Immunopathology Respiratory Syncytial Virus Role for Innate IFNs in ...

The Journal of Immunology

Role for Innate IFNs in Determining Respiratory Syncytial Virus Immunopathology1 Teresa R. Johnson,† Sara E. Mertz,* Negin Gitiban,* Sue Hammond,* Robin LeGallo,* Russell K. Durbin,* and Joan E. Durbin2* Respiratory syncytial virus (RSV) is the major cause of severe lower airway disease in infants and young children, but no safe and effective RSV vaccine is yet available. The difficulties involved in RSV vaccine development were recognized in an early vaccine trial, when children immunized with a formalin-inactivated virus preparation experienced enhanced illness after natural infection. Subsequent research in animal models has shown that the vaccine-enhanced disease is mediated at least in part by memory cells producing Th2 cytokines. Previously we had observed enhanced, eosinophilic lung pathology during primary infection of IFNdeficient STAT1ⴚ/ⴚ mice that are incapable of generating Th1 CD4ⴙ cells. To determine whether these effects depended only on Th2 cytokine secretion or involved other aspects of IFN signaling, we infected a series of 129SvEv knockout mice lacking the IFN-␣␤R (IFN-␣␤Rⴚ/ⴚ), the IFN-␥R (IFN-␥Rⴚ/ⴚ), or both receptors (IFN-␣␤␥Rⴚ/ⴚ). Although both the IFN-␥Rⴚ/ⴚ and the IFN-␣␤␥Rⴚ/ⴚ animals generated strong Th2 responses to RSV-F protein epitopes, predominantly eosinophilic lung disease was limited to mice lacking both IFNRs. Although the absolute numbers of eosinophils in BAL fluids were similar between the strains, very few CD8ⴙ T cells could be detected in lungs of IFN-␣␤␥Rⴚ/ⴚ animals, leaving eosinophils as the predominant leukocyte. Thus, although CD4ⴙ Th2 cell differentiation is necessary for the development of allergic-type inflammation after infection and appears to be unaffected by type I IFNs, innate IFNs clearly have an important role in determining the nature and severity of RSV disease. The Journal of Immunology, 2005, 174: 7234 –7241.

R

espiratory syncytial virus (RSV)3 is the major cause of severe lower airway disease in infants and young children, but no safe and effective RSV vaccine is yet available. The difficulties involved in RSV vaccination became evident in early vaccine trials, when children immunized with a formalininactivated virus preparation experienced enhanced illness after natural infection (1, 2). Subsequent studies in animal models have shown that unlike natural infection, formalin-inactivated RSV priming leads to the development of a Th2 memory response, which is thought to mediate the immunopathology seen after virus challenge (3–5). This is of particular concern in light of recent studies showing a link between severe RSV disease in childhood and the development of asthma later in life (6). To test this hypothesis, in previous studies we examined whether mice lacking the ability to develop a Th1 response would develop eosinophilic airway disease after primary RSV infection (7). For that purpose we studied mice lacking STAT1 (8), the major mediator of IFN-␣␤ and IFN-␥ signaling. Th1 differentiation of naive CD4⫹ cells de*Columbus Children’s Research Institute and Department of Pediatrics, Ohio State University College of Medicine and Public Health, Columbus, OH 43205; and †Vaccine Research Center, National Institutes of Health, Bethesda, MD 20892 Received for publication December 3, 2004. Accepted for publication March 24, 2005. 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. 1 This work was supported by Grant AI47226 from the National Institutes of Health (to J.E.D.). 2 Address correspondence and reprint requests to Dr. Joan E. Durbin, Department of Pediatrics, Ohio State University College of Medicine and Public Health, Columbus Children’s Research Institute, 700 Children’s Drive, Room WA4014, Columbus, OH 43205-2696. E-mail address: [email protected] 3 Abbreviations used in this paper: RSV, respiratory syncytial virus; BAL, bronchoalveolar lavage; i.n., intranasally; IP-10, IFN-induced protein 10; I-TAC, IFN-induced T cell ␣ chemoattractant; MIG, monokine induced by IFN-␥; RNAP, RNase protection assay.

Copyright © 2005 by The American Association of Immunologists, Inc.

pends on the presence of the transcription factor T-bet, the induction of which is STAT1 dependent (9, 10). In fact, RSV-infected STAT1⫺/⫺ mice did develop Th2 cytokine responses and severe eosinophilic lung disease without priming, and this correlated with high levels of IL-13 production by virus-specific CD4⫹ T cells. Although these results clearly demonstrated the link between the development of Th2 cells and enhanced disease, we were struck by the fact that IFN-␥⫺/⫺ animals studied in parallel with STAT1⫺/⫺ mice were much less affected. This result was unanticipated, because we had assumed that mice lacking IFN-␥ would also default to a Th2 CD4⫹ T cell differentiation pathway. We concluded that although CD4⫹ T cells producing Th2 cytokines were necessary for an allergic-type, eosinophilic response to develop, other factors could modulate this effect. Because STAT1 mediates signaling through both the IFN-␣␤ and IFN-␥ receptors (11), we wondered whether type I (␣␤) IFNs were acting to diminish the pathology seen in STAT1 animals. We hypothesized that either IFN-␣␤ was affecting CD4⫹ T cell differentiation, or other effects of type I IFN were acting to limit eosinophilic disease. For a more rigorous approach to these questions, we have compared the innate and adaptive responses to acute RSV infection in a series of 129SvEv knockout mice lacking the IFN-␣␤R, the IFN␥R, both receptors, or STAT1⫺/⫺. In this study we have observed that both mouse strains lacking the IFN-␥R produce virus-specific Th2 lymphocytes in response to RSV infection regardless of IFN-␣␤ signaling. Nonetheless, significant differences remain between mice lacking both IFNRs and mice that retain the IFN-␣␤R. As found previously in the BALB/c mouse, inflammatory infiltrates from infected mice deficient in IFN-␥ signaling remained primarily lymphocytic, whereas bronchoalveolar lavage (BAL) specimens from IFN-␣␤␥R⫺/⫺ (IFN-␣␤␥R⫺/⫺) mice consisted primarily of eosinophils, with a notable lack of CD8⫹ T cells. The degree of eosinophilia seen in the lungs of IFN-␣␤␥R⫺/⫺ mice therefore reflects a paucity of responding CTLs as well as a large 0022-1767/05/$02.00

The Journal of Immunology number of eosinophils presumably elicited by IL-5- and IL-13producing CD4⫹ cells. This study demonstrates that although Th2 cytokines are essential for the development of an allergic-type inflammatory response, they are not the sole determinant of virally induced immune pathogenesis.

