% reduction in cell growth was observed in the four cell lines tested (Fig. 5B). Likewise, FACS analysis indicated that JNK1 siRNA transfection increased apoptosis in the four cell lines tested (Fig. 5C).

Fig. 5.

JNK1 siRNA inhibits growth of ovarian cancer cells. A, effect of JNK1 siRNA on JNK1 expression in different ovarian cancer cells. Western blot of extracts from cells transfected with control and JNK1 siRNA and probed with JNK and β-actin antibodies. B, effect of JNK1siRNA on cell growth. Seventy-two hours after transfection, total cell number was counted on the different groups and results are expressed as % growth inhibition relative to the cells transfected with the control siRNA. Columns, mean; bars, SD. C, cell death after JNK1 silencing. Cells transfected with JNK1 siRNA were stained with FITC Annexin V and propidium iodide, and cell death was assessed by FACS.

Fig. 5.

JNK1 siRNA inhibits growth of ovarian cancer cells. A, effect of JNK1 siRNA on JNK1 expression in different ovarian cancer cells. Western blot of extracts from cells transfected with control and JNK1 siRNA and probed with JNK and β-actin antibodies. B, effect of JNK1siRNA on cell growth. Seventy-two hours after transfection, total cell number was counted on the different groups and results are expressed as % growth inhibition relative to the cells transfected with the control siRNA. Columns, mean; bars, SD. C, cell death after JNK1 silencing. Cells transfected with JNK1 siRNA were stained with FITC Annexin V and propidium iodide, and cell death was assessed by FACS.

Close modal

Therapeutic effect of in vivo JNK1 silencing

To determine if JNK1 silencing specifically was responsible for the in vivo antitumor effects, we used JNK1 siRNA incorporated into DOPC nanoliposomes in HeyA8-bearing (n = 5) and SKOV3ip1-bearing (n = 10) nude mice according to the following groups: (a) control-siRNA DOPC liposomes, (b) JNK1-siRNA DOPC liposomes, (c) docetaxel + control-siRNA DOPC liposomes, and (d) JNK1-siRNA DOPC liposomes + docetaxel. Either JNK1 siRNA-DOPC or docetaxel significantly decreased tumor weight, which was further reduced in the combination group (P < 0.05 for HEYA8; Fig. 6B; and P < 0.001 for SKOV3ip1; Fig. 6E). JNK1 siRNA-DOPC also induced a significant decrease in the number of nodules in both HEYA8 (P < 0.05; Fig. 6A) and SKOV3ip1 (P < 0.01) models. The number of nodules was further decreased in the combination groups (HEYA8, P < 0.05; Fig. 6A; and SKOV3ip1, P < 0.001; Fig. 6D).

Fig. 6.

In vivo therapeutic efficacy of JNK1 siRNA in combination with docetaxel. Nude mice were injected i.p. with HeyA8 (A-C) or SKOV3ip1 (D-F) cells and randomly allocated to one of the following groups, with therapy beginning 1 wk after tumor cell inoculation: control siRNA DOPC, control siRNA DOPC + docetaxel, JNK1 siRNA DOPC, and JNK1 siRNA DOPC + docetaxel. A and D, number of nodules; B and E, mean tumor weight; bars, SD. C and F, individual tumor weights. G, Kaplan-Meier survival curves for PFS for all stage tumors based on pJNK (JNKp183_5) levels.

Fig. 6.

In vivo therapeutic efficacy of JNK1 siRNA in combination with docetaxel. Nude mice were injected i.p. with HeyA8 (A-C) or SKOV3ip1 (D-F) cells and randomly allocated to one of the following groups, with therapy beginning 1 wk after tumor cell inoculation: control siRNA DOPC, control siRNA DOPC + docetaxel, JNK1 siRNA DOPC, and JNK1 siRNA DOPC + docetaxel. A and D, number of nodules; B and E, mean tumor weight; bars, SD. C and F, individual tumor weights. G, Kaplan-Meier survival curves for PFS for all stage tumors based on pJNK (JNKp183_5) levels.

