The Interleukin-33–Group 2 Innate Lymphoid Cell Axis Represents a Potential Adjuvant Target To Increase the Cross-Protective Efficacy of Influenza Vaccine

ABSTRACT

Interleukin-33 (IL-33) is a multifunctional cytokine that mediates type 2-dominated immune responses. In contrast, the role of IL-33 during viral vaccination, which often aims to induce type 1 immunity, has not been fully investigated. Here, we examined the effects of IL-33 on influenza vaccine responses. We found that intranasal coadministration of IL-33 with an inactivated influenza virus vaccine increases vaccine efficacy against influenza virus infection, not only with the homologous strain but also with heterologous strains, including the 2009 H1N1 influenza virus pandemic strain. Cross-protection was dependent on group 2 innate lymphoid cells (ILC2s), as the beneficial effect of IL-33 on vaccine efficacy was abrogated in ILC2-deficient C57BL/6 Il7r^Cre/+ Rora^fl/fl mice. Furthermore, mechanistic studies revealed that IL-33-activated ILC2s potentiate vaccine efficacy by enhancing mucosal humoral immunity, particularly IgA responses, potentially in a Th2 cytokine-dependent manner. Our results demonstrate that IL-33-mediated activation of ILC2s is a critical early event that is important for the induction of mucosal humoral immunity, which in turn is responsible for cross-strain protection against influenza. Thus, we reveal a previously unrecognized role for the IL-33–ILC2 axis in establishing broadly protective and long-lasting humoral mucosal immunity against influenza, knowledge that may help in the development of a universal influenza vaccine.

IMPORTANCE Current influenza vaccines, although capable of protecting against predicted viruses/strains included in the vaccine, are inept at providing cross-protection against emerging/novel strains. Thus, we are in critical need of a universal vaccine that can protect against a wide range of influenza viruses. Our novel findings show that a mucosal vaccination strategy involving the activation of lung ILC2s is highly effective in eliciting cross-protective humoral immunity in the lungs. This suggests that the biology of lung ILC2s can be exploited to increase the cross-reactivity of commercially available influenza subunit vaccines.

INTRODUCTION

Current vaccination strategies against influenza are undoubtedly inadequate. Influenza viruses have high mutation rates, and consequently, vaccine strain mismatch occurs frequently. This issue, coupled with strain-specific immunity conferred by the influenza vaccine, necessitates annual updates of vaccine composition and yearly vaccination. However, even when circulating strains are well matched to the vaccine strain, the vaccine effectiveness is in the range of ∼40 to 60% (1). During mismatched seasons, vaccine effectiveness can be as low as 10%. Furthermore, seasonal influenza vaccines are thought to be ineffective against novel pandemic strains. Thus, a significant clinical need exists for a universal influenza vaccine.

The consensus in the vaccine field is that T helper 1 (Th1)-based vaccines are required for protection against intracellular pathogens and that T helper 2 (Th2)-based vaccines are important for extracellular pathogens. As such, inappropriate induction of Th2 immunity during viral infection is considered pathological and unwanted. Indeed, the failure of a respiratory syncytial virus (RSV) vaccine trial in 1967 was partly attributed to Th2 priming that led to an exaggerated Th2 response upon RSV infection (2). This failed vaccine trial has stigmatized Th2-biased immunity in the viral vaccine field. It is therefore not surprising that generally, efforts to develop a universal flu vaccine have focused on designing influenza vaccines that can promote type 1 or type 1/type 2 balanced immunity.

Interleukin-33 (IL-33) is a tissue-derived alarmin that can potently induce type 2 immune responses via the activation of ST2 (IL-33 receptor)-expressing immune cells such as Th2 cells and group 2 innate lymphoid cells (ILC2s). Respiratory influenza virus infection is known to elicit such danger signals (3, 4). Recent studies by us and others have begun to address the importance of IL-33 and IL-33-dependent ILC2s during influenza virus infection: ILC2s were shown to promote recovery from primary influenza virus infection by restoring airway epithelial integrity and tissue homeostasis (5, 6). The pathological role of IL-33 during influenza virus infections has also been demonstrated: IL-33 and ILC2s were shown to promote influenza-induced exacerbation of lung disease in a mouse model of allergic lung inflammation (7, 8). Consistent with findings in animals, genome-wide association studies have identified several genes in the IL-33–ILC2 axis as susceptible markers for allergic diseases (9–12). Clinical studies in patients with allergic diseases have observed elevated protein levels of IL-33 (13, 14). Furthermore, experimental viral challenge in asthmatic patients has been shown to trigger IL-33 and Th2 cytokine responses that correlate with the severity of asthma exacerbation (15). Presumably as a result of upregulated IL-33 signaling, ILC2s found in asthmatic patients exhibit increased numbers and a heightened active state (16–19). Thus, substantial evidence exists suggesting a detrimental role of IL-33 and ILC2s during viral infection and Th2-mediated diseases. Recent studies have begun to address the specific role of IL-33 in driving antiviral immunity (20–22); however, it remains unexplored whether the IL-33–ILC2 axis contributes to the development of cross-reactive immunity against influenza.

