Porcine SK-6 cells (35) were kindly provided by Professor Maurice Pensaert (University of Gent, Belgium). The synthesis and purification of the mucosal adjuvant c-di-AMP was described in Ebensen et al. (30).

Rep-HA and Rep-NP constructs were already described elsewhere (24–26) and are schematized in Figure 1A. They were derived from plasmid pA187-1 that carries a full-length cDNA copy of the genome of the CSFV strain Alfort/187 (CSFV parent) (36) from which the E^rns coding sequence was deleted (ΔE^rns) to engineer the original ΔE^rns RepRNA (RepRNA). The Rep-NP was obtained by insertion of the NP gene from influenza virus A/chicken/Yamaguchi/7/2004 (H5N1) (37) via a NotI restriction endonuclease site at the 3′ end of the N^pro coding sequence, upstream of an internal ribosomal entry site (IRES) from encephalomyocarditis virus (EMCV) that permits re-initiation of the translation of the downstream polyprotein. For Rep-HA, a copy of the CSFV C coding sequence was placed between the N^pro and the HA glycoprotein coding sequence of the same influenza virus with HA lacking the 5’-terminal 16 codons of the signal peptide. For this, a N^pro-C-HA DNA cassette was generated by PCR and inserted between the ClaI site of N^pro and the NotI site. This ensures the correct ER translocation of the HA glycoprotein by the E^rns signal peptide encoded at the 3-terminal region of C.

VRPΔE^rns carrying a genome with a complete deletion of the E^rns coding sequence were produced typically by transfection of 8 x 10⁶ SK-6(E^rns) cells with 1 μg A187-Δ^Erns replicon RNA. The SK-6(E^rns) cells express the E^rns protein required for the generation of VRPΔE^rns by trans-complementation (38, 39). The infectious titre of the VRP was determined in SK-6(E^rns) cells by end-point dilution and expressed in 50% tissue culture infective doses (TCID₅₀)/ml according to standard protocols. Based on RT-qPCR quantification of infectious CSFV (of known titre) and of the corresponding full-length genomic in vitro transcripts (1µg CSFV genome RNA transcripts = 1.5 x 10¹¹ molecules), 1 TCID₅₀ of CSFV corresponds approximately to 10³ genome equivalents (Hinojosa and Ruggli, unpublished).

The generation of chitosan-based nanoparticle delivery formulations, and the PEI-based polyplex nanoparticles were as described previously (23–27).

Briefly, for polyplexes, RepRNA incorporated into polyplex formulations were as follow: [Rep-NP/PEI-4,000 (1:3)] and [Rep-HA/PEI-40,000 (1:2)/(Arg)₉]. For Rep-NP, RepRNA : PEI-4,000 (weight:weight) ratio of (1:3) was mixed by vortexing (4 s, 10 mM HEPES buffer, pH7.4). After 30 min of incubation at room temperature (RT), volumes were adjusted with serum-free Opti-MEM^®. For Rep-HA, PEI-40,000 solution was first mixed to 0.5 µM of (Arg)₉ and further incubated for 30 min at RT. Then, RepRNA was added to the [PEI/(Arg)₉] core with a RepRNA : PEI (weight:weight) ratio of (1:3), incubate for 30 min at RT. Volumes were adjusted with serum-free Opti-MEM^®.

All batches of RepRNA production used for the trial were tested by a functional assay adapted to confocal microscopy; the number of cells (expressing CSFV E2) and the gene of interest (GOI) HA or NP were monitored. This assay was described elsewhere (24, 25) and employed reference SK-6 cells that were mixed with RepRNA and electroporated immediately. These cells are very efficient at propagating CSFV and provide a reliable reference to support RepRNA replication (23, 24, 26, 35, 39, 40). SK-6 cell growth is facilitated by Eagle’s Minimal Essential Medium (MEM Earle’s, consisting of MEM supplemented with Earle’s salts, 2 mM l-glutamine, and 7% (v/v) pestivirus- and Mycoplasma-free horse serum). The readouts were measured after 48 h; illustrations of E2⁺NP⁺ foci following electroporation of SK-6 cells are shown in Figure 1B. The criterion for inclusion of a given RepRNA batch in the trial is detection of E2⁺GOI⁺ foci in the fourth well of a tenfold serial dilution (10^-4 μg of RepRNA per 8 x 10² SK-6 transfected cells seeded on a pre-formed monolayer of 3 x 10⁵ cells). Ultimately, all the positively-selected RepRNA batches were thawed the day of administration in pigs.

