Abstract
Human respiratory syncytial virus (RSV), a member of the Paramyxoviridae family, is the most important viral agent of pediatric respiratory tract disease worldwide. Human airway epithelial cells (AEC) are the primary targets of RSV. AEC are responsible for the secretion of a wide spectrum of cytokines and chemokines that are important mediators of the exacerbated airway inflammation triggered by the host in response to RSV infection. NF-κB is a key transcription factor responsible for the regulation of cytokine and chemokine gene expression and thus represents a potential therapeutic target. In the present study, we sought to delineate the role of RSV-induced reactive oxygen species in the regulation of the signaling pathways leading to NF-κB activation. First, we demonstrate that besides the well-characterized IκBα-dependent pathway, phosphorylation of p65 at Ser536 is an essential event regulating NF-κB activation in response to RSV in A549. Using antioxidant and RNA-interference strategies, we show that a NADPH oxidase 2 (NOX2)-containing NADPH oxidase is an essential regulator of RSV-induced NF-κB activation. Molecular analyses revealed that NOX2 acts upstream of both the phosphorylation of IκBα at Ser32 and of p65 at Ser536 in A549 and normal human bronchial epithelial cells. Similar results were obtained in the context of infection by Sendai virus, thus demonstrating that the newly identified NOX2-dependent NF-κB activation pathway is not restricted to RSV among the Paramyxoviridae. These results illustrate a previously unrecognized dual role of NOX2 in the regulation of NF-κB in response to RSV and Sendai virus in human AEC.
Respiratory syncytial virus (RSV),3 an enveloped, negative-sense ssRNA virus of the Paramyxoviridae family, is the most important viral agent of pediatric respiratory tract disease worldwide. Clinical manifestations following RSV infection range from rhinitis, otitis, or pneumonia to bronchiolitis, an acute lower respiratory tract infection associated with cough and wheezing and substantial morbidity and mortality (reviewed in Refs. 1 and 2). Moreover, RSV infection is associated with long-term complications, such as recurrent wheezing and asthma (reviewed in Ref. 3).
The mechanisms underlying RSV-induced airway diseases and long-term consequences are still largely unknown. However, experimental evidence suggests that an excessive inflammatory response triggered by the host plays a major role in the development of the clinical manifestations of RSV infection (4). Airway epithelial cells (AEC) of large and small airways are the primary targets of RSV infection and replication. In vitro and in vivo studies revealed that AEC are mainly responsible for the secretion of a wide spectrum of cytokines and chemokines which include, but are not restricted to, RANTES, MIP-1α, MCP-1, IL-8, and IFN-γ-inducible protein 10, that have profound immune and inflammatory functions, and thus determine the elimination or progression of the infection and/or inflammation of airway mucosa (4, 5, 6, 7, 8, 9, 10). Chemokine levels in the lung of RSV-infected mice parallel the intensity of lung cellular inflammation (11). Moreover, IL-8, RANTES, and MIP-1α levels in bronchoalveolar lavages of children infected with RSV correlate with the severity of the disease (12, 13). Thus, chemokines are considered major players in the RSV-induced respiratory pathogenesis.
The NF-κB transcription factor plays a pivotal role in inflammatory processes triggered by various stimuli through regulation of the expression of genes encoding numerous proinflammatory cytokines and chemokines (reviewed in Ref. 14). Substantial data support a role of NF-κB in the regulation of cytokine and chemokine genes, such as TNF-α, RANTES, and IL-8, in AEC following RSV infection (4, 5, 6, 7, 8, 9, 10). Thus, NF-κB is considered a preferred target in the development of therapeutic interventions aimed at limiting the inflammatory response. However, the lack of a complete understanding of the pathways leading to its activation is a barrier in achieving this goal.
NF-κB is a ubiquitous family of homo- and heterodimers of Rel proteins, consisting of p65, c-Rel, RelB, p50, and p52. IκB forms a complex with NF-κB dimers, mainly p65/p50 in epithelial cells, that shuttles between the nucleus and the cytoplasm, with a predominant cytoplasmic localization in unstimulated cells (15). The central event of the classical NF-κB activation cascade following exposure to proinflammatory stimuli is the activation of the IκB kinase (IKK) complex, composed of two catalytic subunits IKKα and IKKβ and a regulatory subunit IKKγ (16). In turn, IKK phosphorylates IκBα at Ser32 and Ser36 to promote its polyubiquitination and subsequent proteasome-mediated degradation, thereby allowing the freed NF-κB to translocate to the nucleus, bind to κB consensus sequences, and transactivate target genes (17). A negative feedback loop involves NF-κB-dependent new synthesis of IκB (15). Recently, additional steps in the activation cascade have been identified, which involve phosphorylation and acetylation of the Rel subunits to fine-tune the control of nuclear localization, DNA-binding affinity, coactivator/corepressor association, and transactivation capacity (reviewed in Refs. 18 and 19). Thus far, inducible phosphorylation of c-Rel in the transactivation domain and of p65 at Ser residues 276 or 311 in the Rel homology domain or at Ser residues 468, 529, 535, or 536 in the transactivation domain have been observed in a stimuli- and cell type-dependent manner (reviewed in Refs. 18 and 19). In some conditions, direct phosphorylation of NF-κB subunits was found to define an alternative mechanism of NF-κB activation independently of IκBα degradation. Rather, phosphorylation of NF-κB subunits was found to control NF-κB nuclear accumulation by inducing NF-κB/IκBα complex dissociation (20, 21, 22, 23).
RSV infection induces a persistent activation of NF-κB, which likely leads to excessive NF-κB-mediated inflammatory gene expression (24, 25). Activation of the classical NF-κB activation pathway in RSV-infected AEC is supported by in vitro and in vivo studies that revealed IKK activation, proteolysis of IκB, and p65 DNA binding (24, 25, 26, 27, 28, 29, 30, 31, 32). Interestingly, a temporal dissociation between IκB phosphorylation and NF-κB DNA-binding activity suggested the existence of an IκB-independent regulation induced by RSV (26). RSV-induced phosphorylation of p65 at Ser276 and Ser536 was previously reported, but its physiological significance in NF-κB activity was not evaluated (33).
