|Home | About | Journals | Submit | Contact Us | Français|
Interferon (IFN) therapy in humans often causes flu-like symptoms by an unknown mechanism. Poly ICLC is a synthetic dsRNA and a Toll-like receptor 3 (TLR3) agonist with a strong IFN-inducing ability. In this work, we analyzed the effect of poly ICLC on pulmonary responses to influenza and respiratory syncytial virus (RSV) infections in the cotton rat (Sigmodon hispidus) model. Viral replication, pulmonary inflammation, and expression of IFN, TLR, and chemokines were monitored and compared. Antiviral effect of poly ICLC against influenza virus and RSV was best achieved at high poly ICLC concentrations that, in the absence of virus infection, induced a strong IFN response. The antiviral doses of poly ICLC, however, also increased lung inflammation, an unexpected finding because of the reported poly ICLC safety in BALB/c mice. Similarly, in contrast to murine model, pathology of RSV infection was increased in cotton rats treated with poly ICLC. Augmented lung inflammation was accompanied by an earlier induction of IFN and TLR responses and a stronger chemokine expression. Overall, these findings indicate significant association between antiviral IFN action and pulmonary inflammation and highlight important animal model-specific variations in the potential of IFN to cause pathology.
Respiratory viral diseases caused by influenza and respiratory syncytial virus (RSV) are multifactorial disorders in which both virus-induced damage and host response to infection contribute to disease pathogenesis. The effective treatment against influenza and RSV is difficult to find, as antiviral therapy often inhibits viral replication, but does not cure pulmonary inflammation (Prince and others 2000; Ottolini and others 2003; Boukhvalova and others 2007b; Zheng and others 2008). When administered prior to infection, however, antivirals abolish both virally induced damage and pulmonary pathology (Prince and others 1985). These findings together suggest that early events triggered after respiratory viral infection define ensuing inflammatory response, even when viral replication is inhibited within days of infection.
Influenza and RSV represent major health threats and are associated with substantial mortality. The discovery of interferons (IFN) and demonstration of their broad-spectrum antiviral action provided hope for these and other viral infections. However, subsequent discovery of toxicity associated with IFN therapy in humans has limited its use therapeutically to a few select diseases (Samuel 2001). The relationship between antiviral IFN response and its toxic effect is still not understood. Viruses induce IFN shortly after infection as a part of broader cytokine response involved in inflammation. Although lung inflammation is first and foremost a protective response to infection or injury, it has the potential to turn harmful if excessive (de Jong and others 2006).
Antiviral IFN response is initiated by the host upon detection of pathogen-associated molecular patterns, such as dsRNA (produced in the course of viral replication), highly structured ssRNA, and viral glycoproteins (Akira and others 2006). Viral glycoproteins are recognized by members of a family of Toll-like receptors (TLR), including TLR2 and TLR4 (Haynes and others 2001; Murawski and others 2009). Cytoplasmic dsRNAs are detected by 2 helicases, RIG-I and MDA5 (Akira and others 2006). Even though it has been difficult to obtain direct evidence that dsRNA is produced in cells infected with negative-strand RNA viruses (Weber and others 2006), RIG-I plays a major role in IFN induction in these cells (Akira and others 2006; Kato and others 2006; Bhoj and others 2008). dsRNA is also recognized by TLR3, implicated in biology of influenza and RSV infections (Guillot and others 2005; Rudd and others 2005). Type I IFNs (IFN-α/β) are produced early following viral infection by a variety of cells and mediate antiviral activity by limiting viral replication and spread, as well as by modulating innate and adaptive immune responses (Smith and others 2005). Type II IFN, IFN-γ, is produced primarily by T cells and NK cells and mediates its antiviral effect by enhancing function of macrophages, NK cells, neutrophils, and other components of the immune system. Synthetic dsRNA poly IC (polyriboinosinic–polyribocytidylic acid) was shown to be among the strongest inducers of IFN (De Clercq 1981; Alexopoulou and others 2001). Although effective in rodents, poly IC, however, turned out to be a poor inducer of IFN in humans and non-human primates. This problem was overcome with the discovery of poly ICLC, poly IC stabilized with poly-l-lysine carboxymethyl cellulose (Levy and others 1975).
