pHBV1.2×, a plasmid which provides all HBV transcripts, was used to infect PNHHs as previously described[12]. This construct was similar to that as described by Guidotti et al[13]. Mammalian expression vector for NIK and NIK DN (aa 624-947) was provided by DV Goeddel (Turarik Inc.)[14], for TRAF2 and TRAF2 DN (aa 241-501) by Dr. SY Lee[15].

We transfected HepG2 cells with pHBV1.2× constructs for generation of HBV using Fugene 6 transfection reagent (Roche) as instructed by the manufacturer. After transfection, the cells were cultured for 5 d and harvested. HBV particles in the harvested media were cleared and concentrated through ultracentrifugation using PST55Ti rotor (Hitachi) for 1 h at 220 000 g with 3 mL cushion buffer containing 20 g/L sucrose, 50 mmol/L Tris-HCl pH 7.5, and 30 mmol/L NaCl. After ultracentrifugation, the pellet was resuspended with 1 × PBS. The resuspended viral solution was filtered with a 0.2 μm pore filter (Millipore). The titer of HBV solution was adjusted to 10⁹ virus genome equivalent (v.g.e.) per mL. PNHHs were infected with the above virus solution at about 100 v.g.e. per cell. Using this method, the efficiency of HBV infection to PNHHs was generally 50%[16].

Healthy parts of a liver from a patient who underwent hepatic resection for an intrahepatic stone at the Inje University Paik Hospital, Pusan, Korea was obtained and used as the source of hepatocytes. The removed tissue was immediately placed in Hank’s balanced salt solution (HBSS) and processed for cell culture. Isolation of hepatocytes was performed using a two-step collagenase perfusion technique[17,18]. The isolated hepatocytes were resuspended in a nutrient medium containing 90 mL/L Williams’ E and 10 mL/L Medium 199, supplemented with 10 μg/mL insulin, 5 μg/mL transferrin, 10^-7 mol/L sodium selenite, and 50 mL/L FBS.

A test was performed with the isolated DNA to determine whether the HBV- infected PNHHs formed cccDNA. Amplification with primers specific for both outside regions was performed with the isolated DNA because the region was specific for cccDNA rather than partially double-stranded DNA found in viral particles. The primers used are (5′-CTATGCTGGGTCTTCCAAATT-3′) which anneals to near the codon for amino acid 80 in human HBc open reading frame (ORF) and (5′-TTTCTGTGTAAACAATATCTG-3′) which anneals to near the codon for amino acid 680 in HBV Pol ORF. Therefore, if cccDNA was present in the isolated DNA, an amplified 1 kb product could be obtained. In this test, DNA isolated from mock-infected PNHHs was used as the negative control. A PCR test was performed with RNA to confirm whether HBV transcripts were produced. Reverse transcription amplification was performed with primers specific for the epsilon regions: (5′-CAACTTTTTTTCACCTCTGCCTA-3′) which anneals to DR1 and the reverse primer (5′-GATCTCGTACTGAAAGGAAAGA-3′). In addition, to detect HBV genome, real-time PCR was also performed as previously described[12].

RNA was isolated using TRIzol reagent (Life Technologies) according to the manufacturer’s instructions. With the total isolated RNA, reverse transcription was performed with Cy5-labeled dUTP in the experimental sample and Cy3-labeled dUTP in the control. Fifty micrograms of RNA, 1.5 ng oligo dT primer, and 1 ng control RNA containing lambda DNA sequences with a poly A sequence at the 3′ ends for reverse transcription, were mixed and volume of the mixture was adjusted to 20 μL. The mixture was incubated at 70°C for 5 min. After incubation, the mixture was quickly cooled on ice. With this whole reaction mixture, the labeling reaction was performed under the following conditions: 1 × reverse transcription buffer, 0.6 mmol/L Cy3- or Cy5-dUTP, 40 U of RNase inhibitor (Roche), 50 U of AMV-RT (Roche) and a dNTP mix containing 1 mmol/L dATP, 1 mmol/L dGTP, 1 mmol/L dCTP, and 0.4 mmol/L dTTP at 42°C for 1 h. After 1 h, 50 U of AMV-RT was added to the reaction mixture and the mixture was further incubated for 1 h for complete reverse transcription. Reverse transcription was stopped by the addition of 5 μL 0.5 mol/L EDTA. The synthesized cDNA was purified using a chromaspin column (Clonetech) as instructed by the manufacturer and precipitated with ethanol. Both Cy3- and Cy5- labeled cDNAs were resuspended with 100 μL hybridization buffer, containing 6 × standard saline citrate (SSC), 2 g/L sodium dodecyl sulfate (SDS), 5 × Denhardt solution, and 1 mg/mL salmon sperm DNA. The labeled cDNA was used for hybridization to the cDNA microarray chip at 62°C. The chip was arrayed using a GMS417 arrayer (Genetic MicroSystems Inc., Woburn, MA) with 4224 cDNAs and internal standards such as tubulin and actin and external standards such as lambda DNA. After 16-18 h of hybridization, the hybridized array was washed twice at 58°C for 30 min with washing bufferIcontaining 2 × SSC and 2 g/L SDS and washed once with washing buffer II containing 0.05 × SSC at room temperature for 5 min.

