Wen Wang and Gang Xu contributed equally to this work
H. Ren; Z.-T. Qi, Department of Microbiology, Shanghai Key Laboratory of Medical Biodefense, National Innovation Alliance for Hepatitis & Liver Cancer, Second Military Medical University, Shanghai 200433, China
As a therapeutic or chemopreventative agent for various cancers, all-trans retinoic acid (atRA) has been reported to inhibit growth, induce apoptosis or cause differentiation. It was found that atRA could protect hepatocellular carcinoma (HCC) cells against cell death induced by serum starvation. Furthermore, it was found that atRA could enhance cell adhesion, but had no effect on the cell cycle and apoptosis. Using an Illumina Human HT-12 v4 expression microarray, 207 upregulated and 173 downregulated genes were identified in HepG2 cells treated with atRA. The most upregulated genes are cytochrome P450 family 26 subfamily A polypeptide 1 (CYP26A1), histidine triad nucleotide binding protein 3 (HINT3), miR-1282 and cytochrome P450 family 26 subfamily B polypeptide 1 (CYP26B1), which showed more than fivefold greater expression. Using Gene Ontology analysis, the greatest significance was found in extracellular-matrix-related molecular functions and the cellular component in upregulated genes. The upregulation of collagen 8A2 (COL8A2) was further confirmed using quantitative RT-PCR and western blotting. Knockdown of COL8A2 blocked enhancement in the early stage of cell adhesion by atRA treatment. Re-expression of COL8A2 in COL8A2-knocked-down HCC cells reversed the effect of small interfering RNA–COL8A2. In addition, COL8A2 could increase HCC cell migration and invasion. Thus, COL8A2 was identified as the key protein involved in the enhancement of cell adhesion of atRA under serum-free conditions. In conclusion, atRA protects HCC cells against serum-starvation-induced cell death by enhancing cell adhesion, and COL8A2 plays an important role in HCC cell migration and invasion.
Hepatocellular carcinoma (HCC) is the fifth most common cancer in the world, and the third leading cause of cancer deaths, with a poor prognosis in most cases. Surgical resection improves survival in the minority of patients with HCC and standard chemotherapy usually has no beneficial outcome on HCC patients . It is therefore extremely important to search for new therapeutic modalities.
Retinoic acid (RA), a vitamin A derivative, acts as an inhibitor of carcinogenesis by blocking the promotion of initiated or transformed cells via three mechanisms: induction of apoptosis, the arrest of further growth of abnormal cells and induction of abnormal cells to differentiate back to normal . All-trans retinoic acid (atRA) has been successfully used in the ‘cancer differentiation therapy’ of human acute promyelocytic leukemia and combined with anthracyclins cures 70–80% of patients , therefore atRA is currently utilized as a therapeutic or chemopreventative agent for various cancers. Several groups have reported that retinoid analogs with agonistic or antagonistic activity can inhibit cell growth [4-7], induce apoptosis [8, 9] or cause differentiation [5, 10]. Other groups have noted the capacity of retinoids to inhibit mammary carcinogenesis in animal models [11-13]. With regard to HCC, atRA was also reported to improve patient survival and clinical outcome in patients with inoperable HCC [14, 15]. Moreover, a phase II trial of a synthetic atRA combined with transarterial chemoembolization versus transarterial chemoembolization alone in advanced HCC patients is currently underway. However, HCC cells occasionally show resistance to atRA cytotoxicity , which indicates a much more complicated mechanism for the effect of atRA on HCC tumorigenesis. In this study, it was found that atRA could protect HCC cells against cell death induced by serum starvation, and COL8A2 was identified as playing a key role in this atRA protection effect.
atRA protected hepatocellular carcinoma cells from cell death induced by serum starvation
In HepG2 cells, cell numbers began to decrease at 36 h, decreased to > 50% at 48 h and to nearly 0 at 96 h after serum starvation. In the presence of 10 μm atRA, HepG2 cell numbers began to decrease at 60 h, decreased to < 30% at 120 h and this was maintained for > 168 h after serum starvation (Fig. 1A). From 36 h after serum starvation, there is a significant difference between atRA-treated HepG2 cells and controls. Moreover, a dose-dependent protection effect was observed when 5 or 20 μm atRA was also used for treatment. From 72 h after serum starvation, there is a significant difference between the 5 μm atRA-treated group and the 10 or 20 μm atRA-treated group. From 96 h after serum starvation, there is a significant difference between the 10 μm atRA-treated group and the 20 μm atRA-treated group (Fig. 1A).