Side effects of IFN therapy in humans are often collectively called “flu-like syndrome,” and include fever, fatigue, myalgias, nasal congestion, and cough (Haeuber 1989). dsRNA extracted from influenza-infected lungs or synthetic dsRNA can also induce a complex of flu-like symptoms (Majde and others 1991; Fang and others 1999). It is possible that the IFN response induced by viral dsRNA plays a crucial role in defining lung injury and systemic manifestations of respiratory viral infections, but this connection has not been established. Influenza, RSV, synthetic dsRNA, and IFN all can modulate TLR expression in vitro in several cell types (Miettinen and others 2001; Guillot and others 2005; Ritter and others 2005; Groskreutz and others 2006). Changes in pulmonary TLR expression have the potential to significantly affect lung microenvironment and the associated inflammatory response (Didierlaurent and others 2007) and may serve as a link between IFN response and pulmonary inflammation.
In this work, we have compared and contrasted IFN-inducing, antiviral, and proinflammatory effect of poly ICLC on influenza and RSV infection in the cotton rat, Sigmodon hispidus, model. The cotton rat is considered to be a reliable small animal model of RSV pathogenesis and has recently been developed as a model of influenza infection (Ottolini and others 2005). Our results indicate that antiviral effect of poly ICLC against influenza and RSV infection is accompanied by a rapid IFN activation, altered expression of pulmonary TLR, and extensive pulmonary inflammation. To the best of our knowledge, this work represents the most comprehensive analysis to date on the association of IFN-TLR expression axis with pulmonary inflammatory response.
Inbred S. hispidus cotton rats were obtained from a colony maintained at Virion Systems, Inc. Animals were housed in large polycarbonate cages with a bedding of paper mill by-products and were fed a standard diet of rodent chow and water. The colony was monitored for antibodies to adventitious respiratory viruses and other common rodent pathogens and no such antibodies were found. All studies were conducted under applicable laws and guidelines and after approval from the Virion Systems, Inc. Institutional Animal Care and Use Committee (IACUC).
Stock of influenza A/Wuhan/359/95 (H3N2 subtype) virus containing 108 TCID50/mL was prepared by Novavax, Inc. (Rockville, MD) using MDCK cells. Influenza titers in lung tissue were determined by TCID50 assays on monolayers of MDCK cells as previously described (Ottolini and others 2005; limit of detection 2.6 TCID50/g). The prototype A/Long and A/A2 strains of RSV were obtained from the American Type Culture Collection (VA). Viruses were propagated in HEp-2 cells and serially plaque-purified to reduce defective-interfering particles (Gupta and others 1996). A single pool of each RSV A/Long (107.6 pfu/mL) and RSV A/A2 (107 pfu/mL) was used for the studies described herein. RSV viral titers in the lungs of infected animals were determined by plaque assay (Prince and others 1978) (limit of detection 2 log10 pfu/g).
Poly ICLC solution (Hiltonol, 2 mg/mL) was kindly provided by Dr. Andre Salazar (Oncovir, Inc.) (Salazar and others 1996).
For the dose studies of poly ICLC, groups of 3 cotton rats were treated intranasally (i.n.) with 100 μL poly ICLC (1, 10, or 40 μg/animal) or with 100 μL saline and sacrificed 1, 2, 5, 8, or 12 days later. Lungs were extracted and processed for mRNA expression and histopathology analysis. For the dose-dependency studies of antiviral effect of poly ICLC, groups of 3 cotton rats were treated i.n. with 1, 10, or 40 μg of poly ICLC (or saline) and inoculated i.n. with 100 μL A/Wuhan/359/95, 107 TCID50 per animal 24 h later. Cotton rats were sacrificed 1 day after influenza infection for pulmonary viral load quantification.
