Dickkopf-4 is frequently overexpressed in epithelial ovarian carcinoma and promotes tumor invasion

Frozen primary EOC tissues ( n = 33), benign epithelial ovarian tumors ( n = 33) and archival paraffin-embedded EOC samples ( n = 239) were obtained from patients with primary epithelial ovarian tumors, aged 28 to 64 years, at Shengjing Hospital (Shenyang, China), from May 2009 to April 2014. Samples were selected for the study based on the criteria and followed up as previously reported [ 19]. After surgical treatment, the 239 ovarian carcinoma patients, whose archival paraffin-embedded samples were selected for immunohistochemical analysis, were followed up from May 1, 2009 to April 30, 2014. These 239 ovarian carcinoma patients were followed from the date of first resection surgery to the date of ovarian cancer recurrence or last observation. 99 patients were recurrent, 53 patients were lost to follow-up, and 87 patients were alive. Ethical approval for human subjects was obtained from the research ethics committee of ShengJing Hospital. Informed consent was obtained from patients enrolled in this study.

Tissue RNA and cell RNA, extracted from frozen paired tissues, were reverse transcribed using a TaKaRa RNA PCR kit. Quantitative real-time PCR was performed using the real-time PCR system 7300 (Applied Biosystems). PCRs were performed in triplicate. The relative levels of mRNA were analyzed by the 2 − ΔΔCt method. ΔCt = Ct (target)-Ct(β-actin). The mean expression level of DKK4 in normal tissues was used as control and considered as a value of 1.0, as described by Jarboe [ 20]. Using the following primers:

DKK4: 5′-TGGACTTCAACAACATCAGGAG-3′(forward)

DKK4: 5′-GGTATTGCAGTCCGTGTCAG-3′(reverse).

β-actin: 5′-CTTAGTTGCGTTACACCCTTTCTTG-3′(forward);

β-actin:5′-CTGTCACCTTCACCGTTCCAGTTT-3′ (reverse).

Tissue protein was extracted from frozen tissues and cells protein was lased with sodium dodecyl sulfate buffer. Proteins were extracted with RIPA buffer. Proteins with same concentration were separated on a 10% SDS-PAGE and then transferred to PVDF membranes. The membranes were blocked with 3% BSA in Tris-buffered saline with tween (TBST) and incubated with primary antibody DKK4 (1:500, R&D), anti-c-jun (1:400) (Abcam), anti-p-c-jun (1:350) (Santa Cruz), anti-JNK1/2 (1:400) (Abcam), anti-p-JNK (1:500) (Santa Cruz) followed by incubation with secondary antibody. Protein expression was visualized using enhanced chemiluminescence. Comparison between different groups were made by determining the specific protein/β-actin ratio of the immunoreactive area with densitometry. The tests were performed in triplicate.

Samples were fixed by formalin, embedded in paraffin and cut into sections (4–6 μm thick). Then the sections were deparaffnized, and stained using a streptavidin-biotin immunoperoxidase technique. The nucleus was stained with hematoxylin. DKK4 antibody (Abcam) was diluted as 1:250. Staining was scored as negative, weak, or strong, and rated by by multiplication of the intensity and positivity scores. (intensity: no staining, score 0; light yellow staining, score 1; moderate yellow staining, score 2; strong yellow staining, score 3; percentage < 5%, score 0; between 5% and 25%, score 1; between 26% and 50%, score 2; between 51% and 75%, score 3; >75%, score 4). Scores of 9–12 were considered as “strong”, scores of 5–8 were considered as “weak”, and scores of 0–4 were considered as “negative”, as mentioned by Hao et al. [ 21]. All of the stained sections were reviewed by two independent pathologists.

37.5 × 10 ⁴ (6-well) HO-8910 and SKOV-3 cells were incubated in RPMI 1640 medium without antibiotics. After 12 h, 100 nmol/L (6-well) siRNA oligonucleotides was transfected using lipofectamine™ 2000 transfection reagent (Invitrogen). siRNA were designed: DKK4-specific siRNA (sense 5′-AGGAAGCAGAGA AACCCGGC-3′). Control cells were transfected with control siRNA.