In HepG3B and Huh7cells, similar protection effects were observed using atRA treatment (Fig. 1B,C), except that HepG3B cells were more sensitive to serum starvation, whereas Huh7 cells were more resistant. Serum starvation resulted in detachment of cells followed by apoptotic cell death, 10 μm atRA treatment was able to prolong cell survival for 4.5, 2.3 and 3.8 days in HepG2, HepG3B and Huh7 cells, respectively (Fig. 1D).
Enhancement of cell adhesion by atRA
In order to explore the mechanism of atRA protection against cell death induced by serum starvation, cell cycle, apoptosis and adhesion assays were performed in HepG2 cells. There were no significant differences in cell cycle (Fig. 2A) and apoptosis (Fig. 2B) assays between groups with or without atRA treatment. In the cell adhesion assay, atRA treatment resulted in an enhancement in the early stage of cell adhesion to the substratum. HepG2 cells treated with 10 μm atRA exhibited ~ 31, 72, 87.5, 96.5 and 100% cell adhesion at 15 min, 30 min, 60 min, 2 h and 4 h after seeding, respectively, which was significantly higher than the percentage of cell adhesion seen in the control group (Fig. 2C). In HepG3B and Huh7 cells, similar enhancements in cell adhesion at the early stage were observed (Fig. 2D,E).
Differentially expressed genes and Gene Ontology analysis
To identify the possible genes involved in the protection effect of atRA, the Illumina Human HT-12 v4 expression microarray was used to compare gene expression between HepG2 cells treated, or not, with 10 μm atRA. In total, 380 differentially expressed genes were identified (fold change ≤ -2 or ≥ +2, P ≤ 0.05), including 207 upregulated and 173 downregulated genes. Thirteen upregulated genes and eight downregulated genes showed changes in expression of more than threefold (Table 1). The most upregulated genes are cytochrome P450 family 26 subfamily A polypeptide 1 (CYP26A1), histidine triad nucleotide binding protein 3 (HINT3), miR-1282 and cytochrome P450 family 26 subfamily B polypeptide 1 (CYP26B1), which showed more than fivefold greater expression in the atRA treatment group.
Table 1. Deregulated genes in RA-treated HepG2 cells
Homo sapiens cytochrome P450, family 26, subfamily A, polypeptide 1 (CYP26A1), transcript variant 2, mRNA.
Homo sapiens histidine triad nucleotide binding protein 3 (HINT3), mRNA.
Homo sapiens microRNA 1282, MIR1282, microRNA.
Homo sapiens cytochrome P450, family 26, subfamily B, polypeptide 1 (CYP26B1), mRNA.
Homo sapiens hemoglobin, alpha 2 (HBA2), mRNA.
Homo sapiens EPH receptor A10 (EPHA10), transcript variant 2, mRNA.
Homo sapiens ribosomal protein L13-like (RPL13L), non-coding RNA.
Homo sapiens energy homeostasis associated (ENHO), mRNA.
Homo sapiens coiled-coil domain containing 101 (CCDC101), mRNA.
Homo sapiens dynein, axonemal, heavy chain 8 (DNAH8), mRNA.
Homo sapiens cyclin I family, member 2 (CCNI2), mRNA.
Homo sapiens small nucleolar RNA, H/ACA box 68 (SNORA68), small nucleolar RNA.
Homo sapiens bradykinin receptor B2 (BDKRB2), mRNA.
Homo sapiens ATG16 autophagy related 16-like 2 (S. cerevisiae) (ATG16L2), mRNA.
Homo sapiens family with sequence similarity 169, member A (FAM169A), mRNA.
Homo sapiens RAB30, member RAS oncogene family (RAB30), mRNA.
