Infection-Triggered Regulatory Mechanisms Override the Role of STAT 4 in Control of the Immune Response to Influenza Virus Antigens

Abstract

Accurate control of the balance of the T1 and T2 cells during antiviral immunity is essential for optimizing immune effector functions and for avoiding potentially severe immunopathology. We examined the in vivo role of the signal transducer and activator of transcription (STAT) 4 in regulating the T1/T2 balance during the response to live influenza virus and isolated viral proteins. We found that the differentiation of gamma interferon (IFN-γ)-producing Th1 and Tc1 cells after inoculation of live virus occurred independently of STAT 4 expression. Influenza virus-specific T2 and Tc2 responses were well controlled in such STAT 4-deficient mice unless IFN-γ was eliminated as well. In contrast, the STAT 4-dependent signaling pathway played a more essential role in regulating the T1/T2 balance after immunization with viral proteins and, in particular, inactivated nonreplicating virus. Pulmonary infection was cleared even in the absence of both functional STAT 4 genes and functional IFN-γ genes because virus-neutralizing antibodies were still generated, consistent with a substantial redundancy in different antiviral effector pathways. Thus, replicating agents such as live influenza virus can elicit IFN-γ and control T2 immunity independently of STAT 4, whereas the profile of immunity to isolated proteins is more reliant on an intact STAT 4 signaling pathway.

Protective immunity against viruses is dependent on conventional major histocompatibility complex (MHC) class I- and II-restricted T cells that recognize specific epitopes initially presented on antigen-presenting cells. The nature of the specific epitopes and the degree of antigen-presenting cell activation subsequent to recognition of infection-associated danger motifs determine the magnitude and profile of the antiviral response. T helper 1 CD4 (Th1) and cytotoxic CD8 type 1 (Tc1) cells that are defined by production of cytokines such as gamma interferon (IFN-γ), tumor necrosis factor alpha, and lymphotoxin beta and can directly disable virus-infected cells are thought to be of paramount importance for protection against viruses (1). In addition, Th1 and interleukin-4 (IL-4)-producing Th2 cells regulate the expansion, differentiation, and isotype switching of B cells that produce virus-neutralizing antibodies. Considerable evidence to support this model comes from the study of immunity to one of the most common human pathogens, the influenza virus. Intriguingly, during flu infection, IFN-γ-producing Th1 and Tc1 as well as IL-4-producing Th2 cells are generated, suggesting that a balanced T1 and T2 cell line response is more favorable for effective recovery from virus infection (7). Whereas the T1 response generally enables viral clearance through IFN-γ-dependent and cytotoxic mechanisms, the T2 arm may limit the associated delayed-type hypersensitivity reaction (4) that can impair the functional capacity of the lung (16). In the absence of a functional IL-4 gene, the T1-orchestrated local response in the lung is overwhelming, leading to pronounced immunopathology despite successful clearance of influenza virus (4). Conversely, administration or overexpression of the T2 cytokine IL-4 in the lungs impairs the local CD8⁺ T-cell response (4, 22). Furthermore, in the absence of IFN-γ the protective memory against highly virulent shift variants is clearly reduced (5). In view of the above arguments, control of the T1/T2 balance during the response to influenza virus may be crucial for effective recovery from infectious disease.

The STAT 4 transcription factor can play a major role in the commitment and differentiation of T1 cells (17). STAT 4 is functionally located downstream of the transcription factor T-bet and has been shown to mediate IL-12-triggered production of IFN-γ by CD4⁺ Th cells (18, 32). Interestingly, STAT 4 seems to be more important for differentiation of Th1 than for Tc1 cells (8). The contribution of the IL-12-associated STAT 4 signaling pathway to protection against intracellular bacteria and protozoan pathogens has been previously reported (3, 30). However, it has been shown previously that STAT 4-independent mechanisms of T1 differentiation are functional under certain conditions (19).