The transwell migration (Corning, USA) was performed using a chamber system with matrigel gel membrane (8.0 lm pore) from BD system (BD, USA). The 2 × 10 ⁴ cells were incubated into a 24-well plate with the upper chamber with 1% FBS medium, and the bottom was covered with the medium containing 20% FBS and 10 μg/ml of bovine fibronectin (chemoattractant) (Hyclone). The cells, migrated for 24 h, were fixed and counted in 10 high-powered (×200) fields under a microscope. The experiment was repeated three times.

To detect the ability of JNK on ovarian cancer cell invasion. 4 × 105 cells were pretreated with the JNK inhibitor SP600125 (0, 5 or 10 μM) in RPMI 1640 medium containing 5% FBS for 30 min before being added to the Matrigel-coated Transwell inserts. The cells, migrated for 12 h, were fixed and counted in 10 high-powered (×200) fields under a microscope as mentioned by Gonzalez-Villasana V et al. [ 22]. The experiment was repeated three times.

Cells were washed with PBS (pH 7.4), fixed in methanol, rinsed and permeabilized with PBS containing 0.1% Triton X-100. Fixed cells were blocked for 1 h in 3% BSA and incubated with FITC-phalloidin (Sigma) for 30 min and then washed with PBS. The DNA dye DAPI (Molecular Probes) was used as nuclear stain. Images were obtained using a Laser confocal microscopy in 5 high-powered (×1000) fields. Multiple cells were categorized in each experimental point.

We compared the expression of DKK4 mRNA and protein in human EOC tissues and benign ovarian tumor tissues by qRT-PCR and western bolt. We found that the relative fold of DKK4 mRNA was significantly increased in EOC tissues (3.63 ± 2.84) than that in benign ovarian tumor tissues (1.66 ± 1.36) ( p = 0.001; Fig. 1a). The relative level of DKK4 protein was significantly upregulated in EOC tissues (0.86 ± 0.01) than that in benign ovarian tumor tissues (0.37 ± 0.03) ( p < 0.0001; Fig. 1b and c).

The result of immunohistochemistry analysis showed that DKK4 was positively expressed in epithelial ovarian cancer samples. DKK4 was strong expressed in 148/239 ovarian cancer samples, weak expressed in 72/239 ovarian cancer samples, while negative expressed in only 19/239 ovarian cancer samples. Meanwhile the strong expression of DKK4 protein in ovarian cancer samples was positively correlated with late FIGO stage with p = 0.005 (Fig. 2a, Table 1). The strong expression of DKK4 protein were not associated with age, cell differentiation or lymphatic metastasis in patients with epithelial ovarian cancer (all p > 0.05) (Table 1).

The mean ± SD of the mean disease-free survival time for the entire group of 239 patients was 43.45 ± 1.18 (95% CI = 41.14–45.77) months. The mean disease-free survival time for patients with strong expression of DKK4 (38.32 ± 1.33 (95% CI = 35.73–40.92) months) was significantly lower as compared with that with weak or negative DKK4 expression (49.15 ± 1.67 (95% CI = 45.88–52.42) months, ( p < 0.0001, log-rank test)) (Fig. 2b). The mean disease-free survival time for patients with late FIGO stage (37.78 ± 1.43 (95% CI = 34.97–40.58) months) was significantly lower as compared with that with early FIGO stage (48.70 ± 1.59 (95% CI = 45.58–51.82) months, ( p < 0.0001, log-rank test)) (Fig. 2c). Unadjusted Cox regression revealed DKK4 level (HR = 2.10, 95% CI = 1.33–3.33, P = 0.001) and FIGO stage (HR = 2.18, 95% CI = 1.41–3.37, P < 0.0001) were independent disease-free prognostic factors for epithelial ovarian carcinoma patients. This association ((DKK4 level (HR = 2.18, 95% CI = 1.37–3.46, P = 0.001) and FIGO stage (HR = 2.21, 95% CI = 1.41–3.46, P = 0.001)) was also significant in the multivariate Cox model adjusted for age, cell differentiation, and lymph node metastasis. However, age (HR = 0.75, 95% CI = 0.50–1.13, P = 0.170), cell differentiation (HR = 0.80, 95% CI = 0.54–1.20, P = 0.29), and lymph node metastasis (HR = 1.25, 95% CI = 0.82–1.91, P = 0.30) were not significantly correlated with disease-free survival rates.

