Background
Hepatocellular carcinoma (HCC), the fourth most commonly diagnosed cancer in human and one of the leading cause of cancer-related death worldwide, progresses via multiple steps accompanied by a series of progressive alterations in gene expression [
1]. Great strides have been made in exploring effective diagnostic and therapeutic strategies for HCC, but there has been no significant reduction in its morbidity and mortality. According to the International Statistical Classification of Diseases and Related Health Problems, 10th revision (ICD-10), an estimated 466,100 new cases of liver cancer and 422,100 deaths related to liver cancer occurred in China in 2015 [
2]. Therefore, the changes in gene expression and related genetic factors involved in HCC progression are still insufficiently understood, and there is an unmet clinical need for novel potential therapeutic targets.
Activating transcription factor 3 (ATF3), an ATF/cAMP-responsive element-binding protein (CREB) family member, was found to be involved in a broad spectrum of cellular stresses, including DNA damage [
3], cellular injury [
4], oxidative stress and oncogenic stimuli [
5], and was also shown to regulate diverse cellular functions by either binding to the ATF/CREB cis-regulatory element or interacting with other proteins, such as p53 and NF-κB [
6]. Several reports indicated that ATF3 expression is downregulated in a variety of human cancers, including colon cancer [
7], liver cancer [
8], multiple myeloma [
9], neuroblastoma [
10], bladder cancer [
11], prostate cancer [
12], malignant glioma [
13] and non-small cell lung carcinoma [
14]. ATF3 may inhibit tumor formation by inducing cell cycle arrest and apoptosis [
15]. In colon cancers, ATF3 plays a role in regulating genes downstream of protein kinase-like endoplasmic reticulum kinase (PERK) and PERK-eIF2α signaling during instances of endoplasmic reticulum stress [
16]; on the other hand, ATF3 plays an important role in berberine-induced apoptosis [
17]. Taketani et al found that when colon cancer cells encountered DNA damage, ATF3 and p53 could synergistically act on the promoter of the DR5 gene to eventually promote TRAIL-mediated apoptosis [
18]. Ri et al found that in multiple myeloma (MM), patients with higher ATF3 and ATF4 expression had longer progression-free survival (PFS) than those with lower ATF3 expression; this observation was also confirmed in chromatin immunoprecipitation (Ch-IP) experiments with MM cell lines and patient specimens [
9]. ATF3 also suppresses the development of prostate cancer induced by silencing of the tumor suppressor Pten in a mouse model, and knockdown of ATF3 expression promotes the activation of the oncogenic AKT signaling pathway [
19]. Although aberrant ATF3 expression is frequently found in human cancers [
20], it is imperative to explore the function and mechanism of this gene, particularly in HCC.
In this study, we identified ATF3 as a tumor suppressor for inhibiting cell proliferation and metastasis in HCC. Because ATF3 is a transcription factor, we also screened and identified its downstream target genes and found that ATF3 exerted its suppressive activities through upregulating CYR61 expression. Clinically, both ATF3 and CYR61 were downregulated in HCC tissues compared with corresponding adjacent tissues, and their expression was positively correlated. ATF3 was also significantly associated with intrahepatic metastases and overall survival (OS).
Methods
Cell lines and cell culture
The human HCC cell line SMMC-7721 and the immortalized hepatocyte cell line L-02 were purchased from the Cell Bank of the Institute of Biochemistry and Cell Biology, Chinese Academy of Sciences (Shanghai, China). MHCC-97 L, MHCC-97H and MHCC-LM3 cells were provided by the Liver Cancer Institute of Zhongshan Hospital, Fudan University (Shanghai, China). PLC/PRF/5, SK-Hep1 and human embryonic kidney (HEK)-293 T cells were purchased from the American Type Culture Collection (Manassas, VA, USA). Huh-7 cells were obtained from the Riken Cell Bank (Tsukuba, Japan). Li-7 cells were purchased from Shanghai Sixin Biological Technology. HCC-LY5 and HCC-LY10 cell lines were established in our laboratory. All the above cell lines were cultured in Dulbecco’s modified Eagle’s medium (DMEM, Gibco, New York, USA) supplemented with 10% fetal bovine serum (FBS, Gibco) along with penicillin (100 U/mL, Sigma-Aldrich) and streptomycin (100 μg/mL, Sigma-Aldrich) at 37 °C in a humidified incubator containing 5% CO2. All the cells were authenticated and characterized by their respective suppliers.