The in vivo role of T-cell regulatory factors cannot be easily predicted based solely on in vitro experimentation. In addition, it is difficult to extrapolate from one class of pathogens to another in regard to the importance of various regulatory pathways, due to the complex and differential nature of danger motifs shaping the immune response. Thus, we asked the question of whether STAT 4 was important for determining in vivo differentiation of MHC class I- and class II-restricted T cells in response to live influenza virus compared to isolated viral antigens or, alternatively, whether STAT 4-independent mechanisms are sufficient. Furthermore, we wanted to know whether a functional STAT 4-dependent signaling pathway was a prerequisite for the development of protective anti-influenza virus immunity. In the present study, we addressed these issues by using mice lacking STAT 4. We found that replicating virus maintained a mostly unaltered T1/T2 balance by generating IFN-γ independently of STAT 4, whereas immunization with noninfectious influenza virus antigens resulted in a more pronounced T2 shift in the absence of STAT 4. After live virus infection, a similar T2 shift was observed only when both STAT 4 and IFN-γ were eliminated but even then the shift did not affect viral clearance, underlining an extensive redundancy at the level of antiviral effector pathways.

MATERIALS AND METHODS

Mice.

Gene-targeted mice (i.e., mice STAT 4 deficient on a BALB/c background [5] [generously provided by Michael Grusby, Harvard Medical School] and IFN-γ knockout mice [11]) and wild-type mice were housed and bred at La Jolla Institute of Allergy and Immunology under specific-pathogen-free conditions according to National Institutes of Health and Animal Care Committee regulations. The gene-targeted mice were genotyped by PCR using DNA extracted from tails and the following sets of primers: (i) for STAT 4-deficient mice on the H-2^d background, STAT 4 sense-STAT 4 antisense (5′-CCTACTGGCAGAGAGTCTTTTCC-3′ and 5′-GGTTGTAGATCAGGAAGGTAGC-3′) and STAT 4 sense-neosense (5′-CCTACTGGCAGAGAGTCTTTTCC-3′ and 5′-GGATTGCACGCAGGTTCTCCG-3′); and (ii) for IFN-γ-deficient mice on the H-2^d background, wild-type sense-antisense (5′-AGAAGTAAGTGGAAGGGCCCAGAA-3′ and 5′-AGGGAAACTGGGAGAGGAGAAATA-3′) and knockout sense-antisense (5′-TCAGCGCAGGGGCGCCCGGTTCTTT-3′ and 5′-ATCGACAAGACCGGCTTCCATCCGA-3′).

The STAT 4 × IFN-γ doubly deficient mouse strain was generated by crossing STAT 4 with IFN-γ-deficient homozygotes. The F₁ mice were intercrossed, and the F₂ gene-deficient mice were selected. In all experiments, we used gene-deficient mice and gene-competent littermates (gender matched) between 8 and 12 weeks old.

Virus, antigens, and immunization.

For immunization and viral challenge, we used the A/WSN/32 H1N1 strain of influenza virus. Viral antigen was prepared by exposure of sucrose gradient-purified WSN virus to short-wave UV light until a complete loss of infectious activity was achieved. The virus was quantified by standard hemagglutination of chicken red blood cells or by measurement (using the biuret reaction) of total protein. Additional antigens used in the experiments were recombinant immunoglobulins (Igs) (mouse IgG2b) bearing dominant MHC class I peptide (nucleoprotein [NP] 147-155 [IgNP]) and class II peptide (hemagglutinin [HA] 110-120 [IgHA]) epitopes inserted within the CDR3 region of the V_H segment (34, 35). For peptides, we used the previously characterized dominant NP-derived, MHC class I-restricted K^d epitope (147-155), in addition to the dominant HA-derived, MHC class II-restricted I-E^d epitope (110-120). For antigen immunization, the UV-inactivated virus was suspended in sterile phosphate-buffered saline and administered intraperitoneally (20 μg/mouse). The recombinant immunoglobulins were emulsified in complete Freund's adjuvant (CFA) (1/1 [vol/vol]) and injected intraperitoneally (50 μg/dose/Ig construct).

Infectious challenge and measurement of viral titers.