DKK4 siRNA plasmid was transfected into SKOV-3 and HO-8910 cells lines, respectively. The knockdown efficiency of DKK4 protein in both SKOV-3 and HO-8910 cells were also confirmed by western blot (all p < 0.0001) (Fig. 3a and b). We detected the effect of DKK4 siRNA on ovarian cancer cell invasion (control siRNA vs. DKK4 siRNA) and the invasion ability of normal SKOV-3 and HO-8910 cells. Our results showed that DKK4 knockdown significantly decreased the incidence of invasion in ovarian cancer cells (SKOV-3: control siRNA (177.97 ± 29.59) vs. DKK4 siRNA (49.43 ± 23.57), p < 0.0001; HO-8910: control siRNA (167.63 ± 11.91) vs. DKK4 siRNA (53.23 ± 4.41), p < 0.0001) (Fig. 3c and d).

Previous studies found that the activation of JNK pathway could promote ovarian cancers progression [ 23, 24]. We detected the expression of JNK and c-JUN protein in 10 ovarian cancer tissues and 10 benign ovarian tumors. Our results showed that the levels of JNK and c-JUN protein in cancer tissues were both strong (Fig. 4a). Our results are consistent with previous studies [ 25, 26]. Then, we detected the activity of JNK and c-JUN in DKK4 siRNA cells. Our results showed that the phosphration of c-jun in DKK4 siRNA group was significantly decreased as compared with those in control siRNA group (DKK4- siRNA SKOV-3 group vs. control siRNA group, p = 0.001; DKK4- siRNA HO-8910 group vs. control siRNA group, p < 0.0001) (Fig. 4b and d). The phosphration of JNK in DKK4- siRNA group was significantly decreased as compared with those in siRNA control group (DKK4- siRNA SKOV-3 group vs. control siRNA group, p < 0.0001; DKK4- siRNA HO-8910 group vs. control siRNA group, p < 0.0001) (Fig. 4c and d). The band intensity of p-c-JUN or p-JNK was normalized to each corresponding band of c-JUN or JNK, respectively. The results indicated that DKK4 could promote JNK activation. Meanwhile, the inhibition of JNK activity, blocked by JNK specific inhibitor (SP600125), could decrease the invasive ability of ovarian cancer cells (SP600125 10 μM SKOV-3 group (42.43 ± 3.23) vs. control group (180.63 ± 9.67), p < 0.0001; SP600125 5 μM SKOV-3 group (70.97 ± 3.40) vs. control group (180.63 ± 9.67), p < 0.0001; SP600125 10 μM HO-8910 group (42.57 ± 3.56) vs. control group (179.83 ± 11.03), p < 0.0001;SP600125 5 μM HO-8910 group (71.17 ± 5.82) vs. control group (179.83 ± 11.03), p < 0.0001) (Fig. 4f and g). These results indicated that DKK4 could promote ovarian cancer cell invasion through promoting JNK activation.

Many evidence indicated that actin filaments played an important role in promoting cell invasion [ 27, 28]. The activation of JNK pathway was known to be involved in modulating cytoskeleton like actin filaments [ 29, 30]. We examined the formation of actin by using phalloidin staining. Our results found that the majority of DKK4 silenced cells lost their actin filaments as compared with that in control siRNA groups (Fig. 5).