MicroRNA-140 Inhibits the Epithelial-Mesenchymal Transition and Metastasis in Colorectal Cancer.

MicroRNA-140, a cartilage-specific microRNA, has recently been implicated in the cancer progression. However, the comprehensive role of miR-140 in the invasion and metastasis of colorectal cancer (CRC) is still not fully understood. In this study, we confirmed that miR-140 downregulates SMAD family member 3 (Smad3), which is a key downstream effector of the TGF-β signaling pathway, at the translational level in the CRC cell lines. Ectopic expression of miR-140 inhibits the process of epithelial-mesenchymal transition (EMT), at least partially through targeting Smad3, and induces the suppression of migratory and invasive capacities of CRC cells in vitro. miR-140 also attenuates CRC cell proliferation possibly via downregulating Samd3. Furthermore, overexpression of miR-140 inhibits the tumor formation and metastasis of CRC in vivo, and silenced Smad3 has the similar effect. Additionally, miR-140 expression is decreased in the clinical primary CRC specimens and appears as a progressive reduction in the metastatic specimens, whereas Smad3 is overexpressed in the CRC samples. Taken together, our findings suggest that miR-140 might be a key suppressor of CRC progression and metastasis through inhibiting EMT process by targeting Smad3. miR-140 may represent a novel candidate for CRC treatment.

Introduction

Metastasis is the biological feature and leading cause of death by malignant tumor. It consists of sequential and interrelated steps, and many of them are favored by epithelial-mesenchymal transition (EMT). Obtaining the mesenchymal state endows cells with enhanced migration and invasion, induces stemness, and prevents apoptosis and senescence, allowing their subsequent roles in the initiation of metastasis. The EMT process has been reported in a variety of epithelial malignancies, including colorectal cancer (CRC).2, 4

CRC is the third leading cause of death from cancer, and it has been estimated to be responsible for 50,000 deaths per year in the United States. Although surgery combined with adjuvant therapy has increased the prognosis of CRC, metastasis still occurs in approximately half of all CRC patients at the time of diagnosis or after intended curative treatment, giving rise to a poor 5-year survival rate of slightly greater than 10%.6, 7 A good explanation of metastasis of CRC would be helpful for the development of effective therapeutic strategies.

MicroRNAs (miRNAs) are noncoding small RNA molecules that post-transcriptionally regulate protein expression via binding to the 3′ UTR of the target genes.8, 9 Some studies have shown the importance of miRNAs in the regulation of EMT either as an inhibitor or as a promoter.2, 10, 11 SMAD family member 3 (Smad3) is one of the key downstream factors of transforming growth factor β (TGF-β signaling pathway, which is a critical inducer of the EMT in the malignant tumor, such as in CRC.12, 13 TGF-β binds with its two transmembrane receptors, type I (TGFBRI) and type II (TGFBRII), leading to the phosphorylation of Smad2 and Smad3, which translocate to the nucleus with Smad4, where it can regulate the downstream EMT-related gene expression.14, 15, 16 Importantly, Smad3 has been confirmed as the direct target of microRNA-140-5p (miR-140) in a pluripotent mouse embryonic fibroblast cell line and alveolar epithelial cells (AECs).17, 18 We previously found that the Smad3 protein is downregulated by miR-140 in CRC cells using western blot.

miR-140 was first shown to be specifically expressed in the cartilage of developing zebrafish and mouse embryos during both long and flat bone development.20, 21 Later, Miyaki et al. reported that miR-140 loss induces osteoarthritis (OA)-like changes in vivo. Our laboratory is the first one that investigated the mechanism of miR-140 in the tumorigenesis, to the best of our knowledge. We found that miR-140 expression is reduced in the CRC specimens and that the ectopic expression of miR-140 can suppress cell proliferation and induce cell-cycle G1 and G2 phase arrest, mediated in part through the downregulation of HDAC4 in the CRC and osteosarcoma cells. Several recent studies have demonstrated that miR-140 can inhibit tumor invasion and metastasis in hepatocellular carcinoma (HCC) and non-small cell lung cancer (NSCLC) through downregulating TGFBRI, FGF9, or IGF1R.24, 25 miR-140 can suppress CRC stem cell (CSC) survival and its metastatic potential by inhibiting Smad2 and autophagy. Downregulation of miR-140 induces EMT and promotes esophageal cancer cell invasion via targeting Slug. Zhang et al. reported that miR-140 inhibits the progression of CRC by targeting VEGFA. Our recent work showed that miR-140 suppresses CRC invasion and metastasis by targeting ADAMTS5 and IGFBP5. However, the impact and comprehensive mechanism of miR-140 in the invasion and metastasis of CRC still needs to be elucidated.

In this study, we found that Smad3 is the direct target of miR-140 in the CRC cell lines via the luciferase assay, and miR-140 regulates the expression of Smad3 at the translational level. Ectopic expression of miR-140 inhibits the process of EMT, at least partially through targeting Smad3, and therefore suppresses the migratory and invasive capacities of CRC cells in vitro. miR-140 attenuates CRC cell proliferation, possibly via downregulating Samd3. Subcutaneous or tail vein injection of HCT 116 cells transfected with lentiviral vectors of miR-140 or short hairpin RNA (shRNA) against Smad3 significantly decreases the growth of local tumor nodules or inhibits the pulmonary metastasis potential in BALB/c mice, respectively. Furthermore, the expression level of miR-140 was significantly decreased in the clinical primary CRC specimens compared to the corresponding adjacent normal colorectal tissues. A progressive reduction of miR-140 was found from the primary tumors to the metastatic tumors. The expression level of Smad3 protein was significantly decreased in the corresponding adjacent normal colorectal tissues compared to the clinical primary CRC specimens. miR-140 might be a key inhibitor of CRC progression and metastasis through suppressing the EMT process by targeting the downstream Smad3 of the TGF-α signaling pathway.