In the past decade, the concept that subtoxic levels of reactive oxygen species (ROS), such as superoxide anion and hydrogen peroxide, act as cellular switches for signaling cascades leading to regulation of gene expression has emerged. ROS regulate a variety of physiological processes, including cell proliferation, apoptosis, and immune and proinflammatory responses (reviewed in Ref. 34). There is now compelling evidence that ROS participate in the regulation of NF-κB activation in a cell type- and stimuli-specific manner (reviewed in Ref. 35). Several data support a role of ROS in the activation of NF-κB in response to RSV infection in AEC. First, RSV infection was shown to trigger ROS production in AEC (36). Moreover, antioxidants, including butylated hydroxyanisol (BHA) or N-acetylcysteine (NAC), blocked IL-8, RANTES, or MCP-1 gene expression in human AEC (29, 36, 37, 38). Use of NAC suggested that redox-sensitive p65 DNA binding was responsible for the observed IL-8, RANTES, and MCP-1 gene expression (29). However, conflicting results proposed that the AP-1 transcription factor, but not NF-κB, was responsible for the redox regulation of RSV-induced IL-8 gene expression in AEC (39).
ROS-generating enzymes include the mitochondrial respiratory chain, 5-lipoxygenase, xanthine oxidase, and the NADPH oxidases. The most potent source of ROS is the recently discovered family of NADPH oxidase enzymes, which are now known to be functionally expressed in a number of cells, including AEC (reviewed in Ref. 40). In phagocytes, the NADPH oxidase is a multisubunit complex composed of a membrane-bound flavocytochrome b558, consisting of NADPH oxidase 2 (NOX2)/gp91phox and p22phox subunits, as well as of the cytosolic regulatory subunits, p47phox and p67phox. The NOX2-containing NADPH oxidase complex is widely reputed for its role in bacteria killing during phagocytosis (reviewed in Ref. 41). Six functionally distinct homologs of NOX2, namely NOX1, NOX3, NOX4, NOX5, DUOX1, and DUOX2, and homologs of the cytoplasmic subunits, NOXO1 and NOXA1, have been identified (reviewed in Ref. 40). Recent functional data have emerged that suggest the involvement of several of these isoforms in the innate host response to invading microorganisms, including innate immune and proinflammatory responses (reviewed in Ref. 42). Surprisingly, while the role of NADPH oxidase enzymes in virus infections, including infection by RSV, is suggested by the use of antioxidants or cells derived from patients with a genetic defect in one of the phagocytic NADPH oxidase subunits (reviewed in Ref. 42), the identity of the NADPH oxidase responsible for ROS production and its functional significance are barely documented. Interestingly, it was recently shown that the absence of NOX2 led to reduced virus titer, increased Th1 cytokines in the airways, and a reduced inflammatory infiltrate into the lung parenchyma in influenza virus-infected mice (43). Similarly, antioxidant treatment was recently found to improve the final outcome of RSV infection in mice by significantly reducing cytokine and chemokine production and recruitment of inflammatory cells, especially neutrophils, to the lung (44), but the source of ROS in this context remains elusive.
Given the observed potential of targeting ROS production to limit the RSV-induced inflammatory response (44), we examined the molecular mechanisms involved in the redox regulation of cytokine and chemokine gene expression in AEC. This study focuses on the identification of the biological source of ROS and its implication in the regulation of NF-κB. To determine whether this redox regulation was limited to RSV or can be extended to other members of the Paramyxoviridae family, we also evaluated the existence of this redox-dependent signaling mechanism in the context of Sendai virus (SeV) infection, which is widely used as a model of Paramyxoviridae in gene regulation studies. We demonstrate for the first time that a NOX2-containing NADPH oxidase is essential for RSV- and SeV-induced NF-κB activation in human AEC. NOX2 not only plays a role in the regulation of the classical IκBα-dependent NF-κB-signaling pathway, but also regulates p65Ser536 phosphorylation, which appears to be essential for NF-κB activity following RSV and SeV infection. Altogether, this study identifies, for the first time, NOX2 as a specific source of ROS responsible for the oxidant-dependent regulation of NF-κB observed in RSV and SeV infection, thereby contributing to a better comprehension of NF-κB activation pathways. Moreover, these results highlight a novel role of NOX2 in host defense in nonphagocytic cells.
Materials and Methods
Reagents
Diphenyleneiodonium (DPI), DMSO, Tween 20, and BSA were purchased from Sigma-Aldrich. Human thyroid total RNA, human fetal kidney total RNA, human colon total RNA, and human spleen total RNA were purchased from BD Clontech. Oligonucleotides used in PCR (Table I) were purchased from Invitrogen. RNA-interference (RNAi) oligonucleotides (Table II) were purchased from Dharmacon.
