The mechanism of ATF3 repression of epithelial‐mesenchymal transition and suppression of cell viability in cholangiocarcinoma via p53 signal pathway

Abstract The aim of this research was to determine the underlying mechanism of activating transcription factor 3 (ATF3) on cell proliferation, invasion, migration and epithelial‐mesenchymal transition (EMT). The differentially expressed mRNAs in cholangiocarcinoma (CC) and its adjacent tissues were screened by microarray analysis, and the expression of ATF3 was detected through Quantitative real time polymerase chain reaction (qRT‐PCR) and Western blot. The expression of EMT markers and p53‐related proteins was analysed by Western blot. Analyses using the Cell Counting Kit‐8 and TUNEL were performed to assess the rate of apoptosis and cell proliferation. Scratch wound and transwell assays were performed to study cell migration and invasion. Activating transcription factor 3 was restrained in CC cell lines and tissues and inhibited EMT while activating the p53 signalling pathway. Knockdown of ATF3 promoted cell proliferation but reduced the rate of apoptosis by inhibiting p53 signalling. Cell migration and invasion can be strengthened by ATF3 through activating the p53 signalling pathway.

injury and oxidative stress. 7 Recent studies indicated that ATF3 is strongly associated with cancer development. 8 Depending on the tumour type, ATF3 may induce tumour cell apoptosis or improve tumour cell survival. 9 Xin et al verified that ATF3 enhances EMT in breast cancer cells, 10 while other studies revealed that ATF3 plays a tumour suppressing role in many different cancer types, including colon cancer, esophageal squamous cell carcinoma (ESCC) and hepatocellular carcinoma (HCC). 9,11 Very little research has directly addressed the role of ATF3 in human CC; therefore, we hypothesized that ATF3 might repress the process of EMT to suppress the development of CC.
p53, a downstream target of ATF3, was the first tumour suppressor gene discovered and can inhibit cancer cells through multiple pathways, such as promoting cancer cell apoptosis and restraining the cell cycle. 2 p53 is primarily regulated by the E3 ubiquitin ligase Murine Double Minute 2 (MDM2), which binds p53 at its transactivation domain to block p53-mediated transcriptional regulation while simultaneously promoting polyubiquitination and proteasome-dependent degradation of p53. 12 Recent studies have reported that ATF3 can modulate the activity of p53. 13 A central leucine zipper domain (Zip) in ATF3 mediates protein-protein interactions. Activating transcription factor 3 and p53 are bound through Zip, stimulating p53 tumour inhibitor activity independent of ATF3 transcription. 14 The aim of our study was to investigate the effects of ATF3 on cell viability via activating the p53 signalling pathway.
This research explored the differentially expressed mRNAs in CC compared to its adjacent tissues and analysed the expression of ATF3, EMT markers and p53-related proteins. Cell migration, the apoptosis rate, proliferation and invasion were also analysed in this study. We found that ATF3 repressed EMT and restrained cell migration, proliferation and invasion while enhancing cell apoptosis via activating p53 signalling.

| Microarray analysis
The gene expression profiles of eight pairs of tumour tissues and adjacent tissues (seven pairs of stage I-II, one pair of stage III-IV) obtained from The Cancer Genome Atlas (TCGA) (https://cancerge nome.nih.gov/) were analysed in this study. Differentially expressed mRNAs between normal and cancerous bile duct specimens were screened using the significance analysis of microarrays (SAMR) package in R software, and |log 2 fold change (FC)| > 2 and false discovery rate < 0.05. Cluster analysis was then performed to confirm whether the identified mRNAs could be used to robustly classify normal and CC specimens.

Genes
Primer sequence 5′-3′ F I G U R E 1 Activating transcription factor 3 (ATF3) was down-regulated in bile duct tumour tissues. A, The heat map of differentially expressed mRNAs in normal and cholangiocarcinoma (CC) specimens showed that the expression level of ATF3 was reduced in stage I-II and III-IV CCs. B, Quantitative real time polymerase chain reaction (qRT-PCR) was performed to confirm the expression of ATF3 in normal and CC tissues. **P < 0.01. C, Western blot (left) and qRT-PCR (right) were conducted to determine protein and mRNA levels of epithelialmesenchymal transition markers and p53 in normal and CC tissues, respectively. D, qRT-PCR was utilized to determine the mRNA level of ATF3 in immortalized human bile duct epithelial cells (HIBEpic) and CC cell lines (HuCCT1, RBE, QBC939 and FRH0201). *P < 0.05, **P < 0.01. E, Western blotting results revealed that ATF3 protein level was reduced in CC cell lines (HuCCT1, RBE, QBC939 and FRH0201) compared to bile duct epithelial cells (HIBEpic). *P < 0.05, **P < 0.01

