Sequential Treatment with Activin and Hepatocyte Growth Factor Induces FOXM1 to Promote Colorectal Cancer Liver Metastasis

Background Cancer stem cells (CSCs) are involved in liver metastasis in colorectal cancer (CRC). Activin and hepatocyte growth factor (HGF) are important regulators of stem cell properties. This study was performed to explore the effect of activin and HGF on CRC invasion and metastasis. The key genes involved in the action of activin and HGF in CRC were identified. Methods HCT116 CRC cells were sequentially treated with activin and HGF and examined for migration and invasion in vitro and liver metastasis in vivo. RNA sequencing was performed to identify differentially expressed genes in response to activin and HGF. Results Sequential treatment with activin and HGF-enhanced CRC cell migration, invasion, and metastasis. CXCR4 and AFP expressions were increased by activin and HGF treatment. Knockdown of FOXM1 blocked liver metastasis from HCT116 cells pretreated with activin and HGF and suppressed CXCR4 and AFP expression. Activin alone increased the mRNA and protein expression of FOXM1. In contrast, HGF alone enhanced the phosphorylation of FOXM1, without altering the total protein level of FOXM1. SMAD2 was required for activin-mediated FOXM1 induction. FOXM1 transactivated CXCR4 by directly binding to the promoter of CXCR4. Additionally, CXCR4 regulated AFP expression through the NF-κB pathway. Conclusions Sequential treatment with activin and HGF accelerates CRC invasion and liver metastasis, which involves the upregulation and activation of FOXM1 and induction of CXCR4 and AFP.


Introduction
Colorectal cancer (CRC) is the third most common cancer globally and the second leading cause of cancer-related deaths [1]. CRC has an extremely high propensity to spread to the liver, consequently leading to a poor prognosis [2]. Most CRC patients with liver metastasis have lost the opportunity to receive curative resection and die of the disease within 5 years [2]. Hence, it is signifcant to develop therapeutic strategies that can prevent liver metastasis in CRC patients [3]. However, the mechanism underlying CRC liver metastasis remains unclear.
Cancer stem cells (CSCs), which represent a small subset of tumor cells (1-10%), are responsible for tumor growth, relapse, and treatment resistance [4,5]. CSCs have been shown to be involved in CRC liver metastasis [5]. Activin and hepatocyte growth factor (HGF) have the capacity of inducing liver progenitor cells from human embryonic stem cells [6]. Activin can promote stem cells to diferentiate into defnitive endoderm cells expressing characteristic markers, such as CXCR4, FOXA2, SOX17, and CER [7]. In turn, HGF induces the diferentiation of defnitive endoderm cells into liver progenitor cells, which are characterized by alphafetoprotein (AFP) expression [7]. Recent studies have suggested that activin is overexpressed in stage IV CRC [8] and plays an essential role in CRC plasticity and metastasis [9]. Moreover, the activin pathway enhances CRC stem cell selfrenewal and tumor progression [10]. Autocrine HGF/c-Met activation can induce common adenocarcinoma cells to dediferentiate to a CSC state [11]. Given the activities of activin and HGF in regulating CSC, we hypothesized that activin and HGF might have an important impact on CRC metastasis through the modulation of cancer stemnessrelated factors.
To test the hypothesis, this study was conducted to determine the efect of activin and HGF on CRC invasion and metastasis. Transcriptome analysis was performed to identify key genes involved in the action of activin and HGF in CRC.

Patients and Tissue Samples.
Parafn-embedded primary tumor and paired liver metastasis samples were obtained from CRC patients with liver metastasis who were treated at Bishan Hospital of Chongqing (Chongqing, China). Te samples were analyzed using immunohistochemical staining with CXCR4 and AFP primary antibodies (Abcam, Cambridge, MA, USA) and an Alexa Fluor-conjugated secondary antibody (1 : 500; Molecular Probes, Eugene, OR, USA). Te Medical Ethics Committee of Bishan Hospital of Chongqing approved the study protocol. Te patients provided informed consent and had not received radiation or chemotherapy before surgery.

