The present study was approved by the Ethics Committee of the First Affiliated Hospital of Guangxi Medical University (Nanning, China) with the approval no. BBMCEC2012063 and was conducted in accordance with the guidelines of the Helsinki Declaration. The patient cohort of 98 colorectal cancer tissues and 15 normal tissues were collected from September 2013 and September 2014 at The First Affiliated Hospital of Guangxi Medical University where colorectal cancer was histologically diagnosed and staged according to the National Comprehensive Cancer Network (NCCN) colorectal cancer classifications (18). These 98 patients included 60 (61.2%) males and 38 (38.8%) females and their ages ranged from 25 to 83 years with a mean age of 56 years. There were 36 tumors located in the left-sided colon and 62 in the right-side colorectum. Four tumors were mucinous adenocarcinoma and 94 were tubular adenocarcinoma and there were 10 patients at stage I, 35 at stage II, 26 at stage III, and 27 at stage IV of the disease.

For IHC the sections (4–5 µm) were deparaffinized in xylene and rehydrated in a series of ethanol solutions and then immunostained following the procedures of a previous study (19) with a primary antibody against human tricellulin (cat. no. SAB1306444; Sigma-Aldrich; Merck KGaA) at a dilution of 1:100. The immunoreactivity score (IRS) system as described in a previous study (20) was used to semi-quantify the immunostaining data, i.e., the staining intensity (SI) × the percentage of positive cells.

Human colorectal adenocarcinoma cell lines HT29, HCT116, RKO, SW620, Caco2 and HCT-8 and a normal human colon mucosal epithelial cell line NCM460 were obtained from the Cell Resource Center, Shanghai Institute of Biochemistry and Cell Biology at the Chinese Academy of Sciences (Shanghai, China). The cell lines had been validated by performing short tandem repeat (STR) profiling. The cells were cultured in RPMI-1640 medium (Corning, Inc.) supplemented with 10% FBS (Biological Industries) in a humidified incubator containing 5% CO₂ at 37°C.

A tricellulin shRNA (targeting sequences of 5′-GATGAGCAGATTGCCACATCA-3′) was cloned into a pcDNA6. 3-EGFP vector (Invitrogen; Thermo Fisher Scientific, Inc.). The vector or negative control vector (lenti-EGFP-miR) was used to generate lentiviruses in 293 cells, and the lentiviruses were used to infect HCT116 cells and named as tricellulin knocked down cells (TRIC-KD) or negative control cells (NC). In addition, lentiviral vector carrying human tricellulin cDNA (lenti-tricellulin-IRES-EGFP) or vector-only was used to infect HCT116 cells for tricellulin-overexpression (TRIC-OE) in cells. The efficiency of tricellulin knockdown or overexpression was confirmed by RT-qPCR and western blotting. To determine the effects of tricellulin on regulation of the canonical NF-κΒ pathway, we treated these stable cells or parental cells (CON) with 50 ng/ml TNF-α (R&D Systems) or 100 µM pyrrolidinecarbodithioic acid (PDTC; Sigma-Aldrich; Merck KGaA) for 24 h for subsequent experiments.

Fresh colorectal cancer tissues were cryosectioned and fixed in 4% paraformaldehyde and then immunostained following the procedures of a previous study (21) with a primary antibody against tricellulin (product no. 48-8400; Invitrogen; Thermo Fisher Scientific, Inc.) at a dilution of 1:200 at 4°C overnight. Subsequently, the sections were washed and incubated with the respective secondary antibody (Alexa Fluor^® 594 goat anti-rabbit; product no. 8889S; 1:1,000; Cell Signaling Technology, Inc.) and 4′,6-diamidino-2-phenylindole (DAPI) (Wuhan Boster Biological Technology, Ltd.).

To visualize the cell actin cytoskeleton, FITC-Phalloidin was utilized to stain HCT116 cells (NC, TRIC-KD, and TRIC-OE). In brief, cells (1×10⁴) were seeded onto coverslips and grown overnight and then fixed and permeabilized in the same way as the cryosections. The cells were next incubated with the Alexa Fluor^® 488 Phalloidin working solution (Beijing Solarbio Science & Technology, Inc.) for 30 min at room temperature in the dark and then with DAPI for 10 min. Images were obtained from the stained cells with a fluorescence Olympus BX53F microscope (magnification, ×600; Olympus Corporation).

