Type I IFN Modulates Innate and Specific Antiviral Immunity

Mice

The production of Stat1^−/− mice on a CD1 background has been previously described (6). For comparison with mice disrupted for the IFN-γ gene on the C57BL6 background and with IFN-α receptor knockout or IFN-α-γ receptor double knockout mice on the 129Sv/Ev background, Stat1^−/− mice were backcrossed to C57BL6 or to 129Sv, respectively (8th backcross generation). Similar results to those described here were also observed with Stat1^−/− mice carrying a distinct mutation (7) that were maintained on a 129Sv/Ev background (data not shown). IFN-γ^−/− mice (12) were purchased from The Jackson Laboratory, Bar Harbor, ME. Mice with targeted mutations at the IFN-α receptor locus (IFNAR^−/−) and at both the IFN-α receptor and IFN-γ receptor loci (IFNAR/GR^−/−) (13, 14) were kindly provided by Michel Aguet (Lausanne Switzerland) and Robert Schreiber (Washington University, St. Louis, MO). Wild-type C57BL6, 129SvEv, and CD1 mice were purchased from Taconic Laboratories (Germantown, NY). Animals were bred and maintained under specific pathogen-free conditions. All work with animals conformed to guidelines approved by the Institutional Animal Care and Use Committee of New York University School of Medicine.

Viruses

Influenza A/WSN (H1N1) and NA/B-NS viruses were grown in Madin-Darby bovine kidney cells and titered by plaque assay on Madin-Darby bovine kidney cells, as previously described (10). Influenza A/PR/8/34 (H1N1) (PR8) was grown in the allantoic cavities of 10-day-old embryonated eggs (SPAFAS, Preston, CT) and titered by plaque formation on Madin-Darby canine kidney cells in the presence of trypsin or by hemagglutination (HA)³ assay. Virus production assays to determine tissue culture infectious dose (TCID) were performed on confluent monolayers of Madin-Darby canine kidney cells in 96-well plates infected with dilutions of viral suspensions. Following 2–3 days of incubation, supernatants were titrated by HA, and the lowest concentration of virus to produce an infection that scored positive by HA was defined as 1 TCID. The HA assay was done in V-bottom 96-well plates. Serial 2-fold dilutions of each sample were incubated for 1 h on ice with an equal volume of a 0.5% suspension of chicken erythrocytes (Truslow Farms, Chestertown, MD). Wells containing an adherent, homogeneous layer of erythrocytes were scored as positive.

Infections

Mice 6–10 wk old were anesthetized and inoculated intranasally (i.n.) with 50 μl virus diluted in PBS. For determination of LD₅₀ in mice of different genotypes, various doses of PR8, ranging from 10 to 10⁵ TCID per animal, were administered and animals were monitored daily for signs of illness. Survival was scored after 10 days. For rechallenge experiments, mice were initially inoculated with 10³ TCID of the attenuated NA/B-NS viral strain, and serum samples were obtained after 3 wk. Animals with positive influenza virus-specific Ab titers and control animals that had been inoculated with PBS alone were rechallenged with 10⁵ PFU WSN, and animals were monitored as described above.

For comparison of virus replication, clearance, and pathogenesis, animals were infected with 10² or 10³ TCID of PR8. Groups of 3–5 mice of each genotype were sacrificed at 3- to 4-day intervals. One lung from each animal was homogenized in 2 ml PBS, and the presence of influenza virus was quantified by virus production/HA assay. The second lung from each animal was divided. One lung lobe was inflated and fixed in 10% buffered formalin for histological examination. The remainder was homogenized in TRIZOL (Life Technologies, Gaithersburg, MD) for RNA isolation and evaluation of cytokine transcripts. Samples for microscopic examination were paraffin embedded after fixation and 5-μm sections were stained with hematoxylin and eosin.

Cytokine determinations

Spleen cells after lysis of RBC were cultured at 2 × 10⁶ cells/ml for 24 h in the presence of soluble anti-CD3 Ab 2C11. Subsequently, cells were washed, incubated with 50 U/ml IL-2 in the absence of Ab for 7 days, replated at 2 × 10⁶ cells/ml, and then left untreated or restimulated with anti-CD3 Ab. Supernatants were harvested from triplicate cultures after 48 h. Production of IFN-γ, IL-4, IL-5, and IL-10 was measured by cytokine-specfic ELISA (PharMingen, San Diego, CA).

