Introduction
Hepatocellular carcinoma (HCC) is one of the most common types of malignant digestive system tumors and is associated with a high mortality rate.1 Occurrence and development of HCC involves the alteration of many genes and signaling pathways, but its pathogenic mechanism has not been fully elucidated.2 To gain a comprehensive understanding of the genetic alterations that occur during HCC initiation, many researchers have analyzed the HCC genome using whole-genome sequencing strategies.3,4 To date, several genes, including telomerase reverse transcriptase (TERT), tumor protein P53 (p53), AT-rich interactive domain 1A (ARID1A), cyclin dependent kinase inhibitor 2A (CDKN2A), catenin beta 1 (CTNNB1), axin 1 (AXIN1), and cyclin D1 (CCND1), among others, have been shown to be related to HCC.5
ARID1A, its encoding gene located on chromosome 1p36.11, represents a subunit of the switch/sucrose non-fermentable (SWI/SNF) chromatin remodeling complex.6 Chromatin remodeling complexes modify chromatin structures and regulate the transcription of genes to control various cellular processes.7 Inactivating mutations in ARID1A have been identified in a wide variety of cancers, suggesting that it functions as a tumor suppressor.8 However, its anticancer mechanisms of action in HCC are not fully understood.
MYC is a transcription factor encoded by the c-MYC gene that regulates an estimated 15% of genes in the human genome.9 MYC is an oncoprotein that contributes to the malignancy of many aggressive cancers.10 The c-MYC locus is the most frequently amplified locus across all human cancers, leading to MYC overexpression.11 MYC is frequently overexpressed in patients with HCC12–14 and experimental overexpression of MYC in the livers of mice can lead to the development of HCC.15
In this study, the role of ARID1A in HCC progression was investigated. Using in vitro cell models and in vivo mouse models, ARID1A deficiency was shown to accelerate the development and progression of liver cancer. Furthermore, mechanistic studies revealed that ARID1A inhibits proliferation in HCC cells via the down-regulation of c-MYC transcription.
Methods
Cell culture
The Bel7404 cell line was a gift from Professor Cang (Zhejiang University, Hangzhou, China). The Huh7 and HepG2 cell lines were purchased from American Type Culture Collection (ATCC, Manassas, VA, USA). Huh7 and Bel7404 cells were grown in Roswell Park Memorial Institute (RPMI) 1640 media (Invitrogen, Grand Island, NY, USA) supplemented with 10% fetal bovine serum, 1% penicillin/streptomycin and 1% glutamine in an incubator maintained at 37°C with 5% CO2. HepG2 cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) (Corning Life Science, Corning, NY, USA).
Plasmid and lentivirus
The clustered regularly interspaced short palindrome repeats/CRISPR-associated protein 9 (CRISPR/Cas9) system (PSpCas9(BB)-2A-Puro; PX459) was purchased from Addgene (Watertown, MA, USA). Knockout cells were generated by transfecting the cells with CRISPR/Cas9 using Lipofectamine™ 2000 (Invitrogen), according to a previous report.16 Individual cells were selected to generate monoclonal cell lines (Bel7404 ARID1A KO-1 and KO-2). LentiCRISPR v2-sgARID1A (Addgene) and pLenti-puro-ARID1A (Addgene) lentiviral vectors were used to knockout and overexpress ARID1A, respectively. To generate stable transfectants, the lentiviral vector, the psPAX2 packaging plasmid (10 µg), and the pMD2.G envelope plasmid (10 µg) were transfected into 293T cells using the standard calcium phosphate transfection method. Lentivirus soups were collected and concentrated by density gradient after 48 h for immediate use, or were frozen at −80°C for later use.
In vivo experiment and hydrodynamic tail vein injection
The animal experiment closely adhered to the Zhejiang University guide for the care and use of laboratory animals. For DEN treatment, 14 day-old male C57BL/6 mice were administered with a single intraperitoneal injection of DEN (25 µg/g body weight). A pX459 vector co-expressing an sgRNA targeting Arid1a, Pten or p53 was cloned. Vectors for hydrodynamic tail vein injections were prepared using the EndoFreeMaxi Kit (Qiagen, Hilden, Germany). For hydrodynamic liver injection, plasmid DNA suspended in 2 mL saline was injected into 8 week-old male C57BL/6 mice via the tail vein within 6–7 sec. The amount of injected DNA was 60 µg for sgArid1a, and 60 µg each for sgArid1a+sgPten+sgp53. An equal amount of pX459 was used as a control for each experiment.