Vector construction and lentivirus infection
The coding sequences (CDS) of human
ATF3 and
CYR61 were amplified and cloned into the pWPXL lentivirus vector (Addgene, USA), pWPXL-
ATF3 and pWPXL-
CYR61 fusion expression clones were successfully obtained. shRNAs targeting
ATF3 or
CYR61 as well as a negative control (shNC) were obtained from GeneChem (Shanghai, China). The
CYR61 sequence spanning 1322 bp near the transcriptional start site (TSS) as well as its truncated and mutated variants were amplified and cloned into the pGL3 vector (Promega, Madison, WI). The target primer sequences are listed in Additional file
1: Table S1. All constructs were verified by DNA sequencing. HEK-293 T cells were transfected with these plasmids using Lipofectamine™ 2000 (Invitrogen) along with the packaging and envelope plasmids psPAX2 and pMD2.G (Addgene, USA) according to the manufacturer’s protocol. Virus particles were harvested 48 h after transfection. The HCC cells were infected with recombinant lentivirus in a 0.1% polybrene (Sigma-Aldrich) solution.
Quantitative real-time polymerase chain reaction (qRT-PCR)
Total RNA from human primary HCC tissues and cell lines was isolated using TRIzol reagent (Invitrogen, USA) and then reverse-transcribed into cDNA using a PrimeScript™ RT Reagent Kit (TaKaRa, Japan). qRT-PCR using SYBR Premix Ex Taq (TaKaRa, Japan) was performed with an Applied Biosystems 7500 (software version 2.0.5) real-time PCR system (Thermo Scientific) in triplicate, and the values were normalized to those of the housekeeping gene
GAPDH. The comparative CT (2
-ΔΔCT) method was applied to analyze the qRT-PCR data. The primer sequences used to quantify the target genes are provided in Additional file
1: Table S1.
Immunoblotting
Whole-cell or tissue extracts were lysed in RIPA buffer (Thermo Scientific, USA) containing protease inhibitor and phosphatase inhibitor (Roche, Switzerland). The resulting lysates were then electrophoresed by 12% SDS-PAGE and transferred to 0.2-μm polyvinylidene difluoride membranes (Merck Millipore). These membranes were blocked in 5% nonfat milk for 2 h at room temperature, incubated with primary antibodies at 4 °C overnight and then treated with HRP-linked secondary antibodies for 1.5 h at room temperature. After the membranes were washed with phosphate-buffered saline containing Tween 20 (PBST), the protein bands were visualized using a Pierce ECL development system (Thermo Scientific, USA) via a chemiluminescence analyzer (Bio-Rad, USA) for different exposure times. The antibody information is listed in Additional file
1: Table S2, and β-Actin (A3854, Sigma-Aldrich) was used as a loading control.
Cell proliferation and colony formation assays
Infected cells were seeded in 96-well plates (2000 cells/well) in triplicate and cultured for 7 days to assess proliferation with the Cell Counting Kit-8 (CCK-8, bimake, USA) and the 3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide (MTT) assay. The absorbance was measured at 450 nm (CCK-8) and 570 nm (MTT). For the colony formation assay, cells were seeded in 6-well plates (1000 cells/well) and allowed to form colonies. Once colonies were visible (> 50 cells), they were fixed with 4% paraformaldehyde and stained with Giemsa (Sigma-Aldrich, USA), and the number of colonies per well was counted.
Apoptotic assay
Annexin V-PE and 7-AAD (both available from BD Biosciences, San Jose, CA, USA) staining was used to visualize apoptotic cells according to the manufacturer’s instructions. Briefly, 1 × 106 cells were seeded in 6-well plates and incubated overnight. Cells were then collected and washed twice with PBS and resuspended in 200 μl of 1× binding buffer. Next, 5 μl of the Annexin V-PE and 7-AAD solution were added, and samples were incubated for 30 min at RT and analyzed by flow cytometry.
Migration and invasion assays
HCC cells were plated into 6-well culture plates (1 × 106 cells/well) and incubated overnight. Then, the monolayer of cells was scratched to form vertical wound. After we washed the cells twice with 1× PBS and replaced the medium with DMEM containing 2% FBS and 1 mM thymidine (Sigma-Aldrich, USA), cells that migrated into the wound area were captured using an inverted fluorescence microscope (Carl ZEISS, Germany) at 100× magnification and images of the wound at 0 h, 24 h and 48 h after scratching were obtained. The migratory ability of the cells was determined by calculating the ratio of the healing width at 48 h to the wound width at 0 h.