For measurement of cytokine transcripts from lymphocytes, splenocytes were cultured for 4 days with plate-bound anti-CD3 (2C11) plus anti-mouse CD28 (37.51) Abs and IL-2. Cells were washed and cultured for an additional 48 h with IL-2. After restimulation with plate-bound Ab for 6 h, cells were lysed with TRIZOL (Life Technologies). RNA was isolated from the splenocyte lysates and from lung tissue-TRIZOL homogenates after phenol and phenol-chloroform extractions. Cytokine transcripts were assayed by RNase protection with the use of RiboQuant multiprobes according to the manufacturer’s protocols (PharMingen). Splenocyte RNA (5–8 μg/sample) and lung RNA (20–30 μg/sample) were used. For intracellular staining of cytokine protein, cells were pretreated with brefeldin A (1 μg/ml) for 3 h to enhance accumulation of secreted proteins followed by surface staining with anti-CD4-Tricolor or anti-CD8-Tricolor Abs (Caltag, South San Francisco, CA). Cells were subsequently fixed and permeabilized with Fix & Perm reagent (Caltag), followed by staining with anti-IFN-γ-FITC and anti-IL-5-PE Abs (Caltag).

Preparation of bone marrow-derived macrophages

Marrow harvested from four femurs of each mouse strain were plated in a six-well tissue culture dish in DMEM supplemented with 10% mouse L cell-conditioned medium and 20% FCS (Life Technologies). The medium was changed after 48 h, and adherent cells were allowed to grow to near confluence. For FACS analysis, cells were dislodged by scraping and then stained with FITC-conjugated hamster anti-mouse CD11b (clone 500-A2, Caltag). Individual wells were left untreated, treated with LPS at 5 μg/ml (from Escherichia coli strain 0127:B8, Sigma, St. Louis, MO), or infected with 10⁶ PFU influenza virus A/WSN/33. After 24 h incubation, cells were lysed in TRIZOL and RNA extracted with phenol-chloroform.

Cellular cytotoxicity assay

CTL activity from mice that had survived infection with influenza virus was measured on autologous cells. Spleen cells from virus-infected mice were treated to lyse RBC and divided into three samples. One sample was irradiated for use as stimulator cells by infecting with influenza A/PR8/34 virus at a multiplicity of infection of 4. Stimulators were incubated together with a second sample of spleen cells (responders) at a ratio of 4 × 10⁶ stimulators plus 4 × 10⁶ responders/ml for 5 days at 37°C. A third sample of splenocytes was stimulated with 25 μg/ml Salmonella typhosa LPS for 5 days at 2 × 10⁶ cells/ml and used as targets in cytotoxicity assays. Virus-specific cytotoxicity was measured by mixing spleen cells at a ratio of 40:1 with autologous target cells labeled with sodium [⁵¹Cr]chromate, either uninfected or infected with influenza PR8 virus at a multiplicity of infection of 10. Cells were incubated for 4 h at 37°C and pelleted, and radioactivity released into supernatants was measured. Maximal release was measured by incubating target cells with 0.5% Nonidet P-40 and spontaneous release by incubating target cells with medium alone (routinely <10% of maximal release). Percent cytotoxicity was calculated as [(test release − spontaneous release)/(maximal release − spontaneous release)] × 100. Ab-redirected lysis, a measure of activated CD8 T lymphocytes, was determined by similar assays but using ⁵¹Cr-labeled P815 tumor cells coated with anti-CD3 Ab 2C11 as targets as previously described (15).

CTL activity from immunized mice was measured against heterologous targets pulsed with influenza virus antigenic peptide. Immunity to influenza virus was generated in C57BL6/J mice (wild-type and Stat1^−/−) by i.p. injection of 250 hemagglutinating units of influenza virus A/PR/8/34 per animal 10 d before harvest of spleens. Spleen cells were isolated as described above and stimulated in vitro for 5 days by cocultivation with irradiated (2000 rad) influenza virus-infected spleen cells (1 × 10⁸ spleen cells were infected with 1 × 10³ hemagglutinating units A/PR8/34). In vitro-stimulated CTL activity was measured against ⁵¹Cr-labeled EL4 (H-2^b) target cells incubated with a 100 nM concentration of a synthetic peptide representing aa 366–374 of A/PR8/34 nucleoprotein (16) in a standard 4-h chromium release assay, as described above.

