MicroRNA-34a is a tumor suppressor in choriocarcinoma via regulation of Delta-like1

The BeWo cells and JEG-3 cells (American Type Culture Collection, Manassas, VA) were cultured respectively in F12K medium or DMEM medium, (Invitrogen, Carlsad, CA) supplemented with 10% fetal bovine serum (FBS), 50 U/ml of penicillin and 50 μg/ml of streptomycin (Invitrogen). For force-expression of miR-34a, 1 × 10⁵ cells were seeded in 12-well culture plates 1 day before transfection either with 50 nM of precursor of miR-34a (pre-miR-34a) or pre-miR-Scramble (Negative Control #1, Ambion, Austin, TX) by Lipofectamine 2000 (Invitrogen). For activation of Notch signaling, a Notch NCID expression plasmid (pCDNA6-Notch NCID, a kind gift from Prof. Jon Aster, Brigham and Women’s Hospital and Harvard Medical School, Boston, Massachusetts, USA) was used. In control experiments, the cells were transfected with an empty vector (pCDNA6).

Cell proliferation was estimated by the CyQuant® cell proliferation assay (Invitrogen) according to the manufacturer’s protocol. Fluorescence signal with excitation at 485 nm and emission at 530 nm was measured by a microplate reader (Tecan Group Ltd, Männedorf, Switzerland).

We used the BD Matrigel Invasion Chamber (8-μm pore size; BD Biosciences, Franklin Lakes, NJ) to quantify cell invasion. The transfected cells in FBS-free culture medium were seeded onto the upper chamber while the lower chamber was filled with normal FBS-containing medium. For DLL1 stimulation, 2.5 μg of recombinant DLL1 (R&D systems, Minneapolis, MN) was added to the upper chamber. In the control experiment, the same volume of DMSO was added to the cells. For antibody inhibition, 5 μg of polyclonal anti-DLL1 antibody (Santa Cruz Biotechnology, Santa Cruz, CA) was added to the upper chamber during seeding, and fresh antibody was added every 24 hours. After 48 hours, the cells remained in the upper chamber were removed by cotton swabs, whilst those that had invaded through the matrix between the two chambers were visualized by staining with 0.1% of crystal violet (Sigma-Aldrich, St Louis, MO). To quantify the invasion result, the dye was dissolved in 10% acetic acid and the absorbance was measured by a microplate reader. Parallel experiments on cell proliferation were performed to estimate the effect of cell proliferation on the results of cell invasion.

Total RNA was prepared by using the mirVana^TM miRNA Isolation Kit (Ambion) according to the manufacturer’s protocol. For assaying mRNA, first-strand cDNA was synthesized by the High Capacity cDNA Reverse Transcription kit (Applied Biosystems, Foster City, CA) and the target gene expression was quantified by the TaqMan® Gene Expression Assays (Applied Biosystems) using an Applied Biosystems 7500 Detection system (Applied Biosystems). The expression of mRNA was determined from the threshold cycle (Ct), and the relative expression levels were calculated by the 2^-ΔΔCt method [14]. The relative expression levels were normalized with the expression of 18S mRNA. For measuring miRNAs, the first-strand cDNA was synthesized by the TaqMan® MicroRNA Reverse Transcription kit (Applied Biosystems) and the miRNA expression was quantified by the TaqMan® MicroRNA assay (Applied Biosystems). The relative expression of miR-34a was calculated as the above and the levels were normalized with the expression of the small RNA RNU6B.