Gene . | Sense (S)/Antisense (AS) . | Sequence (5′-3′) . | Thybb . |
---|---|---|---|
NOX1 | S | gtacaaattccagtgtgcagaccac | 62 |
AS | cagactggaatatcggtgacagca | ||
NOX1L | S | tggaggaattaggcaaagtg | 63 |
AS | caaaggaggttttctgtttcag | ||
NOX2 | S | tgttcagctatgaggtggtga | 60 |
AS | tcagattggtggcgttattg | ||
NOX3 | S | tcacaaactggtcgcctatg | 63 |
AS | agggttccttgccagaaaat | ||
NOX4 | S | ctcagcggaatcaatcagctgtg | 62 |
AS | agaggaacacgacaatcagccttag | ||
NOX5 | S | gtgctacatcgatgggccttatg | 65 |
AS | ccccgtgatggagtctttcttct | ||
NOX5L | S | ggaggatgccaggtggctccggt | 64 |
AS | agccccactaccacgtagccc | ||
DUOX1 | S | cgacattgagactgagttga | 62 |
AS | ctggaatgacgttaccttct | ||
DUOX2 | S | aacctaagcagctcacaact | 62 |
AS | cagagagcaatgatggtgat | ||
p22phox | S | cgctggcgtccggcctgatcctca | 60 |
AS | acgcacagccgccagtaggtagat | ||
p47phox | S | tgccaactacgagaagacctc | 62 |
AS | acagaaccaccaaccgctct | ||
p67phox | S | cggacaagaaggactggaag | 62 |
AS | acatgcagccaatgttgaag | ||
NOXO1 | S | ttctctgtgcgctggtcaga | 62 |
AS | tcttgagctgcctgaattcgt | ||
NOXA1 | S | tgggaggtgctacacaatgtg | 60 |
AS | ttggacatggcctcccttag | ||
GAPDH | S | accacagtccatgccatcac | 58 |
AS | tccaccaccctgttgctgta | ||
Actin | S | acaatgagctgctggtggct | 58 |
AS | gatgggcacagtgtgggtga | ||
RANTES | S | taccatgaaggtctccgc | 60 |
AS | gacaaagacgactgctgg | ||
TNF-α | S | cagagggcctgtacctcatc | 55 |
AS | ggaagacccctcccagatag |
Gene . | Sense (S)/Antisense (AS) . | Sequence (5′-3′) . | Thybb . |
---|---|---|---|
NOX1 | S | gtacaaattccagtgtgcagaccac | 62 |
AS | cagactggaatatcggtgacagca | ||
NOX1L | S | tggaggaattaggcaaagtg | 63 |
AS | caaaggaggttttctgtttcag | ||
NOX2 | S | tgttcagctatgaggtggtga | 60 |
AS | tcagattggtggcgttattg | ||
NOX3 | S | tcacaaactggtcgcctatg | 63 |
AS | agggttccttgccagaaaat | ||
NOX4 | S | ctcagcggaatcaatcagctgtg | 62 |
AS | agaggaacacgacaatcagccttag | ||
NOX5 | S | gtgctacatcgatgggccttatg | 65 |
AS | ccccgtgatggagtctttcttct | ||
NOX5L | S | ggaggatgccaggtggctccggt | 64 |
AS | agccccactaccacgtagccc | ||
DUOX1 | S | cgacattgagactgagttga | 62 |
AS | ctggaatgacgttaccttct | ||
DUOX2 | S | aacctaagcagctcacaact | 62 |
AS | cagagagcaatgatggtgat | ||
p22phox | S | cgctggcgtccggcctgatcctca | 60 |
AS | acgcacagccgccagtaggtagat | ||
p47phox | S | tgccaactacgagaagacctc | 62 |
AS | acagaaccaccaaccgctct | ||
p67phox | S | cggacaagaaggactggaag | 62 |
AS | acatgcagccaatgttgaag | ||
NOXO1 | S | ttctctgtgcgctggtcaga | 62 |
AS | tcttgagctgcctgaattcgt | ||
NOXA1 | S | tgggaggtgctacacaatgtg | 60 |
AS | ttggacatggcctcccttag | ||
GAPDH | S | accacagtccatgccatcac | 58 |
AS | tccaccaccctgttgctgta | ||
Actin | S | acaatgagctgctggtggct | 58 |
AS | gatgggcacagtgtgggtga | ||
RANTES | S | taccatgaaggtctccgc | 60 |
AS | gacaaagacgactgctgg | ||
TNF-α | S | cagagggcctgtacctcatc | 55 |
AS | ggaagacccctcccagatag |
Amplification specificity of primers for the corresponding NADPH oxidase was verified using Amplify3X software.
Hybridization temperature (°C).
Target Genes . | Sequences . |
---|---|
NOX1a | |
Target | 5′-nngcacaccuguuuaacuuug-3′ |
Sense oligonucleotide | 5′-gcacaccuguuuaacuuuguu-3′ |
Antisense oligonucleotide | 5′-caaaguuaaacaggugugcuu-3′ |
NOX2(1)a | |
Target | 5′-nngaagacaacuggacaggaa-3′ |
Sense oligonucleotide | 5′-gaagacaacuggacaggaauu-3′ |
Antisense oligonucleotide | 5′-uuccuguccaguugucuucuu-3′ |
NOX2(2)a | |
Target | 5′-nnguggaugccuuccugaaau-3′ |
Sense oligonucleotide | 5′-guggaugccuuccugaaauuu-3′ |
Antisense oligonucleotide | 5′-auuucaggaaggcauccacuu-3′ |
NOX5b | |
Target | 5′-nngguggacuuuaucuggauc-3′ |
Sense oligonucleotide | 5′-gguggacuuuaucuggaucdtdt-3′ |
Antisense oligonucleotide | 5′-gauccagauaaaguccaccdtdt-3′ |
CTRLa | |
Target | 5′-nncauagcguccuugaucaca-3′ |
Sense oligonucleotide | 5′-cauagcguccuugaucacauu-3′ |
Antisense oligonucleotide | 5′-ugugaucaaggacgcuauguu-3′ |
Target Genes . | Sequences . |
---|---|
NOX1a | |
Target | 5′-nngcacaccuguuuaacuuug-3′ |
Sense oligonucleotide | 5′-gcacaccuguuuaacuuuguu-3′ |
Antisense oligonucleotide | 5′-caaaguuaaacaggugugcuu-3′ |
NOX2(1)a | |
Target | 5′-nngaagacaacuggacaggaa-3′ |
Sense oligonucleotide | 5′-gaagacaacuggacaggaauu-3′ |
Antisense oligonucleotide | 5′-uuccuguccaguugucuucuu-3′ |
NOX2(2)a | |
Target | 5′-nnguggaugccuuccugaaau-3′ |
Sense oligonucleotide | 5′-guggaugccuuccugaaauuu-3′ |
Antisense oligonucleotide | 5′-auuucaggaaggcauccacuu-3′ |
NOX5b | |
Target | 5′-nngguggacuuuaucuggauc-3′ |
Sense oligonucleotide | 5′-gguggacuuuaucuggaucdtdt-3′ |
Antisense oligonucleotide | 5′-gauccagauaaaguccaccdtdt-3′ |
CTRLa | |
Target | 5′-nncauagcguccuugaucaca-3′ |
Sense oligonucleotide | 5′-cauagcguccuugaucacauu-3′ |
Antisense oligonucleotide | 5′-ugugaucaaggacgcuauguu-3′ |
Sequences of NOX1-, NOX2(1)-, NOX2(2)-, and CTRL-RNAi were predesigned by Dharmacon.
