Gene expression profiles in an hepatitis B virus transfected hepatoblastoma cell line and differentially regulated gene expression by interferon-α

Abstract

AIM: To study interactions between hepatitis B virus (HBV) and interferon-α in liver-derived cells.

METHODS: mRNAs were separately isolated from an HBV-transfected cell line (HepG₂2.2.15) and its parental cell line (HepG₂) pre- and post-interferon-α (IFN-α) treatment at 6, 24 and 48 h, followed by hybridization with a cDNA microarray filter dotted with 14 000 human genes. After hybridization and scanning of the arrays, the data were analyzed using ArrayGauge software. The microarray data were further verified by Northern blot analysis.

RESULTS: Compared to HepG₂ cells, 14 genes with known functions were down-regulated 3 to 12-magnitudes, while 7 genes were up-regulated 3-13 magnitudes in HepG₂2.2.15 cells prior to IFN-α treatment. After interferon-α treatment, the expression of four genes (vascular endothelial growth factor, tyrosine phosphate 1E, serine protein with IGF-binding motif and one gene of clathrin light chain) in HepG₂2.2.15 were up-regulated, while one gene encoding a GTP-binding protein, two genes of interferon-induced kinases and two proto-oncogenes were further down-regulated. Interestingly, under IFN-α treatment, a number of differentially regulated genes were new ESTs or genes with unknown functions.

CONCLUSION: The up-regulated genes in HepG₂2.2.15 cell line suggested that under IFN-α treatment, these repressed cellular genes in HBV infected hepatocytes could be partially restored, while the down-regulated genes were most likely the cellular genes which could not be restored under interferon treatment. These down-regulated genes identified by microarray analysis could serve as new targets for anti-HBV drug development or for novel therapies.

Key Words: $[Keywords]

INTRODUCTION

Viral hepatitis B continues to be a significant global problem. It is estimated that there are approximately 350 million chronic hepatitis B virus (HBV) carriers world-wide. Furthermore, chronic hepatitis B patients are at high risk of developing liver cirrhosis and hepatocellular carcinoma with high mortality rates.

In the past two decades, interferon-α (IFN-α) has proven effective in the treatment of chronic hepatitis B patients. However, sustained responses were observed in only one-third of chronic hepatitis B patients[1-3]. In the normal physiological state, interferon-α is expressed at a low level and is induced to high levels by viral or bacterial infections and the exposure to double-stranded RNA[4]. In contrast to other types of viral infections, IFN-α was either non-detectable or present in extremely low levels in HBV chronically infected patients[5,6]. Additionally, decreased synthesis of 2’-5’ A synthetase, belonging to a family of enzymes induced by IFNs, has been reported in chronic hepatitis B patients[7,8]. To date, mechanisms underlying defective production of IFN or defective responses to IFN in chronic hepatitis B patients have not been fully elucidated.

Development of microarray technology has provided a powerful tool for study of the complicated biological process in cells which results in altered global gene expression. cDNA microarray has been used to analyze virus-host cell interactions in Human Immunodeficiency Virus (HIV), Human Cytomegalovirus (HCMV), Herpes Simplex Virus (HSV) and Influenza Virus infected cell cultures[9-13]. By comparing virus infected cell cultures with non-infected cell cultures, a number of differentially expressed genes were identified. For HBV, although several studies have been undertaken to clarify the effects of HBV infection on hepatocytes[14-16], the global effects of all HBV proteins or HBV replication on host cells, especially the interaction between HBV and interferon remain unclear. Therefore, in this study, cDNA microarray filters dotted with 14 000 human genes were used to analyze transcriptional changes between an HBV DNA-transfected cell line (HepG₂2.2.15)[17,18] and its parental cell line (HepG₂) pre- and post- IFN treatment. Based on analysis of altered mRNA expression, several IFN-differentially regulated genes including new ESTs or genes with unknown functions were further investigated for their potential antiviral activity.

MATERIALS AND METHODS

Cell culture

The HBV DNA transfected hepatoblastoma cell line HepG₂2.2.15, that has been confirmed to produce infectious virus[18,19], and its parental cell line HepG₂ were maintained in Dulbecco’s modified Eagle’s medium supplemented with 10 mL/L fetal calf serum and antibiotics. The cells were incubated at 37 °C in 50 mL/LCO₂. Cell viability was estimated by the trypan blue dye exclusion method.