Primary porcine blood DCs (bDCs) were isolated/derived as described (41). Then, they were grown in 8-well fibronectin-coated Lab-Tek^® chambers (Nunc, Wiesbaden, Germany). Before exposure to VRP or [RepRNA/PEI], bDCs were washed to remove serum. After 1-3h, cells were washed again and cultured (39°C; DMEM/10%-porcine-serum/50 U/ml-GM-CSF/100 U/ml-IL-4) for 96 h. Labeling used anti-E2 (HC/TC 26, kindly provided by Irene Greiser-Wilke, Hannover, Germany) and anti-NP (HB6J; ATCC, Rockville MD, USA) antibodies. DC-surface staining used WGA-Alexa633 (Molecular Probes/Invitrogen), 10 min on ice, then washed, fixed (4% PFA, 10 min), permeabilized (Perm/Wash-Buffer-I, BD Biosciences, Allschwil, Switzerland). Slides were mounted in Mowiol for confocal microscopy.

The immunization of mice and rabbits using the chitosan-based nanoparticle (NGA) delivery formulations has been described previously (23). Briefly, RepRNA was incorporated into NGA, and 100 µl were used per dose for vaccination of Balb/c mice and New Zealand white rabbits. Each dose comprised 0.2 µg of Rep-HA in 50 µl plus 0.2 µg of Rep-NP in 50 µl; mice received one dose for each RepRNA mixed together, while rabbits received twice this dose. The NGA vaccines were adjuvanted with 0.5 µg c-di-AMP adjuvant. Animals were vaccinated subcutaneously at days 0, 14, and 28, and bled from the tail vein (mice) or ear (rabbits). Anti-HA and anti-NP titers were assessed by indirect enzyme-linked immunosorbent assay using recombinant HA (H5 Vietnam 2004) and NP as antigen.

The immunization of mice using the polyplex delivery formulations has also been described previously (24, 25). Briefly, C57BL/6 (H-2b) female mice (6–8 week old) were purchased from Harlan (Germany). OVA-TCR transgenic mice C57BL/6-Tg(TcraTcrb)1100Mjb/J (OT-I) were bred at the animal facility of the HZI. All mice were kept under SPF conditions in compliance with the guidelines of the Institutional Animal Use and Care Committee. OT-I (CD8) T cells were sorted from cervical LNs of naive mice (Thy1.1⁺) using Miltenyi CD8 T cell isolation kits. Then, 2.5 × 10⁵ OT-1 CD8 T cells were injected intravenously into recipient C57BL/6 mice, 6–8 week old (Thy1.1⁻) (Day-1). At day 0, mice were vaccinated intrapulmonary with 7.5 μg Rep-OVA and 7.5 μg adjuvant c-di-AMP. Cervical LNs were collected at day 7.

All pigs were obtained from the specific-pathogen-free (SPF) breeding facility of the IVI. Male and female 10-week-old Large White SPF pigs were assigned randomly to three groups of five animals (n=5) housed in separate stables. The experiment was blinded.