Influenza A/PR/8/34 virus clearance was unimpaired in Stat1^−/− mice

We have previously demonstrated an increased sensitivity of Stat1^−/− mice relative to wild-type animals to infection by human influenza virus strain A/PR/8/34 (10). Stat1^−/− mice infected i.n. with this virus displayed an LD₅₀ 10-fold lower than their wild-type counterparts. This increased sensitivity could not be explained by increased virus replication in the absence of IFN responsiveness, because peak virus titers were comparable in lungs harvested from infected wild-type or mutant animals. When viral titers were determined throughout the course of the infection, the kinetics of viral clearance were comparable in wild-type, Stat1^−/−, and IFNAR^−/− mice (10).

To characterize the cellular immune response to influenza virus, CTL activity was measured with the use of splenocytes derived from C57BL6 mice immunized in vivo by i.p. injection of a sublethal dose of PR8 virus (Fig. 1⇓). Both wild-type and Stat1^−/− mice generated cytotoxic lymphocytes capable of killing target cells pulsed with the immunodominant influenza virus peptide presented by H-2D^b (16). Cytotoxicity measurements over a range of peptide concentrations or with virus-infected target cells showed no significant differences between wild-type and Stat1^−/− effector cells, indicating no impairment of either the specificity or the efficiency of the cytotoxic response in the absence of Stat1 (Fig. 1⇓ and data not shown). Antiviral immunity was also measured in mice that had survived respiratory infection with an attenuated strain of influenza virus, NA/B-NS (17). Splenocytes from animals of both genotypes were harvested, stimulated in vitro with influenza-virus infected spleen cells, and assayed for Ag-specific cytotoxicity against autologous infected blasts. As a second measure of the abundance of CTL generated during influenza virus infection, Ab-redirected cellular cytotoxicity was also measured against P815 tumor cells. In all cases, significant influenza virus-specific cellular cytotoxicity was detected in recovered mice, but no significant differences were observed to indicate a deficiency in the absence of Stat1 (data not shown). This antiviral cytotoxic response likely contributed to the equal virus elimination observed in infected mice of both genotypes. Increased sensitivity to influenza virus infection in vivo in the absence of Stat1 was therefore not due to loss of any direct antiviral effect of IFNs on either virus replication or virus elimination by the host.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

Stat1^−/− mice are competent to produce influenza virus-specific CTL. Wild-type (WT) (□, ○) and Stat1^−/− C57BL6/J mice (▪, •) were immunized with a sublethal dose of influenza virus A/PR/8, and splenocytes were assayed for cytotoxic activity against EL4 target cells either without (▪, □) or pulsed with NP_366–374 peptide (•, ○). Specific target cell lysis was measured at various E:T ratios, as indicated, and results from one representative mouse of a group of four of each genotype are shown.

Influenza virus induced a distinct pathological process in Stat1^−/− mice

To compare disease progression in mice lacking type I and/or type II IFN responsiveness, we monitored pulmonary histopathology in wild-type (C57BL6, 129SvEv, and CD1 backgrounds), IFNAR^−/− (129SvEv background), IFN-γ^−/− (C57BL6 background), and Stat1^−/− mice (C57BL6, 129SvJ, and CD1 backgrounds). Five animals of each genotype were infected i.n. with 10² TCID of PR8 virus, and lung tissues were harvested and formalin fixed at 6 days postinoculation. Dramatic differences in disease pathology were revealed by comparative lung histology, and very different inflammatory responses to influenza virus infection were observed dependent on the genotype of the infected animals (Fig. 2⇓). Although some variation was noted among the three different strain backgrounds tested, presumably reflecting differences in virus susceptibility not related to the IFN system, significant differences in disease pathology were detected only in mutant strains. Moreover, the major differences in host response dependent on targeted loss of IFN responsiveness were replicated with the Stat1 mutation bred onto each of the different strain backgrounds 129, C57BL6, and outbred CD1.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