For virus challenge, gene-targeted mice or competent littermates were infected under Metofane anesthesia with doses previously defined as sublethal (10% lethal dose [LD₁₀]) in wild-type mice. The dose corresponded to 10⁴ 50% tissue culture infective doses (TCID₅₀) of WSN virus, administered via the nasal route. At various intervals after infection, the mice were sacrificed and their lungs were harvested, homogenized, and stored at −70°C. The virus titers were measured by 48-h incubation of serial dilutions of homogenized tissue samples with permissive MDCK cells, followed by standard hemagglutination with chicken red blood cells. The endpoint titers were estimated in triplicate measurements by interpolation and expressed as TCID₅₀/organ. The titration of live WSN influenza virus grown on permissive MDBK cells prior to infectious challenge was carried out in a similar manner.

Measurement of antibody responses.

The antibody response was measured by enzyme-linked immunosorbent assay (ELISA) and hemagglutination inhibition (HI) assay. In brief, wells were coated with antigen (8 μg of sucrose-purified WSN virus/ml) and blocked with SeaBlock (Pierce, Rockford, Ill.). Serial dilutions of sera were incubated in coated wells for at least 2 h at room temperature. After washing, the assay was developed with 1:1,000 polyclonal goat anti-mouse IgG antibody coupled with alkaline phosphatase (Sigma) followed by the addition of pNPP substrate and measurement of optical density at 405 nm by using an automatic microtiter plate reader (ThermoMax; Molecular Devices) equipped with SoftMax software. Alternatively, for measuring the antibody isotypes, the wells were incubated with 1:250 biotinylated rat anti-mouse γ₁ or γ_2a monoclonal antibodies (BioSource International, Camarillo, Calif.) overnight at 4°C, followed by incubation with 1:1,000 streptavidin-alkaline phosphatase for 1 h at room temperature.

For the measurement of virus-neutralizing antibodies by standard HI assay, various dilutions of receptor-destroying enzyme (Neuraminidase; Sigma)-treated serum were incubated in 96-well flexible U-bottom plates with defined amounts of live virus diluted in phosphate-buffered saline. The amount of live virus used was determined by hemagglutination in pilot measurements. After 45 min of incubation of virus with diluted serum at room temperature, chicken red blood cells were added. The hemagglutination was read subsequent to an additional 1 h of incubation at room temperature. The HI titer was read as the highest dilution of serum that inhibited hemagglutination. Various controls (nonimmune sera and virus-free wells) were run in parallel. The measurements were done in triplicate.

Measurement of T-cell responses.

For the measurement of cellular responses, splenic cell suspensions were obtained by passing the organ through 70-μm-pore-size nylon Falcon strainers (Becton Dickinson) followed by hypotonic lysis of red blood cells. The lymphocytes from the pulmonary associated lymphoid tissue were isolated by collagenase digestion of lung tissue followed by Ficoll-Paque (Amersham Pharmacia) gradient centrifugation. The T-cell response was measured by enzyme-linked immunospot (ELISPOT) analysis as follows: 96-well 45-μm-thick mixed-cellulose ester plates (Millipore) were coated with 4 μg of rat anti-mouse anti-IFN-γ, IL-2, or IL-4 monoclonal antibodies (catalog no. 554430, 18161D, or 554387, respectively; BD-PharMingen)/ml. After blocking with 10% fetal calf serum in sterile saline for 1 h at 37°C, spleen cell suspensions were added at 5 × 10⁵ cells/well, with or without antigens or peptides. In the case of pulmonary lymphocytes, effector cells were mixed 1:1 with mitomycin-treated, splenic stimulator cells before stimulation. For stimulation, we used graded amounts of antigen (HA, NP peptides, or sucrose-purified WSN virus). At 72 h after initiation of cell culture, the assay was developed with biotinylated rat anti-mouse cytokine antibodies (BD-PharMingen) followed by streptavidin-horseradish peroxidase (BioSource International) and insoluble aminoethyl carbazole substrate. The results were measured using an automatic imaging system (Micromate; Navitar) equipped with multiparametric analysis software (Image Pro; Media Cybernetics). The measurements were carried out in triplicate.

In some cases, the cells were stimulated in regular 96-well microplates (5 × 10⁵ cells/200 μl/well) with or without recombinant IL-12 (1 ng/ml), supernatants were harvested, and cytokines were measured by ELISA using BioSource International kits.

Statistical analysis.

Magnitudes of immune responses were compared using t tests, assuming a normal distribution of the values and equal variances.

RESULTS

Role of STAT 4 in regulating the T1/T2 balance during infection with live influenza virus.

