ITGBL1 promotes cell migration and invasion through stimulating the TGF‐β signalling pathway in hepatocellular carcinoma

Abstract Objectives Integrin beta‐like 1 (ITGBL1) is involved in the migration and invasion of several cancers; however, its roles in the development and progression of hepatocellular carcinoma (HCC) remain largely unknown. Materials and methods Immunohistochemistry staining was used to investigate the expression pattern of ITGBL1 and its prognostic values in HCC patients. The transwell, wound‐healing assays, xenograft and orthotopic mouse models were employed to determine the effects of ITGBL1 on HCC cell migration and invasion in vitro and in vivo. The biological mechanisms involved in cell migration and invasion caused by ITGBL1 were determined with Western blotting and RT‐PCR methods. Results ITGBL1 expression was significantly increased in HCC tissues compared to adjacent normal tissues. Patients with higher ITGBL1 expression were associated with more reduced overall survival. ITGBL1 overexpression promoted migration and invasion in SMMC‐7721 and HepG2 cells in vitro and in vivo, whereas knockdown or knockout ITGBL1 in CSQT‐2 cells significantly reduced cell migration and invasion abilities. In SMMC‐7721 cells, ITGBL1 overexpression stimulated TGF‐β/Smads signalling pathway, along with the KRT17 and genes involved in the epithelial‐mesenchymal transition (EMT). In contrast, ITGBL1 knockout inhibited the TGF‐β/Smads signalling pathway in CSQT‐2 cells. Conclusions These findings suggested that ITGBL1 promoted migration and invasion in HCC cells by stimulating the TGF‐β/Smads signalling pathway. ITGBL1 could be a promising prognostic biomarker, as well as a potential therapeutic target in HCC.


| INTRODUC TI ON
Liver cancer ranks as the second leading cause of cancer-related mortalities worldwide, which led to an estimated 810 000 deaths per year. 1 Importantly, 90% of primary liver cancer patients have been diagnosed as hepatocellular carcinoma (HCC). 2 Various risk factors of HCC have been identified, including hepatitis B virus (HBV) or hepatitis C virus (HCV) infection, exposure to aflatoxin B1 (AFB1) and/or aristolochic acid and a history of diabetes, etc 2 ; these factors have provided efficient targets for HCC prevention. The prognosis of HCC patients is usually poor, with the 5-year post-operative survival rate being 25%-50% for patients with early-stage HCC. 3 However, according to the Surveillance, Epidemiology and End Results (SEER) database, 4,5 the 5-year survival rate of HCC patients with distant metastases was dropped below 5%. After surgery, about 70% of patients have been shown to relapse or develop distant metastasis within 2 years. 3 Thus, identification of novel therapeutic targets or prognosis prediction biomarkers may provide clues for the development of novel treatment strategies.
ITGBL1 (integrin beta-like 1, also termed as OSCP or TIED) encodes a beta integrin-related protein that is a member of the epidermal growth factor (EGF)-like protein family. ITGBL1 was first cloned and characterized in 1998 from the overlapping cDNA clones of foetal lung, human umbilical vein endothelial cell (HUVEC) and osteoblast cDNA libraries. 6 Previous studies also showed that ITGBL1 was involved in the development and progression of several cancers. In breast cancer tissues, ITGBL1 is co-expressed with bone remodelling-and bone metastasis-related genes, 7 and it could promote the bone metastasis of breast cancer cells through activating the TGF-β signalling pathway. 8 In colorectal cancer (CRC), the expression of ITGBL1 was upregulated, 9,10 and the primary tumours could release ITGBL1-rich extracellular vesicles to induce the activation of resident fibroblasts in remote organs, which promotes the metastatic cancer growth. 11 In our previous gene expression profiling data set of HBV-related liver fibrosis tissues, ITGBL1 was positively associated with the fibrosis stage and identified as a key regulator of fibrogenesis in patients with HBV infection. 12 In addition, ITGBL1, together with CD24, CXCL6, EHF, LUM and SOX9, has been suggested to act as a predictive biomarker of cirrhosis in patients with chronic HBV infection. 13 Whether ITGBL1 is involved in the pathogenesis and progression of HCC and the related underlying mechanisms remain unclear.
In the current study, we aimed to determine the gene expression profiling of ITGBL1 and its prognostic values in HCC patients. The roles of ITGBL1 in HCC cell migration and invasion and the corresponding underlying molecular mechanisms were also determined.
This study may unveil the new insights into the progression of HCC and provide novel prognosis biomarkers or potential therapeutic targets for HCC patients.

