FAT4 regulates the EMT and autophagy in colorectal cancer cells in part via the PI3K-AKT signaling axis

In our study, 100 formalin-fixed, paraffin-embedded CRC samples and paired adjacent benign tissues were randomly collected from 72 men and 28 women with a mean age of 60 years at the Second Affiliated Hospital of Nanchang University, Nanchang, China. Informed consent was obtained from the study participants prior to their participation, and the study was highly supported by the Medical Research Ethics Committee of the Second Affiliated Hospital of Nanchang University.

Human CRC cell lines (HCT116, LOVO and SW480) from the Chinese Type Culture Collection were cultured in RPMI 1640 medium (RPMI 1640, Gibco, USA) supplemented with 10% fetal bovine serum (FBS, Gibco, USA) and 2% penicillin/streptomycin (HyClone, Shanghai, China). All the cell lines were incubated under standard conditions at 37 °C in an atmosphere contaning 5% CO2.Once the cells reached approximately 80% confluence, they were passaged by trypsinization.

For the induction and inhibition of autophagy, the cells were treated with 250 nM Torin 1 (Sigma Aldrich, Missouri, USA) and 2 μM MHY1485 (Sigma Aldrich, Missouri, USA), respectively, and to regulate PI3K activity, the cells were treated with 150 nM wortmannin (Sigma Aldrich, Missouri, USA) and 50 μg/mL 740Y-P (Cayman, Michigan, USA).

We utilized the TRIzol reagent (Invitrogen, Carlsbad, CA, USA) for the extraction of total RNA from tissue samples and cultured cells. Quantitative RT-PCR was performed using a PrimeScript RT reagent kit (TaKaRa, Dalian, China) and SYBR Premix Ex Taq (TaKaRa, Dalian, China). The following primer sequences were selected: FAT4, forward, 5’-TTGAAGGAAGGAGAACCCAT-3′, and reverse, 5’-TCGTCCAATAGTAAAGAGGC-3′.

The CRC tissue samples were fixed with formalin and embedded in paraffin. A slide was coated with an anti-FAT4 mouse polyclonal antibody (1:300 dilution; Santa Cruz, USA), incubated overnight at 4 °C, washed three times with PBS and submerged in the corresponding biotinylated secondary antibody (DAKO, Shanghai, China) for 30 min. Furthermore, 3,3′-diaminobenzidine (DAB) (DAKO, Shanghai, China) was administered to develop the peroxidase reaction, and the tissue sections were also stained with hematoxylin. The IHC intensity was scored as follows: 0, 1, 2, and 3 points indicated no staining, minimal staining, moderate staining and strong staining, respectively. The percentage of positive cells was determined using a previously reported method, and the cells were divided into four groups: no positive cells, < 10% positive cells, 10–50% positive cells and > 50% positive cells [4].

In brief, the cells were seeded in 96-well plates (1 × 10³ cells/well) and cultured for 48 h. An MTT dye solution (5 mg/ml, Sigma, NY, USA) was added to each well, and the cells were incubated for another 4 h. Dimethyl sulfoxide (150 μl) was subsequently added, and the plates were placed in an ultraviolet spectrophotometer to measure the OD value of each well at 490 nm. The level of cell proliferation was examined at 1, 2, 3, 4, 5, 6 and 7 days after seeding. Cell invasion and migration were assessed using Boyden chambers (Millicell, Millipore, USA). Initially, 2 × 10⁵ cells in serum-free medium were cultured in the upper chamber of a 24-well plate, and the corresponding medium supplemented with 10% FBS was placed in the lower chamber. After 36 h of incubation, the migrated cells were fixed with 4% paraformaldehyde for 10 min, stained with 0.1% crystal violet for 10 min, air dried and counted under a light microscope. The degree of cell invasion was similarly tested using inserts precoated with Matrigel (BD Biosciences, USA) to simulate circumstances requiring invasion. Each experiment was repeated three times.