Results

miR-140 Regulates Smad3 in CRC Cells

Bioinformatics analysis indicated that the 3′UTR of Smad3 mRNA has a binding site of miR-140. To investigate the direct binding and repression effect between miR-140 and Smad3, a luciferase assay was carried out. As Figure 1A shows, cotransfection of miR-140 and Smad3-3′UTR-wt (miR-140+Smad3-3′UTR-wt) reduced the luciferase activity, compared with cotransfection of negative miRNA and Smad3-3′UTR-wt (NC+Smad3-3′UTR-wt), and such suppressive effect was not found in the cotransfection of miR-140 and Smad3-3′UTR-mut (miR-140+ Smad3-3′UTR-mut). We then overexpressed miR-140 by transient transfection in HCT116 and RKO cells using Oligofectamine, and the successful transfection of miR-140 was confirmed by real-time qRT-PCR (Figure S1). The siRNA against Smad3 (siSmad3) was used as a positive control. CRC cells (Figure S1, Con) and the negative miRNA (Figure S1, NC) were the negative controls. We accessed the expression level of Smad3 mRNA using real-time qRT-PCR. As Figure 1B shows, there was only a slight decrease in Smad3 mRNA expression after miR-140 transfection; by contrast, knockdown of Smad3 dramatically induced the Smad3 mRNA degradation. We next measured the protein level of Smad3 using western blot. Overexpression of miR-140 significantly reduced the Smad3 protein expression in both of these two cell lines compared with the negative controls, while Smad3 siRNA transfection had a similar effect (Figure 1C). The active Smad3, phosphorylating Smad3 (p-Smad3) was reduced as well (Figure 1C). The real-time qRT-PCR and western blot results indicated that miR-140 significantly inhibits the expression of Smad3 protein without considerable degradation of mRNA.

Figure 1

Smad3 Is the Direct Target of miR-140 in CRC Cells

(A) A predicted miR-501 binding site existed in the 3′UTR of Smad3 mRNA (wt) and its mutant type (mut) were generated in the binding site. Transfection of miR-140 mimic reduced the firefly luciferase activity of Smad3-3′UTR-wt compared to the Smad3-3′UTR-mut; the negative miRNA transfection with Smad3-3′UTR-wt (Negative miRNA+Smad3-3′UTR-wt) was also the negative control. n = 3; *p < 0.05; **p < 0.01. (B) Ectopic expression of miR-140 decreased the mRNA level slightly, whereas siRNA against Smad3 dramatically degraded the Smad3 mRNA by real-time qRT-PCR. HCT116 and RKO cells were transfected with miR-140 mimic (miR-140) using Oligofectamine. CRC cells alone (Con) and negative miRNA (NC) were the negative controls, and siRNA against Smad3 (siSmad3) was the positive control. GAPDH was used as an endogenous control. n = 3; **p < 0.01. (C) Transfection of miR-140 or Smad3 knockdown inhibited Smad3 and phosphorylating Smad3 (p-Smad3) protein levels using western blot analysis. The quantitative results of western blot were analyzed by Gel-Pro Analyzer 4.0 software, and GAPDH was used as an endogenous control. n = 3; **p < 0.01. (D) Knockdown of endogenous miR-140 restored the expressions of Smad3 and p-Smad3 by western blot analysis. HCT116 and RKO cells were transfected with miR-140 inhibitor (140i) using Lipofectamine 2000. CRC cells (Con), scrambled miRNA inhibitor (i-NC), and co-transfection of miR-140 inhibitor and siRNA against Smad3 (140i+si) were the negative controls. The quantitative results of western blot were analyzed by Gel-Pro Analyzer 4.0 software, and GAPDH was used as an endogenous control. n = 3; **p < 0.01. The data are presented as mean ± SD.

To further confirm Smad3 is regulated by miR-140, we performed the loss-of-function experiment by knocking down the endogenous miR-140 with inhibitor of miR-140 in HCT116 and RKO cells. CRC cells (Figure 1D, Con) and scrambled miRNA inhibitor (Figure 1D, i-NC) were used as the negative controls. Meanwhile, to more accurately prove the relationship between miR-140 and Smad3, we co-transfected miR-140 inhibitor and siRNA against Smad3 (140i+si) into the CRC cells and considered it as a negative control. Our results showed that knocking down endogenous miR-140 could restore the expressions of Smad3 and p-Smad3, compared to the negative controls (Figure 1D). Based on these data, Smad3 is directly regulated by miR-140 in the translational level in the CRC cells.

Subsequently we examined the expressions of TGF-β signaling-pathway-related proteins Smad2 and Smad4 using western blot. Smad2 has previously been confirmed to be the direct target of miR-140. Our results showed that ectopic expression of miR-140 reduced Smad2 protein expression in both of these cell lines compared with the negative controls, while Smad4 protein was also reduced (Figure S2). Smad4 was not found to be the predicting target of miR-140, using miRNA bioinformatics analysis. A recent study demonstrated that Smad4 protein is decreased in shRNA against Smad2-transfected bone-marrow-derived mesenchymal stem cells (BMSCs). In combination with these data, miR-140 could inhibit the TGF-β signaling-pathway-related proteins Smad2 and Smad3 directly and indirectly downregulate Smad4, possibly via decreased Smad2 expression.

miR-140 Represses the CRC Cell Migration and Invasion

Since Smad3 was confirmed to be the target of miR-140, we then investigated the role of miR-140 in the migration and invasion in CRC cells using scratch-wound and Transwell chamber assays. HCT116 and RKO cells were transfected with miR-140 mimic or siRNA against Smad3, and the successful transfection of miR-140 or Smad3 was confirmed by real-time qRT-PCR (data not shown). The scratch-wound assay and Transwell chamber assay without Matrigel showed that miR-140 overexpression significantly attenuated the migratory capacity in these two cell lines (Figures 2A and 2B). Meanwhile, the Transwell chamber assay with Matrigel demonstrated that miR-140 overexpression dramatically decreased the number of CRC cells passing through the membrane (Figure 2C). These repressive effects of miR-140 on migration and invasion were similar to those induced by Smad3 siRNA in CRC cells (Figures 2A–2C).

