MCPIP3 as a Potential Metastasis Suppressor Gene in Human Colorectal Cancer

Monocyte chemotactic protein induced protein 3 (MCPIP3) belongs to the Cys–Cys–Cys–His (CCCH)-zinc finger protein family and contains a highly conserved CCCH-zinc finger domain and a Nedd4-BP1 YacP nuclease (NYN) domain. Previous studies showed that MCPIP3 inhibits the expression of proinflammatory genes, such as vascular cell adhesion molecule (VCAM)-1, in human endothelial cells, but the roles and functions of MCPIP3 in cancer cells are still unknown. In human colorectal cancer specimens, we found that the messenger RNA expression of MCPIP3 was significantly downregulated in cancer tissues compared to adjacent normal tissues (18/25; average fold change of 8.18). Two cell models were used to demonstrate the anti-migration activity of MCPIP3. First, Tet-on T-REx-293/HA-MCPIP3 cells were used to examine whether MCPIP3 can change epithelial–mesenchymal transition (EMT)-related gene expressions. Second, we used two human colorectal cancer cell lines, SW620 and HCT116, to prove the role of MCPIP3 in regulating EMT-related gene expressions. We found that overexpression of MCPIP3 inhibited cell migration according to a wound-healing assay and Transwell invasion assay and vimentin expression, and increased E-cadherin expression in these two cell lines. These results suggest that MCPIP3 might play a negative role in cell migration of human colorectal cancer cells.


Introduction
Colorectal cancer (CRC) ranks third among all cancers in mortality. It is also the third most diagnosed malignant tumor worldwide [1]. The incidence of CRC is related to lifestyle, dietary habits, chronic bacterial infections, and chronic inflammation such as Crohn's disease. Known risk factors

Downregulation of MCPIP3 in Human CRC Tissues
In total, 25 pairs of T3 CRC tissues and adjacent normal tissues were collected from patients with CRC. After total RNA was extracted, real-time reverse transcription-polymerase chain reaction (RT-PCR) was employed to analyze the expression of MCPIP3 mRNA. As shown in Figure 1A, the relative level of MCPIP3 mRNA in 18/25 (72.0%) tumor tissues showed a decrease in the MCPIP3 mRNA level, compared with that of the adjacent normal tissues; specifically, the average expression in adjacent normal tissues was approximately 8.18-fold that in tumor tissues. In the remaining seven patients, MCPIP3 mRNA expression in tumor tissues was higher than that in peripheral normal tissues, rendering an average expression in tumor tissues 3.17-fold that in adjacent normal tissues. Data obtained from The Cancer Genome Atlas (TCGA) were also analyzed and found that levels of MCPIP3 mRNA were significantly decreased in majority of CRC samples as compared with normal tissue samples ( Figure 1C). The results suggest that MCPIP3 expression in CRC tumor tissues is lower than that in normal tissues. Thus, MCPIP3 may play a role in inhibiting tumor growth or metastasis.

Downregulation of MCPIP3 in Human CRC Tissues
In total, 25 pairs of T3 CRC tissues and adjacent normal tissues were collected from patients with CRC. After total RNA was extracted, real-time reverse transcription-polymerase chain reaction (RT-PCR) was employed to analyze the expression of MCPIP3 mRNA. As shown in Figure 1A, the relative level of MCPIP3 mRNA in 18/25 (72.0%) tumor tissues showed a decrease in the MCPIP3 mRNA level, compared with that of the adjacent normal tissues; specifically, the average expression in adjacent normal tissues was approximately 8.18-fold that in tumor tissues. In the remaining seven patients, MCPIP3 mRNA expression in tumor tissues was higher than that in peripheral normal tissues, rendering an average expression in tumor tissues 3.17-fold that in adjacent normal tissues. Data obtained from The Cancer Genome Atlas (TCGA) were also analyzed and found that levels of MCPIP3 mRNA were significantly decreased in majority of CRC samples as compared with normal tissue samples ( Figure 1C). The results suggest that MCPIP3 expression in CRC tumor tissues is lower than that in normal tissues. Thus, MCPIP3 may play a role in inhibiting tumor growth or metastasis. The clinical data and mRNAseq data of colorectal adenocarcinoma (COADREAD) were from Firebrowse (http://firebrowse.org/) data portal with TCGA data version 2016_01_28. Box plots of MCPIP3 expression in colorectal cancer using independent tumor (n = 373) and normal (n = 51) data. RSEM, RNA-seq by expectation maximization. ** p < 0.01.
In addition, we also analyzed the relationships between MCPIP3 mRNA expression level and patient survival as well as MCPIP3 mRNA expression level and tumor stage of human CRC using the database of The Human Protein Atlas (https://www.proteinatlas.org/). Interestingly, high levels of MCPIP3 mRNA were correlated with high survival rate of CRC patients ( Figure S1A). The more malignant tumors tended to express lower amounts of MCPIP3 at the mRNA level. However, the significant difference was found only between stage I and stage III as well as stage I and stage IV ( Figure S1B). The results suggest that MCPIP3 expression in CRC tumor tissues is lower than that in normal tissues and can be used to predict patient survival and might be involved in inhibition of tumor migration and invasion. Thus, MCPIP3 may play a role in inhibiting tumor growth or metastasis.

