Role of Jagged1/STAT3 signalling in platinum‐resistant ovarian cancer

2.1. Tissue samples

This study was approved by the ethics committee of Renmin Hospital of Wuhan University (Wuhan, China), and each participant provided written informed consent. A total of 42 patients were enrolled from the Department of Obstetrics and Gynaecology, Renmin Hospital of Wuhan University between 2016 and 2017. All of the eligible patients fulfilled the following criteria: (a) Platinum chemotherapy‐sensitive cases: complete remission (CR) was achieved after adjuvant chemotherapy, and the interval between the last chemotherapy treatment and recurrence was greater than or equal to 6 months. (b) Platinum‐resistant cases: complete remission was achieved after adjuvant chemotherapy, and the recurrence time was <6 months after the last chemotherapy treatment. (c) Refractory cases of platinum‐based chemotherapy: the best response to platinum‐based chemotherapy during disease progression (PD) or after at least six courses of platinum‐based chemotherapy was a partial remission (PR) or disease stability (SD). Groups (b) and (c) are collectively referred to as platinum‐based chemotherapy‐resistant cases in this study. The study group consisted of 33 platinum chemotherapy‐sensitive cases and nine platinum‐based chemotherapy‐resistant cases. All sample tissue blocks were fixed for 12‐24 hours at room temperature in a 4% neutrally buffered paraformaldehyde solution, dehydrated, embedded in paraffin, and then sliced into 4‐μm sections for immunohistochemical staining.

2.2. Cell lines and culture

A cisplatin‐sensitive human epithelial ovarian cancer cell line (OV2008) and its resistant variant (C13K) were gifts from professor Ma Ding (Cancer Biology Research Center, Huazhong University of Science and Technology, China). Cells were maintained in RPMI‐1640 (Jenom, Hangzhou, China) complete medium with 100 U/mL penicillin/streptomycin (Beyotime Institute of Biotechnology, Haimen, China) and 10% foetal bovine serum (FBS; Gibco‐BRL, Invitrogen Life Technologies) at 37°C in a humidified atmosphere containing 5% CO₂.

2.3. RNA interference

Small interfering (siRNA) sequences directed against Jagged1 and Jagged2 were designed and synthesized by the Guangzhou RiboBio Company (Guangzhou, China). The sequences for Jagged1 and Jagged2 were: 5′‐GAAGAATGTTTCCGCTGAA‐3′ and 5′‐GCAAAGAAGCCGUGUGUAA‐3′, respectively. The empty vector was used as the control. For transfection, the cells were resuspended at a density of 5 × 10⁴ cells/mL and seeded in six‐well plates. When the cells were 60%‐80% confluent, they were transfected using HiPerfect (QIAGEN, Duesseldorf, Germany) according to the manufacturer's instructions. The transfected cells were named C13K/si‐Jagged1, C13K/si‐Jagged2 or C13K/si‐NC depending on the treatment. To explore the effects of Jagged1 knockdown in ovarian epithelial cancer (OEC) tumourigenesis in vivo, small interfering (siRNA) sequences were designed as shRNA and were packaged with lentiviral vectors. Cells were grown to 30%‐40% confluence and incubated with the constructed lentiviral vectors for 8 hours in growth medium containing10 μg/mL of polybrene (Santa Cruz). Three days after infection, the medium was changed to fresh RPMI 1640 containing 3 μg/mL puromycin (Santa Cruz). Puromycin‐resistant colonies were used for subsequent studies in vivo.

