miR-875-5p exerts tumor-promoting function via down-regulation of CAPZA1 in esophageal squamous cell carcinoma

Cell culture and treatment

Human ESCC cell lines, YSE2, KYSE30, KYSE70, KYSE140, KYSE150, KYSE180, KYSE410, KYSE450, KYSE510, COLO680 were provided by Professor Y. Shimada of Kyoto University. YSE2, KYSE30, KYSE70, KYSE140, KYSE180, KYSE410, KYSE450, KYSE510, COLO680 were cultured in 90% RPMI1640 medium, with 10% fetal bovine serum and antibiotics, at 37 °C, 5% CO₂; KYSE150 was cultured in the medium of 1:1 mixture of Ham’s F12 and RPMI-1640 containing 2% FBS and antibiotics at humidified atmosphere with 37°C, 5% CO₂.

Stable and transient transfection

Stable recombinant plasmids over-expressing CAPZA1(T) and CAPZA1(G) by lentiviral infection and matched control vectors were constructed by Obio Technology Co. Ltd (Shanghai, China). Plasmids over-expressing CAPZA1 CDS and matched control vectors pcDNA3.1were constructed by Generay Biotech Co. Ltd (Shanghai, China). The target genes [CDS-3′UTR(T) and CDS-3′UTR(G)] were inserted into the lentiviral vector (H145 pLenti-EF1a-EGFP-F2A-Puro-CMV-MCS). Luciferase pGL3 overexpression vectors [3′UTR(T) and 3′UTR(G)] and pGL3 control plasmids were constructed by Generay Biotech Co. Ltd (Shanghai, China). miR-875-5p mimics, miR-875-5p inhibitors, biotinylated miR-875-5p, and matched negative controls (NCs) were synthesized by Guangzhou RiboBio Co. Ltd. (Guangzhou, China). For stable transfection, cells at 70%–80% density were infected with lentivirus using Polybrene in serum-free medium, then stable expressing cells were selected with an optimal concentration of puromycin. For transient transfection or co-transfection, cells were transfected in serum-free medium using Lipofectamine2000 following the manufacturer’s protocol (Invitrogen, Carlsbad, California, USA) when 70%–80% confluence was reached. The cells were then harvested and followed by PCR or Western blot analysis.

Cell proliferation assays

As previously descripted, cell proliferation was detected by using the xCELLigence Real-Time Cell Analyzer (RTCA)-MP system (Acea Biosciences/Roche Applied Science). 2,000 cells per well were suspended in E-Plate 96 (Roche Applied Science) with 100 µl medium and monitored in real time for 96 h. Cell index was read automatically every 15 min and the recorded growth curve were shown as cell index SD.

The colony formation assays were employed to test the colony formation ability of cells. 500–600 cells were seeded per well in 60 mm plates and grown for 10–14 days. Then cells were fix with methanol for 15 min and stained with 1% crystal violet-acetic acid solution for 20 min, colonies were visualized and quantitated by G:box (Syngene).

Invasion and migration assays

A Matrigel™  invasion chamber (BD Biosciences, Bedford, MA, USA) was used to evaluate the migratory activity of ESCC cells. A total of 3 × 10⁴ cells/ml were suspended with serum-free RPMI basal medium and seeded in a 24-well chamber, then filled with 600–800 µl of RPMI-1640 medium with 20% FBS per well beneath the chamber. After overnight incubation, the cells which had migrated through the Matrigel™  membrane were fixed with methanol and stained with 1% crystal violet-acetic acid solution for 20 min.

For invasion assays, Matrigel™  matrix solution (1% Matrigel™  matrix diluted with serum-free RPMI basal medium) was prepared, the plates were incubated for 2 h, and then invasive activity was detected as described above. The number of invasive and migrating cells were quantified by fluorescence microscopy.

DNA, RNA extraction and quantitative real-time PCR

Total DNA of cells was isolated by extraction kit according to the standard procedure (Tiangen, Beijing, China). Total RNA of cells was extracted with TRIzol reagent according to the standard procedure. Reverse transcription was performed by using Superscript II-reverse transcriptase kit (Invitrogen) following the manufacturer’s instructions. The real-time PCR was performed in triplicate with the Premix Ex Taq kit (Takara) and a 7300 real-time PCR system (Life Technologies) according to manufacturer’s instructions. Thermal cycling conditions were as follows: the first step, 95 °C for 30s and the ensuing 40 cycles, 95 °C for 5s, 65 °C for 31 s, and melt curve step: 95 °C for 15s, 65 °C for 1min, 95 °C for 15s. The PCR primers were provided by Invitrogen. GAPDH and U6 mRNA was employed as an endogenous control for mRNA.

3′UTR primers for PCR:

F:5′ GCCTCATGGAATACTGTTGAACC 3′;

R:5′  GGATAGATCACTCTCTCACC 3′.