Materials and Methods Virus, cells, and virus quantitation A laboratory stock of RSV A2 (originally obtained from R. Chanock, National Institutes of Health, Bethesda, MD) was propagated in HEp-2 cells (American Type Culture Collection) as previously described (12). Infectivity was assayed using a fibroblast cell line derived from STAT1⫺/⫺ mice, designated NY3 (8). For determination of lung titers, lung tissue was weighed, and samples from individual mice were homogenized in 1 ml of PBS using a PowerGen125 homogenizer (Fisher Scientific). Dilutions of lung supernatants were inoculated onto subconfluent monolayers of NY3 cells, and after 1 h, plates were covered with 0.5% methylcellulose in medium and incubated for 36 h at 37°C. In our experience, NY3 and HEp2 cells give comparable titers. Monolayers were fixed with 2% buffered formaldehyde and stained with crystal violet.

Mouse infection Mice on the 129SvEv background lacking the type I IFNR subunit 1, ifnar1⫺/⫺ (IFN-␣␤R⫺/⫺) (13), the type 2 IFNR subunit 1, ifngr1⫺/⫺ (IFN␥R⫺/⫺) (14), or both genes, ifnnar1⫺/⫺/ifngr1⫺/⫺ (IFN-␣␤␥R⫺/⫺) (15) were provided by R. Schreiber (Washington University, St. Louis, MO) and originally derived by M. Aguet (Swiss Institute for Imperial Cancer Research, Lausanne, Switzerland). Strain-matched controls were purchased from Taconic Farms. Mice with a targeted disruption of the stat1 gene (8) were backcrossed onto the 129SvEv strain, and animals of the ninth backcross generation were used in these experiments. Cohorts of 6- to 8-wk-old animals were lightly anesthetized and inoculated intranasally (i.n.) with 107 PFU of RSV A2. Groups of four to six animals were used for each data point.

BAL Mice were killed, and the tracheas were isolated by blunt dissection. One milliliter of PBS was injected and slowly withdrawn through a blunt tip needle inserted and secured in the airway. After the total cell number was determined, each BAL fluid sample was centrifuged, and the supernatants were stored at ⫺70°C until further use. The pellets were resuspended in a small volume of PBS, and cytospins were prepared using a Shandon Cytospin centrifuge (Thermo Electron). BAL cell counts were made using a hemocytometer; only cells excluding trypan blue (Invitrogen Life Technologies) were counted. The numbers of eosinophils and lymphocytes were calculated by multiplying the total BAL cell numbers by the corresponding fractions derived from differential counts. The numbers of CD4⫹ T cells, CD8⫹ T cells, and B cells per sample were calculated using flow cytometry data to determine the percentage of each cell type within the lymphocyte gate and total lymphocyte numbers. Cells marking as lymphocytes, taken together, comprised a total percentage of BAL cells similar to differentials based on morphology.

Flow cytometry of BAL cells BAL samples were incubated with allophycocyanin-conjugated Abs to CD8, PE-conjugated Abs to CD4, and FITC-conjugated Abs to B220 (BD Pharmingen) for 30 min at room temperature. Samples were washed and resuspended in FACS buffer. Data were collected on a FACSCalibur (BD Biosciences) flow cytometer and analyzed with CellQuest software (BD Biosciences). Dead cells were excluded on the basis of forward and side light scatter.

Pathology Lungs were harvested from mice 8 days after infection with RSV and were inflated with 10% buffered formalin. Paraffin sections were stained with H&E. Differential cell counts were performed on cytospins of BAL cell pellets harvested from each animal. Cytospins were stained with WrightGiemsa, and monocyte/macrophages, lymphocytes, neutrophils, and eosinophils were enumerated based on cell morphology and staining.

Identification of an RSV-F peptide epitope A series of 57 overlapping peptides corresponding to the 574-aa RSV B strain, F protein sequence (16) was synthesized by Chiron Mimotopes.

7235 Each peptide was 20 aa in length and overlapped the adjacent peptide by 10 residues. Individual peptides were dissolved in 10% DMSO to a concentration of 4 mg/ml and then pooled into sets of 10. As an initial screen, peptide pools were used to stimulate splenocytes harvested from wild-type 129SvEv mice infected i.n. 28 days previously with 106 PFU of RSV. After 7 days of peptide stimulation, splenocytes (0.5 ⫻ 105) from RSV-immune animals were mixed with 1.5 ⫻ 105 gamma-irradiated, peptide-pulsed (10 ␮g/ml), naive splenocytes in AIM-V medium containing 2.5% heat-inactivated mouse serum (Stellar BioSystems). Each sample was plated in triplicate wells of a round-bottom, 96-well plate. Two days after peptide restimulation, 1 ␮Ci of [3H]thymidine was added to each well. After an additional 12-h incubation at 37°C, [3H]thymidine incorporation was measured using a Wallac Microbeta scintillation counter.

ELISPOT IFN-␥ ELISPOT assay kits were purchased from U-Cytech. Splenocytes harvested from RSV-immune 129SvEv mice (wild-type, IFN-␣␤R⫺/⫺, IFN-␥R⫺/⫺, and IFN-␣␤␥R⫺/⫺) were plated directly or after CD4⫹ cell depletion. Depletion was performed using mouse CD4 Dynabeads (Dynal Biotech) according to the manufacturer’s instructions. Splenocytes were suspended in AIM-V medium containing 2.5% heat-inactivated mouse serum and plated at various dilutions in 96-well plates. Viral or irrelevant (SIINFEKL) peptides were included at a concentration of 10 ␮g/ml. After 24 h at 37°C, the cells were removed, and IFN-␥-containing spots were visualized by sequential incubations with a biotinylated polyclonal antiIFN-␥ Ab, an anti-biotin tertiary Ab, and the detection reagent according to the manufacturer’s protocol.

Measurement of cytokine transcripts by RNase protection assay (RNAP) Immune splenocytes were placed in culture and stimulated with 10 ␮g/ml peptide for 1 wk; 6 h after restimulation, RNA was purified using TRIzol (Invitrogen Life Technologies). RNAPs were performed using the mCK-1 RiboQuant kit purchased from BD Pharmingen. Assays were performed according to the instructions provided by the manufacturer using 15 ␮g of lymphocyte RNA/sample. Signal intensity over background was quantitated using the Storm PhosphorImager (Molecular Dynamics) and was normalized to an L32 internal control using ImageQuant software.

Cytokine and chemokine assays BAL supernatants were sent to Linco Research to determine concentrations of IL-5 and IFN-␥. This method uses beads conjugated to analyte-specific Abs, which are detected by the Luminex100 system. IL-13, IFN-␥, eotaxin, and IFN-induced protein 10 (IP-10) concentrations were measured using ELISA kits purchased from R&D Systems. IFN-␥ concentrations for individual samples were consistent between the assay systems.

Statistical analysis Data are expressed as the mean ⫾ SEM. Comparisons between groups were based on ANOVA, calculated using SigmaStat software (SAS Institute). Dunnett’s procedure, which adjusts for multiple comparisons against a predefined group, was applied only when there was a significant overall difference between groups. A value of p ⬍ 0.05 was considered statistically significant.