Cell Culture and Transfection.
Human CRC cell line HCT116 was obtained from the American Tissue Culture Collection (Manassas, VA, USA) and cultured under the recommended conditions. According to the manufacturer's protocol, transfections were performed in 70-80% confuent cells using Lipofectamine 3000 (Invitrogen, Carlsbad, CA, USA).

Activin and HGF Treatment.
HCT116 cells were treated sequentially with activin and HGF, as reported previously [12]. Briefy, at 70% confuence, HCT116 cells were exposed to 100 ng/mL activin A (Peprotech, Rocky Hill, NJ, USA) for 5 days. Te cells were then treated with 20 ng/mL HGF (Peprotech) for another 5 days. Afterward, the cells were tested for gene expression, migration, and invasion.

RNA Extraction and Real-Time PCR Analysis.
Total RNA was isolated using RNAiso Plus (Takara, Shiga, Japan), and 1 μg RNA was used to synthesize cDNA with the Prime-Script ™ RT reagent Kit with gDNA Eraser (Takara). SYBR Green master mix (Bio-Rad, Hercules, CA, USA) was used for real-time PCR. β-Actin was used as an internal control. Te primers used for real-time PCR are listed in Table 1.

Transwell Migration and Invasion
Assays. Cell migration was tested using the 24-well transwell with polycarbonate flters. Activin-and HGF-treated cells were suspended in a serum-free medium and added to the upper chamber at a density of 6 × 10 3 cells/well. Dulbecco's modifed Eagle medium containing 10% fetal bovine serum was added to the lower chamber. After incubation for 24 h, the cells were fxed with 4% formaldehyde and stained with crystal violet for 5 min. Cell invasion was assayed using a similar method. Matrigel (BD Biosciences, San Jose, CA, USA) was added to each well and incubated for 5 h at 37°C before the cells were seeded to the upper chamber of the transwell. Te cells that invaded through the Matrigel were counted after culturing for 48 h. Te average number of migrating or invading cells was calculated in fve felds (200x) on each membrane.

In vitro Wound-Healing Assay.
Te wound-healing assay was performed using a 12-well cell culture plate. A 200-μL sterile pipette tip was used to make a scratch in a cell layer of approximately 100% confuence. After incubation for 48 h in a serum-free medium, the rate of wound closure was measured.
2.9. RNA-Sequencing (RNA-Seq) Analysis. Next-generation sequencing libraries were constructed by depleting ribosomal RNA using the Human/Mouse/Rat Ribo-Zero rRNA Removal Kit for Illumina (Epicenter, Madison, WI, USA). At elevated temperatures, the poly(A)+ or poly(A)− RNA fractions were cleaved into small fragments by divalent cations to ensure purifcation. A cDNA library was constructed by reverse transcription of the cleaved RNA fragments. Te average insert size for the paired-end libraries was 250-350 base pairs. Te cDNA libraries were sequenced on an Illumina NovaSeq ™ 6000 system (LC-Bio Technology CO., Ltd., Hangzhou, China) following the recommended protocol. Te gene expression data have been deposited in the Gene Expression Omnibus (GEO) database under accession number GSE205986.

Animal Studies.
Animal experiments were performed according to the NIH Guide for the Care and Use of Laboratory Animals. Te protocol was approved by the Institutional Animal Care and Use Committee of Chongqing Medical University. Five-week-old female nude (nu/nu) mice were purchased from GemPharmatech Co., Ltd. (Nanjing, China). HCT116 cells with indicated treatments (5 × 10 5 cells/mouse) were injected into the spleen of the nude mice using a 29-gauge needle. Animals were monitored daily for adverse efects. Each group contained 10 nude mice. Six weeks after cell injection, the mice were sacrifced. An autopsy was performed to check for tumor metastasis. Metastatic lesions in the liver were counted.
2.11. Luciferase Reporter Assay. Te CXCR4 promoter (−1995 to +2 bp) and its truncated fragments (−912 to +2 bp, −613 to +2 bp, and −421 to +2 bp) were cloned into a pGL3-Basic vector. All constructs were verifed by sequencing. For the luciferase reporter assay, cells were plated in 24-well plates and cotransfected with luciferase reporter constructs, FOXM1-expressing plasmid, and pRL-TK Renilla luciferase control reporter vectors. Luciferase activity was measured 36 h after cell transfection using the dual-luciferase reporter assay system (Promega, Madison, WI, USA) according to the manufacturer's instructions.