Figure 2

miR-140 Represses the Migratory and Invasive Capacities of CRC Cells

(A) The scratch-wound assay showed that miR-140 overexpression or silenced Smad3 attenuated the migratory capacity in HCT116 and RKO cells compared to the negative controls, CRC cells (Con) and negative miRNA (NC). **p < 0.01. (B) miR-140 overexpression or silenced Smad3 decreased the migratory capacity measured by the Transwell migration assay. **p < 0.01. (C) The Transwell invasion assay showed that the ectopic expression of miR-140 or silenced Smad3 decreased the CRC cells passing through the membrane. **p < 0.01. (D) The scratch-wound assay showed that miR-140 knockdown enhanced the migratory capacity in HCT116 and RKO cells compared to the negative controls (Con, i-NC, and 140i+si). **p < 0.01. (E) The transwell migration assay showed that miR-140 knockdown promoted the CRC cell migratory capacity. **p < 0.01. (F) The transwell invasion assay showed that miR-140 knockdown enhanced the CRC cell invasive capacity. **p < 0.01. Each independent experiment was repeated three times, and data are presented as mean ± SD. Similar results were obtained, and the representative photographs are shown.

On the contrary, knocking down of endogenous miR-140 expression by the inhibitor significantly enhanced the migratory and invasive capacities of HCT116 and RKO cells (Figures 2D–2F). These results, therefore, indicated that miR-140 is a repressor of migration and invasion in CRC.

miR-140 Suppresses the CRC Cell Proliferation

We also investigated the impact of Smad3 on the CRC cell proliferation. First, a CCK-8 assay was performed. As Figure 3A shows, either upregulated miR-140 or silenced Smad3 inhibited the CRC cell growth significantly when compared to the negative controls. We next performed the soft agar colony formation assay and counted the colonies on day 14 after plating. The results showed that both enforced expression of miR-140 and silenced Smad3 suppressed the number of colonies remarkably, as compared to the negative controls (Figure 3B). On the contrary, knockdown of the endogenous miR-140 reversed the suppressive effect on the cell growth and colony formation capacities mediated by miR-140 overexpression (Figures 3C and 3D). Taken together, these findings revealed that manipulated expression of miR-140 inhibits the CRC cell proliferation via downregulation of Smad3.

Figure 3

miR-140 Suppresses CRC Cell Growth

(A) Both enforced expression of miR-140 and silenced Smad3 decreased the cell proliferation and the number of colonies remarkably, as compared to the negative controls (Con and NC) measured by CCK-8 assay. **p < 0.01. (B) The colony formation assay showed that miR-140 overexpression and silenced Smad3 decreased the number of colonies as compared to the negative controls (Con and NC). **p < 0.01. (C) The CCK-8 assay showed that miR-140 knockdown promoted the CRC cell proliferative potential as compared to the negative controls (Con, i-NC, and 140i+si). **p < 0.01. (D) The number of colonies was increased by miR-140 knockdown when compared to the negative controls (Con, i-NC, and 140i+si). **p < 0.01. Each independent experiment was repeated three times, and data are presented as mean ± SD.

miR-140 Inhibits EMT in CRC Cells

Given the relationship between miR-140 and Smad3, and the suppression of miR-140/Smad3 in the CRC cell migration and invasion, we then investigated the role and the underlying mechanism of miR-140 in the EMT process. The expressions of the epithelial marker E-cadherin and the mesenchymal marker vimentin were examined by western blot. miR-140 overexpression increased the level of E-cadherin and decreased the level of vimentin compared with the negative controls, and siRNA against Smad3 resembled the effect of miR-140 overexpression (Figure 4A). Conversely, knockdown of miR-140 suppressed the expression of E-cadherin and promoted the expression of vimentin (Figure 4B). These changes in EMT-associated markers were further confirmed by the immunofluorescence assay (Figures 4C and 4D). Our data suggested that miR-140 represses the EMT process and the subsequent invasive phenotype in CRC cells.

Figure 4

miR-140 Inhibits EMT in CRC Cells

(A) Enforced expression of miR-140 or siSmad3 increased E-cadherin protein levels and decreased vimentin and Zeb1 protein levels compared with the negative controls by western blot analysis. (B) Knockdown of miR-140 reversed the protein expressions of E-cadherin, vimentin, and Zeb1 compared to miR-140 overexpression. (C) Immunofluorescence assay showed that miR-140 overexpression and Smad3 knockdown promoted E-cadherin expression (red) and suppressed vimentin expression (green) as compared to the negative controls (Con and NC). (D) E-cadherin expression was inhibited and vimentin expression was increased by miR-140 knockdown as compared to the negative controls (Con, i-NC, and 140i+si). Scale bars, 25 μm. **p < 0.01. Each independent experiment was repeated three times, and data are presented as mean ± SD.