Results Control of RSV replication is STAT1 dependent We have previously observed a strain-dependent increase in virus production after i.n. infection of STAT1-deficient mice (7). In the current study we examined the relative RSV susceptibilities of additional strains of knockout animals (wild-type, IFN-␣␤R⫺/⫺, IFN-␥R⫺/⫺, IFN-␣␤␥R⫺/⫺, and STAT1⫺/⫺), all on a 129SvEv background. Groups of 10 animals from each genotype were inoculated i.n. with 1 ⫻ 107 PFU of RSV. The titers shown in Fig. 1 are derived from lungs taken on day 5 after infection, the peak of virus production, and on day 8 when virus would normally be cleared. In vivo, both IFN-␣␤ and IFN-␥ are induced by virus infection, albeit with somewhat different kinetics, and we expected that either cytokine alone would induce an antiviral state (17). This prediction was borne out; mice lacking either the IFN-␣␤R or the IFN-␥R demonstrated titers similar to those in wild-type mice, although titers from STAT1⫺/⫺ animals were increased 100-fold.

7236

INNATE IFNs AND RSV IMMUNOPATHOLOGY and focused instead on comparing the responses of mice lacking either or both IFNRs to those of strain-matched controls. We had previously described pronounced eosinophilic lung disease after primary RSV infection of STAT1⫺/⫺ and IFN␣␤␥R⫺/⫺, but not IFN-␣␤R⫺/⫺, IFN-␥R⫺/⫺, or IFN-␥⫺/⫺, mice, as judged by microscopic examination of infected tissues (7). For a more quantitative picture, we collected BALs from mice of each genotype at 5 days (Fig. 3A) and 8 days (Fig. 3B) after i.n. RSV

FIGURE 1. Control of RSV replication is STAT1 dependent, but IFN independent. Cohorts of 4- to 6-wk-old 129SvEv mice of each genotype were infected i.n. with 107 PFU of RSV A2. Titers shown in this figure (expressed as PFU per gram of lung tissue) derive from plaque assay of lung homogenates collected on day 5 after inoculation. Only STAT1⫺/⫺ mice produced virus at levels greater than those seen in wild-type animals (p ⬍ 0.001) as determined by one-way ANOVA. By day 8, virus was below the level of detection in all strains (data not shown).

By day 8 after infection, no virus could be detected in lungs of any strain (data not shown). The surprising finding was that mice unable to respond to either type of IFN showed no detectable increase in virus production relative to wild-type controls. This nonequivalence of the IFN-␣␤␥R⫺/⫺ and STAT1⫺/⫺ mutations was not entirely unexpected, because STAT1 is essential for many, but not all, IFN signaling pathways (8, 18). What was surprising, and unlike other viruses, was the relative STAT1 dependence and IFN independence of RSV replication in vivo (19). Altered inflammatory responses in lungs of IFN-deficient mice Along with the increased viral load there was an exacerbated inflammatory response in the lungs of STAT1⫺/⫺ mice (Fig. 2). There were no significant differences in the numbers of BAL leukocytes recovered from wild-type, IFN-␣␤R⫺/⫺, IFN-␥R⫺/⫺, or IFN-␣␤␥R⫺/⫺ mice on days 5 and 8 after infection. However, samples from STAT1⫺/⫺ mice showed 15- and 5-fold greater numbers of inflammatory cells at 5 and 8 days after infection, respectively ( p ⬍ 0.001 on days 5 and 8). Although we have also found an enhanced inflammatory response to RSV by STAT1⫺/⫺ mice on the BALB/c background, where differences in viral titers are much smaller (7), the 100-fold increase in virus found in the STAT1⫺/⫺ 129SvEv animals (Fig. 1) undoubtedly influences many immune parameters. Because we wanted to determine the effects of IFN, rather than virus load, on the outcome of RSV infection, we omitted STAT1⫺/⫺ mice from the following studies

FIGURE 2. Lung lavages from infected STAT1⫺/⫺ mice were more cellular than those from IFN-␣␤R⫺/⫺, IFN-␥R⫺/⫺, IFN-␣␤␥R⫺/⫺ mice or strain-matched controls. BALs were collected from infected animals of each genotype on days 5 and 8 after RSV infection, and cell counts were determined for individual animals. Only STAT1⫺/⫺ mice had a statistically significant increase in the number of BAL cells (p ⬍ 0.001) at either time point. We have consistently seen a 2-fold increase in the number of inflammatory cells present in BALs from IFN-␥R⫺/⫺ mice on day 8, as has been reported by others (39), but it was not statistically significant in this analysis.

FIGURE 3. The composition of BAL infiltrates differed significantly between RSV-infected wild-type and knockout mice. BALs were collected from RSV-infected animals 5 and 8 days after i.n. inoculation with 107 PFU of RSV, and differential cell counts were performed on WrightGiemsa-stained cytospins. At the earlier, day 5, time point (A) neutrophils were found in mice lacking either or both IFNRs, but not wild-type animals (p ⬍ 0.05). Differences in percent lymphocytes between IFN-␣␤R⫺/⫺ and IFN-␣␤␥R⫺/⫺ and controls were also significant (p ⬍ 0.05). By day 8 (B). BAL from IFN-␣␤R⫺/⫺ animals were similar to that in wild-type mice. IFN-␥R⫺/⫺ and IFN-␣␤␥R⫺/⫺ mice had elevated percentages of neutrophils (p ⬍ 0.05) and eosinophils (p ⬍ 0.001) compared with wild-type mice. The proportion of lymphocytes in IFN-␣␤␥R⫺/⫺ mice was significantly decreased (p ⬍ 0.05). C, Day 8 BAL eosinophil composition was compared in three independent experiments; the percentage of eosinophils differs significantly (p ⬍ 0.05) between IFN-␥R⫺/⫺ and IFN-␣␤␥R⫺/⫺ mice in all cases.