Each pig was immunized by intradermal (i.d.) injection of Rep-HA and Rep-NP, co-administered or not with 250 μg of c-di-AMP as adjuvant. During the past few years, we have accumulated data confirming that naked RepRNA is unable to penetrate into DCs for translation, nor does it show evidence for translation in vivo. Accordingly, a control group of naked RepRNA alone would have provided no new insights. Instead, groups were constituted as follow: naked RepRNA in association with c-di-AMP (group “RepRNA + c-di-AMP”), reference points for the assay; RepRNA formulated into PEI polyplex delivery vehicles in association with c-di-AMP (group “[RepRNA/PEI] + c-di-AMP”); or RepRNA packaged into VRPs (group “VRP, no adjuvant”). Within each vaccine of the groups “RepRNA + c-di-AMP” and “[RepRNA/PEI] + c-di-AMP”, 50 µg of Rep-HA and 50 µg Rep-NP were applied in a total volume of 0.5 to 0.8 ml each, distributed in five to eight different spots of 0.1 ml each in the dermis of the right and left neck, respectively. For the “VRP” group, 5x10⁶ TCID₅₀ of VRP-[Rep-HA] and 5x10⁶ TCID₅₀ of VRP-[Rep-NP] were applied in a total volume of 0.5 to 0.8 ml each, distributed in five to eight different spots of 0.1 ml each in the dermis of the right and left neck, respectively. 5x10⁶ TCID₅₀ of VRP corresponds approximately to 5x10⁹ viral genomes (see above), which represents a weight of 33 ng of RNA (1µg CSFV genome = 1.5x10¹¹ genome molecules, see above). Thus, compared with VRP, approximately 10³ times more RepRNA (50 µg) was applied with the PEI polyplex delivery system.

A total of three immunizations were applied with 3 weeks interval. At day 56, animals were euthanized by electrical stunning and subsequent exsanguination, blood was taken and superficial cervical dorsal lymph nodes (LNs) were collected (Figure 1C). Importantly, naked RepRNA co-administered with c-di-AMP, the VRPs and the PEI-complexes were well tolerated by the animals, and no side effects were observed over the 56-day observation period (data not shown).

PBMC were isolated by centrifugation on a Ficoll-paque density gradient (1.077 g/L, GE Healthcare. Chicago, IL, USA). Harvested LNs were washed with sterile PBS^+/+, cut in small pieces, and incubated in 15 ml of pre-warmed PBS^+/+ containing 0.36 mg/ml collagenase D (Sigma-Aldrich) and 100 µg/ml DNase I (Sigma-Aldrich) during 30 min under agitation at 37°C. The enzymatic reaction was stopped by adding 50 ml of PBS^-/- with 5 mM EDTA solution. The cells were then filtered through a 100 µm and 70 µm filter (BD Biosciences, Allschwil, Switzerland), washed three times with PBS^-/- (4°C). Cell count and viability was done with Türk’s solution under the microscope.

Isolated PBMCs or LN cells for T cell and B cell restimulation assays (1 x 10⁶ cells/ml) were cultured in DMEM/10%FBS with 1 µg/ml recombinant HA/NP (Immune Technology Corp., New York, U.S.A.), 1 µg/ml recombinant E2 [produced in cultures of insect cells infected with the baculovirus vector (42)], or without (unstimulated). As a positive control, the cells from all samples were cultured with 10 µg/ml of Concanavalin A (Con A) (Sigma). Importantly, those experiments were performed in the absence of antibiotics. After 3 or 5 days at 39°C, cells were harvested and analyzed by flow cytometry (FCM).

FCM combination staining for T cell and B cell proliferation by CellTrace™ Violet (Life Technology) dilution employed: CD3 (BB23-8E6-8C8) (BD Biosciences), CD4 (PT90A), CD8 (76-2-11), CD21 (BB6.11C9.6), and IgM (PIG 45A) (all from VMRD, Germany). 5x10⁵ events were stained for each sample and the whole tube was acquired. Analysis gate was placed on FSC^high cells, as proliferating T cells are located at this area (24). Proliferating CellTrace^low cells were scored as a percentage of the FSC^high cell gate. Data were acquired using FACSCanto II (Becton-Dickenson), and analyzed with Flowjo software (version 9, Treestar, USA).

Anti-HA and anti-NP antibody titers were assessed by indirect ELISA as described previously (23).