STAT 4 is an exclusive transcription factor that controls the IL-12-mediated induction of IFN-γ expression subsequent to the binding of the IL-12 receptor (IL-12R) by IL-12, both in T cells from spleen and pulmonary tissue (Table 1) (18, 32). In addition, STAT 4-mediated signaling is responsible for limiting IL-4 production in spleen and pulmonary tissue (Table 1). Thus, in the absence of functional STAT 4, the IL-12-induced IFN-γ production was abolished. In addition, IL-12 failed to elicit a decrease in IL-4 production by STAT 4-deficient lymphocytes (Table 1).

TABLE 1.

IL-12/STAT 4-dependent production of IFN-γ by lymphocytes from spleen and pulmonary tissue

Origin of lymphocytes (mouse group)	Concn (pg/ml) of:
	IFN-γ		IL-4
	Nile	rIL-12	Nil	rIL-12
Pulmonary tissuea
Wtc	31 ± 20d	875 ± 125	92 ± 27	27 ± 11
STAT 4−/−	5 ± 5	10 ± 10	54 ± 25	69 ± 19
Spleenb
Wt	30 ± 16	208 ± 50	48 ± 22	13 ± 7
STAT 4−/−	3 ± 3	6 ± 4	38 ± 15	33 ± 11

^aPulmonary lymphocytes were obtained by collagenase digestion of lung tissue and separation on a Ficoll gradient.

^bLymphocytes were prepared from spleen by hypotonic lysis of red blood cells.

^cWt (wild-type) mice were STAT 4+/+ littermates.

^dLymphocytes from naive mice were incubated for 24 h at 5 × 10⁵ cells/200 μl/well with or without optimal amounts of recombinant IL-12 (rIL-12) (1 ng/ml). The concentration of IFN-γ and IL-4 in supernatant was measured by ELISA, and the results are expressed as means ± SEM of triplicates experiments (in picograms per milliliter).

^eNil, culture medium only.

Since the differentiation of T1 cells can proceed via STAT 4-dependent and -independent pathways, we sought to investigate the role of STAT 4 signaling in the control of T-cell responses to influenza virus. To this end, we infected STAT 4-deficient (−/−) mice and STAT 4-competent (+/+ and +/−) littermates with a dose of influenza virus (strain A/WSN/32 H1N1) that is minimally sublethal (LD₁₀) in wild-type mice. Similar to the +/+ and +/− counterparts, the STAT 4−/− mice mounted strong class I- and class II-restricted, virus-specific T-cell responses, easily detectable beyond day 7 postchallenge (Fig. 1 and data not shown). The STAT 4-competent homozygote and heterozygote mice displayed similar cytokine profiles of virus-specific T cells. The STAT 4−/− mice displayed some reduction of NP-specific IFN-γ-producing Tc1 cells (Fig. 1A) and HA-specific IFN-γ-producing Th1 cells in spleens (Fig. 1B). However, T2 responses were generally not augmented. In particular, when overall T-cell immunity was assessed after stimulation in vitro on UV-inactivated virus (displaying most epitopes), no significant alteration of T1/T2 responses was noted (Fig. 1C). Whereas in the absence of functional STAT 4, the generation of HA-specific IL-4-producing Th2 cells was enhanced in both spleen and pulmonary tissue (Fig. 1B), the in vitro T1 recall responses to an extensive set of class II-restricted dominant and subdominant epitopes (whole UV-inactivated WSN virus) were not substantially different in STAT 4-competent and -deficient mice (Fig. 1C). Thus, the STAT 4 signaling pathway was not essential for in vivo differentiation of class I- and class II-restricted T cells committed to Tc1 and Th1 phenotypes, respectively (Fig. 1A to C), after infection with influenza virus.

FIG. 1.