| qRT-PCR
RNA from both tissues and cells was extracted using Trizol reagent (Beyotime, Shanghai, China) according to the manufacturer's guidelines. The PrimeScript TM RT reagent kit (Takara, Tokyo, Japan) was used to generate cDNA. qPCR was performed using SYBR Premix Ex Taq ™ GC (Takara) and an Applied Biosystems 7500 real-time PCR system (Applied Biosystems, Foster City, CA, USA). GAPDH was used as an internal control for ATF3 and repeated in triplicate. Samples were normalized to internal controls, and FCs were obtained using the 2 −ΔΔCT method. The primer sequences used are listed in Table 1. F I G U R E 2 Effects of si-ATF3 and ATF3-pcDNA 3.1 on activating transcription factor 3 (ATF3) expression A, QBC939 and FRH0201 cells were transfected with ATF siRNAs or ATF3-pcDNA 3.1 for 48 h, and then, qRT-PCR was performed to determine knockdown efficiency. **P < 0.01. B, Western blot showed that the protein expression of ATF3 was inhibited in QBC939 and FRH0201 cells transfected with si-ATF3. **P < 0.01, compared with the vector-only control group. Activating transcription factor 3 protein was overexpressed in QBC939 and FRH0201 cells transfected with ATF3-pcDNA 3.1. ***P < 0.001, compared with the vector-only control group. C, QBC939 and FRH0201 cells were transfected with ATF siRNAs or ATF3-pcDNA 3.1 for 48 h, and then, qRT-PCR was performed to determine mRNA level of p53. **P < 0.01. D, QBC939 and FRH0201 cells were transfected with ATF siRNAs or ATF3-pcDNA 3.1 for 48 h, and then, Western blot was performed to determine protein level of p53. **P < 0.01. (E,F) qRT-PCR were conducted to determine the mRNA levels of EMT markers in QBC939 and FRH0201 cells. *P < 0.05, **P < 0.01.

| TUNEL assay
We seeded 2 × 10 5 cells in a 24-well plate and incubated them for 24 hours. After fixation with 4% paraformaldehyde, the cells were infiltrated in 3% hydrogen peroxide/methanol for 10 minutes. Cells were then permeabilized with 0.5% Triton for 5 minutes, incubated with 50 μL of TUNEL reaction liquid and incubated for 1 hour at 37°C. The cells were imaged using the EVOS FL fluorescence microscope (Thermo Fisher Scientific), and five fields of view were selected for quantification.
The apoptosis rate was quantitated as the TUNEL-positive cell number.

| Scratch wound assay
Six-well plates were seeded with 10 5 cells, cultured until 60% confluency, and then scratched with a plastic filter tip to make a

| Statistical analysis
Data are represented as the mean ± SD and analysed by GraphPad differences of treatment groups were calculated using ANOVA and Student's t test. P < 0.05 was considered statistically significant.

| ATF3 is expressed at a low level in CC cell lines and tissues
According to the microarray analysis, ATF3 was markedly repressed in different CC tissues ( Figure 1A). The expression of ATF3 was markedly decreased in CC tissues compared with normal tissues, as detected by qRT-PCR (P < 0.01, Figure 1B). Figure 1C indicates that EMT-and p53-associated proteins were significantly altered in CC tissues. In addition, both mRNA expression (P < 0.01, Figure 1D) and protein expression (P < 0.01, Figure 1E) of ATF3 were significantly lower in the four human CC cell lines compared with the bile duct epithelial cell line HIBEpic, which verified that ATF3 was not highly expressed in CC cells.
The QBC939 and FRH0201 cell lines were transfected with both si-ATF3 and ATF3-pcDNA3.1. The relative expression of ATF3 mRNA and protein was analysed by Western blot as well as qRT-PCR ATF3 expression was significantly down-regulated after knockdown (P < 0.01) compared to QBC939 and FRH0201 cells (P < 0.01, Figure 2). These results suggested that si-ATF3 inhibited F I G U R E 4 Activating transcription factor 3 (ATF3) suppresses cell migration and invasion (A-C) The scratch wound assay revealed that cell the migration capacity of QBC939 and FRH0201 cells in the MX69 group was decreased but was increased in the si-ATF3-treated group. The migration capacity was not significantly different between the MX69 and si-ATF3-treated groups. *P < 0.05, **P < 0.01, compared with the vector group; #P < 0.05, ##P < 0.01, compared with the siRNA groups. (D-F) The Transwell assay showed that MX69 inhibited invasion of QBC939 and FRH0201 cells and that knockdown with si-ATF3 had the opposite effect. The invasion capacity was not significantly different in the MX69 and si-ATF3 groups. **P < 0.01, ***P < 0.001, compared with the vector group; #P < 0.05, ##P < 0.01, ###P < 0.001, compared with the siRNA-treated groups.
the expression of ATF3, whereas ATF3-pcDNA 3.1 promoted ATF3 expression in CC cells; thus, the transfection was successful.