Preparation of cell samples for cDNA microarray

The cells were separately seeded into T-75 flasks at 40% confluence (4×104 cells/cm²) one day prior to the experiment. After overnight culture, two flasks of cells for each cell line were harvested before treatment with IFN-α for RNA preparation and were regarded as 0 h cells for IFN-α treatment. The other cells were exposed to 5×10³ IU/mL of IFN-α (Calbiochem, San Diego, CA, USA) and harvested 6, 24 and 48 h after IFN-α treatment. The culture media were removed and stored at -70 °C. For each flask of cells, 5 mL of TRIzol Reagent (Gibco BRL) was added directly into the flask.

RNA extraction

Total RNA was extracted using standard TRIzol RNA isolation protocol (Gibco BRL). The isolated total RNA was divided into two aliquots, one for RNA slot analysis and the other for mRNA preparation.

Generation of microarray

Generation of microarrays and microarray analysis were carried out by the Institute of Biochemistry and Cell Biology, CAS as described previously[20]. Human cDNA clones were derived from liver, hepatocarcinoma cell lines and hypothalamus-pituitary-adrenal libraries[21] or purchased from Research Genetics (Huntsville, AL, USA). A cDNA array was assembled with 14 000 cDNA clones representing the same number of independent cDNA clusters, of which 7 565 clusters were homologous to that in the UniGene Database. All cDNA fragments were amplified and verified by gel electrophoresis and sequencing. The average length of the cDNA fragments was -1 kb. PCR products were precipitated in isopropanol, redissolved in 10 μL of denaturing buffer (1.5 mol/L NaCl, 0.5 mol/L NaOH) and spotted onto 8×12 cm Hybond-N nylon membranes (Amersham Pharmacia, Buckinghamshire, UK) using an arrayer (BioRobotics, Cambridge, UK). Each spot carried -100 nL in volume and was 0.4 mm in diameter; each cDNA fragment was spotted as two different spots (double-offset). Lambda phage and pUC18 vector DNA were spotted as negative controls.

Eight housekeeping genes encoding ribosomal protein S9 (RPS9), β-actin (ACTB), glyceraldehydes-3-phosphate dehydrogenase, hypoxanthine phosphoribosyltransferase 1, M_r 23 000 highly basic protein (RPL13A), ubiquitin C, phospholipase A2, and ubiquitin thiolesterase (UCHL1) were each evenly distributed in 12 spots on each 8×12 cm array as an intramembrane control. Hybridization data were considered invalid if the intensity of the darkest spot exceeded 1.5 times that of the weakest spot among the 12 spots representing the same gene.

mRNA isolation and probe preparation

The poly (A)⁺ mRNA was isolated from the total RNA using a poly (dT) resin (Qiagen, Hilden, Germany). Approximately 1-2 μg of mRNA were labeled in a reverse transcription reaction in the presence of 200 μCi [α³³P] deoxyadenosine 5’-triphosphate (DuPont NEN, Boston, MA, USA) using Moloney murine leukemia virus reverse transcriptase as recommended by the manufacturer (Promega Corp., Madison, WI, USA).

Hybridization and Image Procession

Prehybridization was carried out in 20 mL of prehybridization solution (6×SSC, 5 g/L SDS, 5×Denhardt’s and 100 μg/mL denatured salmon sperm DNA) at 68 °C for 3 h. Overnight hybridization with the ³³P-labeled cDNA in 6 mL of hybridization solution (6×SSC, 5 g/L SDS, and 100 μg/mL denatured salmon sperm DNA) was followed by stringent washing (0.1×SSC, 5 g/L SDS, at 65 °C for 1 h). Membranes were exposed to Phosphor Screen overnight and scanned with an FLA-3 000A Plate/ Fluorescent Image Analyzer (Fuji Photo Film, Tokyo, Japan). Radioactive intensity of each spot was linearly digitalized to 65 536 gray-grade in a pixel size of 50 μm in the Image Reader and recorded using the Array Gauge software (Fuji Photo Film, Tokyo, Japan). After subtraction of background (3 ± 3) chosen from an area where no cDNA was spotted, genes with intensities > 10 were considered as positive signals to ensure that they were distinguished from background with statistical significance > 99.9%. Normalization among arrays was based on the sum of background-subtracted signals from all genes on the membrane[22].