The Journal of Immunology inoculation and performed differential cell counts on WrightGiemsa-stained cytospins made from individual animals (four to six mice per group). Lavage fluids obtained from wild-type mice consisted almost entirely of monocytes/macrophages and lymphocytes, with the proportion of lymphocytes increasing by day 8 when tissue inflammation was well developed. In IFN-␣␤R⫺/⫺, IFN-␥R⫺/⫺, and IFN-␣␤␥R⫺/⫺ mice, neutrophils were present on day 5, but decreased to near background levels by day 8. Eosinophils were present only in the lungs of IFN-␥R⫺/⫺ and IFN␣␤␥R⫺/⫺ animals. In the double-receptor knockout mice, eosinophils appeared earlier and consistently made up a significantly larger proportion of the total infiltrate. These differences were consistent over three similar experiments, one of which is pictured in Fig. 3, A and B. When these data were analyzed by ANOVA, all three experiments (Fig. 3C) showed significant differences between the IFN-␥R⫺/⫺ and IFN-␣␤␥R⫺/⫺ strains ( p ⬍ 0.05). Another significant and consistent difference among the different genotypes was the relative dearth of lung lymphocytes in IFN␣␤␥R⫺/⫺ animals on day 8 ( p ⬍ 0.05), shown in Fig. 3B. Photomicrographs corresponding to these BAL data are shown in Fig. 4. H&E-stained lung sections from infected wild-type (A and D), IFN-␥R⫺/⫺ (B and E), and IFN-␣␤␥R⫺/⫺ (C and F) mice 8 days after inoculation are pictured. As predicted from the equivalent BAL cell numbers (Fig. 2), the overall extent of inflammation was similar in all four strains. This can be seen at the lower (⫻200) magnification. All sections show pronounced perivascular infiltrates with foci of peribronchial, peribronchiolar, and alveolar inflammation. There was little evidence of cytopathic effect or epithelial cell necrosis. Appreciable differences between strains were apparent only at higher (⫻400) magnification, where the very different characters of the infiltrates are evident. Consistent with the data in Fig. 3, the inflammatory process is entirely lymphocytic in wild-type mice (Fig. 4D) and predominantly lymphocytic in IFN␥R⫺/⫺ mice (Fig. 4E), but is overwhelmingly eosinophilic in IFN␣␤␥R⫺/⫺ animals (Fig. 4F). This relates to both the relative increase in eosinophils and the relative decrease in lymphocytes seen in the double-knockout animals. Lung sections from IFN-␣␤R⫺/⫺ mice (not pictured in this figure) were similar to those from wildtype mice, but with an overall decrease in lymphocytic inflammation. On day 5 after infection, the total number of lymphocytes present in the lungs of all animals was quite small (data not

FIGURE 4. Tissue inflammation reflects BAL composition. Tissue sections are from wild-type (A, ⫻200; D, ⫻400), IFN-␥ R⫺/⫺ (B, ⫻200; E, ⫻400), and IFN-␣␤␥ R⫺/⫺ (C, ⫻200; F, ⫻400) animals 8 days after i.n. RSV inoculation. Eosinophils are large cells filled with bright pink granules; lymphocytes are small, blue cells with very little cytoplasm. Several eosinophils are marked with black arrowheads.

7237 shown), but by day 8 we observed a large expansion of T cells in wild-type and IFN-␥R⫺/⫺ mice. This increase was primarily due to the expanding CD8⫹ cell population, as calculated from cell counts, BAL differentials, and flow cytometry data (Fig. 5A). Interestingly, this expansion was very limited in mice lacking the IFN-␣␤R, although the difference in CD8⫹ T cell number was statistically significant ( p ⬍ 0.05) only for the IFN-␣␤␥R⫺/⫺ animals. This observation is consistent with data from other laboratories showing a role for type I IFNs in the activation and survival of CTLs (20, 21). There were no significant differences in numbers of BAL CD4⫹ T cells among strains. Fig. 5B shows the number of eosinophils present in BALs taken from each strain on day 8 postinfection. These were very similar in IFN-␥R⫺/⫺ and IFN␣␤␥R⫺/⫺ animals, leading us to conclude that the eosinophilic predominance in IFN-␣␤␥R⫺/⫺ mice reflects primarily a lack of CD8⫹ T cells. CD4⫹ T cell differentiation in IFN-deficient knockout mice Having determined that there were significant qualitative and quantitative differences between RSV lung disease in IFN-␥R⫺/⫺ and IFN-␣␤␥R⫺/⫺ mice, we sought to determine the role of IFNs in CD4⫹ T cell differentiation. It has been established that STAT1mediated induction of the transcription factor T-bet is required for the development of Th1 lymphocytes (10). Consistent with this finding, we have observed Th2 cytokine production by STAT1deficient CD4⫹ T cells responding to an RSV-F protein epitope that stimulates primarily IFN-␥ in wild-type mice (7). Although a large body of evidence demonstrates the importance of IFN-␥ in Th1 differentiation in vitro, in vivo studies using mice deficient in IFN-␥ signaling have not shown a clear Th2 predisposition (22– 24). In light of the uncertainty regarding the role of IFN-␥ in Th1 differentiation, and with the differing disease patterns observed in IFN-␥⫺/⫺ or IFN-␥R⫺/⫺ and STAT1⫺/⫺ mice infected with RSV (7) or influenza (25), it was important to learn whether virus-specific CD4⫹ T cells from RSV-immune IFN-␣␤R⫺/⫺, IFN-␥R⫺/⫺, and IFN-␣␤␥R⫺/⫺ mice were primarily Th1 or Th2.

FIGURE 5. Numbers of BAL lymphocytes and eosinophils on day 8 after infection. A, Total numbers of lymphocytes were calculated using total numbers of BAL cells taken from each animal and percentages of total lymphocytes (cytospin differentials) and lymphocyte subsets (FACS analysis) of each BAL sample. Difference in numbers of total lymphocytes and CD8⫹ T cell numbers were statistically significant (p ⬍ 0.05) only for the IFN-␣␤␥R⫺/⫺ animals. Total numbers of eosinophils (B) were similarly calculated.

7238 To perform this experiment in the 129SvEv animals, a dominant viral epitope recognized by RSV-immune CD4⫹ lymphocytes in this strain was required. Overlapping peptides corresponding to the RSV-F protein sequence were mixed to make five pools of 10 peptides and one pool of seven peptides, and these peptide pools were assayed for their ability to stimulate the proliferation of immune splenocytes. Because pool 1 stimulated the most robust proliferation response (data not shown), we next screened peptides 1–10 individually (Fig. 6). From this secondary screen, peptide 9, corresponding to aa 81–100 (QELDKYKNAVTELQLLMQNT), was chosen for additional characterization. An IFN-␥ ELISPOT assay performed with undepleted or CD4⫹ T cell-depleted immune splenocytes demonstrated that CD4⫹ T cells were required for an IFN-␥ response to this peptide. Undepleted cultures produced an average of 85 spots/106 splenocytes compared with an average of 1 spot/106 splenocytes in CD4⫹ T cell-depleted samples. Depletion of CD8⫹ T cells had no effect on IFN-␥ production (data not shown). Cytokine production by immune CD4⫹ T cells from mice of each genotype was assessed by quantitation of cytokine gene transcripts after peptide stimulation of cultured splenocytes. Cells were harvested from two or three mice of each strain that had been inoculated i.n. with 106 PFU of RSV 28 days previously. After 7 days of culture with peptide 9, the splenocytes were washed and restimulated for a 6-h period before harvesting. Equal amounts of RNA from each sample were analyzed in a multi-probe RNAP, and signals were quantitated using the Storm PhosphorImager and ImageQuant software. Relative levels of cytokine transcripts made by each strain are shown in Fig. 7. We concluded from this set of experiments that although RSV does not appear to generate a strong Th1 response to F protein in the wild-type mouse, mice lacking the IFN-␥R alone or lacking both IFNRs default to a Th2 differentiation pathway in our system. Role of IFNs in lung cytokine and chemokine production The earliest response to virus is initiated by infected lung epithelial cells. A number of cytokines and chemokines are induced by RSV (26 –28), and these inflammatory mediators directly affect the trafficking of lymphocytes and granulocytes to the site of infection. Chemokines promoting both Th1 and Th2 cell migration are produced by the infected epithelium in vitro, but in vivo, their synthesis is also affected by the presence of other cytokines. Many of the chemokines attracting Th1 cells (IFN-inducible T cell ␣ chemoattractant (I-TAC), IP-10, monokine induced by IFN-␥ (MIG)) are induced by IFNs (29 –32), and IL-13 is known to increase the