2.4. Cell viability and proliferation assay

Cell viability and IC50 values (drug concentration causing 50% inhibition of cell growth) were measured by the CCK‐8 assay (Beyotime Institute of Biotechnology, Haimen, China). Briefly, cells were suspended in RPMI1640 medium at a density of 5 × 10⁴ cells/mL and seeded into 96‐well plates and incubated for 24 hours at 37°C. The CCK‐8 assay was performed before transfection and then at 24, 48, 72 and 96 hours following transfection for the control groups. For the treatment groups, following removal of the spent culture medium, 100 μL of prepared medium containing various concentrations of DAPT (an inhibitor of the γ‐secretase complex, and it can indirectly inhibit the Notch pathway) (0, 2.5, 5, 10, 20 or 40 μmol/L), an inhibitor of the γ‐secretase complex, and it can indirectly inhibit the Notch pathway, was added. Each treatment was conducted in triplicate. Following incubation at 37°C for 24 hours, 10 μL of CCK‐8 solution was added to each well and the cells were incubated for 1‐2 hours. The absorbance of the wells was then measured at 450 nm using a microplate spectrophotometer (Victor3 1420 Multilable Counter; Perkin Elmer, Inc., Waltham, MA, USA). The viability of the treated group was expressed as a percentage of the untreated control group, which was designated as 100%. Cell growth curves were plotted as follows: Cellular growth (%) = OD450 of experimental well)/OD450 of control well.

2.5. Three‐dimensional (3D) spheroid formation

A total of 5 × 10³ cells was seeded on Matrigel (50 μL/cm², cat. 356234; Corning Incorporated, NY, USA) coated 24‐well plates, and the media were refreshed every 2‐3 days. The cell forming 3D spherical structure was photographed (×200 magnifications) at 2‐day intervals for 2 weeks.

2.6. Wound‐healing assay

Cells were seeded and grown in six‐well plates to complete confluence. A wound injury was made with the yellow tip of a sterile micropipette, and the detached cells were removed by washing with phosphate buffer solution (PBS). Cells were then incubated with the prepared medium containing various concentrations of DAPT (0, 2.5, 5, 10, 20 or 40 μmol/L) and allowed to migrate for up to 24 hours. Images were taken at 4 hours intervals by inversion microscope (BX51; Olympus Corporation, Tokyo, Japan). Relative wound density was calculated by ImageJ Software 1.51 (https://imagej.nih.gov/ij/) as the metric to quantify cell migration.27

2.7. Cell migration and invasion assays

Migration and invasion assays were performed by using 24‐well Transwell plates (Corning Incorporated) as instructed by the manufacturer. Briefly, cells (5 × 10⁴ cells/well for the migration assay and 1 × 10⁵ cells/well for the invasion assay) were plated in serum‐free medium in the upper chambers and in RPMI 1640 containing various concentrations of DAPT (0, 2.5, 5, 10, 20 or 40 μmol/L) with 10% FBS placed in the bottom wells and incubated at 37°C for 24 hours. The migrated or invaded cells were fixed with 4% paraformaldehyde followed by staining with 1% crystal violet. Three random microscopic fields (×100) of the chamber were photographed and counted.

2.8. Bimolecular fluorescence complementation assay

The slides of C13K cells were fixed with 4% formaldehyde for 15 minutes and after washing with PBS, the cells were incubated with 0.1% Triton X‐100 for 5 minutes at room temperature. The slides were blocked with 2% goat serum, followed by incubation overnight with the primary antibodies (a mixture of mouse polyclonal anti‐STAT3 antibody and rabbit polyclonal anti‐Jagged1 antibody) at 4°C. The next day, the cells were incubated with the secondary antibody mixture containing fluorescent‐labelled goat antimouse polyclonal IgG secondary antibody (CY3, 1:100, BA1032, BOSTER Biological Technology Co. Ltd., Wuhan, China) and goat anti‐rabbit polyclonal IgG secondary antibody (FITC, 1:50; BA1105, BOSTER Biological Technology Co. Ltd.) at room temperature for 60 minutes, and the nuclei were stained using Diamidine phenyl indole (DAPI) at room temperature for 10 minutes. The cells were observed under a fluorescence microscope (Olympus Corporation).