Western blot
Cells were lysed in NETN lysis buffer and 30 µg total protein was run on a gel using sodium dodecyl sulfate polyacrylamide gel electrophoresis. After electrophoresis, proteins were transferred onto a polyvinylidene fluoride membrane (Millipore, Burlington, MA, USA). After blocking for 1 h, membranes were rinsed with Tris-buffered saline with Tween-20 (TBST) three times and incubated in the corresponding primary antibody at 4°C overnight (antibodies listed in Supplementary Table 2). The membranes were then rinsed three times with TBST and incubated with secondary antibodies. Finally, membranes were incubated with an enhanced chemiluminescence system (ThermoFisher Scientific, Waltham, MA, USA). The bands were detected by ChemiDoc XRS Image System (Bio-Rad Laboratories, Hercules, CA, USA).
Immunohistochemistry
Tissues were fixed in 4% paraformaldehyde (PFA) and embedded in paraffin. Paraffin sections (4 µm) were dewaxed by xylene and rehydrated in decreasing concentrations of ethanol. Epitope retrieval was performed in 10 mM citrate buffer (pH 6.0) at 95°C for 20 m in a microwave oven. Endogenous peroxidase activity was blocked for 10 min by 0.3% H2O2 in phosphate-buffered saline (PBS). Tissue sections were incubated with antibodies overnight (shown in Supplementary Table 2). Next, the sections were incubated with a diluted biotinylated goat anti-rabbit IgG antibody (Vector Laboratories, Burlingame, CA, USA) for 30 m at room temperature. Chromogenic reactions were carried out according to the protocols provided in the ImmPACT™ DAB kit (Vector Laboratories).
Immunofluorescence (IF)
Cells were plated and grown on glass slides, washed with PBS, and fixed with 4% PFA for 15 m. Cells were then washed again with PBS and treated with 0.2% Triton X-100 (ThermoFisher Scientific) in PBS for 10 m to permeabilize the cells. After washing with PBS again, the cells were blocked with 2% bovine serum albumin (ThermoFisher Scientific) in PBS at room temperature for 1 h, then incubated with primary antibodies (Supplementary Table 2) at 4°C overnight. Cells were washed once with PBS, then incubated with secondary IF-specific antibodies at room temperature for 1 h. The cells were observed under a laser scanning confocal microscope (LSM710; Zeiss, Oberkochen, Germany). 4′,6-diamidino-2-phenylindole (DAPI) and F-actin were used as staining controls.
Quantitative reverse-transcription polymerase chain reaction (qPCR)
TRIzol reagent (Invitrogen) was used to extract the total RNA from cells, and 2 µg of total RNA was used for reverse transcription. The Bio-Rad CFX96 system was used to conduct the qPCR and calculate the expression of mRNA. Data were normalized to glyceraldehyde 3-phosphate dehydrogenase (GAPDH) and relative expression was assessed using the ΔΔCt method. All primers used are shown in Supplementary Table 3. Experiments were performed in triplicate.
3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-2H-tetrazolium bromide (MTT) cell viability assay
HCC cells were plated in 24-well plates at a density between 2×103 and 5×103 cells per well. After treatment, culture media was removed and 500 µL of MTT (0.5 mg/mL) per well was added and cells were incubated at 37°C with 5% CO2 for 1 h. The absorbance at 570nm was detected. Cell viability was calculated using the formula: [optical density (OD; sample) − OD (blank)] / [OD (control) − OD (blank)]. Experiments were repeated at least three times.
Analysis of The Cancer Genome Atlas (TCGA) data
The subset of data from TCGA Liver Hepatocellular Carcinoma (source data from Genome Data Analysis Centre [GDAC] Firehose) of the cbioportal.org website was analyzed. Specifically, on the home page of the website, “liver” was selected, then “Liver Hepatocellular Carcinoma (TCGA, Firehose Legacy)”. “Explore Selected Studies” was chosen, “hepatocellular carcinoma” in cancer type was detailed, and “ARID1A” was entered as the gene. The cBioPortal source code is freely available under the GNU Lesser GPL open-source license and is hosted by Google code (http://code.google.com/p/cbio-cancergenomics-portal/ ).17
Statistical analysis
Data are presented as mean±standard deviation (SD) or standard error of the mean (SEM) from independent experiments. Statistical analyses included the Student’s t-test and chi-squared test using GraphPad Prism 6 (GraphPad Software, Inc., La Jolla, CA, USA). A p-value <0.05 was considered statistically significant.
