Comprehensive Genomic Profiling Identifies FAT1 as a Negative Regulator of EMT, CTCs, and Metastasis of Hepatocellular Carcinoma

Introduction

Liver cancer ranks sixth in relation to new cancer cases and fourth in the most prevalent cause of death worldwide.¹ According to reports, only 12% of patients with liver cancer in China survive for 5 years.² Approximately 80% of patients with hepatocellular carcinoma (HCC) have a history of chronic hepatitis B virus or hepatitis C virus infection.³ Patients with advanced HCC have a high rate of recurrence and metastasis, together with the limited benefit of conventional treatment and a poor prognosis.⁴

The epithelial mesenchymal transition (EMT) is required for metastasis in various cancers.^5,6 The prognosis of patients with cancer is impacted by the phenotypic shifting of cells, which promotes drug resistance and tumor recurrence.⁷

Circulating tumor cells (CTCs) play a critical role in establishing the spread and relapse of HCC.⁸ CellSearch^® (Menarini-Silicon Biosystems) is the first FDA-approved device for automated CTC detection and involves using anti-epithelial cell adhesion molecule (EpCAM) antibodies.⁹ Despite the excellent CTC isolation effectiveness of CellSearch^®, in epithelial malignant tumors, certain EMT-phenotype CTCs are very aggressive and can evade the positive enrichment methods for epithelial markers (such as EpCAM and CK).¹⁰ HCC CTCs express both epithelial markers, including EpCAM¹¹ and CK19,¹² and mesenchymal markers, including vimentin.¹³ Here, we optimized the CTC isolation method by selecting combinations of antibodies against various surface markers on HCC CTCs to prevent their escape during the capture process. The quantity of CTCs identified in patients with HCC is significantly associated with their TNM stage.¹⁴ Therefore, we enrolled ten patients with advanced HCC to participate in this study. Previous reports have demonstrated that the FAT1 protein inhibits the development of malignancy by regulating the EMT process in tumor cells where the FAT1 gene functions as a tumor suppressor gene;^15–17 however, Zhu et al demonstrated that FAT1 and POU2F1 mediate the growth and metastasis of HCC cells.¹⁸ Meng et al suggested that, in conjunction with GPC3, the expression of genes associated with EMT is co-regulated by FAT1 in HCC cells, and thus promotes HCC cell metastasis.¹⁹ This study indicated that the effect of FAT1 on cancer progression may be cell type or microenvironment specific. It is likely that HCC cells in hypoxia in synergy with GPC3 can promote cell metastasis and progression. However, the role and mechanism of FAT1 action on HCC CTCs for distal metastasis of CTCs in the blood microenvironment remains to be clarified. Therefore, we interrogated CTCs which obtained from HCC patients and then found that FAT1 played a negative regulatory role in metastasis of HCC CTCs. In addition, we found that FAT1 expression was higher in tumor tissues in 6 patients, however, the remaining patients had lower FAT1 expression in the tumor tissues when compared with the adjacent non-tumor tissues. HCC cells under hypoxic conditions promoted tumor migration and growth through upregulation of snail and vimentin and downregulation of E-Cadherin by the combined action of FAT1 and GPC3.¹⁹ FAT1 also promoted the anti-apoptotic ability of HCC and thus also promoted the progression of HCC.²⁰ On the other hand, knockdown of FAT1 also led to rapid loss of basal cell polarity as well as adhesion and tight junctions, thus promoting tumor progression.²¹ In this study, we captured CTCs from HCC patients using our modified CTC capture method,²² and extracted HCC tissues and CTC DNAs for next generation sequencing (NGS). We then combined the sequencing data from tissue and CTC to screen the FAT1 gene of interest, and cultured the HCC cell lines in a simulated blood environment and collected the treated tumor cells with magnetic beads. We found that the migration and invasion ability of the treated tumor cells were enhanced after knocking down FAT1, suggesting that interfering with FAT1 may promote metastasis of HCC CTCs.