2.9. Western blot analysis

The cell lysates were electrophoresed on sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS‐PAGE) and transferred onto polyvinylidene fluoride (PVDF) membranes (EMD Millipore, Billerica, MA, USA). The membranes were blocked with 5% non‐fat milk solution before primary antibody incubation overnight at 4°C. After washing, the membrane was incubated with horseradish peroxidase‐conjugated secondary antibodies (1:10 000 dilution; LICOR Biosciences, Lincoln, NE, USA) for 1 hour. Finally, the staining intensity was visualized and quantified by an Odyssey imaging system (LICOR Biosciences, Lincoln, NE, USA). Experiments were performed in triplicate. The following primary antibodies were used: mouse polyclonal anti‐STAT3 antibody (1:1000; cat. no. 9139), mouse polyclonal anti‐phospho‐STAT3 (Tyr705;1:2000; cat. no. 9145), rabbit polyclonal anti‐phospho‐STAT3 (Ser727; 1:1000; cat. no. 94994), mouse polyclonal anti‐E‐cadherin antibody (1:1000; cat. no.14472), mouse polyclonal anti‐N‐cadherin antibody (1:1000; cat. no.13116), rabbit monoclonal anti‐vimentin antibody (1:1000; cat. no. 5741) and rabbit monoclonal anti‐cleaved Notch1 antibody (1:1000; cat. no. 4147), all obtained from Cell Signaling Technology, Inc. (Danvers, MA, USA); rabbit polyclonal anti‐Jagged1 antibody (1:500; cat. no. ab7771), rabbit polyclonal anti‐Jagged2 antibody (1:2000; cat. no. ab109627), rabbit monoclonal anti‐Notch1 antibody (1:500; cat. no. ab8925), rabbit monoclonal anti‐Notch2 antibody (1:500; cat. no. ab8926) and rabbit monoclonal anti‐Twist1 antibody (1:1000; cat. no. ab50581) were obtained from Abcam, Inc. (Cambridge, USA). Rabbit monoclonal anti‐cleaved Notch2 antibody (1:100; cat. no. 40517) was obtained from Signalway Antibody LLC (Maryland, USA). An anti‐β‐actin antibody (1:1000; cat. no. 4970; Cell Signaling Technology, Inc.) served as an endogenous reference.

2.10. Reverse transcription‐polymerase chain reaction (RT‐PCR)

Total RNA from DMSO‐treated control and PEITC‐treated cells was isolated using RNeasy kit (Qiagen). First‐strand cDNA was synthesized using Superscript reverse transcriptase (Invitrogen‐Life Technologies) with oligo (dT)20 primer; semi‐quantitative and real‐time PCR were performed (Primers in Table S1).

2.11. Immunoprecipitation

Cells were solubilized in lysis buffer. An equal amount of each protein lysate was incubated with anti‐STAT3 or anti‐Jagged1 antibody overnight at 4°C, followed by incubation with 10 μL of protein A‐sepharose beads for 2 hours. The immune complexes were analysed by Western blot analysis with anti‐Jagged1 or anti‐STAT3 antibody.

2.12. Immunohistochemistry

An ElivisionTM super HRP IHC kit (cat. no. Kit‐9921; Fuzhou Maxim Biotechnology Development Co., Ltd.) was used for immunohistochemistry. The paraffin‐embedded sections were deparaffinized, rehydrated and antigen retrieval was performed by heat mediation in 0.01 M sodium citrate, pH 6.0. The sections were incubated overnight with mouse polyclonal anti‐STAT3 antibody (1:500; Cell Signaling Technology, Inc.), rabbit polyclonal anti‐Jagged1 antibody (1:200; Abcam, Inc.), rabbit monoclonal anti‐Notch1 antibody (1:150; Cell Signaling Technology, Inc.), or rabbit monoclonal anti‐Notch2 antibody (1:100; Cell Signaling Technology, Inc.) at 4°C following the instructions on the product datasheets for the primary antibodies. Following incubation with streptavidin peroxidase (Maxim Biotechnology Development Co., Ltd.) for 10 minutes at room temperature, secondary antibodies were added for 30 minutes at 37°C. The immune reaction was then visualized using 3,3'‐diaminobenzidine (DAB) (Fuzhou Maxim Biotechnology Development Co., Ltd., Fuzhou, China). The expression intensity was evaluated by a semi‐quantitative system to calculate the percentage of positive neoplastic cells: 0 points, no positive cells; 1 point, 1% to 25%; 2 points, 26% to 50%; 3 points, 50% to 75%; 4 points, >75%. Then, ≤2 points were judged as negative for expression, 2 points, were judged as positive expression.