Homo sapiens apolipoprotein L, 3 (APOL3), transcript variant beta/a, mRNA.
Homo sapiens InaD-like (Drosophila) (INADL), mRNA.
Homo sapiens solute carrier family 37 (glycerol-3-phosphate transporter), member 3 (SLC37A3), transcript variant 1, mRNA.
Homo sapiens RNA, 5S ribosomal 9 (RN5S9), ribosomal RNA.
The top 15 predominant functions of up- and downregulated genes within the three Gene Ontology (GO) categories (molecular function, cellular component and biological process) were assessed (Table 2). In terms of ‘molecular function’ and ‘cellular component’, extracellular matrix (ECM) was enriched in the upregulated gene lists, as expected. The differentially expressed genes contributing to enriched ECM included ELN, FCN3, FCN1, COL8A2, FREM2, COL11A2, MXRA8 and COL13A1, none of which has been reported to be upregulated by RA treatment in HCC.
Table 2. GOs analysis of deregulated genes
Extracellular matrix structural constituent
Large conductance calcium-activated potassium channel activity
Phosphatase activator activity
Proteinaceous extracellular matrix
Large conductance calcium-activated potassium channel activity
Positive regulation of T-helper cell differentiation
Positive regulation of alpha-beta T-cell differentiation
Phosphate ion binding
Regulation of interleukin-13 secretion
CD8 receptor binding
Regulation of T-helper 2 cell cytokine production
CD4 receptor binding
Positive regulation of interleukin-4 production
Regulation of interleukin-10 secretion
Peptide hormone receptor binding
Positive regulation of interleukin-10 secretion
Regulation of interleukin-5 secretion
Detection of mechanical stimulus involved in sensory perception of pain
T-cell receptor binding
Positive regulation of T-helper 2 cell cytokine production
Positive regulation of interleukin-5 secretion
Positive regulation of interleukin-13 secretion
The expression of COL8A2 was upregulated by atRA
In real-time RT-PCR, COL8A2 was upregulated 4.2-fold in HepG2 cells treated by atRA, whereas COL11A2, MXRA8 and FCN3 were upregulated ~ 2.8–3.2-fold (Fig. 3A), and the other four genes showed no significant difference. In western blotting, COL8A2 was confirmed to be upregulated significantly (4.8-fold) in HepG2 cells treated by RA (Fig. 3B). In HepG3B and Huh7 cells, COL8A2 was also significantly upregulated to a similar extent in RT-PCR and western blotting assays (data not shown).
COL8A2 was involved in enhancement of cell adhesion by atRA
At first, the effect of COL8A2 on cell growth was evaluated in HepG2 cells. The expression of COL8A2 was deceased 3.8-fold in HepG2 cells transfected with 20 nm small interfering RNA (siRNA)–COL8A2 and increased 4.5-fold by transfection of pCOL8A2 in COL8A2 knocked-down HepG2 cells (Fig. 3B). Knockdown of COL8A2 blocked the majority of protection effect by atRA treatment in HepG2 cells, compared with nonrelative siRNA transfected cells (Fig. 3C). To validate the effect of COL8A2 on cell adhesion, an adhesion assay was performed in HepG2 cells transfected with siRNA–COL8A2. As shown in Fig. 3D, knockdown of COL8A2 blocked enhancement of the early stage of cell adhesion by atRA treatment in HepG2 cells, compared with nonrelative siRNA transfected cells. Meanwhile, knockdown of COL8A2 showed no effect on regular HepG2 cell growth or adhesion (Fig. 3C,D). In order to confirm the effect of COL8A2, rescue experiments were performed in cell adhesion and cell growth assays. Re-expression of COL8A2 in COL8A2 knocked-down HepG2 cells reversed the effect of siRNA–COL8A2 (Fig. 3C,D). In HepG3B and Huh7 cells, similar results were obtained. These results indicated that COL8A2 is most likely involved in enhancement of cell adhesion by atRA.