Western blotting analysis

Transfected cells were collected using lysis buffer (1 ×PBS + 4% NP-40 + 0.2% proteinase inhibitor) and lysed by centrifugation at 12,000 rpm for 20 min at 4 °C. The supernatants were added with 2 × SDS–PAGE sample loading buffer and boiled for 5 min. Eighty micrograms of protein was fractionated by SDS–PAGE, then transferred to polyvinylidene difluoride membranes. The membranes were incubated with relative primary antibodies overnight at 4 °C, then secondary antibodies were incubated and quantified using Image Quant software (GE Healthcare Biosciences, Pittsburg, PA, USA).

Biotin-labeled pull-down assays

Biotinylated miR-875-5p (Guangzhou RiboBio Co, Ltd, China) pull-down assay with target mRNAs was performed as described (Christoffersen et al., 2009; Shi, Han & Spivack, 2013; Wynendaele et al., 2010). Cells were transfected with control miRNA and biotin-labeled miR-875-5p and whole-cell lysates were harvested. Simultaneously, Streptavidin beads (Invitrogen) were coated with 10 µl per sample yeast tRNA (Ambion) and incubated 2 h at 4 °C. Then the beads were washed with 500 µl lysis buffer three times and resuspended with 50 µl lysis buffer. Sample lysates were mixed with pre-coated beads (50 µl per sample) and incubated with rotator overnight at 4 °C. Beads were pellet down by centrifugation at 5,000 rpm for 1min at 4 °C to remove unbound materials and then washed five times with 500 µl ice-cold lysis buffer. Next, 750 µl of TRIzol (Invitrogen) and 250 µl nuclease-free water was added to both input and pull-down samples to isolate RNA. Finally, qPCR was employed to detect the expression ratio.

Statistical analysis

GraphPad Prism v.5 software (GraphPad Software, Inc., La Jolla, CA, USA) and R version 4.0.2 software were used for statistical analyses. Data of colony numbers and mRNA expression levels of CAPZA1 were analysed by ANOVA. Data of colony formation, cell invasion and migration numbers of miR-875-5p, luciferase activity and the enrichment of CAPZA1 mRNAs were analysed by Student’s t-test (two-tailed). Statistical differences are presented as the mean ± standard deviation (SD). A p value <0.05 was considered statistically significance.

miR-875-5p promoted cell proliferation and metastasis of ESCC cells

Given that miRNAs can function as oncogenes or tumor suppressors in the progression of human cancers (Chen, 2005; Esquela-Kerscher & Slack, 2006; Nohata et al., 2013), we first assessed the functions of miR-875-5p in ESCC cells. According to the TCGA and cBioProtal for Cancer Genomics database, we found that the gene which miR-875-5p encodes was amplified 6% and 5% in EAC and ESCC, respectively, while approximately 1% were deleted in EAC and ESCC (Fig. 1A). To determine the function of miR-875-5p in ESCC, we first introduced miR-875-5p mimics and miR-875-5p negative control(NC) into ESCC cells, then we performed colony formation and transwell assays in KYSE180 and KYSE510 cells. As shown in Figs. 1B to 1G, the expression of miR-875-5p significantly increased colony formation of ESCC cells compared with NC transfected cells, indicating that miR-875-5p promoted the proliferation ability of ESCC cells. In transwell assays, miR-875-5p markedly enhanced the migration and invasion of ESCC cells compared with NC group in KYSE180 and KYSE510 cells (Figs. 1H to 1S), which indicated that miR-875-5p facilitated metastasis of ESCC cells.

miR-875-5p targeted the CAPZA1 3′UTR containing the T>G alteration at rs373245753

Because SNPs may affect the interactions between miRNAs and their target genes, SNPs are thought to interfere with the mRNA expression of target genes by perturbing miRNA-mediated gene regulation (Chen et al., 2008). We used TargetScan database and predicted that miR-875-5p targeted the CAPZA1 3′UTR containing the T>G alteration at rs373245753 (Fig. 2A), which was show in our previous study involving WGS and WES on ESCC (Song et al., 2014). CAPZA1 is located at chromosome 1 in humans with a length of 52167 bp (Fig. 2B). In addition, we employed a pair of PCR primers based on the 3′UTR and performed sequencing of the amplification products in 10 ESCC cell lines (YSE2, KYSE30, KYSE70, KYSE140, KYSE150, KYSE180, KYSE410, KYSE450, KYSE510, and COLO680). We showed that the T>G change at rs373245753 in the 3′UTR did not exist in these ESCC cell lines, indicating the need to heterogeneously-express the rs373245753 of 3′UTR in ESCC cells for further study (Fig. 2C). Then, the full length of CAPZA1 3′UTR fragments containing the rs373245753 T or G allele were cloned into the luciferase reporter vector, pGL3, to compare the luciferase activities between the two alleles (Fig. 2D). The pGL3 vectors containing the rs373245753 T or G allele of the CAPZA1 3′UTR were co-transfected in parallel with NC and miR-875-5p mimic to KYSE180 cells and luciferase activity was assayed. As shown in Fig. 2E, for the pGL3-CAPZA1(T) construct, the miR-875-5p mimic significantly reduced luciferase activity compared with NC, while no difference was observed in the pGL3-CAPZA1(G) group. Luciferase activity was measured and normalized to firefly luciferase and the reference is Ranilla luciferase. These findings indicated that miR-875-5p directly targeted the CAPZA1 3′UTR(T), and negatively regulated CAPZA1 expression, while the variant allele attenuated these effects, thus allowing increased CAPZA1 expression in the presence of this variant allele.