Results
ARID1A deficiency accelerates liver tumorigenesis in mice
To explore the roles of ARID1A in HCC initiation, Arid1a was knocked out in mice using the CRISPR/Cas9 system, as previously described.18 A pX459 vector co-expressing a single guide (sg)RNA targeting Arid1a (Arid1a target sequence presented in Supplementary Table 1, termed sgArid1a) and Cas9 was cloned. In vitro, sgArid1a caused the loss of ARID1A in murine Hep1-6 cells (Fig. S1A). A hydrodynamic tail vein injection was used to deliver CRISPR to the livers in mice, which can affect a large proportion of hepatocytes. As shown in Figure S1B, the hydrodynamic injection of an enhanced green fluorescent protein (eGFP) plasmid DNA resulted in liver-specific expression of eGFP in mice.
First, a cohort of wild type (WT) C57BL/6 mice were administered sgArid1a to determine if knockout of Arid1a could induce the development of tumors in the liver. Immunohistochemical (IHC) staining of liver sections using an ARID1A-specific antibody revealed that approximately 10% of hepatocytes were negative for ARID1A, but these cells were surrounded by ARID1A-positive cells (Fig. S1C). Ten months later, five sgArid1a-treated mice were examined. At necropsy, zero hepatic neoplasms were noted in any of the mice (Fig. 1A).
Considering ARID1A may not directly drive liver tumorigenesis, liver damage was induced using a single intraperitoneal injection of diethylnitrosamine (DEN) at 2 weeks of age, followed by sgArid1a or pX459 as control at 6 weeks of age. At 6 months, the liver phenotypes were assessed. With sgArid1a treatment, four of the five mice developed hepatic tumors, whereas no tumors were found in the control group (p=0.048; Fig. 1B).
Considering liver tumorigenesis may result from the accumulation of multiple mutations, mice were treated simultaneously with sgArid1a, sgPten and sgp53 (target sequences listed in Supplementary Table 1). Liver-specific knockout of Pten in mice has been shown to induce lipid accumulation and the incidence of liver cancer.19,20 As shown in Figure S1D, knockout of Pten was successful, as some hepatocytes were negative for PTEN and showed signs of lipid degeneration. At 3 months, the livers were harvested from sgArid1a+sgPten+sgp53 and control pX459+sgPten+sgp53 mice. There were more nodules in the livers of sgArid1a+sgPten+sgp53 mice compared to the control group, although this difference did not reach statistical significance (n=5/group, p=0.067; Fig. 1C).
The above in vivo data suggested that ARID1A deficiency alone cannot cause liver cancer, but ARID1A may play a tumor suppressive role and its loss can accelerate liver tumorigenesis when other pro-oncogenic factors are introduced.
ARID1A mutations are associated with a poorer prognosis in HCC patients
To dissect the function of ARID1A in HCC patients, a human survival analysis was conducted using cBioPortal (https://www.cbioportal.org/ )17,21 with data from TCGA Liver Hepatocellular Carcinoma (source data from GDAC Firehose). Thirty-four mutations of ARID1A were observed in 32 HCC patients (specific mutations listed in Supplementary Table 4). The results of the Kaplan-Meier survival analysis suggested that patients with an ARID1A mutation had a poor prognosis in terms of overall survival (n=365, p=0.008093); however, there was no significant difference in disease-free survival (n=315, p=0.0719; Fig. 2A). Interestingly, the human clinical survey showed that patients with an ARID1A mutation had more adjacent hepatic tissue inflammation compared to those with WT ARID1A (p=0.002194, q=0.0757; Fig. 2B).
Next, expression profiling analysis was performed based on the subset data of TCGA Liver Hepatocellular Carcinoma. Expression profiles of mRNA were displayed using a volcano plot (Fig. 2C). A total of 216 differentially expressed genes are listed in Supplementary Table 5. The mRNA levels of MYC, a known oncogene, were significantly higher in the ARID1A mutation group compared to the WT ARID1A group (p=9.28 e−8, q=1.332 e−4; Fig. 2D).
In summary, these human clinical results revealed that an ARID1A mutation was associated with poorer prognosis in HCC patients. MYC was a candidate gene that is regulated by ARID1A, which may exert its tumor suppressive functions.
ARID1A inhibits HCC cell proliferation and is required for DNA damage repair and apoptosis
To study the function of ARID1A in vitro, the expression levels of ARID1A were first detected in a variety of HCC cell lines. Using western blot analysis, Bel7404 and HepG2 cells were found to be “ARID1A-positive”. In contrast, Huh7 cells were “ARID1A-negative” (Fig. 3A). To test whether ARID1A is crucial for HCC cell proliferation, ARID1A was further manipulated and the proliferation of HCC cells was examined using a MTT assay. The results revealed that overexpression of ARID1A in Huh7 cells inhibited their proliferation and knockout of ARID1A in Bel7404 cells (ARID1A KO cells) resulted in enhanced proliferation (Fig. 3B).
