p38 Mitogen-Activated Protein Kinase-Dependent Hyperinduction of Tumor Necrosis Factor Alpha Expression in Response to Avian Influenza Virus H5N1

Reagents and antibodies.

Specific antibodies against phosphorylated and unphosphorylated forms of p38 MAPK (catalog no. 9211 and 9212), ERK1/2 (catalog no. 9101 and 9102), and JNK (catalog no. 9255 and 9252) were purchased from Cell Signaling Technology, New England Biolabs, Beverly, MA, for Western blot analysis. The specificity of the antibodies had been confirmed by previous reports (20). Anti-IκB-α (sc-203) and anti-actin (sc1616) antibodies were purchased from Santa Cruz Biotechnology, Santa Cruz, CA. Horseradish peroxidase-conjugated secondary antibodies used were anti-rabbit immunoglobulin G (IgG) and anti-mouse IgG (Transduction Lab, BD Biosciences Pharmingen, San Diego, CA) and anti-goat IgG (DAKO). Kinase inhibitors for p38 MAPK (SB203580), ERK1/2 (PD98059), and JNK (SP600159) were purchased from Calbiochem, La Jolla, CA.

Cells and viruses.

Blood mononuclear cells from healthy donors (Hong Kong Red Cross Blood Transfusion Service) were separated by Ficoll-Paque centrifugation and purified by the adherence method as previously described (3). The purity of harvested monocytes was determined by staining with fluorescein isothiocyanate-conjugated anti-CD14 antibody (Beckman Coulter, Fullerton, CA) and analyzed by using flow cytometry. Monocytes were seeded in 24-well (0.5 × 10⁶ cells/well) or 6-well (2 × 10⁶ cells/well) tissue culture plates for studies of RNA and protein, respectively. Monocytes were differentiated in RPMI 1640 (Gibco, Invitrogen, Grand Island, NY) supplemented with 5% heat-inactivated autologous plasma. Differentiated macrophages were obtained after 14 days of culture as previously described (3). Influenza viruses A/HK/483/97 (H5N1), isolated from a patient with H5N1 infection in 1997, and A/HK/54/98 (H1N1) were used for the experiments. The viruses were grown in MDCK cells and purified by preadsorption to and elution from turkey red blood cells (3). Virus infectivity was determined by titration on MDCK cells.

Virus infection.

The procedure of influenza virus infection was described in our previous study (3). Macrophages were infected with virus at a multiplicity of infection of 2 for 30 min at 37°C. The supernatant containing the virus inoculum was then removed, and the cells were incubated in macrophage serum-free medium (Gibco, Invitrogen, Grand Island, NY) supplemented with 0.6 μg/ml penicillin, 60 μg/ml streptomycin, and 2 μg/ml l-1-tosylamide-2-phenylethyl chloromethyl ketone-treated trypsin (Sigma, St. Louis, MO). The mock-treated control was incubated with the buffer under parallel conditions.

Real-time reverse transcription (RT)-PCR analysis of mRNA in virus-infected macrophages.

Total RNA was DNase treated and then reverse transcribed by using a TaqMan reverse transcription reagent kit (Applied Biosystems, Foster City, CA) according to the manufacturer's protocol. Analysis of real-time RT-PCR was performed with reference to the comparative C_T, cycle number to threshold, as described in user bulletin no. 2 of the ABI Prism 7700 Sequence Detection System. Briefly, the levels of TNF-α mRNA and 18S rRNA (18S), the reference gene, were assayed by TaqMan gene-specific ASSAYS-ON-DEMAND reagent kits (Applied Biosystems, Foster City, CA). The C_T of TNF-α was normalized to that of 18S in each sample (i.e., ΔC_T). The mRNA level of virus-infected cells was relative to the uninfected cells (i.e., ΔΔC_T), and the amount was given by 2^−ΔΔCT. The quantity of the mRNA was statistically analyzed by two-tailed, paired t test.

The mRNA levels of H5N1 matrix and β-actin genes were assayed by SYBR Green PCR reagent kit according to the manufacturer's protocol. (Applied Biosystems, Foster City, CA). In brief, the cDNA was amplified with gene-specific primers and their sequences were as follows: H5N1 matrix protein forward, 5′-TTACTCAACTGGTGCGCTTG-3′; H5N1 matrix protein reverse, 5′-GTTGGTGGTAGTCGCCATCT-3′; human β-actin forward, 5′-TCACCCACACTGTGCCCATCT-3′; human β-actin reverse, 5′-GAACCGCTCATTGCCAATGG-3′. At the end of the PCR, dissociation curve analysis was performed to examine the specificity of the PCR product. The SYBR Green PCR data were statistically analyzed by the C_T method as described above. The C_T of the matrix gene was normalized to that of β-actin in each sample. The mRNA levels of the matrix gene in H5N1-infected cells with kinase inhibitors were relative to that of virus-infected cells without the inhibitors.

Preparation of cytoplasmic and nuclear proteins.

