A total of 86 paraffin-embedded specimens of primary CRC were obtained by surgical resection at the Department of Pathology, Xiangya Hospital (Changsha, China) between February and December, 2011. The patient clinical data are presented in Table I. Patients were not pre-treated with radiotherapy or chemotherapy prior to surgery. All patients were followed-up until October 30, 2017, and their complete clinical data were collected. Overall survival (OS) is defined as time to death, irrespective of cause, and censored is defined as loss to follow-up. All procedures were performed according to the ethical guidelines of Xiangya Hospital. Informed consent was obtained from all participating patients. For the use of human samples, the protocol for this study was approved by the Xiangya Hospital Ethics Committee (Changsha, China). The 7th edition of the AJCC TNM staging system for CRC was used to classify tumor stage.

Sections (5-μm-thick) were prepared from the formalin-fixed and paraffin-embedded tissues. Immunohistochemical staining was performed as follows: After dewaxing and rehydrating the tissue slides, antigen retrieval was performed using boiling citrate buffer (pH 6.0) in a pressure cooker for 2.5 min. The slides were subsequently incubated with 3% H₂O₂ for 15 min at room temperature. Following incubation with rabbit anti-HLTF polyclonal antibody (1:100; ab183042; Abcam, Cambridge, UK) overnight (4°C), the slides were washed 3 times with PBS to remove excess primary antibodies. The tissues were subsequently incubated with horseradish peroxidase-conjugated anti-mouse/anti-rabbit secondary antibody (cat. no. SPN-9000; undiluted; included in the Biotin-Streptavidin HRP Detection System, Beijing Zhongshan Golden Bridge Biotechnology, Beijing, China) for 30 min at 37°C. 3,3′-Diaminobenzidine (DAB) was used to visualize positive immune reactions and hematoxylin was used as a counterstain. The primary antibody was omitted for the negative control. Briefly, the sections were scored using a Fourier scale according to the percentage of positive cells and staining intensity: (-) or 0, tissue specimens without staining (0-10%); (+) or 1, tissue specimens with weak staining (10-25%); (++) or 2, tissue specimens with moderate staining (25-50%) and (+++) or 3, tissue specimens with strong staining (>50%). (-) and (+) were defined as low expression, while (++) and (+++) were defined as high expression or overexpression (18). Two board-certified clinical pathologists, who were blinded to the clinical parameters, independently evaluated the staining results. Any disagreement between the two evaluators was resolved by re-evaluation and careful discussion.

The HCT-8, HCT-116, RKO, SW480 and HT-29 cell lines were obtained from the Cancer Research Institute, Central South University (Changsha, China). LoVo cells were purchased from the Cell Bank of the Chinese Academy of Sciences (Shanghai, China). All cells were cultured in RPMI-1640 medium containing 10% fetal bovine serum (FBS) and maintained in an incubator at 37°C in an atmosphere containing 5% CO₂ (both from Gibco/Thermo Fisher Scientific, Inc., Waltham, MA, USA). The cells were seeded in 6-well plates at a density of 5×10⁴/cells/ml and cultured for 24 h prior to exposure to TGF-β1 (cat. no. AF-100-21C-2; Peprotech, Inc., Rocky Hill, NJ, USA) at the concentration of 0, 0.5, 5 ng/ml. Cells were collected after 48 h for use in western blot analysis.

TRIzol reagent (Invitrogen/Thermo Fisher Scientific, Inc.) was used to extract total RNA from the cells. A Prime Script RT reagent kit with a gDNA Eraser (Takara Biotechnology Co., Ltd., Dalian, China) was used to perform the reverse transcription. The resulting cDNA was used for qPCR with a CFX96 Real-Time System (Bio-Rad Laboratories, Inc., Hercules, CA, USA) with SYBR Premix Ex Taq II (Takara Biotechnology Co., Ltd.). All processes were performed according to the corresponding manufacturer’s instructions. The 2^-ΔΔCq method (19) was used to calculate the relative abundance of RNA for each gene compared with GAPDH expression. Each reaction was performed in triplicate. The primer sequences used are presented in Table II.

