Immunohistochemical staining was performed using the streptavidin-biotin peroxidase complex with a commercially available SP-kit (Beijing ZhongshanJinqiao Biotechnology Limited Company, China, SP-9000). A rabbit anti-human TFF1 monoclonal antibody (Abcam-Epitomics, United States, 2801-1 dilution: 1:100), a rabbit anti-human TFF3 monoclonal antibody (Abcam-Epitomics, United States, 2816-1 dilution: 1:200), and a mouse anti-human TWIST1 monoclonal antibody (Abcam, United States, ab135180 dilution: 1:100) were used. As a negative control, PBS was used in place of the primary antibody. After deparaffinization and rehydration, the sections were heated in a microwave oven for 10 min at 100 °C in 10 mmol/L citrate buffer (pH 6.0). The sections were then incubated sequentially with fresh 3% hydrogen peroxide in PBS, 10% normal goat serum, primary antibody, biotinylated goat anti-rabbit IgG, and streptavidin-peroxidase; the slides were washed with PBS three times before each step. The visualization of the sites of peroxidase binding was achieved with diaminobenzidine. The sections were counterstained with hematoxylin. All slides were interpreted by two independent observers in a blinded fashion.

The immunohistochemistry results were evaluated according to a histoscore that combines the intensity of the immunoreaction with the scope of the positive staining area. The intensity (I) of staining was scored as 0 (no staining), 1 (weak), 2 (moderate), or 3 (strong). The density (D) of staining was scored as 1 (less than 10%), 2 (10%-50%), 3 (50%-80%), or 4 (80%-100%) according to the percentage of positively stained regions in relation to the total cancer area. The final immunohistochemical staining scores (0-12) for each tumor were the product of I × D. Samples with I × D ≤ 4 were considered weakly positive; those with I × D ≥ 4 were considered strongly positive.

All cell culture media were purchased from HyClone, Thermo. Three human colon carcinoma cell lines with different invasive potentials (HT29, SW620, and LoVo) and an immortalized human epithelial cell line (HIEC) were gifts from the laboratory of Xi Jing, Digestive Disease Hospital (Fourth Military Medical University). All cell lines were maintained in RPMI 1640 growth medium supplemented with 10% heat-inactivated fetal bovine serum, 1% penicillin, and 1% streptomycin. HIEC cells were also supplemented with human insulin (0.1 U/mL).

The four colorectal cancer cell lines were harvested separately in ice-cold PBS. Total protein was extracted, separated by SDS-PAGE, and transferred onto PVDF membranes. The membranes were blocked with 5% (w/v) dried skimmed milk powder in Tris-buffered saline (blocking solution) for 1 h at room temperature. Then, the membranes were probed with primary antibodies against TFF1 (1/500 dilution), TFF3 (1/300 dilution), and TWIST1 (1/500) in blocking solution overnight at 4 °C. After a TBST wash, the membranes were incubated with the corresponding secondary antibody (1/2000 dilution) in blocking solution for 1 h at room temperature. Immunoreactive proteins were detected using enhanced chemiluminescence.

The mRNA levels of target genes in colon cancer cell lines were compared with those in the immortalized human normal intestinal epithelial cell line (HIEC). Quantitative real-time RT-PCR (qRT-PCR) was performed as previously described[14]; Brilliant SYBR Green QRT-PCR Master Mix as part of a 2-Step kit (Stratagene, La Jolla, CA, United States) was used. The PCR amplification was performed using a Bio-Red Real Time PCR system. The primers used for qRT-PCR were as follows: GAPDH forward (F): 5′-AGCCTTCTCCATGGTGGTGAA-3′, GAPDH reverse (R): 5′-ATCACCATCTTCCAGGAGCGA-3′; TFF1 (F): 5′-AATAAGGGCTGCTGTTTCG-3′, TFF1 (R): 5′-ACTCCTCTTCTGGAGGGAC-3′; TFF3 (F): 5′-CTGCTGCTTTGACTCCAGGAT-3′, TFF3 (R): 5′-CAGCTGGAGGTGCCTCAGAA-3′; TWIST1 (F): 5′-CATGTCCGCGTCCCACTAG-3′, TWIST1-(R): 5′-TGTCCATTTTCTCCTTCTCTGG-3′; E-cadherin (F): 5′-GAGTGCCAACTGGACCATTCAGTA-3′, E-cadherin (R): 5′-AGTCACCCACCTCTAAGGCCATC-3′; vimentin (F): 5′-CAGGCAAAGCAGGAGTCCAC-3′; vimentin (R): 5′-GCAGCTTCAACGGCAAAGTTC-3′; β-catenin (F): 5′-TGAGTGTCATGAAGTGCACAGGAG-3′, β-catenin (R): 5′-AACAGGCTGATGGTGCCAGAG-3′; snail (F): 5′-CGCGCTCTTTCCTCGTCA-3′, snail (R): 5′-TCCCAGATGAGCATTGGCAG-3′. The Ct (threshold cycle) values of the target gene amplifications were normalized to those of the GAPDH control. All reactions were performed in triplicate in a 25-μL reaction volume. The PCR amplification program consisted of 30 s of an initial denaturation at 95 °C followed by 40 cycles of PCR at 95 °C for 5 s, and then 60 °C for 30 s. Standard curves were drawn and the relative amount of target gene mRNA was normalized to that of GAPDH. Specificity was verified by melt curve analysis. The comparative CT method was used to calculate the relative quantification of gene expression.

