Comparative Analysis of Genes Induced by Respiratory Syncytial Virus and DsRNA in Human Epithelial Cells

Epithelial cells are the primary target of respiratory viral infections and play a pivotal role in virusinduced lung infl ammation and in anti viral immune response. A common signal for the presence of viral infections and induction of infl ammation is recognition of double stranded RNA (dsRNA). Thus far, there has not been a high-throughput transcrptome analysis of RSVor dsRNA-induced genes in primary human bronchial epithelial cells (PHBE), nor there has been a comparison between dsRNAand RSV-induced genes. To establish the transcriptome profi les and to determine the contribution of dsRNA in the induction of infl ammation during respiratory virus infection, we compared the gene expression profi les of PHBE cells that were infected with Respiratory Syncytial Virus (RSV) or were treated with dsRNA. Our transcriptome analysis showed that RSV infection and and dsRNA treatment induced up-regulation of 2024 and 159 genes in PHBE respectively. Comparison of genes revealed that RSV and dsRNA commonly induced 75 genes in PHBE cells. The common up-regulated genes were functionally grouped in multiple response pathways involved in infl ammation and immune responses. Interestingly, there were several previously unreported genes that were up-regulated in primary human epithelial cells that are relevant to a TH2 allergic phenotype. This comparison of a high-throughput gene expression study offers a comprehensive view of transcriptional changes induced by dsRNA and RSV, and importantly compares dsRNA-induced genes with RSV-induced genes in PHBE cells.


Introduction
RSV is the most common cause of childhood respiratory viral infections, and can lead to severe bronchiolitis, which is associated with an increase in the induction of asthma (Singh et al., 2007).Previous transcriptomic studies have reported gene expression analysis of RSV-infected cell lines (Zhang et al., 2001), RSV-infected mice (Janssen et al., 2007) and RSV-infected infant whole and cord blood (Fjaerli et al., 2006;Fjaerli et al., 2007).Thus far, there has not been a transciptome analysis of genes that are induced during RSV infection in primary human bronchial epithelial cells, which are the primary target of RSV and other respiratory viruses.
During viral infections, including infections with respiratory viruses, dsRNA can be formed, which acts as a potent inducer of infl ammation.Experimentally, dsRNA, both intracellular and extracellular, has been widely used to mimic viral-induced inflammation.Previous high-throughput gene analysis using dsRNA in the form of poly I:C have reported the transcriptomeic changes in human glioma GRE cells (Geiss, 2001), peripheral blood mononuclear cells (PBMC's) (Huang et al., 2006) and rat pancreatic beta cells (Rasschaert et al., 2003).However, to date there has been no characterization or comparison of RSV-and Limmon dsRNA-induced genes in primary human airway epithelial cells.
In this study, the comprehensive gene expression profiles of PHBE treated with dsRNA or infected with RSV were analyzed using total cellular RNA, hybridized onto high-density oligonucleotide arrays (Agilant Human Array 1A version 2), which contain 20,000 probes.Interestingly, ours results reveal novel RSV-and dsRNA-induced genes.Some of the novel RSV-induced genes identifi ed in our result including ADAM8 and IL13RA2.ADAM8 has been reported to be induced in experimental asthma in mice, and in patients with severe and moderate asthma (Foley et al., 2007;King et al., 2004;Matsuno et al., 2006;Richens et al., 2007) and IL-13RA2 is important for IL-13 signaling and TGF-μ1-mediated lung fi brosis (Fichtner-Feigl et al., 2007;Fichtner-Feigl et al., 2006).
We also identifi ed novel dsRNA induced genes such as SOCS3 and PRDM1 (BLIMP-1).A report by Seki et al. (2003) showed that SOCS3 was important in the induction and maintenance of allergic TH2 reactions.BLIMP-1 is required for epidermal cell c-myc regulation and for the development of the effector function of B and T lymphocytes ( Lin, 2000;Lin, 1997;Martins, 2008;Messika et al., 1998;Zhou et al., 2007).
Importantly, study showed that 75 genes were commonly induced by RSV infection or dsRNA treatment in PHBE cells.These genes were functionally grouped in multiple response pathways such as infl ammation, apoptosis, transcription regulation, cell signaling, and cell differentiation.The induction of gene expression was confi rmed by determination of changes in expression of select proteins.This is, to our knowledge, is the fi rst study that has analyzed global gene expression in PHBE cells after dsRNA treatment and RSV infection.This characterization of the common genes validates the use of dsRNA as viral mimetic to detect a subset of proinfl ammatory signal during RSV infection, and may likely contribute to developing specifi c anti-infl ammatory agents.
The aim of this study is to unravel the gene expression profi le of human epithelial cell induced by RSV infection and DsRNA treatment.