In the present study, we investigated the potential impact of IL-33 on influenza vaccine-induced immunity. Using a mouse model of influenza, we show that IL-33 significantly augments the cross-reactivity of commercially available influenza vaccines in an ILC2-dependent manner. We demonstrate that intranasal (i.n.) administration of an influenza vaccine together with IL-33 activates lung ILC2s at the time of immunization, resulting in a Th2 cytokine environment that promotes the production of mucosal antibodies, including IgA, which is primarily responsible for heterologous immunity. Our findings therefore suggest that lung ILC2s participate in the induction of IgA-dependent cross-protection against influenza.

RESULTS AND DISCUSSION

IL-33 potentiates inactivated influenza vaccine immunogenicity.

To investigate the role of IL-33 in vaccine-induced immunity, we utilized a commercially available inactivated influenza vaccine, Fluzone [Sanofi Pasteur’s influenza A (H1N1) 2009 monovalent vaccine] as a vaccine antigen since this vaccine does not trigger IL-33 responses (Fig. 1A) and is generally considered to be poorly immunogenic when delivered intranasally (i.n.) (23). Therefore, this i.n. influenza vaccination model allows the evaluation of the immunological effects of exogenous IL-33 on vaccine efficacy (Fig. 1B). This vaccination model demonstrates that even with two doses, Fluzone i.n. vaccination fails to provide protection against homologous infection with the H1N1 A/California/04/2009 (CA04) strain, whereas the incorporation of IL-33 is highly protective (see Fig. S1A to D in the supplemental material). Consistent with the survival data, weight loss patterns were similar in Fluzone-vaccinated mice and mock phosphate-buffered saline (PBS)-vaccinated mice. The observed protection was not due to the long-lasting effects of IL-33 on innate immunity because IL-33 administration without Fluzone did not confer protection (Fig. 1C to E). These results show that inactivated antigens administered i.n. are poorly protective unless supplemented with mouse IL-33. An IL-33 dosing analysis revealed that a single vaccination with the lowest dose tested for IL-33 (0.005 μg) was sufficient to provide complete protection against lethal homologous CA04 infection (Fig. 1C to E). The observed protection was associated with enhanced viral clearance (Fig. 1F). We next tested the potency of protective immunity by challenging mice with viral doses of up to 4 × 10⁵ PFU of CA04, which is at least 200× the 50% lethal dose (LD₅₀). Control mice that received Fluzone alone progressively lost weight, and all succumbed to homologous infection with a low lethal dose of 4 × 10³ PFU (Fig. S1E to H). In contrast, mice that were vaccinated with Fluzone plus IL-33 (Fluzone+IL-33) showed initial weight loss when challenged with high lethal doses of CA04, but all regained their body weight and fully recovered. An ideal pandemic influenza vaccine must be able to rapidly induce protective immunity. We therefore conducted a short-term protection study. Mice vaccinated with Fluzone+IL-33 were fully protected against CA04 infection (Fig. S1I and J). Thus, we conclude that IL-33 signaling can strongly enhance vaccine-induced strain-specific immunity.

FIG 1.

IL-33 potentiates inactivated influenza vaccine immunogenicity. (A) Protein levels of IL-33 in the BALF at day 1 after i.n. vaccination (n = 5 mice per group). (B to F) Schematic diagram of the influenza vaccination and homologous CA04 infection protocol (B), mortality (C) and morbidity (D and E) of CA04-infected mice (n = 8 mice per group), and pulmonary viral burdens at day 3 postinfection (n = 4 to 5 mice per group) (F). (G) Diagram depicting a vaccination-heterologous challenge model. (H to J) BALB/c mice were i.n. immunized once (H and I) or twice (J) and infected with H1N1 heterologous strain PR8 virus (n = 8 to 9 mice per group). Infected mice were monitored for mortality and morbidity. Maximum weight loss was calculated for statistical analysis. Data are representative of results from two independent experiments. (K and L) Levels of cytotoxicity (K) and total protein (L) in the BALF on day 6 after PR8 challenge (8 mice per group). Results are representative of data from two (A and H to J) or three (K and L) independent experiments. The center and error bars show the means and SD. Statistical significance was determined by one-way ANOVA with Bonferroni’s multiple-comparison test (A, E, and F), a log rank Mantel-Cox test (C, H, and J), or an unpaired two-tailed t test with Welch’s correction (I, K, and L). *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001. OD₄₉₀, optical density at 490 nm.