Chromatin Immunoprecipitation (ChIP) Assay.
Cells were fxed with 1% formaldehyde for 10 min at room temperature. Sonicated cell lysates were subjected to immunoprecipitation using 5 μg anti-FOXM1 (Cell Signaling Technology) or isotype control IgG. Immunoprecipitated DNA was extracted and subjected to PCR analysis. Primers used for ChIP are listed in Table 1.

Statistical Analysis.
All statistical analyses were performed using SPSS software (version 22.0, SPSS, Inc., Chicago, IL, USA). Diferences between the two groups were determined by Student's t-test. Diferences among groups were assessed using a one-way analysis of variance. Data are expressed as the means ± standard deviation. P < 0.05 was considered statistically signifcant.

Sequential Treatment with Activin and HGF Promotes CRC Invasion and Metastasis.
To investigate the efect of sequential treatment with activin and HGF on CRC biology, HCT116 CRC cells were treated initially with activin for 5 days, followed by HGF for another 5 days. Transwell assays revealed that activin and HGF treatment increased the migration and invasion abilities of HCT116 cells (Figures 1(a) and 1(b)). In vitro wound-healing assays demonstrated that the motility of HCT116 cells was enhanced in response to activin and HGF treatment ( Figure 1(c)). To determine whether activin and HGF treatment had a long-term impact on CRC cells, HCT116 cells were sequentially treated with activin and HGF and then injected into the spleen of nude mice. As shown in Figure 1(d), activin and HGF treatment of HCT116 cells facilitated the development of metastatic lesions in the liver. Tese data indicate the prometastatic efect of activin and HGF on CRC cells.

Activin and HGF Treatment Induces CXCR4 and AFP
Expression in CRC Cells. CXCR4 and AFP are recognized as biomarkers of endoderm cells and liver progenitor cells, respectively. Analysis of data from the Human Cancer Metastasis Database (http://hcmdb.i-sanger.com/index) revealed that CXCR4 and AFP were expressed at higher levels in liver metastases than in primary colon tumors ( Figure 2(a)). Immunohistochemical analysis confrmed that CXCR4 and AFP levels were increased in liver metastases compared to primary colon tumors ( Figure 2(b)). Next, we checked the efect of activin and HGF treatment on the expression of CXCR4 and AFP in CRC cells. Activin treatment alone signifcantly upregulated the expression of CXCR4 but not AFP in HCT116 cells (Figures 2(c) and 2(d)). Sequential treatment with activin and HGF led to a signifcant increase in the expression of both CXCR4 and AFP in HCT116 cells (Figures 2(c) and 2(d)).

FOXM1 is Involved in the Prometastatic Activity of Activin and HGF in CRC.
To explore the mechanism underlying the prometastatic efect of activin and HGF on CRC, transcriptome sequencing was performed in HCT116 cells subjected to activin alone or in combination with HGF treatment. RNA-Seq analysis revealed that the gene expression profles signifcantly difered in response to activin and HGF treatment (Figure 3). A total of 2,590 genes were diferentially expressed between the control and activintreated cells and 1,302 genes between the activin-treated Table 1: Primers used in this study.