RNA slot analysis

Total cellular RNA, 10 μg/sample was mixed with 7 μL formaldehyde, 20 μL formamide and 2 μL 20×SSC (1×SSC = 0.15 mol/L NaCl plus 0.015 mol/L sodium citrate). After being denatured at 65 °C for 15 min and immediately cooled in an ice bath, the RNA sample was applied to the nylon membrane (Boehringer Mannheim, Germany) with Minifold I (Schleicher & Schuell). After fixation at 120 °C for 30 min, the membrane was pre-hybridized in 20 mL of prehybridization solution (6×SSC, 5 g/L SDS, 5×Denhardt’s and 100 μg/mL denatured salmon sperm DNA) at 42 °C for 6 h, and followed by hybridization with 20 mL hybridization solution (6×SSC, 5 g/L SDS, 5×Denhardt’s and 100 μg/mL denatured salmon sperm DNA) including [α-³²P]dCTP labeled cDNA probes (Random labeling kit, Boehringer Mannheim, Germany) at 42 °C for 16 h. The cDNA probes represented the unique sequence of IFN related P27, KIAA0919, beta3gal-T5 and FLJ12673. After stringent membrane washing process (0.1×SSC, 5 g/L SDS at 68 °C for 30 min, thrice), autoradiographs were exposed to X-ray film (Fijifilm, Japan) overnight at -70 °C. Quantitative analysis was undertaken by scanning the intensity of the membranes. In order to normalize the total RNA quantity in each blot, an [α³²-P]dCTP labeled β-actin probe was hybridized to the membrane again after the first probe was removed. The change of regulated gene expression between different cell samples could be detected by comparing the intensity ratio of specific gene and β-actin.

RESULTS

Validation of microarray results

Similar to previous study, results from the reproducible analysis confirmed that the replicated experiments were in concordance with an R² (square of Pearson correlation coefficiency, measuring similarity in gene expression pattern) of 0.97-0.98 on scatterplot[20].

To confirm the altered expression level, 4 genes (interferon related genes P27, KIAA0919, β-3-gal-T5 gene and FLJ12673 human cDNA clone) were selected and tested by RNA slot analysis (Figure 1). The ratios of the signal intensity of specific genes and β-actin and their comparison with microarray result are listed in Table 1. It is observed that the results of RNA slot analysis are aligned with those from the microarray analyses (Figure 1 and Table 1).

Table 1 Comparison of microarray and RNA slot analyses on 4 selected genes.

Accession	X67325		AB023136		AA622328		NM_018300
ID	Ratio of arrays	Ratio of slot	Ratio of arrays	Ratio of slot	Ratio of arrays	Ratio of slot	Ratio of arrays	Ratio of slot
6 h/0 h	3.6	Undetectable	3.4	1.3	2.0	0.5	1.1	2.9
24 h/0 h	5.9	> 31	7.2	3.5	3.9	1.6	4.9	3.4
48 h/0 h	6.1	> 33	8.4	6.2	3.8	5.7	1.0	1.3
Description	IFN-inducible P27		Sec15B(vesicle traffic protein)		beta3gal-T5 gene		cDNA FLJ12673

Figure 1 RNA slot analyses of IFN-inducible P27, KIAA0919 (Sec15B), beta3gal-T5 and FLJ12673 genes in HepG₂ 2. 2.15 be-fore and after IFN treatment Four genes (interferon related genes P27, KIAA0919, β-3-gal-T5 gene and FLJ12673 human cDNA clone) were selected and tested by RNA slot analysis to confirm the altered expression level indicated by the results of the microarrays.10 μg/sample total cellular RNA was blotted to the nylon membrane. The membrane was then hybridized with [α-³²P]dCTP labeled cDNA probes. Autoradiographs were exposed to X-ray film overnight at -70 °C. The β-actin gene was used as the control to normalize the total RNA quantity.