COL8A2 increased HCC cell migration and invasion
To verify the importance of cell adhesion enhancement, the effects of COL8A2 on the migration of HCC cells were investigated. HepG2 cells that migrated to the lower surface of the filter were increased 31.6% by pCOL8A2 transfection and decreased 30.2% by siRNA–COL8A2 transfection (Fig. 4A). Moreover, to substantiate this observation, a Matrigel invasion assay in transwell culture chambers was carried out to determine the effect of COL8A2 on the invasion of HCC cells. HepG2 cells that passed through the Matrigel were increased 77.8% for pCOL8A2 transfection and decreased 59.6% for siRNA–COL8A2 transfection (Fig. 4B). Together, these results support a critical role for COL8A2 in the invasion of HCC cells. Upregulation of COL8A2 expression may be involved in the invasive phenotype of HCC cells by promoting cell migration and adhesion.
atRA has been shown to possess antitumoral effects by inhibiting proliferation, inducing apoptosis and/or differentiation , so that natural and synthetic retinoids have been used in chemoprevention and differentiation therapy [2, 18]. Surprisingly, it was found that atRA could prolong the survival of HCC cells (HepG2, Hep3B and Huh7) under serum-free conditions, which is quite different from the classic inhibitory effect of atRA. Using serum starvation, it was possible to verify whether HCC growth is dependent to a greater degree on nutrients and growth factors found in the environment. Serum starvation would provide a good model of the nutrient-deprived tumor . In addition, HCC cells invade interstitial connective tissue and proliferate, and serum starvation has been used as a model of the microenvironment of interstitial connective tissue . atRA was reported to promote migration and invasion in neuroblastoma SH-SY5Y cells . Thus, this protection effect of atRA might be involved in tumor invasion or provide evidence for the side effects of atRA treatments.
Further studies excluded the effect of atRA on the cell cycle or apoptosis but validated the enhancement of cell adhesion of atRA under serum-free conditions. atRA was reported to affect cell–ECM interactions by increasing adhesion to the substrate and the expression of integrin receptors in normal and transformed cells [22, 23]. To identify possible new genes that were involved in the effect of atRA under serum-free conditions, a human mRNA expression microarray was used to screen the differentially expressed genes in HepG2 cells treated with atRA. Among the upregulated genes, CYP26A1, HINT3, miR-1282 and CYP26B1 showed more than fivefold greater expression in the atRA treatment group. The cytochrome P450 enzymes CYP26 are believed to partially regulate cellular concentrations of atRA via oxidative metabolism and hence affect retinoid homeostasis and signaling . As the atRA hydroxylases, CYP26A1 and CYP26B1 were upregulated to metabolize atRA during atRA treatment. HINT3 has been classified as a member of the histidine triad nucleotide binding protein subfamily, which was found to be associated with Cdk7 and Kin28 . miR-1282 was a newly identified microRNA, which was only reported to be deregulated in myocardium of heart failure patients on left ventricular assist device support . The significant upregulation of HINT3 and miR-1282 expression in atRA-treated HepG2 cells indicated possible biological significance, which is worthy of further study. The integrin family and laminin which were reported to be involved in the effect of RA on the cell adhesion [27-29] showed no significance in this study, probably because of the different serum conditions.
Using GO analysis, the greatest significance was the enrichment of ECM-related molecular functions and cellular components in upregulated genes, which was consistent with the enhancement of cell adhesion found using atRA. Eight ECM-related genes, ELN, FCN3, FCN1, COL8A2, FREM2, COL11A2, MXRA8 and COL13A1, were found to be upregulated in microarray analysis. The upregulation of COL8A2, COL11A2, MXRA8 and FCN3 was confirmed by real-time RT-PCR and the most significant upregulation of COL8A2 was further confirmed by western blotting in all three HCC cells. These data indicated that atRA may affect ECM remodeling by upregulating COL8A2 and other molecules to prolong the cell survival.
