MicroRNA-362 negatively and positively regulates SMAD4 expression in TGF-β/SMAD signaling to suppress cell migration and invasion

Abstract

Cell migration and invasion are modulated by epithelial-to-mesenchymal transition (EMT) and the reverse MET process. Despite the detection of microRNA-362 (miR-362, both the miR-362-5p and -3p species) in cancers, none of the identified miR-362 targets is a mesenchymal or epithelial factor to link miR-362 with EMT/MET and metastasis. Focusing on the TGF-β/SMAD signaling pathway in this work, luciferase assays and western blot data showed that miR-362 targeted and negatively regulated expression of SMAD4 and E-cadherin, but not SNAI1, which is regulated by SMAD4. However, miR-362 knockdown also down-regulated SMAD4 and SNAI1, but up-regulated E-cadherin expression. Wound-healing and transwell assays further showed that miR-362 knockdown suppressed cell migration and invasion, effects which were reversed by over-expressing SMAD4 or SNAI1, or by knocking down E-cadherin in the miR-362 knockdown cells. In orthotopic mice, miR-362 knockdown inhibited metastasis, and displayed the same SMAD4 and E-cadherin expression profiles in the tumors as in the in vitro studies. A scheme is proposed to integrate miR-362 negative regulation via SMAD4, and to explain miR-362 positive regulation of SMAD4 via miR-362 targeting of known SMAD4 suppressors, BRK and DACH1, which would have resulted in SMAD4 depletion and annulment of subsequent involvement in TGF-β signaling actions. Hence, miR-362 both negatively and positively regulates SMAD4 expression in TGF-β/SMAD signaling pathway to suppress cell motility and invasiveness and metastasis, and may explain the reported clinical association of anti-miR-362 with suppressed metastasis in various cancers. MiR-362 knockdown in miR-362-positive cancer cells may be used as a therapeutic strategy to suppress metastasis.

Keywords: EMT/MET; SMAD4; TGF-β/SMAD signaling pathway; cell migration/invasion; metastasis.; microRNA-362.

Figure 1

MiR-362 regulates expression of SMAD4 and E-cadherin of the TGF-β/SMAD signaling pathway. (A) Relative expression levels of miR-362 in cancer cell lines determined by qRT-PCR. (B) Predicted miR-362 targeting of the SMAD4, SNAI1 and CDH1 transcripts. Top panel: Predicted miR-362-3p (vertical green bars) and -5p (red bars) target sites in the 3'-untranslated regions (3'UTR) of the transcripts. The boxed 3'UTR segments were cloned into the pmirGLO vector for luciferase analysis. The GenBank accession numbers of the transcript sequences are also shown. AAA, polyA tail. The scale bars (in bp) indicating nucleotide positions are only for the 3'UTR sequences. Bottom panel: Validation of miR-362 targeting determined by luciferase assays. In each case, a wild-type (WT) or seed sequence-mutated (Mut) luciferase construct was co-transfected into MCF7 cells in triplicates in the presence (“+”) or absence (“-'') of a miR-362-5p or -3p mimic, or a sequence-scrambled negative control (NC) mimic, before performing the dual luciferase assays. The fold change in relative luciferase activity was plotted. The mean ± SEM data shown were obtained from three independent experiments. (C, D) MiR-362 over-expression (C) down-regulated expression of SMAD4 and E-cadherin but not SNAI1 (D). MiR-362 over-expression (OE) was achieved by transient transfection of a miR-362-3p or -5p mimic in MCF7 cells; NC was a scrambled oligonucleotide. (E, F) MiRNA-362-knockdown (E) led to E-cadherin over-expression but suppressed SMAD4 and SNAI1 expression (F). MiR-362 knockdown (KD) was achieved by transduction of a miR-362-3p or -5p sponge vector in HCT-15 cells; NC was a construct with a scrambled sequence insertion. In (D) and (F), quantification of the western blot data (mean ± SEM) from three independent experiments is shown in the right panels of the blots. In all subfigures, *p < 0.05, **p < 0.01 and ***p < 0.001 relative to the negative controls (NC).