Transcriptional factors from the zinc-finger E-box-binding homeobox factor (Zeb) family have been reported to be the key regulators of EMT through interacting with Smads to enhance the TGF-β signaling pathway.31, 32 Compared to Zeb2, Zeb1 is reported to synergize with Smad-mediated transcriptional activation.33, 34 Previous studies have demonstrated that TGF-β signaling leads to the activation of Smad proteins, especially Smad3, and then induces the expression of E26 transformation-specific 1 (Ets1), an upstream component of Zeb1, which, in turn, stimulates the promoter activity of Zeb1. Zeb1 can repress the transcription of E-cadherin through binding to its E-box region of promoter.31, 32, 33, 34 Therefore, we examined the expression of Zeb1 using western blot. The results showed that upregulation of miR-140 inhibited the expression of Zeb1 protein, compared to the negative controls, and that siRNA against Smad3 showed the same trend as miR-140 overexpression did (Figure 4A). Furthermore, knockdown of miR-140 reversed the expression level of Zeb1 (Figure 4B). miR-140 suppresses the migratory and invasive capacities by inhibiting the EMT process, and this is achieved, possibly, through targeting Samd3 and subsequent reduced Zeb1 expression to interrupt the TGF-β signaling pathway.

miR-140 Inhibits the CRC Progression and Metastasis In Vivo

Given the observed effects of miR-140 on CRC cell proliferation, migration, and invasion in vitro, we subsequently explored whether miR-140 inhibits the tumor progression and metastasis in vivo. First, we infected HCT116 cells with either lentiviral vectors of miR-140 (Len-140) or shRNA against Smad3 (shSmad3). At 72 hr after infection, the green fluorescence was viewed under the microscope (Figure S3A). The successful transfection of miR-140 and siRNA against Smad3 was then confirmed by real-time qRT-PCR (Figures S3B and S3C). We then injected the cells subcutaneously on the backs of BALB/c mice. The subcutaneous tumor growth was monitored for 4 weeks, and the mice were sacrificed. We found that the local tumor developed in all the mice. However, the tumor volumes in the miR-140-injected group were markedly smaller than those in the negative-control-injected group at all the time points (Figures 5A and 5B), and silenced Smad3 by shRNA reduced the tumor volume similarly with the miR-140-injected group.

Figure 5

miR-140 Inhibits the CRC Development and Metastasis In Vivo

HCT116 cells were transduced with either lentiviral vectors of miR-140 (Len-140) or shRNA against Smad3 (shSmad3), and cells infected with lentivirus control were the negative control (Len-NC). (A) The HCT116 cells transduced with lentiviral vectors of miR-140 (Len-140), negative control (Len-NC), and shRNA against Smad3 (shSmad3) were injected subcutaneously on the backs of BALB/c mice respectively, and the subcutaneous tumor growth was monitored for 4 weeks (n = 6 for each group). (B) miR-140 overexpression or silenced Smad3 markedly decreased the tumor volume compared with the negative control. The values were shown as mean ± SD. **p < 0.01. (C) The local tumors were fixed, paraffin embedded, cut into sections, and stained by immunohistochemistry. The expressions of Smad3, p-Smad3, Zeb1, and vimentin in the groups treated with either miR-140 or shRNA-Smad3 were significantly decreased as compared to the negative control, whereas E-cadherin was upregulated. Representative images were shown (left panel: 100×; right panel: 200×). Scale bars, 50 μm. (D) These proteins were examined by western blot in the mouse tumor tissues and showed the same tendency as immunohistochemistry did. Representative images are shown. (E) HCT116 cells infected with lentiviral vectors were injected into BALB/c mice through the tail vein and monitored for 8 weeks. Lungs and liver were collected for H&E staining. The negative control group induced the visible pulmonary metastasis in all mice (n = 6). Overexpressing miR-140 by lentivirus induced pulmonary metastasis in 4 out of 6 mice, including 1 case of macroscopic metastatic nodules and 3 cases of microscopic nodules (n = 6). Knocking down Smad3 induced microscopic pulmonary metastasis in 4 out of 6 mice, and the remaining 2 mice failed to form pulmonary metastasis (n = 6). The liver was barely involved in all of the three groups. Representative images are shown (upper panel: 100×; lower panel: 200×). Scale bars, 50 μm.

We subsequently used immunohistochemistry to detect the expressions of Smad3, p-Smad3, Zeb1, E-cadherin, and vimentin in the three groups of mouse xenograft tumor tissues. As Figure 5C shows, the expressions of Smad3, p-Smad3, and the downstream effector Zeb1 in the groups treated with either miR-140 or shRNA-Smad3 were significantly decreased as compared to the negative control. The EMT marker E-cadherin was upregulated, whereas vimentin was downregulated. We also measured the levels of these proteins using western blot in the mouse tumor tissues. As expected, their expressions showed the same tendency as immunohistochemistry did (Figure 5D). These data indicated that the miR-140/Smad3 axis inhibits the tumor progression in vivo.

To further understand the impact of miR-140 on the CRC metastasis, we injected HCT116 cells infected with lentiviral vectors into BALB/c mice through the tail vein. Eight weeks later, we euthanized the mice and dissected out the lungs and liver for H&E staining. We found that the negative-control group induced the visible metastatic nodules in the lungs in all six mice (6/6; Figure 5E). However, overexpressing miR-140 by lentivirus induced pulmonary metastasis in 4 out of 6 mice, including 1 case of visible metastatic nodules and 3 cases of microscopic nodules (Figure 5E). As for knockdown of Smad3, 4 out of 6 mice showed the microscopic pulmonary metastatic nodules, and the remaining 2 mice were not involved (Figure 5E; Table 1). The liver was barely involved in all of the three groups. Together, these results demonstrated that miR-140 has a suppressive impact on CRC metastasis.