Differentially expressed genes in HepG₂2.2.15 cell line and HepG₂ cell line prior to IFN-α treatment

To study the basic differentiated expressed genes between HepG₂ 2.2.15 cell line and its parental cell line HepG_2, systematic comparison of gene expression profiles from both cell lines before treatment with IFN-α was done at 0 h. Totally 89 genes were either up- or down-regulated three-fold (Table 2). In these 89 genes, 22 genes have defined functional activities (Table 3) while 67 genes were new ESTs with unknown functions. Fifteen of the 22 known genes were down-regulated by three- to twelve-fold, while 7 genes were up-regulated by three- to thirteen-fold. 4 genes could be catalogued in four classes, such as cell cycle (Y13467), ligands and receptors (AF022375), phosphatase (U12128), and transcription factors (D87258).

Table 2 Gene expression profiling in the microarray analysis.

Ratio	Gene differences in HepG22.2.15			Gene differences in HepG2			Gene differences in HepG22.2.15/HepG2
	6 h/0 h	24 h/0 h	48 h/0 h	6 h/0 h	24 h/0 h	48 h/0 h	0 h	6 h	24 h	48 h
3	34	48	45	83	103	48	89	40	33	36
5	10	18	12	37	37	23	27	12	8	11

Table 3 Differently expressed known genes in HepG₂ 2.2.15 and HepG₂ cell lines before IFN-α treatment.

Accession ID	HepG2	HepG22.2.15	Ratio	Description	Unigene
Cell cycle
Y13467	101.95	29.54	-3.45	RB18A	Hs.15589
Ligands and receptors
AF022375	16.17	5.19	-3.12	vascular endothelial growth factor	Hs.73793
Phosphatase
U12128	152.03	27.26	-5.58	tyrosine phosphatase 1E (PTP1E)	Hs.211595
Transcription factors
D87258	156.76	37.78	-4.15	serin protease with IGF-binding motif	Hs.75111
Miscellaneous
X69962	242.93	73.77	-3.29	FMR-1	Hs.89764
X56597	1.64	21.98	13.40	humFib mRNA for fibrillarin	Hs.99853
U76713	403.90	123.59	-3.27	apobec-1 binding protein 1	Hs.81361
U50410	189.94	27.03	-7.03	heparan sulphate proteoglycan (OCI5)	Hs.119651
U23070	5.35	16.82	3.14	putative transmembrane protein (nma)	Hs.78776
S68805	4.12	12.74	3.09	L-arginine:glycine amidinotransferase	Hs.75335
M22430	13.63	50.17	3.68	RASF-A PLA2	Hs.76422
M16961	221.65	61.50	-3.60	alpha-2-HS-glycoprotein alpha and beta chain	Hs.75430
M13928	21.15	163.32	7.72	delta-aminolevulinate dehydratase	Hs.1227
L40157	14.58	92.98	6.38	endosome-associated protein (EEA1)	Hs.2864
L24123	1 309.05	430.92	-3.04	NRF1 protein (NRF1)	Hs.83469
D00096	7.61	31.28	4.11	Prealbumin	Hs.194366
AJ006835	150.29	49.21	-3.05	U17 small nucleolar RNA host gene	Hs.196769
AF073308	237.11	78.45	-3.02	nonsyndromic hearing impairment protein (DFNA5)	Hs.13530
AF050127	823.97	71.21	-11.57	hypoxia inducible factor (aHIF)	Hs.197540
AC002045	237.87	67.71	-3.51	Chromosome 16 BAC clone CIT987SK-A-589H1	Hs.253182
AB027013	15.62	2.31	-6.76	Nucleosome Assembly Protein 1-like 2	Hs.66180
AB023136	544.25	44.18	-12.32	KIAA0919	Hs.44175

Differentially regulated genes in HepG₂2.2.15 cell line and HepG₂ after IFN-α treatment

Six hours after treatment with interferon-α, significant changes in mRNA expression were observed in both HepG₂ and HepG₂2.2.15 cells, with 83 and 34 genes up- or down-regulated by at least three-fold (Table 2). A moderate increase in changes of cellular gene expression was detected at 24 h, at which time, 103 and 48 genes were up- or down-regulated respectively by at least three-fold (Table 2). Approximately half of these genes could be classified in functional categories, including interferon responsive proteins, signal transducing genes such as kinase, phosphatase and ligand as well as receptors (Table 4). It is interesting to note that the number of differentially expressed cellular genes in HepG₂ and HepG₂2.2.15 cell lines could be decreased by interferon-α treatment from 89 to 40, 33 and 36 genes at 6, 24, 48 h post treatment (Table 2). The differentially regulated gene expressed profile in HepG₂ 2.2.15 and HepG₂ cell lines after IFN-α treatment are available at http://www.mvlab-fudan.cn/microarray databases.

