We designed a genetically modified rMVA encoding the model antigen OVA and flagellin (rMVA-OVA-flagellin), with the aim to further improve immunogenicity upon i.n. application of our rMVA vaccine, also encoding OVA (rMVA-OVA). Since the priming of adaptive immune responses depends on innate recognition of the vaccine, we addressed the innate response induced after i.n. administration of our novel vaccine.

Following i.n. immunization with rMVA-OVA-flagellin, cytokine production was measured in the BAL. Previous kinetic experiment (3, 6, 12, and 24 h) demonstrated an optimal cytokine response at a 24-h time-point (data not shown), which was, therefore, chosen for the study. Our results show that inflammasome-specific cytokines IL-1β (Figure 1A) and IL-18 (Figure 1B) could be detected. Mice immunized with rMVA-OVA also exhibited production of IL-18 (Figure 1B), but no IL-1β (Figure 1A).

Enhanced TNF-α could be observed in rMVA-OVA-flagellin-immunized mice compared to rMVA-OVA (Figure 1C), suggesting a TLR-mediated innate immune sensing. The levels of IL-6 were similar after rMVA-OVA and rMVA-OVA-flagellin immunization (data not shown).

Together, these results indicate that i.n. rMVA-OVA-flagellin immunization in vivo can trigger both inflammasome and TLR-mediated innate immune responses.

Myeloid cells, including DCs, macrophages, or neutrophils, are an important source of IL-1β. Having established IL-1β production in the BAL of rMVA-OVA-flagellin-immunized mice, we addressed the potential in vivo contribution of DCs in our vaccination model.

Two different models were used, where the in vivo role of DCs could be investigated. First, rMVA-OVA-flagellin-immunized Flt3-L^−/− mice, known to harbor drastically reduced numbers of dendritic cells, exhibited a decrease of IL-1β compared to wt mice (wt: 940 ± 85 pg/ml; Flt3-L^−/−: 305 ± 75 pg/ml) (Figure 2A). Similar decrease of IL-18 (Figure 2B) and IL-6 (Figure 2C) could be observed, although never abrogated, whereas TNF-α (Figure 2D) was unaffected. Next, we investigated conditional ablation of DCs using CD11c-DTR mice. In order to deplete lung DCs, CD11c-DTR mice were treated with DT i.n. prior to immunization with rMVA-OVA-flagellin. In our model, we could observe that DT treatment led to the depletion of the CD103⁺CD11b⁻ DC subset (Figure S1 in Supplementary Material). BAL of DT-treated CD11c-DTR mice exhibited significant reduction of IL-1β compared to non-depleted controls (CD11c-DTR + PBS: 644 ± 130 pg/ml; CD11c-DTR + DT 256 ± 116 pg/ml) (Figure 2E), whereas IL-18 (Figure 2F), TNF-α (Figure 2G), and IL-6 (Figure 2H) were unchanged after DT treatment.

Altogether, these data indicate that mucosal DCs are a major source of IL-1β, in response to i.n. MVA-flagellin immunization.

Having described the in vivo role of DCs in rMVA-OVA-flagellin lung innate immune recognition, we investigated cytokine production of in vitro bone marrow (BM) derived DCs cultured in the presence of Flt3-L (FL) (FL-DCs) transduced with rMVA-OVA, rMVA-OVA-flagellin or recombinant flagellin.