Distinct inflammatory response to influenza virus in Stat1^−/− and IFNAR^−/− mice. Photomicrographs show histological sections of lungs from wild-type 129 mice, Stat1^−/− 129 mice, IFN-γ^−/− C57BL6 mice, and IFNAR^−/− 129 mice 6 days after i.n. infection with 100 TCID of influenza virus A/PR/8. A–D, ×20 views of infected lungs; E–H, ×100 magnifications to show infiltrating cells. A and E, Lungs from wild-type mice showing peribronchial and bronchiolar cuffs of lymphocytes. m, a smooth muscle cell of the airway. B and H, Diffuse inflammatory process involving the entire lung of the Stat1^−/− animals with inflammatory cells consisting entirely of macrophages, neutrophils, and eosinophils infiltrating between the epithelial lining of an airway (ep) and its muscular coat (m). C and F, IFN-γ^−/− lung inflammation is peribronchiolar and largely lymphocytic. D and G, IFNAR^−/− mouse lung with patchy bronchopneumonia of mixed peribronchial and perivascular infiltrates consisting of lymphocytes and acute inflammatory cells. In addition to bronchitis, the infiltrating cells extended into lung parenchyma.

In wild-type animals, infected airways were surrounded by a small cuff of responding lymphocytes by day 6 post inoculation (Fig. 2⇑, A and E). Inflammation in IFN-γ^−/− mice was similar to wild-type, although somewhat more cellular (Fig. 2⇑, C and F), consistent with previous reports showing little differential sensitivity to influenza virus infection in animals lacking IFN-γ (18). In contrast, lungs from infected Stat1^−/− mice (Fig. 2⇑, B and H) showed a markedly exacerbated inflammatory process. Infiltrates were much more cellular and diffuse. In addition to the increased number of inflammatory cells, the character of the infiltrates was changed in the Stat1^−/− mice. Although small numbers of lymphocytes were identified in the infected tissue, the majority of responding cells were granulocytes, and infiltrating macrophages, neutrophils, and eosinophils were abundant (Fig. 2⇑H). Interestingly, whereas the inflammatory process detected in lungs harvested from IFNAR^−/− animals (Fig. 2⇑D) was significantly worse than that in wild-type or IFN-γ^−/− mice, the severity was consistently intermediate between that observed in wild-type relative to Stat1^−/− animals. Similar to the Stat1^−/− mice, the responding inflammatory cells in PR8 virus-infected IFNAR^−/− mice were primarily (>50%) granulocytes (Fig. 2⇑G). However, fewer infiltrating cells and less tissue destruction were observed relative to infected Stat1^−/− tissue (compare Fig. 2⇑B and Fig. 2⇑D). The histopathology of PR8 virus-infected lungs from IFNAR/GR^−/− mice was comparable with that seen in the Stat1^−/− animals (data not shown), suggesting that both IFN-α and IFN-γ contributed to regulation of the inflammatory response to influenza virus infection.

Humoral immune response to influenza virus infection in Stat1-deficient and wild-type mice

Serum samples were taken from wild-type and Stat1^−/− mice before and after i.n. inoculation with the attenuated influenza virus strain, NA/B-NS (17), and were assayed for the presence of virus-specific Abs. Influenza virus-specific Ig was detected at equivalent levels in animals of both genotypes (Fig. 3⇓A). Interestingly, whereas wild-type animals produced Abs against influenza virus of both IgG1 and IgG2a subclasses, no IgG2a virus-specific Abs were detected in serum from infected Stat1^−/− mice. Despite these differences in Ab isotype, both wild-type and Stat1^−/− mice displaying detectable anti-influenza virus Abs were fully protected from rechallenge with a lethal dose of WSN virus (data not shown).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3.

Enhanced Th2-type Ig production in Stat1^−/− CD1 mice. A, Influenza virus-specific Ig levels were measured by ELISA. Data represent averages of sera from three wild-type (□) or Stat1^−/− mice (▪). B, Circulating IgE levels are increased in Stat1^−/− mice. IgE levels in sera from naive (left) or influenza virus-infected mice (right) were measured by ELISA. The levels of IgG2a and IgE were significantly different between wild-type and Stat1^−/− (p < 0.01).