Table 4 Functional catalogues of 3 fold regulated genes in HepG₂ and HepG₂ 2.2.15 cell lines post IFN-α treatment.

Catalogue	3-fold regulated gene number in HepG22.2.15			3-fold regulated gene number in HepG2
	6 h/0 h	24 h/0 h	48 h/0 h	6 h/0 h	24 h/0 h	48 h/0 h
Cell cycle	-	1	1	-	-	-
Coagulation	-	1	1	-	1	-
Cytokine/receptors	1	2	2	1	1	-
Extracellular matrix and cell adhesion	-	-	1	-	1	-
GTP binding proteins and signal transduction	2	-	1	-	1	1
Interferon	8	6	3	6	5	5
Kinase and phosphatase	-	-	1	-	2	-
Ligands and receptors	-	1	2	-	-	-
Mitochondrial	-	1	1	-	2	-
Phosphatase and phosphodiesterase	2	2	2	1	2	1
Protooncogenes	-	1	-	1	-	-
Transcription factors	-	-	-	1	1	1
Viral and clathrin receptors	-	1	1	-	1	-
Miscellaneous	7	8	8	14	18	13
New ESTs	14	24	21	59	68	27
Total	34	48	45	83	103	48

Kinetics of IFN-treated gene expression profile in 23 genes from HepG₂ and HepG₂2.2.15 cell lines

Twenty three selected genes, belonging catalogues of interferon induced proteins kinase/phosphatase, ligands and receptors, phosphatase, protein degradation, viral and clathrin receptors, protooncogenes and transcription factors respectively, were further compared between HepG₂2.2.15 and its parental cell line HepG₂ to study the effect of interferon-α treatment on HBV and cellular gene expressions (Table 5). From this table, it is noted that most of these selected genes with known functions have two- three-fold expression regulated in both HepG₂ cell line and HepG₂2.2.15 cell post interferon treatment. However, five genes (U02680, U07358, U52969, U50529 and U70321), were up-regulated only in HepG2 cell line, while down-regulated in HepG₂2.2.15 cells.

Table 5 Kinetics of IFN-treated gene expression profiles in 23 selected genes from HepG₂ and HepG₂2.2.15 cells.

Accession ID	HepG22.2.15 cell lines			HepG2 cell lines			Description
	6 h/0 h	24 h/0 h	48 h/0 h	6 h/0 h	24 h/0 h	48 h/0 h
Interferon
X57352	3.8	3.1	2.5	6.5	8.1	6.4	1-8U gene from interferon-inducible gene family
X57351	3.4	2.2	2.2	4.3	4.4	4.4	1-8U gene from interferon-inducible gene family
M87503	2.3	1.7	1.7	3.2	2.4	2.6	IFN-responsive transcription factor subunit
M33882	2.4	1.6	1.5	2.5	2.1	1.6	p78 protein
M11810	9.1	5.7	4.4	8.4	7.7	3.9	(2’-5’) oligo A synthetase E gene
L07633	2.0	1.3	1.5	2.0	2.7	1.3	(clone 1950.2) interferon-gamma IEF SSP 5111
J04164	6.7	5.0	2.9	6.4	7.0	6.5	Interferon-inducible protein 9-27
D28137	3.0	3.3	2.8	2.7	2.6	1.8	BST-2
AJ225089	3.9	1.3	1.1	1.9	2.1	1.2	2’-5’ oligoadenylate synthetase 59 kDa isoform
Kinase/Phosphate
U07358	-1.3	-1.3	-2.5	1.1	2.4	1.4	Protein kinase (zpk)
U02680	-1.1	1.4	-2.5	1.5	7.1	1.6	Protein tyrosine kinase
AF032437	-1.1	-2.5	-1.7	-1.3	-3.3	-2.0	Mitogen activated protein kinase activated protein kinase
Ligands and receptors
D86096	2.3	4.6	3.5	1.9	1.8	2.3	Prostaglandin E receptor EP3 subtype
AF022375	1.8	1.8	3.2	-1.2	-1.1	1.1	Vascular endothelial growth factor
Phosphatase
U12128	3.0	4.8	4.7	-1.2	-1.1	-1.2	Protein tyrosine phosphatase 1E (PTP1E)
Protein degradation
Z14977	2.2	1.2	1.4	2.7	1.3	1.6	Major histocompatibility complex encoded
							proteasome subunit LMP2
AF061736	2.6	1.5	1.4	2.3	2.3	2.0	Ubiquitin-conjugating enzyme RIG-B
Protooncogenes
U52969	1.1	-2.0	-1.3	2.2	1.6	2.1	PEP19 (PCP4)
U50529	-2.5	-5.0	-2.5	1.7	1.5	2.2	BRCA2 region, mRNA sequence CG016
Transcription factors
D87258	1.6	2.3	2.5	1.0	-1.1	-1.5	Serin protease with IGF-binding motif
Viral and clathrin receptors
X81636	1.1	5.5	5.2	-1.1	6.4	-1.7	Clathrin light chain a
U70321	-1.3	-2.0	1.1	2.0	1.2	1.6	Herpesvirus entry mediator
AF079221	1.0	1.2	2.9	1.7	2.2	1.9	BCL2/adenovirus E1B 19kDa-interacting protein 3 a

