Human CC cell lines JEG-3 (derived from a gestational CC) and JAR (derived from a male fetus CC) were purchased from the American Type Culture Collection, and the human umbilical vein endothelial cell (HUVEC) line was kindly provided by Dr Jer-Tsong Hsieh (University of Texas Southwestern Medical Center, Dallas, TX, USA). These cells were cultured in DMEM (Gibco; Thermo Fisher Scientific, Inc.) with 10% fetal bovine serum (Gibco; Thermo Fisher Scientific, Inc.). All cells were cultured at 37˚C in a humidified atmosphere with 5% CO₂. The cells used in the present study were of the third generation after resuscitation and passage. When cells reached 50% confluence, they were treated with 100 µM forskolin (Abcam) or the vehicle (0.1% DMSO) for 48 h at 37˚C. DMSO was purchased from Sigma-Aldrich; Merck KGaA. Notch-1 signaling was blocked with 10 µM DAPT (Abcam) for 24 h.

JAR and JEG-3 cells treated with forskolin or DMSO were scratched in 6-well plates with a 200 µl pipette tip when the cell confluence reached ~90%. Cells were incubated in serum-free medium (SFM), and the wound was photographed at 0 and 24 h under a phase-contrast microscope (×10 magnification; Olympus Corporation) and the total wound area was measured.

The effects of forskolin on cell migratory and invasive abilities were assessed using 8-µm-pore Transwell inserts (EMD Millipore). For the migration assay, a 200 µl SFM suspension with 5×10⁴ JAR or JEG-3 cells pre-incubated with forskolin or DMSO for 48 h were added to the upper chambers. The lower chambers were filled with 800 µl medium containing 10% fetal bovine serum. For cell invasion assay, the upper chambers were pre-coated with 50 µl Matrigel (Sigma-Aldrich; Merck KGaA) for 4 h. A 200 µl SFM suspension with 8×10⁴ JAR or 5×10⁴ JEG-3 cells pre-incubated with forskolin or DMSO for 48 h were added to the upper chambers. For the HUVEC migration assay, 200 µl HUVECs suspended in SFM at 2×10⁵ cells/ml were seeded into the upper chamber, and 1×10⁴ CC cells pretreated with forskolin or DMSO for 48 h were added to the lower chamber. Following 24-h incubation at 37˚C, cells that had migrated to the lower chamber were fixed with 4% paraformaldehyde for 25 min and stained with crystal violet for 10 min at room temperature. The upper chamber was wiped with a cotton swab, and cell numbers were counted in five randomly selected fields per well under a phase-contrast microscope (x200 magnification; Olympus Corporation).

The assay was performed as previously described by Zhan et al (6,27). A total of 200 µl Matrigel (BD Biosciences) was added to 24-well plates and incubated at 37˚C for 1 h. CC cells (1×10⁵ cells/well) suspended in SFM were seeded into Matrigel-coated wells following treatment with forskolin or DMSO for 48 h. After 6 and 24 h, tubule structure formation was observed, and the number and completeness of the tubule were assessed under a phase-contrast microscope (x200 magnification; Olympus Corporation).

Total RNA was harvested using the TRIzol^® reagent (Thermo Fisher Scientific, Inc.) according to the manufacturer's protocol and quantified by spectrophotometry. The samples were reverse-transcribed to cDNA using the PrimeScript™ RT Reagent kit (Takara Bio, Inc.). The RNA sample was incubated with 2 µl 5X PrimeScript RT Master Mix at 37˚C for 15 min, followed by 85˚C for 5 sec and terminated at 4˚C. QPCR was performed using the SYBR^® Green PCR Master Mix (Takara Bio, Inc.) on a CFX96 real-time PCR system (Bio-Rad Laboratories, Inc.). The PCR protocol was 94˚C for 10 min, followed by 40 cycles of 94˚C for 10 sec and 60˚C for 30 sec. The primer sequences used in the present study are listed in Table I. Data were normalized to GAPDH expression and all reactions were performed in triplicate. Relative gene expression was calculated using the 2^-ΔΔCq method (28).