Figure 2

MiR-362 promotes cell migration and invasion in vitro. (A) Wound‑healing assays of MCF7 cells untransfected (UT) or transfected with a miR‑362‑5p or -3p, or a negative control (NC) mimic. The images were captured before (0 h) and 24 h after wound healing began. Quantification of the data obtained from three independent experiments is shown in the bar charts on the right, normalizing to data of the untransfected cells. (B, C) Effects of miR-362 over-expression (B) or knockdown (C) on cell migration and invasion in transwell assays. MiR-362 over-expression and knockdown were achieved as described in Fig. 1C above. Representative images captured 24 h post-transfection (left panels) and the mean ± SEM from three independent experiments relative to the negative control groups (right panels) are shown; *p< 0.05, **p< 0.01 relative to the negative controls (NC).

Figure 3

SMAD4 and SNAI1 overexpression and E-cadherin knockdown restore the cell migration and invasion phenotypes in miR-362-knockdown cells in vitro. (A, B) Effects of SMAD4 and SNAI1 over-expression on cell migration and invasion in the miR-362-knockdown HCT-15 cells. SMAD4 and SNAI1 over-expression (OE) was achieved by transient transfection of the respective expression plasmid constructs, followed by western blot analysis (A) and effects on cell migration and invasion were assayed in transwell assays (B). (C, D) Effects of E-cadherin knockdown on cell migration and invasion in the miR-362-knockdown HCT-15 cells. E-cadherin knockdown was achieved by transfection of a siCDH1 siRNA (C) and effects on migration and invasion (D) were assayed in transwell plates. In the over-expression experiments, a blank (NC) expression plasmid was used as a control; in the siRNA knockdown experiments, a validated negative control (NC) siRNA was used. Data of mean ± SEM from three independent experiments are shown. *p<0.05; **p<0.01 values were calculated relative to the negative control samples.

Figure 4

Similar in vivo expression patterns of SMAD4, SNAI1 and E-cadherin as in in vitro on miR-362 knockdown. (A) Tumors developed in ectopic mice generated by subcutaneous injection of miR-362-knockdown cells, or cells transduced with a negative control (NC) sponge vector. Total RNAs and protein lysates were prepared from the tumors for further analysis. (B) Knockdown miR-362 expression in the mouse tumors determined by qRT-PCR. Data were derived from analysis of three different tumors harvested from three independently generated mice using different cell batches (see Suppl. Fig. S1). (C, D) Protein expression in the tumors of the miR362-knockdown ectopic mice as shown in western blots (C) and immunohistochemical staining (D). In (D), scale bar: 20 μm.

Figure 5

MiR-362 knockdown suppresses metastasis in vivo. (A) Tumor growth and metastasis of untreated HCT-15 cells (NC, negative control) in an orthotopic mouse model generated as described in Materials and Methods (Table 1). Images of tumors derived from two independently generated mice are shown; two metastatic nodes of mice #28 are also shown. In the images, yellow arrows indicate the initial tumor masses at the site of injection whereas red arrows indicate the metastasized tumor loci. (B) Suppressed tumor growth and metastasis of miR362-knockdown HCT-15 cells in orthotopic mice (see also Table 1). In each experimental group, images on the left and right were obtained under white or fluorescent light, respectively.

Figure 6

MiR-362-driven TGF-β/SMAD signaling modulation of cell migration and invasion: a proposed scheme. The scheme depicts the normal miR-362 negative regulation supported by data described in this work (A), and also a proposed mechanism of positive regulation of SMAD4 (B) in modulating the TGF-β signaling pathway and down-regulated expression of the EMT gene, SNAI1, and concurrently up-regulated expression of the MET gene, E-cadherin, leading to suppressed cell migration and invasion in vitro and cancer cell metastasis in vivo. See Discussion section for a full description of the scheme. Up- (green) and downward (red) pointing arrows indicate up- and down-regulation, respectively. Solid-line arrows and blunt-ends indicate activation and suppression, respectively, of the indicated events. The circled P denotes phosphorylation. Blue lines and arrows indicate predicted miR-362 targeting not further investigated in this work. In (B), the predicted BRK mRNA-miR-362-3p alignment shown at the top was derived by MicroT-CDS and that of DACH1 mRNA-miR-362-5p by Targetscan and MicroT4-CDS (see also Suppl. Figure S2 for further details).