| HCC patient recruitment
The formalin-fixed, paraffin-embedded (FFPE) tumour and adjacent normal tissue samples from 98 HCC patients, who received the curative surgery treatment in the Eastern Hepatobiliary Surgery Hospital of the Second Military Medical University between July 2012 and February 2014, were collected. 14 The patients who did not receive any anti-cancer treatments before they underwent curative resection surgery were included in this study. If the patients who had a distant metastasis or (and) a history of other malignancies were excluded in the study. And if the patients who had received the palliative treatments or (and) the liver transplant treatment were also been excluded in the study. Patients were also excluded if they were unwilling to participate in the study. Based

| Western blotting analysis
The proteins from the liver tissues and cells were extracted using

| Immunohistochemistry (IHC) staining
Tissue microarrays (TMAs) of 98 HCC patients were constructed as reported previously. 14 IHC staining was performed to detect the expression of ITGBL1 for each patient. The TMAs were dewaxed twice with xylene for 20 minutes each time and rehydrated with gradient ethanol solutions for 5 minutes at a time. After reheated for 0.5 hours at room temperature, the TMAs were then treated with 3% hydrogen peroxide in methanol to quench the endogenous peroxidase activity for 30 minutes at room temperature. Then, the TMAs were boiled with the citric acid solution (pH = 6.0) for antigen retrieval at 92°C for 40 minutes. After natural cooling, the TMAs were washed with 1 × PBS and then blocked using 1% foetal bovine serum (FBS) for 30 min at room temperature. And the TMAs were incubated with anti-ITGBL1 antibodies (1:100; cat. no. NBP1-82473; Novus Biologicals) at 4°C overnight. Then, the TMAs were washed with 1 × PBS for three times and incubated with peroxidase polymerlabelled goat anti-rabbit secondary antibodies (cat. no. HAF008; Novus Biologicals) at room temperature for 1 hour. The TMAs were washed three times with 1 × PBS and then incubated with diaminobenzidine, counter-stained with haematoxylin, dehydrated and cover-slipped. The TMAs were examined using a Leica CTR5000 microscope (Leica Microsystems).
The immunoreactivity score (IRS) was calculated by multiplying the staining intensity with the percentage of positively stained cells. 15 The patients of HCC patients were divided into two groups according to the median IRS score of ITGBL1 expression level in cancer tissues (higher vs lower).

| Construction of a CRISPR/Cas9 lentivirus vector targeting ITGBL1
The sgRNA sequences of ITGBL1 were designed using the online CRISPR Design Tool (http:/tools.genome-engineering.org), and three different sgRNA1, sgRNA2 and sgRNA3 were selected (Table S2). 16 Synthetic sgRNA oligonucleotides were cloned into the lentiCRISPRv2 construct at BsmBI restriction sites (cat. no.

| RNA extraction and real-time PCR
The TRIzol reagent (cat. no. 16 096 040; Thermo Fisher Scientific) was used to extract the total RNAs from the liver tissue specimens or cells following the manufacturer's instructions. The reverse transcription and RT-PCR were performed using the PrimeScript™RT Master Mix (cat. no. RR036A; Takara) and the QuantiNova SYBR Green PCR Kit (cat. no. 208054; Qiagen), respectively. The amplification parameters were 95°C for 2 minutes, followed by 40 cycles of 95°C for 5 seconds and 60°C for 10 seconds, according to the manufacturer's protocol. The mRNA level of GAPDH was used as an internal control. The gene-specific RT-PCR primers used in this study are provided in Table S3. The relative expression of the mRNA levels was calculated using the ΔΔCt (2 −ΔΔCT ) method. 17

| Wound-healing assay
SMMC-7721, HepG2 and CSQT-2 cells were plated into 6-well culture plates at a density of 5 × 10 5 cells/well and grown until confluence overnight. After scratching three parallel lines onto the confluent cell layer, the cells were washed three times with 1 × PBS and culture with DMEM medium containing 10% FBS. Images of migrating cells were sequentially taken 0, 24 and 48 hours after the scratch. The relative wound region among different sample groups was evaluated and compared.

| Transwell assay
The transwell assays were performed using the transwell cham- The numbers of migrating or invading cells were counted in three randomly selected fields. The differentially expressed genes between the two groups' cells were analysed using the DESeq2 R package (1.16.1). Genes with a false-discovery rate (FDR) adjusted P-value of <.05, as estimated using the DESeq2 package, were assigned as differentially expressed genes. The differentially expressed genes between the conditions are listed as Table S4.