2.13. Xenograft tumourigenesis in nude mice

Athymic nude mice (BALB/c nu/nu) (female, 5‐week old) were purchased from Beijing Vital River Laboratory Animal Technology Cooperation (Beijing, China) and were acclimated for 7 days in the laboratory before experimentation. To establish the capability of invasiveness and migration, 1 × 107 Jagged1 knockdown, negative control and C13K cells were injected into the tail vein of the mice (n = 4/group). Body weight was measured once a week. On the day of harvest the lung, liver and spleen were analysed by the haematoxylin and eosin staining method. All animal studies were approved by the ethics committee of the Renmin Hospital of Wuhan University.

2.14. Statistical analysis

Statistical analysis was performed with the statistical Package for Social Science (SPSS Release 22.0; SPSS Inc., Chicago, IL, USA). The differences observed between the control and treated groups for cell proliferation, viability, cell migration and invasion ability, the mRNA and proteins expression were analysed using either one‐way ANOVA or unpaired Student's t tests (two‐tailed). Chi‐squared test was used to analyse the protein expression intensity between platinum‐resistant group and platinum responsive group. The results are expressed as the mean ± standard deviation from triplicate experiments and a value of P < 0.05 was considered to be statistically significant.

3.1. Notch pathway in platinum‐resistant ovarian cancer is important for the malignant phenotype

In this study, we first examined the cytotoxic effect of cisplatin on OV2008 and C13K cells by using a CCK‐8 assay. The IC50 value was used to represent the level of cytotoxicity. The IC50 values for the OV2008 and C13K cells were 23.11 ± 0.97 μmol/L and 39.43 ± 1.19 μmol/L, respectively (Figure 1A), which suggested that the C13K cells were more resistant to cisplatin‐induced cytotoxicity compared with the OV2008 cells. To investigate whether the Notch pathway was involved in the cisplatin resistance of ovarian cancer, we first examined the protein expression levels of this pathway. Western blot analysis determined that the expression levels of Notch1/2 and cleaved Notch1/2 in C13K cells were significantly higher than in OV2008 cells. Moreover, the Jagged1 protein level and mRNA level were also highly expressed in C13K cells (Figure 1B and Figure S1). To confirm whether these findings were consistent with that in actual human tumours, the relative genes' protein expression levels were examined by IHC of the tissues in the platinum‐resistant group and platinum responsive group. The results showed that Notch1 and Notch2 were expressed in all tumour samples from the platinum‐resistant group and most of the tumour samples from the platinum responsive group, and in addition, Notch1 and Notch2 positive staining intensities were higher in the platinum‐resistant group than in the platinum responsive group (Figure 1C and Table S2).

To illuminate the role of the Notch pathway in C13K cells, DAPT was applied and its effects on cell proliferation and migratory ability were examined. First, we examined the effect of DAPT on cell migration and invasion abilities. Wound healing assays showed that the wound density in C13K cells was significantly higher after DAPT exposure (Figure 1D and Figure S2). Moreover, the Transwell migration assay confirmed that DAPT treatment greatly suppressed the migratory ability of the C13K cells (Figure 1E and Figure S3) and their invasive ability (Figure 1F and Figure S4). Therefore, these findings implicated the Notch pathway as playing an important role in the migration and invasive abilities of cisplatin‐resistant ovarian cancer cells. We also checked whether DAPT could inhibit the proliferation of the C13K cells. However, we found that the cell proliferation rates up to 72 hours after DAPT treatment showed no significant change (Figure 1G). In case a continuous exposure to DAPT was not sufficient, we also maintained continuous exposure of C13K cells to different concentrations of DAPT for one to four passages and we observed that the cell proliferation was gradually decreased following passage 2 in a dose‐dependent manner (Figure 1H). These results suggest that the Notch pathway in cisplatin‐resistant ovarian cancer is significant in increasing the cell malignant phenotype.