Table 1

Comparison of Pulmonary Metastasis Rate in Different Mice Models

	Len-NC	Len-140	shSmad3
Pulmonary metastasis rate (%)	100 (6/6)	66.7 (4/6)	66.7 (4/6)
p Value		<0.05a	<0.05a

Compared to Len-NC.

miR-140 Is Associated with CRC Progression and Metastasis

To further determine the significance of miR-140 on CRC progression and metastasis, we detected the expression level of miR-140 on a cohort of CRC specimens, including 31 CRC cases, 11 cases with lymph node metastasis, and 2 cases with liver metastasis, by real-time qRT-PCR. We found that miR-140 was significantly downregulated in the primary CRC tissues compared to the adjacent normal mucosa (Figure 6A, p < 0.05). We then selected the 13 out of 31 cases of CRC patients with lymph node or liver metastasis and extracted RNA from the metastatic tumor tissues. After comparing miR-140 expression levels in the paired primary and metastatic tumor tissues, we found that miR-140 was significantly downregulated in the lymph node and liver metastatic tumors (Figure 6B, p < 0.05). These results further confirmed that miR-140 might be a key repressor of CRC development and metastasis

Figure 6

miR-140 Expression Is Reversely Correlated with CRC Development and Metastasis, Whereas Smad3 Expression Is Positively Correlated with CRC Development

(A) A cohort of CRC specimens, including 31 paired CRC tissues and adjacent colorectal mucosa, were selected, and the expression levels of miR-140 were measured by real-time qRT-PCR. MiR-140 was significantly downregulated in the primary CRC tissues compared to the adjacent normal mucosa. (B) Then, 13 out of 31 cases of CRC patients with lymph node or liver metastasis were selected, and RNA was extracted from the metastatic tumor tissues. miR-140 was significantly downregulated in the lymph node, and liver metastatic tumors were compared with the paired primary tumor tissues. The expression level of miR-140 was normalized by the internal control RNU6B in each sample. (C) Immunohistochemistry analysis of Smad3 expression was performed in the CRC cohort. Smad3 protein was overexpressed in the CRC tissues compared to the adjacent normal mucosa. Each experiment was repeated 3 times; data are presented as means ± SD. *p < 0.05. Scale bars, 50 μm.

Smad3 Is Associated with the Development of CRC

We also performed immunohistochemistry to assess the expression of Smad3 protein in the CRC cohort. Smad3 protein was located in the cytoplasm of colorectal glandular cells and CRC cells. Compared with the adjacent normal colorectal tissues, Smad3 protein was overexpressed in the CRC samples (Figure 6C, p < 0.05), indicating that Smad3 is related to the development of CRC. We noticed a tendency that miR-140 was downregulated, whereas Smad3 was upregulated, in the CRC tissues compared to the normal colorectal tissues, further implying the negative regulatory relationship between miR-140 and Smad3 in CRC.

Discussion

The regulation of EMT by miRNAs has recently been highlighted.2, 10, 11 In the present study, we revealed a molecular mechanism of CRC development, invasion, and metastasis mediated by miR-140 through interrupting the EMT process.

EMT is an indispensable process in the development, differentiation, and repair of multiple tissues and organs; however, inappropriate reactivation of EMT can promote carcinoma progression through a variety of mechanisms.2, 3, 10, 11, 35 The TGF-β signaling pathway, an essential driver of EMT, can stimulate tumor invasion and metastasis in a Smad3-dependent manner.36, 37 The knockdown of Smad3 in MDA-MB-231 cells can prolong the latency and delay the bone metastasis of breast cancer. Here, we confirmed that Samd3 was the direct target of miR-140 in the CRC cell lines via the luciferase assay and that the active p-Smad3 was reduced simultaneously (Figure 1). Enforced miR-140 expression inhibited the CRC cell migration and invasion, and the inhibitory effect was due to the suppression of Smad3 (Figure 2). It has been reported that the active phosphorylated Smad2/3 binds to Zeb1 to enhance the effect of the TGF-β signaling pathway on the EMT process.33, 34 Our results showed that miR-140 repressed the EMT process of CRC cells as a consequence of inhibition of Smad3, p-Smad3, and subsequent reduced Zeb1 expression (Figure 4), thus playing a critical role in the suppression of CRC cell migration and invasion. Consistent with this finding, a recent study found that downregulation of miR-140 induces EMT and promotes esophageal cancer cell invasion by targeting Slug. We further determined the impacts of miR-140 on the CRC metastasis by constructing a mouse metastatic model. Overexpression of miR-140 significantly inhibited the metastatic potential of CRC cells in BALB/c mice (Figure 5; Table 1). A recent study also reported that CRC stem cells with highly expressed miR-140 produce smaller tumor burden and fail to form the lung or liver metastasis, and this is mediated by the suppression of Smad2 and autophagy. Our results both in vitro and in vivo provide new evidence to support the involvement of miR-140 in the suppression of CRC invasion and metastasis via Smad3. We also confirmed the decrease of Smad2 protein by miR-140 in our study. Smad4 protein was decreased, possibly, due to the inhibition of Smad2 induced by the overexpression of miR-140. Taken together, miR-140 inhibits EMT, possibly, via directly targeting TGF-β signaling-pathway-related proteins Smad2 and Smad3 and via indirectly downregulating Smad4, resulting in the suppression of migration, invasion, and metastasis in the CRC. Except the aforementioned targets of miR-140, VEGF-A, ADAMTS5, and IGFBP5 have been confirmed to be involved in the inhibition of CRC invasion and metastasis induced by miR-140.28, 29 Thus, miR-140 inhibits CRC invasion and metastasis through regulating multiple mRNAs and might be a key suppressive regulator.