Materials and Methods

Patient Tissues and Peripheral Blood Samples Collection

Ten patients were diagnosed with HCC at the Shanghai Sixth People’s Hospital. This study was conducted in accordance with the guidelines of the Declaration of Helsinki. All patients provided signed informed consent before sample collection. Peripheral blood was collected in 10-mL EDTA vacutainers before surgery, 7.5 mL of blood per case was centrifuged at 2500 rpm for 10 min within 24 h of collection, and the plasma layer was separated and transferred to EP tubes. After surgery, patient paired tumor and non-tumor tissues were also made into paraffin-embedded tissue samples. This study was approved by the ethics committee of Shanghai Jiao Tong University Affiliated Sixth People’s Hospital (No. 2022–020; China Clinical Trial Registration Center - Registration No. ChiCTR2200055847), and the patients signed the written consent after receiving oral and written information.

Cell Culture

The cell lines of Hep3B and MHCC97H were purchased from the Shanghai Cell Bank of Chinese Academy of Sciences. Hep3B and MHCC97H cells were cultured in Dulbecco modified Eagle medium (DMEM, Gibco, USA) containing 10% fetal bovine serum (FBS; HyClone, Logan, UT, USA) in a humidified incubator containing 5% CO₂/95% air at 37°C. After magnetic separation, the cells were added to simulated blood²³ cultured for 72 h, during which 20 ng/mL of TNF-α was added, and finally the cells were collected for subsequent in vitro and in vivo experiments.

Western Blot

Protease inhibitors (Roche, Switzerland) and lysis buffer (RIPA, Beyotime, Shanghai, China) were used to lyse whole cells on ice. The lysate was centrifuged for 15 min at 12,000 g and 4°C, and the di-octanoic acid (Sigma) technique was used to determine the content of the supernatant. For Western blot, 50 µg of protein was separated by electrophoresis on NUPAGE 10%–12% Transferring Bis-Tris gels (Invitrogen, CA, USA) before transferring to polyvinylidene difluoride (PVDF, 0.45 µm) membranes with a continuous current of 350 mA for 70/120 min. After being blocked for 2 h in a TBST buffer that included 5% skim milk powder and Tris-buffered saline, the membranes were incubated with the following specific antibodies at 4°C overnight: p-ERK1/2, ERK1/2, E-cadherin, N-cadherin, vimentin, and snail (Abcam, Cambridge, UK). Horseradish peroxidase-conjugated secondary antibodies (Sigma) were used to examine the blots, and a LAS4000 device was used to detect chemiluminescence (Fuji). Beta (β)-actin (Proteintech, Group, Wuhan, China) was applied as a loading control. The assays were conducted in triplicate.

Plasmid Constructs and Transfection

We used two lentiviral vector sequences (5′-GCAGCTGGAGAATATGATATT-3′, 5′-GCAACCGGCTCTCTCTATACT-3′) and a control sequence (5′-TTCTCCGAACGTGTCACGTTTC-3′) named FAT1sh1, FAT1sh2, and FAT1NC, respectively. The shRNAs were implanted into the vector pGLV-FAT1-Puro and co-transfected into Hep3B and MHCC97H cells. FAT1-overexpressed cell clones were generated by stably transfecting Hep3B cells with the lentivirus expression vector. Next, 2 µg/mL of puromycin was used to select stable cell lines over 2 weeks (Invitrogen). Hep3B and MHCC97H cells were cultured in DMEM (10% FBS) and placed in 4-well glass bottom culture plates with 1.5×10⁴ cells/well. After 24 h, cells were transfected with plasmids (1 µg per well) using Lipofectamine 2000 transfection reagent (Invitrogen, Carlsbad, CA, USA; 2 µL per well) in accordance with the manufacturer’s instructions. The new medium was applied after 4 h, and cells were cultured for 36 h before being exposed to 1.5 µg/mL of cytochrome B (Sigma, St. Louis, MO, USA) for 12 h.