The balance of T1/T2 T-cell responses to live influenza virus in the absence of functional STAT 4 genes. Female STAT 4−/− and STAT 4-competent (wt) littermates were infected via the intranasal route with a sublethal dose of 10⁴ TCID₅₀ of A/WSN/32 H1N1 influenza virus. At 4 weeks after infection, the mice were sacrificed and the frequency of cytokine-producing T cells reacting to MHC class I-restricted NP-K^d (panel A) and MHC class II-restricted HA-I-E^d (B) peptides or whole UV-inactivated WSN virus (C) was evaluated by ELISPOT analysis. The data are expressed as the number of spot-forming cells (SFC)/organ after subtraction of background (means ± standard errors of the mean [SEM], n = 3/group; data are representative of two independent experiments). (D) Frequency (SFC/10⁶ responder splenocytes [SFC/10^6 R]) of IL-4-producing cells from STAT 4-deficient or STAT 4/IFN-γ doubly deficient mice treated similarly. The values significantly different from those corresponding to control mice are marked with an asterisk.

We hypothesized that a STAT 4-independent pathway responsible for induction of IFN-γ during infection with influenza virus provides additional control over the magnitude of the T2 response. To address this question, we generated and evaluated doubly deficient (STAT 4−/− IFN-γ−/−) mice. Indeed, the induction of IL-4-producing Th2 and, surprisingly, of Tc2 cells also was profoundly amplified when generation of both STAT 4-dependent and -independent IFN-γ was inhibited (Fig. 1D). Together, the data show only a minor role for STAT 4 in fine-tuning the T1/T2 balance during infection with influenza virus in vivo. In contrast, the control of the T2 component during influenza virus infection is heavily dependent on STAT 4-independent IFN-γ production.

Role of STAT 4 in regulating the T-cell profile during the response to noninfectious influenza virus antigens.

Since double-strand RNAs act as “danger motifs” during infection with RNA viruses such as influenza virus (33), possibly bypassing the requirement for a functional IL-12/STAT 4 pathway, we speculated that STAT 4 would exert a higher degree of control over T-cell responses generated after immunizing with viral antigens or UV-inactivated, nonreplicating virus. As shown in Fig. 2A, the generation of both NP-specific Tc2 and whole-virus-specific Th2 cells was significantly (three to fivefold) increased in the absence of a functional STAT 4 signaling pathway, in contrast to a much more limited increase (a twofold increase with HA and no increase with UV-inactivated virus) measured subsequent to viral infection (Fig. 1). Inactivated influenza virus may retain limited fusogenic ability, resulting in cytoplasmic delivery of antigens. Thus, immunization with inactivated virus triggers modest Tc1 responses (Fig. 2A). However, in contrast to viral infection, the limited Tc1 response triggered by inactivated virus to the dominant NP epitope was completely dependent on a competent STAT 4-signaling pathway (Fig. 2A). We next tested whether the STAT 4 signaling pathway would control the cytokine profile of class I- and class II-restricted T-cell responses to the dominant NP- and HA-derived T-cell epitopes delivered as proteins alone (Fig. 2B). FcγR-mediated delivery of HA and NP inserted within the variable segment of an IgG2a (IgHA+IgNP [34, 35]) and emulsified in CFA resulted in dual priming of class II- and class I-restricted T cells (Fig. 2). The STAT 4 signaling pathway had a profound effect on the generation of all Tc MHC class I-restricted responses in this context (via cross-priming), as shown by the five- to eightfold increase in the generation of NP-specific IL-4-producing Tc2 in the STAT 4 knockout mice (Fig. 2C). In addition, the STAT 4 signaling pathway was responsible for limiting the HA-specific Th2 response as well as the overall T2 population reactive to whole virus (Fig. 2D and E). As shown by the measurement of virus-specific Tc and Th cells producing IL-2, STAT 4 had little if any effect upstream of T-cell commitment to T1 or T2 cells (Fig. 2). Thus, in the absence of viral infection, the STAT 4-signaling pathway exerts major control on the T1/T2 balance during the immune response to viral antigens. Conversely, viral infection triggers STAT 4-independent mechanisms that mostly override the control of T-cell responses to class I- and class II-restricted epitopes exerted via the STAT 4 signaling pathway.

FIG. 2.