The current influenza vaccination strategies provide suboptimal protection against mismatched seasonal influenza virus strains and provide no protection against novel pandemic strains (24). Therefore, we evaluated the impact of IL-33 on the cross-reactivity of Fluzone using another strain of H1N1 influenza virus, A/PR/8/1934 (PR8) (Fig. 1G). The 2009 Fluzone formulation contains antigens derived from the H1N1 2009 pandemic strain, so it is expected to be ineffective against PR8 challenges. As expected, Fluzone alone did not provide any protection against the PR8 virus after single or double doses of immunizations (Fig. 1H to J). In striking contrast, Fluzone+IL-33 prevented significant mortality in mice infected with the PR8 virus. This was unexpected because it is generally believed that seasonal flu vaccines have only a limited, if any, capacity to provide cross-protection (1). The observed cross-protection was associated with reduced immunopathology (Fig. 1K and L). Cytotoxicity testing, which quantifies the extent of dying and damaged cells, revealed that pulmonary cytotoxicity was significantly reduced in Fluzone+IL-33-vaccinated mice following lethal heterologous PR8 infection (Fig. 1K). Similarly, the total protein level, which is another marker of lung injury, was decreased in Fluzone+IL-33-vaccinated mice (Fig. 1L). This cross-protection persisted for at least 24 weeks, suggesting that IL-33 can enhance not only cross-reactivity but also the longevity of antiviral immunity (Fig. S1K to N).

IL-33 promotes humoral mucosal immunity.

Inactivated influenza vaccines, such as Fluzone, are thought to provide immunity primarily by inducing antibodies. Therefore, we next determined the impact of IL-33 on humoral immunity. Analysis of bronchoalveolar lavage fluid (BALF) for Fluzone-specific antibodies after vaccination revealed that IL-33 increased IgG, IgA, and IgM levels (Fig. S1O and P). The enhanced mucosal antibody response was observed using the i.n. route of administration but not with intramuscular (i.m.) vaccination. Consistent with this finding, the same vaccine formulation failed to cross-protect when administered via the i.m. route (Fig. S1Q). Thus, IL-33 signaling selectively enhances humoral immune responses to mucosally delivered antigens. The enhanced antibody titers persisted after infection, and this was correlated with enhanced B cell numbers in the airway (Fig. 2A to C). Consistent with increased local B cell responses, the viral burden was significantly reduced in Fluzone+IL-33-vaccinated mice (Fig. 2D). To determine whether antibodies and B cells play a role in our vaccination model, we depleted B cells at the time of vaccination (Fig. S2A). CD20⁺ B cell depletion led to a marked, albeit not complete, reduction in Fluzone-specific serum antibody levels, and this was associated with a moderate reduction in the survival rate (Fig. S2B to D). In contrast, CD8⁺ T cell depletion had no effect on survival, suggesting that memory CD8⁺ T cells do not play a significant role in cross-protective immunity in Fluzone+IL-33-vaccinated mice (Fig. S2E to G). Consistent with this finding, IL-33 did not upregulate the expression of granzyme B, a marker of cytotoxic T cell effector function, upon influenza virus infection (Fig. S2H). Since the beneficial effect of IL-33 on influenza vaccine-induced antiviral immunity was evident only with mucosal antigen delivery, we hypothesized that local IL-33 delivery may preferentially stimulate mucosal humoral immunity. IgA is the predominant antibody in mucous secretions, particularly in the upper respiratory tract (25), and previous studies have demonstrated the contributions of IgA to cross-protection against heterologous influenza virus infections (26, 27). Thus, we next determined the role of IgA in our vaccination model using IgA^−/− mice. Notably, Fluzone+IL-33 i.n. vaccination did not provide protection in IgA^−/− mice against heterologous PR8 challenge (Fig. 2E) despite having comparable Fluzone-specific IgG and enhanced IgM levels compared to wild-type (WT) control mice (Fig. 2F). The absence of IgA resulted in an increased viral burden compared to that in WT mice vaccinated with Fluzone+IL-33 (Fig. 2G). Consistent with the notion that mucosal, but not systemic, humoral immunity is important for cross-protection, passive serum transfer failed to protect against heterologous infections (Fig. S2I and J). Collectively, these results demonstrated that IL-33 enhances the cross-reactivity of inactivated influenza vaccines by eliciting mucosal IgA responses.