IL-1β was solely produced from rMVA-OVA-flagellin-stimulated FL-DCs, with marginal secretion following cytosolic delivery of recombinant flagellin (Figure 3A). This response was completely abrogated in Nlrc4^−/− FL-DCs. Similarly, the IL-18 response in rMVA-OVA-flagellin-stimulated FL-DCs was also strongly reduced in Nlrc4^−/− FL-DCs (Figure 3D), but unaffected in rMVA-OVA-stimulated wt and Nlrc4^−/− FL-DCs. This highlights the role of the NLRC4 inflammasome in MVA-OVA-flagellin-induced IL-1β and IL-18 production from DCs. When we investigated the cytokines IL-6 (Figure 3B) and TNF-α (Figure 3C), both cytokines were produced from rMVA and rMVA-flagellin-stimulated FL-DCs. MyD88^−/− FL-DCs, unable to respond to various TLRs including TLR5, stimulated with rMVA-OVA-flagellin showed significant reduction of IL-6 (Figure 3B) and TNF-α (Figure 3C), while still produced in Nlrc4^−/− FL-DCs, confirming the role of the TLR pathway in rMVA-OVA-flagellin innate immune recognition by DCs. Moreover, secretion of active IL-1β after NLRC4 activation requires caspase-1. Therefore, we investigated the role of caspase-1 in rMVA-OVA-flagellin-stimulated wt FL-DCs treated with the caspase-1 inhibitor, VX-765. A clear inhibition of IL-1β was observed (Figure 3E), suggesting inflammasome-dependent caspase-1 activation in response to rMVA-OVA-flagellin.

To elucidate if the differential detection of flagellin in MVA via TLR and inflammasome-dependent ways, described above for murine cells, could also be found for human cells, the human monocytic cell line, THP-1, known to naturally express moderate levels of TLR5 (23) was used. Notably, rMVA-OVA-flagellin induced highest levels of IL-1β (Figure 3F) and TNF-α (Figure 3G) in THP-1 cells. To confirm the role of NLRC4, we employed THP-1 cells with NLRC4 overexpression. Indeed, these cells produced elevated levels of IL-1β (Figure 3H), IL-18 (Figure 3I), TNF-α (Figure 3J), and IL-6 (Figure 3K) when they were stimulated with rMVA-OVA-flagellin but not rMVA-OVA. This response was fully dependent on the genetic activity of the virus since the induction of IL-1β was completely lost upon heat-inactivation of the MVA-OVA-flagellin (Figure 3L).

These data indicate both a role of the inflammasome and TLR pathway in innate immune recognition of rMVA-OVA-flagellin by DCs, together with cytosolic delivery of flagellin essential for optimal cytokine responses. Importantly, the data are consistent in both murine and human cells.

We further elucidated the role of NLRC4 activation by rMVA-OVA-flagellin in vivo. Wt and Nlrc4^−/− mice were i.n. immunized and IL-1β was significantly produced in the BAL of wt mice, while abrogated in Nlrc4^−/− immunized mice (Figure 4A). No loss of neither IL-6 (Figure 4B) nor TNF-α (Figure 4C) was observed in the BAL of rMVA-OVA-flagellin Nlrc4^−/− immunized mice. Interestingly, higher TNF-α levels were observed in Nlrc4^−/− mice compared to wt, which would indicate a negative role of the NRLC4 inflammasome on the NF-κb-mediated pathway of TNF-α production in our model.

Furthermore, the role of NLRC4 in ex vivo isolated lung cells was also investigated (Figures 4D–F). Lung cells from wt or Nlrc4^−/− immunized mice were cultured ex vivo, and IL-1β, IL6, and TNF-α were determined. Results show elevated IL-1β production from lung cells of wt mice in response to rMVA-OVA-flagellin-immunization, which was reduced in that of Nlrc4^−/− mice (Figure 4D). In contrast, no significant changes in IL-6 (Figure 4E) or TNF-α (Figure 4F) were observed.

Together, these results show that rMVA-OVA-flagellin recognition in the lung and IL-1β production is indeed dependent on the NLCR4 inflammasome.

rMVA vaccine, also encoding OVA-flagellin leads to enhancement of innate immune recognition after i.n. application. Therefore, we addressed whether adaptive immune responses would also be influenced by this adjuvant effect in vivo. Our vaccination schedule consists of an i.n. prime with either rMVA-OVA or rMVA-OVA-flagellin, followed by a homologous boost. At day 35, antigen-specific CD8 responses at systemic and mucosal sites were assessed.