CAPZA1(T) decreased the proliferation of ESCC cells and the function of CAPZA1 CDS was not reversed by the mimic of miR-875-5p

Because the specific functions of CAPZA1 in ESCC have not been reported, we first determined the functions of CAPZA1(T) and CAPZA1(G) in ESCC. Cells were stably overexpressed with CAPZA1(T) and CAPZA1(G) by lentiviral infection and we further assessed the functions in KYSE180 and KYSE510 cells. Cell proliferation ability was detected using an xCELLigence Real-Time Cell Analyzer (RTCA)-MP system (Roche, Basel, Switzerland) and growth curves showed that CAPZA1(T) reduced the rate of cell proliferation in KYSE180 and KYSE510 cell lines, whereas CAPZA1(G) substantially abolished these effects (Figs. 3A and 3B). Colony formation abilities in both KYSE180 and KYSE510 cells were markedly suppressed by CAPZA1(T) overexpression, but substantially increased in the CAPZA1(G) overexpression cell lines (Figs. 3C to 3J). Furthermore, we detected the function of CAPZA1 CDS in ESCC cells. Cells were overexpressed with CAPZA1 CDS and cell proliferation ability were assessed in KYSE180 cells. Cell growth curves and colony formation assays showed that CAPZA1 CDS significantly reduced the rate of cell proliferation in KYSE180 cells (Figs. 3K to 3N). Additionally, we wondered that whether the suppressive effect was reversed by miR-875-5p. Cells were co-transfected with CAPZA1 CDS and miR-875-5p, then cell growth curves and colony formation assays showed that the function of CAPZA1 CDS was not reversed by miR-875-5p (Figs. 3O to 3R). These results suggested that CAPZA1 (T) decreased the proliferation of ESCC cells and the function of CAPZA1 CDS was not reversed by the mimic of miR-875-5p.

miR-875-5p differentially regulated CAPZA1 expression in the presence of the SNP, rs373245753 T

miRNAs can regulate gene expression and generate mRNA cleavage or translation repression by base pairing with their target mRNAs (Bartel, 2004; Lee & Vasudevan, 2013). To determine whether miR-875-5p preferentially regulates CAPZA1 protein and mRNA levels of the CAPZA1 (T) and CAPZA1(G) transcripts, pGL3 vectors containing the rs373245753 T or G of the CAPZA1 3′UTR and miR-875-5p mimic were co-transfected into KYSE180 and KYSE510 cells. We found that the protein levels were markedly reduced in CAPZA1(T) cells compared to CAPZA1(G) and controls in KYSE180 and KYSE510 cells, which suggested that miR-875-5p might bind to CAPZA(T) and represses the protein expression of CAPZA1 (Figs. 4A and 4B). In addition, qPCR assays showed that miR-875-5p could target CAPZA1(T) and suppressed CAPZA1 mRNA levels in KYSE180 and KYSE510 cells (Fig. 4C and 4D). These observations indicated that miR-875-5p targets CAPZA1(T) and negatively regulates CAPZA1 mRNA expression.

The SNP, rs373245753 G, disturbed the binding of miR-875-5p and CAPZA1 mRNA

Increasing evidence has revealed that SNPs in the 3′UTRs of genes targeted by miRNAs can disturb or obstruct miRNA binding and consequently influence regulation of target genes, which might be associated with cancer (Sethupathy & Collins, 2008). To determine whether the SNP, rs373245753 T>G, influences the binding of miR-875-5p and CAPZA1 mRNA, RNA affinity pull-down assays were performed (Christoffersen et al., 2009; Shi, Han & Spivack, 2013; Wynendaele et al., 2010). The pGL3 vectors containing the rs373245753 T or G allele of the CAPZA1 3′UTR were co-transfected with control or biotin-labeled miR-875-5p into KYSE180 and KYSE510 cells. Cell lysates were incubated with streptavidin-coated beads. RNA was harvested from the pull-down materials and amplified with CAPZA1 by qPCR. A biotin-labeled scrambled miRNA served as the control in these experiments. Firstly, qPCR assays were performed to measure the levels of pull-down and input CAPZA1 mRNA in the materials pulled down by biotin-miR-875-5p in ESCC cell lines (Figs. 5A to 5D). We found that CAPZA1(T) was markedly elevated pulled down by biotin-miR-875-5p than CAPZA1(G) groups, which indicated that the SNP would disturbed the binding of miR-875-5p and CAPZA1 mRNA (Figs. 5A and 5B). Subsequently, the levels of CAPZA1 mRNA were significantly enriched in the pull-down material isolated from KYSE180 and KYSE510 cells following co-transfection biotin-labeled miR-875-5p with CAPZA1(T) compared to CAPZA1(G). We obtained that CAPZA1(T) were more highly bound with biotin-labeled miR-875-5p than CAPZA1(G) (Figs. 5E and 5F). Taken together, these results illustrated that the SNP, rs373245753 T>G, disturbed the binding of miR-875-5p and CAPZA1 mRNA.