SLCO1B3 promotes colorectal cancer tumorigenesis and metastasis through STAT3

Solute carrier organic anion transporter family member 1B3 (SLCO1B3) is a gene that encodes an organic anion-transporting polypeptide (OATP) 1B3, a membrane-bound multi-specific transporter in hepatocytes. SLCO1B3 was first reported in hepatocytes. Later, it was found that its expression is higher in colorectal cancer (CRC) than in the adjacent normal tissue. However, the role of SLCO1B3 in CRC is not well elucidated. In this study, the correlation between SLCO1B3 and the overall survival (OS) of CRC patients was evaluated using data from the GEO database. This study evaluated the relationship between SLCO1B3 and the clinicopathological characteristics and prognosis of CRC patients. The effects of SLCO1B3 knockdown, on human CRC cell proliferation, migration, and invasion in vitro and CRC tumorigenesis and metastasis in vivo were also examined. In addition, next-generation sequencing was used to identify SLCO1B3 mediators. The results confirmed the association between SLCO1B3 and poor OS of CRC patients, and SLCO1B3 was identified as the top hub gene associated with the OS. The study showed that high SLCO1B3 expression was associated with poor tumor differentiation, advanced disease stage, tumor invasion, lymph node metastasis, and poor OS. Next-generation sequencing revealed that SLCO1B3 knockdown affected the expression of several genes involved in cancer invasion, metastasis, and DNA repair. Moreover, the western blot analysis showed that SLCO1B3 knockdown downregulated p-STAT3, MMP-2, and MMP-9. In summary, we demonstrated that SLCO1B3 acts as a novel carcinogen in the CRC that drives the CRC tumorigenesis and metastasis. SLCO1B3 inhibitors, alone or in combination with current drugs, may have therapeutic benefits in CRC.


INTRODUCTION
Colorectal cancer (CRC) is the fourth most commonly diagnosed cancer and the second leading cause of cancer-related mortality worldwide [1]. CRC is a highly aggressive malignancy with rapid progression. Its pathogenetic heterogeneity confers an escape mechanism to radiation, chemotherapy, and targeted therapies [2]. Despite recent advances in early diagnosis and treatment, the CRC prevalence and death rate in China have shown a significant upward trend over the past ten years [3], which sparked the need for effective management of CRC. Understanding the molecular mechanisms involved in CRC pathogenesis is key to discover new targets and therapeutic agents for CRC.
The SLCO1B3 gene encodes an organic aniontransporting polypeptide (OATP) 1B3, a member of the liver-enriched OATP superfamily. The OATP1B3 was first identified as a membrane-bound multi-specific transporter in hepatocytes responsible for the uptake of endogenous and xenobiotic substances [4,5]. Later studies found that a variant of the liver-type SLCO1B3 mRNA (Lt-SLCO1B3) is expressed in human cancer tissues and cell lines [6]. This cancer-type SLCO1B3 mRNA (Ct-SLCO1B3) has a different transcription initiation site from the Lt-SLCO1B3, and its translated product (Ct-OATP1B3) mainly localized in the cytoplasm of cancer cells. Importantly, Ct-SLCO1B3 was detected significantly higher than Lt-SLCO1B3 in human colon cancer tissues [6]. However, the potential diagnostic and/or prognostic value and the functional role of SLCO1B3 in human CRC remain unexplored.
This study evaluated the relationship between SLCO1B3 and the clinicopathological characteristics and prognosis of CRC patients. We also investigated the function of SLCO1B3 in human CRC cell proliferation, migration, and invasion in vitro and in CRC tumorigenesis and metastasis in vivo. The molecular mechanisms underlying the function of SLCO1B3 in CRC were also explored.

