Specific Imbalanced Inflammatory Response Human Endothelial Cells ...

5 downloads 0 Views 811KB Size Report
whelming proinflammatory but not antiviral response, whereas NFATC4 was found to regulate transcription of specifically H5N1- induced genes. We describe for ...
The Journal of Immunology

H5N1 Virus Activates Signaling Pathways in Human Endothelial Cells Resulting in a Specific Imbalanced Inflammatory Response Dorothee Viemann,*,†,‡ Mirco Schmolke,x Aloys Lueken,*,‡ Yvonne Boergeling,x Judith Friesenhagen,* Helmut Wittkowski,*,† Stephan Ludwig,‡,x and Johannes Roth*,‡ H5N1 influenza virus infections in humans cause a characteristic systemic inflammatory response syndrome; however, the molecular mechanisms are largely unknown. Endothelial cells (ECs) play a pivotal role in hyperdynamic septic diseases. To unravel specific signaling networks activated by H5N1 we used a genome-wide comparative systems biology approach analyzing gene expression in human ECs infected with three different human and avian influenza strains of high and low pathogenicity. Blocking of specific signaling pathways revealed that H5N1 induces an exceptionally NF-kB–dependent gene response in human endothelia. Additionally, the IFN-driven antiviral program in ECs is shown to be dependent on IFN regulatory factor 3 but significantly impaired upon H5N1 infection compared with low pathogenic influenza virus. As additional modulators of this H5N1-specific imbalanced gene response pattern, we identified HMGA1 as a novel transcription factor specifically responsible for the overwhelming proinflammatory but not antiviral response, whereas NFATC4 was found to regulate transcription of specifically H5N1induced genes. We describe for the first time, to our knowledge, defined signaling patterns specifically activated by H5N1, which, in contrast to low pathogenic influenza viruses, are responsible for an imbalance of an overwhelming proinflammatory and impaired antiviral gene program. The Journal of Immunology, 2011, 186: 164–173.

I

nfluenza A viruses still pose a major threat due to their pandemic potential. There is a huge natural virus reservoir in birds, which provides a pool of viral genes that contribute to the generation of novel pandemic virus strains. The highly pathogenic avian H5N1 virus is the first example of an avian virus reported to have infected and killed several hundreds of humans (1). Human-adapted influenza A viruses cause primarily infections of the upper respiratory tract with limited virus replication. Highly pathogenic avian influenza viruses (HPAIVs) rather cause systemic infections with hemorrhagic sepsis in poultry. In humans, clinical manifestations of H5N1 infections are a systemic inflammatory response syndrome (SIRS) leading to multiorgan failure (2–6). There are two characteristics of H5N1 infections: an unusual strong cytokine response (7, 8) and a neuro- as well as endothelial *Institute of Immunology, ‡Interdisciplinary Center of Clinical Research, and xInstitute of Molecular Virology, University of Muenster; and †Department of Pediatrics, University Hospital of Muenster, Muenster, Germany Received for publication December 28, 2009. Accepted for publication October 29, 2010. This work was supported by the Bundesministerium fu¨r Bildung, Wissenschaft, Forschung und Technologie Zoonosis Network FluResearchNet (Grant 01KI07130). D.V. and J.R. were supported by grants from the Interdisciplinary Center of Clinical Research University of Muenster. Address correspondence and reprint requests to Dr. Dorothee Viemann, Institute of Immunology, University of Muenster, Roentgenstr. 21, D-48149 Muenster, Germany. E-mail address: [email protected] The online version of this article contains supplemental material. Abbreviations used in this paper: BS, binding site; EC, endothelial cell; FC, fold change; FPV, A/FPV/Bratislava/79 (H7N7) fowl plague virus; HPAIV, highly pathogenic avian influenza virus; IKK2kd, kinase-dead IkB kinase 2; IRF, IFN regulatory factor; ISRE, IFN-stimulated response element; MDCK, Madin-Darby canine kidney; MOI, multiplicity of infection; PCA, principal component analysis; p.i., postinfection; PR8, A/PR8/34 (H1N1); qRT-PCR, quantitative real-time RT-PCR; siRNA, short interfering RNA; SIRS, systemic inflammatory response syndrome; wt, wild-type. Copyright Ó 2010 by The American Association of Immunologists, Inc. 0022-1767/10/$16.00 www.jimmunol.org/cgi/doi/10.4049/jimmunol.0904170

cell (EC) tropism (9–13). Recently, a-2,3-linked sialic acid receptors preferentially bound by avian influenza virus were shown to be present on human ECs, suggesting a crucial involvement of ECs in the characteristics of systemic H5N1 diseases (14). Isolated suppression or knockdown of the cytokine response did not protect against lethal outcomes (13, 15, 16). In human bronchial epithelial cells, high pathogenic H5N1 variants replicate stronger than low pathogenic variants or the human H3N2 virus but elicit a much weaker and delayed IFN response (17). The overall pattern of proinflammatory and antiviral signaling pathways activated by H5N1 in specific tissues might be the clue to the pathogenicity of HPAIVs. The cardinal problem of treating H5N1 infections with neuraminidase inhibitors is that it has to be initiated very early, otherwise virus replication and dissemination has already progressed too far (6). Evaluating the role of endothelial response programs with respect to proinflammatory and antiviral strategies and delineating the responsible signaling cascades will allow the development of novel therapeutic strategies. To approach to these objectives, we performed comparative global gene expression studies of ECs infected with HPAIV of the H7N7 and H5N1 subtype and a low pathogenic human influenza virus of the H1N1 subtype to identify H5N1 specific gene response subpatterns. In line with recent findings of our group, we confirm a remarkable NF-kB dependency of the H5N1-induced gene expression program (18), which we demonstrate to be much weaker or not relevant for infections with the other influenza strains. Concurrently, infections with H5N1 were characterized by a striking attenuation of the IFN-b–dependent antiviral response. Furthermore, we reveal HMGA1 as a novel transcription factor to be specifically responsible for the overwhelming proinflammatory but not antiviral response in H5N1 infections, whereas NFATC4 was found to be a transcriptional regulator of specifically H5N1induced genes. We describe for the first time, to our knowledge,

The Journal of Immunology

FIGURE 1. Infectibility of HUVECs with different strains of influenza viruses. Lysates of uninfected HUVECs (ctrl) and PR8-, FPV-, and H5N1infected HUVECs were immunoblotted 3, 5, and 8 h p.i. for viral proteins PB1 (86 kDa), M1 (27 kDa), and NS1 (26 kDa). Immunostaining of ERK1/ 2 served as protein loading control.

defined signaling patterns specifically activated by H5N1, which, in contrast to low pathogenic influenza viruses, are responsible for an imbalance of an overwhelming proinflammatory and impaired antiviral gene program.

Materials and Methods Viruses and cells The human H5N1 influenza A virus isolate A/Thailand/KAN-1/2004 was used with kind permission of Pilaipan Puthavathana, Bangkok, Thailand. The avian influenza virus A/FPV/79/Bratislava (H7N7, fowl plague virus [FPV]) and the human influenza virus strain A/Puerto-Rico/8/34 (Giessen variant) were obtained from the Institute of Virology in Giessen, Germany. Viruses were propagated on Madin-Darby canine kidney (MDCK) II cells.