Close modal

JNK and pJNK expression in human ovarian carcinoma

Given the paucity of data regarding JNK expression in human ovarian carcinoma, we examined JNK and pJNK in 116 epithelial ovarian cancers obtained at primary debulking surgery by reverse phase protein array. Out of the 116 specimens, 50% (58) showed high levels of pJNK. Low and high expression were simply defined using the mean expression value for pJNK. We found that pJNK was significantly associated with PFS (Fig. 6G), but not with overall survival. JNK expression was not significantly associated with either PFS or overall survival time. Because both residual disease at debulking surgery and stage at diagnosis were also significantly associated with PFS time, a Cox model for PFS was constructed using pJNK, stage, and residual disease. In this model, both residual disease (P = 0.01) and pJNK (P = 0.04) remained significant predictors of PFS. When the survival analysis was confined to those with stages 3 or 4 cancer only, PFS was still significant (Supplementary Fig. S4). In addition, pJNK was significantly more highly expressed in stage 3 and 4 cases versus stage 1 and 2 cases (t test P = 0.024).

The key finding from our study is that activated JNK1 is present in a substantial proportion of ovarian cancers, and is predictive of decreased PFS. Inhibition of JNK-1 in ovarian cancer cells leads to growth inhibition and apoptosis resulting in antitumor efficacy in vivo. To inhibit JNK1 activity, we used the newly developed WBZ_4 compound, which suppressed both basal and UV-induced phosphorylation of JNK1 in ovarian cancer cells. WBZ_4 showed even greater therapeutic activity in combination with docetaxel.

JNKs are members of the well-known family of mitogen-activated protein kinases. Despite the controversial roles described for JNKs in mediating cell death as well as cell survival and tumorogenesis, several lines of evidences support the notion that under basal conditions, JNK is necessary to promote progression through the cell cycle and therefore to promote cell proliferation (1219). JNK was even shown to be necessary for the development of gastric cancer, indicating the tumorigenic potential of this protein and the importance of its inhibition in this model (13). Work done in other models and cell lines arrived to similar conclusions.

Given the role of several kinases in cell survival, proliferation, differentiation, apoptosis, and transformation, tyrosine, and in general, kinase inhibitors have emerged as strong agents for the treatment of different types of cancers and other diseases (40, 41). However, because of the great degree of conservation present across kinases, it is difficult to control specificity of the inhibitors and in many cases; this leads to undesired toxicity (42, 43). Here, we used WBZ_4, a compound recently derived from imatinib after applying a new conceptual model to curve its activity toward c-Kit and JNK1. We showed earlier that nonconserved sites among different targets with low water residence time have a higher propensity to dehydrate, therefore called dewetting hotspots. Given that these hotspots are present in nonconserved residues, they serve as a discriminator among paralogs. Targeting these unique dewetting hotspots allows for the design of potentially more specific kinase inhibitors (20, 21).

Earlier work using this paradigm led to the synthesis of WBZ_4, which indeed was more specific for c-Kit and was able to bind in vitro to c-Jun NH2-terminal Kinase 1 (21), a property absent from the parental compound. Previously, we showed that WBZ_4 is highly effective for the management of gastrointestinal stromal tumors due to inhibition of c-Kit activity; however, it does not have any effect on chronic myeloid leukemia where imatinib is widely used. Although imatinib lacks activity in ovarian carcinoma, WBZ_4 is highly active. This finding is likely reflective of JNK1 inhibition because ovarian cancers largely lack c-Kit expression. This contention is supported by the fact that JNK1 siRNA-DOPC was able to inhibit tumor growth to an extent that was similar to WBZ_4.

In summary, we show the functional, clinical, and biological significance of JNK-1 in ovarian carcinoma. Moreover, we provide evidence for a novel inhibitor of JNK that merits further clinical development.

No potential conflicts of interest were disclosed.