FIG 2.

IL-33 promotes humoral mucosal immunity. (A to D) BALB/c mice were i.n. vaccinated with Fluzone with or without IL-33 and infected with 500 PFU of PR8 virus 30 days after vaccination (A), and 6 days after infection, BALF samples were analyzed for levels of Fluzone-specific antibodies (B), CD19⁺ B cells (C), and viral burdens (D) (n = 8 mice per group). Data in panel B are representative of results from two independent experiments. (E) WT and IgA^−/− mice were immunized and PR8 infected for a 20-day survival analysis (n = 7 to 8 mice per group). (F and G) Antibody (ab) titers (F) and viral loads (G) in BALF samples were determined 6 days after infection (n = 5 to 6 mice per group). Results are representative of data from two (G) or four (E and F) independent experiments. The center and error bars show the means and SD. Statistical significance was determined by an unpaired two-tailed t test with Welch’s correction (B to D), one-way ANOVA with Bonferroni’s multiple-comparison test (F and G), or a log rank Mantel-Cox test (E). *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001.

Given the known association between Th2 cytokines and IgA responses (28–30), the IL-33-mediated enhancement of IgA responses to vaccine antigens may involve the induction of Th2 cytokines. Indeed, cytokine analysis showed that i.n. Fluzone+IL-33 vaccination promoted a type 2 cytokine environment with a statistically significant increase in Th2 cytokine levels (Fig. S3A and B). Type 2 cytokines possess diverse immunomodulatory functions (31). For example, IL-5 plays a role as a maturation factor to promote IgA responses (28). Similarly, IL-13 acts at different stages of the B cell maturation pathway (29, 30). Thus, we next examined the importance of Th2 cytokines in the enhanced mucosal antibody responses in Fluzone+IL-33-vaccinated mice by treating IL-4/13-deficient mice with an anti-IL-5 antibody at the time of vaccination (Fig. 3A). IL-4/13^−/− mice that received the IL-5 neutralizing antibody had significantly lower Fluzone-specific mucosal antibodies, including IgA, than in IgG-treated controls (Fig. 3B), thus establishing a causal relationship between Th2 cytokines and upregulated antibody responses.

FIG 3.

IL-33 promotes mucosal immunity in an ILC2-dependent manner. (A) Schematic of anti-IL-5 mAb treatment of WT and IL-4/13^−/− mice vaccinated with Fluzone with or without IL-33. (B) Fluzone-specific antibodies were measured in BALF samples from vaccinated mice at week 4 postvaccination (n = 5 to 9 mice per group). (C) Schematic for BALF collection from vaccinated WT and ILC2-deficient mice. (D) HAI titers of BALF samples were determined prior to infection (4 to 6 mice/group). (E and F) BALF samples were analyzed for Fluzone-specific antibodies before (E) and after (F) PR8 infection (n = 4 to 6 mice per group [D and E] and 7 to 8 mice per group [F]). (G) CD19⁺ B cell expression in BALF on day 6 after PR8 (500 PFU) challenge (n = 7 to 8 mice per group). (H) Survival and weight loss of naive mice i.n. inoculated with immune BALF together with 250 PFU of PR8 (n = 5 mice per group). Representative results from two (B, F, and G), three (E), or four (H) experiments are shown. The center and error bars show the means and SD. The data were analyzed by one-way ANOVA with Bonferroni’s multiple-comparison test (B and D to G) or a log rank Mantel-Cox test (H). *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001.

IL-33 promotes mucosal immunity in an ILC2-dependent manner.