Moreover, we investigated the clinical relevance of miR-140 by comparing the expression level of miR-140 on a cohort of CRC specimens with and without metastasis using real-time qRT-PCR. We found that miR-140 was significantly downregulated in the primary CRC tissues as compared to the adjacent normal mucosa (Figure 6A). This is consistent with our previous study and a recent study.23, 26 Interestingly, we found that miR-140 was progressively downregulated in the lymph node and liver metastatic tumors as compared to the primary CRC tumors (Figure 6B). In accordance with our findings, Zhai et al. also showed the same trend of miR-140 expression in 18 archival CRC patient samples with metastasis. The clinical significance of miR-140 on the CRC samples further confirms the role of miR-140 in the CRC metastasis. We also examined the expression of Smad3 protein in the CRC cohort and found that Smad3 was significantly overexpressed in the CRC specimens compared to the adjacent normal colorectal tissues (Figure 6C). In line with our results, Korchynskyi et al. reported that Smad3 is upregulated in the CRC tissues, compared to the epithelial mucosa of normal colon, using immunohistochemistry. These findings suggest that Smad3 overexpression is correlated with the development of CRC.

In addition to the inhibitory effect of miR-140 on the CRC invasion and metastasis, we revealed a function of miR-140 in the growth of CRC in vitro and in vivo. Our in vitro experiments showed that miR-140 suppresses the cell proliferation and colony formation capacity of CRC cells via downregulation of Smad3 (Figure 3). It is well known that miRNAs exert their regulatory function on targeting multiple mRNAs. Previously, our group has reported that miR-140 inhibits CRC cell proliferation by the suppression of HDAC4. Zhai et al.’s study showed that the suppressive effect of miR-140 on CRC cell proliferation is partially due to the downregulation of Smad2. We further examined the function of miR-140 in CRC development and found that miR-140 overexpression remarkably reduces the tumor burden and that silenced Smad3 has a similar effect (Figures 5A and 5B). Taken together, our work reveals a novel regulatory mechanism of miR-140 in CRC growth, invasion, and metastasis.

Recently several studies have suggested that miR-140 is a tumor suppressor in other solid tumors, including HCC, NSCLC, and esophageal cancer through targeting some oncogenes.24, 25, 27 Judging from the combination of previous CRC studies and our present results, miR-140 might have the potential to be a therapeutic candidate for treating cancer.23, 26 Since the first miRNA, lin-4, was discovered in 1993, multiple miRNAs have been revealed as oncogenes or tumor suppressors in tumorigenesis and progression. The famous miR-34a has become the first miRNA to start the clinical trial, opening a novel era in cancer treatment. Compared to traditional gene-based therapy, miRNAs have the ability to regulate several cellular pathways simultaneously and make them suitable for the treatment of the multipathway-induced diseases such as cancer.

In conclusion, in this study, we provided the experimental evidence both in vitro and in vivo to support the suppressive effect of miR-140 on CRC growth, invasion, and metastasis. This is achieved largely due to the downregulation of Smad3 and the subsequent repression of the EMT process. The reduced expression of miR-140 in the primary and metastatic samples of CRC further supports the role of miR-140 in CRC progression and metastasis. As miRNA-based therapy is currently in clinical trial, our results indicate that miR-140 may be a novel candidate for treating CRC.

Materials and Methods

Cell Lines and Culture

The human CRC cell lines HCT116 and RKO were purchased from the Cell Bank of Type Culture Collection, Chinese Academy of Sciences, Shanghai, China. HCT116 cells were maintained in McCoy’s 5A media (GIBCO), and RKO cells were cultured in DMEM media (GIBCO) containing 10% fetal bovine serum (FBS; GIBCO) in a 5% CO2 container at 37°C.

Clinical CRC Specimens

Specimens from 31 CRC patients who underwent surgical resection in the First Affiliated Hospital of Dalian Medical University, Dalian, China, were enrolled in this study. The use of human tissue samples was in accordance with institutional ethics requirements and was approved by the Medical Ethics Committee, The First Affiliated Hospital of Dalian Medical University. Each patient sample contains normal colon mucosa, which were resected within at least 5 cm of the tumor margin. Eleven cases had lymph node metastasis, and 2 cases had liver metastasis. All patients provided written informed consent prior to participation in this study. The primary and metastatic CRC specimens were histologically examined by H&E staining.

miRNA and siRNA Transfections

HCT116 and RKO cells were plated in six-well plates at 2 × 105 per well and then transfected with 100 nM of either miR-140 mimics or negative miRNA (Invitrogen) after 24 hr by Oligofectamine (Invitrogen) according to the manufacturer’s protocols, respectively. Small interfering RNA (siRNA) against Smad3 (Invitrogen) was transfected by Oligofectamine at a final concentration of 100 nM in Opti-MEM I Reduced Serum Media (Life Technologies) according to the manufacturer’s instructions. The transfected cells were harvested for RNA isolation and protein extraction at 24 hr and 48 hr post-transfection, respectively.

miR-140 Knockdown

To knock down the endogenous miR-140, RKO and HCT116 cells were transfected with 100 nM scrambled miRNA inhibitor or miR-140 inhibitor (Invitrogen) using Lipofectamine 2000 (Invitrogen) in six-well plates (2 × 105 cells per well), respectively. Co-transfection of miR-140 inhibitor and siRNA against Smad3 in the CRC cells was used as a negative control.

Luciferase Assay

The wild-type or mutant double-stranded oligonucleotides of Smad3 3′UTR were synthesized and inserted into the sites of SacI and XhoI of the GP-miRGLO vector (GenePharma, Shanghai, China). The sequences are listed in Table S1.

HCT116 and RKO cells were plated in triplicate in 24-well plates. After 24 hr, cells in each well were co-transfected with 100 nM miR-140 mimic or negative miRNA and 200 ng GP-miRGLO constructs containing wild-type or mutant 3′UTR of Smad3 mRNA using Lipofectamine 2000. 20 ng pRL-TK Renilla plasmid (GenePharma, Shanghai, China) was used as the normalization. All the cells were harvested at 48 hr after transfection, and the luciferase activity was measured using the Dual-Luciferase Reporter Assay System (Promega). Firefly luciferase activity was normalized to Renilla internal control for each condition and then compared to the negative miRNA-transfected wells.