Total protein from the CRC cell lines and tissues was isolated, and the protein concentrations were measured using the BCA protein assay kit (Pierce, IL, USA). The protein extracts were washed in lysis buffer, separated by 10% SDS-PAGE and transferred to polyvinylidene fluoride membranes (Millipore, Bedford, Massachusetts, USA). The membranes were blocked for 1 h in TBS-Tween (0.1%) with 2% BSA for the detection of phosphorylated proteins or in 5% skimmed milk for the detection of other proteins and then incubated with the following primary antibodies: FAT4 (1:300 dilution; Santa Cruz, USA), E-cadherin (1:1000 dilution; Santa Cruz, USA), N-cadherin (1:3000 dilution; Santa Cruz, USA), vimentin (1:1000 dilution; Santa Cruz, USA), Twist 1 (1:300 dilution; Santa Cruz, USA), β-catenin (1:5000 dilution; Santa Cruz, USA), LC3 (1:300 dilution; Santa Cruz, USA), P62 (1:300 dilution; Santa Cruz, USA), PI3K (1:300 dilution; Santa Cruz, USA), AKT (1:300 dilution; Santa Cruz, USA), p-AKT (1:300 dilution; Santa Cruz, USA), mTOR (1:300 dilution; Santa Cruz, USA), GSK-3β (1:300 dilution; Santa Cruz, USA), p-GSK-3β (1:300 dilution; Santa Cruz, USA), C-Myc (1:300 dilution; Santa Cruz, USA), unc-51-like kinase 1 (ULK1; 1:1000 dilution; Sigma-Aldrich, USA), p70(S6K) (1:1000 dilution; R&D Systems, USA), phospho-p70(S6K) (1:1000 dilution; R&D Systems, USA) and β-actin (1:1000 dilution; Beyotime, China). The primary antibodies were detected using specific secondary antibodies. The protein bands were visualized using an ECL detection reagent (Proteintech, Hubei, China), and the band densities were measured using imaging software.

Lentiviral constructs for shRNA-mediated FAT4 (LV-FAT4-shRNA-puromycin) silencing and FAT4 (LV-FAT4-puromycin) overexpression were acquired from Shanghai Genechem Co, Ltd., China. The FAT4 shRNA vector sequence was as follows: (forward) 5’-GCTTTGTATAAAGTGGAGATT-3′. The three above-mentioned cell lines were cultivated in six-well plates at a density of 0.6 × 10⁵ cells per well (20–30% density) on the day prior to lentiviral transduction. LV-FAT4-shRNA-puromycin was used for the transduction of SW480 cells at a multiplicity of infection (MOI) of 40 using polybrene (10 μg/ml), which enhanced the infection efficiency (Genechem, China). Additionally, LOVO and HCT116 cells were transfected with LV-FAT4-puromycin at a MOI of 20 (LOVO) and 10 (HCT116). Moreover, two nontarget negative control viruses, puromycin-LV (Genechem, China), were used to transduce cells using the above-described methods to neutralize the influence of the viral vector. After 12 h of incubation, the medium was removed, and the appropriate fresh medium was added. After 48 h of incubation, the appropriate concentrations of puromycin were added to the cell cultures to kill cells that had not been transduced with the virus. At the indicated time points, the cells were used for protein and mRNA analyses and other assays.

The cells were cultivated on appropriate six-well coverslips, immobilized with 4% paraformaldehyde for 20 min, permeabilized with 0.2% Triton X-100 (Beyotime, Zhenjiang, Jiangsu, China) and then blocked in 10% goat serum at 37 °C for 30 min. Primary antibodies targeting LC3 (1:300 dilution; Santa Cruz, USA) and E-catenin (1:300 dilution; Santa Cruz, USA) were added to the coverslips. After incubation at 4 °C overnight, the coverslips were washed three times in PBS and incubated with the appropriate secondary antibody (SA00009-2Cy3-goat anti-rabbit IgG for LC3 and E-catenin, Proteintech) for 2 h. The nuclei were then labeled with DAPI for 10 min, and images were captured using a confocal microscope (× 600).