Gene screening and the expression of SLCO1B3 in human CRC and its relationship with disease progression and survival
To further assess the mechanism of CRC acceleration, we downloaded the GSE123734 dataset from the GEO database, which comprised CRC samples. According to data in GSE123734, we found that a high expression level of SLCO1B3 is associated with a lower overall survival rate ( Figure 1A). Cytoscape and cytoHubba were used to catch the Hub genes. Consequently, the SLCO1B3 gene with the highest score was considered a hub gene involved in the cell-matrix adhesion and MMPs pathway GSE123734 ( Figure 1B). In the collected clinical samples compared with adjacent normal tissues, the human CCR tissues exhibited significantly higher SLCO1B3 expression as revealed by qRT-PCR (P < 0.01, Figure 1C). SLCO1B3 is highly expressed in poorly differentiated CRC, whereas reduced expression has been noticed in moderately well-differentiated CRC. The difference is statistically significant (P < 0.01, Figure  1D). Correlation analysis revealed that high tumorous SLCO1B3 was associated with advanced disease stage, tumor invasion, lymph node metastasis, poor tumor differentiation, and low overall survival (Table 1 and Figure 1E). However, SLCO1B3 expression was not associated with gender, age, tumor size, or tumor location in these patients (Table 1).

SLCO1B3 expression in CRC cell lines and select effective interfere sequence
We found that high SLCO1B3 expression in human CRC tissues was associated with advanced disease, relationship between OS and Ct-SLCO1B3 expression in CRC patients was assessed by assessing the mechanism of SLCO1B3 in colorectal cancer acceleration. GSE123734 dataset from the GEO database was downloaded that comprised of colorectal cancer samples. (B) Cytoscape and cytoHubba were used to catch the Hub genes based on GSE123734. As a result, the SLCO1B3 gene with the highest score was considered a hub gene involved in the cell-matrix adhesion and MMPs pathway on GSE123734. (C) The Ct-SLCO1B3 expression in cancer and adjacent normal tissues by qRT-PCR. n=96. **P < 0.01. (D) In poorly differentiated CRC, higher expression of SLCO1B3 and in moderately welldifferentiated CRC reduced expression of SLCO1B3 has been noticed significantly. **P < 0.01. (E) Overall survival analysis of SLCO1B3 expression in cancer tissues based one the collected clinical sample revealed that the high expression level of SLCO1B3 was associated with a lower overall survival rate. The median SLCO1B3 expression level was used as the cutoff for splitting high-expression and low-expression. tumor invasion, lymph node metastasis, and poor patient survival. We investigated the role of this gene in CRC tumorigenesis in vitro and in vivo. We examined SLCO1B3 expression in the NCM460 normal human colon epithelial cell line and the SW480, SW620, HT29, HCT116, and RKO human CRC cell lines. The HCT116, HT29, and SW480 CRC cell lines showed significantly higher SLCO1B3 expression than the NCM460 cell line (Figure 2A, 2B). To evaluate the function of SLCO1B3 in human CRC cell proliferation, migration, and invasion in vitro, we generated HCT116 and SW480 cells transiently transfected with si1-SLCO1B3, si2-SLCO1B3, si3-SLCO1B3, or si-NC. The SLCO1B3 expression was effectively reduced with si1-SLCO1B3 or si2-SLCO1B3, but not with si3-SLCO1B3 transfection ( Figure 2C, 2D). Subsequently, the HCT116 and SW480 cell lines were used to perform the followed research, transfected by si1-SLCO1B3 and si2-SLCO1B3.

SLCO1B3 knockdown inhibits human CRC cell proliferation, migration, and invasion in vitro
Compared with the corresponding control cells, SLCO1B3-silenced HCT116 and SW480 cells exhibited significantly reduced migration and invasion abilities as indicated by the Transwell and wound healing assays (P < 0.05; Figure 3A-3C). In addition, the MTS assay revealed decreased proliferation of the SLCO1B3silenced cells than their corresponding control (P < 0.05, Figure 3D, 3E).