Cells, tissue culture conditions, antibodies and reagents.
The primary human bronchial epithelial (PHBE) were grown as a monolayer at 37 o C in a 5% CO2 humidifi ed chamber in a serumfree epithelial cell growth medium from Cambrex (Lonza, Walkersville, MD).The synthetic dsRNA, poly I:C, was purchased from Sigma Chemicals (St. Louis, MO), dissolved in phosphate-buffered saline, and was used at concentrations indicated in each fi gure legend.The dsRNA structural fi delity of poly I:C was determined by the ability of each dsRNA batch to activate the dsRNAspecifi c protein kinase PKR in in vitro kinase assays (Imani, 1988).

RNA extraction and RT-PCR
For RNA extraction, cells were left untreated, were treated with 1 μg/ml poly I:C, or were infected with RSV at multiplicity of infection (MOI) of 2.5 plaque forming unit (pfu)/ cell.After 2 hr for dsRNA and 24 hr for RSV, total cellular RNA was isolated using the TRIzol reagent (Invitrogen, Carlsbad, CA) according to the manufacturers' instructions.The quality of RNA was evaluated by electrophoresis and only high quality RNA was used in the microarray and RT-PCR experiments.
For real-time PCR, fi rst strand cDNA was synthesized using superscript reverse transcriptase (Invitrogen, Carlsbad, CA) using 1 μg total RNA.Following reverse transcription, 1 μl of cDNA was amplifi ed by real-time PCR.Each experiment was performed in duplicate in 96 well plates by using 1x Sybr Green master mix (Bio-Rad, Hercules, CA) in a fi nal volume of 25 μl.Amplifications were performed with the following protocol: 95 o C for 3 minutes followed by 50 cycles of 94 o C for 10 sec and 60 o C for 30 sec.

Microarray, linear amplifi cation, labeling Protocol and feature extraction
Gene expression analysis was conducted using Agilent Human1Av2 arrays (Agilent Technologies, Palo Alto, CA).Total RNA was amplifi ed using the Agilent Low RNA Input Fluorescent Linear Amplifi cation Kit protocol.Starting with 500ng of total RNA, Cy3 or Cy5 labeled cRNA was produced according to manufacturer's protocol.For each two color comparison, 750 ng of each Cy3 and Cy5 labeled cRNAs were mixed and fragmented using the Agilent In Situ Hybridization Kit protocol.Hybridizations were performed for 17 hours in a rotating hybridization oven using the Agilent 60-mer oligo microarray processing protocol.Slides were washed as indicated in this protocol and then scanned with an Agilent Scanner.Data was obtained using the Agilent Feature Extraction software (v7.5), using defaults for all parameters.

Rosetta Resolver, Genespring GX, and Ingenuity pathways analysis.
Images and GEML files, including error and p-values, were exported from the Agilent Feature Extraction software and deposited into Rosetta Resolver (version 4.0, build 4.0.1.0.7.RSPLIT) (Rosetta Biosoftware, Kirkland, WA).The resultant ratio profi les were combined into ratio experiments as described (Dai et al., 2002).Intensity plots were generated for each ratio experiment and genes were considered "signature genes" if the p value was < 0.001.
Data from Agilent Feature Extraction software were also loaded to genespring (version 7.3.1)and fi ltered for markers and fl agged spots.The data were then adjusted for dye swap experiment, normalized and fi ltered based on fold change.
Genes that differentially regulated by dsRNA and RSV infection in primary epithelial cell, BEAS-2B and A549 were analyzed through the use of Ingenuity Pathways Analysis (Ingenuity ® Systems, www.ingenuity.com).
The Functional Analysis identifi ed the biological functions and/or diseases that were most signifi cant to the data set.Genes from the dataset that met the normalized cut off of 1.5 fold and were associated with biological functions and/or diseases in the Ingenuity Pathways Knowledge Base were considered for the analysis.Fischer's exact test was used to calculate a p-value determining the probability that each biological function and/or disease assigned to that data set is due to chance alone.
The Functional Analysis of a network identifi ed the biological functions and/or diseases that were most signifi cant to the genes in the network.The network genes associated with biological functions and/ or diseases in the Ingenuity Pathways Knowledge Base were considered for the analysis.Fischer's exact test was used to calculate a p-value determining the probability that each biological function and/ or disease assigned to that network is due to chance alone.