NOX5-specific RNAi was previously described (81 ).
Plasmids
pCMV-flag-p65 and pCMV-flag-p65S536A plasmids were obtained from Dr. M. Servant (University of Montreal, Montreal, Quebec, Canada). The pRL-null reporter plasmid was obtained from Promega. The P2(2X)TK-pGL3 NF-κB luciferase reporter construct was obtained from Dr. J. Hiscott (McGill University, Montreal, Quebec, Canada). Plasmids used for cytokine copy number determination, generated by cloning of the +169 to +872nt fragment of the TNF-α transcript (NM_000594) and the +65 to +332nt fragment of the RANTES transcript (NM_002985) into the pCR2.1-TOPO vector using EcoRI, were a gift from Dr. D. Lamarre (University of Montreal, Montreal, Quebec, Canada).
Virus
Initial stock of RSV A2 strain was obtained from Advanced Biotechnologies. Amplification was performed in HEp-2 cells (American Type Culture Collection (ATCC)) at a multiplicity of infection (MOI) of 0.1 until a 50% cytopathic effect was observed. Virus was purified on a 30% sucrose cushion after precipitation using polyethyleneglycol (45). Virus titer was determined by methylcellulose plaque assays as previously described (46). SeV Cantrell strain was obtained from Charles River Laboratories.
Cell culture, plasmid transfection, and luciferase assays
HEp-2 cells (ATCC) were cultured in DMEM medium (Invitrogen Life Technologies) supplemented with 10% heat-inactivated FBS (HI-FBS; Invitrogen Life Technologies). A549 cells (ATCC), used as a cell line model of human AEC, were cultured in F-12 nutrient mixture (Ham) medium (Invitrogen Life Technologies) supplemented with 10% HI-FBS and 1% l-glutamine (Invitrogen Life Technologies). Normal human bronchial epithelial cells (NHBE) were obtained from Clonetics, cultured in BEGM medium (Clonetics), and used between passages 2 and 4.
Transfection of A549 cells was performed with the TransIT-LT1 Transfection Reagent (Mirus) according to the manufacturer’s instructions. For luciferase assays, subconfluent cells in 24-well plates were transfected with 50 ng of pRL null reporter (Renilla luciferase, internal control), 100 ng of P2(2X)TK-pGL3 NF-κB reporter construct (firefly luciferase), and the indicated amounts of expression plasmids. Cells were assayed for reporter gene activities after 24 h using the Dual Luciferase Reporter Assay System (Promega) according to the manufacturer’s protocol. Relative luciferase activities represent the ratio of firefly luciferase normalized to Renilla luciferase and are presented as fold over the nonstimulated condition.
Infections
Subconfluent A549 cells were infected with RSV at a MOI of 3 in culture medium containing 2% HI-FBS. SeV infection (40 HAU/106 cells) was conducted for 2 h in serum-free medium and was further cultured for the indicated time in complete medium. Infection of NHBE cells was performed similarly in BEGM medium. In experiments where DPI was used, the reagent or the corresponding vehicle DMSO was added at the indicated concentration for 1 h before infection and maintained at this concentration throughout the infection except in Fig. 4F.
RNAi oligonucleotide transfections
RNAi oligonucleotide (Table II) transfection was performed as previously described (47) using Oligofectamine reagent (Invitrogen) and pursued for 62 h before viral infection as described above. Where RNAi transfection preceded a luciferase reporter gene assay, transfection with reporter plasmids using TransIT-LT1 Transfection Reagent as described above was performed 48 h after transfection of the RNAi.
Immunoblot analysis
Whole cell extracts (WCE) were prepared in Nonidet P-40 lysis buffer as previously described (48) and subjected to SDS-PAGE electrophoresis and immunoblot analysis as performed in Ref. 49 using the primary Abs anti-IκBα-phospho-Ser32 (1/1,000; Cell Signaling), anti-p65-phospho-Ser536 (1/1,500; Cell Signaling), anti-actin (1/10,000; Chemicon International), anti-IκBα (1/1,000; Cell Signaling), anti-p65 (1/400; Santa Cruz Biotechnology), anti-Flag M2 (1 μg/ml; Sigma-Aldrich), anti-SeV (1/10,000; obtained from Dr. J. Hiscott, McGill University, Montreal, Quebec, Canada), or anti-RSV (1/2,000; Chemicon International) diluted in PBS containing 0.5% Tween 20 and either 5% nonfat dry milk or 5% BSA for phosphospecific Abs. Immunoreactive bands were visualized by ECL using the Western Lightning Chemiluminescence Reagent Plus (PerkinElmer Life Sciences). In between IκBα-phospho-Ser32 and IκBα Abs, as well as p65-phospho-Ser536 and p65 Abs, the membrane was stripped in 0.2% SDS, 62.5 mM Tris-HCl (pH 6.8), 0.1 mM 2-ME for 20 min at 50°C, washed three times in PBS, and blocked in blocking solution.
RNA extraction, RT-PCR, and real-time PCR
Total RNA was extracted from A549 using the RNAqueous-4PCR Isolation kit (Ambion) and treated to remove genomic DNA using either the DNase1 treatment included in the RNAqueous-4PCR Isolation kit or the reagent included in the QuantiTect Reverse Transcription kit (Qiagen). Total RNA (1 μg) was subjected to reverse transcription using 0.5 μg of oligo(dT)12–18 primers (Invitrogen) and 200 U of Moloney murine leukemia virus reverse transcriptase (Invitrogen) or using the QuantiTect Reverse Transcription kit. PCR amplifications were performed using the QuantiTect SYBR Green kit (Qiagen) in the presence of 2 μM specific primers, except for NOX2 amplification, which was performed using 1 μM specific primers, and β-actin, TNF-α, and RANTES amplifications, which were performed using 0.4 μM specific primers. MgCl2 concentration in all assays was 2.5 mM, except for NOX1L amplification, which required 5 mM MgCl2. Sequences and annealing temperatures for gene-specific primers are listed in Table I. The absence of genomic DNA contamination was verified with each of the primer sets by PCR in a reaction without reverse transcriptase. When analyzed on agarose gel, PCR were performed in a T Gradient Cycler (Biometra) for the indicated number of cycles previously established to be in the linear detection range for each gene. For real-time PCR analyses, detection was performed on a Rotor-Gene 3000 Real-Time Thermal Cycler (Corbett Research). For TNF-α and RANTES gene expression, standard curves of absolute quantification expressed as copy number and PCR efficiencies were obtained using serial dilutions of pCR2.1-TOPO-TNF-α and pCR2.1-TOPO-RANTES plasmids. For β-actin, NOX1L, NOX2, and NOX5 gene expression, standard curves and PCR efficiencies were obtained using serial dilutions of cDNA prepared from cells used as a positive control as described in Fig. 5. TNF-α and RANTES data are presented as absolute copy numbers normalized to β-actin used as a reference gene. NOX1L expression was normalized to β-actin and is presented as relative fold expression of NOX1L from the NOX1-RNAi sample vs the control-RNAi sample. Relative NOX2 and NOX5 fold expression values were determined applying the ΔΔ cycle threshold (Ct) method (50).