Next, the cellular responses to DNA damage were investigated. An MTT assay revealed better cell viability in ARID1A KO cells after ionizing radiation (IR) treatment, suggesting that the knockout of ARID1A in HCC cells caused the cells to become more resistant to IR stress (Fig. 3C, top). Results from western blot and immunofluorescence (IF) assays revealed that ARID1A KO cells could not accumulate the same level of gamma histone 2 A variant X after radiation treatment compared to WT Bel7404 cells (Fig. 3C, middle). Results from the apoptosis assay, where cleaved poly (ADP-ribose) polymerase (c-PARP) was detected by western blotting, also revealed that knockout of ARID1A could dramatically decrease the levels of c-PARP after IR treatment (Fig. 3C, bottom), suggesting that the tumor suppressive role of ARID1A could be the result of inducing apoptosis in damaged cells.
Taken together, the above data suggest that ARID1A inhibits HCC cell proliferation and is required for DNA damage repair and apoptosis.
ARID1A inhibits cell proliferation of HCC cells via downregulation of MYC transcription
Expression profiles of mRNA indicated that MYC might contribute to the ARID1A-dependent regulation of cell proliferation. To test this hypothesis concerning the molecular mechanism, transcript levels of MYC were detected using qPCR. Consistent with the gene expression profiling analysis, knockout of ARID1A increased the level of MYC mRNA in Bel7404 cells, while overexpression of ARID1A resulted in a decrease of MYC mRNA in Huh7 cells (Fig. 4A). Using western blotting, knockout of ARID1A increased the level of MYC protein in Bel7404 cells (Fig. 4B). In addition, knockdown of MYC in Bel7404 cells could reverse the proproliferative effect caused by knockout of ARID1A (Fig. 4C).
Taken together, these data presented in Figure 4 suggest that MYC can be regulated by ARID1A and this contributes to the regulation of cell proliferation.
Discussion
HCC is one of the most common types of liver cancer, and accounts for 90% of all primary liver cancers.22 However, effective treatments for HCC are lacking due to its heterogeneity.23 HCC patients carry mutations in numerous genes, including ARID1A. In the TCGA data set, 8.22% (30/365) of HCC patients carry an ARID1A mutation, and most of these are inactivating mutations. The clinicopathologic significance of ARID1A expression in HCC has been investigated previously, and it was revealed that 12.17% of HCC tumors (14/115) were ARID1A-negative and that loss of ARID1A was significantly associated with larger tumors.24
In this study, the knockout of Arid1a alone could not initiate liver cancer in mice, suggesting that it is not a cancer-driver gene. However, knockout of Arid1a accelerated liver tumorigenesis when DEN or a combination knockout of Pten and p53 were introduced, suggesting that liver tumorigenesis is a multistep process that requires various other factors. Similar results have been reported by others, such as the finding of mice with homozygous or heterozygous deletions in Arid1a not developing ovarian lesions but mice with an Arid1a and Pten double-knockout developing ovarian endometrioid cancer.25
Further TCGA data analysis showed that ARID1A mutations are associated with a poorer prognosis in HCC patients, indicating that ARID1A may have prognostic value. Expression profiles of mRNA showed that MYC transcription was significantly higher in patients with an ARID1A mutation. ARID1A mutations resulted in abnormal chromatin remodeling that diverted gene transcription. Previously, others have also reported that mutant ARID1A is able to promote cell proliferation by triggering the phosphoinositide 3-kinase/protein kinase B (PI3K/AKT) signaling pathway,26,27 which affects the expression of other cell cycle regulators, such as the MYC gene.28 Consistent with previous reports, the contribution of MYC to the antiproliferative effect of ARID1A in HCC was definitively shown in the present study.
The role of ARID1A in DNA damage repair and apoptosis was also investigated. ARID1A knockout resulted in disruptions of DNA damage repair and apoptosis following IR exposure. Loss of ARID1A led to disrupted SWI/SNF function, which caused enhanced mutagenesis due to the defective DNA repair and aberrant apoptosis evasion.29,30 Thus, tumorigenesis of HCC with an ARID1A mutation is complex and involves an intricate network of mechanisms, including cell proliferation, DNA damage repair and apoptosis signaling pathways.