Overexpression of MCPIP3 Inhibits Cell Migration in T-REx-293/HA-MCPIP3 Cells
To examine the role of MCPIP3 in cells, a T-REx-293/HA-MCPIP3 cell line was established by transducing an HA-MCPIP3 gene into Tet-on T-REx-293 cells. HA-tagged MCPIP3 is expressed under the presence of the tetracycline derivative doxycycline (Dox). To determine whether MCPIP3 The clinical data and mRNAseq data of colorectal adenocarcinoma (COADREAD) were from Firebrowse (http://firebrowse.org/) data portal with TCGA data version 2016_01_28. Box plots of MCPIP3 expression in colorectal cancer using independent tumor (n = 373) and normal (n = 51) data. RSEM, RNA-seq by expectation maximization. ** p < 0.01.
In addition, we also analyzed the relationships between MCPIP3 mRNA expression level and patient survival as well as MCPIP3 mRNA expression level and tumor stage of human CRC using the database of The Human Protein Atlas (https://www.proteinatlas.org/). Interestingly, high levels of MCPIP3 mRNA were correlated with high survival rate of CRC patients ( Figure S1A). The more malignant tumors tended to express lower amounts of MCPIP3 at the mRNA level. However, the significant difference was found only between stage I and stage III as well as stage I and stage IV ( Figure S1B). The results suggest that MCPIP3 expression in CRC tumor tissues is lower than that in normal tissues and can be used to predict patient survival and might be involved in inhibition of tumor migration and invasion. Thus, MCPIP3 may play a role in inhibiting tumor growth or metastasis.

Overexpression of MCPIP3 Inhibits Cell Migration in T-REx-293/HA-MCPIP3 Cells
To examine the role of MCPIP3 in cells, a T-REx-293/HA-MCPIP3 cell line was established by transducing an HA-MCPIP3 gene into Tet-on T-REx-293 cells. HA-tagged MCPIP3 is expressed under the presence of the tetracycline derivative doxycycline (Dox). To determine whether MCPIP3 regulates cell growth, 1 µg/mL Dox was employed to treat T-REx-293/HA-MCPIP3 cells for 1, 2, 4, and 5 days. A 3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2-H-tetrazolium bromide (MTT) assay was used to measure viable cells, revealing that the viability of Dox-treated cells was slightly lower than that of control cells; however, the difference was nonsignificant ( Figure S2). The experimental results suggest that in T-REx-293/HA-MCPIP3 cells, overexpression of MCPIP3 did not influence cell survival.
A wound-healing assay was used to understand whether MCPIP3 expression affects cell migration. T-REx-293/HA-MCPIP3 cells were treated with Dox (1 µg/mL) for 16 h. A microscope was used to observe cell migration and count the amount of cell migration. Results showed that overexpression of MCPIP3 inhibited cell migration (Figure 2), verifying the capability of the protein to downregulate cell migration. regulates cell growth, 1 μg/mL Dox was employed to treat T-REx-293/HA-MCPIP3 cells for 1, 2, 4, and 5 days. A 3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2-H-tetrazolium bromide (MTT) assay was used to measure viable cells, revealing that the viability of Dox-treated cells was slightly lower than that of control cells; however, the difference was nonsignificant ( Figure S2). The experimental results suggest that in T-REx-293/HA-MCPIP3 cells, overexpression of MCPIP3 did not influence cell survival.
A wound-healing assay was used to understand whether MCPIP3 expression affects cell migration. T-REx-293/HA-MCPIP3 cells were treated with Dox (1 μg/mL) for 16 h. A microscope was used to observe cell migration and count the amount of cell migration. Results showed that overexpression of MCPIP3 inhibited cell migration (Figure 2), verifying the capability of the protein to downregulate cell migration. Cells were treated with 1 μg/mL of doxycycline (Dox) for 16 h, and the number of cells migrating was determined by a wound-healing assay (magnification ×200). A representative photo and quantified data are shown. Each value represents the mean ± SE from three independent experiments. * p < 0.05 vs. doxycycline-free cells.