165 MDCKII were cultured in MEM (PAA Laboratories, Pasching, Austria) containing 10% v/v FCS and 100 U/ml penicillin/0.1 mg/ml streptomycin (13 penicillin/streptomycin) (Life Technologies, Carlsbad, CA). A/PR8/ 34 (H1N1) (PR8) was cultivated in 11-d-old embryonic chicken as described previuosly (19). Allantoic liquid containing viral particles was harvested 50 h postinfection (p.i.). HUVECs were obtained from Cambrex (Charles City, IA) and cultured as described elsewhere (20, 21). HUVECs between passages 5 and 7 were infected with a multiplicity of infection (MOI) of 5 of each viral strain. For the plaque assay, MDCK cells grown 100% confluent in six-well dishes were washed with PBS and infected with serial dilutions of culture supernatants in PBS containing 0.2% bovine serum albumin, 1 mM MgCl2, 0.9 mM CaCl2, 100 U/ml penicillin, and 0.1 mg/ml streptomycin for 30 min at 37˚C. The inoculum was aspirated, and cells were incubated with 2 ml MEM medium containing 0.2% BSA and antibiotics supplemented with 0.6% agar (Oxoid, Cambridge, U.K.), 0.3% DEAE-dextran (GE Healthcare Bio-Sciences AB, Uppsala, Sweden), and 1.5% NaHCO3 at 37˚C, 5% CO2 for 2 to 3 d. Virus plaques were visualized by staining with neutral-red or Coomassie blue (0.1% Coomassie Brilliant blue G-250 in 40% methanol and 10% acetic acid). The FNX amphotropic retrovirus producer cells were kindly provided by G. Nolan, Stanford, CA, and cultured as described elsewhere (21). Retroviral infection of HUVECs with the pCFG5-IEGZ vector containing an insert of kinase-dead IkB kinase 2 (IKK2kd) was performed as described (18, 21).

Immunofluorescence microscopy To determine the infectious doses resulting in an 80% infection rate, HUVECs were infected with different MOIs of each virus strain and fixed with 3.7% formaldehyde. They were stained against an influenza A group-specific nucleoprotein (mouse mAB OBT0053, clone 1A52.9; Serotec, Dusseldorf, Germany) and counterstained with DAPI to determine the proportion of nucleoproteinpositive HUVECs. Goat anti-mouse Cy5-labeled secondary Ab was pur-

FIGURE 2. Differential gene profiles induced in HUVECs p.i. with different influenza strains. A, PCA comparing gene profiles of uninfected HUVECs (ctrl; black vector cloud) and PR8- (light gray), FPV- (medium gray), and H5N1-infected HUVECs (dark gray) 5 h p.i. Vector clouds representing a gene profile of one experiment are positioned in a three-dimensional vector space according to variance to each other. B, Venn diagram indicates the number and overlap of genes induced 5 h p.i. by H5N1, FPV, and PR8. Data are based on microarray analyses of three independent experiments. Plotted are functional gene groups according to gene ontology annotations overrepresented in C, the intersection group of genes induced by PR8, FPV, and H5N1, or in D, the group of genes only upregulated by H5N1. Statistical significance (y-axis) was determined by applying the Fisher’s exact test.

166 chased from Dianova (Hamburg, Germany).The fluorescence was detected using the Axioskop microscope from Zeiss (Go¨ttingen, Germany).

Western blot For immunoblotting, cells were lysed in RIPA buffer and separated by SDSPAGE as described (22). Western blot staining was performed with anti-

H5N1 TRANSFORMS IMMUNE RESPONSES IN ENDOTHELIA ERK2 (rabbit IgG, C-14; Santa Cruz Biotechnology, Heidelberg, Germany), anti-M1 (GA2B; AbD Serotec, Oxford, U.K.), anti-IFN regulatory factor (IRF) 3 (rabbit p-Ab; Zymed, San Francisco, CA), anti-IRF7 (F-1), anti-HMGA1 (FL-95), anti-NFATC4 (H-74), and anti-PB1 Ab (goat pAB, vK20) (all from Santa Cruz Biotechnology). A murine mAb against influenza A virus NS1 was generated by V. Wixler (Institute of Molecular Virology, Muenster, Germany).

FIGURE 3. Kinetic course of gene expression in HUVEC after HPAIV infection. A–H, HUVEC monolayers were infected with PR8 (rhombus), FPV (squares), and H5N1 (triangles) viruses. Gene expression changes of antiviral genes (A, IRF7; B, IFNB1; C, CXCL9; and D, IFIH1) and inflammatory genes (E, IL8; F, CCL2; G, CXCL2; and H, VCAM) were determined 5, 8, 10, and 16 h p.i. by qRT-PCR and plotted as mean of FCs compared with uninfected control HUVECs (n = 3).

The Journal of Immunology DNA microarray hybridization and statistical data analyses Total cellular RNA was isolated from three independent experiments with wild-type (wt) HUVECs and three independent experiments with HUVECs that had been infected with empty retroviral expression vector or a vector containing the dominant-negative mutant of IKK2 (IKK2kd). Cells were infected for 5 h with the three influenza strains PR8, FPV, and H5N1 (RNeasy kit, Qiagen, Hilden, Germany). Samples were processed for microarray hybridization using Affymetrix Human Genome U133 Plus 2.0 Gene Arrays according to the manufacturer’s instructions (Affymetrix, Santa Clara, CA). Fluorescent signals were detected by the GeneChip Scanner 3000 and recorded and computed by GeneChip Operating Software version 1.4 (Affymetrix). For a more sophisticated data analysis, we used the Expressionist Suite software from GeneData (Basel, Switzerland) as described (20). Genes with a fold change (FC) of .2.0 and a p value # 0.05 (paired t test) were considered. On/off-regulated genes were evaluated as described (20) considering genes with on/off ratios of 0:3, 0:2, 1:3, 3:0, 2:0, and 3:1, respectively. From this group of on/off-regulated genes, we only included regulations with a high fold-change of $5 and a p value of , 0.05 to exclude on/off phenomenons occurring around the background threshold. Microarray data are Minimum Information About a Microarray Experiment compliant and deposited in Gene Expression Omnibus database (http://www.ncbi.nlm.nih.gov/geo/; accession number GSE13637). Parts of the data set concerning H5N1-infected HUVECs have also been used in a previous study of our group (18).