To characterize effect of poly ICLC on influenza and RSV infection, cotton rats were inoculated i.n. with 100 μL of poly ICLC (40 μg or 1 μg per animal) or saline and infected i.n. with influenza virus (A/Wuhan/359/95, 107 TCID50 per animal in 100 μL) or RSV 24 h after treatment. RSV A/Long (105.6 pfu per animal) or RSV A/A2 (105 pfu per animal) was administered i.n. in 100 μL volume in all experiments on the effect on viral replication. RSV A/Long (105 pfu per animal) was used for all experiments on cytokine induction and pathology. Animals were sacrificed by CO2 inhalation on various days after infection (4 animals per group). A group of animals was sacrificed prior to infection (0 h) and served as uninfected control. The lungs were removed from the thorax and processed for viral titrations, RNA extraction, or histopathology analysis.
Lungs were prepared for histopathology analysis as previously described (Prince and others 2001). Each lung section was scored for peribronchiolitis, perivasculitis, alveolitis, and interstitial pneumonitis (Prince and others 2001). The maximum possible value for each lesion was 100. Individual lesion scores were added to yield a cumulative histopathology score for each animal. Random and blind scoring was applied to all samples.
Animals were sacrificed by CO2 inhalation. Lungs were extracted and bronchoalveolar lavage (BAL) was collected by sequentially infusing 4 aliquots of 2 mL PBS (pH 7.4) each into the right lung and immediately aspirating BAL into a separate syringe (via 3-way stopcock (Discofix, Braun)). BAL was centrifuged for 5 min at 400g and supernatant was collected and stored at −80°C until assayed for total protein content by Bio-Rad Protein Assay. BAL cells were treated with Red Blood Cell Lysing Buffer (Sigma) for 5 min, washed once with PBS (pH 7.4), and finally resuspended in 500 μL PBS with 0.1% BSA. The total number of cells recovered was determined using a hemocytometer. An aliquot containing 105–106 cells was used to make a cytospin preparation, which was then stained with Diff-Quik (Dade Behring) to estimate the cell differential. At least 200 cells were counted on each slide. Cells were differentiated using standard morphological characteristics. All differential cell counts were performed blind.
Expression of IFN-α, IFN-β, IFN-γ, IP-10, and Mx-1 mRNA was analyzed in total lung homogenates by RT-PCR followed by Southern blot procedure with chemiluminescent detection as previously described (Blanco and others 2002; Pletneva and others 2006).
Fragments of cotton rat cDNA for TLR1, TLR2, TLR3, TLR4, TLR5, and TLR7 were obtained from a lung cDNA pool of RSV-infected, poly ICLC-treated cotton rats (for TLR1, TLR3, TLR4, TLR5, and TLR7), or activated cotton rat spleen/macrophage mixed cDNA library (for TLR2) using degenerate primers for these TLR. The following GenBank accession numbers were assigned: TLR1, FJ475825; TLR2, FJ475829; TLR3, FJ475826; TLR4, FJ475827; TLR5, FJ475828; TLR7, FJ475830. Primers for real-time PCR of TLR fragments were designed using Primer3 software (http://primer3.sourceforge.net/) (Table 1). Resulting amplicons were sequenced to verify TLR identity. Sequences of cotton rat chemokines GRO, MCP-1, and MIP-1α were described by us previously (Blanco and others 2004). Primers for the real-time PCR analysis of chemokine expression (Table 1) were designed using Primer3 software. For all real-time PCRs, the Bio-Rad iQ™ SYBR Green Supermix was used in a final volume of 25 μL with final primer concentrations of 0.5 μM. Melting curves were completed for each primer pair and no nonspecific amplification was noted under conditions tested. Amplifications were performed on a Bio-Rad iCycler for 1 cycle of 95°C for 3 min, followed by 40 cycles of 95°C for 10 s, 60°C for 10 s, and 72°C for 15 s. The baseline cycles and cycle threshold (Ct) were calculated by the iQ5 software in the PCR Base Line Subtracted Curve Fit mode. Real-time PCR reactions were analyzed and results were normalized by the β-actin mRNA level in each sample as previously described (Boukhvalova and others 2007a). The standard curves were developed using serially diluted cDNA sample from the lungs of poly ICLC-treated animals.