For the TEM studies, the cells were exposed to RF-EMFs (1, 2, or 4 W/kg) for 24 h, harvested, washed twice with PBS (pH 7.4) at room temperature, and fixed with 2.5% glutaraldehyde. The cells were then washed three times with PBS, fixed with 1% osmic acid for 2–3 h, washed three times with PBS, dehydrated, embedded in paraffin, cut into 70-nm-thick sections using an Ultrathin slicing machine (Leica EM UC6; Leica Microsystems, Wetzlar and Mannheim, Germany), and stained with uranyl acetate-lead citrate. The cells were then observed using a transmission electron microscope (JEM1230, JEOL Ltd., Tokyo, Japan) for the detection of autophagic vacuoles.

In our study, eight-week-old male BALB/c nude mice were purchased from the Laboratory Animal Center of Wuhan. All the animal experiments adhered to the National Institutes of Health Guide for the Care and Use of Laboratory Animals and were performed using protocols approved by the Medical Research Ethics Committee of the Second Affiliated Hospital of Nanchang University. SW480 cells (2 × 10⁶) were transduced with a FAT4 shRNA vector (LV-FAT4-shRNA-puromycin), and HCT116 and LOVO cells were transduced with a FAT4-overexpression vector (LV-FAT4-puromycin) and suspended in 100 μL of basal medium. Corresponding negative control vectors were also transduced into all three cell lines. The cells were then subcutaneously injected into the right flank of the nude mice to induce the formation of xenograft tumors. The mice were allowed to recover for 8 weeks with normal feeding and watering, and the tumors were then excised for further analysis. Specifically, the tumor volume was calculated using the formula V (mm³) = length×width²/2, and the mice were then anesthetized, euthanized, and humanely disposed.

The data were analyzed using Prism 5 (GraphPad Software, San Diego, California, USA), and the values are presented as the means ± SEMs or SDs. A two-tailed unpaired t-test was used to compare two groups. SPSS 17.0 software (SPSS Inc., Cary, North Carolina, USA) was also used for the data analyses. A P value < 0.05 was assumed to indicate a statistically significant difference, and differences between P < 0.05 and P < 0.01 are denoted by single (*) and double (**) asterisks, respectively.

In a preliminary study, we estimated the FAT4 expression levels in 100 random CRC patients, and our immunohistochemistry results revealed that FAT4 expression was decreased in the nuclei of the CRC sections compared with the nuclei in the nonmalignant tissues (Fig. 1a). In addition, a qRT-PCR analysis showed that FAT4 expression was lower in CRC tissues compared with nonmalignant tissues (Fig. 1c), and a western blotting assay confirmed that FAT4 protein expression was low in CRC tissues (Fig. 1b). To further understand the underlying relationship between low FAT4 expression and CRC, we detected the expression of FAT4 in three CRC cell lines, HCT116, SW480 and LOVO. The results revealed that SW480 cells exhibited high expression of FAT4, whereas the other two cell lines showed relatively low expression (Fig. 1d). To elucidate the pathological effects of FAT4 in CRC cells, we examined the proliferation of these cells through MTT assays and found that the proliferation capacity of SW480 cells with suppressed FAT4 expression was higher than that of control SW480 cells; however, the overexpression of FAT4 in HCT116 and LOVO cells led to a decline in the proliferation ability of these cells (Fig. 1e).