Transwell Migration and Invasion Assays

During the Transwell migration experiment, the upper chamber of the Transwell was loaded with 2×10⁵ cells in 300 μL serum-free DMEM, and the bottom chamber was loaded with 500 μL of serum-free medium. The cells were fixed with 4% formalin solution, stained with Gentian Violet, and incubated for 4 h at 37°C. Non-migrating cells were retained at the top layer and removed by wiping them with a cotton swab. Ten randomly selected microscopic regions were used to count the average number of cells that passed through the filter. Transwell chambers (Corning) were also used for in vitro invasion tests. Briefly, the cells were coagulated for 1 h at 37°C before covering the Transwell system with 300 μL of serum-free DMEM and cells (2 × 10⁵). Next, 500 μL of the serum-free medium was added to the lower layer and cultured for 24 h at 37°C. Crystal violet was used to stain fixed cells, which were then counted in ten randomly selected microscopic regions.

Animal Experiments

This study was approved by the ethics committee of Shanghai Jiao Tong University Affiliated Sixth People’s Hospital (No. 2022–020), all animal experiments were performed in accordance with the National Institutes of Health Guidelines for the Care and Use of Laboratory Animals and regulations of the ethics committee of Shanghai Jiao Tong University Affiliated Sixth People’s Hospital. For in vivo experiments, 5×10⁵ FAT1 knockdown stable Hep3B or Hep3BNC cells and FAT1-WT overexpressing MHCC97H or MHCC97HNC cells were administered subcutaneously into the right armpit of male BALB/c nude mice. The tumor size was measured every 7 days using calipers, and the tumor volume was calculated according to the following formula: volume = 1/2 × (width² × length). The tumors were removed and measured after the mice were euthanized and sacrificed after 5 weeks.

RNA Isolation and Real-Time Quantitative PCR (qRT–PCR)

Cells cultured in the growth medium for 48 h were serum-starved for 24 h before total RNA extraction using TRIzol Reagent (Invitrogen) following the manufacturer’s instructions. The PrimeScriptTM RT reagent kit (Toyobo, Osaka) was used to reverse-transcribe 1 g of total RNA into cDNA after determining the RNA concentration. Subsequently, cDNA was used for qRT–PCR using the SYBR-Green PCR Master Mix (Takara, Osaka). qRT–PCR was performed on Hep3B, Hep3BNC, MHCC97H, MHCC97HNC, and the matching FAT1-depleted cells to determine alterations in the relevant genes. GAPDH was used as a control.

RNA Sequencing

RNA sequencing was performed on Hep3B, Hep3BNC and corresponding FAT1-depleted cells to monitor changes in their downstream genes. Huzhou Lieyuan Medical Laboratory Co. Ltd. handled the extraction of total RNA and sequence analysis. GSEA was used to evaluate the different genes in comparison with FAT1 knockdown and normal cells.

Hematoxylin and Eosin (HE) Staining

HE staining was used to confirm the presence of lung metastatic nodules in paraffin blocks made from 10% buffered formalin-fixed tumor and pulmonary lung tissue samples.

Immunohistochemistry (IHC)

Serial sections of the FFPE specimens were generated at a thickness of 5 μm. Antigen repair was performed in 0.01 M citrate buffer (pH 6.0) in a pressure cooker for 30 min, then 3% hydrogen peroxide for 5 min of pretreatment. After fixing with 10% buffered formalin, the tissue slices were incubated for 24 h at 4°C with anti-FAT1 (Sigma, 1:300 dilution), E-cadherin (CST, USA), Ki-67 (CST, USA), N-cadherin (Maixing, China), Snail (CST, USA), p-ERK1/2 (CST, USA), Vimentin (CST, USA). Sections were washed with phosphate buffered saline (PBS) before being exposed to the secondary antibody for 20 min at 37°C. The levels of FAT1 protein expression were evaluated with Aperio Cytoplasma software by IHC.

CTC Assay

The plasma in EP tubes was placed on a magnetic separation rack and left for 15 min after being stirred seven times over a 28 min incubation period at 25°C, before aspirating the supernatant. The samples were washed twice with 2 mL PBS, before adding 10 μL DAPI, CK19-FITC, and CD45-PE staining solutions sequentially and incubating for 30 min at 25°C. Following incubation, the samples were distributed equally on anti-dislodging slides, dried naturally, and photographed and counted under a fluorescent-microscopic device. Captured by Ep-LMB and CK⁺/CD45⁻/DAPI⁺ were defined as epithelial CTC (^ECTC), and captured by Vi-LMB and CK⁺/CD45⁻/DAPI⁺ were defined as mesenchymal CTC (^MCTC). The ^MCTC ratio was defined as the proportion of ^MCTCs to all CTCs in each patient sample (7.5 mL).