Overexpression of MCPIP3 Affects EMT-Related Protein Expressions in T-REx-293/HA-MCPIP3 Cells
The EMT process is vital to tumor cell migration. To determine whether MCPIP3 changes EMTrelated molecule expressions to influence cell migration, we overexpressed MCPIP3 in T-REx-293/HA-MCPIP3 cells and analyzed expressions of crucial EMT-related molecules, such as Ecadherin (an epithelial marker protein) and vimentin (a mesenchymal marker protein). Figure 3A shows that expressions of the MCPIP3 and E-cadherin proteins increased with a prolonged Dox treatment duration. In contrast, expressions of the vimentin and VCAM-1 proteins declined with a prolonged Dox treatment duration. In a test of whether overexpression of MCPIP3 influences Ecadherin and vimentin mRNA expressions, as shown in Figure 3B, Dox-treated cells demonstrated increased E-cadherin mRNA expression and decreased vimentin mRNA expression. In addition, overexpression of MCPIP3 enhanced mRNA expressions of other epithelial markers, such as ZO-1, and repressed mRNA expressions of other mesenchymal markers, such as fibronectin. The mRNA expressions of other EMT-related transcription factors, such as ZEB1, Snail1, and Slug, were not affected by the overexpression of MCPIP3. Results suggest that MCPIP3 can inhibit cell migration by changing EMT-related gene expressions.

Overexpression of MCPIP3 Inhibits Cell Migration and Affects EMT-Related Gene Expressions in Human CRC Cells
To further confirm whether MCPIP3 regulates cancer cell migration and changes EMT-related molecule expressions in human CRC cells, we selected 10 human CRC cell lines, namely SW48, SW480, SW620, SW948, HCT8, HCT15, HT29, HCT116 p53 wild-type, HCT116 p53 −/− , and DLD1. E-Cadherin and vimentin protein expressions of these cells lines were tested, revealing that in all human CRC cell lines, MCPIP3 protein expression was low (Figure 4), among which the HT29 strain exhibited relatively higher expression of MCPIP3. Next, the SW620 and HCT116 cell lines were Cells were treated with 1 µg/mL of doxycycline (Dox) for 16 h, and the number of cells migrating was determined by a wound-healing assay (magnification ×200). A representative photo and quantified data are shown. Each value represents the mean ± SE from three independent experiments. * p < 0.05 vs. doxycycline-free cells.