Statistical analyses
All analyses were performed using SigmaStat 3.5 software.
Results
Phosphorylation of p65 at Ser536 is a key event in the activation of NF-κB triggered by RSV and SeV infection in A549
Activation of the classical NF-κB pathway, involving IKK-mediated phosphorylation of IκBα and its subsequent degradation, is well documented in the context of RSV and SeV infections (49, 51). However, the importance of p65 phosphorylation is still barely documented in these contexts. Thus, we first re-evaluated the activation of NF-κB by monitoring not only IκBα but also p65 phosphorylation in RSV- and SeV-infected A549. IκBα phosphorylation at Ser32 and p65 phosphorylation at Ser536 were analyzed by immunoblot using phosphospecific Abs. As shown in Fig. 1, IκBαSer32 and p65Ser536 phosphorylations were both detected with similar kinetic patterns during RSV (Fig. 1, A and C) or SeV (Fig. 1, B and D) infection. P65 phosphorylation at Ser276 was neither detected in RSV nor SeV infection (data not shown). Thus, our data reveal that RSV and SeV infection of A549 trigger activation of NF-κB at two levels, namely IκBα and p65Ser536 phosphorylation.
To determine whether p65Ser536 phosphorylation is essential for NF-κB activity in the context of RSV and SeV infection, the effect of Ser536 mutation into Ala (p65S536A), which abrogates phosphorylation at this particular residue, was analyzed by luciferase reporter gene assay. A549 were cotransfected with the NF-κB-responsive P2(2X)TK-pGL3 luciferase reporter gene and the pRL-null Renilla internal control plasmids, together with the empty vector, or vector encoding flag-tagged-p65 or -p65S536A and were either left untreated or infected with RSV or SeV. As shown in Fig. 2, expression of p65 in noninfected cells resulted in a 53-fold induction of the NF-κB reporter activity, whereas p65S536A only exhibited a minimal effect with a 5-fold induction. RSV and SeV infection of A549 cells resulted in a 14- and 13-fold stimulation of the reporter activity, respectively. These inductions were strongly enhanced by the ectopic expression of p65 to 191- and 179-fold, respectively. In contrast, RSV and SeV stimulation of the promoter in the presence of p65S536A only reached 43- and 35-fold, respectively. Together, these results demonstrate that phosphorylation of p65 at Ser536 during RSV and SeV infection in A549 cells is essential to trigger full activation of NF-κB transactivation potential.
DPI inhibits RSV- and SeV-induced NF-κB activation
In a first attempt to determine whether NF-κB activation in the context of infection by RSV and SeV was dependent on ROS production by a NADPH oxidase, we analyzed the effect of DPI, an inhibitor of flavoproteins widely used to target NADPH oxidases. The expression profile of NF-κB target genes, TNF-α and RANTES, was monitored by real-time PCR in A549 cells either left untreated or infected with RSV or SeV for 6 h in the absence or presence of 30 μM DPI. As shown in Fig. 3, A and B, TNF-α mRNA levels were strongly induced following RSV and SeV infection of A549 cells. However, pretreatment with DPI reduced TNF-α expression by 66 and 78% in RSV and SeV infection, respectively. Analysis of RANTES gene expression gave similar results, with reduction of RSV and SeV induced RANTES mRNA levels by 4 and 1.6 log, respectively (Fig. 3, C and D).
As p65Ser536 phosphorylation appears to be essential for NF-κB activation (Fig. 2), we next assessed whether abrogation of NF-κB-dependent gene expression by DPI correlated with inhibition of p65Ser536 phosphorylation. A549 cells were infected with RSV for 8 h in the presence of increasing doses of DPI. As shown in Fig. 4,A, RSV-induced p65Ser536 phosphorylation was inhibited by DPI in a dose-dependent manner reaching 72.6 ± 8.2% inhibition at 30 μM DPI. Kinetic analysis of p65Ser536 phosphorylation during RSV infection in the absence or presence of 30 μM DPI confirmed the inhibition of p65Ser536 phosphorylation by DPI over time (Fig. 4,B). Similarly, DPI inhibited SeV-induced p65Ser536 phosphorylation in a dose-dependent manner and over the course of infection (Fig. 4, C and D). Importantly, the effect of DPI is not attributable to an effect on viruses. Indeed, neither RSV nor SeV replication were affected by DPI treatment as demonstrated by the detection of viral protein level by immunoblot (Fig. 4,E). Moreover, identical effects were observed when DPI was present in the medium during the infection (protocol I; Fig. 4,F) or when DPI was preincubated with the cells and washed off before infection (protocol II; Fig. 4 F). Altogether, these results demonstrate the importance of ROS, most likely from a NADPH oxidase origin, in NF-κB activation through p65Ser536 phosphorylation in RSV- and SeV-infected A549.