Total cellular RNA was isolated from cells using the Eastep™ Super Total RNA Extraction Kit and reversely transcribed into cDNA using a Reverse Transcription Kit (both from Promega Corporation). qPCR was performed using the Power SYBR Green Master Mix (Promega Corporation) on an 7500 ABI Real-time PCR system (Applied Biosystems; Thermo Fisher Scientific, Inc.). The primers used were tricellulin forward, 5′-CCAGCTATAGCGCCAGATCTCAA-3′ and reverse, 5′-CAGACACCGGCTTATCCCATTC-3′; and GAPDH forward, 5′-GCACCGCAAGGCTGAGAAC-3′ and reverse, 5′-TGGTGAAGACGCCAGTGGA-3′, which were all obtained from Takara Bio, Inc.. The cycling parameters were as follows: Denaturing at 95°C for 10 min, 40 cycles of denaturing at 95°C for 15 sec, primer annealing at 60°C for 1 min and extension temperature at 95°C for 15 sec; final extension at 60°C for 15 sec and final denaturing at 95°C for 15 sec. Results were quantified by using the 2^−ΔΔCq method (22).

Cells were lysed using the RIPA lysis buffer (Beyotime Institute of Biotechnology) and quantified by using the bicinchoninic acid (BCA) protein assay kit (Beyotime Institute of Biotechnology). Equal amounts (30 µg/lane) of protein samples were separated in 10 or 12% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) gels and transferred onto polyvinylidene fluoride membranes (PVDF; EMD Millipore).

Western blotting was performed according to our previous study (23); after being blocked in a 5% skim milk for 1 h at the room temperature, the membranes were incubated with a primary antibody against GAPDH (cat. no. 60004-1-Ig; 1:3,000), MMP2 (cat. no. 10373-2-AP; 1:800), vimentin (cat. no. 10366-1-AP; 1:500), Snail (cat. no. 13099-1-AP; 1:500), caspase-3 (product no. 19677-1-AP; 1:800) (all from ProteinTech Group, Inc.), tricellulin (cat. no. 48-8400; 1:800; Invitrogen; Thermo Fisher Scientific, Inc.), p65 (product no. 8242S; 1:1,000), p-p65 Ser536 (product no. 3033; 1:1,000), E-cadherin (product no. 3195; 1:1,000), survivin (product no. 2808S; 1:1,000), or Bcl-2 (product no. 15071S; 1:1,000) (all from Cell Signaling Technology, Inc.) overnight at 4°C. Subsequently, the membranes were incubated with a secondary antibody (1:10,000; LI-COR IRDye^® 680RD Goat anti-Rabbit P/N 926-68071 or LI-COR IRDye^® 680RD Goat anti-Mouse P/N 926-68070; LI-COR Biosciences) at room temperature for 1 h. The protein bands were quantified by using Odyssey infrared imaging (LI-COR Biosciences). The final data were semi-quantification was carried out using ImageJ software (version 1.50i; National Institutes of Health).

The cells (1×10⁵) were added to the upper chamber of a Transwell (8-µm pores) with noncoated membranes (migration assay) or Matrigel-coated Transwell chambers (invasion assay) with serum-free medium, while a cell culture medium containing 20% FBS was added to the bottom chambers. After 24 or 48 h culture, the migrated or invasive cells at the bottom of the membrane were fixed in 4% methanol for 20 min at room temperature, and stained using 0.1% crystal for 10 min at room temperature, and counted from five-randomly selected fields under the 100X objective of an inverted microscope (Nikon Corporation).

Cells were seeded into a 24-well flat bottom plate at a density of 1×10⁵ cells/well and grown for 24 h. Next, the cells were washed with phosphate-buffered saline (PBS), fixed in 4% methanol, and incubated with Hoechst 33258 stain solution (Beijing Solarbio Science & Technology, Inc.) for 10 min at room temperature. The cell apoptotic morphology was observed under an Olympus BX53F fluorescence microscope (magnification, ×600).