Cells at 80-90% confluence were washed three times with cold PBS and lysed in RIPA buffer (50 mM Tris pH 8.0, 150 mM NaCl, 0.1% SDS, 1% NP-40, and 0.5% sodium deoxycholate) containing protease inhibitors (1 mM PMSF; Sigma-Aldrich; Merck KGaA). Protein samples were quantified by bicinchoninic acid assay, and 20 µg of protein was loaded per lane, separated by SDS-PAGE (8-12% gel) and electrophoretically transferred to nitrocellulose membranes (EMD Millipore). The membranes were blocked with 5% skimmed milk at room temperature for 1 h and incubated with primary antibodies against GAPDH (cat. no. KG-5G4; Kangchen; 1:1,000), syncytin-1 (cat. no. SC130888; Santa Cruz; 1:200), VEGFA (cat. no. ab185265; Abcam; 1:1,000), VE-cadherin (VE-cad; cat. no. ab205336; Abcam; 1:1,000), MMP-2 (cat. no. ab37150; Abcam; 1:1,000), Hes family BHLH transcription factor 1 (HES1; cat. no. ab108937; Abcam; 1:500), cleaved Notch1 (NICD; cat. no. 4147; Cell Signaling Technology, Inc.; 1:1,000), N-cadherin (N-cad; cat. no. 13116; Cell Signaling Technology, Inc.; 1:500), zinc finger E-box-binding homeobox 1 (ZEB1; cat. no. 3396; Cell Signaling Technology, Inc.; 1:1,000), E-cadherin (E-cad; cat. no. sc71009; Santa Cruz Biotechnology, Inc.; 1:1,000), cyto-keratin (CK19; cat. no. sc376126; Santa Cruz Biotechnology, Inc.; 1:1,000), vimentin (cat. no. sc80975; Santa Cruz Biotechnology, Inc.; 1:400) and MMP-9 (cat. no. sc10737; Santa Cruz Biotechnology, Inc.; 1:1,000) at 4˚C overnight, followed by washing with TBS containing 0.1% Tween-20 and 1 h incubation with a horseradish peroxidase (HRP)-conjugated goat anti-rabbit IgG (H+L) secondary antibody (cat. no. sc2301; Santa Cruz Biotechnology, Inc.; 1:800) at room temperature. Protein bands were visualized using the Odyssey detection system (LI-COR Biosciences) and quantified using ImageLab software (version 4.1; Bio-Rad Laboratories, Inc.).

Animal care and protocols were performed according to the guidelines of the Institutional Animal Care and Use Committee of Xi'an Jiaotong University. A preliminary experiment indicated that xenografts in nude mice were easier to establish using JEG-3 cells compared with JAR cells (29); therefore, JEG-3 cells were used for the in vivo study. A total of 12 5-week-old athymic nude mice (Charles River Laboratories, Inc.) were randomly divided into two groups; the mice in the forskolin group were injected intra-peritoneally with 5 mg/kg forskolin diluted in a PBS/DMSO solution (15:0.1) five times on days −4, −2, 2, 4 and 6 (relative to JEG-3 cell injection), and the control groups were injected with equal volumes of DMSO (30). A total of 5×10⁶ JEG-3 cells were injected subcutaneously into nude mice to establish the xenograft model on day 0. Xenograft tumors were harvested after 3 weeks and used for western blotting, hematoxylin and eosin (H&E) staining and CD31/periodic acid-Schiff (PAS) double staining (Fig. 1).

Immunohistochemistry was performed on 4-µm-thick sections. Sections were deparaf-finized using xylene and dehydrated using ethanol (100, 95, 80, 70 and 50% for 3 min each). Antigen retrieval was performed using citric acid for 5 min at 121˚C. After blocking with 3% hydrogen peroxide for 30 min at room temperature, the sections were incubated with an anti-CD31 antibody (cat. no. ab28364; Abcam; 1:100) overnight at 4˚C and anti-rabbit ChemMate™ EnVision™/HRP secondary antibody for 1 h at room temperature, followed by staining with DAB for 40 sec (cat. no. GK500705; Dako; Agilent Technologies, Inc.). The sections were incubated with PAS (0.5% periodic acid for 10 min and Schiff solution for 15 min) at room temperature prior to H&E counterstaining (hematoxylin for 5 min and 0.5% eosin for 1-3 min). VM structures were defined as CD31-negative PAS-positive structures.

Data are presented as the mean ± standard error of the mean of at least three independent experiments. GraphPad Prism software (version 6.0; GraphPad Software, Inc.) was used to analyze differences between two groups with Student's t-test, and among multiple groups with one-way ANOVA followed by Tukey's post hoc test. P<0.05 was considered to indicate a statistically significant difference.

