Six- to eight-week-old female CB6F1/OlaHsd mice (Harlan Laboratories, The Netherlands) were housed in ventilated cages with free access to food and water. This strain is a first-generation hybrid derived from crossing two strains of mice, the BALB/c (which typically exhibits T2-biased immune response) and C57BL/6 (which exhibits T1 biased response). The balanced T1/T2 immunological profile of this strain is particularly relevant for our study’s objectives of evaluating how different adjuvants and immunization routes can skew the immune response. All mice were housed under standardized pathogen-free conditions at the Experimental Biomedicine Animal Facility, University of Gothenburg.

The use of mice in this study was performed in accordance with the regulations set forth by the Ethical Committee for Animal Experimentation in Gothenburg, Sweden and in accordance with European Community Directive 86/609. All the techniques and procedures were refined to provide for maximum comfort and minimal stress to mice. Experiments performed were approved by the Ethical Committee for Animal Experimentation in Gothenburg, Sweden under license 64/2015.

Groups of mice were either s.c. or i.n. immunized three times, at 21-day intervals. Treatment mice were vaccinated with 5µg recombinant H56 (SSI; diluted in 10mM Tris) combined with either 5µg CTB-CpG (produced as previously described (30)), diluted in Phosphate buffered solution [PBS]), or CAF01 (SSI; diluted in 10mM Tris buffer) mixed in a 1:1 ratio with H56 as recommended by manufacturer. Control mice were administered H56 alone. S.c. administrations were given in volumes of 200 µL, with two 100 µL doses on each side of the base of the tail. Intranasal administrations constituted a volume of 12 µL divided into two equal doses to each nostril. Intranasal administrations adjuvanted with CAF01 were given in small volumes repeatedly 30 minutes apart, comprising a total volume of 40 µL. Unless stated otherwise, heterologous Prime-Boost schedules comprised of two subcutaneous administrations of H56+CAF01 as the prime, and a booster dose of H56 adjuvanted with either CAF01, CTB-CpG or 2 µg mmCT (produced as previously described (29)) administered i.n. A homologous prime-boost immunization scheme consisted of repeated i.n. administrations of H56 adjuvanted with either CAF01, CTB-CpG, mmCT or 2 µg CT (List biological laboratories, Inc; diluted in PBS) 21 days apart.

Blood samples were collected by puncturing the vena saphena located in the hindleg two weeks after the final immunization. Samples were left to coagulate at room temperature (RT), then centrifuged 10,000 × g for 10 min at 4°C. Sera were frozen at −20°C until further analysis.

Spleens and lungs were harvested from euthanized mice two weeks post final immunization. Lungs were perfused with 0.1% heparin-PBS, then dried with a towel before one lung from each animal was collected for antibody measurement and the other for FACS analysis. For antibody measurements the lungs were chopped and put into Eppendorf tubes. 30 μL of 20% Saponin was added to each Eppendorf tube, it was then vortexed and stored at 4°C overnight. Next, the lung samples were centrifuged 13,000 × g for 10 min and the supernatant stored at -20°C until antibodies were measured with an ELISA. For FACS analysis, one lung from each animal was cut into smaller pieces, resuspended in liberase (1:5 dilution of Liberase [Sigma-Aldrich] in 1640 RPMI [Invitrogen]) in a gentleMACS C tube (Miltenyi Biotech) and dissociated using a gentleMACS dissociator (Miltenyi Biotech). Lungs were then digested for 1hr, 37°C in digestion solution comprising of liberase (1:5 dilution) mixed with 0.5% DNase in RPMI medium (supplemented with 5-10% FCS [Fetal Calf Serum]). Organs were forced through a 70µm nylon mesh cell strainer and single-cell suspensions were collected. Red blood cells from the spleen were lysed with addition of ammonium chloride followed by incubation for 5 minutes at 37°C. RPMI (without FCS) was then added and spun at 500g for 5 minutes. The supernatants were then discarded and the cells were washed again with RPMI. The splenocytes were resuspended in complete RPMI (supplemented with HEPES, penicillin-streptomycin, L-Glutamine, sodium pyruvate and non-essential amino acids and 10% FCS).