Data were analyzed statistically using 2-way ANOVA followed by the Tukey post hoc test (SPSS 12.0.1). A value of P < 0.05 was considered to be significant.
To verify biological activity of poly ICLC in the cotton rat model, induction of IFN response by poly ICLC was measured (Fig. 1A). Type I and II IFNs induction was assessed by monitoring dynamics of pulmonary IFN-β and IFN-γ mRNA expression, respectively, as well as by measuring up-regulation of mRNA for Mx-1 (a type I IFN-inducible gene) and IP-10 (a chemokine inducible by both type I and type II IFNs). Poly ICLC induced a strong IFN response in the cotton rat lung. Up-regulation of IFN-β, IFN-γ, Mx-1, and IP-10 mRNA was maximal 24–48 h after poly ICLC administration and declined to basal levels by day 8. A pronounced dose dependence of the effect was seen, with 40 μg poly ICLC inducing maximal IFN mRNA responses.
Interferon-inducing effect of poly ICLC was accompanied by an increased cellular infiltration in the lung (primarily monocytes and neutrophils) around bronchioles (peribronchiolitis), small blood vessels (perivasculitis), within lung interstitium (interstitial pneumonitis), and alveolar lumen (alveolitis) (Figs. 1B and and2).2). Maximum histopathology was seen in animals inoculated with 40 μg poly ICLC, and no inflammatory response was detected in the lungs of animals inoculated with 1 μg poly ICLC. The dynamics of the inflammatory response differed from that of the IFN response kinetically, with maximum histopathology observed between days 2 and 5 after the treatment. Other features of pulmonary inflammation, such as an increase in bronchoalveolar lavage (BAL) protein content and an increase in the percentage of neutrophils in BAL, were also detected in poly ICLC-treated animals. BAL protein content increased from 63.0 ± 11.6 μg/mL in saline-treated animals to 134.7 ± 10.3 μg/mL in animals treated with 40 μg poly ICLC 5 days earlier (P < 0.05). Significantly larger number and percentage of neutrophils (17.7% of all BAL cells) was found in the BAL of poly ICLC-treated cotton rats compared to saline-treated animals (Fig. 3A). The number of macrophages was also increased by poly ICLC treatment, with a fraction of macrophages in BAL of poly ICLC-treated animals (~15%) displaying enlarged morphology with more intense staining in the cytoplasm. These cells were later also detected in BAL of influenza- or RSV-infected animals treated with poly ICLC (Fig. 3B).
The antiviral action of IFN against influenza is a staple of IFN function that originally led to discovery of IFN. To determine whether poly ICLC-induced IFN can inhibit influenza replication in cotton rats, animals were inoculated with 1–40 μg of poly ICLC and challenged with A/Wuhan/359/95 influenza virus 24 h later. Lungs were collected for viral titrations 1 day after infection, the time of maximum influenza replication in the cotton rat lung (Ottolini and others 2005). Poly ICLC treatment inhibited influenza replication in a dose-dependent manner, with 40 μg poly ICLC producing maximal, and 1 μg poly ICLC producing minimal inhibition (Fig. 4A). To determine whether the antiviral effect of poly ICLC was accompanied by changes in pulmonary inflammation during influenza infection, pulmonary histopathology was measured in the lungs of cotton rats inoculated with influenza after saline or poly ICLC treatment (40 μg of poly ICLC per animal). Infection of cotton rats with influenza virus led to a strong pulmonary inflammatory response that peaked on day 4 and declined on day 7 post-infection (Figs. 2 and and4B).4B). Treatment of cotton rats with poly ICLC prior to influenza infection had no significant effect on the dynamics or magnitude of pulmonary pathology of the disease (Fig. 4B). Analysis of inflammatory changes in the BAL of these animals revealed a decreased presence of macrophages in the airway spaces (Fig. 3A), although an increased number of enlarged macrophages was seen in the BAL of influenza-infected cotton rats treated with poly ICLC compared to influenza-infected animals treated with saline (data not shown).