IL-33 signals through the ST2 receptor. In the airway, ST2 receptors are highly expressed on lung ILC2s, making this cell type a primary target for IL-33 (32, 33). Furthermore, ILC2s have been shown to be the major source of type 2 cytokines in response to activation by exogenous IL-33 (32, 34–37). Consistent with the literature (38, 39), intracellular cytokine staining showed that ILC2s highly upregulated the type 2 cytokines IL-5 and IL-13 upon i.n. vaccination with Fluzone+IL-33 (Fig. S3C to E). In addition, no IL-5-enhancing effect of IL-33 was observed in ILC2-deficient (Il7r^Cre/+ Rora^f/f [40]) mice immunized with Fluzone+IL-33 (Fig. S3F). These results suggest that ILC2-derived type 2 cytokines at the time of vaccination aid in the establishment of mucosal humoral immunity. To corroborate these results, we next determined whether ILC2s are required for the enhanced antibody responses observed in Fluzone+IL-33-vaccinated mice. ILC2 deficiency reduced hemagglutination inhibition (HAI) antibodies in the BALF of vaccinated mice (Fig. 3C and D). Furthermore, analyses of Fluzone-specific antibodies in the BALF before and after infection demonstrate that ILC2 deficiency completely abrogates the IL-33-mediated enhanced Fluzone-specific antibody responses (Fig. 3E and F). Similarly, mucosal CD19⁺ B cell expansion was also reduced in ILC2-deficient mice (Fig. 3G). To assess the functional impairment of mucosal antibody responses due to ILC2 deficiency, protective efficacies of passive immunization were compared between C57BL/6 WT and ILC2-deficient donor mice with a C57BL/6 background. Unlike immune BALF from WT mice, immune BALF collected from donor ILC2-deficient mice failed to protect naive recipient mice against heterologous PR8 infection (Fig. 3H). In summary, these results demonstrate that the adjuvantive effects of IL-33 on mucosal humoral immunity depend on ILC2s, and ILC2s promote humoral immunity via type 2 cytokine secretion. However, these results do not rule out other mechanisms by which ILC2s may promote humoral immunity. Recent studies have shown that ILC2s can transiently express major histocompatibility complex class II (MHC-II) and that MHC-II-expressing ILC2s can present antigens and interact with CD4⁺ T cells to synergistically promote type 2 immunity (32, 40, 41). Considering these findings, it is plausible that Fluzone+IL-33 vaccination upregulates MHC-II expression on ILC2s, which in turn play a critical role, together with type 2 cytokines, in initiating vaccine responses.

ILC2s potentiate cross-reactive immunogenicity of inactivated influenza vaccines.

To further establish that ILC2s are important for vaccine efficacy, we performed a series of influenza vaccination and challenge experiments, comparing WT and ILC2-deficient mice. ILC2-deficient mice showed significantly reduced survival compared to WT littermates during homologous CA04 and heterologous PR8 infections following a single i.n. Fluzone vaccination with IL-33 (Fig. 4A to C). Additional vaccination (i.e., two doses of i.n. vaccination) did not improve the survival rate in ILC2-deficient mice. Decreased protection was associated with higher viral loads in ILC2-deficient mice (Fig. 4D). It is likely that the uncontrolled viral replication in ILC2-deficient mice contributed to the observed increase in immunopathology (Fig. 4E) since a reduction in amphiregulin, a tissue repair cytokine, was not observed in ILC2-deficient mice (Fig. 4F). Taken together, we propose that ILC2s exert beneficial effects at the time of vaccination to potentiate adaptive immunity. However, it is unlikely that ILC2s directly contribute to protection during the effector phase since the interval between Fluzone+IL-33 vaccination and influenza virus challenge was up to 24 weeks (Fig. S1K to N), and ILC2s do not possess antigen-specific memory responses. In support of this, i.n. Fluzone+IL-33 vaccination did not provide protection in adaptive-immunity-deficient Rag2^−/− mice with an intact ILC2 compartment (Fig. S4A and B). Thus, the beneficial role of ILC2s in vaccine efficacy occurs at the time of immune priming via facilitating the establishment of humoral adaptive immunity.

FIG 4.

ILC2s potentiate cross-reactive immunogenicity of inactivated influenza vaccines. (A) Schematic of single- and two-dose Fluzone vaccination against homologous or heterologous challenge. (B and C) Vaccinated WT and ILC2-deficient mice were infected with CA04 (B) or PR8 (C) for survival analyses (n = 6 to 14 mice per group [B] and 4 to 15 mice per group [C]). (D to F) Levels of viral load (D), cytotoxicity (E), and amphiregulin (F) in BALF on day 6 after PR8 challenge (n = 7 to 8 mice per group). (G) Schematic of vaccination using a 2008–2009 seasonal trivalent flu vaccine against 2009 pandemic strain CA04 virus. (H and I) Vaccinated WT and ILC2-deficient mice were infected with heterologous CA04 virus for survival (H) and weight loss (I) analyses (n = 5 to 9 mice per group). Representative results from two experiments are shown (B [left], D to F, H, and I). The center and error bars show the means and SD. The data were analyzed with a log rank Mantel-Cox test (B, C, and H) or one-way ANOVA with Bonferroni’s multiple-comparison test (D to F and I). *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001.