Forward
Reverse Canadian Journal of Gastroenterology and Hepatology 3   b)). In particular, FOXM1, CXCR4, and AFP were induced by sequential activin and HGF treatment (Figure 3(c)). Gene Ontology analysis revealed that the diferentially expressed genes in response to activin and HGF treatment were enriched in several critical cancer-related processes, including "pathways in cancer," "cell diferentiation," and "signal transduction." Forkhead box M1 (FOXM1) is an important transcription factor that is overexpressed in many cancers [13][14][15]. It has been reported that FOXM1 is associated with cellular senescence [16] and stemness in CRC [17]. Next, we checked whether FOXM1 mediated the efect of activin and HGF on CRC cells. To address this, HCT116 cells were transfected with FOXM1 shRNA before sequential activin and HGF treatment. Transfection with FOXM1 shRNA resulted in signifcant downregulation of endogenous FOXM1 (Figure 4(a)). Both CXCR4 and AFP expression were decreased in FOXM1-depleted cells (Figures 4(a) and 4(b)). In vivo studies showed that FOXM1 depletion impaired the formation of liver metastases from HCT116 cells pretreated with activin and HGF. Te average number of metastatic foci in the FOXM1 depletion and control groups was 9 ± 2 and 2.3 ± 0.58, respectively (Figure 4(c)). Immunohistochemical analysis indicated the downregulation of CXCR4 and AFP in FOXM1-depleted tumors (Figure 4(d)). Collectively, these data suggest that FOXM1 is required for CRC liver metastasis induced by activin and HGF.

Activin and HGF Induces the Expression and Phosphorylation of FOXM1 to Upregulate CXCR4 and AFP.
Treatment with activin alone increased the mRNA and protein expression of FOXM1 in HCT116 cells (Figure 5(a)). In contrast, the expression of FOXM1 remained unchanged in response to HGF alone ( Figure 5(b)). Although HGF treatment did not alter the total level of FOXM1, the phosphorylation of FOXM1 was enhanced by HGF treatment ( Figure 5(b)). Given the stimulation of FOXM1 expression by activin and HGF, we asked whether FOXM1 could modulate the expression of CXCR4 and AFP. Notably, FOXM1 overexpression increased CXCR4 and AFP protein and mRNA levels in HCT116 cells, whereas FOXM1 knockdown reduced CXCR4 and AFP expression (Figures 4(a) and 4(b)). Tese results suggest that FOXM1 is involved in the upregulation of CXCR4 and AFP by activin activity and HGF.

Activin Induces FOXM1 Expression through SMAD2.
It has been reported that activin can bind to SMAD2 [18]. We speculated that activin might increase FOXM1 expression through SMAD2. Interestingly, the knockdown of SMAD2 abrogated the upregulation of FOXM1 by SMAD2 ( Figure 5(c)), indicating the importance of SMAD2 in modulating FOXM1 expression.