\( \mathrm{Relative}\ \mathrm{percentage}\ \mathrm{of}\ \mathrm{wound}\ \mathrm{healing}\ \left(\%\right)=\frac{\mathrm{the}\ \mathrm{wound}\ \mathrm{width}\ \mathrm{of}\ 0\mathrm{h}\hbox{-} \mathrm{the}\ \mathrm{wound}\ \mathrm{width}\ \mathrm{of}\ 48\mathrm{h}}{\mathrm{the}\ \mathrm{wound}\ \mathrm{width}\ \mathrm{of}\ 0\mathrm{h}}\times 100\% \)
A transwell insert with an 8-μm pore filter (Merck Millipore, USA) was precoated with 10% Matrigel (BD Bioscience, USA) and incubated for 30 min at 37 °C. A total of 1 × 105 cells was seeded into the upper chamber of the insert with 200 μl of serum-free media, while the lower chamber was filled with 600 μl of complete medium as a chemoattractant. After 12 ~ 24 h of incubation, the inserts were removed, and cells in the upper chamber that did not migrate were scraped away with a cotton swab. The cells that migrated through the membrane and adhered to its lower surface were fixed with 4% paraformaldehyde for 30 min and stained with a crystal violet solution (Sigma-Aldrich, USA). Invasive cells were photographed and counted using the fluorescence microscope at 200× magnification.
Mouse xenograft model
All treatments were performed under the guidelines of the Shanghai Medical Experimental Animal Care Commission. Forty microliters of ATF3-overexpressing or ATF3-silenced cells in serum-free DMEM/Matrigel (1:1, v/v) at a final concentration of 5 × 107 cells/ml were orthotopically inoculated into the left hepatic lobes of 6~ 8-week-old male mice using a microsyringe. Mice were sacrificed at 6~ 8 weeks after injection. Then, the tumor xenografts within the liver were weighed, and the liver and lung tissues were fixed in 4% buffered formalin and subjected to a routine preparation of paraffin embedding, sectioning, hematoxylin-eosin (H&E) staining and immunohistochemical staining.
RNA-seq
ATF3-overexpressing SK-Hep1 cells and ATF3-silenced SMMC-7721 cells were used for RNA-seq analysis by Shanghai Biotechnology Corporation (Shanghai, China) in duplicate. Part of the differentially expressed genes were available in Additional file
1: Table S3–S4, and the whole raw sequencing data of this study have been deposited at DNA Data Bank of Japan (DDBJ) under the accession number DRA007320 and is available for download at:
https://www.ddbj.nig.ac.jp/index-e.html.
Dual-luciferase reporter assays
Cells were plated in 48-well culture plates to allow adhesion, followed by transient cotransfection with pWPXL or ATF3 plasmids, CYR61 promoters, and the PRL-TK reporter construct using Lipofectamine™ 2000 (Invitrogen). After 48 h, the Renilla and firefly luciferase activities were determined according to the manufacturer’s instructions (Promega).
Ch-IP
The Ch-IP assay was performed in 293 T, SMMC-7721 and Huh-7 cells. The cells were cross-linked with 10% formaldehyde and then quenched with 1 M glycine. After the cells were washed with 1× PBS, they were incubated in Tissue Protein Extraction Reagent (Thermo Scientific) for 5 min in an ice bath and centrifuged at 2000 rpm for 5 min. The sediments were suspended in nuclear lysis buffer, and DNA was sheared into fragments of 200~ 500 bp by sonication. The nuclear lysate was incubated with specific antibody and protein A/G agarose beads (Sigma-Aldrich) at 4 °C overnight on a rotator. After reversing the crosslinks, the DNA was isolated and used for PCR analysis with the primers listed in Additional file
1: Table S1.
Immunohistochemical analysis
Tissue microarrays (TMAs) comprising 236 human primary HCC tissues obtained from the Qidong Liver Cancer Institute were constructed, and staining was performed as previously described [
21]. The samples were photographed using a Leica SCN400 slide scanner (Meyer Instruments, Houston, TX, USA) and analyzed by semiquantitative scoring. Immunohistochemical scores were obtained as follows: the intensity of staining was categorized as 0 or 1 for low or high protein expression, respectively. The antibodies used are listed in Additional file
1: Table S5.
Statistical analysis
Values are presented as the mean ± standard deviation (S.D.) with at least three independent experiments. The data were analyzed using SPSS version 16.0 (SPSS, Inc., Chicago, IL, USA). Data analysis was conducted by paired or unpaired two-tailed Student’s t-tests. The relationship between ATF3 and CYR61 expression was analyzed using Spearman’s correlation coefficient. Survival analysis was performed using the Kaplan–Meier method. Survival prognosis was investigated with univariate Cox regression models followed by a multivariate Cox regression model. Statistical significance was defined as *P < 0.05 and **P < 0.01.