Collagen type VIII, a nonfibrillar short-chain collagen, is a structural component of many ECM [22, 30]. Two highly homologous polypeptides, COL8A1 and COL8A2, form either homotrimeric or heterotrimeric molecules [31, 32]. Type VIII collagen is involved in crosstalk between cells and the surrounding matrix by modulating diverse cellular responses such as proliferation, adhesion, migration, chemotaxis and metalloproteinase synthesis [33, 34]. Increasing evidence indicates that collagen type VIII plays a role in atherogenesis and vascular remodeling, presumably through modification of cell migration and proliferation [35, 36]. The short-chain type VIII collagen was thought to play a key role in angiogenesis and embryonic development of the heart and glomerulonephritis [37-39]. In this study, it was shown that the expression of COL8A2 was significantly upregulated in all three HCC cells by treatment with 10 μm atRA. Using knockdown and re-expression of COL8A2, it was confirmed that COL8A2 was involved in the enhancement the cell adhesion, and thus the protection effect of atRA. Furthermore, COL8A2 was found to be involved in the HCC cell migration and invasion, whereas atRA treatment also showed enhancement of HCC cell migration and invasion, which provided a new understanding of the function of COL8A2. Of course, the other three upregulated genes, COL11A2, MXRA8 and FCN3, may be involved in the effect of atRA, which requires further study.
The physiological level of atRA is ~ 2.7–4.2 ng·mL−1 . The concentration of 5–20 μm reported in this study is three orders of magnitude above the physiological concentration. Although some researchers have reported the effect of physiological concentrations of atRA , 10 μm atRA is commonly reported for tumor cell lines treatments. The concentration of compounds used in in vitro assays is usually higher than the physiological level in vivo. One of possible reason might be that the cell lines (usually from tumor tissues) are more resistant than primary cells. In addition, the effect of any compound depends on the variety of cell lines, some cell lines are sensitive to the drug but others are resistant. As shown in this assay, < 5 μm atRA has no significant effect on HepG2 and Huh7 cells, but 1 μm atRA showed large effects on HepG3B cells.
In conclusion, atRA protects HCC cells against serum-starvation-induced cell death and increases cell migration and invasion by enhancing cell adhesion, and COL8A2 plays an important role in this progress. This study might provide evidence for the effect of atRA on HCC cells and deepen our understanding of HCC.
Material and methods
Cell culture and treatment
HepG2 cells were kindly provided by Mark Feitelson (Temple University, Philadelphia, PA, USA), Huh7 and HepG3B cells were kindly provided by Zhong Jin (Institut Pasteur of Shanghai, Shanghai, China). All three HCC cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum, 100 U·mL−1 penicillin, 100 U·mL−1 streptomycin and 2 mm glutamine (Invitrogen, Carlsbad, CA, USA) at 37 °C in a humidified atmosphere containing 5% CO2. For HepG2 cells, culture plates were precoated with 1 mg·mL−1 rat tail collagen type I (BD Bioscience, Franklin Lakes, NJ, USA) for 2 h at room temperature and washed with culture medium once. For the serum starvation, cells were seeded in culture medium without fetal calf serum for the indicated times. For the atRA treatment, atRA (Sigma-Aldrich, St. Louis, MO, USA) was dissolved in dimethylsulfoxide at 100 mm, stored at −80 °C, and diluted to the indicated concentrations before use. The final concentration of dimethylsulfoxide in the assays was < 0.03%.
Cell growth assay
HepG2, Huh7 and HepG3B cells (1 × 104) were seeded in a 24-well plate in triplicate under serum-free conditions with or without atRA for 7 days. At the indicated time, cells in five random 20× fields were counted using a digital microscope system (IX81; Olympus, Tokyo, Japan). Finally, the attached cells were fixed with 4% paraformaldehyde for 5 min, and then stained with 0.1% Coomassie Brilliant Blue R250 at room temperature for 20 min and photographed.
HepG2, Huh7 and HepG3B cells (1 × 105) were harvested, washed once in NaCl/Pi and fixed in 70% ethanol at 4 °C overnight. Staining for DNA content was performed with 50 mg·mL−1 propidium iodide and 1 mg·mL−1 ribonuclease A at room temperature for 30 min. The cell populations in the G0–G1, S and G2–M phases were measured using Cell Lab Quanta SC flow cytometry (Beckman Coulter, Fullerton, CA, USA) and the data analyzed with flowjo v7.6 software (Tree Star, Inc., Ashland, OR, USA).