Discussion
Te precise mechanism of CRC liver metastasis has yet to be determined. CSCs play an important role in primary tumor initiation, development, and metastasis [21]. Many regulators of CSCs have been found to be involved in CRC liver metastasis [22][23][24]. For instance, Yao et al. [24] reported that FBXW11 promotes stem-cell-like properties and liver metastasis in CRC. Sequential treatment with activin and HGF has been shown to induce the generation of liver progenitorlike cells [8,9]. Rodrigues et al. [25] reported that activin contributes to the dediferentiation of lung carcinoma cells into cancer stem cells. Liu et al. [10] reported that activin is implicated in colorectal cancer stem cell self-renew and tumor progression. Similarly, HGF can enhance the stem cell-like potential of cancer cells [26]. In support of the link between cancer stemness and metastasis, our data show that the 2 stem cell regulators, activin and HGF, can promote CRC invasion and liver metastasis. It has been reported that activin and HGF can induce the generation of liver progenitor cells from human embryonic stem cells, which is characterized by the expression of Canadian Journal of Gastroenterology and Hepatology CXCR4 and AFP [6]. Interestingly, our data show that activin and HGF can also upregulate the expression of CXCR4 and AFP in CRC cells. Moreover, CXCR4 and AFP expression is increased in CRC liver metastases. Tese results suggest that the prometastatic efects of activin and HGF are associated with the induction of CXCR4 and AFP expression. Transcriptome analysis identifes many genes that are deregulated by sequential activin and HGF treatment. FOXM1 ranks among the most diferentially expressed genes. In this study, we focus on FOXM1 because it is involved in cancer stemness and metastasis [27][28][29]. Dysfunction of FOXM1 occurs in several malignant cancers and is highly correlated with poor prognosis and enhanced metastasis [13][14][15]. Te FOXM1-regulatory network is a critical predictor of poor prognosis in 18,000 cancer cases across 39 human malignancies [30]. Overexpression of FOXM1 can promote cancer cell migration and invasion and establish a premetastatic niche [31]. Regarding CRC biology, FOXM1 expression levels correlate with cancer progression, lymph node, liver metastasis, and high TNM stages [31]. Moreover, FOXM1 has been associated with stem celllikeself-renewal in CRC [32]. Several studies have revealed the regulation of FOXM1 by activin and the HGF pathway [33,34]. Activin A reduces FOXM1 expression in rat hepatocytes [33]. FOXM1 binds to the Met gene promoter and enhances HGF/Met signaling in pancreatic ductal adenocarcinoma cells by transcriptionally increasing Met expression [34]. Our data show that the expression of FOXM1, similar to CXCR4 and AFP, is upregulated in liver metastases of CRC. Knockdown of FOXM1 reduces CXCR4 and AFP expression levels in HCT116 cells. Most importantly, depletion of FOXM1 blocks liver metastasis of CRC cells with activin and HGF treatment in nude mice. Taken together, our data validate that FOXM1 is essential for the prometastatic efect of activin and HGF on CRC cells.
FOXM1 is a ubiquitously expressed transcription factor and has the capacity to trigger the transcription of a large number of genes [15][16][17]. Protein phosphorylation is required for FOXM1 function [16]. Joshi et al. [35] reported that MELK-mediated FOXM1 phosphorylation plays an essential role in the proliferation of glioma stem cells. Our data show that activin alone increases the expression of FOXM1 in CRC cells. Moreover, HGF alone enhances the phosphorylation of FOXM1 without altering the total level of FOXM1. Tese results provide a rationale for the treatment of CRC cells with sequential activin and HGF.
Biochemical studies further demonstrate that SMAD2 is implicated in the activin-induced upregulation of FOXM1. AFP expression can be induced by CXCL12/CXCR4 signaling, which is blocked by an NF-κB inhibitor. In glioblastoma multiforme cells, CXCL12/CXCR4 upregulates FOXM1 expression, promoting tumor migration and invasion [36]. Our results support that FOXM1 can directly increase CXCR4 expression in CRC cells by binding to the promoter of CXCR4. Hence, FOXM1 and CXCL12/CXCR4 may form a positive feedback loop in the CRCLM; however, this hypothesis requires further investigation. In rat hepatocytes, activin A strikingly downregulates FOXM1 expression. Our results demonstrate that activin activates and increases FOXM1 expression. Te opposite efects of activin on FOXM1 induction may be attributed to the diferent selected cell lines, suggesting that the functions of activin and FOXM1 in mature hepatocytes radically difer from those in CRC cells. CXCR4 is a marker of normal and malignant stem cells, and CXCL12/CXCR4 signaling is a crucial regulator of stem cell trafcking and cancer cell metastasis [37]. CXCR4-overexpressing CRC cells have been Relative protein level (e) Figure 6: FOXM1 binds to the promoter of CXCR4, and CXCR4 regulates AFP expression through the NF-κB pathway in CRC cells. (a) Luciferase reporter assays performed using the reporter constructs harboring diferent lengths of CXCR4 promoter fragments. * P < 0.05. (b) ChIP assays demonstrate that FOXM1 is enriched in the promoter of CXCR4. (c-d) Western blot and qPCR analysis of the efect of CXCR4 ligand CXCL12, CXCR4 inhibitor AMD3100, or NF-κB inhibitor NBD peptide on the expression of AFP. * P < 0.05vs. control. (e) Western blot analysis of the efect of CXCR4 ligand CXCL12 or CXCR4 inhibitor AMD3100 + CXCL12 on the expression of nuclear p65.
In conclusion, our results indicate that sequential treatment with activin and HGF promotes CRC invasion and liver metastasis. FOXM1 mediates the prometastatic efects of activin and HGF on CRC, which involves the induction of CXCR4 and AFP. Our work provides novel therapeutic targets for inhibiting liver metastasis of CRC.

Data Availability
All the data used to support the fndings of this study are included within the article.

Ethical Approval
Te study was conducted in accordance with the Declaration of Helsinki. Te study was approved by the Medical Ethics Committee of Bishan Hospital of Chongqing. Signed informed consent was obtained from the patients or their family members.