167 As an additional bioinformatics tool, we applied principal component analysis (PCA), which is a mathematical procedure that reduces the enormous number of possibly correlated gene expression values to a smaller number of uncorrelated variables called principal components (23). This method results in location of individual microarray data sets in a virtual three-dimensional vector space allowing comparisons of individual virusinduced expression patterns regarding variations (different location) and similarities (close location) in gene expression in an observer-independent manner. For three-dimensional performance, we chose data transformation into three principal components. To identify overrepresented functional categories of genes, we compared the distribution of gene ontology annotations on the Affymetrix U133 Plus 2.0 Array (Affymetrix) with the gene group of interest applying Fisher’s exact test. In case of genes that are represented by two or more probe sets, only one transcript was taken into account to avoid potential bias. For the promoter analyses, we took advantage of a computational method for transcriptional regulatory network inference, Computational Ascertainment of Regulatory Relationships Inferred from Expression, described by Haverty et al. (24). Briefly, microarray data and promoter sequence data derived from TRANSFAC database 7.0 are merged and checked against each other. Thereby, promoter regions of a group of unregulated genes are compared with the promoters of a group of regulated genes regarding transcription factor binding sites (BSs) to compute the relative overabundance of cis-elements in the group of regulated genes. So, overrepresented transcription factor BSs are detected and defined as potential regulators of

FIGURE 4. Role of NF-kB for the gene expression profiles induced in HUVECs by different influenza strains. A, Lysates of wt HUVECs, empty vectorinfected HUVECs, and HUVECs expressing IKK2kd mutant were immunoblotted against viral M1 protein (27 kDa) 5 h p.i. with PR8, FPV, and H5N1. Blot shows comparable expression of M1 in wt and empty vector-infected HUVECs. Blocking NF-kB did not significantly change M1 expression. PCA comparing gene profiles induced by PR8 (B), FPV (C), or H5N1 (D) virus in empty vector-transfected HUVECs (black clouds), in empty vector-transfected and virus-infected HUVECs (dark gray clouds), and in virus-infected HUVECs transfected with IKK2kd (light gray clouds). Data illustrate the strong NFkB dependence of gene expression changes induced by H5N1. FCs of gene expression (y-axis) in monolayer of empty vector- (black bars) or IKK2kdtransfected (gray bars) HUVECs determined by qRT-PCR 5 h p.i. with PR8 (E, F), FPV (G, H), and H5N1 (I, J). Bars represent means of FCs 6 SD compared with uninfected control HUVECs (n = 5). Asterisks mark significant inhibition of gene inducibility by H5N1 resulting from blocking NF-kB. *p , 0.05; **p , 0.005; ***p , 0.001.

168 a gene group of interest according to the statistical significance of overrepresentation.

Quantitative real-time RT-PCR cDNA was synthesized from 4 mg total RNA using RevertAid H Minus M-MuLV Reverse Transcriptase (Fermentas, St. Leon-Rot, Germany). Specific primers for each gene (sequences shown in Supplemental Table XI) were designed using Primer Express software (Applied Biosystems, Foster City, CA) and obtained from MWG Biotech (Ebersberg, Germany). Quantitative real-time RT-PCR (qRT-PCR) was performed using the QuantiTect SYBR Green PCR kit (Qiagen) as described (20) and data acquired with the ABI PRISM 7900 (Applied Biosystems). Gene expression was normalized to the endogenous housekeeping control gene RPL9, and relative expression of respective genes was calculated using the comparative threshold cycle method as described (25).

Short interfering RNA experiments A total of 1.2 3 105 cells were seeded on six-well plates. Twenty-four hours later, each well was transfected with 200 pmol short interfering RNA (siRNA; Qiagen) and 6 ml Oligofectamine (Invitrogen) in 200 ml OptiMEM (Life Technologies). The siRNAs used in this study consisted of 21nucleotide dsRNAs, each strand of which contained a 19-nucleotide target sequence and a two-uracil overhang at the 39 end and were predesigned for IRF3, IRF7, GATA6, and NFATC4 and validated for HMGA1 and negative control siRNA (Supplemental Table XII). Efficiency of mRNA or protein knockdown was measured by qRT-PCR or Western blot, respectively. Forty-eight hours after siRNA transfection, HUVECs were infected with 5 MOI of H5N1. In case of additional IFN-b pretreatment, HUVECs were incubated with 100 U/ml IFN-b1a (ImmunoTools, Friesoythe, Germany). Then, 5 h p.i., RNA was collected and processed for qRT-PCR.

Statistical analyses The results of all qRT-PCRs were assessed by Student t test and are expressed as means 6 SD.

Results Infection of primary ECs with HPAIV We used three influenza virus isolates to infect monolayers of HUVECs and analyzed the effects on gene expression changes: A/ PR8/34 (H1N1) (PR8) (a low pathogenic human reference influenza strain), FPV (highly pathogenic for birds) (26, 27), and an H5N1 isolate A/Thailand/KAN-1/2004 (an HPAIV isolated from a fatal human case) (28, 29). Infectious doses of all virus strains were adjusted to result in 80% of infected ECs as determined by immunofluorescence staining of an influenza A group-specific nucleoprotein (data not shown). Efficient infection was confirmed by immunoblotting of the viral proteins PB1, M1, and NS1 3, 5, and 8 h p.i., respectively (Fig. 1). Although at 5 h p.i. there were slight differences in viral protein expression suggesting an advantaged H5N1 propagation, the accumulation of viral proteins at 8 h p.i did not greatly differ between the different strains, indicating that the replication efficiency of all three viruses was at least similar within the first replication cycle. To map primary effects on gene expression, we therefore chose the time point of 5 h p.i., well within the first replication cycle of all three viruses. For 24 h p.i., we found a significant trend of 1 log level higher viral titers in supernatants of H5N1-infected HUVECs compared with PR8 and FPV infections (data not shown). Gene profiling of HPAIV-infected HUVECs After 5 h p.i., we processed total RNA of infected and uninfected control HUVECs for microarray hybridization. We exclusively focused on upregulated genes because unspecific cap-snatching mechanisms significantly contribute to the process of gene downregulation by influenza virus (30, 31). PCA displaying all influenzainducible genes on the particular gene profiles showed consistence of the profiles within the same experimental group and clear separation of the different experimental groups (Fig. 2A). PR8 induced

H5N1 TRANSFORMS IMMUNE RESPONSES IN ENDOTHELIA the highest number of genes (418 genes) due to a high proportion of predominantly unspecific gene regulations with low FCs between 2.0 and 4.0 followed by the H5N1 strain (149 genes) and the FPV virus (80 genes) (Fig. 2B, Supplemental Tables I–III). Forty-seven genes were induced by all influenza viruses (Supplemental Table IV), whereas 33 genes were specifically induced by H5N1, 22 genes specifically by FPV, and 291 genes specifically by PR8. Functional clustering according to gene ontology annotations revealed that genes induced by all viruses primarily belonged to inflammatory viral response genes and cell–cell signaling genes (Fig. 2C). A similar functional pattern was computed for PR8specific genes, whereas FPV-specific genes had functions assigned to metabolism, cell proliferation, and cell cycle control (Supplemental Fig. 1). In contrast, in the group of H5N1-specific genes, an overwhelming overrepresentation of chemokines and chemotactic genes was observed (Fig. 2D). Chemokines specifically induced by H5N1 included CCL7, CXCL1, CXCL5, CXCL6, and IL8 (Supplemental Table V). Table I. Transcription factor profiles of H5N1-induced gene programs in HUVECs Transcription Factor Binding Site