FIGURE 6. Identification of a viral peptide recognized by RSV-immune 129SvEv splenocytes. Activation of immune cells by individual peptides was assayed by measuring [3H]thymidine incorporation after in vitro peptide stimulation. Peptides derived from the F protein sequence of RSV B were combined in pools. Because pool 1 (no. 1–10) induced the highest level of proliferation, the peptides making up this pool were assayed individually. Peptide 9, from pool 1, corresponds to aa 81–100 of the RSV-F protein.

INNATE IFNs AND RSV IMMUNOPATHOLOGY

FIGURE 7. Memory CD4⫹ T cells from RSV-immune IFN-␥R⫺/⫺ and IFN-␣␤␥R⫺/⫺ mice produced high levels of Th2 cytokines. Splenocytes from RSV-immune animals of each genotype were expanded in culture. RNA was harvested from each culture after peptide stimulation, and RNAP was used to determine relative levels of cytokine transcripts. Signal intensity was quantitated by PhosphorImager and normalized to levels of ribosomal protein L32.

synthesis of eotaxin (33, 34), a major eosinophil attractant. We wondered whether the altered inflammatory responses in the absence of IFN-␣␤ and/or IFN-␥ signaling correlated with an altered chemokine profile. To answer this question, BAL supernatants from infected mice were collected 5 days after i.n. inoculation with 107 PFU of RSV and were assayed for the presence of IP-10. IP-10 was chosen because it can be induced by either type I or type II IFNs (29, 32). Fig. 8 shows the data obtained for both eotaxin and IP-10 assayed at two time points in all four strains of mice. Although no IP-10 could be detected in mice lacking both IFNRs (IFN-␣␤␥R⫺/⫺), mice retaining either the IFN-␣␤R ((IFN␣␤R⫺/⫺) or the IFN-␥R (IFN-␥R⫺/⫺) continued to produce IP-10. Eotaxin was induced in all strains, but eosinophilia occurred only in those lacking virus-specific Th1 cells, that is, the IFN-␥R⫺/⫺ and IFN-␣␤␥R⫺/⫺ mice. In addition, eotaxin levels remained significantly elevated on day 8 only in mice lacking both IFNRs ( p ⬍ 0.001). These same BAL supernatants from each animal were also assayed for the presence of Th2 cytokines. Levels of IFN-␥, IL-5, and IL-13 in lavage samples collected at 5 and 8 days after inoculation were determined by ELISA or by the Luminex100 system (Fig. 9). Although there were increased levels of IL-5 and IL-13 in lungs of IFN-␥R⫺/⫺ and IFN-␣␤␥R⫺/⫺ mice, these differences

FIGURE 8. BAL chemokines IP-10 and eotaxin were assayed by ELISA on day 5 (f) and day 8 (䡺) after primary RSV infection.

The Journal of Immunology

FIGURE 9. BAL cytokines IFN-␥ (A) and IL-13 (B) were assayed by ELISA, and IL-5 (C) was assayed by LINCOplex on day 5 (f) and day 8 (䡺) after primary RSV infection.

were statistically significant only for IL-5. IFN-␥ levels were similarly high for all strains on day 5, probably reflecting production by NK as well as T cells. By day 8, these levels had dropped, most precipitously in the IFN-␣␤R⫺/⫺ animals. Thus, it appears that chemokine synthesis is affected by early IFN production, and this suggests that chemokine levels may be influenced by the presence of Th2 lymphocytes.

Discussion Although many childhood diseases have been largely eradicated or attenuated by successful vaccination programs, RSV remains a major pediatric problem. One of the obstacles in the development of a safe and effective RSV vaccine is the danger that the dominance of a Th2 memory compartment will lead to enhanced disease, rather than protection, after immunization. We had previously observed that in mice lacking a functional STAT1 gene, primary infection with RSV led to severe eosinophilic lung disease that appeared to be mediated by Th2 lymphocytes. In retrospect, this is not surprising, because T-bet, the transcription factor shown to be essential for Th1 development of naive CD4 T cells (35, 36), is itself STAT1 dependent (37, 38). Therefore, even in the setting of primary RSV infection, it appears that host predisposition to Th2 differentiation determines whether an eosinophilic response will develop. This conclusion was inconsistent with the absence of this pathology in infected mice lacking IFN-␥, a cytokine known to activate STAT1 and drive Th1 differentiation of naive CD4⫹ T cells. In vivo experiments to determine whether IFN-␥ itself is essential