TMT-LC-MS/MS was performed by R&S Biotechnology Co., Ltd.. In brief, total protein was extracted from cells and processed in a 6-plex TMT labeling kit (Thermo Fisher Scientific, Inc.) according to a procedure of a previous study (24). Fractionation was then carried out by basic pH reverse-phase liquid chromatography with fraction combining according to a previous study (25). The separated samples were loaded onto the Nano-Aquity UPLC system (Waters Corporation) and the resulting peptide mixture was loaded onto a trap column (2.1×150 mm X Bridge BEH300; Waters Corporation) attached to an analytical column (ZORBAX 300SB-C18 column, 5 µm, 300 Å, 0.1×150 mm; Microm for mass-spectrometer analysis (26). The linear gradient was from 2% D to 80% D in 90 min (solution D: 0.1% formic acid in ACN) and the triple TOF 5600 MS, switched between MS and serial (MS/MS) acquisition. The MS data were acquired using a spectral accumulation time of 250 msec in the mass range of 350-1,300 m/z. The scanning tandem mass spectrometry was from 100–1250 m/z with rolling collision energy. The 20 strongest precursors were selected according to the procedure of a previous study (26).

The data were then analyzed using the free online platform of Majorbio I-Sanger Cloud Platform (www.i-sanger.com). The ratio cutoff values were set to 1.2 and 0.8 for fold-change and more than two unique peptides were used for protein identification. A P-value <0.05 was considered as the threshold to identify significant changes. We then performed Gene Ontology (GO) (27) and used the Kyoto Encyclopedia of Genes and Genomes (KEGG) (28) database to identify the functions or interactions between these differentially expressed proteins (DEPs) in colorectal cancer cells.

A PPI network was constructed using a Search Tool [STRING (29); http://string.embl.de/]. A confidence score of 0.4 was set as the cut-off criterion and the DEPs with connection numbers 10 were considered as hub genes. The top 10 hub nodes with higher degrees were screened using R (version 3.6.0, http://cran.r-project.org/).

All data were presented as the means ± SD and statistically analyzed using the SPSS v17.0 software package (IBM Corp.) and GraphPad Prism v7.0 (GraphPad Software, Inc.). The association between tricellulin expression and clinicopathological characteristics of colorectal cancer patients was analyzed using the Chi-squared (χ²) test, while the survival rates stratified by tricellulin expression were calculated by Kaplan-Meier curves and the log-rank test. Unpaired Student's t-tests and one-way ANOVA tests were used to determine statistical significance between two groups or among more than two groups, respectively. After a one-way ANOVA test was performed, for post hoc evaluation, Tukey's HSD test was used. A P-value <0.05 was considered to indicate a statistically significant difference.

Tricellulin expression was analyzed in 98 cases of colorectal cancer tissues vs. 15 normal mucosae and the results revealed that tricellulin expression was significantly upregulated in colorectal cancer tissues compared to normal mucosae (Fig. 1A). Tricellulin was mainly expressed in the cytoplasm or nucleus in colorectal cancer tissues, which was quite different from that of normal mucosae in the cytoplasm (Fig. 1B). Of these 98 cancer cases, 51 (52.0%) had tricellulin overexpression vs. 2 (13.3%) with tricellulin expression among the 15 normal mucosae (Fig. 1C).

Then, the association of tricellulin expression with clinicopathological characteristics of colorectal cancer patients was assessed and it was revealed that tricellulin expression (52.0%; 51/98) was associated with tumor distant metastasis (M stage) and advanced tumor-lymph node metastasis (TNM stage) (Table I), but not with the sex and age, tumor size, and N stage of patients. The Kaplan-Meier curves and the long rank test revealed that the upregulated tricellulin expression was associated with shorter overall survival (OS; P=0.012; Fig. 1D).