For quantification of the signals, the chips were scanned using an array scanner generation III (Molecular Dynamics) followed by image analysis using ImaGene ver. 3.0 software (BioDiscovery Ltd., Swansea, UK). The signal intensity of each spot was adjusted to obtain more accurate data by subtracting the background signals from the immediate surroundings. In this analysis, a difference in the ratio of more than two folds was considered significant.

HepG2, Huh7, and Chang liver cells were maintained in minimum essential media (Sigma) supplemented with 100 mL/L fetal bovine serum. For reverse transcription-polymerase chain reaction and luciferase reporter assay, cells were seeded in 12- well plates at a density of 0.2 × 10⁶ cells per well and transfected on the following day with the aperopriate DNA and fugene 6 (Roche) as described by the manufacturer. To normalize the total DNA, pUC119 and backbone DNA of pHBV1.2× were used. The transfection efficiency for HepG2 with fugene 6 was usually 10%-20%.

Cells were transfected with pUC119 and pHBV1.2×. After 48 h of transfection, total RNA was extracted with TRIzol reagent (Life Technologies) as described by the manufacturer. cDNA was produced by reverse transcription using the same procedure as cDNA microarray analysis. Following reverse transcription, the synthesized cDNA was amplified with 2.5 U Hot start Taq polymerase (Takara), GAPDH specific primer set, and appropriate primer set. The sequences of the primer set are as follows: TRAF2 specific forward(5′-AGGGGACCCTGAAAGAATAC-3′), TRAF2 specific reverse(5′-CAGGGCTTCAATCTTGTCTT-3′), NIK specific forward(5′-TACCTCCACTCACGAAGGAT-3′), NIK specific reverse(5′-CAAGGGAGGAGACTTGTTTG-3′), LT-α specific forward(5′-AGCTATCCACCCACACAGAT-3′),

LT-α specific reverse(5′-GTTTATTGGGCTTCATCGAG-3′), GAPDH specific forward(5′-ATCATCCCTGCCTCTACTGG-3′), and GAPDH specific reverse (5′-TGGGTGTCGCTGTTGAAGTC-3′). PCR amplification was performed using Gene mp PCR system 2400 (Perkin-Elmer) with 5 min initial denaturation at 95°C and 35 cycles of 50 s at 95°C, 50 s at 60°C, and 50 s at 72°C, followed by 7 min of extension at 72°C. To separate the PCR fragments, 15% agarose gel was used.

To confirm NIK and TRAF2 protein expression, we performed immunoblot analysis with anti-NIK rabbit polyclonal antibody (Santa Cruz) and anti-TRAF2 mouse monoclonal antibody (Santa Cruz). HepG2 cells were seeded in 6- well plates at a density of 0.4 × 10⁶ cells per well. Cells were lysed with RIPA buffer containing 25 mmol/L Tris-HCl (pH 7.4), 150 mmol/L KCl, 5 mmol/L EDTA, 10 mL/L Nonidet P-40 (NP-40), 5 g/L sodium deoxycholate, and 1 g/L SDS and centrifuged at 12 000 g for 10 min. The supernatants were separated by 12% SDS-PAGE protein gel for immunoblot analysis.