In the spleen, no increase in absolute numbers of OVA peptide-specific dextramer positive cells (Figure 5A) between i.n. rMVA-OVA-flagellin and rMVA-OVA-immunized mice was observed (rMVA: 125,000 ± 34,000, rMVA-flagellin: 302,000 ± 90,000). Distinctively, in the lungs (Figure 5B), the data show a significant increase in OVA peptide-specific dextramer positive cells following rMVA-OVA-flagellin immunization (Figure 5B) (rMVA: 191,000 ± 420,004, rMVA-flagellin: 47,000 ± 76,000).

The effector functions of memory CD8⁺ T cells, known to rapidly produce IFN-γ and TNF-α (24) following in vitro restimulation with OVA peptide was assessed. In the spleen, no significant difference in the number of TNF-α⁺ OVA-specific CD8 T cells after rMVA-OVA-flagellin immunization could be observed (Figure 5C) (rMVA-flagellin: 303,000 ± 55,000, rMVA: 178,000 ± 20,000). In contrast, an increase in IFN-γ⁺ OVA-specific CD8 T cells ocurred (rMVA-flagellin: 447,000 ± 75,000, rMVA: 267,000 ± 30,000) (Figure 5C).

In the lungs, a significantly increased number of cytokine-producing OVA-specific CD8 T cells was observed in rMVA-OVA-flagellin-immunized mice (Figure 5D) (rMVA-flagellin: 147,000 ± 24,000 IFN-γ⁺, 103,500 ± 18,000 TNF-α⁺ OVA-specific CTLs; rMVA: 66,000 ± 14,000 IFN-γ⁺ and 47,000 ± 9,000 TNF-α⁺ OVA-specific CTLs).

Altogether, these data demonstrate the adjuvant-enhanced adaptive immunity of rMVA-OVA-flagellin, characterized by an increased neo-antigen-specific CD8 response and frequency of cytokine-producing memory CTLs. This significant enhanced antigen-specific immune response occurs in the lungs, at the site of immunization, with a similar trend observed in the spleen.

In recent years, i.n. immunization was described to elicit both upper-respiratory, genital, and GI tract immune responses (21, 22, 25, 26). Therefore, we determined if rMVA-OVA and rMVA-OVA-flagellin administered i.n. could also induce immune responses in the GI tract.

In order to increase the number of CD8 T cell precursors, we used an adoptive transfer of OVA-specific CD8⁺ T cells from donor CD45.1⁺ OT-I mice into C57BL6 wt recipient mice (Figure 6A), followed a day later by i.n. rMVA-OVA and rMVA-OVA-flagellin immunization. When mesenteric lymph nodes (MLN) were isolated 9 days after immunization, an increase of OT-I CD8⁺ T cells in the MLN was observed (Figure 6B). More specifically, a twofold increase in OT-I CD8⁺ T cells could be detected in rMVA-OVA-flagellin-immunized mice, compared to rMVA-OVA-immunized mice (rMVA: 0.5% ± 0.1, rMVA-flagellin: 1% ± 0.09) (Figure 6B).

These OT-I CD8⁺ T cells also expressed the gut-homing receptors α4β7 and CCR9 (Figure 6C).

Altogether, the increase in OT-I CD8⁺ T cells in the MLN indicates that i.n. MVA-OVA is associated with a CD8 response in the GI tract, which is enhanced when using rMVA-OVA-flagellin.

Having previously described that an i.n. rMVA-OVA-flagellin vaccine induced stronger adaptive T cell responses, we also wanted to determine the contribution of rMVA-OVA-flagellin on systemic and mucosal humoral immune responses.

Serum OVA-specific IgG was measured. Our data show no significant difference in OVA-specific total IgG titers between rMVA-OVA-flagellin and rMVA-OVA-immunized mice (Figure 7A). In order to identify the type of immune response generated, specific IgG isotype responses were measured. Whereas higher serum OVA-specific IgG1 titers were observed in rMVA-OVA-immunized mice (Figure 7A), enhanced OVA-specific IgG2c titers (Figure 7A) were observed after rMVA-OVA-flagellin immunization, suggesting a stronger bias toward a Th1 immune response.