C y t o k i n e a n d c h e m o k i n e p r o t e i n determination.
T h e a m o u n t o f c y t o k i n e s a n d chemokines present in media and cell extract was determined using a multiplexed cytokines assay system (Bio-plex 200) (Biorad, Hercules, CA).In brief, premixed antibody-coupled beads were reconstituted in assay buffer and loaded onto a pre-wetted 96 well plate.The beads were washed once with the provided wash buffer and then diluted standard or samples were added, mixed by vortexing and incubated at room temperature for 60 min.After washing 3 times to remove unbound proteins, a second specific biotinylated detection antibody was added to the beads, mixed and incubated for an additional 30 min.For fi nal detection, after washing, streptavidinphycoerythrin (streptavidin-PE) was added to the mixture and incubated for 10 min.The beads were then washed, resuspended in assay buffer and analyzed using Bio-

Identifi cation of RSV-induced genes in PHBE cells
To identify Genes that induced by double stranded RNA (Poly I:C), global gene expression profiling were performed Limmon by using labeled cRNA derived from total RNA extracted from mock-infected or RSVinfected (2.5 pfu/cell) PHBE cells after 24 hr.The cRNA was hybridized onto Agilent Human1Av2 arrays.Each comparison (mock vs infected) was done in duplicate with dye swap experiments.Data were then vigorously fi ltered to remove the biases.The genes that have more than 1.5 relative fold inductions compare to untreated control were considered differentially regulated genes.
Our result shows that RSV infection induced up-regulation of 2024 genes in PHBE cells.Genes that were highly induced by RSV included, cytokines and chemokines involved in innate and adaptive immunity, matrix metalloproteinases, cell signaling molecules, TGF family members, cell cycle regulatory molecules, and heat shock proteins (see supplement for complete list).We next compared our data with the previously published RSV-induced genes in other cells.The comparison revealed several novel RSVinduced genes that are listed in Table 2.The novel RSV-induced genes that are shown in Table 1 included ADAM8, ADRAC2C, and IL13RA2.This genes are associated with induction of experimental asthma in mice, and are up-regulated in patients with severe and moderate asthma ( Foley et al., 2007;King et al., 2004;Matsuno et al., 2006)

Identifi cation of dsRNA induced genes in PHBE cells
To determine the dsRNA induced gene profiles, PHBE cell were either left untreated or were treated with dsRNA at 1μg/ml for 2 hr prior to RNA extraction.Gene expression profi ling was carried out as above.The results showed that expression of 159 genes in primary epithelial cells was induced by dsRNA treatment (see supplement for complete list).DsRNA induced genes included infl ammatory and regulatory cytokines and chemokines, signal transduction molecules, cell migration molecules, and apoptosis pathways.Some of the highly induced gene included interferon We also identifi ed several novel dsRNA induced genes in PHBE cells.The list of novel dsRNA induced genes is presented in Table 3.Some of these genes such as BMP2, GBP, and RND3 have been implicated in infl ammation.
BMP2 has been shown to be induced and may play a role in BMP/SMAD signaling during airway inflammation (Rosendahl et al., 2002), and RND3/RhoE was reported to be involved in inflammatory response by regulating IRAK/ERK/NF-κB pathway and cytoskeleton organization (Guasch et al., 2007).

Comparison of differentially expressed genes in dsRNA treated cells and RSV infected cells
Since dsRNA is commonly used as viral mimetic, we compared the gene expression profi les of RSV-infected and dsRNA-treated PHBE cells.Data from Venn diagram comparing RSV-and dsRNA-induced genes (Figure 1) revealed that 75 genes were induced in common.The genes that were commonly up-regulated were listed in Table 4. Several genes that are involved in innate immune responses that were highly induced by both RSV and dsRNA in PHBE cells included apoptosis gene (BIRC3), chemokines (IL-8, CXCL1, CXCL11), and interferon induced protein (IFIT1, IFIT2, IFIT4, ISG15).This data indicate that dsRNA can mimic an innate