3.2. Notch pathway is involved in EMT progression in cisplatin‐resistant ovarian cancer cells

First, we found that the morphology of the C13K cells showed changes to a shuttle and stem‐like shape (Figure 2A and Figure S5), which are consistent with morphological EMT features. Western blot analyses revealed that the expression of the epithelial adhesion protein E‐cadherin was lower, while the mesenchymal marker proteins N‐cadherin and vimentin as well as the EMT key modulator Twist1 were up‐regulated in C13K cells (Figure 2B). Furthermore, IHC assay showed that the positive staining intensity of the EMT related mesenchymal proteins (N‐cadherin, vimentin and Twist1) were higher in the platinum‐resistant group than in the platinum responsive group (Figure 2C and Tables S3 and S4). These results suggested that EMT is a critical phenotype change in platinum‐resistant ovarian cancer.

To further explore the role of Notch related to EMT signalling in C13K cells, exposure to a wide concentration range of DAPT was used. The protein expression of Twist1 was inhibited when the C13K cells were treated with 10 μmol/L DAPT for 48 hours, but the E‐cadherin, N‐cadherin and vimentin expression levels were not obviously changed (Figure 2D). The reason for this result was either not a long enough exposure time or the Notch inhibitor had no effects on EMT. Dramatically, when we examined the protein levels after continuous exposure of C13K cells to 10 μmol/L DAPT for one to four passages, we found that E‐cadherin expression recovered and the N‐cadherin and vimentin protein levels were down‐regulated in a dose and time‐dependent manner (Figure 2D). Furthermore, we employed 3D culture to examine the characteristics of cancer stem cells (CSCs), which was an important functional change endowed by EMT. When C13K cells were cultured for 2 weeks, we found that they displayed an aggressive phenotype, showing highly disorganized cell clusters lacking basal polarity, while OV2008 cells and C13K cells treated with DAPT showed a lower aggressive ability and more organized spheroid structures (Figure 2E and Figure S6). These results suggested that the Notch pathway could regulate the occurrence of EMT in cisplatin‐resistant ovarian cancer cells.

3.3. Jagged1 plays a critical role in EMT signalling in C13K cells

Jagged1 is one the most important ligands of the Notch pathway, and studies have demonstrated it has a role in regulating EMT of cancer.28 Jagged2, a homologue of Jagged1, also plays an important role in EMT of lung adenocarcinoma.29 To investigate whether Jagged1 and/or Jagged2 are critical for EMT progression in C13K cells, the cells were transfected with control, Jagged1 and Jagged2 siRNA. Figure 3A and B show strong suppression of Notch1 and cleaved Notch1 after transfection with each target siRNA. Western blot assays showed that the expression of N‐cadherin and vimentin, mesenchymal markers of EMT, were reduced by Jagged1 and Jagged2 knockdown in C13K cells, and the expression of these proteins was lower in C13K/si‐Jagged2 cells compared with C13K/si‐Jagged1 cells. A similar result was found for Twist1, a key transcription regulator of EMT. Although the protein expression of E‐cadherin was not completely restored, a tendency towards reversion was observed following Jagged1 knockdown (Figure 3). These results suggest that Jagged1 and Jagged2 are involved in the Notch pathway, and Jagged1 plays a more important role in EMT signalling in C13K cells.