Overexpression of MCPIP3 Affects EMT-Related Protein Expressions in T-REx-293/HA-MCPIP3 Cells
The EMT process is vital to tumor cell migration. To determine whether MCPIP3 changes EMT-related molecule expressions to influence cell migration, we overexpressed MCPIP3 in T-REx-293/HA-MCPIP3 cells and analyzed expressions of crucial EMT-related molecules, such as E-cadherin (an epithelial marker protein) and vimentin (a mesenchymal marker protein). Figure 3A shows that expressions of the MCPIP3 and E-cadherin proteins increased with a prolonged Dox treatment duration. In contrast, expressions of the vimentin and VCAM-1 proteins declined with a prolonged Dox treatment duration. In a test of whether overexpression of MCPIP3 influences E-cadherin and vimentin mRNA expressions, as shown in Figure 3B, Dox-treated cells demonstrated increased E-cadherin mRNA expression and decreased vimentin mRNA expression. In addition, overexpression of MCPIP3 enhanced mRNA expressions of other epithelial markers, such as ZO-1, and repressed mRNA expressions of other mesenchymal markers, such as fibronectin. The mRNA expressions of other EMT-related transcription factors, such as ZEB1, Snail1, and Slug, were not affected by the overexpression of MCPIP3. Results suggest that MCPIP3 can inhibit cell migration by changing EMT-related gene expressions. selected for subsequent research because SW620 has properties of mesenchymal cells, expressing high vimentin and low E-cadherin; in contrast, HCT116 exhibits epithelial cell characteristics, expressing high E-cadherin and low vimentin.  These two cell lines were transduced with a lentivirus (at a multiplicity of infection (MOI) of 3) carrying the pSIN/HA-MCPIP3 plasmid, and in control experiments, cells were transduced with a lentivirus carrying the pSIN/EGFP plasmid. After SW620 and HCT116 cells were transduced with the lentivirus for 1-3 days, an MTT assay was applied to analyze cell survival. Results revealed that overexpression of MCPIP3 slightly decreased cell survival, but this decline did not reach statistical significance ( Figure S3). In T-REx-293/HA-MCPIP3 cells, overexpression of MCPIP3 also demonstrated a similar trend ( Figure S2). If cells are cultivated with an increased MOI or prolonged