Bronchoalveolar OVA-specific IgA antibody responses were determined and significant enhanced mucosal IgA responses were observed in rMVA-OVA-flagellin-immunized mice (Figure 7B). Furthermore, an increase in BAL anti-OVA-IgG titers, together with enhanced BAL IgG2c antibody responses (Figure 7B) was also observed in those mice, suggesting again a bias toward a Th1 immune response. OVA-specific IgA could also be detected in the LP of immunized mice (Figure 7C) with a clear increase in IgA titers following i.n. rMVA-OVA-flagellin immunization (Figure 7C).

Altogether these data suggest that i.n. rMVA-OVA-flagellin immunization leads to improved serum antibody responses and enhanced mucosal IgA in the upper-respiratory and GI tracts.

Having described the contribution of NLRC4 in rMVA-OVA-flagellin-specific enhanced innate immune responses, we wanted to determine the requirement of NLRC4 for the adjuvant effect observed in i.n. rMVA-OVA-flagellin vaccinated mice.

Nlrc4^−/− mice were immunized and adaptive CD8 T cell responses in the lung determined. Our results show that compared to wt mice, a lack of enhanced OVA-specific CD8 responses could be observed in i.n. rMVA-OVA-flagellin-immunized Nlrc4^−/− mice (Figure 8A).

Moreover, we also determined the number of cytokine-producing OVA-specific CTLs in rMVA-OVA-flagellin-immunized Nlrc4^−/− mice (Figures 8B,C). Convincingly, our results show that the enhancement of cytokine-producing CTLs observed in wt mice does not occur in rMVA-OVA-flagellin-immunized Nlrc4^−/− mice, with similar numbers of IFN-γ⁺ and TNF-α⁺ OVA-specific CTLs as that of rMVA-OVA-immunized mice.

This suggests that NLRC4 plays a preferential role in the rMVA-OVA-flagellin-induced lung CD8 T cell responses following i.n. immunization.

The enhanced T cell responses in the lung of rMVA-OVA-flagellin-immunized mice were largely dependent on functional NLRC4. Accordingly, we wanted to determine the role of NRLC4 in rMVA-OVA-flagellin-induced humoral responses.

First, we determined total serum IgG titers. Similar to wt mice (Figure 7A), there was no difference in total IgG titers between rMVA-OVA and rMVA-OVA-flagellin-immunized Nlrc4^−/− mice (Figure 9A). In contrast, striking differences were observed with regard to IgG2c responses. While rMVA-OVA-flagellin-immunized wt mice showed distinct enhanced serum OVA-specific IgG2c titers (Figure 7A), this effect was completely abrogated in Nlrc4^−/− mice (Figure 9A), suggesting an indispensable role of NLRC4 in rMVA-OVA-flagellin-induced enhanced Th1 responses.

Having observed this striking phenotype, we determined whether mucosal humoral responses might also be impaired in the absence of NLRC4. Our results show that the increased BAL IgA response of rMVA-OVA-flagellin-immunized wt mice (Figure 7B) is also abrogated in Nlrc4^−/− mice (Figure 9B), with similar titers of OVA-specific IgA observed between rMVA-OVA and rMVA-OVA-flagellin-immunized Nlrc4^−/− mice (Figure 9B). Comparable findings were observed in intestinal lavages of Nlrc4^−/− mice (Figure 9C), whereby enhanced IgA responses were also abrogated compared to wt mice (Figure 7C). Thus, this result suggests another essential role of NLRC4 in rMVA-OVA-flagellin-induced mucosal IgA responses.

Since rMVA-OVA-flagellin induced enhanced bias in mucosal Th1 responses in wt mice (Figure 7B), we also addressed the role of NLRC4 in rMVA-OVA-flagellin-specific induced mucosal IgG2c antibody responses. Quite strikingly, a loss of rMVA-OVA-flagellin-induced IgG2c responses was observed in the BAL of Nlrc4^−/− (Figure 9D) compared to wt mice (Figure 7A). This reduction of IgG2c strongly suggests that NLRC4 signaling is a prerequisite for rMVA-OVA-flagellin-induced enhanced IgG2c antibody responses.

Altogether these data point toward a role of NLRC4 in rMVA-flagellin-induced Th1-biased humoral systemic and mucosal immune responses.