| Statistical analysis
The Oncomine database (https://www.oncom ine.org) and The Cancer Genome Atlas (TCGA) HCC data set were used to evaluate the mRNA expression levels of ITGBL1 and KRT17 in the normal and cancer tissues of HCC patients, and compared using Student's t test between the groups. The non-parametric Wilcoxon's rank test was used to compare the protein expression of ITGBL1 in the tumour tissues and the matched non-tumour tissues. The χ 2 test was performed to evaluate the correlations between the ITGBL1 expression and clinicopathological features. The Kaplan-Meier plot, together with the log-rank test, was used to compare the overall survival (OS) between the groups. Univariate Cox regression was used to determine the associations between the clinical features and OS of the patients. To determine the independent prognostic factors associated with the OS of the HCC patients, the multivariate Cox proportional hazards model was applied. One-way analysis of variance (ANOVA) was used to test the overall differences between the groups, and the difference between subgroups was compared by Turkey's honestly significant difference (Turkey's HSD) test. The SPSS (version 22.0) and the GraphPad Prism software (version 6.0, GraphPad) were used for the statistical analyses. Two-sided P-value < .05 was considered as statistically significant for all analyses.

| ITGBL1 expression is upregulated in HCC tumour tissues
To determine the expression pattern of ITGBL1 in HCC tissues, the TCGA and Oncomine databases were searched, and we found that the ITGBL1 mRNA levels were significantly increased in HCC cancer tissues compared to the adjacent normal tissues ( Figure 1A,B).
We confirmed these results in 22 pairs of HCC patients by the RT-PCR method ( Figure 1C). Further, the Western blotting results also suggested that the protein expression level of ITGBL1 was upregulated in eight paired HCC tumour tissues, compared with the adjacent non-tumour tissues ( Figure 1D,E). Using the IHC staining method, we investigated the ITGBL1 expression level in 98 paired HCC tissues. We found that ITGBL1 was highly expressed in HCC tumour tissues, whereas weak or negative expression of ITGBL1 was observed in the adjacent non-tumour tissues (Wilcoxon's signed-rank test, P < .001, Figure 1F). These results implicated that both the mRNA and protein expression level of ITGBL1 were significantly increased in HCC tumour tissues compared to adjacent normal tissues.

| Higher ITGBL1 expression predicts a poorer prognosis in HCC patients
Next, we evaluated the associations of ITGBL1 protein expression levels and the clinicopathological characteristics in 98 HCC patients. We found that ITGBL1 expression level was significantly associated with the tumour encapsulation status (P = .037), but not other clinicopathological characteristics (Table S1). The Kaplan-Meier survival analysis suggested that HCC patients with higher ITGBL1 expression had relatively poorer OS than those with lower ITGBL1 expression (log-rank test, P = .014, Figure 1G). In the univariate analysis, the clinicopathological characteristics including AFP (P = .038), tumour number (P < .001), BCLC stage (P < .001), vascular invasion (P = .003) and ITGBL1 expression (P = .014) were significantly associated with the OS in HCC patients ( Table 1). The multivariate Cox proportional hazards model also suggested that ITGBL1 was an independent prognostic factor for HCC patients (P = .040; Table 1).
These results highlighted that patients with relatively higher ITGBL1 expression level were associated with the poorer OS; the ITGBL1 expression level may, thus, serve as a novel prognostic factor for HCC patients.

| Overexpression of ITGBL1 promotes HCC cell migration and invasion in vitro and in vivo
To investigate the roles of ITGBL1 in HCC cells, we first evaluated the expression of ITGBL1 in several liver cell lines. As shown in Figure    The tail vein injection model also suggested that ITGBL1 knockout in CSQT-2 cells significantly reduces its lung metastasis abilities compared to wild-type cells ( Figure 4H). Besides, the orthotopic liver tumour models also showed that ITGBL1 knockout significantly inhibited the metastatic colonization of CSQT-2 cells in the lungs, compared to the injection of control CSQT-2 cells ( Figure 4I). These data demonstrated that ITGBL1 suppression inhibited the migration and invasion of HCC cells both in vitro and in vivo.