To assess the effects of tricellulin knockdown and overexpression on regulation of colorectal cancer cell malignant phenotypes in vitro, tricellulin expression was first analyzed in human colorectal adenocarcinoma cell lines HT-29, HCT116, RKO, SW620, Caco2, and HCT8 as well as a normal human colon mucosal epithelial cell line, NCM460. It was revealed that the level of tricellulin protein was higher in HT29, HCT116, and SW620 cells than in other all cancerous and normal cells (Fig. 2A). We therefore selected HCT116 with moderate endogenous expression of tricellulin for stable tricellulin knockdown or overexpression. Tricellulin knockdown (TRIC-KD) or overexpression (TRIC-OE) was confirmed by RT-qPCR and western blotting (Fig. 2B-E).

Next, it was determined that TRIC-OE significantly increased HCT116 cell migration and invasion capacities, whereas TRIC-KD reduced tumor cell migration and invasion (Fig. 3A-C). The FITC-Phalloidin staining revealed that tricellulin regulated pseudopodium formation and cytoskeletal reorganization; downregulation of tricellulin expression notably inhibited the length and number of pseudopodia and stress fibers. In contrast, upregulation of tricellulin expression increased the length and number of pseudopodia and stress fibers (Fig. 3D).

To elucidate the potential molecular mechanisms underlying tricellulin, quantitative proteomic analysis was performed to profile DEPs in the TRIC-OE group vs. the NC group. We identified a total of 6,957 unique proteins in the TRIC-OE group and 193 were DEPs vs. the NC group, in which 63 were upregulated and 130 were downregulated. The DEPs volcano plots are presented in Fig. 4A.

To determine their biological and functional properties in the TRIC-OE compared to the NC group, GO was performed and the KEGG database was used. The GO data revealed that tricellulin could regulate the biological process (BP), molecular function (MF), and cellular component (CC). These DEPs were significantly enriched in regulation of the ventricular cardiac muscle cell action potential, peptidyl-cysteine S-nitrosylation, lateral sprouting from an epithelium, epithelial-mesenchymal transition (EMT), and phosphatidylinositol 3-kinase signaling (Fig. 4B). The KEGG analysis revealed that DEPs were mainly enriched in nicotine addiction, basal cell carcinoma, Wnt signaling pathway, circadian entrainment, and pathways in cancer (Fig. 4C).

The PPI network data with the confidence score set at 0.4 revealed that the hub genes with higher degrees included UBE2D1, DTL, TIMP1, CREB1, CYCS, PIK3CA, CENPH, COL1A1, COMMD1, and PTTG1 (Fig. 4D).

EMT, a key developmental regulatory program, plays a critical role in the promotion of cancer progression and metastasis in human carcinomas (30). In our previous study, we demonstrated that the NF-κB signaling was closely associated with colorectal cancer cell EMT (31). Thus, we further determined the role of the NF-κB pathway in mediating the tricellulin oncogenic activities in colorectal cancer cells. The expression of p65, p-p65, mesenchymal markers (N-cadherin, Snail and vimentin), and an epithelial marker (E-cadherin) was analyzed in HCT116 cells after tricellulin overexpression of knockdown. The data revealed that tricellulin overexpression upregulated the levels of vimentin and Snail but downregulated E-cadherin, which was not completely inhibited by an NF-κB inhibitor PDTC (Fig. 5A). In contrast, knockdown of tricellulin expression decreased vimentin and Snail, but increased E-cadherin, which was also not completely activated by an NF-κB activator TNF-α (Fig. 5B). These results indicated that tricellulin may regulate the EMT in colorectal cancer cell through the canonical NF-κB signaling pathways.

Hoechst 33258 staining and western blotting were first performed to further assess the role of tricellulin in apoptosis. The present data revealed that the apoptotic cells had significant nuclear condensation and morphological changes in TRIC-KD cells compared to the negative control group (NC). Conversely, tricellulin overexpression (TRIC-OE) had normal nuclear morphology (Fig. 6A). Assessment of the levels of apoptosis-related proteins revealed that tricellulin overexpression reduced caspase-3 and increased Bcl-2 and survivin levels, whereas knockdown of tricellulin expression reduced Bcl-2 and survivin levels and enhanced caspase-3 levels (Fig. 6B and C). These data indicated that tricellulin enhanced tumor cell survival.