Undoubtedly, the full decoding of the ARID1A tumor suppressive mechanism may have future therapeutic implications. These data support the role of ARID1A in protection against HCC progression. Several targeted therapy drugs should also be considered in future studies, including inhibitors of MYC or PI3K/AKT signaling, PARP inhibitors targeting the DNA damage signaling pathway, and synthetic lethal therapies targeting epigenetic changes in ARID1A mutation-based cancers. Interestingly, it was also shown that patients with an ARID1A mutation had more severe adjacent hepatic tissue inflammation, suggesting that ARID1A is involved in tumor immunity and may be targeted by immunotherapy.
In summary, the current findings further elucidate the tumor-suppressive mechanism of ARID1A in HCC. A loss-of-function ARID1A mutation promotes cell proliferation and disrupts DNA damage repair and apoptosis pathways. Loss of ARID1A may promote HCC progression via the transcriptional up-regulation of MYC.
Supporting information
Fig. S1
CRISPR-mediated mutation of genes in hepatocytes of mice.
(A) In vitro, Hep1-6 cells were infected with lentiCRISPR targeting Arid1a and western blots showed ARID1A and β-tubulin protein expression levels. (B) Hydrodynamic injection of an eGFP plasmid DNA resulted in liver-specific expression of eGFP in C57BL/6 mice. (c) Representative IHC staining of C57BL/6 mice injected with PX459 (control) or sgArid1a. Arrows indicate hepatocytes with negative ARID1A staining, surrounded by ARID1A-positive hepatocytes. (D) Representative IHC staining of C57BL/6 mice injected with PX459 (control) or sgPten plus sgP53. Arrows indicate hepatocytes with negative PTEN staining and lipid accumulation. ARID1A, AT-rich interactive domain-containing protein 1A; eGFP, enhanced green fluorescent protein; PTEN, phosphatase and tensin homolog.
(TIF)
Supplementary Table 1
Single guide RNA target site sequences.
(DOCX)
Supplementary Table 2
Antibodies used in WB, IHC, and IF.
(DOCX)
Supplementary Table 3
Primer sequences.
(DOCX)
Supplementary Table 4
Mutations of ARID1A in TCGA dataset.
(DOCX)
Supplementary Table 5
Differentially expressed genes between Arid1a mutant and WT HCC samples (TCGA database, Firehose).
(DOCX)
Abbreviations
- γ-H2AX:
gamma histone 2 A variant X
- ARID1A:
AT-rich interactive domain-containing protein 1A
- c-PARP:
cleaved poly (ADP-ribose) polymerase
- CRISPR/Cas9:
clustered regularly interspaced short palindrome repeats/CRISPR-associated protein 9
- DAPI:
4′,6-diamidino-2-phenylindole
- DEN:
diethylnitrosamine
- DMEM:
Dulbecco’s modified Eagle’s medium
- eGFP:
enhanced green fluorescent protein
- GAPDH:
glyceraldehyde 3-phosphate dehydrogenase
- GDAC:
Genome Data Analysis Centre
- HCC:
hepatocellular carcinoma
- IF:
immunofluorescence
- IHC:
immunohistochemical
- IR:
ionizing radiation
- KO:
knockout
- MTT:
3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-2H-tetrazolium bromide
- p53:
tumor protein P53
- PBS:
phosphate-buffered saline
- PFA:
paraformaldehyde
- PI3K/AKT:
phosphoinositide 3-kinase/protein kinase B
- PTEN:
phosphatase and tensin homolog
- qPCR:
quantitative reverse-transcription polymerase chain reaction
- RPMI:
Roswell Park Memorial Institute
- SD:
standard deviation
- SEM:
standard error of the mean
- sg:
single guide
- TBST:
Tris-buffered saline with Tween-20
- TCGA:
The Cancer Genome Atlas
- WT:
wild type
Declarations
Acknowledgement
We thank Yong Cang, Tong Ji, and Lidan Hou in the Life Sciences Institute of Zhejiang University for expert technical assistance.
Data sharing statement
No additional data are available.
Funding
This work was supported by Zhejiang Provincial Natural Science Foundation (LQ18H160010, LQ20C010005) and National Natural Science Foundation of China (82003062, 81903004).
Conflict of interest
The authors have no conflict of interests related to this publication.
Authors’ contributions
Study concept and design (JZ, YX), acquisition of data (JZ, YX, GL, XO, DZ), analysis and interpretation of data (JZ, YX, GL, JxZ, XL, QZ), drafting of the manuscript (JZ, YX), critical revision of the manuscript for important intellectual content (JZ, YX), administrative, technical, or material support, study supervision (JZ).