The positive expression of TFF1 and TFF3 was found in the cytoplasm and cytomembrane (Figure 1A), while TWIST1 staining was observed in both the cytoplasm and nuclei of the cells (Figure 1A). TFF1 was mainly expressed in CRC cells with varying intensity and was only occasionally expressed or was negative in matched normal tissues (Figure 1B). TFF3 was distributed diffusely and was expressed as fine granules in most goblet cells in normal tissues, but its expression was decreased in the cytoplasm of CRC cells (Figure 1A and B). We found selective expression of TWIST1 in cancerous tissues but scarcely in their normal counterparts (Figure 1A and B).

Open in New Tab   Full Size Figure   Download Figure

Figure 1 TFF1, TFF3 and TWIST1 expression trends in colorectal cancer (A), adjacent normal tissue (B), metastatic lymph node (C), and liver metastasis (D). Compared with normal mucosal tissue, colorectal cancer tissues expressed significantly higher TFF1 and TWIST1 and lower TFF3. Compared to primary colorectal cancer tissue, there were no statistical differences in TFF3 and TWIST1 expression in metastatic lymph node or liver metastatic tissue. But disparity in TFF1 expression between metastatic lymph node and primary colorectal cancer had statistical significance.

Among the 50 tissues that were positive for TFF1, 19 demonstrated weak staining (38%), 23 demonstrated moderate staining (46%), and 8 demonstrated strong staining (16%) (Table 1). Among the 59 patients with TFF3-positive cancer, 34 showed weak staining (57.6%), 21 showed moderate staining (35.6%), and 4 showed strong staining (6.8%) (data not shown). Of the 41 patients with TWIST1-positive cancer, 18 showed weak staining (43.9%), 16 showed moderate staining (39%), and 7 showed strong staining (17.1%) (Table 1).

Results from the assessment of TFF1, TFF3 and TWIST1 expression in primary tumor tissues, metastatic lymph nodes, and liver metastatic tissues are presented in Table 1. It was shown that 66.7% (50/75), 78.7% (59/75), and 54.7% (41/75) of samples exhibited positive staining for TFF1, TFF3 and TWIST1, respectively. In addition, 27.3% (13/47), 100% (47/47), and 17% (8/47) of normal tissues adjacent to colorectal CRC tissues showed positive staining for TFF1, TFF3 and TWIST1, respectively. Compared with normal mucosal tissues, primary CRC tissues expressed significantly higher levels of TFF1 and TWIST1 but a lower level of TFF3 (Table 1). No significant difference was observed in TFF3 or TWIST1 expression between primary CRC tissues and metastatic lymph nodes or liver metastatic tissues (P = 0.879, P = 0.105 for TFF3 and P = 0.08, P = 0.780 for TWIST1). However, the disparity in the expression of TFF1 between metastatic lymph nodes and primary colorectal cancer was statistically significant (P = 0.01, P = 0.03) (Table 2). The correlation analysis between the expression of TFF1 and TFF3 proteins also showed a negative correlation in CRC (P = 0.007, r = -0.312; Table 3). We also found that the expression levels of TFF3 were stronger in mucinous and signet-ring cell cancers than in other cancer types, which might illustrate that the more mucus that is secreted, the stronger the TFF3 expression.

The results illustrating a correlation between TFF1, TFF3, and TWIST1 expression and clinicopathological variables are presented in Table 4. Higher expression of TFF3 and TWIST1 was significantly correlated with lymph node metastasis (P = 0.034, P = 0.000) and advanced stage (P = 0.031, P = 0.003). In contrast, TFF1 expression was correlated with differentiation (P = 0.029).