Identification of Genetic Variants

Ten FFPE and CTC samples from patients with HCC were submitted to Huzhou Lieyuan Medical Laboratory Co. Ltd. A total of 610 genes were included in the assay. The NGS assay of FFPE and CTC samples was conducted following previously published protocols.^24,25

Statistical Analysis

SPSS statistical software (version 22.0, IBM SPSS, USA) was used to analyze the data. Results are shown as mean ± SEM, and three independent tests were conducted. FAT1 protein levels were examined in tumors and matched non-tumors using the rank sum test. The data between the two groups were evaluated using an unpaired t-test, and differences from multiple groups were examined using one-way ANOVA. P-values < 0.05 were considered statistically significant.

Results

Identification of FAT1 from Genomic Analysis of CTC and Tissue DNAs

We optimized the CTC capture method using EpCAM-supplemented CTC enrichment and detection technology. CTCs and tumor tissues from patients with HCC were analyzed for genetic mutations, as shown in the workflow (Figure 1A). The mean age of 10 patients was 62 years (range, 46–79 years), and the number of patients at clinical stage IIIB and IVA was 8 (80%) and 2 (20%), respectively. We performed immunofluorescence analysis with anti-pan keratin antibodies to determine the number of CTCs in the blood of patients with HCC. CTCs were detected in all ten patients, with a median CTC count of 6 CTCs/mL (range: 4–9). Representative images of CTCs from patients are presented in Figure 1C. NGS genomic profiling of CTCs and tumor tissues from patients with HCC is shown in Figure 1B. Mutations were detected in all patients with HCC, with a median number of mutations detected in peripheral blood CTCs and HCC tissues of 7 (range: 1–29) and 6.5 (range: 2–19), respectively. CTC and histogenetic analyses showed that TP53 was the most commonly altered gene, followed by ALK, ARID1A, and FAT1. TP53 and ARID1A are tumor suppressor genes, while ALK is a proto-oncogene, all of which are associated with malignancy.^26–28 Combinatorial analysis of the genomic mutation profiles obtained from CTC and tissue DNA showed some common genetic alterations, including FAT1 (Figure 1D and E). We next examined FAT1 protein expression in malignant and adjacent normal tissues of ten patients with HCC using IHC analysis. We found that FAT1 expression was higher in tumor tissues in 6 patients, however, the rest 4 patients had lower FAT1 expression in the tumor tissues when compared to the adjacent non-tumor tissues. Overall, the levels of FAT1 protein expression in cancer tissues were lower than those in adjacent normal tissues in this section (Figure 1F–G). Microarray datasets GSE365²⁹ were downloaded from Gene Expression Omnibus (GEO) database. Reading expression profile data was conducted in R environment using GEOquery package.³⁰ The GSE365 dataset contained 87 HCC tissue samples and 87 non-cancerous samples. By exploring this dataset, we found that FAT1 expression was higher in tumor tissues and metastases than normal tissues in some hepatitis B virus-positive HCC patients, while the remaining patients showed the opposite pattern of FAT1 expression, which was consistent with our above results. These results indicated that FAT1 expression showed a high degree of heterogeneity in the HCC population.

Figure 1 Genomic analysis identified FAT1. (A) Flow chart of the protocol for genomic analysis of CTC and tissue DNA. (B) CTC and tissue genomic variants in patients with HCC. (C) Fluorescence images of captured CTCs. (D) The top five genes with the highest mutation frequency. (E) Combined analysis of CTC and tissue gene mutations. (F) Analysis of FAT1 protein levels in ten patients with HCC (n = 10, rank sum test, *P < 0.05). (G) Representative IHC staining for FAT1 from HCC tissues and adjacent non-tumor tissues. Upper row (I, from left to right): FAT1 expression was higher in tumor tissues than in adjacent non-tumor tissue; lower row (II, from left to right): FAT1 expression was lower in tumor tissues than in adjacent non-tumor tissues.