Scratch-Wound Assay

To evaluate the effect of miR-140 on CRC cell migration, a scratch-wound assay was performed in the HCT116 and RKO cells, miRNAs, miRNA inhibitors, or siRNA-transfected CRC cells. Upon reaching confluence, the monolayer cells were scraped with a 10-μL pipette tip to generate a scratch wound. The wounded surface was washed with 1× PBS to remove the cell debris and cultured in serum-free media. After 72 hr, the wound closures were photographed with a BZ-8100 microscope (Keyence, Osaka, Japan).

Migration and Invasion Assays

The in vitro migration and invasion assays were performed using a Transwell chamber (Corning Costar), according to the protocol of the manufacturer. The HCT116 and RKO cells and miRNAs, miRNA inhibitors, or siRNA-transfected CRC cells were trypsinized and used for the migration and invasion assays at 48 hr after transfection, respectively.

For the Transwell migration assay, the inserts were first rehydrated with serum-free media for 2 hr at room temperature. After rehydration, the media was carefully removed, and 200 μL of cell suspension with 3 × 105 cells in serum-free media was seeded into the upper chambers (24-well insert; pore size, 8 μm), respectively, while 600 μL of media supplemented with 10% FBS was added in the lower chambers. 24 hr after incubation, the non-migrating cells on the top side of the inserts were gently removed with a cotton swab. The inserts were then fixed by 4% methanol for 30 min and stained with 0.05% crystal violet for 15 min. The average migratory cells were counted by picking up 5 fields randomly each time under the microscope. The experiments were repeated 3 times.

For the Transwell invasion assay, the inserts were coated with a thin layer of Matrigel. Briefly, 60 μL Matrigel was diluted by precooled serum-free media in 1:4 ratios and was added to the middle of the bottom of the Transwell chamber and incubated for 1 hr at 37°C. The remaining protocol was the same as for the Transwell migration assay.

Immunofluorescence

The parental and transfected CRC cells were trypsinized, resuspended, and seeded on the coverslips; cultured for 48 hr; and then fixed with 4% formalin, permeabilized in 0.5% Triton X-100, blocked with 2% BSA, and thereafter incubated with mouse anti-E-cadherin monoclonal antibody (mAb) (1:50; #610181, BD Transduction Laboratories) or rabbit anti-vimentin mAb (1:50; ab92547, Abcam) overnight at 4°C, respectively. The cells were then incubated with rhodamine (tetramethylrhodamine [TRITC])-conjugated goat anti-mouse IgG ([immunoglobulin G] H+L) (1:50; Proteintech) or Alexa Fluor 488-conjugated goat anti-rabbit IgG (heavy chain and light chain [H+L]) (1:100, Proteintech). Nuclei were stained using DAPI (Sigma-Aldrich). The protein labeling was visualized using a laser scanning confocal microscope (Leica).

CCK-8 Assay

Cells were replated in 96-well plates at 1 × 103 per well after transfection with miRNA mimics, siRNA against Smad3, or miRNA inhibitors for 24 hr. At 1, 2, 3, 4, and 5 days after transfection, the absorbance at 450 nm was measured after incubation with 10 μL Cell Counting Assay Kit-8 solution (Sigma) for 2 hr. All experiments were performed three times independently.

Colony Formation Assay

Cells were trypsinized, counted, and seeded in the six-well plates in triplicate at 5 × 102 cells per well for the colony formation assay. During the colony growth, the culture medium was replaced every 2 days. After 14 days’ culture, the colonies were stained with 0.02% crystal violet for 1 hr and then were counted under an inverted phase-contrast microscope (Olympus IX71).

RNA Extraction

Total RNAs, including miRNAs, were isolated from the CRC cells with or without transfection, using TRIzol reagent (Invitrogen), according to the manufacturer’s instructions.

For human CRC formalin-fixed, paraffin-embedded (FFPE) specimens, approximately 0.005 g for each sample was extracted. These samples were deparaffinized with xylene, hydrated by ethanol, and digested with proteinase K. Total RNA was isolated using the High Pure RNA Paraffin Kit (Roche, Germany), according to the manufacturer’s instructions.

Real-Time qRT-PCR

For real-time qRT-PCR analysis of miRNAs, cDNA was synthesized using the TaqMan microRNA Reverse Transcription Kit (Life Technologies). Real-time qRT-PCR analysis was performed on an MX3000P real-time PCR system (Agilent). The primers for miR-140 and endogenous control RNU6B were purchased from Ambion. The gene expression CT values of miRNAs from each sample were calculated by normalizing to RNU6B, and relative quantitation values were plotted.

For real-time qRT-PCR analysis of mRNA expression, cDNA was synthesized using the PrimeScript RT Reagent Kit with gDNA Eraser (TaKaRa, Dalian, China). The qRT-PCR amplification for mRNAs was performed with SYBR Premix Ex TaqII (TaKaRa) on an Agilent MX3000P instrument. The gene expression CT values of mRNAs from each sample were calculated by normalizing to GAPDH, and relative quantitation values were plotted.

Protein Extraction and Western Blot

CRC cells with or without transfection were scraped and lysed in 1× RIPA buffer (Sigma), and the proteins were extracted. Equal amounts of protein extract (50 μg) were denatured with 10% SDS-PAGE, transferred to polyvinylidene fluoride (PVDF) membranes (Millipore), and blocked with 5% non-fat milk in TBS-Tween (TBST; 0.1 M, pH 7.4). Protein abundance of GAPDH (1:500, AP50811, Abgent) served as a control for protein loading.