Various subunits of NADPH oxidases are expressed in RSV- and SeV-infected A549
Several NADPH oxidase subunits might be responsible for the observed DPI-mediated inhibition of p65 phosphorylation at Ser536 during RSV and SeV infection. As expression of NADPH oxidase subunits appears to be cell-type specific, we monitored which of the NOX1–5 and DUOX1–2 catalytic subunits, p22phox and regulatory subunits p47phox, p67phox, NOXO1, and NOXA1 were expressed in A549 during RSV and SeV infection. For this purpose, A549 were infected with RSV and SeV and NOX mRNA expression was analyzed by RT-PCR (Fig. 5,A) using primers specific for each subunit (Table I). NOX1, NOX2, and NOX5 mRNA were detected in unstimulated A549 and their expression remained steady during virus infection. Using a specific set of primers, the long isoform of NOX1, NOX1L (52, 53), was found to follow the same pattern of expression as total NOX1. In contrast, the long isoform of NOX5, NOX5L, was not detected in any conditions, thus suggesting that the NOX5 mRNA represent the NOX5S isoform (54). NOX3 and NOX4 mRNA were neither detected in unstimulated nor in virus-infected A549. It is noteworthy that DUOX2 mRNA, but not DUOX1 mRNA, was expressed at the basal level and its expression was stimulated by RSV and SeV infection. p22phox mRNA expression was detected at an equal level in all conditions. As shown in Fig. 5 B, NOXO1, NOXA1, p47phox, and p67phox regulatory subunits mRNA were all expressed at the basal level and p67phox mRNA was strongly inducible following RSV and SeV infections. In conclusion, these results demonstrate that NOX1L, NOX2, NOX5S, and DUOX2 catalytic subunits of NADPH oxidase family members, p22phox, and regulatory subunits NOXO1, NOXA1, p47phox, and p67phox are expressed in A549 during RSV and SeV infection.
RNA interfering with NOX2 expression inhibits NF-κB transactivation potential in RSV and SeV infection
Based on the observation that several members of the NOX/DUOX family of enzymes, namely NOX1L, NOX2, and NOX5S, are expressed at the basal level in A549, we hypothesized that one of these isoforms might be responsible for the observed redox-dependent regulation of NF-κB following RSV and SeV infection (Figs. 3 and 4). To verify our hypothesis, specific interfering RNA (RNAi) oligonucleotides (Table II) were used in A549 to down-regulate the expression of NOX1, NOX2, and NOX5. As shown in Fig. 6,A, NOX1-, NOX2- (NOX2(1)), and NOX5-RNAi efficiently inhibited the corresponding NOX expression by 82, 64, and 90%, respectively, as evaluated by real-time PCR. The effect of interference with the expression of each isoform on RSV- and SeV-induced NF-κB transcriptional activity was evaluated using a luciferase reporter gene assay. RNAi-transfected A549 were further transfected with the P2(2X)TK-pGL3 luciferase reporter as well as the pRL-null Renilla luciferase (internal reference) constructs and were either left untreated or infected with RSV or SeV. As shown in Fig. 6,B, interference with NOX1 and NOX5 expression did not affect endogenous NF-κB ability to stimulate gene transcription. However, as shown in Fig. 6, C and D, when NOX2 expression was reduced by NOX2(1)- RNAi, RSV-, and SeV-induced NF-κB transactivation potential was significantly inhibited. RSV-induced NF-κB activity in a time-dependent manner with a 26- ± 1.37-fold stimulation at 16 h postinfection (hpi), while in the presence of NOX2(1)-RNAi, NF-κB stimulation remained lower over time reaching a 9.67- ± 1.62-fold at 16 hpi (Fig. 6,C). Similar results were observed in SeV infection (Fig. 6 D). Indeed, stimulation of NF-κB activity was of 11.94- ± 1.16-fold in CTRL-RNAi-transfected A549 at 16 hpi, whereas it was decreased to 6.68- ± 0.3-fold in NOX2(1)-RNAi transfected cells. Taken together, these results demonstrate that NOX2 is an essential component of the signaling pathway triggering NF-κB activation following RSV and SeV infection in A549.
Interference with NOX2 expression inhibits RSV- and SeV-induced NF-κB activation through diminution of IκBαSer32 and p65Ser536 phosphorylation in A549
Inhibition of NF-κB transactivation potential (Fig. 6, C and D) in the presence of NOX2-RNAi suggested that upstream events leading to NF-κB activation might be altered by the absence of NOX2 expression. Therefore, RNAi-mediated NOX2 inhibition was further used to investigate the role of NOX2 in NF-κB activation following RSV and SeV infection. A549 were transfected with CTRL- or NOX2-specific RNAi, NOX2(1), and were then left untreated or infected with RSV or SeV for various times. IκBα phosphorylation at Ser32 and p65 phosphorylation at Ser536 were evaluated by phosphospecific immunoblot and the ratio of phospho-IκBα vs total IκBα and phospho-p65 vs total p65 was determined by densitometric quantification (Fig. 7). RSV- (Fig. 7,A) and SeV-induced (Fig. 7,B) IκBαSer32 and p65Ser536 phosphorylations were significantly inhibited by reduced NOX2 expression. An identical experiment with a second NOX2-specific RNAi, NOX2(2) (Fig. 8,A), yielded similar results (Fig. 8,B). To ascertain that the observed effects on NF-κB activation were not resulting from an effect of the RNAi on virus replication, viral protein expression in various conditions was evaluated by immunoblot using anti-RSV (Figs. 7,A and 8B) and anti-SeV (Fig. 7 B) Abs. Neither NOX2(1)- nor NOX2(2)-RNAi interfered with virus replication. NOX1- and NOX5-RNAi did not alter p65Ser536 phosphorylation (data not shown), thus demonstrating the specific role of NOX2 in this pathway. In conclusion, our results demonstrate that NOX2 plays an essential role in RSV- and SeV-induced NF-κB regulation at two critical steps, namely IκBαSer32 and p65Ser536 phosphorylation in A549.