FAT1-Knockdown-Induced EMT Promotes HCC Cell Migration and Invasion

RT-qPCR and Western blot were used to evaluate the mRNA and protein concentrations of FAT1 in 2 HCC cell lines (Figure 2A), before clarifying the true effects of FAT1 knockdown and overexpression in these cells individually (Figure 2B and C). Knockdown of FAT1 enhanced cell division (Figure 2D) and the establishment of colonies (Figure 2E) in Hep3B and MHCC97H cell lines, while overexpression of FAT1 significantly slowed cell growth and colony formation (Figure 2F–G). Next, we evaluated the impact of FAT1 on the invasive cell motility of cultured cells. Interestingly, FAT1 knockdown markedly facilitated the colonization and mobility of Hep3B and MHCC97H cells. FAT1 knockdown in Hep3B and MHCC97H cells significantly enhanced tumor cell colonization and mobility, as assessed by the Transwell assay (Figure 3A) and wound healing migration assay (Figure 3B). However, we observed the opposite trend in exogenous FAT1-expressing cells (Figure 3C). We also measured the level of EMT markers in FAT1 knockdown cells (Figure 3D). E-cadherin mRNA expression was markedly reduced by FAT1 knockdown, which in turn caused N-cadherin, vimentin, and snail mRNA levels to increase. We further validated this trend with Western blotting (Figure 3E).

Figure 2 The function of FAT1 affects the proliferation and colony formation of HCC cells. (A) FAT1 mRNA expression and protein level profile in two HCC cell lines were evaluated using RT-PCR and Western blotting. (B) In Hep3B and MHCC97H cells, qPCR was used to determine the knockdown effectiveness of FAT1. (C) The effectiveness of FAT1 overexpression in Hep3B and MHCC97H cells was validated by Western blotting and RT-PCR. (D) Hep3B and MHCC97H cell proliferation are stimulated by FAT1 knockdown. (E) The FAT1-knockdown group had considerably more colonies in HCC cells than the control group. (F) The MTT experiment showed that overexpressing FAT1 prevented Hep3B and MHCC97H cells from proliferating. (G) Hep3B and MHCC97H cell colony formation were inhibited by FAT1 overexpression. All the data obtained is presented in the format of three separate experiments and is expressed as the mean ± standard deviation. **P < 0.01.

Figure 3 In HCC cells, FAT1 dysregulation of EMT facilitates invasion and migration. (A) Hep3B and MHCC97H cells treated with lentiviruses encoding FAT1-shRNA in transwell migration and invasion studies. Hep3B and MHCC97H cell migration and invasion were markedly enhanced by FAT1 knockdown. (B) A wound healing experiment was used to examine the cell migration of Hep3B and MHCC97H cells treated with lentiviruses encoding FAT1shRNA. (C) FAT1 overexpression decreased cell migration and invasion in co-culture assay tests of Hep3B and MHCC97H cells treated with lentivirus expressing FAT1-WT. (D) RT-PCR was used to determine the relative mRNA levels of E-cadherin, N-cadherin, vimentin, and snail in cells that either overexpressed or knocked down FAT1. GAPDH served as the control. (E) Levels of EMT-related protein were measured using a Western blot assay. Beta (β)-actin served as a control. Independent experiments were repeated at least three times. *P < 0.05, **P < 0.01.