Our data suggest that lung ILC2 activation, via exogenous IL-33 at the time of immunization, can significantly broaden the cross-protective efficacy of conventional influenza vaccines. To further assess the degree of IL-33–ILC2-enhanced cross-reactivity of anti-influenza immunity, we utilized 2008–2009 Flulaval, a seasonal trivalent influenza vaccine that failed to prevent the 2009 influenza pandemic (42). We found that Flulaval was cross-protective against the pandemic CA04 strain when IL-33 was codelivered i.n. (Fig. 4G to I). However, this effect was completely abolished in ILC2-deficient mice. In summary, these results suggest that ILC2s can potentiate the cross-reactivity of seasonal influenza vaccines to an extent that they can provide protection against pandemic influenza virus strains. This demonstrates the potential benefit of designing vaccines to exploit the biology of ILC2s.

There is evidence that ILC2s can acquire trained immunity. A recent report has shown that primed ILC2s respond more efficiently during secondary allergen exposures (43). This memory response is mediated by ILC2-activating cytokines and not antigens since ILC2s do not possess antigen-specific receptors. Therefore, reactivation of Fluzone+IL-33-primed ILC2s by epithelium-derived IL-33 during influenza virus infection may contribute to better survival. Indeed, Fluzone+IL-33 vaccination resulted in heightened type 2 cytokine responses during influenza virus infection in WT but not ILC2-deficient mice (Fig. S4C). Increased type 2 cytokines were also detected in Fluzone+IL-33-vaccinated IgA^−/− mice that succumbed to the infection (Fig. S4D). Thus, this finding suggests that type 2 cytokine responses, possibly mediated by ILC2s, during the effector phase do not contribute to the observed cross-protection, or they are insufficient in the absence of IgA to promote survival. Of note, in addition to the secretion of type 2 cytokines, ILC2s also secrete amphiregulin, a cytokine that has been shown to promote recovery from influenza (5). However, amphiregulin responses during influenza virus infection were suppressed in Fluzone+IL-33-vaccinated WT mice and enhanced in ILC2-deficient mice (Fig. 4F). Therefore, the possibility that i.n. Fluzone+IL-33 vaccination generates memory ILC2s that contribute to protection via amphiregulin and/or type 2 cytokine secretion during the effector phase of influenza virus infection is unlikely.

In conclusion, we have revealed a previously unappreciated capacity of ILC2s to guide broadly protective humoral responses to influenza vaccination. Specifically, we demonstrate that IL-33-mediated activation of ILC2s is a critical early event that is necessary for the establishment of mucosal humoral immunity, which in turn is responsible for cross-strain protection against influenza. Our findings are striking and counterintuitive given that the consensus in the vaccine field is that Th1-biased immunity is required for protection against intracellular pathogens and that Th2-biased immunity is important for extracellular pathogens. As such, the inappropriate induction of Th2 immunity during viral infections is considered pathological and unwanted. Indeed, excessive Th2 responses have been implicated in virus-induced asthma exacerbation and in vaccine-enhanced RSV disease (8, 15, 44). However, in our system, despite the induction of Th2-biased inflammation both at the time of priming (Fig. S3B) and during the effector phase (Fig. S4C and E), vaccinated mice were resistant to influenza-associated mortality and morbidity. Our data clearly demonstrate the benefit of IL-33–ILC2-driven Th2 priming in providing cross-protection against influenza. Thus, ILC2s represent a novel target for a mucosal vaccination strategy to facilitate the optimal activation of downstream adaptive immunity.

MATERIALS AND METHODS

Mice.

Specific-pathogen-free male and female BALB/c, C57BL/6, and Rag2^−/− (designated RAG2^−/−) mice (45), 6 to 8 weeks old, were purchased from The Jackson Laboratory (Bar Harbor, ME) or Charles River Laboratories (Raleigh, NC). BALB/c Il4^−/− Il13^−/− mice (46) and ILC2-deficient C57BL/6 Il7r^cre/+ Rora^fl/fl mice (40) were provided by Andrew N. J. McKenzie (MRC Laboratory of Molecular Biology, Cambridge, UK). BALB/c Il4^−/− Il13^−/−, C57BL/6 Il7r^cre/+ Rora^fl/fl, C57BL/6 Il7r^cre/+ Rora^+/fl, C57BL/6 Il7R^+/+ Rora^fl/fl, BALB/c Iga^−/− (47), and C57BL/6 Iga^−/− (47) mice were maintained at the Albany Medical College Animal Resources Facility. For experiments involving ILC2-deficient C57BL/6 Il7r^cre/+ Rora^fl/fl mice, C57BL/6 Il7r^cre/+ and C57BL/6 Il7R^+/+ Rora^fl/fl mice were used as WT littermate control mice. Animals were sex and age matched whenever possible. Most animals were used at 6 to 10 weeks of age. The mice were maintained in the Animal Resources Facility at Albany Medical College. All experimental procedures were performed according to the protocols approved by the Institutional Animal Care and Use Committee at Albany Medical College.