Alteration of the T1/T2 T-cell profile subsequent to immunization with nonreplicating influenza virus antigens in the absence of functional STAT 4 genes. STAT 4−/− or -competent (wt) littermates were immunized intraperitoneally with 50 μg of UV-inactivated WSN virus (A and B) or a mixture of 50 μg of UV-inactivated WSN virus plus 50 μg of recombinant IgG encompassing the major NP and HA class I- and II-restricted epitopes, respectively (50 μg of IgHA + 50 μg of IgNP), emulsified in CFA (C to E). The mice were sacrificed 4 weeks after immunization, and the number of cytokine-producing T cells reacting with NP- and UV-inactivated virus (in the case of nonlive virus immunization) or reacting with NP, HA, or live WSN virus (in the case of recombinant IgG immunization) was measured by ELISPOT analysis. The results are expressed as means ± SEM of SFC/spleen (n = 4/group; data are representative of two independent measurements). The values significantly different from those corresponding to control mice are marked with an asterisk.

The effect of STAT 4 and IFN-γ on the antibody response and clearance of influenza virus.

We next tested whether the lack of STAT 4 and IFN-γ during influenza virus infection would have effects that extend beyond the differentiation of cytokine-producing class I- and II-restricted T cells. Upon infection with the same sublethal dose of WSN influenza virus, all mice mounted virus-specific IgG responses even in the absence of functional STAT 4 or IFN-γ genes (Fig. 3A). In addition, virus-specific IgG1 and IgG2a antibodies were easily detected, indicative of simultaneous generation of T2 and T1 cells, respectively. Consistent with the T-cell response data, STAT 4−/− and, particularly, STAT 4−/− IFN-γ−/− double-knockout mice displayed moderately increased IgG1 antibody levels and a higher IgG1/IgG2a ratio (Fig. 3A). Comparing IgG isotype patterns and T-cell profiles supports the notion that STAT 4 controls B-cell differentiation and isotype switching indirectly through the differentiation of T cells into T1 and T2 subsets.

FIG. 3.

Control of isotype switching and pulmonary virus clearance by STAT 4 and IFN-γ. (A) Gene-targeted STAT 4 and IFN-γ singly or doubly deficient mice and gene-competent littermates were infected with nonlethal doses of strain WSN influenza virus. The IgG, IgG1, and IgG2a antibody responses were measured by ELISA applied to serum samples. The results are expressed as means ± SEM of endpoint titers (n = 4/group) at 4 weeks after infection. The data shown here are similar to titers measured at 1 and 3 weeks after infection. In addition, we expressed the ratio between mean titers of IgG1 and IgG2a antibodies for each mouse strain (right y axis). (B) Gene-targeted mice deficient in STAT 4 and/or IFN-γ genes and gene-competent littermates were infected via the respiratory tract with 10⁴ TCID₅₀ of WSN influenza virus. At 7 and 14 days postinfection, the mice were sacrificed and pulmonary virus titers were measured individually. Results are expressed as means ± SEM of total pulmonary virus (TCID₅₀) for each group (n = 4/group). Titers below 80 were indistinguishable from zero. The results are representative of two independent experiments.

Since STAT 4 and IFN-γ regulate induction of antiviral protective mechanisms, we next tested whether the STAT 4-deficient mice can clear pulmonary infection with influenza virus. As shown in Fig. 3B, STAT 4/IFN-γ doubly deficient mice completely cleared pulmonary virus within 2 weeks after infection with a challenge dose that causes only minimal lethality in STAT 4- and IFN-γ-competent mice as well. We next tested whether the humoral response in the absence of functional STAT 4/IFN-γ genes still included generation of virus-neutralizing antibodies. As shown in Table 2, the STAT 4−/− mice mounted similar titers of virus-neutralizing antibodies within 7 days after challenge. Additional IFN-γ deficiency did not have a measurable impact on the development of virus-neutralizing antibodies. Thus, induction of neutralizing antibodies paralleled the development of the whole IgG antibody response rather than being dependent on a defined IgG1/IgG2a ratio. This strongly suggests that humoral immunity is critical for the ability of mice to clear influenza virus infection even in the presence of a pronounced T2 bias, as observed in doubly deficient STAT 4−/− IFN-γ−/− mice.

TABLE 2.