| ATF3 suppresses EMT in CC cells
The epithelial markers α-catenin and E-cadherin showed robust down-regulation, while the expression of the mesenchymal markers Vimentin and Fibronectin increased after down-regulation of ATF3 (P < 0.01, Figure 2D-F). Transcriptional expression of the three canonical EMT markers Snail1, Slug and Twist was up-regulated after ATF3 knockdown (P < 0.01, Figure 2D-F), which was further confirmed with similar western blotting results (Figure 3). Scratch wound and Transwell assays also demonstrated that MX69 inhibited the cell invasion and migration ability of QBC939 and FRH0201 cells, whereas si-ATF3 had an opposite effect that did not change in the MX69 and si-ATF3 group (P < 0.01, Figure 4). From these experiments, we conclude that ATF3 can partially repress EMT in CC cells.

| ATF3 activates the p53 signalling pathway
qRT-PCR assays indicated that ATF3 knockdown might inhibit the expression of p53 and its target genes such as Bax, p21 and p53 upregulated modulator of apoptosis (PUMA) (P < 0.01, Figure 5A,B).
Western blotting indicated that low expression of ATF3 might repress the expression of p53 and the downstream proapoptotic protein Bax. MDM2 antagonized p53 activity and was highly expressed in both QBC939 and FRH0201 cells (P < 0.01, Figure 5C, D). Collectively, these results indicated that ATF3 activates the p53 signalling pathway and therefore affects the progression of CC.

| ATF3 suppresses cell viability via the p53
signalling pathway MX69 (MDM2 suppressor/p53 activator) markedly enhanced expression of p53, while si-ATF3 inhibited expression of p53. However, p53 expression levels did not significantly change when cells were cultured with MX69 and si-ATF3 (P < 0.01, Figure 6A,B). The CCK-8 assay revealed that MX69 inhibited proliferation of QBC939 and FRH0201 cells in contrast to the si-ATF3-treated group. Cell proliferation was almost identical in the vector-only control group after MX69 or si-ATF3 treatment (P < 0.01, Figure 6C,D). Similarly, the TdT-mediated dUTP Nick-End Labeling (TUNEL) assay illustrated that QBC939 and FRH0201 cells transfected with MX69 exhibited a higher apoptosis rate than the control group. Although si-ATF3 decreased the rate of apoptosis, there was no observable change in cells treated with both MX69 and si-ATF3 (P < 0.01, Figure 7). In conclusion, ATF3 inhibits F I G U R E 5 Activating transcription factor 3 (ATF3) activates the p53 signalling pathway (A) QBC939 cells were transfected with ATF siRNAs for 48 h, and then, qRT-PCR was performed to determine mRNA levels of p53 and its target genes such as Bax, p21 and p53 upregulated modulator of apoptosis (PUMA). **P < 0.01. B, FRH0201 cells were transfected with ATF siRNAs for 48 h, and then, qRT-PCR was performed to determine mRNA levels of p53 and its target genes such as Bax, p21 and PUMA. **P < 0.01. C, QBC939 cells were transfected with ATF siRNAs for 48 h, and then, Western blot was performed to determine protein levels of p53 and its target genes such as Bax, p21 and PUMA. **P < 0.01. D, FRH0201 cells were transfected with ATF siRNAs for 48 h, and then, Western blot was performed to determine protein levels of p53 and its target genes such as Bax, p21 and PUMA. **P < 0.01. cell invasion, proliferation and migration while increasing the apoptosis of cancer cells in the bile duct via the p53 signalling pathway. The potential mechanism of ATF3 in EMT is shown in Figure 8. We also investigated the interactions between ATF3 and p53. The P53 gene encodes a tumour suppressor protein that contains DNA binding, transcriptional activation and oligomerization domains. 17,18 A recent study showed that ZNF545 acted as a functional tumour suppressor in multiple myelomas via activating the p53 signalling pathway. 19 Yan et al described that activating the p53 pathway in CNPY2 knockout cells reduced cell growth, migration, colony formation and angiogenesis but increased cellular apoptosis in colorectal tumours. 20 Furthermore, a regulatory feedback mechanism between p53 and ATF3 was previously proposed. 21