HepG2, Huh7 and HepG3B cells (1 × 105) were harvested, washed once in NaCl/Pi, resuspended and incubated with FITC Annexin V (Promega, Madison, WI, USA) for 15 min at 4 °C in the dark, according to the manufacturer's instructions. After staining, cells were incubated with propidium iodide for 5 min at 4 °C in the dark and then analyzed using Cell Lab Quanta SC flow cytometry and the data were analyzed by flowjo v7.6 software.
After treatment for 24 h with or without 10 μm atRA under serum-free conditions, 1 × 104 HepG2, Huh7 and HepG3B cells were harvested and seeded in a fresh 24-well plate for the indicated times of 15 min, 30 min, 60 min, 2 h, 4 h, 8 h, 12 h or 24 h, unattached cells were rinsed gently with culture medium, and the remaining attached cells were counted using a digital microscope system (IX81; Olympus). At 30 min, 60 min, 2 h and 8 h, the remaining attached cells were also were fixed with 4% paraformaldehyde for 5 min, and then stained with 0.1% Coomassie Brilliant Blue R250 at room temperature for 20 min and photographed.
Total RNA extraction
Total RNA was isolated using TRI Reagent combined with the RNeasy Tissue kit protocol (Qiagen, Valencia, CA, USA) according to the manufacturer's recommendations. The RNA concentrations and the A260/A280 ratio were assessed with a multiplate reader (Synergy 2; BioTek, Winooski, VT, USA). The 28S/18S ratio and the RNA integrity number were assessed with a Bioanalyzer 2100 (Agilent Technologies, Wilmington, DE, USA). An A260/A280 ratio of 1.9, a 28S/18S ratio of 1.8 and an RNA integrity number of 5 were the minimum requirements for inclusion in the expression analysis.
Gene expression profiling and analysis
The expression levels of HepG2 cells treated (or not) with atRA or analyzed using the Illumina Human HT-12 v4 expression microarray (Illumina, San Diego, CA, USA) platform. In brief, 500 ng of total RNA was amplified and biotin-labeled with the Illumina TotalPrep-96 RNA Amplification Kit (Ambion, Austin, TX, USA). A total of 750 ng of labeled complementary RNA was hybridized to Illumina's HumanHT-12 v4 expression BeadChips and then imaged using a BeadArray Reader according to manufacturer's instructions.
The raw data were obtained with bead studio software (Illumina). Data quality was assessed based on the positive and negative control probes on each array, and by inspection of the probe intensity distributions. Data were normalized using the quantile normalization method. A moderated t-test implemented in the limma library of bioconductor (Bioconductor, Seattle, WA, USA) was applied to test differential expression, and a false discovery rate adjustment of the P-value was performed to correct for multiple testing. Probes were considered significantly different if the adjusted P-value was < 0.05 and the fold change difference between groups was at least 2.
The total differentially expressed genes were subjected to GO. GO was applied to analyze the main function of the differential expression genes according to the GO, which is the key functional classification of National Center for Biotechnology Information. In general, Fisher's exact test and v2 test were used to classify the GO category, and the false discovery rate was calculated to correct the P-value. Values of P were computed for the GOs of all the differential genes. Enrichment provides a measure of the significance of the function, which helps us to find those GOs with more concrete function description in the experiment.
Real-time quantitative PCR
To verify the microarray results, eight upregulated ECM-related genes, elastin (ELN), ficolin 3 (FCN3), ficolin 1 (FCN1), COL8A2, FRAS1-related extracellular matrix protein 2 (FREM2), collagen 11A2 (COL11A2), matrix remodeling associated 8 (MXRA8) and collagen 13A1 (COL13A1), were selected for real-time quantitative PCR assays. The primers are listed in Table S1. Total RNA (2 μg) was reverse transcribed with the SuperScript III First-Strand Synthesis System with oligo(dT) (Invitrogen). SybrGreen RT-PCR kit (Takara Bio, Shiga, Japan) was used to validate expression analysis data. Reactions were incubated in a 96-well optical plate at 95 °C for 5 min, followed by 40 cycles at 95 °C for 15 s and 60 °C for 1 min. PCRs were run on a StepOne Plus real-time PCR machine (Applied Biosystems, Foster City, CA, USA) and the data were analyzed by sds v2.3 software (Applied Biosystems, Foster City, CA, USA). GAPDH was used as the control.