The nuclear extract preparation method used was a modified form of that of Schreiber et al. (29). In brief, macrophages (2 × 10⁶ cells/well) were harvested at different times after infection and washed with phosphate-buffered saline. The cells were treated with cold buffer A (10 mM HEPES [pH 7.9], 10 mM KCl, 0.1 mM EDTA, 0.1 mM EGTA, 1 mM dithiothreitol [DTT], protease inhibitor cocktail) on ice for 15 min. Then, NP-40 was added to a final concentration of 0.6% to the cell lysate and removed with a cell scraper. The cytoplasmic portion of the lysate was harvested by centrifugation for 1 min at 4°C. The nuclear pellet was resuspended in 50 μl of ice-cold buffer C (20 mM HEPES [pH 7.9], 0.4 M NaCl, 1 mM EDTA, 1 mM EGTA, 1 mM DTT, protease inhibitor cocktail), and the nuclei were disrupted on ice with agitation for 15 min. The nuclear protein lysate was harvested by centrifugation for 5 min at 4°C, and all protein samples were frozen in aliquots at −70°C. The harvested proteins were quantified with a Coomassie blue staining kit (Pierce, Rockford, IL).

Electrophoretic mobility shift assay (EMSA).

Nuclear protein (2 μg) was incubated with 20 fmol of [γ-³²P]ATP-labeled, double-stranded κB enhancer oligodeoxynucleotides, with sequences obtained from the region of the human immunodeficiency virus type 1 long terminal repeat (5′-GATCAGGGACTTTCCGCTGGGGACTTTCC-3′ and 5′-GGAAAGTCCCCAGCGGAAAGTCCCTGATC-3′) in binding buffer (25 mM HEPES [pH 7.9], 0.5 mM EDTA, 0.5 mM DTT, 1% NP-40, 12% glycerol, 50 mM NaCl) and 0.5 μg of poly(dI-dC) for 30 min at 37°C. A pair of oligodeoxynucleotides encoding the mutated κB binding sites (5′-GATCACTCACTTTCCGCTGCTCACTTTCC-3′ and 5′-GGAAAGTGAGCAGCGGAAAGTGAGTGATC-3′) was included in parallel experiments to test for specific binding of the DNA-protein complexes. After incubation, the reaction complex was separated from the free labeled probe in a 5% nondenaturing polyacrylamide gel using Tris-glycine buffer (50 mM Tris, 200 mM glycine, 1 mM EDTA, pH 8.5). The gel was dried for 2 h in a gel dryer (Bio-Rad) and exposed to Fuji X-ray film.

Western blot analysis.

For Western blot analysis, 30 μg of cytoplasmic protein was heat denatured in sample buffer (125 mM Tris [pH 6.8], 4% sodium dodecyl sulfate, 20% glycerol, 5% β-mercaptoethanol, 0.01% bromophenol blue), separated by 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis, and transferred to a nitrocellulose membrane for assaying phosphorylated kinases by using ECL solution (Amersham Biosciences, Buckinghamshire, United Kingdom). Blots were stripped and reprobed using different antibodies against the respective nonphosphorylated kinase to determine the total amount of the indicated kinase.

Quantitative analysis of TNF-α by enzyme-linked immunosorbent assay (ELISA).

The supernatant samples of macrophage cultures were collected at 6 h after infection and irradiated with UV light before the levels of TNF-α were measured by a specific TNF-α assay kit (R&D Systems, Minneapolis, MN) as previously described (3).

H5N1 and H1N1 influenza viruses induce NF-κB activation in human macrophages.

To investigate whether NF-κB is involved in H5N1-induced cytokine expression, we first measured the level of NF-κB activation by EMSA in the nuclear extract of H5N1- and H1N1-infected cells at 0.5 and 2 h after infection. Over the time course, there was a strong activation of NF-κB to comparable levels by both H5N1 and H1N1 viruses (Fig. 1A, lanes 2 to 5). This result was reproduced in macrophages from three healthy donors, suggesting that the finding was not donor specific (data not shown). The specificity of NF-κB binding was shown by incubating the nuclear extracts with 5 pmol of unlabeled κB oligodeoxynucleotide probe or with 20 fmol of the radioactively labeled, mutated κB probe (Fig. 1A, lanes 6 and 7). Our results indicate that H5N1 did not induce higher levels of NF-κB activation in primary blood macrophages compared to H1N1 virus.

To further investigate the activation of the NF-κB signaling pathway in cytokine dysregulation, we examined the IκB-α degradation in the cytoplasmic portion of virus-infected cells by Western blot analysis. The protein levels of IκB-α in the H5N1- or H1N1-infected cells at 0 h were the same as in mock-treated cells (Fig. 1B, lane 1). IκB-α expression was reduced by 15 to 20% at 0.5 and 2 h after infection in both H5N1- and H1N1-infected cells (Fig. 1B, top, lanes 2 to 5). Similar results were found in three different donors (data not shown). Equal loading of protein samples was shown by reprobing the blot with an anti-actin antibody (Fig. 1B, bottom). Together with the EMSA, our results clearly demonstrated that H5N1 and H1N1 induce similar levels of NF-κB activation and IκB-α degradation.

H5N1 activates p38 MAPK and ERK1/2 phosphorylation but not JNK in human macrophages.