Universe of H5N1-induced genes Myogenic MADS factor MEF-2 Fetal Alz-50 clone 1 IRF8 IRF1 IRF Retroviral TATA box ISRE IRF7 E4BP4 GATA6 Hepatic leukemia factor Related to serum response factor, C4 TATA binding protein C/EBPd HMGA1 Meis-1b/HOXA9 heterodimeric binding Serum response factor Forkhead box D3 Meis-1a/HOXA9 heterodimeric binding Pax-6 Cell division control protein 5 NF-kB Homeo domain factor Pbx-1 HNF-3 NFATC4 STAT5B 58 KDA repressor protein Pit-1 Sex-determining region Y gene product Specifically H5N1-induced genes GATA6 STAT5A TTF-1, TITF1 STAT5B Androgen receptor POU3F2 TGIF Upstream stimulating factor NFATC4 Serum response factor Transcriptional repressor CDP HMGA1

p Valuea

1.52E-22 1.21E-21 9.25E-21 9.25E-21 3.34E-20 6.48E-17 1.07E-15 5.47E-14 3.75E-12 1.19E-10 2.29E-10 3.42E-10 1.31E-09 1.74E-09 2.07E-09 2.52E-09 4.95E-09 1.02E-07 1.15E-07 2.34E-07 3.79E-07 3.92E-07 5.67E-07 8.10E-07 8.59E-07 1.29E-06 4.41E-06 4.41E-06 7.41E-06 9.39E-16 1.87E-15 3.67E-13 1.56E-12 6.70E-11 6.83E-08 6.88E-08 3.83E-07 1.14E-05 2.16E-04 4.48E-04 1.98E-03

Boldface text indicates BSs that were not computed for any PR8- or FPV-related gene profile and additionally identified in the analysis of the entire H5N1-induced gene profile as well as in the analysis of specifically H5N1-induced genes. a p values indicate the significance of overrepresentation of binding sites in promoters of H5N1-regulated genes compared with unregulated genes. ISRE, IFN-stimulated response element.

The Journal of Immunology To screen whether the overrepresentation of inflammatory genes after H5N1 infections is temporal or sustained compared with PR8 and FPV infections, we performed qRT-PCRs of representative genes of the antiviral and inflammatory gene program 5, 8, 10, and 16 h after HPAIV infection (Fig. 3). Expression differences detected 5 h p.i. also persisted 16 h p.i. Differences in genes contributing to the antiviral response like IRF7, IFNB1, CXCL9, and IFIH1 were similar 5 h as well as 16 h p.i. They were strongest induced by PR8 virus and significantly lower by FPV and H5N1 virus (Fig. 3A–D). We also confirmed microarray data to that effect that H5N1 infections compared with PR8 and FPV viruses caused significantly stronger or even specific inductions of inflammatory genes like IL8, CCL2, CXCL2, and VCAM in HUVECs beginning 5 h p.i. (first replication cycle) and becoming clearer 16 h p.i. So, the gene profiles identified within the first replication cycle of all three viruses represent reliable data of primary HPAIV effects on endothelial gene expression. Role of NF-kB for gene expression changes in influenza virus-infected HUVECs With chemokines being exceedingly overrepresented among H5N1-specific genes and knowing that NF-kB is crucial for a

169 proinflammatory response in ECs (20), we blocked the NF-kB signaling pathway by introducing retrovirally an IKK2kd mutant into HUVECs (20, 21, 32). NF-kB blockade did not significantly attenuate viral M1 expression of each viral strain, indicating appropriate replication of all viruses in this EC system (Fig. 4A). PCA of microarray data raised in these HUVECs 5 h p.i. demonstrated that blocking of NF-kB did not significantly alter the global gene profiles induced by PR8 (Fig. 4B). For FPV, we found a slight shift toward the control profiles, suggesting a minor dependence on NF-kB (Fig. 4C) whereas H5N1-induced endothelial gene profiles significantly shifted back toward controls if NF-kB was blocked (Fig. 4D), confirming recent findings of our group in terms of an essential impact of NF-kB signaling in H5N1 infections (18). However, the NF-kB dependence of the H5N1 virus did not reach the same level that we observed in a previous study for TNF-induced gene expression changes in human ECs, which are virtually to 100% dependent on NF-kB (20). To confirm microarray data concerning NF-kB dependence and virus specificity, we performed qRT-PCRs (Supplemental Table VI). Genes could be classified into three different groups: first, H5N1-strain specific genes; second, influenza virus inducible genes with NF-kB dependence predominantly in the H5N1 strain; and third, genes induced in an NF-kB–dependent mode by all influenza

FIGURE 5. Role of IRF3 and IRF7 in H5N1-mediated gene expression changes in HUVECs. A, Western blot shows the detection of IRF3 (50 kDa) in HUVECs transfected with IRF3 siRNA (24 h, 48 h, 72 h) and nontargeting control siRNA. Transfection of HUVECs with IRF3 siRNA knocks down IRF3 protein expression for 48 h, whereas after 72 h, IRF3 protein expression reappears. Detection of b-actin serves as protein loading control. B, Demonstration of lasting successful knockdown of IRF3 protein expression in HUVEC by siRNA pre- and p.i. with H5N1 virus. Detection of b-actin serves as protein loading control. C, Western blot shows induction of IRF7 (63 kDa) protein expression and knockdown by siRNA in uninfected and H5N1-infected HUVECs. D, Western blot shows the detection of IRF7 in control siRNA-transfected HUVECs, overexpression of IRF7 in IRF3 siRNA-transfected HUVECs, and successful knockdown of IRF7 in IRF7 siRNA-transfected HUVECs. Detection of b-actin serves as protein loading control. FCs of mRNA expression (y-axis) in H5N1-infected compared with uninfected HUVECs transfected with untargeted control siRNA (black bars), IRF3-directed siRNA (E, gray bars), and IRF7-directed siRNA (F, gray bars) determined by qRT-PCR. IRF7* in F shows the induction of IRF7 in HUVECs pretreated with IFN-b before siRNA transfections and H5N1 infection. Bars represent means 6 SD (n = 5). *p , 0.005; **p , 0.001.

170 virus. Primarily proinflammatory genes such as CCL5, CXCL10, CXCL11, TNFSF10, CXCL2, TLR3, and VCAM belonged to the third group of generally virus-inducible NF-kB–dependent genes (Supplemental Table VI). IL8 and CCL2 proved true as H5N1specific NF-kB–dependent genes 5 h p.i. (Fig. 4E, 4G, 4I). From the second group of generally influenza-inducible genes with superior NF-kB dependence in the H5N1 profile, we selected five genes for qRT-PCR (SELE, IFNB1, CXCL9, IFIH1, and IRF7). We verified that the induction of none of these genes was NF-kB dependent in the case of PR8 infections (Fig. 4F). FPV only induced SELE and IRF7 in an NF-kB–dependent manner (Fig. 4H). In contrast, virtually all of these genes were NF-kB dependently induced by H5N1 (Fig. 4J). Interestingly, in H5N1-infected ECs also, genes that launch an antiviral response, including IFNB1, CXCL9 and IFIH1, were NF-kB dependently induced, whereas this was not the case or tendentiously even contrary in PR8 or FPV infections. Besides the NF-kB dependence, it was very striking that the IFNB1 response induced by the H5N1 strain in ECs was significantly weaker compared with PR8 and FPV (Fig. 4F, 4H, 4J).