Validation of microarray result with real time PCR
To validate the results of the gene array experiments, we used RT-PCR and Bio-Plex assay.Initially the novel RSV-induced genes (ADAM8, IL13RA2, CCNA1 and ADRA2C) novel dsRNA-induced genes (RAND3, BMP2 and SOCS3), and several commonly induced genes were verifi ed by RT-PCR (Figure 2).We also verifi ed 28 genes from the 75 genes that were found to be up-regulated by RSV and dsRNA in PHBE cells and all of the 27 genes that were found to be up-regulated by RSV and dsRNA in PHBE cells and all of the 27 genes that were shown to beup-regulated by microarray analysis were proven to be upregulated by Sybr Green RT-PCR (Table 5).
We next confirmed up-regulation of several cytokines and chemokines by protein determination using Bio-Plex assay, this      Limmon included TNF-α, IP10/CXCL10 and RANTES/ CCL5.Data in Figure 3 showed that cytokine protein expression was signifi cantly increased in infected or treated cells, which validated the data obtained from microarray experiments.

Functional analysis of differentially expressed genes
To investigate the relation and interaction of genes/molecules that were differentially regulated by dsRNA and RSV, we analyzed the up-regulated genes with Ingenuity Pathways Analysis (IPA, Ingenuity Systems).Functional analysis of 75 genes that were commonly induced by RSV and dsRNA in PHBE showed that the up-regulated genes are involved in various diseases and cellular activation pathways.The highest signifi cant diseases and molecular pathways corresponded to these genes are infl ammatory disease Figure 4A.From 75 commonly upregulated genes, 20 genes were determined by IPA to be related to infl ammatory disease (Table 1).
TNF-α has been reported to be induced by dsRNA in vitro and in vivo, and is among the most important cytokines that play a role in viral induced infl ammation.Therefore, particularly important nodes in the network are the genes that are known to interact or induced by TNF-α (Figure 4B).

Discussion
During viral infections, a large repertoire of genes is induced to prevent viral replication and spread.The initial site of infection for respiratory viruses is the airway epithelium; therefore, determining the viral-induced transcriptional events in these cells can shed lights on the initial steps leading to anti viral immune response.The establishment of the anti viral state requires coordinate transcriptional activation of many genes of innate and adaptive immunity.Previous reports have shown that viral infection of various cell types or treatment with poly I:C induced several genes, including type I IFN's.TNF-α RANTES, IL-1, IL-4, IL-6 and IL-8.
DsRNA is recognized as signal to the presence of viral infection.Experimentally, dsRNA in the form of poly I:C has been widely used as a viral mimetic to induce an anti-viral state.Thus far, there have been several reports on the transcriptomic profi ling of virus infections and dsRNA treatment in cells.What has been lacking is a global expression profi ling in PHBE cells, which are the natural target of respiratory viral infections.In addition, there has not been a comparison between RSV infection and dsrNA treatment to assess the contribution of dsRNA-induced events during virus infections in these cells.
Our global transcriptomic analysis identifi ed 2024 genes that were upregulated by RSV and 159 genes that were induced by dsRNA treatment.A comparison of these genes revealed that 75 genes were commonly up-regulated in PHBE cells after treatment with dsRNA or infection with RSV.Interestingly, dsRNA induced similar pattern of infl ammatory cytokines and chemokines to RSV infection.These included type I IFN's, TNF-α, IL-1, IL-6, IL-8, CCL-5, CXCl1, CXCL2, CXCL3, CXCL10, and CXCL11.However there were a number of genes that were specifi cally induced only by dsRNA or RSV (see table 3).Our results also revealed several dsRNA-and RSV-induced genes that have not been reported to be induced in in epithelial cells.These novel genes that are induced by RSV and or dsRNA included ADAMs, ILR13R2, CCNA1, ADRA2C, and BMP2, RND3, SOCS3 and BLMP-1.
ADAMs such as ADAM8 and ADAM10 are disintegrins metalloproteinases that play an important role in releasing soluble CD23 membrane bound precursor (Fourie et al., 2003).The release of the soluble form of CD23 has been reported to increase IgE production (Christie et al., 1997;Kilmon et al., 2001;Saxon et al., 1990), which is the key component of allergic reactions.In as much IgE is critical for allergic reactions, this observation is consistent with previous reports showing that RSV and dsRNA are Limmon associated with induction of allergic immune responses (Akira, 2006;Graham, 2000;von Herrath, 2003).Based on epidemiologic and experimental evidence, expression of ADAM8 has been shown to be up-regulated in a murine model of asthma, and in patients with severe and moderate asthma (Foley et al., 2007;King et al., 2004;Matsuno et al., 2006).Taken together, these data may suggest a novel mechanism for the association of virus infections with induction and exacerbations of allergic diseases such as asthma.
Another interesting gene that has not been reported previously that was upregulated by RSV in ours transcriptomic analysis was the alpha 2 subunit of IL-13 receptor (IL-13RA2).Recent reports have shown that IL-13RA2 maybe decoy receptor for IL-13 which can abrogate a stimulatory signal generated from IL-13 and and IL-13RA1 interaction (Wu, 2003).IL-13 is a pleiotrophic cytokine with potent immune and epithelial cell-regulatory functions such as suppression of infl ammation, epithelial growth, and mucus secretion (Hershey, 2003;Holgate, 2003;Kasaian, 2008).Based on the accumulated data, it is tempting to speculate that up-regulation of the decoy IL13RA2 during RSV infection maybe underlying factor for the immuno-deviation and severe bronchiolitis associated with this infection.Furthermore, IL-13RA has been suggested to play important role in TGF-β1 mediated fibrosis and could be considered for new therapeutic target (Fichtner-Feigl et al., 2007;Fichtner-Feigl et al., 2006).