The survival analysis indicated that the 3- and 5-year survival rates of the 75 patients were 71% (55 patients) and 54% (43 patients), respectively. The overall survival of the patients with higher TFF3 or TWIST1 expression levels was significantly less than that of the patients with lower TFF3 or TWIST1 expression levels (P = 0.042 for the TFF3 group; P = 0.003 for the TWIST1 group; Figure 2A and B). TFF1 expression demonstrated no correlation with patient survival (P = 0.952; Figure 2D). In contrast, higher expression of TFF3 and TWIST1 was correlated with an even worse overall survival (Figure 2C). Furthermore, a univariate analysis showed that the survival rate had a close relationship with age, infiltration depth, lymph node metastasis, distant metastasis, TFF3 expression, TWIST1 expression, and TNM stage (P < 0.05). A multivariate analysis demonstrated that only age, tumor stage, and TFF3 expression were correlated with survival (Table 5).

Open in New Tab   Full Size Figure   Download Figure

Figure 2 Correlation between protein expression in colorectal cancer and overall survival. The survival analysis indicated that the overall survival of the patients with higher TFF3 or TWIST1 expression was significantly less than that of the patients with lower TFF3 or TWIST1 expression (A and B); Higher expression of the TFF3 and TWIST1 correlated with even worse overall survival (C); TFF1 expression had no correlation with the survival (D).

Based on the immunohistochemistry results, we determined gene expression in colon cancer cell lines with different metastatic potentials. Our results showed that the expression levels of TFF3 and TWIST1 in cancer cell lines were higher than those in the normal human intestinal epithelial cell line (P < 0.05). Moreover, the expression intensities were inclined to rise both at the protein and mRNA levels with the increase in metastatic potential. However, TFF1 expression demonstrated the opposite tendency (Figure 3).

Open in New Tab   Full Size Figure   Download Figure

Figure 3 The mRNA and protein expression of TFF1, TFF3, TWIST1, E-cadherin, vimentin, and β-catenin in colon cancer cell lines with different invasion potentials. The expression of TFF3 and TWIST1 in cancer cell lines was higher than that in normal human intestinal epithelial cell line, and the expression intensity was inclined to rise with the increase in metastatic potential both at protein and mRNA levels. But TFF1 expression had the opposite tendency. The TFF1 protein and mRNA expression decreased and TWIST1 and TFF3 expression increased gradually with the increase in metastatic potential of cell lines. E-cadherin and β-catenin expression tended to decrease, while vimentin and TWIST1 expression was inclined to rise with the increase in metastatic potential. ^aP < 0.05, ^bP < 0.01.

Based on the above-mentioned outcome, we determined whether TFF1 and TFF3 expression was related to EMT. We investigated E-cadherin, vimentin, β-catenin and Snail, which are classic markers and critical transcription factors of EMT in different cell lines. It was observed that the expression levels of TFF1 protein and mRNA were decreased and that the expression levels of TWIST1 and TFF3 increased gradually with the increase in the metastatic potential of the cell lines. It was also observed that E-cadherin and β-catenin expression tended to decrease, while that of vimentin, TWIST1 and Snail tended to increase with the increase in metastatic potential (Figure 3).

Despite that TFFs are primarily secreted by the epithelium of the gastrointestinal tract, their role in the progression of colorectal cancer (CRC) and their prognostic value have not been extensively expounded. TWIST1 is a key transcription factor that regulates epithelial-mesenchymal transition (EMT).

Recent evidence indicates a pivotal role of TFFs in the oncogenic transformation, growth and metastasis of human solid tumors.TFF1 and TFF3 are found to co-express notably in the tumors of the human mammary gland and co-regulate each other in a positive feedback loop. The studies also show that TFF3 continuously up-regulates the expression of TWIST1 in HT29 cell line, which hints the involvement of TFFs in the process of EMT.

Reports on TFFs and TWIST1 expressions in CRC and correlations with metastasis and survival are quite rare. This research findings suggest that The expressions of TFF3 and TWIST1 are associated with CRC patients’ survival after a curative resection and pivotal predictors of disease progression. TFF3 may be associated with the invasiveness of the CRC.

This research outcomes indicate that TFF3 and TWIST1 expression plays a vital role in the development of CRC and these two proteins might be important markers for disease progression and prognosis of CRC. The correlation of TFF3 with the EMT process may hint a potential target of CRC treatment.

EMT is a process by which epithelial cells lose their cell polarity and cell-cell adhesion, and gain migratory and invasive properties to become mesenchymal stem cells; these are multipotent stromal cells that can differentiate into a variety of cell types. EMT is essential for numerous developmental processes including mesoderm formation and neural tube formation. EMT has also been shown to occur in wound healing, in organ fibrosis, and in the initiation of metastasis for cancer progression.

The studies reported in this manuscript are descriptive/correlative in nature and are well organized. Each discussion presented is intriguing and has some guiding significance for clinical practice and future research.