H56-specific IgG antibody levels in the serum samples were evaluated using an ELISA. Flat 96-well MaxiSorp plates (Nunc) were coated with 0.5 µg/ml H56 in carbonate buffer (pH 9.6) and incubated at 4°C overnight. This was then washed thrice with wash buffer (PBS with 0.2% Tween-20), and then blocked with blocking buffer (PBS with 2% Bovine Serum Albumin [BSA]) for 2 hours at room temperature. The plates were then washed another three times, followed by incubation with serially diluted serum samples (diluted in PBS with 1% BSA), starting with a dilution of 1:3 for IgA evaluation and 1:10 for IgG evaluation. Background wells without serum samples were included as controls. After incubation, plates were washed another three times with washing buffer. Dilutions of HRP-conjugated secondary anti-mouse antibodies (Invitrogen) were made in PBS with 1% BSA as follows: 1:20000 for anti-mouse IgG antibody, and 1:5000 for anti-mouse IgA antibody. Secondary antibodies were then added to wells and incubated for 1hr at room temperature. This was then followed by washing five times with washing buffer. For the last wash, wash buffer was left in wells for 1 minute before discarding. TMB Ready-to-use substrate (Kim-En-Tec) was then added to the wells. Reaction was stopped after 10 minutes with stopping solution (H₂SO₄ 2N). Absorbance was then read at 450nm. Endpoint titers were assessed by plotting curves as Log₁₀ values on GraphPad PRISM and fitting sigmoidal curves without constraint.

Lung cells and splenocytes were harvested from euthanized mice as described above. Cell suspensions were resuspended in complete RPMI (supplemented with HEPES, penicillin-streptomycin, L-Glutamine, sodium pyruvate, non-essential amino acids and 10% FCS), adjusted to 1 x 10⁷ cells/mL and were seeded in v-bottom 96-well plates. For antigen recall stimulation, 2 µg/ml of H56 was added to each well. A negative medium-only control and a positive control of PMA (Sigma-Aldrich; 50 ng/mL) + ionomycin (Sigma-Aldrich; 1µM) were included for each sample. Single stain-only controls were also included for compensation purposes, as well as an unstained control. Anti-CD28 (BD Biosciences) and anti-CD49d (BD Biosciences) antibodies were added to a final concentration of 1 µg/mL. The plates were then incubated for 2 hours at 37°C. Brefeldin A (Sigma-Aldrich) was then added to the wells to a final concentration of 10 µg/ml and incubated for 5 hours at 37°C using ThermostatPlus heatblock (Eppendorf) programmed to cool down to 4°C for the next day. The plates were then spun at 2200 rpm for 3 mins at 4°C. The supernatants were then discarded, and the cells resuspended in FACS buffer (PBS with 1% FCS). The cells were spun at 2200 rpm for 3 minutes and supernatants discarded. They were then blocked with anti-CD16/32 antibody (BD Pharmingen) for 10 min at room temperature. Cells were then surface stained with CD4:PE-CF594 (eBioscience), CD44:V450 (BD Biosciences) and CCR6:AF647 (eBioscience) (all dilutions were 1:200) and incubated for 15 minutes in the dark. Cells were then spun at 2200 rpm for 3 mins at 4°C, supernatant discarded, and washed again in FACS buffer as before. For intracellular staining, cells were fixed with the addition of Cytofix/Cytoperm (BD Biosciences) and incubated for 20 minutes at 4°C in the dark. Perm Wash was then added to wells, and plates spun at 2200 rpm for 3 minutes and supernatant discarded. Intracellular antibodies were then diluted 1:200 in Perm Wash Buffer: IFNγ:FITC (eBioscience), TNFα:PE-Cy7 (eBioscience) and IL-17:PerCP-Cy5.5 (eBioscience) and added. Plates were incubated for 20 minutes at 4°C in the dark. Cells were washed thrice in Perm Wash, resuspended in FACS buffer and then subjected to flow cytometry on LSR II (BD Biosciences). Cytokine-secreting CD4+ T cells were then evaluated by assessing IFN-γ, TNF-α & IL-17 positive gates within the CD4+ CD44^High cell population. Boolean combination gates were used to assess the frequency of each cytokine co-expressing subset and its frequency in total T cell population.

Th1/Th2/Th17-specific cytokines were measured in H56-induced proliferated T cells using bead array. The BD™ CBA mouse Th1/Th2/Th17 Cytokine Kit (BD Biosciences) was used as per manufacturer’s instructions to evaluate levels of IL-2, IL-4, IL-6, IFN-γ, TNF-α, IL-17A, and IL-10. Briefly, capture beads for each cytokine were combined and mixed with dilutions of splenocytes or recombinant standard. These were then incubated with PE-conjugated detection antibodies to form sandwich complexes. The intensity of PE fluorescence of each sandwich complex correlates with the concentration of cytokine. Samples were then run on LSR II (BD Biosciences), and FCAP Array™ software was utilized to generate results in graphical and tabular format.

Data analyses were performed using Graphpad Prism 7.0. Statistical comparison was done using non-parametric Mann-Whitney test, Kruskal-Wallis test, or an ordinary one-way ANOVA. Differences were considered significant (*) at p value <0.05.