C57BL/6J (H-2^b) mice were purchased from Janvier Labs. C57BL/6-Tg (TcraTcrb) 1100MjbJ (OT-I), and B6.SJL-Ptprca Pepcb/BoyJ (CD45.1) mice were obtained from the University of Zurich and bred to obtain CD45.1⁺ OT-I mice, respectively. CD11c-DTR, Myd88^−/−, Nlrc4^−/−, and Flt3-L^−/− mice were obtained from the University of Zurich. Mice were bred and maintained either in the animal facilities at Bavarian Nordic GmbH or at the University of Zurich according to institutional guidelines.

All recombinant virus vectors used for this study were based on a cloned version of MVA-BN^® in a bacterial artificial chromosome (BAC). The generation of MVA recombinants was carried out as described recently (24, 34). MVA-BN^® was developed by Bavarian Nordic and is deposited at the European Collection of Cell Cultures (ECACC) (V00083008). Briefly, the The pS promoter was cloned upstream of the open-reading frame for chicken OVA. The pHyb promoter was developed and described by Baur et al. (35) and comprises a late element from the promoter directing the expression of the AT1 protein in cowpox virus and five tandemly arranged early elements derived from a modified p7.5 promoter. The pHyb promoter was cloned upstream of the open-reading frame for Salmonella typhimurium FliC, hereafter referred to as Flagellin. Infectious viruses were reconstituted from BACs by transfecting BAC DNA into BHK-21 cells and superinfecting them with Shope fibroma virus as a helper virus. After three additional passages on primary chicken embryo fibroblasts, helper virus free MVA-OVA, MVA-OVA-flagellin viruses were obtained. All viruses used in animal experiments were purified twice through a sucrose cushion.

For intranasal immunizations, mice were anesthetized and immunized with a total volume of 50–60 µl containing 3.5 × 10⁷ TCID₅₀ of the various MVA recombinants. Mice received two immunizations on days 0 and day 21 with 3 weeks interval. Immune responses were evaluated 2 weeks after the final immunization at day 35. Serum, BAL, and intestinal lavages were taken on the final day for antibody titers. Groups of five mice were used.

For lung DC depletion, CD11c-DTR mice were treated i.n. with 100 ng diphtheria toxin (DT; Sigma-Aldrich) in 50 µl a day prior to immunizations.

CD45.1⁺ OT-I mice were used as cell donors for adoptive cell transfer into syngeneic CD45.2⁺ C57BL/6J recipient animals. Lymphocytes were isolated from spleen and OT-I CD8⁺ T cells purified using EasySep kits from StemCell Technology according to the manufacturer’s protocol. 1 × 10⁴ OT-I CD8⁺ T cells were injected intravenously via the tail vein in a 100-µl volume a day prior to i.n. immunization.

Spleen and lung cells were harvested from mice and incubated with 1 mg/ml collagenase/DNase in RPMI 1640 media containing 2% FCS for 30 min at 37°C on a shaker. Single-cell suspensions were then prepared by mechanically disrupting the organs through a 70-µm filter, washed, and counted. Spleen cells were then subjected to red blood cell lysis (Erylysis, Sigma-Aldrich), whereas lung cells were isolated by centrifugation with 44% Percoll, then washed and counted.

Mesenteric lymph nodes were harvested in RPMI 1640 media containing 2% FCS, disrupted mechanically through a 70-µm cell strainer. Mononuclear cells were washed and counted.