DISCUSSION

DNA microarray is an important and powerful tool for surveying changes in cellular gene expression on a large scale. Several groups have used this approach to analyze virus-host cell interactions in HBV infection. Rho et al[15] constructed a HepG₂ cell line stably expressing HBx (HepG2-HBx), and performed cDNA microarray analysis on 588 cellular cDNAs to study the effect of HBx on the transcriptional regulation in the human liver cell. To gain more information on how HBV proteins regulate host cellular gene expression, Otsuka et al[16] examined the differences in the gene expression profiles of HepG₂2.2.15 and HepG₂ cells, using an in-house cDNA microarray dotted with 2 304 cellular genes. To identify hepatocellular genes whose transcriptional regulation is tightly linked with IFN-γ- and IFN-α/β-mediated inhibition of HBV replication, Chisari et al[23] used DNA microarrays and a high-throughput cDNA differential display method (total gene expression analysis (TOGA) to examine the gene expression profiles of HBV transgenic mouse livers before and after the intrahepatic induction of IFN-α/β and IFN-γ. To further understand molecular mechanisms underlying HBV pathogenesis and identify novel cellular genes as new targets for anti-HBV drug development or for novel therapies, we used cDNA microarray filters dotted with 14 000 human genes to analyze transcriptional changes between an HBV DNA-transfected cell line (HepG₂ 2.2.15) and its parental cell line (HepG₂) pre- and post-IFN treatment. The HepG₂2.2.15 cell line, that has four copies of HBV genome integrated in the cellular chromosome and can persistently secrete HBV virions, was used for HBV infected liver-derived cells[17,18]. The parental hepatoblastoma cell line, HepG₂, from which HepG₂2.2.15 cell line was generated, was used as the uninfected cell control. In this study, the differentially expressed cellular genes detected between these two cell lines could be due to the integration of HBV DNA, the replication of HBV or the expression of HBV proteins. In chronic HBV infections, the virus usually persists in hepatocytes similar to the status of HBV in HepG₂2.2.15 cell line; the differentially expressed cellular genes in HepG₂2.2.15 cell line could reveal similar changes in the hepatocytes of HBV chronically infected patients.

Prior to interferon-α treatment, compared to HepG₂ cells, 15 known genes were more than three-fold down-regulated, while 7 genes were more than 3-fold up-regulated (Table 3) in HepG₂ 2.2.15 cells. To concentrate on the down-regulated genes, 4 genes belonging to the categories of cell cycle, ligands and receptors, phosphatase and transcription factors were down-regulated more than three-fold in HepG₂ 2.2.15 cells (Table 3). In a microarray study on HCMV-host cell interactions compared to non-infected human fibroblasts, down-regulated genes belonging to these categories were also reported in HCMV infected cells[10]. Though identical down-regulated genes from these categories were not found in the HCMV infected and HBV transfected cells, changes in the expression of cellular genes in these categories could reflect the common adverse effects of virus infections on host cells. In future studies, when more microarray analyses on virus-host interactions will be performed, it may be possible to identify significant changes of cellular gene expression due to HBV-hepatocyte interactions.