We first immunized mice either s.c. or i.n. with H56 adjuvanted with either CAF01 or CTB-CpG to evaluate their systemic humoral immune response. We observed that while s.c. administration of both H56+CAF01 and H56+CTB-CpG resulted in strong serum H56-specific IgG antibody levels, i.n. administration of H56+CAF01 resulted in a poor vaccine-specific humoral response compared to H56+CTB-CpG ( Figure 1A ). Further, i.n. administration of H56+CTB-CpG enhanced both vaccine-specific IgG and IgA antibody levels in the lungs, which was not observed with i.n. administration of H56+CAF01 ( Figure 1B ).

We then isolated splenocytes of vaccinated mice and restimulated with H56 to evaluate T cell-associated cytokine response. Interestingly, splenocytes from mice immunized s.c. with H56+CAF01 generated stronger IL-17, TNF-α, IFN-γ, IL-1+, IL-6 and IL-2 cytokine levels than that from s.c. H56+CTB-CpG-immunized mice ( Figure 1C ). This suggests that while parenteral administration of both vaccines results in a similar systemic humoral response, H56+CAF01 may induce a stronger peripheral T cell response. In contrast, i.n. administration of the two vaccines resulted in distinct cytokine profiles. While splenocytes from H56+CAF01-immunized mice had higher levels of IL-6 and IL-2 levels than H56+CTB-CpG-immunized mice, the latter showed stronger IL-17 (p<0.01), IFN-γ and IL-10 levels (suggestive of a Th17-skewed immune response) ( Figure 1D ).

These results collectively point to the suitability of CTB-CpG as a mucosal adjuvant because its i.n administration elicited systemic T cell immune responses, as well as humoral responses in both the lungs and periphery. In contrast, while s.c. administered CAF01 induced robust systemic cellular and humoral responses, poor mucosal responses were induced upon i.n. administration.

We sought to determine whether a heterologous prime-boost immunization scheme consisting of a strong parenteral adjuvant (CAF01) as the prime, and strong mucosal adjuvants (CTB-CpG and mmCT) as i.n. boosters would enhance the mucosal immune response. We compared this to repeated homologous i.n. administrations of the same adjuvant, and also included CT as a gold standard adjuvant for homologous mucosal immunization. Mice were immunized either homologously with repeated i.n. administrations of the same adjuvant, or heterologously with a parenteral prime (H56+CAF01) and a mucosal boost of either H56+CAF01, H56+CTB-CpG or H56+mmCT ( Figure 2A ). To evaluate the humoral response induced by the different prime-boost regimens, we measured IgG and IgA antibody levels in serum and perfused lungs. We observed that the H56-specific IgA antibody response induced in the lungs were highest in the group that received repeated intranasal CT immunizations. However, IgA levels across all 3 heterologous prime-boost immunization groups were comparable ( Figure 2B ). Compared to the homologous i.n. groups, with the exception of the combination of H56+CAF01 s.c./H56+CAF01 i.n., heterologous prime-boost scheme did not provide an added benefit in inducing vaccine-specific IgA antibody levels in the lungs ( Figure 2B ).

This trend was similarly observed in the induction of serum and lung H56-specific IgG antibody levels. While heterologous vaccine regimes induced H56-specific IgG antibody levels in both sera ( Figure 2C ) and lungs ( Figure 2D ), it was not superior to that seen by their homologous intranasal counterparts – with the exception of H56+CAF01 s.c./H56+CAF01 i.n. in the lungs. Furthermore, the IgG levels in the lungs strongly correlated (r=0.71, p<0.0001) with IgG antibody levels in the serum ( Figure 2E ) suggesting transudation of serum IgG antibody from the blood circulation in to the lung mucosa, a phenomenon that has been reported previously (33). Taken together, these results demonstrate that heterologous prime-boost may be beneficial to evoke a balanced systemic and mucosal responses that are non-inferior to that observed with homologous i.n. prime-boost combinations with strong mucosal adjuvants.

Next, we sought to investigate whether the different prime-boost strategies would differentially skew the Th profile in the lung mucosa. To this end, CD4+ T cells were retrieved from mice immunized with different prime-boost regimens two weeks after the last immunization and performed CD4+ T cell-specific and intracellular cytokine analysis upon H56 restimulation. The group that received homologous CT i.n. administrations had the highest frequency of antigen-specific CD4+ T cells (CD44^high) in lungs compared to all other groups ( Figure 3A ). We observed that i.n CT immunization exhibited a strong Th17 profile, as indicated by the secretion of IL-17 and TNFα ( Figure 3B ).