Spleen and lung lymphocytes were stained with PE-conjugated MHC class I H-2k^b dextramers loaded with OVA_257–264-peptide (SIINFEKL) (Immudex), for the detection of OVA-specific CD8 T cells, respectively, followed by anti-CD8α, CD44, and CD4. Antibodies were purchased from BD Biosciences, eBioscience, or BioLegend. For detection of adoptively transferred OT-I CD8⁺ T cells and gut-homing receptors, MLN cells were stained with anti-CD8α, CD45.1, CD44, α4β7, and CCR9. For intracellular cytokine staining, spleen and lung mononuclear cell suspensions were incubated with 2.5 µg/ml of MHC class I restricted peptides (OVA_257–264) (Genscript) for 5–6 h at 37°C in complete RPMI in the presence of 1 µg/ml GolgiPlug (BD Biosciences). Cells were stained as before with anti-CD8α, CD44, CD4, IFN-γ, and TNF-α. Intracellular cytokine staining of IFN-γ and TNF-α were performed after fixation/permeabilization (BD Cytofix/Cytoperm, BD Biosciences). For live/dead discrimination cells were stained before fixation (LIVE/DEAD fixable violet dead cell staining kit, Life Technologies). For analysis, all cells were gated on live CD8⁺ T lymphocytes following live/dead staining. All cells were acquired using a digital flow cytometer (LSR II, BD Biosciences) and data were analyzed with FlowJo software (Tree Star).

Serum OVA-specific IgGs were determined. For detection of OVA-specific IgA, the BAL fluid was collected by flushing the trachea three times with 1 ml of cold PBS/0.5 mM EDTA with a cannula, centrifuged to pellet the cells from the BAL fluid, and collected for ELISA. Intestinal washes were collected by flushing the distal part of the small intestine with 1 ml PBS containing trypsin inhibitor, 50 mM EDTA and 0.1% BSA. The solution was spun at 4,000 g for 15 mins at 4°C. ELISAs were performed by coating 96-well plates with 5 µg/ml OVA, followed by blocking with PBS containing 5% FCS/0.05% Tween20. IgG, IgG1, IgG2c, and IgA were detected using HRP-conjugated antibodies, followed by TMB substrate. Absorbance was measured at 450 nm. ELISA titers were determined using linear regression analysis and Log10 titers calculated.

Bone marrow (BM)-derived DCs cultured in the presence of Flt3-L (FL) (FL-DCs) were prepared as follows. BM cells obtained from flushed femurs and tibias, cultured in the presence of human recombinant FL for 8 days. FL-DCs (5 × 10⁵ DC/well) were stimulated with 4 TCID₅₀/cell of rMVA, rMVA-flagellin, or with 1 µg/ml recombinant flagellin from S. typhimurium (Invivogen). FL-DCs were pre-treated with the transfection reagent Lyovec (Invivogen) for 15 mins. FL-DCs were pre-treated for 1 h with the caspase-1 inhibitor VX-765 (Invivogen), followed by infection with 4 TCID₅₀/cell of rMVA or rMVA-flagellin. Cells were incubated overnight and the supernatant was collected and measured for cytokine production using the Luminex.

Lung cells from PBS, rMVA, or rMVA-flagellin immunized mice (wt or Nlrc4^−/−), were harvested as described previously. A total of 5 × 10⁵ cells/well were cultured ex vivo with 10 ng/ml IL-4, GM-CSF and rat IFN-γ overnight in RPMI media containing 10% FCS. The supernatant was collected and measured for cytokines using the ProcartaPlex^® Multiplex Immunoassay (Affymetrix).

The human monocytic cell lines THP-1 and THP-1-NLRC4 (Invivogen), stably overexpressing NLRC4 and naturally expressing TLR5 (23) were cultured in RPMI containing 10% FCS. Cells were stimulated with 4 TCID₅₀/cell of rMVA, rMVA-flagellin overnight. Viruses were heat-inactivated for 10mins in a 70°C water-bath, and THP-1 and THP-1-NLRC4 cells stimulated with 4 TCID₅₀/cell of rMVA, rMVA-flagellin overnight. The supernatant was collected and measured for cytokine production using the Luminex.

Bronchoalveolar and harvested supernatants were measured for cytokines using the ProcartaPlex^® Multiplex Immunoassay (Affymetrix) according to the manufacturer’s protocol. Data were acquired on a MAGPIX^® system (Invitrogen). Results were analyzed using the Masterplex 2010 software (Hitachi Solutions).

Statistical analysis using one-way or two-way ANOVA with multiple comparisons was performed using GraphPad Prism 7 software. p value < 0.05 was considered significant.