The effects of SLCO1B3 knockdown on CRC tumorigenesis in vivo
We investigated the role of SLCO1B3 in CRC tumorigenesis in a mouse xenograft model and subsequently generated HCT116 and HT29 cells. They were then stably transfected with a lentiviral vector carrying sh-SLCO1B3 or sh-Control. SLCO1B3 knockdown was confirmed with western blot analysis ( Figure 4A). Female BALB/c nude mice received HCT116 cells with stable SLCO1B3 knockdown (sh-SLCO1B3) or an equal number of negative controls HCT116 cells (sh-Control) by subcutaneous injection into the front flank (n=6 per group). The mice were euthanized 21 days after the inoculation. We found that xenograft tumors derived from SLCO1B3-silenced HCT116 cells grew significantly slower than those derived from control cells (P < 0.05). This resulted in AGING smaller and lighter tumors after three weeks of growth ( Figure 4B-4D).

SLCO1B3 activates STAT3 in human CRC cells
To investigate the molecular mechanisms underlying the function of SLCO1B3, we used next-generation sequencing to study the effects of SLCO1B3 knockdown on the gene expression profile of CRC cells. A total of 286 mRNAs were differentially expressed in SLCO1B3-silenced HCT116 cells compared with the mRNAs in control HCT116 cells.
Among which 125 were upregulated, and 161 were downregulated ( Figure 5C). The mRNA expression data from this study are available from the GSE163396 dataset of the GEO database ( Figure 5A). The GO functional enrichment analysis ( Figure 5B, 5C), KEGG pathway analysis ( Figure 5D), and GSEA ( Figure 5E) revealed that many of these DEGs are involved in cancer invasion, metastasis, and DNA repair.
In a previous study, we found that STAT3, an important driver for CRC tumorigenesis and metastasis [7][8][9], was  AGING overexpressed in human CRC tissues compared with adjacent non-tumor colorectal tissues [10]. The coexpression analysis revealed a positive association between the two ( Figure 5F). In our in vitro experiments, SLCO1B3 knockdown in HCT116 cells downregulated p-STAT3 (P < 0.01, Figure 6A, 6B), while total STAT3 remained unchanged. This indicated that SLCO1B3 activates the STAT3 pathway by promoting STAT3 phosphorylation. MMP-2 and MMP-9 are downstream targets of STAT3 that play critical roles in CRC invasion and metastasis [11]. SLCO1B3 knockdown also decreased MMP-2 and MMP-9 (P < 0.01). Thus, the pro-CRC properties of SLCO1B3 were at least partially mediated by STAT3 and its downstream mediators such as MMP-2 and MMP-9. Interleukin-6 (IL-6), a cytokine released from CRC cells, activates STAT3 to drive CRC pathogenesis and metastasis [12]. p-STAT3, MMP-2, and MMP-9 downregulated by SLCO1B3 silencing were restored by IL-6 ( Figure 6A). Female BALB/c nude mice received HT29 cells with stable SLCO1B3 knockdown (sh-SLCO1B3) or an equal number of negative controls HT29 cells (sh-Control) by injection into the tail vein (n=6 per group). The mice were euthanized 21 days after the injection ( Figure 7A, 7B). H&E staining of the liver Figure 7A and lung Figure  7B tissues showing metastatic nodules. The expression of p-STAT3, STAT3, MMP-2, and MMP-9 in metastatic tumors by western blot analysis ( Figure 7C). **P < 0.01 vs. sh-Control.