Oligonucleotides were synthesized according to the nucleotide sequence of potential miR-34a binding regions identified by TargetScan5.2 on DLL1 (355–361 of DLL1 3’UTR, NCBI reference sequence: NM_005618.3). Specific primers were purchased from Invitrogen (Forward: 5’-TCCTCGAGAA TTAGAAACAC AAACACTGCC TGCGGCCGCT G-3’ and Reverse: 5’-CAGCGGCCGC AGGCAGTGTT TGTGTTTCTA ATTCTCGAGG A-3’). The DNA fragment was cloned into the Xho I and Not I sites of the pSiCheck™-2vector (Promega, Madison, WI). The vector was transfected with either pre-miR-34a or pre-Scramble into the cells (Ambion). At 48-hour post-transfection, the cells were lysed and the luciferase activities in the lysate were measured by the Dual Luciferase Reporter Assay System (Promega). The effect of the miRNA was measured by the activity of the Renilla luciferase normalized to that of the firefly luciferase. To test the specificity of the interaction between miR-34a and 3’UTR of DLL1, the miR-34a seed binding region on the 3’UTR of DLL1 was mutated. The mutant construct was generated with specific primers (Forward: 5’-TCCTCGAGAA TTAGAAACAC AAAGAGTACT TGCGGCCGCT G-3’ and Reverse: 5’-CAGCGGCCGC AAGTACTCTT TGTGTTTCTA ATTCTCGAGG A-3’; underlined regions denote the mutated sequences) and cloned into the pSiCheck™-2vector as described above.

BeWo and JEG-3 cells transfected with pre-miR-34a or pre-Scramble were seeded at a density of 20 cells/cm² in normal culture medium as stated as the above and allowed to grow for 2 weeks. The colonies were then stained with 0.1% crystal violet (Sigma-Aldrich), washed with PBS and their number was counted. Images of the colonies were scanned with a gel documentation system (AlphaImager® HP, Alpha Innotech Corporation, San Leandro, CA).

The study protocol was approved by the Committee on the Use of Live Animals in Teaching and Research at the University of Hong Kong. BeWo cells were transfected either with 50 nM of miR-34a miRCURY LNA™ knockdown probe or control (Exiqon, Vedbaek, Denmark). The transfected BeWo cells (1 × 10⁶) were resuspended in 100 μl of PBS, mixed with 100 μl of matrigel (BD Biosciences), and injected subcutaneously into both sides of the posterior flanks of 4- to 6-week-old female B-17/Icr-scid (SCID) mice. The animals were sacrificed after 4 weeks. Four mice were used in each experiment and the experiment was repeated for 5 times independently.

Cell lysates were prepared as described [15]. The protein expression of DLL1, NOTCH1 and β-actin were detected using specific anti-DLL1 (Santa Cruz, sc-9102), anti-NOTCH1 (Santa Cruz, sc-6014) and anti-β-actin antibodies (Santa Cruz, sc-47778). The denatured protein samples were resolved on a 8% denaturing SDS-PAGE and transferred to a nitrocellulose membrane. The membrane was blocked with Tris-buffered saline containing 5% nonfat milk and 0.5% Tween 20 (blocking buffer) at room temperature for 1 hour. Hybridization was performed at 4°C overnight (1^oAb 1:1000 for DLL1 and NOTCH1, 1:10000 for β-actin), followed by extensive washing and incubation with appropriate horseradish peroxidase-conjugated secondary antibody (1:2500) in blocking buffer for 1 hour at room temperature. The protein bands were detected by chemiluminescence detection.

Tissues preparation and immunohistochemistry were performed as described [16]. Briefly, antigen retrieval was performed by heating the sections in 1X target antigen retrieval solution (Dako, Glostrup, Denmark). Non-specific binding was blocked by incubating the tissue sections in PBS containing 5% serum (Sigma-Aldrich) and 0.1% Tween 20. DLL1 immunoreactivities were detected by successive incubation with specific antibody against DLL1 (Santa Cruz), biotinylated polyclonal rabbit anti-goat IgG (Dako) and Strep ABComplex/Horseradish Peroxidase HRP (Vector Laboratories, Burlingame, CA). Signal was visualized with 3,3’-diaminobenzidine (Dako).

We first studied the biological effect of miR-34a in two choriocarcinoma cell lines BeWo and JEG-3 through transfection of pre-miR-34a. We examined the level of miR-34a in the cells at day 3 and day 10 post-transfection, and found that the pre-miR-34a transfected cells had at least ~80-fold higher levels of miR-34a than the control (Additional file 1: Figure S1). Ectopic expression of miR-34a did not significantly affect BeWo cells proliferation in the first 72-hour post-transfection (Figure 1A), but a significant reduction was observed after 7 days (168 hours) of transfection (p < 0.05). The colony formation ability in low seeding density was evaluated 2 weeks post-transfection and the pre-miR-34a transfected cells had around 2–3 times lower colony-forming ability than the scramble precursor transfected cells (Figure 1B, p < 0.05).