In line with the results of our previous experiment, HCT116 and LOVO cell lines, which manifested relatively low nascent FAT4 expression, were transduced with FAT4-overexpressing constructs, and SW480 cells were transduced with shRNA to knockdown FAT4 expression. It is well accepted that E-cadherin, N-cadherin and vimentin are considered classical EMT markers, and the immunofluorescence and western blotting results demonstrated that E-cadherin expression was enhanced in FAT4-overexpressing HCT116 and LOVO cell lines and dysregulated in FAT4-knockdown SW480 cells compared with the levels in the corresponding untreated cells. In contrast, N-cadherin, vimentin, C-Myc, β-catenin and Twist1 expression showed an opposite trend (Fig. 2a-c, Fig. 3a-c). Furthermore, the function of FAT4 in gastric cancer was previously studied [3]. The EMT requires cancer cells to survive independently from the primary tumor site without a nutrient support system, and as a result, these cells might be more sensitive to autophagy [5]. Autophagy might play a crucial role in cell invasion and migration. To investigate this phenomenon, we further explored the function of FAT4 in autophagy. The overexpression of FAT4 significantly elevated the LC3-II/LC3-I ratio and ULK1 level and decreased P62 expression, as indicated by western blotting. In contrast, the knockdown of FAT4 in SW480 cells decreased the LC3-II/LC3-I ratio and ULK1 level and promoted the expression of P62 (Fig. 3a-c). Through immunofluorescence, we observed that FAT4-overexpressing HCT116 and LOVO cells showed strong LC3 expression, whereas FAT4-knockdown SW480 cells exhibited weak LC3 expression. E-cadherin showed the opposite results in the transduced cells: high expression in the FAT4-overexpresing HCT116 and LOVO cells and low expression in the FAT4-knockdown SW480 cells (Fig. 2a-c). Additionally, as observed by TEM, autophagosome formation was significantly induced in the HCT116 and LOVO cells transduced with the FAT4-overexpressing constructs, whereas the inhibition of FAT4 expression in SW480 cells resulted in a decreased formation of autophagosomes (Fig. 4a-c). Together, these findings provide strong evidence showing that FAT4 plays a crucial role in the EMT and autophagy in CRC.

Based on the above-described results, we hypothesized that FAT4 played a significant role in the EMT and autophagy to affect the tumor characteristics of CRC cells and that this activity partially relies on the PI3K-AKT signaling pathway. Based on these assumptions, we aimed to detect the relevant effectors within this pathway through western blotting. First, we evaluated PI3K expression in these three cell lines, and the results indicated that PI3K expression was notably downregulated in the two FAT4-overexpressing cell lines compared with the FAT4-knockdown SW480 cells (Fig. 5a-c). Furthermore, the level of p-AKT expression was reduced after FAT4 overexpression, and SW480 cells with suppressed FAT4 expression showed increased levels of this protein (Fig. 5a-c). In addition, as demonstrated by the western blotting assays, the FAT4-overexpressing HCT116 and LOVO cells exhibited higher levels of GSK-3β and lower levels of mTOR expression compared with those in the corresponding control cells. The level of p-p70S6K expression was reduced following FAT4 overexpression, and the suppression of FAT4 expression in SW480 cells increased the expression of p-p70S6K (Fig. 5a-c). However, the knockdown of FAT4 led to decreased GSK-3β and increased mTOR expression, and the phosphorylation of GSK-3β (p-GSK-3β) showed the opposite trend to GSK-3β, which revealed that the regulatory effects of FAT4 can be partially attributed to the PI3K-AKT-mTOR and PI3K-AKT-GSK-3β signaling pathways (Fig. 5a-c). Although the carcinostatic function of FAT4 is specific, further investigation is urgently needed to illustrate the mechanism through which FAT4 affects the EMT and autophagy in tumors via the PI3K-AKT signaling pathway. Thus, we applied wortmannin, a PI3K-AKT signaling pathway inhibitor, and the activator 740Y-P to the cells showing stable FAT4 knockdown or overexpression and evaluated the expression of PI3K, AKT, p-AKT, mTOR, GSK-3β, p-GSK-3β and β-catenin by western blot analysis. The results showed that in the FAT4-overexpressing HCT116 and LOVO cells, the expression levels of PI3K, p-AKT, mTOR, p-GSK-3β and β-catenin were increased after treatment with the PI3K activator 740Y-P, but these levels were reduced in the FAT4-knockdown SW480 cells after treatment with the PI3K inhibitor wortmannin (Fig. 6a-c). In contrast, GSK-3β showed the opposite results (Fig. 6a-c). In addition, the LC3-II/LC3-I expression ratio was decreased in the FAT4-overexpressing HCT116 and LOVO cells after 740Y-P treatment, but the levels of P62 were augmented in these cells (Fig. 6a-c). Additionally, wortmannin treatment inhibited the PI3K-AKT signaling pathway, leading to reduced expression of PI3K, p-AKT, mTOR, β-catenin, and p-GSK-3β, whereas the GSK-3β level was increased in FAT4-silenced SW480 cells. In contrast, the LC3-II/LC3-I expression ratio was increased after wortmannin treatment, resulting in inhibition of the PI3K-AKT-mTOR signaling pathway and thereby the induction of autophagy, as demonstrated by a decreased level of the negative autophagy marker P62. Considering these results, we conclude that the regulatory effects of FAT4 on autophagy and the EMT can be partially attributed to the PI3K-AKT-mTOR and PI3K-AKT-GSK-3β signaling pathways.