We thank Robert Langley, Donna Reynolds, Yunfang Wang, Nicholas Jennings, and the Lopez Laboratory members for helpful discussions and/or technical assistance.

Grant Support: NIH grants (CA109298, CA110793, and CA092115), The Marcus Foundation, the University of Texas M.D. Anderson Cancer Center Specialized Programs of Research Excellence in Ovarian Cancer (P50CA083639), a Program Project Development Grant from the Ovarian Cancer Research Fund, Inc., the Zarrow Foundation, and the Betty Ann Asche Murray Distinguished Professorship.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1
Davis
RJ
. 
Signal transduction by the JNK group of MAP kinases
.
Cell
2000
;
103
:
239
52
.
2
Harper
SJ
,
LoGrasso
P
. 
Signalling for survival and death in neurones: the role of stress-activated kinases, JNK and p38
.
Cell Signal
2001
;
13
:
299
310
.
3
Chang
L
,
Karin
M
. 
Mammalian MAP kinase signalling cascades
.
Nature
2001
;
410
:
37
40
.
4
Hibi
M
,
Lin
A
,
Smeal
T
,
Minden
A
,
Karin
M
. 
Identification of an oncoprotein- and UV-responsive protein kinase that binds and potentiates the c-Jun activation domain
.
Genes Dev
1993
;
7
:
2135
48
.
5
Kyriakis
JM
,
Banerjee
P
,
Nikolakaki
E
, et al
. 
The stress-activated protein kinase subfamily of c-Jun kinases
.
Nature
1994
;
369
:
156
60
.
6
Kyriakis
JM
,
Avruch
J
. 
Mammalian mitogen-activated protein kinase signal transduction pathways activated by stress and inflammation
.
Physiol Rev
2001
;
81
:
807
69
.
7
Potapova
O
,
Haghighi
A
,
Bost
F
, et al
. 
The Jun kinase/stress-activated protein kinase pathway functions to regulate DNA repair and inhibition of the pathway sensitizes tumor cells to cisplatin
.
J Biol Chem
1997
;
272
:
14041
4
.
8
Kuan
CY
,
Yang
DD
,
Samanta Roy
DR
,
Davis
RJ
,
Rakic
P
,
Flavell
RA
. 
The Jnk1 and Jnk2 protein kinases are required for regional specific apoptosis during early brain development
.
Neuron
1999
;
22
:
667
76
.
9
Sabapathy
K
,
Hu
Y
,
Kallunki
T
, et al
. 
JNK2 is required for efficient T-cell activation and apoptosis but not for normal lymphocyte development
.
Curr Biol
1999
;
9
:
116
25
.
10
Sabapathy
K
,
Jochum
W
,
Hochedlinger
K
,
Chang
L
,
Karin
M
,
Wagner
EF
. 
Defective neural tube morphogenesis and altered apoptosis in the absence of both JNK1 and JNK2
.
Mech Dev
1999
;
89
:
115
24
.
11
Sabapathy
K
,
Hochedlinger
K
,
Nam
SY
,
Bauer
A
,
Karin
M
,
Wagner
EF
. 
Distinct roles for JNK1 and JNK2 in regulating JNK activity and c-Jun-dependent cell proliferation
.
Mol Cell
2004
;
15
:
713
25
.
12
Alexaki
VI
,
Javelaud
D
,
Mauviel
A
. 
JNK supports survival in melanoma cells by controlling cell cycle arrest and apoptosis
.
Pigment Cell Melanoma Res
2008
;
21
:
429
38
.
13
Shibata
W
,
Maeda
S
,
Hikiba
Y
, et al
. 
c-Jun NH2-terminal kinase 1 is a critical regulator for the development of gastric cancer in mice
.
Cancer Res