H5N1 TRANSFORMS IMMUNE RESPONSES IN ENDOTHELIA IRF3 is essential for the initiation of an IFN response to H5N1 virus infections in HUVECs To identify additional transcriptional regulators crucial for H5N1mediated gene expression, we performed promoter analyses (Table I). This bioinformatic approach revealed that in ECs, next to NFkB also transcription factor BSs of IRFs, namely IRF consensus BSs, IRF8, IRF1, IFN-stimulated response element, and IRF7, are significantly overrepresented in promoters of H5N1-induced genes (Table I) as well as in the group of PR8- and FPV-induced genes (Supplemental Tables VII, IX). In other cell systems than human ECs, IRF3 is important for the induction of the IFN–STAT1/2 pathway, whereas IRF7 rather overtakes the tasks of IRF3 in the second stage (33, 34). To examine the relevance of IRF3 and IRF7 in ECs, we knocked them down by targeted siRNA (Fig. 5A–D; mean of 70% inhibition of IRF3 mRNA expression and 85% inhibition of IRF7 mRNA expression according to qRT-PCR). IRF3 expression remained readily knocked down for 48 h after siRNA transfection (Fig. 5A) and also after H5N1 infection (Fig. 5B). Because IRF7 is only hardly expressed in resting HUVECs, we

FIGURE 6. HMGA1 and NFATC4 regulate H5N1-mediated gene expression changes in HUVECs. A, Western blot shows knockdown of HMGA1 (18 kDa) protein expression by siRNA compared with control siRNA-transfected HUVECs pre- and p.i. with H5N1 virus. Detection of b-actin serves as protein loading control. B, FC of mRNA of NF-kB dependently induced genes in H5N1-infected compared with uninfected HUVEC transfected with control siRNA (black bars) or HMGA1-directed siRNA (gray bars) determined by qRT-PCR (n = 10). C, Western blot shows knockdown of NFATC4 (140 kDa) protein expression by siRNA in HUVEC pre- and p.i. with H5N1 virus. Detection of b-actin serves as protein loading control. D, FC of mRNA of NF-kB dependently induced genes in H5N1-infected compared with uninfected HUVECs transfected with control siRNA (black bars) or NFATC4-directed siRNA (gray bars) determined by qRT-PCR (n = 10). FC of mRNA of NF-kB independently induced genes in H5N1-infected compared with uninfected HUVECs transfected with control siRNA (black bars), HMGA1-directed siRNA (E, gray bars), or NFATC4-directed siRNA (F, gray bars) determined by qRT-PCR (n = 6). B and D–F, Bars represent means 6 SEM. *p , 0.05; **p , 0.005.

The Journal of Immunology tested its inducibility in uninfected HUVECs by IFN-b and verified successful and lasting knockdown of IRF7 after H5N1 infection (Fig. 5C). Interestingly, IRF3 knockdown resulted in a considerable IRF7 overexpression (Fig. 5D). qRT-PCRs in H5N1infected control and knockdown cells confirmed that IRF3 is essential for the induction of IFNB1 and CXCL9 in ECs by H5N1 virus (Fig. 5E). Surprisingly, some proinflammatory genes, such as CXCL11, VCAM, and SELE were strongly inducible in HUVECs with blocked IRF3. Knockdown of IRF7 in HUVECs neither changed the level of IRF3 expression nor did it significantly influence H5N1-mediated gene expression changes 5 h p.i. (Fig. 5F). IRF7 mRNA itself showed no relevant changes after IRF7 knockdown, maybe due to the very low basal expression level. We therefore enhanced IRF7 expression by IFN-b treatment before IRF7 knockdown and H5N1 infection and found a clear inhibitory effect on IRF7 mRNA induction by H5N1 virus (Fig. 5F, IRF7*). HMGA1 and NFATC4 represent novel transcriptional regulators in the response of H5N1 virus-infected HUVECs Next, we performed a promoter analysis of genes exclusively induced upon H5N1 virus infection (Table I) in ECs. The same dual analysis was performed for PR8- and FPV-induced genes (Supplemental Tables VII–X). So, we established virtually virus strainspecific endothelial transcription factor profiles. We filtered out those specifically H5N1-relevant transcription factors that are also overrepresented in the promoter regions of the entire H5N1induced gene profile but not relevant for any PR8- or FPV-related gene profile. In that way, three candidates of H5N1-specifc transcription factors were identified (gray shaded in Table I): GATA6, HMGA1, and NFATC4. Using an siRNA approach, we could not confirm any regulatory role of GATA6 for H5N1-induced genes (data not shown). In contrast, transfection of HMGA1-directed siRNA in HUVECs (Fig. 6A, mean of 80% inhibition of HMGA1 mRNA expression) led to a pattern of inhibited or decoupled gene transcription upon H5N1 virus infection that differed significantly from the effects of NF-kB or IRF3 blocking on gene transcription (Figs. 4I, 4J, 5E, 6C). The induction of several proinflammatory genes as SELE, CXCL11, TNFSF10A, and CXCL2 was dependent on HMGA1 (Fig. 6B). Except for SELE, these genes belonged to the group of generally NF-kB–dependent and virus-inducible genes not dependent on IRF3 (Fig. 5E, Supplemental Table VI). In contrast, genes such as IL8, IFIH1, or IRF7 that were H5N1 specifically dependent on NF-kB (Fig. 4I, 4J) and not IRF3 dependent (Fig. 5E) were not affected by HMGA1 knockdown (Fig. 6B). Interestingly, the induction of IFNB1 and CXCL9, both strictly IRF3 and H5N1 specifically NF-kB dependent (Fig. 4C, 4J), significantly increased upon knockdown of HMGA1 (Fig. 6B). The impact of NFATC4 knockdown in HUVECs (Fig. 6C, mean of 50% inhibition of NFATC4 mRNA expression) was differential compared with the knockdown of NF-kB, IRF3, or HMGA1 (Figs. 4, 5E, 6B) and resulted in specific changes of the H5N1-induced gene profile in HUVECs. The induction of all selected H5N1inducible genes was suppressed when NFATC4 was knocked down (Fig. 6D). According to microarray data, H5N1 infection does not induce HMGA1 mRNA and only slightly NFATC4 mRNA (Supplemental Tables I, V), whereas at the protein level, induction of none of could be observed (Fig. 6A, 6C). To clarify the relation between NF-kB and HMGA1 and NFATC4, we focused on genes specifically induced by H5N1 in an NF-kB–independent mode as CCL7, CXCL5, and CXCL6 (Supplemental Table VI). All genes were induced by H5N1 in an HMGA1- and NFATC4-dependent mode (Fig. 6E, 6F). Data demonstrate that HMGA1 and NFATC4 modify

171 the H5N1-induced proinflammatory as well as antiviral gene profile independent of NF-kB. The interaction of IRF3, NF-kB, HMGA1, and NFATC4 in H5N1-infected HUVECs results in an imbalanced inflammatory response with an overwhelming activation of common and specific proinflammatory gene programs and an impaired antiviral gene program compared with low pathogenic influenza strains (Fig. 7).