In contrast to influenza, antiviral potential of IFN against RSV infection remains controversial (Sperber and Hayden 1989; Merolla and others 1995; Guerrero-Plata and others 2005; Johnson and others 2005; Rudd and others 2005; Jewell and others 2007). To clarify this issue in the cotton rat model, the poly ICLC dose with the strongest IFN-inducing ability, 40 μg per animal, was tested for its effect on RSV infection (Fig. 5). Cotton rats were treated with poly ICLC and infected with RSV 24 h later. Animals were sacrificed on various days after infection for analysis of viral load (Fig. 5A), pulmonary histopathology (Figs. 2 and and5B),5B), and BAL changes (Fig. 3). Two strains of RSV, A/Long and A/A2, were included in this experiment because of their reported differences with respect to IFN response (Schlender and others 2005).
Poly ICLC reduced replication of RSV in the cotton rat lungs, although the effect was less dramatic than seen with influenza replication (Fig. 5A). Early stages of RSV replication (day 1) were affected more by poly ICLC treatment than later stages (day 4). Poly ICLC was more effective in reducing replication of RSV A/A2 than RSV A/Long strain, suggesting strain-specific variations in sensitivity to IFN action, consistent with earlier reports (Schlender and others 2005; Pletneva and others 2008). RSV was cleared from the lungs of all animals by day 7. The antiviral effect of poly ICLC against RSV infection was dose-dependent, as administration of 1 μg poly ICLC had no effect on replication of either RSV A/Long or RSV A/A2 (data not shown). Administration of poly ICLC 1 day after RSV infection failed to significantly reduce RSV A/Long replication, and caused only transient and moderate (<1 log10 pfu/g) decrease in RSV A/A2 pulmonary levels (data not shown). Thus, poly ICLC has a stronger antiviral potential in the cotton rat model when administered prophylactically rather than therapeutically.
Although weaker, inflammatory response to RSV was kinetically similar to that of influenza infections, peaking on day 4 and declining by day 7 post-infection (Fig. 5B). In contrast to influenza, however, histopathology of RSV disease was significantly increased by prior poly ICLC treatment (Figs. 2 and and5B)5B) and quantitatively approached histopathology induced by influenza infection itself. The most striking difference between RSV-infected animals treated with poly ICLC or with saline was the dramatic increase in infiltration of neutrophils into alveolar spaces (Fig. 2, inserts). A significant increase in neutrophil number and percentage was also seen in the BAL of RSV-infected, poly ICLC-treated animals (Fig. 3A). This increase was accompanied by a concomitant increase in macrophage presence, and an appearance of the enlarged macrophages (Fig. 3B).
Antiviral effect of poly ICLC on respiratory viral infections might have been caused by an augmented IFN production. To address this issue, the dynamics of the pulmonary IFN response in cotton rats infected with influenza virus or RSV with or without poly ICLC treatment was measured. Both influenza and RSV infections in the absence of poly ICLC treatment led to a significant up-regulation of all cytokines analyzed (P < 0.05; Fig. 6). The dynamics of virus-induced cytokines, however, differed between the two viruses. All cytokines in influenza-challenged animals were strongly up-regulated within a day after infection. In contrast, RSV infection led to an early (peak on day 1) increase in expression of IFN-α and IFN-β mRNA, while expression of Mx-1, IP-10, and IFN-γ mRNA was delayed. Maximum expression levels were comparable in RSV- and influenza-challenged animals with the exception of IFN-α mRNA, which was lower in RSV-infected animals (P < 0.05).