Dual role of NOX2 in RSV- and SeV-induced NF-κB regulation in NHBE
Although A549 cells are widely considered a cell line model of human AEC, we verified the existence of the newly identified NOX2-dependent pathway of NF-κB regulation in primary NHBE cells. First, NOX2 mRNA expression was analyzed by RT-PCR in NHBE during RSV and SeV infection. Similarly to the observation made in A549, NOX2 mRNA was detected in unstimulated NHBE and its expression remained steady during RSV and SeV infection (Fig. 9,A). More importantly, interference with NOX2 expression using the NOX2(1) RNAi oligonucleotide (Fig. 9,B) resulted in significant inhibition of both RSV- and SeV-induced IκBαSer32 and p65Ser536 phosphorylation as evaluated by phosphospecific immunoblots and phospho-IκBα vs total IκBα and phospho-p65 vs total p65 ratios (Fig. 9, C and D). These results highlight the important dual role of NOX2 in the regulation of NF-κB during Paramyxoviridae infections in the context of normal AEC.
Discussion
NF-κB plays an essential role in the RSV-induced exacerbated inflammatory response through the regulation of cytokines and chemokines expression and thus constitutes an interesting therapeutic target. However, the molecular mechanisms controlling the persistent RSV-induced NF-κB activation in AEC are still far from being deciphered. Although it has long been a matter of debate, cell-type- and stimuli-specific redox-regulation of NF-κB is now well documented (reviewed in Ref. 35). Previous studies have shed light on the possible ROS-dependent regulation of NF-κB following RSV infection in AEC. First, RSV triggers ROS production in AEC (36). Second, RSV-induced expression of genes encoding the chemokines IL-8, RANTES, and MCP-1, which are known NF-κB-target genes, was found to be dependent on a redox-sensitive pathway(s) in A549 (29, 36, 37, 38). In addition, ROS were found to be involved in the regulation of IκBα degradation, nuclear translocation, and in vitro DNA-binding activity of p65 in RSV-infected AEC (24, 29). Although these studies support a role of ROS in RSV-mediated NF-κB activation, they did not address two major questions: 1) the biological source of ROS and 2) the steps in the NF-κB-signaling pathway that are redox sensitive. Here, we provide the first evidence that RSV triggers NF-κB activation through a NOX2-containing NADPH oxidase in A549 and primary NHBE. This pathway is essential for IκBαSer32 and p65Ser536 phosphorylation events. Furthermore, we demonstrate that this pathway is not restricted to RSV among Paramyxoviridae as similar results were obtained with SeV. Thus, our results highlight a novel signaling pathway mediating NF-κB activation in AEC infected with RSV or SeV that involves a NOX2-containing NADPH oxidase as a central regulator.
Attempts to identify the origin of RSV-induced ROS in AEC were previously limited to a recent study that suggested the involvement of an uncharacterized NADPH oxidase based on the observation that RSV-induced activation of the IFN regulatory factor (IRF)-3 transcription factor was inhibitable by DPI (55). Although the effect of DPI is widely considered as an indicator of a potential role of a NADPH oxidase, other ROS-generating flavoproteins, including the mitochondrial respiratory chain, could explain the effects observed with DPI (56). Therefore, to definitively conclude on the involvement of NADPH-oxidase activity, we used selective RNAi against NOX1, NOX2, and NOX5, and provide the first direct proof that a NOX2-containing enzyme specifically participates in the redox-sensitive regulation of RSV- and SeV-induced NF-κB signaling in A549 and NHBE.
Our study shows that transcripts of NOX1, NOX2, NOX5, and DUOX2 together with the small transmembrane subunit p22phox and the various regulatory subunits, p47phox, p67phox, NOXO1, and NOXA1 are expressed in A549 (Fig. 5). Previous studies describing the expression of NOX/DUOX isoforms in various AEC cell lines, including the Calu-3 submucosal gland cell line, the hTE human tracheal surface epithelial cell line, and the NCI-H292 human pulmonary cell line, highlighted significant discrepancies (57, 58, 59). It is noteworthy that NOX4 was not detected in our study (Fig. 5), while its basal expression and inducibility following exposure to diesel exhaust particules in A549 was recently reported (60). Expression of alternative NOX4 splice variants was also documented (61). Comparison of our results with the one described by Amara et al. (60) should be made with caution. Indeed, in their study NOX4 mRNA was only expressed as fold over ubiquitin and not as absolute values and thus the amount of NOX4 might in fact be very low. Here, the use of primers expected to amplify the various potential variants failed to detect NOX4 mRNA in uninfected and infected A549 (Fig. 5). Similar results were obtained using a second set of primers (data not shown). These variations in NOX/DUOX detection in cell lines is of particular significance as these cell lines are frequently used as a model of AEC. Here, we focused our functional studies on the role of NOX isoforms and demonstrated a selective role of NOX2. To clarify the role of NOX2, it was of importance to verify the expression status of NOX2 in primary AEC. Here, we demonstrate that NOX2 transcripts are expressed in NHBE cells (Fig. 9), thus confirming the relevance of the A549 model. Moreover, RNAi experiments allowed us to confirm the role of NOX2 in NF-κB regulation during the course of RSV and SeV infections in NHBE (Fig. 9).
With the recent description of the NOX/DUOX family, data are emerging that link NADPH oxidase enzymes activity to NF-κB activation and downstream proinflammatory response. In the context of infectious diseases, different NOX isoforms, including NOX4 and NOX2, were found to act downstream of TLR to trigger proinflammatory signals (reviewed in Ref. 42). NOX isoforms were also identified as inflammatory regulators in a few other contexts. Following stimulation of MCF-7 cells by IL-1β, NOX2-dependent ROS production in the endosome was shown to be essential for recruitment of the TNFR-associated factor 6 adaptor to trigger an IKK-dependent NF-κB activation (62). In vascular smooth muscle cells, NOX4 was shown to orchestrate C-reactive protein-mediated inflammatory activities through regulation of AP-1 and NF-κB (63). Taken together, these data and our study reveal that NOX isoforms are important inflammatory signaling molecules in nonphagocytic cells and that NOX constitute potential anti-inflammatory therapeutic targets.