Discussion Diseases caused by HPAIVs are characterized by systemic dissemination leading to an SIRS (2, 4, 6–8, 10). Given that ECs significantly contribute to SIRS and considering the EC tropism of HPAIV, we examined the network of signaling pathways and transcriptional regulators specifically activated upon H5N1 virus infection of HUVECs. For this purpose, we used a comparative approach with three influenza strains of different pathogenicity for birds and humans. PR8 is a low pathogenic human influenza strain, FPV is a H7 virus highly pathogenic for birds, whereas H5N1 viruses are known to infect and kill humans. Within the first viral replication cycle, propagation efficiency of all influenza virus strains was comparable. Each viral strain induced a distinct characteristic gene profile. Functional clustering revealed that genes induced by all strains were inflammatory viral response genes and cell–cell signaling genes. The functional pattern of PR8specific genes was very similar to that of the intersection of generally influenza-inducible genes. Among FPV-specific genes, no relevant highly overrepresented gene groups were observed. On the contrary, in the set of H5N1 specifically induced genes proinflammatory cytokines were enormously overrepresented, confirming the manifold described cytokine storm (7, 8). Our data suggest that ECs may be a major source of cytokines in the pathogenesis of H5N1-induced SIRS. Mircroarray data from PR8-, FPV- and H5N1-infected HUVECs for which NF-kB pathway had been blocked illustrated that NFkB is strictly required for the H5N1-induced gene response but less important for the FPV-induced profile and of minor relevance for PR8-induced genes. Thereby, we confirm the pivotal role of NF-kB signaling for the endothelial response to H5N1 infections worked out in a recent study of our group (18) and demonstrate additionally the specificity of NF-kB dependence for the induction of several inflammatory (IL8, CCL2, and SELE) and important antiviral genes (IFNB1, CXCL9, IFIH1, and IRF7) in H5N1

FIGURE 7. Signaling pathways activated by H5N1 virus in HUVECs. Model figure illustrates the signaling pathways activated by H5N1 virus in HUVECs. HMGA1 and NFATC4 are specifically involved by H5N1 (black arrows), whereas NF-kB and IRF3 are also activated by PR8 and FPV virus (gray arrows). Drawn arrows indicate positive and dotted lines negative regulation.

172 infections compared with low-pathogenic influenza virus. Other data concerning the role of NF-kB in influenza infections are conflicting most likely due to cell-specific responses to influenza infections (35) or due to the use of genetically modified cells or viruses (36). It has been argued that cells with low NF-kB activity were resistant to influenza virus infections (37) and that NF-kB plays a supportive role for virus propagation itself (38). Lee et al. (39) examined signaling pathways in H5N1-infected human macrophages and assigned a superior regulatory function for the induction of the cytokine storm to p38 MAPK but negate an exceptional role for NF-kB. However, their conclusion bases on the assumption of similar levels of NF-kB activation in H5N1 infections compared with a low pathogenic H1N1 virus without performing blocking experiments. In our studies, we generally find an imperative necessity of NFkB signaling to mediate strong proinflammatory gene responses in ECs as shown in this study in H5N1 infections and previously shown for TNF (20) and Candida albicans (32). NF-kB activation obviously hallmarks responses to inflammatory agents with capacity to cause SIRS or sepsis. Besides a differential NF-kB dependence, representatives of antiviral genes as IFNB1 and CXCL9 are significantly weaker induced 5 h p.i. by H5N1 than by PR8 or FPV. In a human bronchial epithelial cell model, Zeng et al. (17) similarly worked out that the pathogenicity of different H5N1 strains correlated with the capability to attenuate the type I IFN response. Our data suggest that H5N1 causes an imbalance between a strong NF-kB–dependent proinflammatory response and an impaired induction of the antiviral IFN program in ECs, which might be the key to the fatality of H5N1 infections. Altogether, this specific gene program apparently ensures a survival and replication advantage for H5N1 virus in ECs as reflected by constantly higher H5N1 virus titers 24 h p.i. compared with PR8 and FPV. Furthermore, our bioinformatic approach suggested IRFs to be equally involved in PR8-, FPV-, and H5N1-induced gene inductions in HUVECS, which could be verified experimentally. As in other cell systems already demonstrated (33, 34, 40, 41), we confirm in this study a crucial role of IRF3 for the induction of IFNB1 in ECs. IRF7 has no regulatory role during the initial phase of virus infection. Thus, both NF-kB as well as IRF3 play outstanding roles in the endothelial signaling network activated by H5N1 viruses. An interesting side finding was that IRF7 is upregulated in IRF3blocked HUVECs. Either IRF7 overexpression virtually compensates the IRF3 loss or it is a matter of an unspecific IFNa-inducing activity of IRF3 siRNA (42). Analyzing promoter regions of specifically H5N1 virus-induced genes indicated that GATA6, HMGA1, and NFATC4 are novel putative transcription factors with H5N1 virus-specific relevance. For GATA6, no biological impact on the H5N1-induced endothelial gene program could be confirmed, demonstrating that in silico approaches help to narrow down the selection of potential transcriptional regulators but are no final proof of relevance. In contrast, HMGA1 came out with an interesting pattern of regulatory functions on H5N1-induced genes that clearly differed from the role of NF-kB and IRF3. Blocking of HMGA1 inhibited the overwhelming induction of proinflammatory genes by H5N1. HMGA1 is reported to play an important role in the formation of the IFN-b enhancesome, crucially launching the type I IFN response (41, 43). Thereby, HMGA1 acts as a positive as well as negative regulator of IFNB1 expression depending on its acetylation status (44, 45). In our setting, the knockdown of HMGA1 rather increased IFNB1 expression, revealing HMGA1 as a negative regulator of the IFN-b enhancesome in H5N1-infected ECs. H5N1 virus might induce repression of the antiviral gene program

H5N1 TRANSFORMS IMMUNE RESPONSES IN ENDOTHELIA by appropriate modulation of the acetylation status of HMGA1 (46). NFATC4 was the third putative H5N1 virus-specific transcription factor. In contrast to other NFAT family members (47), NFATC4 has not been reported until now to regulate inflammatory genes. The factor was hitherto not known to be expressed in cells of the immune system but to regulate cell differentiation of cardiomyocytes and be expressed in perivascular tissue directing the development of ECs (47, 48). We verified that NFATC4 is also expressed in HUVECs. After blocking NFATC4, the H5N1-induced proinflammatory as well as antiviral gene program is inhibited. No correlation to the mode of IRF3 or HMGA1 dependence was detectable. Moreover, both HMGA1 and NFATC4 clearly regulate NF-kB independently H5N1-induced genes. All findings identify HMGA1 and NFATC4 for the first time, to our knowledge, to be transcription factors that independently modulate the proinflammatory as well as antiviral endothelial gene program in an H5N1-specific manner. Taken together, we demonstrate that a systems biology approach is a very useful tool to evaluate proinflammatory and antiviral response programs in ECs. With respect to H5N1 infections of HUVECs, we observed an attenuation of the antiviral type I IFN response and identified HMGA1 and NFATC4 as two novel molecular targets contributing to a strong overbalance of proinflammatory mechanisms. Our data may help to design strategies for the development of novel therapeutic approaches for the treatment of H5N1 infections impeding a devastating proinflammatory response.