Limited impact of STAT 4 and IFN-γ deficiency on the generation of influenza virus-neutralizing antibodies

Genotype	No. of neutralizing antibodiesa
STAT 4+ IFN-γ+	446 ± 206
STAT 4− IFN-γ+	240 ± 138
STAT 4+ IFN-γ−	512 ± 238
STAT 4− IFN-γ−	346 ± 162

^aMice were infected with a dose of 10⁴ TCID₅₀ of influenza virus, and at day 7 after challenge, the neutralizing antibodies were assessed by HI assay. Results are expressed as means ± SEM of hemagglutination-inhibiting titers (n = 5/group; results are representative of two independent measurements).

DISCUSSION

In vivo regulation of the T1/T2 balance during immune responses to microbes may be important for achieving effective recovery from infection while avoiding excessive immunopathology. The T1 arm is thought to be required for effective clearance of intracellular pathogens by facilitating the differentiation of appropriate effector cells and isotype switching to opsonizing or complement-fixing antibodies (1). Defects in T1 immunity result in clinically significant immunodeficiency syndromes (12, 26). In contrast, unregulated T1 responses may cause organ-associated immunopathology or autoimmune reactions (29). The induction and regulation of T1 immunity are dependent on multiple molecular checkpoints (28), including some T2-associated molecules (4).

In vivo studies with gene-targeted mice have shown that the IL-12/STAT 4 signaling pathway is important for effective control of intracellular nonviral pathogens (3, 30, 31). Nevertheless, little is known in regard to its role in in vivo control of T1/T2 immunity during the response to viruses. Influenza virus is an enveloped, negative-single-stranded RNA orthomyxovirus that is a common human pathogen (10). The virus replicates productively in furin-positive epithelial cells of the respiratory tract, resulting in significant tissue damage caused directly by the virus and through associated immunopathology (10). During the normal response against influenza virus, both T1 and T2 cells specific for class I- and class II-restricted epitopes are generated (6). Whereas T1 cells are thought to orchestrate protective immunity against the virus, the T2 component is thought to limit the lung-associated pathology that accompanies infection with influenza virus (4).

During infection with influenza virus, there is a rapid induction of IL-12 (21). Therefore, we explored the role of the STAT 4 signaling pathway in regulating the T-cell profile during influenza virus infection by using gene-targeted mice that lack expression of functional transcription factor. Interestingly, mice defective in STAT 4 expression showed only subtle shifts in T1/T2 immunity (mostly toward HA-specific immunity but not toward overall UV-inactivated viral immunity; Fig. 1) after infection with replicating influenza virus. We observed a modest decrease of NP- and HA-specific Tc1 and Th1 immunity and a slight enhancement of HA-specific Th2 immunity (Fig. 1A and B). This shows that IL-12/STAT 4-independent pathways are sufficient for triggering IFN-γ and T1 immunity during influenza virus infection. In addition, the IL-12/STAT 4 signaling pathway participates only peripherally in limiting the T2 component. Indeed, additional inhibition of the IFN-γ gene resulted in substantial expansion of the T2 population after flu infection (Fig. 1D), demonstrating a potent STAT 4-independent, IFN-γ-dependent regulatory mechanism that controls the profile of T-cell response to both MHC class I- and class II-restricted epitopes in vivo. This may be the main in vivo role of IFN-γ during the infection with influenza virus, since in contrast to Tc1 effectors, the Tc2 cells are not effective in clearing infectious virus (9). Interestingly, a more pronounced role of the IL-12/STAT 4 signaling pathway in controlling the differentiation of T1 cells and limiting the T2 component was evident when the immunogen was nonreplicating; that is, UV-inactivated influenza virus or subunit antigen formulated in CFA is known to induce IL-12-promoting danger signals (2) that are critical to its ability to elicit T1 responses in BALB/c mice, as demonstrated by using IL-12 knockout mice (13). Sialoprotein- or FcγR-mediated delivery of viral epitopes resulted in cross-processing, manifested by the induction of MHC class I-restricted cytokine-producing T cells along with the generation of Th cells (Fig. 2). Under these conditions, STAT 4-deficient mice displayed significantly increased Th2 and Tc2 responses and a substantial shift of the T1/T2 balance (Fig. 2). Thus, the importance of the IL-12/STAT 4 pathway in regulating the T-cell profile is critically dependent on the nature of antigens and the context of immunization. Employment of IL-12/STAT 4-independent pathways of IFN-γ production during influenza virus infection limits the impact of STAT 4 signaling in the regulation of the T1/T2 profile. Concordant with this, STAT 4 had only a limited effect on isotype switching during infection with influenza virus, mostly by limiting the γ₁/γ₂a ratio in conjunction with STAT 4-independent IFN-γ (Fig. 3).