7239 for Th1 development have been complicated by the importance of IFN-␥ in establishing IL-12 responsiveness. Current models of naive CD4⫹ cell activation postulate that STAT1 activation is required for Th1 commitment, but that the STAT4-mediated IL-12 signal is still necessary for IFN-␥ production in response to Ag (10, 38). Because IFN-␥ up-regulates IL-12R expression, a robust Th1 response should require IFN-␥, and we would expect mice lacking this cytokine to mount a Th2 response to virus infection. Nonetheless, in our experiments with either IFN-␥⫺/⫺ BALB/c or 129SvEv IFN-␥R⫺/⫺ mice, eosinophils were present in infiltrates surrounding vessels and airways of infected lungs, but only at low levels. This observation has been made by others (24, 39), who concluded, as we did, that RSV pathogenesis in IFN-␥-deficient animals is similar to that in wild-type animals. We therefore wondered whether, in the absence of IFN-␥, IFN-␣␤ could act in a compensatory manner to inhibit Th2 lymphocyte development and eosinophilic disease (7). To test this hypothesis we needed to determine 1) whether Th2 differentiation was blunted in IFN-␥R⫺/⫺ mice, and 2) whether mice lacking both IFNRs (IFN-␣␤␥R⫺/⫺) responded differently from those lacking one or the other. We approached the question of T cell differentiation in infected IFN-␥-deficient mice by comparing cytokine responses of four mouse strains to a single MHC class II restricted viral peptide. The peptide used (F9), corresponding to aa 81–100 of the RSV-F protein, specifically stimulated the proliferation of and IFN-␥ production by CD4⫹ T cells from RSV-immune wild-type 129SvEv mice. Using an RNAP to survey simultaneously for Th1 and Th2 cytokine production, we observed that although peptide-stimulated T cells from IFN-␥R⫺/⫺ and IFN-␣␤␥R⫺/⫺ animals produced levels of IFN-␥ transcript similar to those in wild-type animals, IL-13 production was increased 10- to 20-fold (Fig. 7). The elevated amounts of IL-13 as well as IL-5 induced by peptide F9 in the absence of the IFN-␥R demonstrated that the IFN-␥ pathway is essential for inhibiting the Th2 response to RSV infection. Having found that both the IFN-␥R⫺/⫺ and the IFN-␣␤␥R⫺/⫺ animals default to a Th2 lymphocyte response after infection, an alternate explanation for the significant differences in lung pathology between these animals was required. Characterization of BAL samples from mice of each genotype at two time points suggested that the major difference between these strains was not an increase in eosinophils, but, rather, an absence of CD8⫹ lymphocytes in both strains of mice lacking the IFN-␣␤R⫺/⫺. In wild-type and IFN-␥R⫺/⫺ mice, there was a large expansion in the number of lymphocytes present in the lungs by day 8 after inoculation. In IFN-␥R⫺/⫺ animals, this expansion occurred in the presence of eosinophils and neutrophils (Fig. 3). In contrast, mice lacking both IFNRs showed little increase in lymphocyte numbers as the infection progressed, and eosinophils were the predominant cell type on day 8. This pattern was seen consistently in three replicate experiments and was significant in each ( p ⬍ 0.05). Evaluation of tissue sections showed a similar picture (Fig. 4), with largely lymphocytic infiltrates in infected IFN-␥R⫺/⫺ mice and predominantly eosinophilic inflammation in lungs of IFN-␣␤␥R⫺/⫺ animals. We have observed that Th2 cytokine production by CD4⫹ T cells is necessary, but not sufficient, for the development of eosinophilic lung inflammation after RSV infection. Other factors contribute to pulmonary eosinophilia during primary infection in mice. It has been shown that a large number of genes are induced by RSV infection of respiratory epithelial cells; among these are chemokines that regulate inflammatory cell trafficking (26, 27, 40, 41). A number of innate immune mediators will impact relative levels of chemokine production by these cells (31), and among these are the IFNs. Both IFN-␣␤ and IFN-␥ are produced early in infection: IFN-␣␤ by the virus-infected cells and innate IFN-␥ by activated

7240 NK cells (17). A number of the chemokines responsible for Th1 and CTL trafficking, among them the CXCR3 ligands I-TAC, MIG, and IP-10, are IFN-inducible genes (31, 32, 41). We wondered whether IP-10, which is known to be induced by both type I and type II IFNs, would be detectable in mice lacking either or both IFNRs and whether differences in chemokine production could be contributing to the differences in pathology. As shown in Fig. 9, IP-10 was detectable in BALs from all animals on day 5, except in those lacking both IFNRs. Eotaxin, another RSV-induced chemokine, is a chemoattractant for activated Th2 cells and eosinophils (42). Although eotaxin was present in BALs from mice of all genotypes at early times, levels were significantly increased only in IFN-␣␤␥R⫺/⫺ mice on day 8 ( p ⬍ 0.001; Fig. 9A). Allergic inflammation is a self-amplifying process, and the elevated eotaxin levels in animals with eosinophilic disease may reflect this situation. If activated Th2 cells and eosinophils secreting IL-13 (43) traffic to the lung, increasing IL-13 concentrations will up-regulate eotaxin production by infected epithelium (34) (see Fig. 10). In summary, our data suggest that although Th2 lymphocytes are necessary for the development of an eosinophilic response to primary RSV infection, they are not sufficient. In this system it is the IFN-␣␤ and IFN-␥ produced during the innate immune response, rather than lymphocyte IFN-␥, that determines disease outcome. RSV lung pathology in IFN-␥R⫺/⫺ animals with an intact type I IFN signaling pathway is similar to that seen in wild-type controls. CD8⫹ T cells, rather than eosinophils, continue to dominate the process. In the absence of both IFNRs, disease is markedly eosinophilic and without the usual expansion of the lung lymphocyte compartment (44, 45). This is consistent with a report from Hussell et al. (46) showing a correlation between the magnitude of the CD8⫹ T cell response and the propensity of a mouse strain to develop eosinophilia. Therefore, the elevated eotaxin levels, the absence of IP-10 induction, and the lack of CTLs in IFN␣␤␥R⫺/⫺ animals work together to create differences between two strains of animals that both generate Th2 responses. Fig. 10 depicts a model showing how the summation of these influences will determine the composition of the inflammatory infiltrate.

FIGURE 10. Innate IFNs shape the inflammatory response by regulating CTL expansion and chemokine induction at the site of infection. Both IFN-␣␤ and eotaxin are synthesized by RSV-infected epithelial cells, and NK cells are the early IFN-␥ producers. In the absence of both IFNRs, induction of IFN-stimulated chemokine genes does not occur, decreasing the influx of IFN-␥-producing cells. Eotaxin, also produced by infected cells, will attract both Th2 cells and eosinophils to the lung, and IL-13 production by these cells will further amplify allergic-type inflammation.

INNATE IFNs AND RSV IMMUNOPATHOLOGY We propose that innate IFN-␣␤ and -␥ affect local chemokine production by increasing levels of the Th1 and Tc1 chemoattractants (I-TAC, IP-10, and MIG) as well as stimulating CD8⫹ T cell proliferation and activation. An important role for IFN-␣␤ in CTL survival and IFN-␥ production has also been shown in other systems (20, 21, 47). If virus-specific Th1 cells and CTLs are present, they will be drawn to the site of infection and augment local IFN-␥ production. If activated Th2 lymphocytes are present, they will preferentially accumulate in response to eotaxin, increasing local IL-13 and eotaxin production and thereby amplifying an allergictype inflammatory response. In animals lacking the IFN-␥R⫺/⫺, it appears that the presence of type I IFNs does not block eosinophil accumulation, but can still inhibit a primarily allergic response by promoting CD8⫹ T cell accumulation. It has been suggested that in addition to their roles as CXCR3 agonists, I-TAC, IP-10, and MIG may also act to block Th2 cell and eosinophil accumulation by binding the CCR3 expressed by those cell types and acting as antagonists (48). If this is a biologically significant process, this antagonism could also play a role in shaping the inflammatory response. Overall, our findings in the mouse model suggest that although the link between RSV and asthma (6) is probably due to the ability of this virus to promote Th2 lymphocyte differentiation, it may be exacerbated by the limited type I IFN induction by RSV infection in vivo (49, 50).