Within the heterologous prime-boost groups, the number of antigen-experienced CD4+ T in lungs of the groups receiving H56+CAF01 s.c./H56-mmCT i.n were higher than in the other heterologous prime-boost groups ( Figure 3A ). In addition, the prime-boost combination of H56+CAF01 s.c./H56+CTB-CpG i.n. resulted in a stronger Th1-skewed profile than the other heterologous groups, as evidenced by stronger IFN-γ-secreting CD4+ T cells ( Figure 3B ). This is further supported by higher levels of IFN-γ+TNFα double-secreting or IFN-γ single-secreting cells in H56+CAF01 s.c./H56+CTB-CpG i.n. than the other heterologous immunization groups ( Figure 3C ). In contrast, a combination of H56+CAF01 s.c./H56+mmCT i.n. resulted in a polyfunctional Th1/Th17 immune phenotype, as evidenced by strong levels of IL-17+TNFα double secreting cells ( Figure 3C ) and higher levels of CCR6+ CD4+ T cells (p<0.05, Figure 3D ). Further, cytokines detected in the cell culture supernatant from H56-restimulated lung-derived CD4+ T cells also supported that heterologous H56+CAF01 s.c./H56+CTB-CpG i.n. mainly induced a Th1 response whereas H56+CAF01 s.c./H56+mmCT i.n. was more Th17 biased ( Supplementary Figure 1 ). We also observed that while the heterologous H56+CAF01 s.c./H56+mmCT i.n group engendered similar levels of H56-responding CD4+ T cells as those received homologous H56-mmCT intranasal administrations ( Figure 3A ), the latter combination induced a higher frequency of H56-specific CCR6+ CD4+ T cells ( Figure 3D ).

These results suggest that by varying the mucosal booster, a uniquely Th1-skewed or Th1/Th17-skewed mucosal response can be induced. Importantly, replacing the parental prime with either H56+CTB-CpG or H56+mmCT (followed by i.n. immunization with the same adjuvant) induces limited mucosal T-cell responses ( Supplementary Figure 2 ). These results suggest that the enhanced mucosal immunity induced by the parenteral administration of CAF01 followed by a mucosal administration, was unique to CAF01. Moreover, this effect is not seen when a parenteral CAF01 administration is preceded by intranasal CTB-CpG ( Supplementary Figure 3 ), indicating that the order of administration is important.

Next, we examined the longevity of the mucosal immune responses induced by the heterologous parenteral prime/mucosal boost immunization. Thus, we carried out a 3-month long study with the following groups: homologous parenteral immunization (H56+CAF01 s.c.), heterologous prime-boost (H56+CAF01 s.c. followed by H56+mmCT i.n.) and homologous mucosal immunization (repeated H56+mmCT i.n.). Three months after the last immunization, half of the mice received an additional i.n. booster with mmCT for evaluation of mucosal recall responses ( Figure 4A ). Firstly, we observed detectable levels of H56-specific cytokine-secreting CD4⁺ T cells in the lungs of all immunized groups 3 months following the last immunization ( Figure 4B ). These levels were further increased after the 3-month booster in the groups that had been parenterally primed with CAF01, whereas no significant booster effects were observed in the groups that had only received repeated i.n. immunization ( Figure 4B ). Further, the frequency of IFN-γ-secreting and TNFα-secreting CD4+ T cells post-boost in mice that received H56+CAF01 s.c./H56+mmCT i.n. immunization was higher than that observed in mice with repeated i.n. mmCT immunizations. This indicates that while we had previously observed no added benefit of a CAF01 s.c/mmCT i.n combination compared to repeated mucosal mmCT vaccinations in terms of humoral ( Figure 2 ) and cellular immune responses in the mucosa ( Figure 3 ) 2 weeks post-final immunization, CAF01 s.c/mmCT i.n immunization did result in a more profound T cell anamnestic response 3 months post-final immunization. Moreover, while homologous s.c. administration of H56+CAF01 showed a comparable trend to that of heterologously administered H56+CAF01s.c/H56+mmCT i.n in terms of mucosal T cell response, the former elicited undetectable mucosal IgA antibody levels, even upon the 3-month booster ( Figure 4C ). In contrast, the heterologous immunization regimen induced enhanced mucosal IgA antibody levels indicating the induction of a strong mucosal memory B cell response; however these levels were not different from that observed in the group that received repeated i.n. administrations ( Figure 4C ). Collectively, these results indicate that the heterologous prime-boost immunization with H56+CAF01 s.c./H56-mmCT i.n. is beneficial to induce potent long-lived H56-specific CD4+ T cell and IgA responses that can be pulled to the mucosal tissue upon i.n. booster.