FAT1 Inhibits Proliferation Associated with the MAPK/ERK Signaling Pathway

We next conducted RNA sequencing of FAT1 knockdown Hep3B to identify alterations in their gene transcription levels. Differential genes (P<0.05) were enriched in processes by gene ontology analysis, including focal adhesion, regulation of immune response process, cell–substrate junction, and location (Figure 4A). Bioinformatics analysis of these genes indicated enrichment in cancer-related pathways (Figures 4B and S1). Notably, some genes in the tight junction signaling pathway (RAB13, etc.) associated with EMT showed varying levels of upregulation. We hypothesized that FAT1 regulates these signaling pathways to stimulate cell division and EMT. To validate this, the effects of FAT1 on MAPK/ERK signaling activity were examined. As anticipated, the amount of phosphorylated ERK1/2 (p-ERK1/2) was dramatically elevated by the suppression of FAT1. To block MEK1/2 activity in the knockdown group, we used the highly specific inhibitor U0126, then performed an MTT analysis and an EMT marker test to illustrate that FAT1 controls cell proliferation and EMT through the signal pathways. U0126 restricted the knockdown impact of FAT1, preventing continued cell proliferation (Figure 5A) and dysregulating the levels of EMT-associated proteins (Figure 5B). Additionally, U0126 inhibited the ability of Hep3BFAT1sh and MHCC97HFAT1sh cells to migrate and invade (Figure 5C).

Figure 4 Critical cancer pathway components were altered in FAT1 knockdown cells compared to control cells, as analyzed by RNA sequencing. (A) GSEA of the significantly enriched GO terms indicated that the functional annotation items of crucial genes with differential expression were enriched in focal adhesion, regulation of the immune response process, cell–substrate junction and other aspects. The figure shows the comparison of Hep3BNC and Hep3BFAT1sh. (B) GSEA of the significantly enriched KEGG terms showed that crucial signaling pathways were altered in FAT1 knockdown cells, including the MAPK signaling pathway, the Wnt signaling pathway, the Notch signaling pathway, and the tight junction signaling pathway.

Figure 5 The FAT1-knockdown group promoted the process of EMT in tumor cells through MAPK/ERK signaling pathway. (A) The proliferation of tumor cells in the FAT1-knockdown group treated with U0126 was significantly slowed down. (B) Expression levels of EMT-related proteins were detected in U0126-treated and untreated cells, respectively. (C) Migration and invasion assays of cells in the FAT1 knockdown group treated with U0126. **P < 0.01.

Inhibitory Effects of FAT1 in vivo Confirm the Effects in vitro

To further demonstrate the tumor-suppressive effects of FAT1 in HCC, we injected 5×10⁵ FAT1 silenced Hep3B or Hep3BNC cells, and 5×10⁵ FAT1-WT overexpressing MHCC97H or MHCC97HNC cells into the right armpit of mice. Two weeks later, we saw that the FAT1-knockdown group markedly promoted tumor development compared to the control group (Figure 6A), whereas overexpression of FAT1-WT significantly inhibited tumor growth (Figure 6B). The strong growth-promoting impact of FAT1 knockdown and the inhibitory effects of FAT1-WT overexpression cells were further verified by IHC of Ki-67, providing evidence that FAT1 may impact tumorigenesis in HCC in vivo. Additionally, we measured the expression levels of relevant proteins in paraffin-embedded sections of nude mice tumor tissues by IHC. Compared to the in vitro study, similar expression patterns were seen in the in vivo experiments, indicating that FAT1 may control EMT during the development of HCC tumors; depletion of FAT1 induced EMT, whereas the overexpression of FAT1 in HCC enhanced the inhibition of EMT (Figure 6C and D). In comparison to the MHCC97HFAT1WT group, the ^MCTC ratio was greater in the Hep3BFAT1sh group (Figure 6E and F), and the number of lung metastases per section was markedly elevated in the Hep3BFAT1sh group (Figure 6G and H).

Figure 6 FAT1 knockdown promoted the EMT process and accelerated tumor development in nude mice. (A) FAT1 knockdown promoted tumor growth and significantly increased tumor size (left) and tumor weight (right) in the FAT1 knockdown group compared to the control group. **P < 0.01. (B). Compared with the control group, the tumors in the FAT1-overexpressing group were significantly reduced in size (left) and tumor weight (right). **P < 0.01. (C) Representative IHC staining for Ki-67, FAT1, p-ERK1/2, E-adherin, N-adherin, vimentin and snail from the control group and FAT1 knockdown group (bar, 100 μm). (D) Representative IHC staining for Ki-67, FAT1, p-ERK1/2, E-adherin, N-adherin, vimentin and snail from the control group and FAT1-overexpressing group (bar, 100 μm). (E) Representative CTC images from two groups of mice. (F) ECTC and MCTC count distribution in two groups of mice. (G) Lung metastatic nodules of mice were examined by H&E staining. (H) Average numbers of individual pulmonary metastases. The FAT1-knockdown group had significantly more pulmonary metastases compared to the FAT1-overexpressing group. ***P < 0.001.