After seeded on 12-well plates, HepG2 cells were co-transfected with appropriate DNA (Figure 5), 0.1 μg pNF-κB-luciferase and 0.1 μg pCMV-β-galactosidase. After 48 h of transfection, cell extracts were prepared and luciferase reporter assay was performed using a luciferase assay system (Promega) as described by the manufacturer. The transfection efficiency was normalized by its galactosidase activity. The assay was triplicated and repeated at least twice.

Open in New Tab   Full Size Figure   Download Figure

Figure 5 Immunoblot assay of TRAF2 and NIK. pUC119 is a backbone DNA about pHBV1.2x. In the HepG2 cells transfected with pHBV1.2x, the protein levels of NIK (A) and TRAF2 (B) were increased. β-actin was used to normalize total protein level.

To confirm HBV infection to PNHHs, DNA was isolated during the RNA purification step with TRIzol reagent as instructed by the manufacturer. Infection was confirmed through PCR-based amplification specific only for cccDNA in nuclei of the infected cells. Figure 1 shows that the amplified product appeared in HBV- infected cells, indicating that HBV did infect PNHHs and that the nucleocapsid was transported into nuclei of the infected hepatocytes (Figure 1A). RT-PCR analysis of the HBV transcripts amplified with primers specific for epsilon and polymerase regions as described in Materials and Methods confirmed the presence of HBV RNA transcripts in the infected cells (Figure 1B). Real time PCR showed that HBV was not detected until 6 d after infection (Figure 1C).

Open in New Tab   Full Size Figure   Download Figure

Figure 1 Confirmation of HBV infection for PNHHs using PCR for cccDNA formation and RT-PCR for production of transcripts. From PNHHs infected with HBV, RNA and DNA were extracted using TRIzol reagent as described in Materials and Methods. As a negative control, mock-infected PNHHs were used. A: With the extracted DNA, cccDNA was confirmed by a PCR analysis as described in Materials and Methods; B: With the extracted RNA, a transcript of HBV was confirmed by RT-PCR analysis as described in Materials and Methods; C: With the extracted RNA, a tanscript of HBV was confirmed by real-time PCR as previously described[19].

The experiments were carried out in triplicate at the infection step for more certain identification of genes differentially expressed by HBV infection. From each of the HBV infected cells for over 8 d, RNA was isolated and analyzed by microarray. As a result, three sets of cDNA array images were obtained. We analyzed the intensity of the raw image through scatterplot analyses. Figure 2A shows scatterplot analyses of log (Cy3 signal × Cy5 signal) vs log2 (Cy3 signal/Cy5 signal). This showed that each plot tended to divert from the general small curve (Figure 2A). But, each scatterplot analysis of Log (Cy3 signal/Cy5 signal) vs Log (Cy5 signal) showed a curve closer to the exponential decay (Figure 2B). Therefore, the data were fitted to an exponential decay curve for Cy3 per Cy5 channel correction (Figure 2C). Through these steps, we obtained a higher confidence ratio of the Cy3 signal compared to the Cy5 signal for each chip. With the ratios obtained, we analyzed the correlation coefficient between the data of the three chips. The correlation coefficient turned out to be more than 0.7 (Figure 3A), suggesting that the relationship between each chip was significant. The correlation coefficient for genes that were differentially expressed more than two folds was more than 0.95 (Figure 3B). Selected genes that were differentially expressed more than two folds, showed a high reproducibility among the triplicate microarray analyses.

Open in New Tab   Full Size Figure   Download Figure

Figure 2 Scatter plot analysis. For normalization of the Cy3 (3D) and Cy5 channel signal (5D) channels, data obtained from the raw image scanning were plotted in a scatter plot using Excel software (Microsoft). A: The X-axis represents Log₂ (3D/5D) and the Y-axis Log₂ (3D/5D); B: The X-axis represents Log (5D) and the Y-axis Log (3D/5D); C: The X-axis represents Log (5D) and the Y-axis Log (3D/5D) F, in which “F” is the function for normalization. The bottom panel shows data with signals fitted to an exponential decay curve.

Open in New Tab   Full Size Figure   Download Figure

Figure 3 Correlation between three sets of PNHHs infected for eight days. A: With the reliable signals in obtained signals, the correlation efficient was calculated between each experiment; B: In addition, another correlation efficient was also calculated with only the selected genes, which were differentially expressed more than 2 folds.