The CRC cells were washed and lysed in strong RIPA buffer supplemented with proteinase and phosphatase inhibitors (Thermo Fisher Scientific, Inc.). Total protein was extracted and the concentration was determined by bicinchoninic acid protein assay. Equal amounts of proteins (40 μg per sample) were loaded and separated by 8% SDS-PAGE. Following electrophoresis, proteins were transferred onto a 0.22 μm PVDF membrane (EMD Millipore, Billerica, MA, USA). Following blocking with 5% non-fat milk, the membranes were incubated with specific primary antibodies at 4°C overnight, washed extensively and incubated with secondary antibodies (1:6,000 dilution; cat. no. 7074S; Cell Signaling Technology, Inc., Danvers, MA, USA) for 1 h. The protein bands were visualized and quantified using an ECL Advanced Detection System (EMD Millipore). GAPDH was used as loading control. The primary antibodies used were as follows: HLTF (1:1,000; cat. no. 14286-1-AP; ProteinTech Group, Inc., Chicago, IL, USA), SMAD4 (1:1,000, cat. no. 38454T), SMAD2/3 (1:1,000, cat. no. 8685T), Vimentin (1:1,000, cat. no. 5741T) (all from Cell Signaling Technology, Inc.), p-SMAD2/3 (1:500; cat. no. wl02305), zinc finger e-box binding homeobox 1 (ZEB1; 1:500; cat. no. wl01657; both from Wanleibio, China) and (GAPDH; 1:5,000; cat. no. SAB2701826; Sigma-Aldrich, Merck KGaA, Darmstadt, Germany).

shRNAs targeting HLTF and a control shRNA were purchased from GeneChem (Shanghai, China). The shRNA vector (GV493), also from GeneChem, was hU6-MCS-CBh-gcGFP-IRES-puromycin. Moreover, the shRNAs obtained targeted the following sequences of HTLF (NM_139048): shRNA 1# targets AGGT GGAGTTGGTTTGAAT and shRNA 2# targets TATTAGAG AACCGGCCTTA. The HCT-8 and HCT-116 cells were infected with the purified lentiviruses containing the shRNA sequence targeting HLTF or the control sequence lentivirus for 24 h. Stably transfected cells were isolated using 2 μg/ml puromycin (Sigma; Merck KGaA) for 7 days. In addition, human SMAD4 siRNA was purchased from GenePharma (Shanghai, China). The SMAD4-siRNA sequences were as follows: sense, 5′-GAGAAGTTCTCAAAGTTAA-3′ and anti-sense, 5′-TTAACTTTGAGAACTTCTC-3′. The pGV141-HLTF plasmid (purchased from GeneChem), is an expression vector, containing HLTF cDNA and the structure of the vector is CMV-MCS-3FLAG-SV40-Neomycin. When the cells are transfected with pGV141-HLTF, they then overexpress HLTF. ViaFect™ Transfection reagent (cat. no. E4981; Promega Corp., Madison, WI, USA) was used to transfect the plasmid into RKO cells according to the manufacturer’s instructions. The siRNA and plasmid were co-transfected into the cells using Lipofectamine^® 2000 reagent (cat. no. 11668027; Invitrogen; Thermo Fisher Scientific, Inc.).

The cells were cultured on 6-well plates with RPMI-1640 containing 10% FBS. When the cell density reached 70-80%, the bottom of the plate was scratched with a 100 μl pipette tip to create a cell-free gap, following which the cells were incubated for 48 (HCT-116 and RKO) or 72 h (HCT-8) with RPMI-1640 medium containing 1% FBS. An inverted Olympus IX50 microscope (Olympus Corp., Tokyo, Japan) was used to obtain phase-contrast images of the wound healing process at different time-points after scratching. The size of the healed wound was then compared with the size of the initial wound.

A total of 5×10⁴ cells suspended in 100 μl serum-free RPMI-1640 medium were seeded into the upper chamber of a Transwell apparatus (Corning Inc., Corning, NY, USA; 8 μm pore) with 50 μl Matrigel (BD Biosciences, Franklin Lakes, NJ, USA). A total of 600 μl RMPI-1640 medium containing 10% FBS was added to the lower chamber. Following incubation at 37°C for 48 h (HCT-116 and RKO) or 60 h (HCT-8), cells in the upper chamber were removed with a cotton swab and cells on the lower surface were fixed in 1% paraformaldehyde followed by staining with 0.1% crystal violet solution (Beyotime Institute of Biotechnology, Haimen, China) at room temperature. The number of invading cells was determined for 5 randomly selected fields (×200 magnificatoin) under a microscope (Leica inverted microscope DMi1; Leica, Wetzlar, Germany). Three independent experiments were performed and the mean was calculated.

GCBI online software (https://www.gcbi.com.cn/gclib/html/index) was used to predict the putative binding sites for HLTF on SMAD4 promoter region.

Data were analyzed using GraphPad Prism 5.0 software (GraphPad Software, Inc., La Jolla, CA, USA). Differences between 2 groups were assessed using a Student’s t-test. For 3 or more groups, one-way ANOVA test was used followed by Tukey’s post hoc test for multiple comparisons. The associations between HLTF expression and the patient clinicopathological characteristics were determined using the Chi-square and Fisher’s exact tests. Overall survival (OS) curves were plotted according to the Kaplan-Meier method, and differences in survival were examined using the log-rank test. P<0.05 was considered to indicate a statistically significant difference.