The majority of studies that report ROS-dependent regulation of NF-κB support a role of ROS in the activation of NF-κB (reviewed in Ref. 35). The use of H2O2 as a stimulus revealed that IKK is under redox control in epithelial cells. H2O2 was found to trigger IKKα and IKKβ phosphorylation and activation in HeLa cells. More recently, IKK activation in response to H2O2 in HeLa cells was shown to be dependent on a Src/protein kinase Cδ/protein kinase D signaling cascade (64, 65). Conflicting results came from studies performed in C10-immortalized alveolar type II cells, where H2O2 inhibited TNF-α- or LPS-mediated IKK activation (66). However, our data showing NOX2-dependent IκBα phosphorylation at Ser32 support a ROS-dependent activation of IKK upon RSV or SeV infection in AEC. Although this discrepancy may only reflect stimulus specificity of the response, one also needs to consider that the use of H2O2 does not necessarily reflect physiological concentrations of ROS triggered by physiological stimuli.
The molecular mechanisms regulating NF-κB activation following RSV infection of AEC have not yet been completely elucidated. A temporal dissociation between NF-κB DNA binding and IκBα degradation (26) suggests that an IκBα-independent mechanism might be involved in early activation of NF-κB. Activation of the noncanonical pathway of NF-κB activation, which involves cleavage of p100 to form the p52 subunit, was recently described following RSV infection (67) and it was proposed that it might be responsible for early activation of NF-κB. Although it might indeed be relevant for early cytokine genes transcription, it does not explain the early p65 DNA binding (26). Our data confirm previous reports (31) that degradation of IκBα is only observed between 12 and 24 hpi following RSV infection. We now show that p65 Ser536 phosphorylation occurs in vivo as early as 3 hpi (Figs. 1 and 7) and that it is essential for RSV- and SeV-induced p65 activation (Fig. 2). The Ser536 residue of p65 is evolutionary conserved and is phosphorylated in response to numerous stimuli, including DNA-damaging agents, T cell costimulation, angiotensin (Ang) II and LPS (21, 68, 69, 70). Substantial literature supports a function of p65Ser536 phosphorylation in the regulation of p65 affinity to IκBα rather than of its transcriptional activity (reviewed in Ref. 18). A study recently suggested that, in fact, Ser536-phosphorylated p65 does not interact at all with cytosolic IκBα and regulates a distinct set of target genes (23). Thus, this pathway constitutes a likely candidate to explain the early NF-κB activation before IκBα degradation triggered by RSV in AEC. Although redox regulation of this pathway was previously suggested by the observed inhibition of TNF-α-induced p65Ser536 phosphorylation by NAC in endothelial cells (71), our data constitute the first identification of the biological source of ROS responsible for its regulation.
Characterization of the signaling molecules targeted by ROS is still barely documented. However, the oxidation of thiol groups in the catalytic sites of kinases and phosphatases appears to be an important posttranslational modification that affects the function of signaling proteins (72). It is thus tempting to hypothesize that NOX2-derived ROS regulate the activation of kinases/phosphatases acting upstream of IκBαSer32 and p65Ser536 in RSV and SeV infections. Oxidation of IKKβ at Cys179 was found involved in negative regulation of NF-κB following anti-inflammatory stimuli (73). Thus, it is more likely that upstream regulators might be targeted by thiol modification to trigger its activation. However, kinases and phosphatases acting upstream of IKK in the context of RSV and SeV infection remain to be identified. Similarly, the signaling pathway leading to p65Ser536 phosphorylation during RSV and SeV infections is still unknown. Five kinases were previously found to phosphorylate p65 at Ser536. IKKα/β phosphorylate p65Ser536 in vivo in response to various stimuli, including LPS, Ang II, TNF-α, or lymphotoxin β receptor signaling (68, 74, 75, 76). Additionally, the IKK homologs TBK1 and IKKε, that have recently been found to be involved in the phosphorylation of the IRF-3 and IRF-7 transcription factors following virus infections (47, 77), also phosphorylate p65 at Ser536 in response to TNF-α or IL-1 (78, 79). Finally, the RSK-1 kinase was also identified as a p65Ser536 kinase following Ang II stimulation or in response to p53 (21, 80). Which of these kinases is involved in RSV- and SeV-induced p65Ser536 phosphorylation and their redox regulation is currently under investigation.
The importance of the redox regulation of cytokines and chemokines production in RSV pathology was recently illustrated by an in vivo study in a mouse model. Indeed, treatment of mice with BHA significantly reduced RSV-induced cytokines and chemokines production and recruitment of inflammatory cells, especially neutrophils, to the lung (44). However, this study used a general antioxidant and did not target the specific source of ROS responsible for NF-κB regulation. Our data now demonstrate that NOX2 represents a potential target to limit the excessive inflammatory response triggered by RSV. During the course of our work, it was shown that absence of NOX2 led to a reduced virus titer, increased Th1 cytokines in the airways and a reduced inflammatory infiltrate into the lung parenchyma in influenza-infected mice (43). These results are in line with the concept raised by our study that inhibiting NOX2 might be a means of controlling the inflammatory response triggered by Paramyxoviridae infection.
Acknowledgments
We thank Loubna Jouan for her help with the real-time PCR experiments, and Ian-Gaël Rodrigue-Gervais and Robert Boileau for their help with statistical analyses. We are also very grateful to Dr. J. Hiscott (McGill University, Montreal, Quebec, Canada), Dr. M. Servant (University of Montreal, Montreal, Quebec, Canada), and Dr. D. Lamarre (University of Montreal, Montreal, Quebec, Canada) for providing reagents used in this study.
Disclosures
The authors have no financial conflict of interest.
Footnotes
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by grants from the Canadian Institutes of Health Research and Fonds de la Recherche en Santé du Québec (to N.G.). K.F. and A.S.-F. were recipients of a studentship from Institut National de la Santé et de la Recherche Médicale Unité 743. N.G. was a recipient of a Tier II Canada Research Chair.
Abbreviations used in this paper: RSV, respiratory syncytial virus; AEC, airway epithelial cell; IKK, IκB kinase; ROS, reactive oxygen species; BHA, butylated hydroxyanisol; NAC, N-acetylcysteine; SeV, Sendai virus; DPI, diphenyleneiodonium; RNAi, RNA interference; MOI, multiplicity of infection; NHBE, normal human bronchial epithelial cell; WCE, whole cell extract; Ct, cycle threshold; hpi, hour postinfection; Ang, angiotensin; NOX2, NADPH oxidase 2; HI-FBS, heat-inactivated FBS.