When poly ICLC was used, almost the same level of cytokine mRNA expression was seen with poly ICLC plus either virus as with poly ICLC treatment alone. The dynamics of poly ICLC-induced response was associated with an earlier increase in IFN-inducible genes and more closely resembled that of influenza than RSV infection. The maximum response when poly ICLC was administered with virus infection did not exceed the level induced by poly ICLC alone. IFN-γ was the only cytokine induced more strongly by viral infections than by poly ICLC treatment, but was lower in virus-infected animals after poly ICLC treatment, suggesting that poly ICLC may reduce adaptive immune responses to viruses.
Although both dsRNA and respiratory viruses were shown to change TLR expression in vitro in cell cultures (Miettinen and others 2001; Guillot and others 2005; Ritter and others 2005; Groskreutz and others 2006), in vivo modulation of TLR expression remains obscure. Yet, changes in TLR expression may be linked to pulmonary damage. Because poly ICLC increased pulmonary inflammation, both by itself and when given prior to RSV, it was possible that the effect of poly ICLC on lung pathology was associated with changes in pulmonary TLR levels. We cloned cotton rat cDNAs for several TLR implicated in responses to dsRNA, influenza, or RSV infection, including TLR1, 2, 3, 4, 5, and 7, and developed real-time PCR assays for measuring their expression. Poly ICLC treatment of cotton rats resulted in elevated expression of TLR1, 2, 3, and 7 mRNA, decreased TLR5 mRNA, and had little effect on TLR4 mRNA levels in the lung (Fig. 7). Both influenza and RSV also caused increases in TLR1, 2, 3, and 7 mRNA levels, but the kinetics and magnitude of induction differed between the two viruses. Specifically, influenza caused stronger up-regulation of TLR expression within 1 day of viral challenge, while RSV-induced TLR up-regulation reached maximum only on day 4 post-infection and was in general lower than that induced by influenza. Expression of TLR5 was transiently suppressed in influenza-, but not in RSV-infected animals. In spite of reported in vitro effect (Monick and others 2003), no increase in TLR4 mRNA expression was seen in RSV-infected cotton rats, suggesting that RSV-induced TLR4 up-regulation might be spatially localized or that it has lesser importance for RSV biology in vivo.
Poly ICLC treatment had little effect on the dynamics of influenza-associated TLR mRNA expression (with the exception of TLR3 mRNA reduction), but altered RSV-induced TLR expression. Expression of TLR1, 2, 3, and 7 changed from the delayed type seen in RSV-infected animals to an early induction type characteristic of influenza-infected animals. TLR5 expression was reduced in poly ICLC-treated, RSV-infected animals compared to animals infected with virus alone. This switch from a delayed to an earlier induction type paralleled similar change in the dynamics of expression of IFN-inducible genes in RSV-infected, poly ICLC-treated animals.
Examination of lung parenchyma and bronchoalveolar lavage of poly ICLC-treated cotton rats revealed that poly ICLC caused increased infiltration of lungs with neutrophils and macrophages. Moreover, the morphology of inflammatory cells, in particular macrophages, noticeably changed following poly ICLC treatment. These results implied that recruitment and activation of neutrophils and macrophages might have been affected by poly ICLC treatment. To determine the possible mechanism responsible for these changes, we measured pulmonary expression of mRNA for chemokines mediating recruitment/activation of neutrophils and monocytes/macrophages, namely GRO, MCP-1, and MIP-1α. Expression of all three chemokines was significantly increased in poly ICLC-treated or influenza-infected cotton rats compared to saline-treated animals (Fig. 8). RSV infection caused a weaker chemokine response. However, expression of all three chemokines in RSV-infected animals was significantly augmented by prior poly ICLC treatment. Thus, poly ICLC may modulate neutrophil and macrophage recruitment and/or activation through changes in pulmonary chemokine levels. Chemokine expression during influenza infection was not further augmented by poly ICLC treatment, with the only exception of MIP-1α mRNA, which may have contributed to the appearance of morphologically distinct macrophages in the airways of infectedanimals.