Acknowledgments We thank Michel Aguet and Bob Schreiber for providing IFN-␣␤R⫺/⫺, IFN-␥⫺/⫺, and IFN-␣␤␥R⫺/⫺ mice; Tom Hamilton for anti-IP-10 antiserum; and Philip R. Johnson for RSV F protein peptides. We also thank Chris Walker, Naglaa Shoukry, Andrew Cawthon, and Christine Biron for helpful discussions, and Karen Watkins for manuscript preparation.

Disclosures The authors have no financial conflict of interest.

References 1. Kim, H. W., J. G. Canchola, C. D. Brandt, G. Pyles, R. M. Chanock, K. Jensen, and R. H. Parrott. 1969. RSV disease in infants despite prior administration of antigenic inactivated vaccine. Am. J. Epidemiol. 89: 422– 434. 2. Chin, J., R. L. Magoffin, L. A. Shearer, J. H. Schieble, and E. H. Lennette. 1969. Field evaluation of a RSV vaccine and a trivalent parainfluenxa virus vaccine in a pediatric population. Am. J. Epidemiol. 89: 449 – 463. 3. Graham, B. S. 1995. Pathogenesis of RSV-augmented pathology. Am. J. Respir. Crit. Care Med. 152: S63–S66. 4. Waris, M. E., C. Tsou, D. D. Erdman, S. R. Zaki, and L. J. Anderson. 1996. RSV infection in BALB/c mice previously immunized with formalin-inactivated virus induces enhanced pulmonary inflammatory response with a predominant Th2-like pattern. J. Virol. 70: 2852–2860. 5. Connors, M., A. B. Kulkarni, C. Y. Firestone, K. L. Holmes, H. C. Morse III, A. V. Sotnikov, and B. R. Murphy. 1992. Pulmonary histopathology induced by respiratory syncytial virus (RSV) challenge of formalin-inactivated RSV-immunized BALB/c mice is abrogated by depletion of CD4⫹ T cells. J. Virol. 66: 7444 –7451. 6. Sigurs, N., R. Bjarnason, F. Sigurbergsson, and B. Kjellman. 2000. Respiratory syncytial virus bronchiolitis is an important risk factor for asthma and allergy at age 7. Am. J. Respir. Crit. Care Med. 161: 1501–1507. 7. Durbin, J. E., T. R. Johnson, R. K. Durbin, S. Mertz, R. S. Peebles, and B. S. Graham. 2002. The role of interferon in respiratory syncytial virus pathogenesis. J. Immunol. 168: 2944 –2952. 8. Durbin, J. E., R. Hackenmiller, M. C. Simon, and D. E. Levy. 1996. Targeted disruption of the mouse Stat1 gene results in compromised innate immunity to viral disease. Cell 84: 443– 450. 9. Lucas, S., N. Ghilardi, J. Li, and F. de Sauvage. 2003. IL-27 regulates IL-12 responsiveness of naive CD4⫹ T cells through Stat1-dependent and -independent mechanisms. Proc. Natl. Acad. Sci. USA 100: 15047–15052. 10. Afkarian, M., J. Sedy, J. Yang, N. Jacobson, N. Cereb, S. Yang, T. Murphy, and K. Murphy. 2002. T-bet is a STAT1-induced regulator of IL-12R expression in naive CD4⫹ T cells. Nat. Immunol. 3: 549 –557. 11. Levy, D., and J. Darnell. 2002. STATS: transcriptional control and biological impact. Nat. Rev. Mol. Cell. Biol. 3: 651– 662. 12. Graham, B. S., G. S. Henderson, Y.-W. Tang, X. Lu, K. M. Neuzil, and D. G. Colley. 1993. Priming immunization determines T helper cytokine mRNA expression patterns in lungs of mice challenged with RSV. J. Immunol. 151: 2032–2040.

The Journal of Immunology 13. Muller, I., U. Steinhoff, L. Reis, S. Hemmi, J. Pavlovic, R. Zinkernagel, and M. Aguet. 1994. Functional roles of type I and type II interferons in antiviral defense. Science 264: 1918 –1921. 14. Huang, S., W. Hendriks, A. Althage, S. Hemmi, H. Bluethmann, R. Kamijo, J. Vilcek, R. M. Zinkernagel, and M. Aguet. 1993. Immune response in mice that lack the IFN-␥ receptor. Science 259: 1742–1745. 15. van den Broek, M. F., U. Muller, S. Huang, M. Aguet, and R. M. Zinkernagel. 1995. Antiviral defense in mice lacking both ␣/␤ and ␥ interferon receptors. J. Virol. 69: 4792– 4796. 16. Johnson, P. R., and P. L. Collins. 1988. The fusion glycoproteins of human respiratory syncytial virus of subgroups A and B: sequence conservation provides a structural basis for antigen relatedness. J. Gen. Virol. 69: 2623–2628. 17. Biron, C., and G. Sen. 2001. Interferons and other cytokines. In Fields Virology, Vol. 1. D. Knipe and P. Howley, eds. Lippincott Williams & Williams, New York, p. 321. 18. Meraz, M., J. M. White, K. C. Sheehan, E. A. Bach, S. J. Rodig, A. S. Dighe, D. H. Kaplan, J. K. Riley, A. C. Greenlunc, D. Campbell, et al. 1996. Targeted disruption of the STAT1 gene in mice reveals unexpected physiologic specificity in the JAK-STAT signaling pathway. Cell 84: 431– 442. 19. Gil, M., E. Bohn, A. O’Guin, C. Ramana, B. Levine, G. Stark, H. Virgin, and R. Schreiber. 2001. Biologic consequences of Stat1-independent IFN signaling. Proc. Natl. Acad. Sci. USA 98: 6680 – 6685. 20. Marrack, P., J. Kappler, and T. Mitchell. 1999. Type I interferons keep activated T cells alive. J. Exp. Med. 189: 521–530. 21. Ogasawara, K., S. Hida, Y. Weng, A. Saiura, K. Sato, H. Takayangi, S. Sakaguchi, T. Yokochi, T. Kodama, M. Naitoh, et al. 2002. Requirement of the IFN-␣/␤-induced CXCR3 chemokine signalling for CD8⫹ T cell activation. Genes Cells 7: 309 –320. 22. Coyle, A., T. S., C. Bertrand, S. Huang, M. Aguet, S. Alkan, and G. Anderson. 1996. Mice lacking the IFN-␥ receptor have impaired ability to resolve a lung eosinophilic inflammatory response associated with a prolonged capacity of T cells to exhibit a Th2 cytokine profile. J. Immunol. 156: 2680 –2685. 23. Schijns, V., B. Haagmans, E. Rijke, S. Huang, M. Aguet, and M. Horzinek. 1994. IFN-␥ receptor-deficient mice generate antiviral Th1-characteristic cytokine profiles but altered antibody responses. J. Immunol. 153: 2029 –2037. 24. Srikiatkhachorn, A., and T. Braciale, J. 1997. Virus-specific memory and effector T lymphocytes exhibit different cytokine responses to antigens during experimental murine RSV infection. J. Virol. 71: 678 – 685. 25. Durbin, J., A. Fernandez-Sesma, C. Lee, T. Rao, A. Frey, T. Moran, S. Vukmanovic, A. Garcia-Sastre, and D. Levy. 2000. Type I interferon modulates innate and specific antiviral immunity. J. Immunol. 164: 4220 – 4228. 26. Domachowske, J., C. Bonville, and H. Rosenberg. 2001. Gene expression in epithelial cells in response to pneumovirus infection. Respir. Res. 2: 225–233. 27. Haeberle, H., W. Kuziel, H. Dieterich, A. Casola, Z. Gatalica, and R. Garofalo. 2001. Inducible expression of inflammatory chemokines in RSV-infected mice: role of MIP-1a in lung pathology. J. Virol. 75: 878 – 890. 28. Zhang, L., M. Peeples, R. Boucher, P. Collins, and R. Pickles. 2002. Respiratory syncytial virus infection of human airway epithelial cells is polarized, specific to ciliated cells, and without obvious cytopathology. J. Virol. 76: 5654 –5666. 29. Ohmori, Y., and T. Hamilton. 1993. Cooperative interaction between interferon (IFN) stimulus response element and ␬B sequence motifs controls IFN␥- and lipopolysaccharide-stimulated transcription from the murine IP-10 promoter. J. Biol. Chem. 268: 6677– 6688. 30. Luster, A., J. Unkeless, and J. Ravetch. 1985. Gamma-interferon transcription regulates an early-response gene containing homology to platelet proteins. Nature 315: 672– 676. 31. Sauty, A., M. Dziejman, R. Taha, A. Iarossi, K. Neote, E. Garcia-Zepeda, A. Hamid, and A. Luster. 1999. The T cell-specific CXC chemokines IP-10, Mig, and I-TAC are expressed by activated human bronchial epithelial cells. J. Immunol. 162: 3549 –3558.