Heavy chain class switch to IgG2a is stimulated in response to IFN-γ (19), suggesting that the lack of this Ig isotype was the result of a direct loss of the ability of cells to respond transcriptionally to IFN-γ in the absence of Stat1. Preferential production of the IgG1 subclass of Abs is characteristic of B lymphocytes maturing under the influence of Th2 cells (20). A hallmark of Th2-influenced responses, which normally characterize allergic, antiparasitic, or autoimmune responses rather than antiviral ones, is the production of IgE. We detected abnormally high levels of total IgE in the serum of influenza virus-infected Stat1^−/− mice (Fig. 3⇑B), and even naive animals showed significant levels of circulating IgE. IgE levels were below the limit of detection (<0.04 ng/ml) in control animals, whether or not they were infected with virus. Thus, Stat1^−/− mice displayed a default Th2 bias that was greatly augmented by viral infection. Neither the enhanced production of IgE nor the complete suppression of IgG2a synthesis were noted in influenza virus-infected IFN-γ^−/− mice (18), although a heavy bias toward IgG1 was noted previously in IFNAR/GR^−/− mice (13). These results strongly suggest that type I IFN in addition to type II IFN controls heavy chain class switch recombination.

Th2 cytokine production in Stat1^−/− mice

T cell responses in the absence of IFN signaling were further characterized by comparison of wild-type and Stat1^−/− splenic lymphocyte differentiation. Th1 and Th2 T cell subsets are defined functionally in terms of the cytokines that they produce (21). The high levels of IgE found even in serum from naive Stat1^−/− mice suggested an intrinsic bias toward Th2-influenced responses. Therefore, we measured cytokines produced by spleen cell cultures isolated from uninfected wild-type and Stat1^−/− mice (Table I⇓). Splenocytes were stimulated in vitro through the TCR and allowed to differentiate for 1 wk in the presence of IL-2. Restimulation of Stat1^+/+ splenocytes led to exclusive production of IFN-γ. In contrast, significant amounts of IL-5 were secreted by rested Stat1^−/− splenocytes even in the absence of restimulation. Moreover, restimulation of Stat1^−/− cells produced significant production of the Th2-type cytokines IL-4, IL-5, and IL-10 not seen in wild-type cultures, in addition to IFN-γ.

This bias toward Th2 cytokine secretion in the presence of an ongoing Th1-like IFN-γ-producing response was also observed when cytokine transcripts from in vitro differentiated splenocytes were assayed by RNase protection. By this method, wild-type splenocytes were found to make mRNA corresponding to the expected IFN-γ and IL-2, as well as IL-10 (Fig. 4⇓A, lane 11). Stat1^−/− splenocytes, differentiated and assayed in parallel, produced transcripts corresponding to IL-5, IL-9, IL-10, and IL-13 as well as comparable amounts of IL-2 and IFN-γ (Fig. 4⇓A, lane 12). Cytokine mRNA profiles varied somewhat with strain background in these experiments. Lymphocytes harvested from wild-type 129 animals were much more Th0 in character than those from C57BL6 or CD1 mice after in vitro differentiation and restimulation with anti-CD3 and anti-CD28. Nonetheless, both Stat1^−/− (Fig. 4⇓A) and IFNAR^−/− (data not shown) splenocytes showed a marked Th2 bias compared with wild-type, strain matched controls.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4.