Discussion

CTCs are the keys to the mechanism of hematogenous metastasis of malignant tumors.³¹ In EMT, the replacement of the epithelial phenotype by a mesenchymal and migratory phenotype is a critical step in the spread of tumors.³² HCC cells with EMT phenotype are more prone to metastasis.³³ Our modified CTC capture method used novel multi-labeled lipid magnetic nanoparticles to isolate HCC CTCs. Using combinatorial analysis of genomic mutation profiles obtained from HCC tissues and CTCs, we identified high-frequency mutations in FAT1 in ten patients with HCC. Based on previous studies,³⁴ we hypothesized that FAT1 has a tight connection to HCC initialing, the EMT process of tumor cells, and distal metastasis of tumors.

The transmembrane protein FAT1 is important in tumors because it governs EMT, cell proliferation, and actin kinematics.³⁵ Focusing on the relationship between FAT1 and spread of HCC cells, and in order to simulate the seeding of HCC CTCs in blood, we purposely cultured Hep3B and MHCC97H cells in a simulated blood environment and induced the cell EMT. Then, we further sorted out the tumor cells by magnetic beads for subsequent studies. Our results showed that FAT1 could influence HCC cell proliferation in vitro, and regulate the EMT process, thus inhibiting HCC invasion and metastasis, supporting earlier discoveries.^36–38 However, our results differ from those of Fu et al³⁹ although the same cell line was used, tumor cells in this study were treated and purified accordingly to simulate the blood environment. We have noted that previous studies on the tumorigenic role of NF-kB in HCC by Pikarsky et al⁴⁰ and Maeda et al⁴¹ showed opposite results. Those findings indicated that the underlying mechanisms about the metastasis of HCC remain to be further explored.

Next, GSEA demonstrated that the MAPK/ERK, Wnt, and Notch signaling pathways were significantly altered after FAT1 knockdown. Cancer cells often exhibit persistent MAPK/ERK signaling pathway activation.⁴² Similarly, alteration in other signaling pathways is also believed to play a role in the development of tumors and the proliferation of cancer cells in various cancers.⁴³ Notably, we detected alterations in the TIGHT JUNCTION signaling pathway, in which RAB13 could inhibit the expression of PKA, thus weakening the intercellular junctions and promoting EMT in tumors,⁴⁴ while RhoA, under the regulation of CUX1, further affected the migration and invasion of tumor cells.⁴⁵ The relationship among RAB13, RhoA, CUX1, and EMT in HCC remains to be explored. Consistent with previous reports, our results suggest that the effects of FAT1 on HCC cells involve the aberrant activation of multiple signaling pathways.^46,47

To determine whether there is a link between FAT1 inactivation and critical signaling pathways, we discovered that HCC had low levels of FAT1 expression, which activated the MAPK/ERK pathway. Meanwhile, we observed that U0126 eliminated FAT1 knockdown effects to inhibit EMT. Moreover, the ratio of ^MCTC and the quantity of pulmonary metastases were dramatically increased in the FAT1-knockdown group compared with the FAT1-overexpressing group in vivo. Deregulation of FAT1 function promoted tumorigenesis in HCC, confirming our hypothesis. However, there are some limitations in our study. The number of enrolled cases was small and all participants were advanced HBV-associated HCC patients. We will continue to expand the sample size of clinical research and improve the in vivo experiments to further validate the results and methodology of the study.

Conclusions

In summary, we investigated the effects of FAT1 on EMT and metastasis of HCC from the perspective of CTCs for the first time. We used magnetic enrichment method to analyze and screen candidate genes, and applied a novel cell culture and purification strategy to conduct this study. Our research demonstrates a connection between the FAT1 tumor suppressor and the EMT process in HCC. Moreover, suppressing FAT1 facilitates the EMT process, which promoted the metastasis of HCC. These discoveries shed light on the processes underlying the onset and progression of HCC.