Overexpression of MCPIP3 Inhibits Cell Migration and Affects EMT-Related Gene Expressions in Human CRC Cells
To further confirm whether MCPIP3 regulates cancer cell migration and changes EMT-related molecule expressions in human CRC cells, we selected 10 human CRC cell lines, namely SW48, SW480, SW620, SW948, HCT8, HCT15, HT29, HCT116 p53 wild-type, HCT116 p53 −/− , and DLD1. E-Cadherin and vimentin protein expressions of these cells lines were tested, revealing that in all human CRC cell lines, MCPIP3 protein expression was low (Figure 4), among which the HT29 strain exhibited relatively higher expression of MCPIP3. Next, the SW620 and HCT116 cell lines were selected for subsequent research because SW620 has properties of mesenchymal cells, expressing high vimentin and low E-cadherin; in contrast, HCT116 exhibits epithelial cell characteristics, expressing high E-cadherin and low vimentin.
These two cell lines were transduced with a lentivirus (at a multiplicity of infection (MOI) of 3) carrying the pSIN/HA-MCPIP3 plasmid, and in control experiments, cells were transduced with a lentivirus carrying the pSIN/EGFP plasmid. After SW620 and HCT116 cells were transduced with the lentivirus for 1-3 days, an MTT assay was applied to analyze cell survival. Results revealed that overexpression of MCPIP3 slightly decreased cell survival, but this decline did not reach statistical significance ( Figure S3). In T-REx-293/HA-MCPIP3 cells, overexpression of MCPIP3 also demonstrated a similar trend ( Figure S2). If cells are cultivated with an increased MOI or prolonged treatment time, we cannot rule out the possibility that overexpression of MCPIP3 can inhibit cell growth and eventually cause cell apoptosis. Since we have found that overexpression of MCPIP3 for 5 days resulted in significant decrease of cell proliferation ( Figure S4). The mechanism for this, however, will require further clarification from a future in-depth study. Western blotting was adopted to determine that the lentivirus facilitated successful expression of MCPIP3 in the two cell lines ( Figure S3B). Subsequently, the lentiviral transduction mode was used to test whether overexpression of MCPIP3 inhibited human CRC cell migration and invasion. For both SW620 and HCT116 cells, the wound-healing assay showed that overexpression of MCPIP3 significantly inhibited cell migration ( Figure 5).  These two cell lines were transduced with a lentivirus (at a multiplicity of infection (MOI) of 3) carrying the pSIN/HA-MCPIP3 plasmid, and in control experiments, cells were transduced with a lentivirus carrying the pSIN/EGFP plasmid. After SW620 and HCT116 cells were transduced with the lentivirus for 1-3 days, an MTT assay was applied to analyze cell survival. Results revealed that overexpression of MCPIP3 slightly decreased cell survival, but this decline did not reach statistical significance ( Figure S3). In T-REx-293/HA-MCPIP3 cells, overexpression of MCPIP3 also demonstrated a similar trend ( Figure S2). If cells are cultivated with an increased MOI or prolonged treatment time, we cannot rule out the possibility that overexpression of MCPIP3 can inhibit cell growth and eventually cause cell apoptosis. Since we have found that overexpression of MCPIP3 for 5 days resulted in significant decrease of cell proliferation ( Figure S4). The mechanism for this, however, will require further clarification from a future in-depth study. Western blotting was adopted to determine that the lentivirus facilitated successful expression of MCPIP3 in the two cell lines ( Figure S3B). Subsequently, the lentiviral transduction mode was used to test whether overexpression of MCPIP3 inhibited human CRC cell migration and invasion. For both SW620 and HCT116 cells, the wound-healing assay showed that overexpression of MCPIP3 significantly inhibited cell migration ( Figure 5).
(A) (B) The NYN domain of MCPIP3 has RNase activity, but the D251N mutation in the NYN domain abolishes its RNase activity. To examine whether RNase activity of the NYN domain involves inhibiting cell invasion, we used a lentiviral expressing the HA-tagged MCPIP3 or MCPIP3-D251N mutant. As shown in Figure 6, the Transwell invasion assay revealed that both MCPIP3 and MCPIP3-D251N significantly inhibited cell invasion and loss of RNase activity decreased the inhibitory activity of MCPIP3 on cell invasion. The results suggest that the RNase activity of MCPIP3 plays a minor role in inhibition of cell invasion. Moreover, the overexpression of MCPIP3 and the MCPIP3-D251N mutant enhanced E-cadherin expression and inhibited vimentin expression as in T-REx-293/HA-MCPIP3 cells (Figure 7). However, there was no significant difference in change of Ecadherin and vimentin expression between MCPIP3-overexpressed and MCPIP3-D251Noverexpressed cells. These results verified that in human CRC cells, MCPIP3 can change EMT-related gene expressions to inhibit cancer cell migration and invasion, and the RNase activity of MCPIP3 might be not involved in the regulation of E-cadherin and vimentin expression. In addition, the Ecadherin can bind with β-catenin and form complexes, and the aberrant expression of E-cadherin/βcatenin complexes is correlated with the EMT processes [21]. Overexpression of MCPIP3 did not change the expression of β-catenin ( Figure S5), suggesting β-catenin is not involved in the EMT processes regulated by MCPIP3. The NYN domain of MCPIP3 has RNase activity, but the D251N mutation in the NYN domain abolishes its RNase activity. To examine whether RNase activity of the NYN domain involves inhibiting cell invasion, we used a lentiviral expressing the HA-tagged MCPIP3 or MCPIP3-D251N mutant. As shown in Figure 6, the Transwell invasion assay revealed that both MCPIP3 and MCPIP3-D251N significantly inhibited cell invasion and loss of RNase activity decreased the inhibitory activity of MCPIP3 on cell invasion. The results suggest that the RNase activity of MCPIP3 plays a minor role in inhibition of cell invasion. Moreover, the overexpression of MCPIP3 and the MCPIP3-D251N mutant enhanced E-cadherin expression and inhibited vimentin expression as in T-REx-293/HA-MCPIP3 cells (Figure 7). However, there was no significant difference in change of E-cadherin and vimentin expression between MCPIP3-overexpressed and MCPIP3-D251N-overexpressed cells. These results verified that in human CRC cells, MCPIP3 can change EMT-related gene expressions to inhibit cancer cell migration and invasion, and the RNase activity of MCPIP3 might be not involved in the regulation of E-cadherin and vimentin expression. In addition, the E-cadherin can bind with β-catenin and form complexes, and the aberrant expression of E-cadherin/β-catenin complexes is correlated with the EMT processes [21]. Overexpression of MCPIP3 did not change the expression of β-catenin ( Figure S5), suggesting β-catenin is not involved in the EMT processes regulated by MCPIP3.