| ITGBL1 regulates the expression of KRT17
To uncover the mechanisms underlying the ITGBL1-mediated regulation of cellular metastasis and invasion, we performed the RNAsequencing analysis of the SMMC-7721 cells with or without ITGBL1 overexpression. A total of 195 differentially expressed genes were identified, including 135 upregulated and 60 downregulated genes ( Figure 5A and Table S4). Eight genes, including MMP17, WNT6, KRT17, FOXQ1, FOS, SMOC1, VEGFA and CA9, had been verified via the RT-PCR method ( Figure 5B). We noticed that the expression of one gene, KRT17, which has been reported to stimulate the migration and invasion of the tumour cells, 18,19 was significantly increased following ITGBL1 overexpression. The Western blotting results also confirmed that the KRT17 expression was significantly increased in the ITGBL1 overexpressing SMMC-7721 cells, but was significantly decreased in the ITGBL1 knockout CSQT-2 cells ( Figure 5C). These results were confirmed by RT-PCR analyses through overexpressing and knockout of ITGBL1 in SMMC-7721 and CSQT-2 cells, respectively ( Figure 5D). Subsequently, we further determined the expression of KRT17 in the TCGA HCC data set and found that the KRT17 mRNA expression level is significantly higher in primary liver tumour tissues than the normal liver tissues ( Figure 5E). Moreover, we also found that the expression of KRT17 was increased in HCC tumours with F I G U R E 1 ITGBL1 expression is upregulated in HCC tissues and associated with the prognosis of HCC patients. A, ITGBL1 expression in the normal liver tissues (n = 50) and primary liver tumour tissues (n = 371), as revealed by the TCGA HCC data set (P < .001). B, ITGBL1 expression in the normal liver (n = 19) and HCC tissues (n = 47), as revealed by the Oncomine database (GSE14323; P < .001). C, Different ITGBL1 mRNA expression levels in 22 paired samples of HCC tumour tissues and adjacent non-cancerous tissues, as determined by RT-PCR analyses (P = .0057). D, Western blotting analyses of the ITGBL1 protein expression levels in eight randomly selected tumour tissues and the matched adjacent non-tumour tissues from HCC patients; GAPDH was used as the internal loading control. E, Greyscale analysis of the ITGBL1 expression level, compared to the GAPDH expression level, in the eight patients (P < .001). T, tumour tissue; N, normal tissue. F, Immunochemistry staining of ITGBL1 in the TMAs with HCC tissues and adjacent non-cancerous tissues (n = 98; P < .001). G, The Kaplan-Meier plot for the overall survival of HCC patients according to ITGBL1 expression levels (higher vs lower, log-rank test, P = .014) the lymph node metastasis ( Figure 5F). Taken together, these results demonstrated that KRT17 might be regulated by ITGBL1 in HCC cells.

| ITGBL1 induces EMT via the TGF-β/Smads signalling pathway
It has been reported that KRT17 is a downstream factor of TGF-β1 and that ITGBL1 may stimulate liver fibrosis through regulating the TGF-β1 signalling pathways. 12 We determined whether ITGBL1 regu-