Mouse model of vaccination and influenza virus infection.

Mice were vaccinated by either the i.n. or i.m. route with 50 μl of the Fluzone 2009 monovalent formulation containing 2 μg of hemagglutinin (HA) derived from A/California/07/09 virus (Sanofi Pasteur, Lyon, France) or 50 μl of the Flulaval 2008–2009 trivalent formulation containing 1.5 μg of each of the HA antigens from influenza viruses A/Brisbane/59/07, A/Uruguay/716/07 NYMC X-175C, and B/Florida/4/06 (ID Biomedical Corporation of Quebec, Canada) in the presence or absence of mouse recombinant IL-33 (BioLegend, San Diego, CA). In some experiments, booster vaccinations were performed 2 or 4 weeks after the first vaccination. One to twenty-four weeks after the last vaccination, the mice were infected i.n. with either CA04 or PR8. Survival and weight loss were monitored daily. At the indicated time points after infection, BALF samples were harvested to determine viral titers via a plaque assay on Madin-Darby canine kidney cell monolayers.

BALF analysis.

At the indicated time points, BALF samples were harvested by lavaging the lungs with 1 ml of PBS. Cell-free BALF supernatants were analyzed for protein levels of interferon gamma (IFN-γ), IL-2, IL-4, IL-5, IL-6, IL-12, and IL-13 using a Bio-Plex mouse cytokine assay (Bio-Rad, Hercules, CA). Protein levels of amphiregulin, IL-33, granzyme B, IL-4, IL-5, IL-13, and IFN-γ were also quantified using commercially available enzyme-linked immunosorbent assay (ELISA) kits from Thermo Scientific (Rockford, IL), R&D Systems (Minneapolis, MN), and BD Biosciences (San Jose, CA). Cytotoxicity levels in the cell-free fraction of the BALF were measured using a CytoTox 96 nonradioactive cytotoxicity assay kit (Promega, Madison, WI). The concentration of total protein in the cell-free BALF was determined using a bicinchoninic acid (BCA) protein assay kit (Thermo Scientific, Rockford, IL).

Antibody analysis.

Immune sera and BALF were analyzed for the presence of vaccine-specific antibodies by an ELISA as described previously (48). Briefly, microtiter plates (96-well MaxiSorp; Nunc) were coated with 1 μg/ml of Fluzone (Sanofi Pasteur) at 4°C overnight. After washing four times with PBS plus 0.05% Tween (Sigma), the plates were blocked with PBS plus 10% fetal calf serum (FCS) for 1 h at room temperature. After washing, the wells were incubated with serial 2-fold sample dilutions for 90 min at room temperature. Bound antibodies were detected by incubation with biotin-conjugated goat anti-mouse antibodies specific for IgG, IgA, or IgM (Caltag Laboratories, Burlingame, CA) for 90 min, and horseradish peroxidase-conjugated streptavidin was then added (Biosource). After a 30-min incubation, the plates were extensively washed, and tetramethylbenzidine substrate (BD Biosciences) was added. The reaction was stopped by adding 2 N H₂SO₄, and the absorbance was measured at 450 nm using a Power-Wave HT microplate reader (BioTek Instruments). The antibody titer is expressed as the reciprocal dilution that gave 50% of the maximum optical density.

Hemagglutination inhibition assay.

BALF samples were tested in a standard hemagglutination inhibition (HAI) assay, as described previously (48). Briefly, serially diluted cell-free BALF samples were mixed with 4 hemagglutination units of PR8 virus in V-bottom 96-well plates. Following 30 min of incubation at room temperature, 0.5% chicken red blood cells (RBCs) (Lampire Biological Laboratories) were added and incubated for an additional hour at room temperature. The HAI titer was defined as the reciprocal of the last dilution that prevented hemagglutination activity.

Flow cytometric analysis.

BALF cells were harvested in 1 ml of PBS. Live BALF cells were enumerated based on trypan blue staining. Red-blood-cell-depleted BALF cells were incubated with eFluor 780 fixable viability dye (Invitrogen, Carlsbad, CA). Cells were then incubated with anti-mouse FcγIII/II receptor (2.4G2 monoclonal antibody [mAb]) to block nonspecific binding of mAbs to Fc receptors. Fc receptor-blocked cells were stained with a mixture of the following anti-mouse surface antigen mAbs: CD19 Brilliant Violet 421 (BV421) (clone 6D5; BioLegend), B220 phycoerythrin (PE)/Cy7 (clone RA3-6B2; eBioscience), and CD8 PE (clone H35-17.2; BD Biosciences). Stained cells were analyzed using a BD FACSCanto flow cytometer or an LSR II instrument (BD Biosciences).