Limmon
A pathway analysis or our data using IPA revealed that TNF-α was a central player in the virus-and dsRNA-induced events (see fi gure 4B).TNF-α is an early and potent inducer of infl ammation, which can regulate innate and adaptive immune responses, monocyte and lymphocyte migration, and tissue damage.We previously reported that induction of TNF-α during RSV infection or dsRNA treatment was regulated by dsRNAactivated protein kinase (PKR) and p28 MAPK activation (Meusel, 2003).In addition, several of the intracellular antiviral pathways are also up-regulated including, RIG-1, MDA5, TLR3 and MX-1.Predictably, our data suggest that a coordinated activation of several anti-viral pathways may be required for an effective response against respiratory viral infections C o l l e c t i v e l y , o u r d a t a s h o w e d that dsRNA could mimic several of the infl ammatory and regulatory cytokines and chemokines that RSV induces.Our data further provide several novel genes that are induced by dsRNA and RSV, which may be associated with the induction of allergic conditions.Our future studies are aimed at further characterizing the regulation of these novel RSV-and dsRNA induced genes in airways epithelial cells.
The expression of the Novel genes that were up-regulated in primary human lung epithelial cell after RSV infection or dsRNA treatment by microarray analysis was confi rmed by RT-PCR.Y axis represents the relative fold increase of expression.X axis represents genes.
The protein expression level of selected genes were analyzed using bioplex assay.Consistent with the increase in RNA expression, protein expression levels of the selected genes also signifi cantly increase.
Shown are selected Ingenuity Pathway Analysis (IPA) assigned functional categories for the genes induced by dsRNA and RSV.The -log (signifi cance) indicates an increase in confi dence for each category.
. The concentration of each cytokine was then determined by measuring magnitude of the fl uorescently labeled reporter molecules associated with individual target proteins.

Figure 1 .
Figure 1.Venn diagram of differentially expressed genes.Comparison of genes induced by RSV infection and dsRNA treatment in PHBE cells.

Figure 4A .
Figure 4A.Functional analysis of differentially expressed genes.

Figure 4B .
Figure 4B.Functional network analysis of the genes induced by dsRNA and RSV using IPA.Functional network of genes that was commonly induced by RSV infection and dsRNA treatment.The top-scoring network consists of nodes, where each node represents a gene.The shape of the node is indicative of the functional class of the gene product.Multiple connections indicate the biological relationships between the nodes.

Table 1 .
List of PCR primers used to validate microarray data 1

Table 2 .
List of signifi cant diseases and cellular function with related genes identifi ed by IPA .

Table 3 .
Novel RSV induced genes Limmoninfl ammatory subset of genes that are upregulated during RSV infection.

Table 4 .
Novel dsRNA induced genes

Table 5 .
Genes that were commonly induced by DsRNA and RSV in PHBE.
Figure 3. Analysis of protein expression.

Table 6 .
RT-PCR validation of gene expression.