Discussion
ATF3, a transcriptional repressor in the ATF/CREB family, contains only one functional domain (the basic leucine zipper domain, bZIP) that binds to the ATF/CREB cis-regulatory element (5’-TGACGTCA-3′) while dimerizing with other ATF/CREB proteins [
22]. Normally, ATF3 is rarely detectable under basal conditions in most cells, but its expression can be immediately induced upon exposure of cells to a broad spectrum of stimuli, such as anticancer drugs, toxic chemicals, endoplasmic reticulum stress, DNA damage, proteasome inhibitors, oxidative stress, oncogenic stimuli, genotoxic agents, homocysteine and ischemia-reperfusion, all of which can induce cell cycle arrest and apoptosis [
23,
24].
ATF3 participates in a number of cellular signal transduction pathways that include proteins such as p53 [
25], TGF-β, NF-κB, Toll-like receptor 4 [
26] and mouse double minute 2 (MDM2) by either interacting with other proteins or binding to the consensus ATF/CREB cis-regulatory element [
5]. Previous studies demonstrated that in response to DNA damage, ATF3 could activate the tumor suppressor p53 and regulate the expression of target genes downstream of p53 [
27].
ATF3 may play different roles in various tumors. Li et al. [
28] found that ATF3 was reduced in esophageal squamous cell carcinoma (ESCC) compared with non-tumor adjacent tissues and ATF3 suppressed ESCC via downregulation of ID1. In bladder cancer, ATF3 suppresses metastasis of bladder cancer by regulating gelsolin-mediated remodeling of the actin cytoskeleton [
11]. Reactivation of ATF3 by pracinostat is a determining factor in the tumor response to the HDACi therapy and ATF3 was a biomarker of tumor response [
29]. On the other hand, upregulation of ATF3 in lung cancer could promote cell proliferation, migration, and invasion [
30]. ATF3 plays contradictory functions in different diseases may be due to the complex tumor microenvironment, such as the community of genomically altered cancer cells, non-neoplastic cells, and a diverse collection of microorganisms, as well as complex crosstalk of intracellular molecules and multi-module activated signal transduction pathways. In hepatocellular carcinoma, ATF3 expression is lower in patients with advanced HCC and capsule invasion [
8]. Consistent with this result, our clinical pathological analysis indicated that ATF3 was relatively downregulated in cancerous tissues compared with corresponding noncancerous liver tissues. However, we observed that ATF3 was negatively correlated with intrahepatic metastasis and was positively associated with the OS of HCC patients (Fig.
7). Our experiments further confirmed that ATF3 could suppress HCC cell proliferation and metastasis both in vitro and in vivo.
Because of ATF3’s role as a transcription factor, identifying the target genes of ATF3 is needed to better understand its physiological significance in HCC. Based on the results of the RNA-Seq analysis in HCC cells with overexpressed or silenced ATF3, we found that the most obviously affected gene is
CYR61, which is a member of the CCN family that acts as an immediate-early gene in fibroblasts following exposure to growth stimuli [
31]. CYR61 expression could be induced during hepatic injuries, and the protein functions to restrict and resolve liver fibrosis [
32]. Recent studies have shown that CYR61 exhibits a protective role in wound healing and tissue repair [
33]. Some reports have stated that CYR61 mediates diverse functions, including extracellular matrix formation, differentiation, cell proliferation, adhesion, migration, survival, angiogenesis and tumorigenesis [
34‐
36], and acts as a tumor suppressor in hepatocarcinogenesis via p53 and the DNA damage response [
37]. Chen et al considered that CYR61 may suppress HCC through both apoptotic and growth inhibitory mechanisms to prevent tumor progression during the very early stages of oncogenesis; one such example is suppression of hepatocarcinogenesis by inhibiting compensatory EGFR-dependent hepatocyte proliferation through integrin α6-ROS-p38 MAPK-mediated activation of p53 [
38]. On the contrary, there were still some other reports showed that CYR61 could lead to the carcinogenesis in the HCC [
39,
40]. In non-small cell lung cancer, CYR61 overexpression remarkably inhibited the proliferation of cancer cells by arresting them in G1 phase, prominently upregulated the expression of p53 and p21 (WAF1), and decreased the activity of cyclin-dependent kinase 2 [
41]. The increased apoptosis via Cyr61-induced caspase-3 activation and depolarization of the mitochondrial membrane suppressed the growth of endometrial cancer cells [
42]. Juric et al found that CYR61 is a physiologic regulator of Fas-mediated apoptosis and that the extracellular matrix microenvironment can modulate Fas-dependent apoptosis through CYR61 expression [
43]. Consistent with these reports, our experiment showed that CYR61 could induce apoptosis and decrease the proliferation of HCC cells. On the other hand, the rescue experiments confirmed that CYR61 knockdown could remarkably decrease the anti-cancer function of ATF3 overexpression.