Distinctive lung cytokine profiles are induced by influenza virus infection in wild-type, Stat1^−/−, IFN-γ^−/−, and IFNAR^−/− mice as determined by RNase protection assays. A, Lung RNA samples from five wild-type (lanes 1–5) and 5 Stat1^−/− (lanes 6–10) CD1 animals 3 d after i.n. inoculation with PR8 virus. In vitro differentiation of splenocytes gave rise to Th1-type cytokine profiles in wild-type (lane 11) and Th2-type cytokine profiles from Stat1^−/− cells (lane 12). B, Lung RNA from 3 wild-type (lanes 1–3) and 3 Stat1^−/− (lanes 4–6) mice 6 days after i.n. inoculation with PR8 virus. C, Twelve IFN-γ^−/− C57BL6 mice were inoculated with PR8 virus. Samples for RNase protection assay were harvested 3 days (lanes 1–4), 6 days (lanes 5–8), and 9 days (lanes 9–12) after i.n. inoculation. D, Lung RNA samples from four wild-type 129 mice 3 days after PR8 virus infection (lanes 1–4). RNA from lungs of PR8 virus-infected IFNAR^−/− 129 mice harvested 3 days (lanes 5–7) and 6 days (lanes 8–10) postinfection. All infections were i.n. using 100 TCID of PR8 virus. RNase protection assays used 25 μg RNA samples and the PharMingen mCK-1 multiprobe template set. Unprotected probe is used as a size marker for each assay (M). E, Lymphocytes producing IFN-γ and IL-5 were scored by intracellular cytokine staining and for cell surface CD4 or CD8 phenotype by flow cytometry. The percentages of CD8 cells (upper) or CD4 cells (lower) producing IL-5 (upper left quadrants) or IFN-γ (lower right quadrants) are indicated.

The production of Th2-type cytokines by stimulated Stat1^−/− splenocytes in the presence of ongoing production of Th1-type cytokines such as IFN-γ suggested a mixed response, possibly due to the absence of IFN-γ-mediated inhibition. Because unfractionated splenocytes were used for the in vitro differentiation and restimulation, the production of IFN-γ could reflect contributions from cell types other than CD4 helper cells, such as CD8 or NK cells. To dissect the contribution of CD4 and CD8 cells to these cytokine profiles, we measured cytokine protein levels on a per cell basis by flow cytometry (Fig. 4⇑E). We scored IFN-γ-producing cells as a marker for Th1-type and IL-5-producing cells as a marker of Th2-type. Approximately one-half the differentiated wild-type CD4 cells produced IFN-γ and very few produced IL-5, indicative of a Th1 response. In contrast, few Stat1^−/− CD4 cells produced IFN-γ, whereas nearly 10% produced IL-5. The majority of CD8 cells of both genotypes produced IFN-γ. Therefore, the majority of Stat1^−/− CD4 cells would appear to be biased toward a Th2 response or at least against a Th1 response, and the vast majority of T lymphocyte-derived IFN-γ produced by Stat1^−/− cultures was not from helper T cells but rather from CD8 cells.

Cytokine production in infected lung tissue

We examined whether this Th2 bias observed with in vitro differentiated Stat1^−/− splenocytes was also present at the site of infection. Fig. 4⇑, A–D, shows cytokine mRNA profiles from lungs of wild-type, Stat1^−/−, IFN-γ^−/−, and IFNAR^−/− mice, harvested at various times after i.n. inoculation with PR8 virus. On day 3 of infection, lungs from wild-type animals (Fig. 4⇑A, lanes 1–6) showed strong induction of IL-6 and IL-15, and a clear but less robust induction of IFN-γ. By day 6 (Fig. 4⇑B), the IL-15 signal had begun to wane, and a delayed IL-10 induction was evident.

This pattern of cytokine gene induction, consistently present in each animal assayed, was distinct from that seen in lungs from PR8 virus-infected Stat1^−/− mice. In Stat1 mutant animals, IL-6 and IFN-γ were also induced strongly by day 3 (Fig. 4⇑A, lanes 6–10), similar to wild-type animals. However, missing from the Stat1^−/− profile at day 3 was the strong induction of IL-15 transcripts, but present was weak induction of the Th2 cytokines IL-5 and occasionally IL-13. Missing from the 6-day infected Stat1^−/− lungs was the induction of IL-10 seen in wild-type samples (Fig. 4⇑B). This pattern of cytokine expression within the infected tissue was quite different from that seen in differentiated splenocytes and could not be classified in terms of a simple Th1-Th2 dichotomy. Therefore, although the presence of Th2 cytokines may have contributed to disease exacerbation in Stat1^−/− animals, it is likely that Th2 cytokines alone were not sufficient and additional factors influenced the altered pathogenesis observed.