Discussion
This study analyzed 25 pairs of human CRC tissue specimens. According to the TNM cancer staging system (T means "the extent of the tumor"; N means "the extent of spread to the lymph nodes"; M means "the presence of metastasis"), these carcinoma tissues were T3 primary tumors, which entails tumor invasion penetrating the muscularis propria to the mucosa. A real-time RT-PCR was applied to analyze tumor tissues and adjacent normal tissues for MCPIP3 mRNA expression, revealing that most tumor tissues expressed a lower level of MCPIP3 than adjacent normal tissues. Thus, MCPIP3 possibly plays a crucial role in CRC cell growth and/or metastasis. In T-REx-293/HA-MCPIP3, SW620, and HCT116 cells, overexpression of MCPIP3 did not influence cell proliferation but did inhibit cell migration and invasion. Furthermore, overexpression of MCPIP3 changed EMTrelated gene expressions, including increasing E-cadherin expression and decreasing vimentin expression. Therefore, the results suggest that MCPIP3 plays a crucial role in downregulating human CRC cell migration.
MCPIP family members contain both a CCCH zinc finger domain (relevant to DNA/RNA binding) and an NYN nuclease domain (which has RNase activity). Previous studies focused on

Discussion
This study analyzed 25 pairs of human CRC tissue specimens. According to the TNM cancer staging system (T means "the extent of the tumor"; N means "the extent of spread to the lymph nodes"; M means "the presence of metastasis"), these carcinoma tissues were T3 primary tumors, which entails tumor invasion penetrating the muscularis propria to the mucosa. A real-time RT-PCR was applied to analyze tumor tissues and adjacent normal tissues for MCPIP3 mRNA expression, revealing that most tumor tissues expressed a lower level of MCPIP3 than adjacent normal tissues. Thus, MCPIP3 possibly plays a crucial role in CRC cell growth and/or metastasis. In T-REx-293/HA-MCPIP3, SW620, and HCT116 cells, overexpression of MCPIP3 did not influence cell proliferation but did inhibit cell migration and invasion. Furthermore, overexpression of MCPIP3 changed EMTrelated gene expressions, including increasing E-cadherin expression and decreasing vimentin expression. Therefore, the results suggest that MCPIP3 plays a crucial role in downregulating human CRC cell migration.
MCPIP family members contain both a CCCH zinc finger domain (relevant to DNA/RNA binding) and an NYN nuclease domain (which has RNase activity). Previous studies focused on