To verify the expression of HLTF in CRC, immunochemistry was performed on CRC tissues at various stages. The results revealed that HLTF expression decreased with the increasing TNM stage (Fig. 1A and B). This suggests that HLTF expression is negatively correlated with CRC progression. To explore the association between HLTF and CRC, we analyzed the association between HLTF expression and the patient clinicopathological characteristics. As shown in Table I, the expression of HLTF was significantly associated with the differentiation status (P<0.0001), TNM stage (P<0.0001), depth of invasion (P<0.0001), lymph node metastasis (P<0.0001) and distant metastasis (P<0.01). However, no significant association was observed between HLTF expression and age (P=0.0927) or sex (P=0.1755). To further assess the prognostic value of HLTF in CRC, we performed a survival analysis. Kaplan-Meier analysis revealed that patients with CRC with a high HLTF expression had a significantly longer OS compared with patients with a low HLTF expression (Fig. 1C, log-rank test, P<0.0001), suggesting that a low HLTF protein expression contributes to CRC progression and is associated with a poor prognosis. Based on these results, we thus concluded that the expression of HLTF is negatively associated with the progression of CRC.

The HCT-8 and HCT-116 cell lines were used to assess the effects of HLTF on cell migration and invasion due to their high expression of HLTF (Fig. 1D). We manipulated the cells by knocking down HLTF (Fig. 2A). The results of wound healing (P<0.01; Fig. 2C) and Transwell assays (P<0.0001; Fig. 2B) demonstrated that HLTF knockdown significantly enhanced the migration and invasion of CRC cells in vitro.

HLTF was overexpressed in RKO cells to assess whether the migration and invasion abilities would decrease. Evidence of HLTF overexpression in RKO cells is shown in Fig. 3A. HLTF upregulation significantly suppressed the migration (P<0.0001; Fig. 3B) and invasion (P<0.0001; Fig. 3C) of the CRC cells. These results confirmed that HLTF overexpression suppressed the migration and invasion of CRC cells in vitro.

In order to elucidate the mechanisms through which HLTF suppresses the migration and invasion of CRC cells, we investigated metastasis-related pathways. The results of western blot analysis revealed that the expression of HLTF, SMAD4 and p-SMAD2/3, but not that of total SMAD2/3 were increased when the HCT-8, HCT-116 and RKO cells were stimulated with TGF-β1 (Fig. 4A), suggesting that HLTF is regulated by the TGF-β signaling pathway. We also investigated the mechanism through which HLTF affects the TGF-β signaling pathway. As shown in Fig. 4B, we observed that HLTF knockdown decreased the expression of SMAD4 and p-SMAD2/3, but not that of total SMAD2/3 in both the HCT-8 and HCT-116 cells. The results of RT-qPCR confirmed that the mRNA expression of SMAD4 was also decreased (Fig. 4D). In addition, when using GCBI online software, we predicted that the SMAD4 gene promoter presented putative binding sites for HLTF (51056636 - 51056653, aaaaAAAAAgagggttc; 51027033 - 51027050, aggcACTAAgaagctga) (Fig. 4C). Furthermore, in the HCT-8 and HCT-116 cells, HLTF knockdown upregulated the expression of Vimentin and ZEB1, which are downstream targets of TGF-β signaling (Fig. 4E). The opposite result was observed in the RKO cells overexpressing HLTF (Fig. 4B, D and E). These results suggest that HLTF is associated with the TGF-β/SMAD pathway, and that a high HLTF expression suppresses the migration and invasion of CRC cells by targeting TGF-β/SMAD signaling.

To confirm whether HLTF affects tumor aggressiveness via the TGF-β/SMAD pathway, SMAD4 was knocked down using siRNA in HLTF-transduced RKO (HLTF-siSMAD4-RKO) cells. With 3 siRNAs targeting SMAD4, we finally selected the most effective siRNA (siRNA1#) for the following experiments (data not shown). As shown in Fig. 5A, the SMAD4 mRNA level was significantly decreased in the HLTF-siSMAD4-RKO cells compared with the HLTF-RKO cells (P<0.01). Furthermore, the results of wound healing (P<0.0001; Fig. 5B) and Transwell assays (P<0.001; Fig. 5C) revealed that the motility of the HLTF-siSMAD4-RKO cells was increased compared with that of the HLTF-RKO cells. These results further support the conclusion that HLTF affects CRC cells via the TGF-β/SMAD signaling pathway.