In vitro studies employing ectopic expression of transcription factors or study of T-cell lines from gene-targeted animals showed that IFN-γ triggers T-bet expression via a STAT-1-mediated pathway (20). T-bet, in turn, enables localized chromatin remodeling and IL-12-mediated STAT 4 activation by genetically regulating Hlx and IL-12β₂R (23, 24). These events profoundly enhance transcriptional activation of IFN-γ, resulting in a positive feedback. However, our results support the notion that STAT 4-independent mechanisms allow substantial differentiation of T1 cells during influenza virus infection. During infection with influenza virus, innate immune cells are exposed to heterogeneous danger motifs borne by double-stranded RNA that bind to TLR-3 and its isoforms (33). Mediators that are in turn produced by innate immune cells, such as IFN-α/β, tumor necrosis factor alpha, and possibly IL-18 and IL-27 (23), may promote the induction of IFN-γ in a STAT 4-independent manner. STAT-1/T-bet-dependent but STAT 4-independent signaling may promote differentiation of T1 cells in such complex circumstances. Previous in vitro studies supported the existence of a T-bet-dependent, STAT 4-independent T1 differentiation pathway (27). Alternatively or in addition, T-bet- and STAT 4-independent mechanisms may be triggered in specific circumstances associated with infection. Recent studies show that the differentiation of T1 cells during influenza virus infection largely proceeds in a manner dependent on T-bet but independent of STAT 4 signaling (M. von Herrath and A. Bot, unpublished data).

In the present study, we examined the requirement for STAT 4 expression in regulation of the clearance of influenza virus subsequent to respiratory infection. Surprisingly, both gene-targeted and -competent control mice infected with doses of infectious virus that are sublethal in wild-type mice were able to completely eliminate the virus and recover from disease (Fig. 3). Further inhibition of the IFN-γ gene did not impair the virus clearance in all conditions studied. This contrasts with the protective effect of IFN-γ during secondary, heterologous challenge with a lethal dose of virus (5). Concordant with these findings, STAT 4 was not required for the generation of influenza virus-neutralizing antibodies (Table 2). This may explain the ability of gene-targeted mice to completely clear sublethal inocula, consistent with previous observations in IFN-γ-deficient mice (14).

In conclusion, our data underline a significant redundancy of protective mechanisms to a common viral pathogen. This contrasts sharply with the stronger dependency of protection after infection with nonviral intracellular pathogens on STAT 4 (11, 12, 22). Finally, we recently showed that STAT 4 expression was required for effective development of autoimmune diabetes but not protection against infection with lymphocytic choriomeningitis virus in a transgenic mouse model (15). However, it was recently shown that IFN-α/β-triggered STAT 4 expression is critical to the differentiation of IFN-γ-producing T1 cells during infection with choriomeningitis virus (25), contrasting dramatically with its more limited role in influenza virus infection. First, this underlines a differential involvement of transcription factors in the regulation of T1/T2 balance during infection with different categories of viruses. Second, these results support the existence of STAT 4-dependent, IL-12-independent (as well as STAT 4-independent) pathways of T1 differentiation.

Together with these previous studies, the data in the present report support the concept that in vivo targeting of defined transcription factors such as STAT 4 alters the pathogenesis of inflammatory conditions while leaving redundant protective mechanisms against common pathogens still intact, even when they are endowed with immune evasion strategies such as drift mutation and shift variation. In addition, multiple transcription factors and signal transduction pathways are likely simultaneously involved during viral infections, because they are part of a redundant network evolving from continuous interaction of the immune system with common pathogens and diversification of danger motif recognition strategies. On the other hand, protective mechanisms against less-common pathogens are more dependent on defined regulatory elements such as STAT 4, which is involved in the control of T-cell profiles. Finally, our results support the concept that precise transcriptional control of the T1/T2 profile during the response to common viruses such as influenza virus is more important for limiting organ-associated immunopathology than for determining the protective nature of the response.