2008
;
68
:
5031
9
.
14
Vivanco
I
,
Palaskas
N
,
Tran
C
, et al
. 
Identification of the JNK signaling pathway as a functional target of the tumor suppressor PTEN
.
Cancer Cell
2007
;
11
:
555
69
.
15
Gross
ND
,
Boyle
JO
,
Du
B
, et al
. 
Inhibition of Jun NH2-terminal kinases suppresses the growth of experimental head and neck squamous cell carcinoma
.
Clin Cancer Res
2007
;
13
:
5910
7
.
16
Potapova
O
,
Gorospe
M
,
Bost
F
, et al
. 
c-Jun N-terminal kinase is essential for growth of human T98G glioblastoma cells
.
J Biol Chem
2000
;
275
:
24767
75
.
17
Potapova
O
,
Gorospe
M
,
Dougherty
RH
,
Dean
NM
,
Gaarde
WA
,
Holbrook
NJ
. 
Inhibition of c-Jun N-terminal kinase 2 expression suppresses growth and induces apoptosis of human tumor cells in a p53-dependent manner
.
Mol Cell Biol
2000
;
20
:
1713
22
.
18
Bost
F
,
McKay
R
,
Dean
N
,
Mercola
D
. 
The JUN kinase/stress-activated protein kinase pathway is required for epidermal growth factor stimulation of growth of human A549 lung carcinoma cells
.
J Biol Chem
1997
;
272
:
33422
9
.
19
Du
L
,
Lyle
CS
,
Obey
TB
, et al
. 
Inhibition of cell proliferation and cell cycle progression by specific inhibition of basal JNK activity: evidence that mitotic Bcl-2 phosphorylation is JNK-independent
.
J Biol Chem
2004
;
279
:
11957
66
.
20
Fernandez
A
,
Sanguino
A
,
Peng
Z
, et al
. 
Rational drug redesign to overcome drug resistance in cancer therapy: imatinib moving target
.
Cancer Res
2007
;
67
:
4028
33
.
21
Fernandez
A
,
Sanguino
A
,
Peng
Z
, et al
. 
An anticancer C-Kit kinase inhibitor is reengineered to make it more active and less cardiotoxic
.
J Clin Invest
2007
;
117
:
4044
54
.
22
Sood
AK
,
Shahin
MS
,
Sorosky
JI
. 
Paclitaxel and platinum chemotherapy for ovarian carcinoma during pregnancy
.
Gynecol Oncol
2001
;
83
:
599
600
.
23
Apte
SM
,
Bucana
CD
,
Killion
JJ
,
Gershenson
DM
,
Fidler
IJ
. 
Expression of platelet-derived growth factor and activated receptor in clinical specimens of epithelial ovarian cancer and ovarian carcinoma cell lines
.
Gynecol Oncol
2004
;
93
:
78
86
.
24
Schmandt
RE
,
Bennett
M
,
Clifford
S
, et al
. 
The BRK tyrosine kinase is expressed in high-grade serous carcinoma of the ovary
.
Cancer Biol Ther
2006
;
5
:
1136
41
.
25
Landen
CN
 Jr.
,
Chavez-Reyes
A
,
Bucana
C
, et al
. 
Therapeutic EphA2 gene targeting in vivo using neutral liposomal small interfering RNA delivery
.
Cancer Res
2005
;
65
:
6910
8
.
26
Halder
J
,
Kamat
AA
,
Landen
CN
 Jr.
, et al
. 
Focal adhesion kinase targeting using in vivo short interfering RNA delivery in neutral liposomes for ovarian carcinoma therapy
.
Clin Cancer Res
2006
;
12
:
4916
24
.
27
Gonzalez-Angulo
AM
,
Stemke-Hale
K
,
Palla
SL
, et al
. 
Androgen receptor levels and association with PIK3CA mutations and prognosis in breast cancer
.
Clin Cancer Res
2009
;
15
:
2472
8
.
28