7241 32. Farber, J. 1997. Mig and IP-10: CXC chemokines that target lymphocytes. J. Leukocyte Biol. 61: 246 –257. 33. Komiya, A., H. Nagase, H. Yamada, T. Sekiya, M. Yamaguchi, Y. Sano, N. Hanai, A. Furuya, K. Ohta, K. Matsushima, et al. 2003. Concerted expression of eotaxin-1, eotaxin-2, and eotaxin-3 in human bronchial epithelial cells. Cell. Immunol. 225: 91–100. 34. Matsukura, S., C. Stellato, S. Georas, V. Casolaro, J. Plitt, K. Miura, S. Kurosawa, U. Schindler, and R. Schleimer. 2001. IL-13 upregulates eotaxin expression in airway epithelial cell by a STAT6-dependent mechanism. Am. J. Respir. Cell Mol. Biol. 24: 755–761. 35. Szabo, S., B. Sullivan, C. Stemmann, A. Satoskar, B. Sleckman, and L. Glimcher. 2002. Distinct effects of T-bet in TH1 lineage commitment and IFN-␥ production in CD4 and CD8 T cells. Science 295: 338 –342. 36. Szabo, S., S. Kim, G. Costa, X. Zhang, C. Fathman, and L. Glimcher. 2000. A novel transcription factor, T-bet, directs Th1 lineage commitment. Cell 100: 655– 669. 37. Szabo, S., B. Sullivan, S. Peng, and L. Glimcher. 2003. Molecular mechanisms regulating TH1 immune responses. Annu. Rev. Immunol. 21: 713–758. 38. O’Shea, J., and W. Paul. 2002. Regulation of Th1 differentiation: controlling the controllers. Nat. Immunol. 3: 506 –508. 39. van Schaik, S., N. Obot, G. Enhorning, K. Hintz, K. Gross, G. Hancock, A. Stack, and R. Welliver. 2000. Role of interferon ␥ in the pathogenesis of primary RSV infection in BALB/c mice. J. Med. Virol. 62: 257–266. 40. Zhang, Y., B. A. Luxon, A. Casola, R. P. Garofalo, M. Jamaluddin, and A. R. Brasier. 2001. Expression of respiratory syncytial virus-induced chemokine gene networks in lower airway epithelial cells revealed by cDNA microarrays. J. Virol. 75: 9044 –9058. 41. Miller, A., T. Bowlin, and N. Lukacs. 2004. Respiratory Syncytial virus-induced chemokine production: linking viral replication to chemokine production in vitro and in vivo. J. Infect. Dis. 189: 1419 –1430. 42. Garcia-Zepeda, E., M. Rothenberg, R. Ownbey, J. Celestin, and L. P. 1996. Human eotaxin is a specific chemoattractant for eosinophil cells and provides a new mechanism to explain tissue eosinophilia. Nat. Med. 2: 449 – 456. 43. Schmid-Grendelmeier, P., F. Altznauer, B. Fischer, C. Bizer, A. Straumann, G. Menz, K. Blaser, B. Wuthrich, and H. Simon. 2002. Eosinophils express functional IL-13 in eosinophilic inflammatory diseases. J. Immunol. 169: 1021–1027. 44. Chang, J., and T. Braciale. 2002. RSV infection suppresses lung CD8⫹ T-cell effector activity and peripheral CD8⫹ memory in the respiratory tract. Nat. Med. 8: 54 – 60. 45. Chang, J., A. Srikiatkhachorn, and T. Braciale. 2001. Visualization and characterization of RSV F-specific CD8⫹ T cells during experimental virus infection. J. Immunol. 167: 4254 – 4260. 46. Hussell, T., A. Georgiou, T. E. Sparer, S. Matthews, P. Pala, and P. J. Openshaw. 1998. Host genetic determinants of vaccine-induced eosinophilia during RSV infection. J. Immunol. 161: 6215– 6222. 47. Nguyen, K., W. Watford, R. Salomon, S. Hofmann, G. Pien, A. Morinobu, M. Gadina, J. O’Shea, and C. Biron. 2002. Critical role for STAT4 activation by type 1 interferons in the interferon-␥ response to viral infection. Science 297: 2063–2066. 48. Loetscher, P., A. Pellegrino, J. H. Gong, I. Mattioli, M. Loetscher, G. Bardi, M. Baggiolini, and I. Clark-Lewis. 2001. The ligands of CXC chemokine receptor 3, I-TAC Mig, and IP-10, are natural antagonists for CCR3. J. Biol. Chem. 276: 2986 –2991. 49. Hall, C. B., R. G. J. Douglas, R. L. Simons, and J. M. Geiman. 1978. Interferon production in children with RSV, influenza and parainfluenza virus infections. J. Pediatr. 93: 28 –32. 50. Roberts, N. J., J. Hiscott, and D. J. Signs. 1992. The limited role of the human interferon system response to RSV challenge: analysis and comparison to influenza virus challenge. Microb. Pathog. 12: 409 – 414.