Each sample was treated with the primary antibodies, including mouse anti-Smad3 mAb (1:1,000; sc-101154, Santa Cruz Biotechnology), rabbit anti-p-Smad3 (Ser423/425) mAb (1:500; #9520, Cell Signaling Technology), mouse anti-E-cadherin mAb (1:1,000; 610181, BD Transduction Laboratories), rabbit anti-vimentin mAb (1:1,000; ab92547, Abcam); rabbit anti-Smad2 mAb (1:1,000; 12570-1-AP, Proteintech), rabbit anti-Smad4 mAb (1:1,000; 10231-1-AP, Proteintech), and rabbit anti-Zeb1 mAb (1:1,000; 21544-1-AP, Proteintech) at 4°C overnight. Membranes were incubated with secondary antibody, horseradish-peroxidase (HRP)-conjugated rabbit/mouse anti-IgG (1:16,000, LI-COR Biosciences), for 2 hr at room temperature. Protein bands were detected by the Odyssey infrared imaging system (LI-COR Biosciences).

Immunohistochemistry

CRC and normal colorectal tissues were sectioned into 4-μm-thick slides and were incubated with rabbit anti-Smad3 mAb (1:100, 25494-1-AP, Proteintech) at 4°C in a humid container. A non-biotin horseradish peroxidase detection system and DAB substrate (Dako) were used to determine the staining. All samples were observed by two pathologists. Both the intensity and extent of immunostaining were considered for scoring. The positive expression of Smad3 protein was found mainly on the cytoplasm of cells, presenting as a brown granular material. The dominant staining intensity was scored as follows: 0 = negative, 1 = weak, 2 = moderate, and 3 = strong. The extent of the positive staining of cells was scored as follows: <25% = 1, 25%–50% = 2, 51%–75% = 3, and >75% = 4. The final score was determined by multiplying the intensity and the extent positivity scores, which yielded a range from 0 to 12. The mean score from each individual was calculated.

Lentivirus Transduction

Lentiviral vectors of miR-140 and negative control and shRNA against Smad3 were purchased from GeneChem (Shanghai, China). HCT116 cells (2 × 105) were plated in the six-well plates and cultured to about 70% confluence. A concentration of 2 μL was used to infect the cells for the subsequent in vivo experiments. The transduction efficiency was determined under a fluorescence microscope (Olympus IX73) after being infected for 3 days.

Mouse Subcutaneous Tumor Implantation and Colon Cancer Metastasis Model

Animal care was conducted according to the protocol and guidelines approved by the Animal Care and Use Committee of Dalian Medical University. Thirty-three 6- to 8-week-old BALB/c mice were purchased from the Experimental Animal Center of Dalian Medical University. Mice were randomly divided into three groups; group 1 was injected with lentivirus miR-140-transduced cells, group 2 was injected with lentivirus control-transduced cells, and group 3 was injected with lentivirus shRNA against Smad3-transduced cells. Each group was injected subcutaneously on the back or through the tail vein. The transduced HCT116 cells were collected and resuspended at 2 × 107/mL in McCoy’s 5A media with Matrigel (1:1). The mice were anesthetized by isoflurane inhalation.

For the subcutaneous tumor implantation model, 100 μL cell suspension was injected subcutaneously on the backs of mice (5 per group). The tumor mass was measured using a caliper, and the tumor volume (V) was calculated by the formula V = length × width2 × 0.5. For the CRC metastasis model, 100 μL cell suspension was injected through the tail vein (6 per group). The mice that were injected subcutaneously were sacrificed at 4 weeks, and those that were injected through the tail vein were sacrificed at 8 weeks. The subcutaneous tumors, livers, and lungs were excised and frozen at −80°C until processing for protein isolation. The frozen mouse tumor tissues were crushed by a sterile steel ball in a cryotube. Then, extraction and solubilization were performed. The portions of tumors were fixed in 10% neutral buffered formalin, and paraffin-embedded and 4-μm sections were stained for H&E. For immunohistochemistry, the consecutive sections were incubated with rabbit anti-Smad3 mAb (1:100; 25494-1-AP, Proteintech), rabbit anti-p-Smad3 mAb (1:100; ab52903, Abcam), rabbit anti-E-cadherin mAb (1:100; 20874-1-AP, Proteintech), rabbit anti-vimentin mAb (1:100; ab92547, Abcam), and rabbit anti-Zeb1 mAb (1:100; 21544-1-AP, Proteintech). The procedures and evaluation of immunostaining were briefly described earlier, in the Immunohistochemistry section.

Statistical Analysis

Statistical analysis was done using the SPSS software, v11.0. The results were expressed as the mean ± SD. Statistical differences were determined using a two-tailed Student’s t test between two groups (paired t test for matched clinical samples and unpaired t test for the unpaired two groups). For comparison of more than two groups, one-way ANOVA followed by a Bonferroni-Dunn test was performed. p < 0.05 was considered statistically significant.

Author Contributions

Conceptualization: B.S. and L.L.; Data Curation: J.L., K.Z., L.Y., and W.Z.; Formal Analysis: J.L., K.Z., L.Y., W.Z., Y.L., J.M., B.W., L.W., and S.F.; Investigation: J.L., K.Z., L.Y., and W.Z.; Methodology: J.L.; Project Administration: B.W.; Resources: J.L. and B.W.; Supervision: B.S. and L.L.; Validation: J.L.; Visualization: J.L. and K.Z.; Writing – Original Draft: B.S.; Writing – Review & Editing: J.L., B.S., and L.L.

Conflicts of Interest

The authors declare no conflict of interest.