Interferon production is among the earliest immune responses triggered by viral infection. There is a compelling amount of evidence suggesting that IFN and dsRNA can induce toxic effect characterized by flu-like symptoms (Majde 2000). This effect, however, seems to be species-specific. Treatment of humans with high doses of dsRNA is accompanied by flu-like syndrome (Brodsky and others 1985; Samuel 2001), while administration of dsRNA to certain strains of mice is safe and anti-inflammatory (Guerrero-Plata and others 2005). Even within the same murine strain (eg, BALB/c), there is some heterogeneity in the reported toxic effect of dsRNA (Wong and others 1995; Wong and others 1999; Guerrero-Plata and others 2005). Some studies attest that both intravenous and intranasal administration of poly ICLC induces toxic effect in BALB/c mice, characterized by hypothermia and body weight loss (Wong and others 1999). Other studies report that antiviral doses of poly ICLC cause no weight loss, and in fact alleviate weight loss and reduce histopathology accompanying RSV infection in BALB/c mice (Guerrero-Plata and others 2005). This work, carried in the cotton rat model, demonstrates that antiviral doses of dsRNA (poly ICLC) indeed induce an influenza-like illness, characterized by increased pulmonary inflammation. The important similarity between dsRNA- and influenza-engaged mechanisms appears to be determined by a particular signature of IFN-inducible genes and pulmonary TLR expression. The fact that this similarity is important for pulmonary inflammatory response is supported by the finding that pathology of milder respiratory infection caused by RSV can be augmented by converting IFN and TLR expression dynamics to resemble poly ICLC- and influenza-induced pattern.
Influenza virus is a strong inducer of both type I and II IFNs and is also easily inhibited by the antiviral action of IFN (Wong and others 1995). For RSV, both induction of IFN response and its antiviral potential remain controversial (Sperber and Hayden 1989; Merolla and others 1995; Guerrero-Plata and others 2005; Johnson and others 2005; Rudd and others 2005; Ramaswamy and others 2006; Jewell and others 2007). This work indicates that RSV stimulates a strong IFN response in vivo, comparable in magnitude to that induced by influenza infection (with the exception of IFN-α mRNA level), but is characterized by a delayed induction of IFN-γ, Mx-1, and IP-10 mRNA. Moreover, IFN can inhibit RSV replication, but antiviral effect is more limited to early stages of RSV disease, possibly due to the ability of RSV to counteract IFN response (Ramaswamy and others 2006). No additive or synergistic increase in IFN levels was observed during combined dsRNA virus treatment, suggesting that the lung capacity for IFN induction is limited and/or is tightly regulated to prevent IFN overexpression.
Ability of synthetic dsRNA to trigger flu-like response in certain animal models depends on IFN-γ. In C57BL/6 mice, acute flu-like symptoms can be induced by intraperitoneal, but not intratracheal, administration of poly IC in spite of an increased type I IFN production. The inability of intratracheally administered poly IC to mimic flu in mice, however, can be circumvented by adding IFN-γ to intratracheal poly IC inoculums (Traynor and others 2004). In the cotton rat model, intranasal poly ICLC treatment increases endogenous IFN-γ production, at the same time increasing pulmonary pathology. Taken together these findings suggest important species-specific variations in IFN response, and prompt a speculation that activation of both type I and II IFNs may mediate pulmonary inflammatory response. Curiously, a recent study of patients with severe acute respiratory syndrome also suggests that type I and II IFNs and IFN-stimulated chemokines contribute to immunopathology of lung disease (Cameron and others 2007).