The effects of IFN-α on the global gene expression of HepG₂ and HepG₂2.2.15 cells revealed interesting differences (data available at http://www.mvlab-fudan.cn/microarray databases). In HepG₂ cells, compared with the expression of cellular genes at 0 h, no known gene was three-fold down-regulated 48 hr after IFN-α treatment. In contrast, 21 genes were up-regulated three-fold or more, and among these, five were IFN-induced genes, one belonging to the GTP-binding protein and signal transduction category, and one belonging to the phosphatase and phosphodiesterase category associated with the IFN-induced cascade. In HepG₂ 2.2.15 cells, after being treated with IFN-α for 48 h, eight genes remained down-regulated three-fold or more. In comparison to the known functional genes, one gene was predicted as encoding the regulatory G protein beta subunit, one was predicted as encoding an antigen of B cell differentiation, one encoding the thymosin beta 4Y isoform, one encoding a LDL-receptor related protein, which was also reported as encoding a human IFN-γ receptor, and one was an exon encoding a protein of the phosphatase and phosphodiesterase category, and the other 3 genes belonged to coagulation, mitochondrial and miscellaneous categories, respectively. These genes remained down-regulated and thus seemed refractory to the treatment of IFN-α. Since 2 of these genes are associated with thymosin and IFN-γ, microarray studies of HepG₂2.2.15 cells treated with IFN-γ and/or cytokines may be able to elucidate whether these defects in cellular functions could be reversed. In HepG₂2.2.15 cells, a number of IFN induced genes were up-regulated post IFN-α treatment; among these, genes p27 and 2’-5’A synthetase E were the top two up-regulated genes, being 6.1 and 9.1 respectively. Compared to the increases in these two genes up-regulated in HepG₂ (being 7.1 and 8.4 respectively), IFN-α was as effective in inducing these two genes in HBV infected cells as that in the uninfected hepatocytes.

To analyze whether there were HBV specific down-regulated cellular genes, which could be restored by IFN-α treatment, 23 matched genes from HepG₂ and HepG₂2.2.15 cell lines were compared post IFN-α treatment. As shown in Table 5, the three genes which were found down-regulated in HepG₂2.2.15 prior to IFN-α treatment (Table 3), namely described as encoding vascular endothelial growth factor, tyrosine phosphatase 1E, serine protease with IGF-binding motif were all up-regulated by IFN-α treatment. In addition, a gene described as coding for clathrin light chain, which belongs to the viral and clathrin receptor category was also up-regulated 5.2 times. However, even after IFN-α treatment, the down-regulated protein kinase zpk and protein tyrosine kinase genes could not be restored (Table 5), which suggested a possible defective link in the IFN-α induced cascade. Measures for restoring the functions of the kinases could help to improve the therapeutic efficacy of IFN-α. The clinical implications of non-restorable down-regulated protooncogenes in HepG₂ 2.2.15 cell lines merit further study.

Analysis of gene expression profiles in HBV-hepatocyte interaction by microarray is an interesting field. The information gathered in this study, though preliminary in scope, provides an important basis for further study. The differentially expressed genes detected in this study should be further evaluated by cell transfection or by in vivo studies and should provide new insight into the molecular mechanisms of HBV infection. In addition, there were approximately 200 unknown genes or ESTs being up- or down-regulated in HepG₂2.2.15 cells. Preliminary transiently co-transfected functional assay has already shown that some of cellular genes significantly regulated by HBV infection and IFN are associated with the inhibition of HBV gene replication and expression (data not shown). It is predicted that other IFN-inducible proteins could be discovered and new mechanisms of this cytokine may be revealed.

ACKNOWLEDGMENTS

This work was supported by the Chinese State Basic Science Foundation (No.1999054105) and Med-X Foundation of Fudan University. Xun Wang and Wei Xiong are Ph.D. students supported by the Chinese Ministry of Education.