Intracellular cytokine staining of lung ILC2s.

Harvested lungs were incubated in 1 ml of RPMI 1640 medium containing 0.1 mg/ml of Liberase, 0.1 mg/ml of DNase, and 5 mM MgCl₂ for 1 h with a constant rocking motion at 37°C. After digestion, red blood cells were depleted using RBC lysis buffer. Total lymphocytes were enumerated using trypan blue exclusion. For in vitro stimulation, cells were incubated with a mixture of 50 ng/ml phorbol myristate acetate (PMA) (Sigma-Aldrich, USA), 500 ng/ml ionomycin (Sigma-Aldrich, USA), and 1× brefeldin A (BioLegend, USA) in RPMI 1640 medium for 4 h. Following stimulation, cells were incubated with the fixable viability dye eFluor 450 (Invitrogen, USA), followed by FcR blocking with anti-mouse CD16/CD32 (BioLegend). Cells were subsequently incubated with surface antibodies for biotinylated anti-mouse-lineage cell-specific antibodies from BioLegend, CD45R/B220 (clone RA3-6B2), CD3 (clone 17A2), CD11b (clone M1/70), Ly6G/Ly6C (clone RB6-8C5), TER-119 (clone TER-119), CD49b (clone DX5), and CD19 (clone 6D5), followed by streptavidin BV510 (BioLegend). Cells were further stained with surface antibodies for ST2 fluorescein isothiocyanate (FITC) (clone DJ8; MD Bioproducts), CD90.2 PE/Cy7 (clone 53-2.1; eBioscience), KLRG1 peridinin chlorophyll protein (PerCP)-EF-710 (clone 2F1; eBioscience), and CD127 allophycocyanin (APC)-eFluor 780 (clone A7R34; eBioscience). For intracellular staining, cells were first fixed and permeabilized with a BD Cytofix/Cytoperm fixation/permeabilization solution kit (Fischer Scientific, USA), followed by incubation with anti-mouse IL-5 PE (clone TRFK; eBioscience) and IL-13 eFluor 660 (clone eBio13A) antibodies. Cells were analyzed using a BD FACSCanto flow cytometer or an LSR II instrument (BD Biosciences).

In vivo immune cell depletion/cytokine neutralization.

For CD20⁺ B cell depletion, mice were treated retro-orbitally (r.o.) and intraperitoneally (i.p.) with 0.3 mg of anti-CD20 mAb (clone SA271G2; BioLegend) or the IgG isotype control on day 2 prior to vaccination. The efficacy of CD19⁺ B220⁺ B cell depletion in the lungs and spleens (ranging from 90.4 to 98.6%) was confirmed 1 day after vaccination by flow cytometry. For CD8⁺ T cell depletion, mice were treated i.p. with 0.5 mg of anti-CD8 mAb (clone 53-6.7; BioXcell) on days 1 and 2 and challenged with PR8 virus on day 3. CD8 depletion of >94% in the lungs and spleens was confirmed by flow cytometry. For IL-5 neutralization, mice were treated i.p. with 150 μg of anti-IL-5 mAb (clone TRFK5; BioXcell) or the IgG isotype control on days −1, 1, 3, 6, 8, 10, and 13 postvaccination. BALF was harvested at 4 weeks postvaccination for antibody analysis.

Passive immune transfer.

Sera were collected from vaccinated mice at 4 weeks postimmunization. BALF from Fluzone+IL-33-vaccinated mice was collected on day 6 after PR8 infection. The immune sera and BALF samples were pooled and heated for 1 h at 60°C. Recipient mice were inoculated i.n. with PR8 or CA04 virus mixed with 5% serum or 80% BALF. Mice were monitored for mortality and body weight for 20 days.

Statistics.

Data were analyzed using GraphPad Prism 6 software, with a P value of <0.05 considered to be statistically significant. All survival data were analyzed with a log rank (Mantel-Cox) test. All other data were analyzed by one-way analysis of variance (ANOVA) with Bonferroni’s multiple-comparison test for comparisons of multiple groups and by an unpaired two-tailed t test with Welch’s correction for comparisons of two groups, as specified in the relevant figure legends. Results are presented as means and standard deviations (SD).

Data availability.

The data that support the findings of the study are available from the corresponding author upon request.