Through a microarray analysis of PNHHs infected with HBV, we obtained the profile of 45 genes that were down regulated more than two folds compared to the control. The 45 down-regulated genes were analyzed classified by function (Table 1). Table 1 shows that many transcription factors related to RNA polymerase II, were down-regulated by HBV infection. In contrast, transcription factors such as C/EBP, which is used for transcription of HBV genes[19,20], were not differentially expressed. That is, the C/EBP expression level was changed less than two folds.

From the analysis by cDNA microarray, 53 up-regulated genes were identified by an increase of more than two folds in their differential expression. Table 2 shows that growth- and tumor-related molecules comprised a proportion of the up-regulated genes. The positive effector genes for tumor and proliferation have found to be GDF11[21] and NOL1[22] and the negative effector genes EXTL3[23] and RAD50. The most interesting genes have found to be the TNF signaling pathway- related genes. LT-α is an inflammatory cytokine and induces the TNF signaling pathway as a ligand for TNF receptor (TNFR). LT-α binds to TNFR and recruits TRAF2. MAP3K14 (NF-κB inducing kinase, NIK) binds to TRAF2 and activates NF-κB[24]. LT-α, TRAF2, and NIK were also up-regulated in the experiment (Table 2). This means that HBV activates NF-κB through up-regulation of LT-α, TRAF2, and NIK.

According to the cDNA microarray data, three genes related to the TNF signaling pathway, LT-α, TRAF2, and NIK, were up-regulated. Upregulation of these genes was confirmed by RT-PCR. For RT-PCR analysis, primer sets specific to LT-α, TRAF2, and NIK, were used and experiments were performed in hepatoma-derived cell lines, including HepG2, Huh7, and Chang liver cells. As a result, the mRNA levels of these three genes in each cell line were increased by pHBV1.2× transfection (Figure 4). In addition to RT-PCR, the expression of NIK and TRAF2 was confirmed at the protein level (Figure 5). The expression of LT-α, was confirmed by immunofluorescence staining analysis (data not shown).

Open in New Tab   Full Size Figure   Download Figure

Figure 4 RT-PCR analysis of selected genes. Mock means untransfected cells. pUC119 is backbone of pHBV1.2x, so pUC119 transfected cells are the negative control for pHBV1.2x transfected cell. In the hepatoma-derived cell lines, HepG2, Huh7, and Chang liver cells, TRAF2, NIK, and LT-α mRNA level in pHBV1.2x transfected cells were up-regulated rather than mock and pUC119 transfected cells.

According to cDNA microarray and RT-PCR analysis, mRNA expression of LT-α, TRAF2, and NIK was up-regulated by HBV. Since their expression was related to NF-κB activation. HBV-mediated NF-κB activation might be involved in the up regulation of these genes. To determine whether these genes actually are involved in HBV-mediated NF-κB activation, we performed a luciferase assay with a pNF-κB-luciferase vector as a reporter plasmid. To elucidate whether HBV-mediated NF-κB activation is dependent on TRAF2 and NIK of three genes, we cotransfected pTRAF2 DN or pNIK DN, the dominant negative form of pTRAF2 or pNIK, with pNF-κB-luciferase, pCMV β-galactosidase, and pHBV1.2× (Figure 6). The experiment for LT-α was performed with anti- LT-α, to neutralize LT-α. The pHBV1.2× produced about a 4.2 folds greater increase in NF-κB luciferase activity than pUC119. However, pHBV1.2× cotransfection with TRAF2 DN or NIK DN produced about a 3.5 or 1.2 fold relative increase (Figure 6) and treatment with anti- LT-α decreased the ratio to less than 4.2 folds (data not shown). These findings led us to think that HBV could activate NF-κB and that this HBV-mediated NF-κB activation might require LT-α, TRAF2, and NIK. DC.

Open in New Tab   Full Size Figure   Download Figure

Figure 6 HBV-mediated NF-κB activation through TRAF2 and NIK. In pHBV1.2x transfected HepG2 cells, NF-κB activity was increased more than pUC119 transfected cells. But in cotransfected cells with pHBV1.2x and TRAF2 DN or NIK DN, NF-κB activity was decreased less than pHBV1.2x transfected cells.