Next, we assessed the action of miR-34a on cell invasion. The miR-34a force-expressed cells were allowed to invade a matrigel membrane for 48 hours. It was found that the invasiveness of the miR-34a force-expressed choriocarcinoma cells was significantly decreased when compared with the control group (Figure 1C).

Since miRNA is non-translational, it must exert its effect through regulating target genes. To determine the target gene of miR-34a, we first used in-silico miRNA target prediction tools to find the potential target of miR-34a. Both TargetScan 5.2 (http://​www.​targetscan.​org/​) and PictTar (http://​pictar.​mdc-berlin.​de/​) predict that the Notch ligand DLL1 is a potential target of miR-34a (Figure 2A). We examined the expression of DLL1 in BeWo and JEG-3 cells upon miR-34a force-expression for 3 days, and found that the DLL1 protein level was greatly reduced by miR-34a but not by the scramble miRNA precursor.

NOTCH1 is a known miR-34a targeted gene in choriocarcinoma cells [15]. We compared the action of miR-34a on the protein expression of NOTCH1 and DLL1. It was found that miR-34a force-expression decreased the level of DLL1 to a greater extent than that of NOTCH1 in both BeWo and JEG-3 cells (Figure 2B). Therefore, we focused our study on DLL1.

We further examined whether there is a direct interaction between miR-34a and DLL1. We constructed a luciferase reporter carrying the 3’UTR of DLL1 and transfected the reporter with either the pre-miR-34a or scramble miRNA precursor into BeWo cells. Force-expression of miR-34a reduced the luciferase reporter activity by more than 50% (p < 0.05, Figure 2C). To determine the specificity of the interaction, another reporter vector carrying a mutation at the putative seed binding sequence was constructed. Force-expression of miR-34a had no significant effect on the reporter activities of the mutant construct, confirming the specificity of the action of miR-34a on DLL1.

To study the mechanism of action of miR-34a on DLL1 expression, we determined the mRNA expression of DLL1 upon miR-34a force-expression, RT-qPCR revealed that the treatment and the control groups had similar levels of the DLL1 mRNA (Figure 2D). The observation indicated that miR-34a regulated DLL1 expression in choriocarcinoma cells through translational inhibition. Similarly, the expression of NOTCH1 mRNA was not affected by miR-34a force-expression. To confirm that the action of miR-34a on DLL1 modulated Notch signaling, we examined the expression of the Notch signaling target gene Hairy Enhancer of Split-1 protein (Hes-1) and found that it was reduced upon force-expression of miR-34a in both cell lines (Figure 2E).

DLL1 treatment significantly increased cell invasion (Figure 3A and B), whilst treatment with anti-DLL1 antibody inhibited around 30% of the invasion. On the other hand, force-expression of NCID increased cell invasion by more than 2-fold. These treatments did not significantly affect proliferation as reflected by the cell proliferation assay (Figure 3C). We next determined the role of Notch signaling activation on the action of miR-34a on cell invasion. As shown in Figure 4A, the effect of force-expression of miR-34a was nearly completely nullified by Notch signaling activation but not by the control treatment. Again, the treatments did not affect cell proliferation (Figure 4B). Moreover, RT-qPCR showed that miR-34a force-expression reduced around 35% of the urokinase-type plasminogen activator (uPA) and 55% of the matrix metalloproteinase-9 (MMP9) expression (Figure 4C). Thus, we concluded that miR-34a force-expression reduced the invasiveness of BeWo cells through DLL1 and the Notch signaling pathway.

To examine whether miR-34a knockdown affects tumor formation in vivo, we subcutaneously inoculated miR-34a knockdown BeWo cells or scramble knockdown cells into SCID mice. Inhibition of miR-34a significantly increased the weight of the xenografts by day 28 when compared with xenografts transfected with scramble control (p < 0.05, Figure 5A-C). Immunostaining showed that miR-34a knockdown increased the expression of DLL1 in the xenografts when compared to the control (Figure 5D).