To confirm the inhibitory function of autophagy in cell chemotaxis, we analyzed the migration and invasion capability of CRC cells transduced with FAT4-overexpressing or FAT4-knockdown constructs. Our previous results demonstrated that FAT4 can promote autophagy and inhibit the migration and invasion of CRC cells via the PI3K-AKT-mTOR and PI3K-AKT-GSK-3β signaling pathways. Additionally, we questioned whether FAT4 can reduce the EMT by increasing the activity of autophagy. The study conducted by Gugnoni indicated that accumulated P62 might ubiquitinate Twist1 to promote the stabilization of Twist1 and the EMT in papillary thyroid carcinomas [4]. However, in CRC, whether FAT4 can regulate autophagy to influence the EMT needs to be confirmed. Thus, we applied MHY1485, an autophagy inhibitor (mTOR promoter), and the activator Torin 1 (mTOR inhibitor) to the cells showing stable FAT4 overexpression or knockdown and evaluated the expression of LC3, P62 and Twist1. The results showed that in the FAT4-overexpressing HCT116 and LOVO cells, the expression levels of P62 and Twist1 were increased after treatment with the autophagy inhibitor MHY1485, but these levels were reduced in FAT4-knockdown SW480 cells treated with the autophagy activator Torin 1. In addition, the LC3-II/LC3-I expression ratio was decreased in the FAT4-overexpressing HCT116 and LOVO cells following MHY1485 treatment, and this ratio was increased in the FAT4-knockdown SW480 cells treated with Torin 1. The inhibition of autophagy is associated with an increased level of P62, which prevents the degradation of Twist1. Transwell assays revealed that cell migration and invasion were promoted after the attenuation of autophagy in the FAT4-overexpressing HCT116 and LOVO cells treated with the autophagy inhibitor MHY1485, whereas the FAT4-knockdown SW480 treated with the autophagy activator Torin 1 showed poor cell migration and invasion (Fig. 7a-c). These findings demonstrate that autophagy can partially inhibit the migration and invasion of CRC cells, and this effect can be partly attributed to the PI3K-AKT-mTOR signaling pathway.

After gaining abundant acknowledgement of the in vitro effects of FAT4, we sought to further elucidate the tumor suppressor role of FAT4 during tumor growth in vivo. Thus, we constructed a tumor xenograft model. Specifically, tumor xenografts were successfully obtained with FAT4-overexpressing HCT116 and LOVO cells, and the tumor volumes obtained with these cells were smaller than those obtained with the control cell lines. In contrast, the tumors derived from the FAT4-silenced SW480 cells were increased compared with those obtained with the non-transduced SW480 cells. The tumor growth curve is demonstrated in the right panel of Fig. 8b. Furthermore, throughout the growth process, the average tumor weight was typically obtained in the groups with reduced FAT4 expression, whereas the FAT4-overexpressing HCT116 and LOVO cells yielded tumors with relatively lighter weights throughout the growth process (Fig. 8a). In conclusion, these data are consistent with the results from the in vitro cell proliferation assay and indicate that FAT4 is associated with CRC progression and tumorigenesis in vivo.