Acknowledgments We thank U. Nordhues for excellent technical assistance.

Disclosures The authors have no financial conflicts of interest.

References 1. Chen, J. M., J. W. Chen, J. J. Dai, and Y. X. Sun. 2007. A survey of human cases of H5N1 avian influenza reported by the WHO before June 2006 for infection control. Am. J. Infect. Control 35: 467–469. 2. Yuen, K. Y., P. K. Chan, M. Peiris, D. N. Tsang, T. L. Que, K. F. Shortridge, P. T. Cheung, W. K. To, E. T. Ho, R. Sung, and A. F. Cheng. 1998. Clinical features and rapid viral diagnosis of human disease associated with avian influenza A H5N1 virus. Lancet 351: 467–471. 3. Tran, T. H., T. L. Nguyen, T. D. Nguyen, T. S. Luong, P. M. Pham, V. C. Nguyen, T. S. Pham, C. D. Vo, T. Q. Le, T. T. Ngo, et al; World Health Organization International Avian Influenza Investigative Team. 2004. Avian influenza A (H5N1) in 10 patients in Vietnam. N. Engl. J. Med. 350: 1179–1188. 4. Beigel, J. H., J. Farrar, A. M. Han, F. G. Hayden, R. Hyer, M. D. de Jong, S. Lochindarat, T. K. Nguyen, T. H. Nguyen, T. H. Tran, et al; Writing Committee of the World Health Organization (WHO) Consultation on Human Influenza A/H5. 2005. Avian influenza A (H5N1) infection in humans. N. Engl. J. Med. 353: 1374–1385. 5. Wong, S. S., and K. Y. Yuen. 2006. Avian influenza virus infections in humans. Chest 129: 156–168. 6. Writing Committee of the Second World Health Organization Consultation on Clinical Aspects of Human Infection with Avian Influenza A (H5N1) Virus, A. N. Abdel-Ghafar, T. Chotpitayasunondh, Z. Gao, F. G. Hayden, D. H. Nguyen, J. de, A. Naghdaliyev, J. S. Peiris, N. Shindo, S. Soeroso, and T. M. Uyeki. 2008. Update on avian influenza A (H5N1) virus infection in humans. N. Engl. J. Med. 358: 261–273. 7. Chan, M. C., C. Y. Cheung, W. H. Chui, S. W. Tsao, J. M. Nicholls, Y. O. Chan, R. W. Chan, H. T. Long, L. L. Poon, Y. Guan, and J. S. Peiris. 2005. Proinflammatory cytokine responses induced by influenza A (H5N1) viruses in primary human alveolar and bronchial epithelial cells. Respir. Res. 6: 135. 8. Cheung, C. Y., L. L. Poon, A. S. Lau, W. Luk, Y. L. Lau, K. F. Shortridge, S. Gordon, Y. Guan, and J. S. Peiris. 2002. Induction of proinflammatory cytokines in human macrophages by influenza A (H5N1) viruses: a mechanism for the unusual severity of human disease? Lancet 360: 1831–1837. 9. de Jong, M. D., V. C. Bach, T. Q. Phan, M. H. Vo, T. T. Tran, B. H. Nguyen, M. Beld, T. P. Le, H. K. Truong, V. V. Nguyen, et al. 2005. Fatal avian influenza A (H5N1) in a child presenting with diarrhea followed by coma. N. Engl. J. Med. 352: 686–691. 10. Klenk, H. D. 2005. Infection of the endothelium by influenza viruses. Thromb. Haemost. 94: 262–265.

The Journal of Immunology 11. Ocan˜a-Macchi, M., M. Bel, L. Guzylack-Piriou, N. Ruggli, M. Liniger, K. C. McCullough, Y. Sakoda, N. Isoda, M. Matrosovich, and A. Summerfield. 2009. Hemagglutinin-dependent tropism of H5N1 avian influenza virus for human endothelial cells. J. Virol. 83: 12947–12955. 12. Perkins, L. E., and D. E. Swayne. 2001. Pathobiology of A/chicken/Hong Kong/ 220/97 (H5N1) avian influenza virus in seven gallinaceous species. Vet. Pathol. 38: 149–164. 13. Szretter, K. J., S. Gangappa, X. Lu, C. Smith, W. J. Shieh, S. R. Zaki, S. Sambhara, T. M. Tumpey, and J. M. Katz. 2007. Role of host cytokine responses in the pathogenesis of avian H5N1 influenza viruses in mice. J. Virol. 81: 2736–2744. 14. Yao, L., C. Korteweg, W. Hsueh, and J. Gu. 2008. Avian influenza receptor expression in H5N1-infected and noninfected human tissues. FASEB J. 22: 733– 740. 15. Droebner, K., S. J. Reiling, and O. Planz. 2008. Role of hypercytokinemia in NF{kappa}B p50 deficient mice after H5N1 influenza A virus infection. J. Virol. 82: 11461–11466. 16. Salomon, R., E. Hoffmann, and R. G. Webster. 2007. Inhibition of the cytokine response does not protect against lethal H5N1 influenza infection. Proc. Natl. Acad. Sci. USA 104: 12479–12481. 17. Zeng, H., C. Goldsmith, P. Thawatsupha, M. Chittaganpitch, S. Waicharoen, S. Zaki, T. M. Tumpey, and J. M. Katz. 2007. Highly pathogenic avian influenza H5N1 viruses elicit an attenuated type I interferon response in polarized human bronchial epithelial cells. J. Virol. 81: 12439–12449. 18. Schmolke, M., D. Viemann, J. Roth, and S. Ludwig. 2009. Essential impact of NF-kappaB signaling on the H5N1 influenza A virus-induced transcriptome. J. Immunol. 183: 5180–5189. 19. Mazur, I., W. J. Wurzer, C. Ehrhardt, S. Pleschka, P. Puthavathana, T. Silberzahn, T. Wolff, O. Planz, and S. Ludwig. 2007. Acetylsalicylic acid (ASA) blocks influenza virus propagation via its NF-kappaB-inhibiting activity. Cell. Microbiol. 9: 1683–1694. 20. Viemann, D., M. Goebeler, S. Schmid, K. Klimmek, C. Sorg, S. Ludwig, and J. Roth. 2004. Transcriptional profiling of IKK2/NF-kappa B- and p38 MAP kinase-dependent gene expression in TNF-alpha-stimulated primary human endothelial cells. Blood 103: 3365–3373. 21. Denk, A., M. Goebeler, S. Schmid, I. Berberich, O. Ritz, D. Lindemann, S. Ludwig, and T. Wirth. 2001. Activation of NF-kappa B via the Ikappa B kinase complex is both essential and sufficient for proinflammatory gene expression in primary endothelial cells. J. Biol. Chem. 276: 28451–28458. 22. Viemann, D., A. Strey, A. Janning, K. Jurk, K. Klimmek, T. Vogl, K. Hirono, F. Ichida, D. Foell, B. Kehrel, et al. 2005. Myeloid-related proteins 8 and 14 induce a specific inflammatory response in human microvascular endothelial cells. Blood 105: 2955–2962. 23. Alter, O., P. O. Brown, and D. Botstein. 2000. Singular value decomposition for genome-wide expression data processing and modeling. Proc. Natl. Acad. Sci. USA 97: 10101–10106. 24. Haverty, P. M., U. Hansen, and Z. Weng. 2004. Computational inference of transcriptional regulatory networks from expression profiling and transcription factor binding site identification. Nucleic Acids Res. 32: 179–188. 25. Liu, W., and D. A. Saint. 2002. A new quantitative method of real time reverse transcription polymerase chain reaction assay based on simulation of polymerase chain reaction kinetics. Anal. Biochem. 302: 52–59. 26. Banks, J., E. Speidel, and D. J. Alexander. 1998. Characterisation of an avian influenza A virus isolated from a human—is an intermediate host necessary for the emergence of pandemic influenza viruses? Arch. Virol. 143: 781–787. 27. Kurtz, J., R. J. Manvell, and J. Banks. 1996. Avian influenza virus isolated from a woman with conjunctivitis. Lancet 348: 901–902. 28. Centers for Disease Control and Prevention (CDC). 2004. Cases of influenza A (H5N1)—Thailand, 2004. MMWR Morb. Mortal. Wkly. Rep. 53: 100–103. 29. Chotpitayasunondh, T., K. Ungchusak, W. Hanshaoworakul, S. Chunsuthiwat, P. Sawanpanyalert, R. Kijphati, S. Lochindarat, P. Srisan, P. Suwan, Y. Osotthanakorn, et al. 2005. Human disease from influenza A (H5N1), Thailand, 2004. Emerg. Infect. Dis. 11: 201–209.