To investigate the role of MAPK in mediating H5N1-induced cytokine expression, we next measured the phosphorylation levels of p38 MAPK, ERK1/2, and JNK in the cytoplasmic protein lysate of virus-infected cells by Western analysis at different times.

Phosphorylation of p38 MAPK was rapidly induced by the H5N1 virus at 1 h postinfection, compared with the H1N1 counterpart, in human macrophages (Fig. 2A). The maximal level of p38 MAPK activation in macrophages was reached at 2 h after H5N1 infection, and the level of activation was sustained at 4 h postinfection (Fig. 2A, lanes 1 to 4). There was a 7.6-fold induction of p38 MAPK phosphorylation in macrophages infected with H5N1 compared with mock-infected cells at 2 h postinfection (Fig. 2A, lanes 3 and 9). In contrast, there was no significant activation of p38 MAPK in H1N1-infected cells throughout the time course (Fig. 2A, lanes 5 to 8). In addition, the activation of ERK1/2 in H5N1-infected macrophages was higher than that in H1N1-infected cells at 2 and 4 h after infection (Fig. 2B, lanes 3 to 4 and 7 to 8). In contrast, there was no significant change in the levels of phosphorylation of JNK in both H5N1- and H1N1-infected cells (Fig. 2C, lanes 1 to 4 and 5 to 8). These experiments were performed in duplicate with macrophages obtained from five different blood donors with similar results. Thus, our results showed that H5N1 significantly induces the phosphorylation of p38 MAPK and moderately activates ERK1/2 but not JNK in primary human macrophages.

Inhibition of MAPK suppresses H5N1 induction of TNF-α in human macrophages.

In light of the H5N1-induced activation of p38 MAPK and ERK1/2, we examined whether these kinase activities are required for cytokine dysregulation by real-time RT-PCR. We first determined the dose-dependent effects of MAPK kinase inhibitors on H5N1-induced TNF-α at 3 h postinfection, i.e., at the peak of TNF-α induction by H5N1, as previously described (3). Our assay system for TNF-α mRNA was validated by the linear relationship of the ΔC_T of TNF-α and 18S to the log input of RNA (Fig. 3A). The slope of the validation plot was 0.0094 (i.e., <0.1), indicating equal amplification efficiencies of these two genes in our detection system.

Primary blood macrophages were mock treated or treated with increasing concentrations of pharmacological inhibitors for p38 MAPK (SB203580) or ERK1/2 (PD98059) for 1 h before virus infection. We did not find any cellular changes including morphology and viability at the indicated concentrations at the time of harvesting. Total RNA was obtained for RT-PCR and analyzed by TaqMan TNF-α assay kit (Fig. 3B). At 3 h postinfection, the H5N1-induced TNF-α mRNA expression decreased significantly with increasing concentrations of SB203580. Specifically, the TNF-α mRNA level in cells treated with 2.5 μM SB203580 was suppressed by 41% (P < 0.05), compared to the H5N1-infected cells without treatment with the kinase inhibitor. In contrast, PD98059 did not show any dose-dependent effect on the H5N1-induced TNF-α level. For example, the TNF-α mRNA level in cells treated with 2.5 μM PD98059 was not significantly suppressed by PD98059 (P > 0.05). The real-time RT-PCR results were reproduced in five separate macrophage cultures obtained from different donors.

We also performed Western analysis to examine the inhibitory effects of kinase inhibitors on the activities of p38 MAPK or ERK1/2 after H5N1 infection. The macrophages were left untreated or treated with SB203580 or PD98059 at 2.5 μM for 1 h. The cytoplasmic proteins of H5N1-infected cells were collected at 1 h after infection. The phosphorylation levels of p38 MAPK or ERK1/2 in macrophages infected with H5N1 were reduced significantly to the basal levels, compared to cells with mock treatment (Fig. 3C and D, lanes 1 and 3).

p38 MAPK regulates TNF-α production in H5N1-infected cells.

To further examine whether activation of p38 MAPK or ERK1/2 in H5N1 infection results in TNF-α production, we examined the quantities of TNF-α protein in the supernatants of kinase inhibitor-treated cells at 6 h after infection. Consistent with the real-time RT-PCR results, SB203580 reduced the TNF-α production by 32% (P < 0.05) whereas PD98059 did not have significant suppressive effects, as shown in Fig. 4. Taken together, our RT-PCR and ELISA results strongly indicated that p38 MAPK plays a statistically significant role in H5N1-induced cytokine dysregulation.

Infectivity of H5N1 is not affected by MAPK inhibitors.

To study the effects of the p38 MAPK and ERK1/2 inhibitors on viral infectivity, we measured the levels of the H5N1 matrix mRNA in virus-infected cells with or without kinase inhibitor treatment at 3 h postinfection by real-time RT-PCR assay. We did not find any significant change in the levels of matrix mRNA in cells with or without inhibitor treatment (P > 0.5; Fig. 5). The specificity of the PCR product was confirmed by using the dissociation curve software and agarose gel electrophoresis (data not shown). The results were derived from triplicate macrophage cultures, and each set of macrophages was obtained from three separate donors.