This conclusion was reinforced by examination of the cytokine response in virus-infected IFN-γ^−/− mice (Fig. 4⇑C). Animals were infected i.n. with PR8 virus, and groups of four were sacrificed at 3-day intervals and analyzed individually. Lung samples harvested at day 3 showed strong induction of IL-6 and IL-15 (Fig. 4⇑C, lanes 1–4). Similar to wild-type animals, IL-10 was induced by day 6. Unlike wild-type animals, however, IL-5 and IL-13 were also occasionally detected in lungs of animals harvested on day 6 (lanes 5 and 8) similar to the pattern observed in Stat1^−/− samples and indicative of the Th2 bias observed in the absence of IFN-γ. This result suggests that the presence of IL-5 and IL-13 were not sufficient to cause significant disease exacerbation, at least in the presence of an intact type I IFN response, and reinforces the conclusion that the mere presence of Th2 cytokines is insufficient to explain the altered pathology of virus-infected Stat1^−/− mice.

Fig. 4⇑D shows results of influenza virus infection in IFNAR^−/− mice. In PR8 virus-infected lungs harvested on day 3, IL-6 and IFN-γ transcripts were detected in all animals (Fig. 4⇑D, lanes 1–5). Juxtaposition of samples from wild-type, background-matched controls (Fig. 4⇑D, lanes 1–5) underscores the reduction in IL-15 induction in the absence of IFN-α responsiveness (Fig. 4⇑D, lanes 6–10). No increase in IL-5 or IL-13 was detected by RNase protection assay in these animals. IL-10 induction was present in the day 6 IFNAR^−/− lung samples, consistent with kinetics seen in wild-type animals. The increased IFN-γ mRNA synthesis seen in the IFNAR^−/− mice from 3 to 6 days (Fig. 4⇑D, lanes 6–10) was consistent between experiments and not seen with other mutant strains.

Macrophage cytokine production in vitro

We examined potential sources of IL-15 mRNA detected in influenza virus-infected lungs of wild-type and IFN-γ^−/− mice. Macrophages were derived from each mouse strain by culturing bone marrow for 7 days in L cell-conditioned media. Homogeneity of cultured cells was estimated to be between 80 and 90% by FACS analysis after staining for CD11b (data not shown). Monolayers of adherent macrophages were treated with LPS (5 μg/ml) or infected at high multiplicity with influenza virus A/WSN/33. RNA was isolated from each culture after 24 h and cytokine transcripts were analyzed by RNase protection. As expected (22, 23), IL-15 was strongly induced by LPS in wild-type macrophages derived from 129SvEv or C57BL6 mice (Fig. 5⇓, left). Cytokine profiles from IFN-γ^−/− cells resembled wild-type controls (Fig. 5⇓A, right). However, no IL-15 induction occurred in LPS-treated IFNAR^−/− or Stat1^−/− macrophages. IL-15 was also induced in wild-type and IFN-γ^−/− cultures in response to virus infection, although to a lesser extent than by LPS. It is likely that type I IFN induction by either stimulus mediates the increased transcription of IL-15. IL-10 induction paralleled that of IL-15 in response to LPS or virus, whereas IL-6 was induced in all strains regardless of genotype only in response to LPS. Thus, the prominent induction of IL-15 observed in infected lungs likely originated from macrophages, either resident or infiltrating. IL-10, found in infected lungs of all but Stat1^−/− mice, may have been elaborated by other cell types in addition to macrophages (24, 25) since IFNAR^−/− macrophages appear deficient in IL-10 production, and yet this cytokine was clearly induced in infected lungs (see Fig. 4⇑D).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 5.

Macrophage cytokine mRNA profiles. Bone marrow-derived macrophages were cultured from wild-type, Stat1^−/−, and IFN-γ^−/− C57BL6 mice (A) as well as wild-type, Stat1^−/−, and IFNAR^−/− 129 mice (B). Macrophages of each strain/genotype were left untreated or treated for 24 h with LPS (5 μg/ml) or 10⁶ PFU influenza virus A/WSN/33, as indicated. IL-15, IL-10, and IL-6 transcripts were strongly induced by LPS in wild-type and IFN-γ^−/− macrophages. Virus produced a more modest induction of IL-15 and IL-10. Only IL-6 production was induced in IFNAR or Stat1^−/− macrophages.