173 30. Katze, M. G., and R. M. Krug. 1984. Metabolism and expression of RNA polymerase II transcripts in influenza virus-infected cells. Mol. Cell. Biol. 4: 2198–2206. 31. Engelhardt, O. G., and E. Fodor. 2006. Functional association between viral and cellular transcription during influenza virus infection. Rev. Med. Virol. 16: 329–345. 32. Mu¨ller, V., D. Viemann, M. Schmidt, N. Endres, S. Ludwig, M. Leverkus, J. Roth, and M. Goebeler. 2007. Candida albicans triggers activation of distinct signaling pathways to establish a proinflammatory gene expression program in primary human endothelial cells. J. Immunol. 179: 8435–8445. 33. Sato, M., H. Suemori, N. Hata, M. Asagiri, K. Ogasawara, K. Nakao, T. Nakaya, M. Katsuki, S. Noguchi, N. Tanaka, and T. Taniguchi. 2000. Distinct and essential roles of transcription factors IRF-3 and IRF-7 in response to viruses for IFN-alpha/beta gene induction. Immunity 13: 539–548. 34. Honda, K., A. Takaoka, and T. Taniguchi. 2006. Type I interferon [corrected] gene induction by the interferon regulatory factor family of transcription factors. Immunity 25: 349–360. 35. Bernasconi, D., C. Amici, S. La Frazia, A. Ianaro, and M. G. Santoro. 2005. The IkappaB kinase is a key factor in triggering influenza A virus-induced inflammatory cytokine production in airway epithelial cells. J. Biol. Chem. 280: 24127–24134. 36. Wang, X., M. Li, H. Zheng, T. Muster, P. Palese, A. A. Beg, and A. Garcı´aSastre. 2000. Influenza A virus NS1 protein prevents activation of NF-kappaB and induction of alpha/beta interferon. J. Virol. 74: 11566–11573. 37. Nimmerjahn, F., D. Dudziak, U. Dirmeier, G. Hobom, A. Riedel, M. Schlee, L. M. Staudt, A. Rosenwald, U. Behrends, G. W. Bornkamm, and J. Mautner. 2004. Active NF-kappaB signalling is a prerequisite for influenza virus infection. J. Gen. Virol. 85: 2347–2356. 38. Wurzer, W. J., C. Ehrhardt, S. Pleschka, F. Berberich-Siebelt, T. Wolff, H. Walczak, O. Planz, and S. Ludwig. 2004. NF-kappaB-dependent induction of tumor necrosis factor-related apoptosis-inducing ligand (TRAIL) and Fas/FasL is crucial for efficient influenza virus propagation. J. Biol. Chem. 279: 30931– 30937. 39. Lee, D. C., C. Y. Cheung, A. H. Law, C. K. Mok, M. Peiris, and A. S. Lau. 2005. p38 mitogen-activated protein kinase-dependent hyperinduction of tumor necrosis factor alpha expression in response to avian influenza virus H5N1. J. Virol. 79: 10147–10154. 40. Au, W. C., P. A. Moore, W. Lowther, Y. T. Juang, and P. M. Pitha. 1995. Identification of a member of the interferon regulatory factor family that binds to the interferon-stimulated response element and activates expression of interferon-induced genes. Proc. Natl. Acad. Sci. USA 92: 11657–11661. 41. Wathelet, M. G., C. H. Lin, B. S. Parekh, L. V. Ronco, P. M. Howley, and T. Maniatis. 1998. Virus infection induces the assembly of coordinately activated transcription factors on the IFN-beta enhancer in vivo. Mol. Cell 1: 507–518. 42. Hornung, V., M. Guenthner-Biller, C. Bourquin, A. Ablasser, M. Schlee, S. Uematsu, A. Noronha, M. Manoharan, S. Akira, A. de Fougerolles, et al. 2005. Sequencespecific potent induction of IFN-alpha by short interfering RNA in plasmacytoid dendritic cells through TLR7. Nat. Med. 11: 263–270. 43. Merika, M., and D. Thanos. 2001. Enhanceosomes. Curr. Opin. Genet. Dev. 11: 205–208. 44. Munshi, N., M. Merika, J. Yie, K. Senger, G. Chen, and D. Thanos. 1998. Acetylation of HMG I(Y) by CBP turns off IFN beta expression by disrupting the enhanceosome. Mol. Cell 2: 457–467. 45. Munshi, N., T. Agalioti, S. Lomvardas, M. Merika, G. Chen, and D. Thanos. 2001. Coordination of a transcriptional switch by HMGI(Y) acetylation. Science 293: 1133–1136. 46. Reeves, R. 2001. Molecular biology of HMGA proteins: hubs of nuclear function. Gene 277: 63–81. 47. Macian, F. 2005. NFAT proteins: key regulators of T-cell development and function. Nat. Rev. Immunol. 5: 472–484. 48. Horsley, V., and G. K. Pavlath. 2002. NFAT: ubiquitous regulator of cell differentiation and adaptation. J. Cell Biol. 156: 771–774.