DISCUSSION
In the present study, the relationship between SLCO1B3 and the clinicopathological characteristics and prognosis of CRC patients was evaluated [13]. Our bioinformatics analysis identified SLCO1B3 as the top hub gene associated with the OS of CRC patients. Thus, SLCO1B3 may serve as a novel therapeutic target for CRC.
In 2008, L. Wooin et al. detected SLCO1B3 expression in most colon tumors [14]. They also found that higher SLCO1B3 expression was associated with lower stage AGING and lower grade tumors and improved 5-year survival within individual tumor grades. This suggested that SLCO1B3 might exhibit antitumor properties in human CRC. However, another 2008 report showed that SLCO1B3 overexpression in CRC cells conferred resistance to drug-induced apoptosis, supporting the protumorigenic function of SLCO1B3 [14]. These seemingly contradictory findings were reported before the discovery of Ct-SLCO1B3 in 2012. In 2012, Nagai and colleagues identified Ct-SLCO1B3 in human colon and lung cancer tissues [6]. The original SLCO1B3 was renamed Lt-SLCO1B3 because of its enriched expression in the liver. Importantly, Ct-SLCO1B3 was detected significantly higher than Lt-SLCO1B3 in both human colon cancer tissues and cells [6,15]. While Lt-OATP1B3 (the translated product of Lt-SLCO1B3) was mostly found as a membrane-bound protein on the cell surface, Ct-OATP1B3 (the translated product of AGING Ct-SLCO1B3) was mainly detected in the cytoplasm of colon cancer cells [15]. In addition, Ct-OATP1B3 and Lt-OATP1B3 exhibited disparate patterns of posttranslational modifications and proteasomal degradation [15]. Compared with Lt-OATP1B3, Ct-OATP1B3 showed poorer transporter activity [15]. Given these differences between the two, it is reasonable to speculate that Ct-SLCO1B3 and Lt-SLCO1B3 may play separate roles in carcinogenesis. Herein, we reported for the first time the function of SLCO1B3 as a carcinogen in CRC. Intriguingly, Ct-SLCO1B3, but not Lt-SLCO1B3, can be induced in response to ambient or chemical hypoxia through the HIF-1a pathway [16]. Thus, Ct-SLCO1B3 may serve as a defense mechanism in cancer cells to sustain tumor growth under hypoxic conditions.
This study found that high SLCO1B3 expression in human CRC tissues is associated with advanced disease stage, tumor invasion, lymph node metastasis, poor tumor differentiation, and poor OS. SLCO1B3 knockdown effectively inhibited human CRC proliferation, migration, and invasion in vitro and curbed CRC tumorigenesis and metastasis in vivo.
Mechanistically, we identified p-STAT3 and its downstream targets MMP-2 and MMP-9 as mediators of the protumorigenic function of SLCO1B3. The STAT3 signaling is overactivated in the CRC, and its downstream mediators such as VEGF, c-Myc and Cyclin D1, MMP-2, and MMP-9 have been linked with tumor angiogenesis, proliferation, invasion, and metastasis, respectively [17][18][19]. Importantly, the activation of STAT3 has been identified as a critical mechanism for resistance to 5-FU [20], the current firstline chemotherapy for advanced CRC [21]. A SLCO1B3 inhibitor would possibly potentiate the efficacy of 5-FU in CRC by inactivating STAT3.
This study has limitations: (i) The primers designed in this study do not specifically recognize CT-OATP1B3, LT-OATP1B3; (ii) Immunohistochemistry in this study is also lacking. Thus, in a future study on this topic, immunohistochemistry should be performed with the antibody that recognizes Ct-OATP1B3.
In summary, we identified SLCO1B3 as a novel carcinogen in the CRC that drives CRC tumorigenesis and metastasis. SLCO1B3 inhibitors, alone or in combination with current drugs, may have therapeutic benefits in the CRC.  All patients included in this study provided written informed consent before test-specific implementation process.

Cell culture
The SW480, SW620, HT29, HCT116, and RKO human CRC cell lines were purchased from the Type Culture Collection of the Chinese Academy of Science (Shanghai, China). The NCM460 normal human colon epithelial cell line was obtained from INCELL (USA). All cell lines were authenticated with short tandem repeat profiling. The SW480, SW620, and HT29 cells were cultured in Dulbecco's modified Eagle's medium (DMEM; Thermo Fisher USA), and the HCT116, RKO, and NCM460 cells were cultured in RPMI-1640 medium (Thermo Fisher). All growth media were supplemented with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin. All cells used in the experiments were mycoplasma-free.