The full-length of COL8A2 (Genbank accession no. NM_005202.2) was amplified with the primers 5′-GGACGCCATGCTGGGGACTCTGAC-3′ (forward) and 5′-TGTGGGGCAGAGCAAGAATCCTGAAAA-3′ (reverse). The PCR product was cloned into mammalian expression vector pcDNA™3.1/myc-his (Invitrogen), confirmed by sequencing analysis and termed pCOL8A2.
Transfection was carried out using FuGene HD transfection reagent (Roche, Indianapolis, IN, USA) following the manufacturer's protocol. In brief, 2 × 104 HepG2, HepG3B and Huh7 cells in 24-well plate were transfected with the indicated plasmid DNA and siRNA (GenePharma, Shanghai, China) and collected 24–48 h after transfection for assay. COL8A2 expression was confirmed by western blotting.
Cell migration and invasion assays
For the cell migration assay, 2 × 105 HepG2 or Huh7 cells treated with atRA or transfected with pCOL8A2 or siRNA–COL8A2 were seeded in the upper chamber of transwell units (Corning Inc, Corning, NY, USA) with 8 μm pore size polycarbonate filter under serum free condition. The lower chamber was filled with 500 μL Dulbecco's modified Eagle's medium containing 10% fetal bovine serum. After incubation for 24 h, cells on the upper surface of the filter were completely removed by wiping with a cotton swab. Then the filters were fixed with 4% paraformaldehyde and stained with 0.1% Coomassie Brilliant Blue R250 for 20 min. The number of cells that migrated through the pores to the lower surface of the filter was counted and analyzed with a digital microscope system (IX81; Olympus). Triplicate samples were acquired and the data were expressed as the average cell number of five fields. For the cell invasion assay, the similar protocol of cell migration assay was used except that the transwell units were precoated with 200 μg·mL−1 Matrigel (BD Biosciences) and incubated overnight. Cells that had invaded the Matrigel and reached the lower surface of the filter were counted.
Western blotting analysis
Protein extracts from HepG2 cells were prepared using a modified RIPA buffer with 0.5% SDS in the presence of proteinase inhibitor cocktail (Complete mini; Roche). Fifty micrograms of proteins were electrophoresed in 10% SDS/PAGE mini gels and transferred onto poly(vinylidene difluoride) membranes (Immobilon P-SQ; Millipore, Billerica, MA, USA). After blocking with 5% nonfat milk, the membranes were incubated with rabbit anti-COL8A2 Ig (1 : 1000 dilution; Abcam, Cambridge, UK) or mouse anti-GAPDH Ig (1 : 5000 dilution; Anbobio, ChangZhou, China) at 4 °C overnight, followed by incubation with horseradish peroxidase-conjugated goat anti-rabbit or goat anti-mouse Ig (1 : 10 000 dilution; KPL, Gaithersburg, MA, USA) for 1 h at room temperature. Finally, signals were developed with Super Signal West Pico chemoluminescent substrate (Pierce, Rockford, IL, USA) and visualized using the GeneGnome HR Image Capture System (Syngene, Frederick, MD, USA).
All data are from at least three independent experiments. The data are presented as the mean ± SD. Comparisons were made by using a two-tailed t-test or one-way analysis of variance for experiments with more than two subgroups. P < 0.05 was considered statistically significant.
We thank research assistants at Genergy Biotech for their assistance in Illumina HumanHT-12 v4 expression microarray performance and Gene Ontology analysis. This work was supported by National S&T Major Project for Infectious Diseases Control (2012ZX10002009-004 and 2012ZX10002003-004-010), Shanghai Leading Academic Discipline Project (B901) and National Key Basic R&D (973) Program of China (2009CB522503); National Science Foundation of China (30921006).