Discussion
This study analyzed 25 pairs of human CRC tissue specimens. According to the TNM cancer staging system (T means "the extent of the tumor"; N means "the extent of spread to the lymph nodes"; M means "the presence of metastasis"), these carcinoma tissues were T3 primary tumors, which entails tumor invasion penetrating the muscularis propria to the mucosa. A real-time RT-PCR was applied to analyze tumor tissues and adjacent normal tissues for MCPIP3 mRNA expression, revealing that most tumor tissues expressed a lower level of MCPIP3 than adjacent normal tissues. Thus, MCPIP3 possibly plays a crucial role in CRC cell growth and/or metastasis. In T-REx-293/HA-MCPIP3, SW620, and HCT116 cells, overexpression of MCPIP3 did not influence cell proliferation but did inhibit cell migration and invasion. Furthermore, overexpression of MCPIP3 changed EMT-related gene expressions, including increasing E-cadherin expression and decreasing vimentin expression. Therefore, the results suggest that MCPIP3 plays a crucial role in downregulating human CRC cell migration.
MCPIP family members contain both a CCCH zinc finger domain (relevant to DNA/RNA binding) and an NYN nuclease domain (which has RNase activity). Previous studies focused on MCPIP1 and MCPIP4. The main function of MCPIP1 is as an antiviral, and its feedback downregulates inflammation. MCPIP4 was speculated to be a tumor-suppressor gene, because its mutation was identified in certain tumors. Previous studies indicated that the CCCH zinc finger domain of MCPIP1 can bind to AU-rich elements (AREs) of mRNA (e.g., TNF-α and GM-CSF), thereby causing these mRNAs to become unstable and degrade [10]. However, it was demonstrated that MCPIP1 can degrade the proinflammatory molecule, IL-1β, through an ARE-independent pathway [22]. The current study demonstrated that MCPIP3 can enhance E-cadherin mRNA expression and inhibit vimentin mRNA expression, and the underlying molecular mechanism seems to be different between regulation of E-cadherin and vimentin by MCPIP3.
Regarding to the downregulation of vimentin by MCPIP3, we noted that the AREs are not present within the 3 UTR of vimentin mRNA [23]. In addition, the MCPIP3-D251N mutant did not lose its inhibitory activity on the vimentin protein expression (Figure 7). Therefore, we think the downregulation of vimentin expression by MCPIP3 might not be associated with the RNase activity of MCPIP3. It has been known that activator protein 1 (AP-1) consensus element is a stronger enhancer located in human vimentin gene promoter, and 12-O-tetradecanoyl-phorbol-13-acetate (TPA) and growth factors can induce vimentin gene expression by activation of AP-1 [24,25]. Previous study also found that MCPIP1 functions as a negative regulator of AP-1 signaling pathways. It is possible that MCPIP3 negatively regulates AP-1 activity as MCPIP1 does, and then down-regulates vimentin expression [26].
On the other hand, the upregulation of E-cadherin by MCPIP3 might be at the transcription level. Many transcription factors that can directly or indirectly inhibit E-cadherin are regarded as EMT-related transcription factors, such as Snail1, Slug (Snail2), ZEB1, and ZEB2. These transcription factors can bind to the promoter of E-cadherin and inhibit E-cadherin gene expression [6]. However, results of the present study show that overexpression of MCPIP3 did not influence mRNA expressions of Snail1, Slug (Snail2), or ZEB1 ( Figure 3B); thus, we think that MCPIP3 does not change E-cadherin expression by changing expressions of these transcription factors. Nonetheless, we cannot exclude the possibility that MCPIP3 may change the transcription activities of these factors, or directly or indirectly binds to a gene promoter and then influence EMT-related gene expressions. Moreover, matrix metalloproteinases (MMPs) of basement membranes, of which MMP-2 and MMP-9 are considered the most essential in cell metastasis [27], can hydrolyze collagen and the extracellular matrix. The present study found that overexpression of MCPIP3 inhibited MMP-2 mRNA expression ( Figure 3B), indicating that MCPIP3 may inhibit MMP gene expression and thus tumor cell metastasis. Previous studies indicated that VCAM-1 expression influences EMT-related gene expressions [28]. High expression of VCAM-1 reduces E-cadherin expression and elevates vimentin expression, thereby enhancing cell adhesion, migration, and invasion. Both a previous study [20] and the current study identified that overexpression of MCPIP3 can effectively inhibit VCAM-1 expression ( Figure 3). Accordingly, overexpression of MCPIP3 can be mediated through down-regulation of VCAM-1 molecules, which in turn helps maintains epithelial phenotype, thereby weakening cell migration and invasion.
Analyzing CRC tissues of clinical patients showed that mRNA expression of MCPIP3 in most CRC tissues was low (Figure 1), but it increased in a small number of CRC tissues. Previous studies indicated that the MCPIP family can negatively regulate macrophage activation [7,8]. Macrophages in later-period inflammation induce MCPIP gene expression, which forms a further feedback loop to inhibit macrophage inflammation. Tumor tissues generally demonstrate infiltration of inflammatory macrophages; thus, the inflammation status involving macrophages may influence MCPIP expression. In this study, a great number of macrophages may have existed in some of tumor tissues, and therefore MCPIP3 expression may have been influenced by these macrophages.
were calculated using the comparative Ct method (2dCt formula) and were normalized to the GAPDH mRNA level [32].

Wound-Healing Assay
Cells were seeded in reservoirs of a Culture-Insert (ibidi GmbH, Martinsried, Germany) which was stuck to the surface of a 12-well cell culture plate. The Culture-Insert was then removed, and cells were continuously cultured for 24 h. Four low-magnification areas (100×) were randomly captured by an inverted microscope (Nikon ECLIPSE TE2000-U, Kanagawa, Japan), and the number of migrated cells was visually counted [34].

Invasion Assay
Cells were seeded in the upper chamber of a Transwell (Corning Costar, Corning, NY, USA), which was coated with Matrigel ® matrix (Corning Life Sciences, Taipei, Taiwan), and medium with 10% fetal bovine serum (FBS) was added to the lower chamber. After 24 h, cells in the top of the filter were removed with a cotton swab, and cells in the bottom of the filter were fixed with 34% paraformaldehyde and stained with 0.1% crystal violet dye. Four low-magnification areas (100×) were randomly captured with an inverted microscope (Nikon ECLIPSE TE2000-U), and the number of invaded cells was counted [34].

Statistical Analysis
Statistical analyses were performed using one-way Student's t-test by SigmaPlot 13.0 (Systat Software Inc., San Jose, CA, USA), and differences were considered significant at p < 0.05.