Cell motility is associated with tumor metastasis. In order to assess whether forskolin could stimulate the invasion and migration of JAR and JEG-3 cells, the present study used wound healing and transwell assays. In the wound healing assay, the migration rates of the forskolin groups at 24 h were significantly higher compared with those of the DMSO groups (P<0.01; Fig. 2A). In the migration assay presented in Fig. 2B, the numbers of migrant cells were ~2-fold higher compared with the control groups when JAR and JEG-3 cells were treated with forskolin (P<0.01). Similarly, the invasive capacity was markedly enhanced when cells treated with forskolin compared with the respective control groups (P<0.01; Fig. 2C). Consistently, RT-qPCR and western blot assays demonstrated significantly increased MMP2 and MMP9 expression following forskolin treatment in the two CC cell lines compared with the corresponding controls (Fig. 2D and E).

VM has been demonstrated to be associated with cell migration, invasion and capillary tube formation (6). In order to investigate the role of forskolin in VM, the present study used a well-established in vitro model of 3D culture. Compared with the control group, JAR and JEG-3 cells treated with forskolin exhibited an enhanced capability of forming typical capillary-like structures on 3D Matrigel medium (P<0.01; Fig. 3A and B), which appeared more typical in JEG-3. As VE-cad and EphA2 act in a coordinated manner as key regulatory elements during the process of VM (31), these two VM-associated markers were detected via western blotting or RT-qPCR; both VE-cad protein levels and EphA2 mRNA levels were upregulated in the forskolin-treated group compared with the control group (P<0.01; Fig. 3C; P<0.05; Fig. 3D), suggesting that forskolin was involved in the formation of vasculogenic-like networks in CC cells. In addition, to reveal the potential roles of forskolin in angiogenesis, an endothelial recruitment assay was performed, which revealed a decreased ability of forskolin-treated CC cells to recruit HUVECs compared with that of DMSO-treated cells (P<0.01; Fig. 3E). In line with these results, the expression of vascular endothelial growth factor α (VEGFA), which serves a key role in tumor angiogenesis (32), was also effectively inhibited by forskolin treatment (Fig. 3F).

In order to investigate the biological variation of trophoblasts during syncytiolization induced by forskolin in CC cells, the morphological changes and EMT markers were observed and detected in the present study. The results demonstrated that the cells were converted to a spindle-like shape and lost cellular cohesiveness following administration of forskolin for 48 h (Fig. 4A and C), and the expression levels of epithelial markers E-cad and CK19 were significantly downregulated, whereas those of mesenchymal markers, including N-cad, Vimentin and ZEB1, were significantly upregulated (Fig. 4B and D).

A previous study has demonstrated that Notch-1 signaling is involved in gestational trophoblastic neoplasia (26). Based on the result that the expression levels of NICD and HES1 were upregulated after forskolin treatment, 10 µM DAPT was added into the culture medium to evaluate whether the Notch-1 signaling pathway was involved in regulating metastatic and VM capabilities in CC cells (Fig. 5A and B). In addition, DAPT could reverse the upregulating of NICD and HES. Consistently, inhibition of the Notch-1 signaling pathway also significantly attenuated the invasive ability of cells in the forskolin and DMSO groups compared with cells without DAPT treatment (Fig. 6A). Upregulation of MMP2, MMP9 and VE-cad expression levels by forskolin was also reversed by DAPT (P<0.01; Fig. 6B and C). Changes in EMT-associated proteins were also evaluated following DAPT interference. Inhibition of the Notch-1 signaling pathway partially reversed the downregulation of E-cad and significantly reversed the upregulation of N-cad in forskolin-treated cells (Fig. 6D and E), which indicated that the Notch-1 signaling pathway may mediate EMT during the syncytiolization process induced by forskolin.

In order to validate the function of forskolin in vivo, JEG-3 cells were subcutaneously implanted into the flanks of nude mice. After 3 weeks, the subcutaneous tumor diameter in the forskolin group was larger compared with that in the DMSO group (Fig. 7A). Tumor sections were stained with CD31 to observe tumor angiogenesis by counting the number of vessels; VM structures formed by tumor cells exhibited a PAS-positive basement membrane (25). Endothelium-dependent vessels and VM channels were observed in the excised tumors; PAS-positive tubular structures lined with CD31-negative cells were indicated, and intratumoral blood vessels were detected in all specimens (Fig. 7B). Consistent with the observation in vitro, forskolin treatment resulted in high expression of MMP2, MMP-9 and VE-cad, as well as the activation of the Notch-1 signaling pathway (Fig. 7C).