3.4. Crosstalk of Jagged1/STAT3, not Jagged2/STAT3, is important for EMT in C13K cells

What could be causing the stronger ability of Jagged1 in regulating EMT in C13K cells? Recently, a study reported that Notch4/STAT3 crosstalk is important for EMT in breast cancer,23 and IHC assays showed that Jagged1 and STAT3 protein were both expressed in the tissues of the cisplatin‐resistant group and the cisplatin responsive group. The STAT3 protein was mainly localized to the cell nucleus with some molecules localized to the cytoplasm and cytomembrane. Jagged1 protein was mainly localized to the cytomembrane and only a few molecules were localized to the cytoplasm and cell nucleus, and the staining intensity of both proteins were stronger in the cisplatin‐resistant group (Figure 4A and Table S5). Therefore, we examined the protein level of STAT3 in C13K cells transfected with Jagged1 and Jagged2 siRNA. Western blot assay showed that the levels of total STAT3 and tyrosine 705‐phosphorylated STAT3 (pY705) were not obviously altered in C13K/si‐Jagged1 and C13K/si‐Jagged2 cells, while the serine 727‐phosphorylated STAT3 (pS727) protein level was significantly reduced by knocking down Jagged1 in C13K cells but not by knocking down Jagged2 (Figure 4B and Figure S7). In addition, we observed similar results with C13K/si‐Jagged1 cells and C13K cells treated with DAPT, with a reduction in pS727 protein in a dose‐dependent manner (Figure 4C).

To further explore the physical association between Jagged1 and STAT3, total cell lysates from C13K cells were immunoprecipitated using an anti‐STAT3 antibody. The co‐IP assay showed that STAT3 physically interacts with Jagged1 but not Jagged2 (Figure 4D). We further performed a bimolecular fluorescence complementation assay in C13K cells and we detected that both STAT3 and Jagged1 are localized to the cytoplasm and cytomembrane, and the interaction of these proteins was confirmed by co‐immunofluorescent staining and confocal microscopy imaging (Figure 4E). These results suggest that Jagged1 could regulate the protein expression of STAT3 and Jagged1/STAT3 crosstalk may play an important role for EMT in C13K cells. Interestingly, Yang et al30 have reported that acquisition of trastuzumab resistance is associated with the formation of the EMT/CSC phenotype and transition of survival signalling through activating an IL‐6/STAT3/Jagged‐1/Notch positive feedback signalling loop in gastric cancer cells crosstalk. Therefore, we exposed C13K cells to a wide concentration range of WP1066, an inhibitor of the JAK/STAT3 pathway. Western blot assay showed that the Jagged1 level was not altered obviously and the levels of N‐cadherin and vimentin, mesenchymal markers of EMT, were inhibited by WP1066 in a dose‐dependent manner (Figure 4F). Based on these results, we could conclude that Jagged1/STAT3 crosstalk is a critical mechanism for EMT in cisplatin‐resistant ovarian cancer.

3.5. Jagged1 knockdown impairs tumour growth and invasion in mouse xenograft models

To explore the effects of Jagged1 knockdown on OEC tumourigenesis in vivo, 1 × 10⁷ OEC cells were injected into the tail veins of each nude mouse. Mice were divided into three groups: mice bearing control C13K cells (group 1), mice bearing C13K/si‐NC cells (group 2) and mice bearing C13K/si‐Jagged1 cells (group 3), and the weight of each mouse was measured once a week. As shown in Figure 5A, there were no obvious differences in body weight amongst these three groups before 4 weeks after injection. However, the first group and the second group began losing weight after the fourth week and developed cachexia at 7 to 8 weeks, while the third group did not show weight loss until the seventh week. This suggests that knockdown of Jagged1 dramatically impaired the tumourigenic growth of C13K cells. At harvest (Figure 5B‐D), mice implanted with C13K/sh‐Jagged1 cells showed no tumour formation in the spleen and only 50 percent (2/4) had liver metastases. In contrast, mice implanted with C13K and C13K/si‐NC cells exhibited aggressive tumour formation in the spleen (3/4) and increased susceptibility to macroscopic metastases in the liver (4/4), and even worse, large areas of necrosis occurred in the spleen and liver, and a large number of metastatic carcinomas were found in the lung in group 1 and 2. Although there were some macroscopic metastases in the lung in group 3, the size of the metastatic carcinoma was one‐fifth that of group 2 and 3. These results confirmed the inhibitory effect of Jagged1 knockdown on cisplatin‐resistant ovarian cancer growth.