Treatment of cotton rats with poly ICLC or infecting them with influenza or RSV caused significant changes in pulmonary TLR expression. Expression of the same TLR was modulated in response to all 3 different stimuli. The particulars of TLR expression changes in cotton rats, such as up-regulation of TLR1, 2, and 3 mRNA and down-regulation of TLR5 mRNA, strikingly resemble changes induced by influenza infection and IFN and dsRNA stimulation of human macrophages and epithelial cells (Miettinen and others 2001; Ritter and others 2005). In general, the dynamics of TLR expression changes in treated cotton rats were more similar for influenza and poly ICLC than for poly ICLC and RSV and paralleled that of IFN response. Poly ICLC treatment modified RSV-induced TLR expression to resemble influenza-associated signature, much the same way it modified RSV-induced IFN expression. Recent in vitro studies indicate that IFNs (both types I and II) regulate TLR expression (Miettinen and others 2001). Moreover, RSV-induced activation of RIG-I pathway in epithelial cells mediates TLR3 induction via secreted IFN-β (Liu and others 2007). These findings may indicate that the way pulmonary TLR microenvironment responds to respiratory viruses is determined in large by IFN response to infection. Alterations in lung TLR levels likely reflect inducible mRNA expression in respiratory epithelial and resident inflammatory cells, as well as an increase in pulmonary cellular infiltrate (neutrophils, like macrophages, express a wide range of TLR molecules (Hayashi 2003)). Rapid changes in pulmonary TLR expression induced by early IFN response may have important implications for viral pathogenesis via chemokine overexpression and macrophage- and neutrophil-associated lung damage (Welliver 2003; Reed and others 2008; Welliver and others 2008). Pulmonary chemokine expression was significantly increased by poly ICLC treatment and was augmented by poly ICLC in RSV-infected animals. Increased expression of chemokines regulating neutrophil and monocyte/macrophage recruitment and activation in RSV-infected, poly ICLC-treated animals led to a significant increase in neutrophil and macrophage presence in the BAL and an overall increase in pulmonary pathology of RSV infection. Neutrophil survival, prolonged during RSV infection (Jones and others 2002) and mediated via TLR pathways (Lindemans and others 2006), may have also contributed to increased neutrophil presence in the airways of RSV-infected animals after poly ICLC treatment.
Overall, this work demonstrates that activation of pulmonary IFN response establishes an antiviral state in the lung, but also promotes tightly regulated proinflammatory state. Even in the case when viral replication is inhibited, early triggering of IFN-inducible genes initiates immune response that could lead to lung inflammation. This proinflammatory effect may be mediated through regulation of pulmonary TLR expression by IFN and through associated changes in chemokine expression. This finding has important implications for clinical science, as it may explain the lack of therapeutic effect of antiviral drugs administered after the onset of respiratory viral disease.
We thank Dr. Stefanie Vogel for critical and insightful reading of this manuscript. We also thank Stephany Venero for technical help with cotton rat TLR cloning, and Fredy and Ana Rivera and Charles Smith for the animal care.
Marina S. Boukhvalova, Virion Systems, Inc., Rockville, Maryland.
Talia B. Sotomayor, Virion Systems, Inc., Rockville, Maryland.
Ryan C. Point, Virion Systems, Inc., Rockville, Maryland.
Lioubov M. Pletneva, Virion Systems, Inc., Rockville, Maryland.
Gregory A. Prince, Virion Systems, Inc., Rockville, Maryland.
Jorge C.G. Blanco, Virion Systems, Inc., Rockville, Maryland.
This work was supported by the Virion Systems, Inc., corporate funds and by the National Institutes of Health (grant AI-057575 to J.C.G.B.). The authors do not have an association that might pose a conflict of interest.