| D ISCUSS I ON
In clinical settings, intrahepatic and extrahepatic metastasis of Western blotting (C) and RT-PCR analyses (D) for the quantification of KRT17 expression levels in SMMC-7721 cells with or without ITGBL1 overexpression and CSQT-2 cells with or without ITGBL1 knockout. E, KRT17 expression levels in the normal liver tissues (n = 50) and primary liver tumour tissues (n = 371), as revealed by the data for HCC samples in the TCGA database. F, KRT17 expression in the normal liver tissues (n = 50), HCC tissues with no regional lymph node metastasis (N0, n = 252) and HCC tissues with 1-3 axillary lymph nodes (N1, n = 4) were determined in TCGA database. * P < .05, ** P < .01 and *** P < .001 indicate a significant difference between the groups and other EMT-related biomarkers through stimulating the TGF-β1/Smads signalling pathway in HCC cells ( Figure 7E). These results suggested that ITGBL1 might serve as a novel prognostic biomarker for HCC patients, and the ITGBL1/TGF-β1/Smads signalling pathways could act as potential targets for metastasis prevention in HCC patients.
F I G U R E 6 ITGBL1 promotes the epithelial-mesenchymal transition in HCC cells. A, Western blotting analysis of EMT-related genes in SMMC-7721 cells with or without ITGBL1 overexpression. B, RT-PCR analysis the EMT-related genes in SMMC-7721 cells with ITGBL1 overexpression (left) and CSQT-2 cells with ITGBL1 knockout (right). Data are shown as the mean ± SD. C, RT-PCR analysis for the quantification of EMT-related genes in SMMC-7721 cells with or without ITGBL1 overexpression following the treatment with TGF-β1. Data are shown as the mean ± SD. D, RT-PCR analysis for the quantification of the expression of EMT-related genes in SMMC-7721 cells with or without ITGBL1 overexpression following the treatment with the TGFBR1/2 dual inhibitor LY2109761. Data are shown as the mean ± SD. * P < .05, ** P < .01 and *** P < .01 indicate a significant difference between the groups F I G U R E 7 ITGBL1 promotes EMT via the TGF-β/Smads signalling pathway. A, Western blotting analysis (left) and the greyscale analysis (right) of genes involved in EMT and downstream of the TGF-β signalling pathway in SMMC-7721 cells (with or without ITGBL1 overexpression) individually treated with TGF-β1 or LY2109761 or the combination of TGF-β1 and LY2109761 treatment. GAPDH was used as the internal loading control. B, Western blotting analysis (left) and the greyscale analysis (right) of EMT-related genes and the genes downstream of the TGF-β signalling pathway in CSQT-2 cells (with or without ITGBL1 knockout) individually treated with TGF-β1 or LY2109761 or the combination of TGF-β1 and LY2109761 treatment. GAPDH was used as the internal loading control. C, RT-PCR analysis for the quantification of the mRNA levels of EMT-related genes and KRT17 in SMMC-7721 cells with or without ITGBL1 overexpression following the treatment with TGF-β1, LY2109761, or a combination of TGF-β1 and LY2109761. Data are shown as the mean ± SD. D, RT-PCR analysis for the quantification of the mRNA levels of EMT-related genes and KRT17 in CSQT-2 cells with or without ITGBL1 knockout following the treatment with TGF-β1, LY2109761, or a combination of TGF-β1 and LY2109761. Data are shown as the mean ± SD. E, A schematic model depicting that ITGBL1 stimulates the expression of KRT17 and induces EMT via TGF-β/Smads signalling pathway, which promotes invasion and metastasis of HCC cells. * P < .05, ** P < .01 and *** P < .001 indicate a significant difference between the groups Previous studies have reported that the increased expression of ITGBL1 is associated with tumorigenesis and poorer prognosis in a variety of cancers. In ovarian cancer, ITGBL1 was upregulated in ovarian cancer tissues compared to adjacent non-cancer tissues, and it was positively correlated with lymph node invasion and advanced FIGO stage. 8,22 In prostate cancer (PCa) patients, ITGBL1 was significantly upregulated and positively associated with lymph node metastasis status. 26 In the current study, we found both the protein and mRNA levels of ITGBL1 were increased in HCC tumour tissues, compared with the adjacent non-tumour tissues. In clinic, higher ITGBL1 was correlated with incomplete encapsulation of HCC patients and it was independently associated with the overall survival of HCC patients.
Previous studies suggested that unencapsulated or incomplete encapsulated tumours could invade directly into the surrounding liver parenchyma, cause the destruction of the extracellular matrix and then migrate into the circulation. 28 These indicated that ITGBL1 acts as an oncogene in HCC and may promote the migration and invasion of HCC However, the expression of KRT17 and its correlation with ITGBL1 in HCC remains unclear. From the TCGA database, we found that the mRNA expression level of KRT17 is higher in primary liver tumour tissues compared to that in normal liver tissues. Besides, we found that the expression of KRT17 increased significantly in ITGBL1-overexpressing SMMC-7721 cells and decreased in ITGBL1knockout CSQT-2 cells. Inhibition of the TGFBR1/II significantly inhibits the ITGBL1 induced expression of KRT17. These results indicated that KRT17 was regulated by ITGBL1 through its regulation on TGF-β/Smads signalling pathway in HCC cells; however, the roles of KRT17 in HCC pathogenesis and progression remain to be elucidated in future.
In summary, our study showed that ITGBL1 was upregulated in HCC and associated with the OS in HCC patients. In HCC cells, ITGBL1 induces increased expression of KRT17 and EMT-related genes through activating the TGF-β/Smads signalling pathway, thereby promoting the migration and invasion of HCC cells. These data provided evidence for ITGBL1 as a promising prognostic factor for HCC, and it may also serve as a novel therapeutic target in HCC.

ACK N OWLED G EM ENTS
The authors acknowledge the members of the Research Laboratory of Clinical Virology, especially Donghua Zhang, Jiehong Jiang, Jia Chen and Yongyan Chen for experimental material assistance.

CO N FLI C T O F I NTE R E S T
The authors declare that they have no competing interests.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.