CRC mouse xenograft model
Female BALB/c nude mice (five weeks old) were purchased from Beijing Vital River Laboratory Animal Technology (Beijing, China). The mouse model was established following the previously published protocol [22]. The mice were randomly assigned to two groups (n=6 per group) to receive HCT116 cells with stable SLCO1B3 knockdown or an equal number of negative controls HCT116 cells by subcutaneous injection into the front flank. This was done to evaluate the effects of SLCO1B3 knockdown, on CRC xenograft tumor growth. Tumor length and width were measured with an external caliper once every three days. Tumor volume was calculated using the formula: tumor volume = (length × width 2 )/2. The mice were euthanized 21 days after the inoculation, and the tumors were harvested, photographed, and weighed. The mice were randomly assigned to two groups (n=6 per group) to receive HT29 cells with stable SLCO1B3 knockdown or an equal number of negative controls HT29 cells by injecting them into the tail vein. This was done to evaluate the effects of SLCO1B3 knockdown, on CRC metastasis. The mice were euthanized 21 days after the injection.
The lung and liver tissues were harvested and examined with hematoxylin and eosin (H&E) staining. The metastatic tumors were collected for western blot analysis. All animal experiments were performed following the international standards-3R principle of AGING animal welfare and with permission of the Experimental Animal Ethics Committee of The Fourth Affiliated Hospital of Hebei Medical University.

H&E staining
H&E staining was conducted following the previously published protocol [23]. The paraffin sections of the mouse liver and lung tissues were dewaxed and stained with hematoxylin (Servicebio, Wuhan, China) for 3-5 minutes. After differentiation, bluing, and dehydration in 85% and 95% alcohol for 5 minutes each, the sections were stained with eosin (Servicebio) for 5 minutes, dehydrated, sealed, and observed under the microscope.

Scratch-wound healing and transwell assays
The cell migration ability was assessed using the scratch-wound healing assay [25]. The width of the wound bed was recorded at 0, 24, and 48 hours after the scratch wound. In addition, the cell migration and invasion abilities were evaluated using the Transwell migration and invasion assays, respectively. In brief, 1.5 × 10 5 cells were loaded into the upper chamber of Transwell inserts with an 8-μm pore membrane (Corning Costar, USA), either uncoated (for migration assessment) or pre-coated with Matrigel, BD, USA, (for invasion assessment). After 48 hours, the cells that had migrated to the lower surface of the membrane were fixed, stained with crystal violet, and counted under an inverted light microscope. Each experiment was performed in triplicate.

Identification of differentially expressed genes (DEGs) by microarray mRNA analysis
The microarray mRNA analysis was performed on HCT116 cells transfected with si1-SLCO1B3 or si-NC. The raw microarray data were converted to gene expression data using the Edge R package [26]. The DEGs (|logFC|>2 and padj<0.05) were identified with the Limma R package.

GSEA, GO, and KEGG analysis of DEGs
The Gene Set Enrichment Analysis (GSEA) was performed using the Category package (version 2.10.1). The functional enrichment Gene Ontology (GO) analysis and the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis were performed with the DAVID online tools. Fisher's exact test was used to select the significant pathways. A P value less than 0.05 was considered statistically significant.

Bioinformatics analysis
The overall survival (OS) and mRNA expression data of CRC patients were retrieved from the GSE123734 dataset from the GEO (Gene Expression Omnibus) database. The relationship between OS and SLCO1B3 expression was analyzed by the Kaplan-Meier method. The CytoHubba application in Cytoscape was used to identify and rank the hub genes. The Cytoscape software was used to generate the protein-protein interaction diagram.

Statistical analysis
Each experiment was performed in triplicate. The results are presented as mean ± standard deviation. The differences between the two groups were evaluated using the two-tailed Student's t-test. The correlation between tumorous SLCO1B3 and the clinicopathological characteristics and prognosis of CRC patients was evaluated using the chi-square test. The Kaplan-Meier method (log-rank test) was used to interpret the survival curves. All statistical analyses were performed using the GraphPad Prism 5.0 and SPSS 21.0 software. A P-value of less than 0.05 was considered statistically significant.

Data availability statement
The mRNA expression data from this study will be deposited in the GSE163396 dataset of the GEO database.