Stemke-Hale
K
,
Gonzalez-Angulo
AM
,
Lluch
A
, et al
. 
An integrative genomic and proteomic analysis of PIK3CA, PTEN, AKT mutations in breast cancer
.
Cancer Res
2008
;
68
:
6084
91
.
29
Hennessy
BT
,
Lu
Y
,
Poradosu
E
, et al
. 
Pharmacodynamic markers of perifosine efficacy
.
Clin Cancer Res
2007
;
13
:
7421
31
.
30
Hu
J
,
He
X
,
Baggerly
KA
,
Coombes
KR
,
Hennessy
BT
,
Mills
GB
. 
Non-parametric quantification of protein lysate arrays
.
Bioinformatics (Oxford, England)
2007
;
23
:
1986
94
.
31
Ip
YT
,
Davis
RJ
. 
Signal transduction by the c-Jun N-terminal kinase (JNK)-from inflammation to development
.
Curr Opin Cell Biol
1998
;
10
:
205
19
.
32
Ennis
BW
,
Fultz
KE
,
Smith
KA
, et al
. 
Inhibition of tumor growth, angiogenesis, and tumor cell proliferation by a small molecule inhibitor of c-Jun N-terminal kinase
.
J Pharmacol Exp Ther
2005
;
313
:
325
32
.
33
Oktay
K
,
Buyuk
E
,
Oktem
O
,
Oktay
M
,
Giancotti
FG
. 
The c-Jun N-terminal kinase JNK functions upstream of Aurora B to promote entry into mitosis
.
Cell Cycle (Georgetown, Tex)
2008
;
7
:
533
41
.
34
Ringel
I
,
Horwitz
SB
. 
Studies with RP 56976 (taxotere): a semisynthetic analogue of taxol
.
J Natl Cancer Inst
1991
;
83
:
288
91
.
35
Hsu
Y
,
Sood
AK
,
Sorosky
JI
. 
Docetaxel versus paclitaxel for adjuvant treatment of ovarian cancer: case-control analysis of toxicity
.
Am J Clin Oncol
2004
;
27
:
14
8
.
36
Huang
S
,
Van Arsdall
M
,
Tedjarati
S
, et al
. 
Contributions of stromal metalloproteinase-9 to angiogenesis and growth of human ovarian carcinoma in mice
.
J Natl Cancer Inst
2002
;
94
:
1134
42
.
37
Gueritte-Voegelein
F
,
Guenard
D
,
Lavelle
F
,
Le Goff
MT
,
Mangatal
L
,
Potier
P
. 
Relationships between the structure of taxol analogues and their antimitotic activity
.
J Med Chem
1991
;
34
:
992
8
.
38
Barasoain
I
,
De Ines
C
,
Diaz
F
. 
Interaction of tubulin and cellular microtubules with Taxotere (RP56976), a new semisynthetic analog of Taxol
.
Proc Am Assoc Cancer Res
1991
;
32
:
329
.
39
Vasey
PA
,
Scottish Gynecologic Cancer Trials Group
. 
Preliminary results of the SCOTROC Trial: a phase III comparison of paclitaxel-carboplatin (PC) and docetaxel-carboplatin (DC) as first-line chemotherapy for stage IC-IV epithelial ovarian cancer (EOC)
.
Proc Am Soc Clin Oncol
2001
;
20
:
202a
.
40
Fabian
MA
,
Biggs
WH
 III
,
Treiber
DK
, et al
. 
A small molecule-kinase interaction map for clinical kinase inhibitors
.
Nat Biotechnol
2005
;
23
:
329
36
.
41
Tibes
R
,
Trent
J
,
Kurzrock
R
. 
Tyrosine kinase inhibitors and the dawn of molecular cancer therapeutics
.
Annu Rev Pharmacol Toxicol
2005
;
45
:
357
84
.
42
Druker
BJ
. 
Molecularly targeted therapy: have the floodgates opened?
Oncologist
2004
;
9
:
357
60
.
43
Knight
ZA
,
Shokat
KM
. 
Features of selective kinase inhibitors
.
Chem Biol
2005
;
12
:
621
37
.

Supplementary data