Statistical analysis was done using the GraphPad Prism 8 software (GraphPad software, La Jolla, CA, USA). The p-values were calculated using one-way ANOVA and Bonferroni nonparametric post-test. Data are presented as box and whisker plots (* p <0.05, ** p <0.01, *** p <0.001).

Comparison of the previously reported chitosan-alginate nanoparticulate (NGA) delivery with VRPs using RepRNA encoding influenza virus HA and NP showed a clear advantage for VRP delivery in rabbits (Figure 2A). In contrast, the NGA delivery appeared to be the more advantageous in mice, in terms of an early induction of specific immunity; the VRPs required three immunizations before a detectable specific antibody response was obtained (Figure 2B). Moreover, VRP efficiency in mice was variable between experiments, with little or no specific antibody induced in certain cases (data not shown).

The readouts obtained with the NGA in mice also showed variation between experiments, as did the use of lipids (27) or polyplexes (24, 26) for delivery. It was discovered that the efficiency of polyplex delivery leading to translation of the RepRNA was highly dependent on the PEI formulation (24), and it appears that the potential for lipoplexes suffers from the same encumbrance (27). Finally, we considered the efficiency of polyplex delivery formulations to induce cellular immune response. Antigen-specific proliferation of OT-I CD8⁺ T cells were assessed 7-days post-immunisation in cells collected from cervical LNs. Reduced CFSE signal from preloaded cells (due to cell division distributing the dye to daughter cells) was employed as readout. In case of the CD8⁺ T-cell compartment, immunisation with polyplex was efficient to induce CD8⁺ proliferative responses compared to PBS and naked RepRNA treated mice. These in vivo results demonstrate that polyplexes combined with the mucosal adjuvant c-di-AMP promotes translation of encoded antigens by DCs to stimulate proliferation of GOI-specific CD8⁺ T cells (Figure 2C), making important to assess their efficiency in the porcine model, due to the closer immunological relationship to humans.

Before proceeding with the assessment of RepRNA delivery in pigs, it was considered important to ascertain the efficiency with which the polyplexes and VRPs delivered RepRNA for translation of the inserted influenza virus genes. This was initially determined using porcine SK-6 cells and blood DCs, relating to our previous publications for chitosan-based, lipid-based, and polyplex-based delivery of the RepRNA (6, 23–27). Figure 2D shows that both the polyplexes and VRPs were able to deliver RepRNA leading to translation of the gene encoding influenza virus NP antigen.

Next, we compared polyplex- and VRP-mediated delivery of RepRNA in pigs with the immunization scheme shown in Figure 1C. Cells collected from peripheral blood and draining LNs of immunized pigs were restimulated in vitro with CSFV E2, or influenza virus HA/NP recombinant proteins. First, we quantified the percentage of FSC^high cells for all unstimulated and restimulated samples, since porcine proliferating lymphocytes exhibit forward scatter of a higher intensity (24) (example shown in Figures 3A, B).

Lymphocyte activity in the peripheral blood was assessed using PBMCs derived from comparators immunized with “RepRNA + c-di-AMP”. Polyclonal restimulation of PBMCs in vitro employed Con A to ascertain the responsiveness of the cells. A high and comparable percentage of FSC^high cells was observed with cells from animals in the vaccination groups, particularly at day 3 (Figure 3C, upper panel, Con A). This indicated that the PBMCs retained a good and comparable intrinsic capacity for proliferation, whether they had been exposed in vivo to “RepRNA + c-di-AMP”, “[RepRNA/PEI] + c-di-AMP” or “VRP”. When antigen-specific stimulation of the PBMC was assessed with E2, HA or NP antigens, PBMCs derived from pigs receiving “[RepRNA/PEI] + c-di-AMP” or “VRP” failed to display a clear enriched FSC^high population in peripheral blood on day 3 and 5 (Figure 3C, upper panel).

Lymphocytes were also derived from draining LNs. Their responsiveness to Con A was similar to that observed in PBMCs (Figure 3C, lower panel), implying a preserved capacity to proliferate for all animals in the study. In contrast to PBMCs, a more notable antigen-specific stimulation was observed, particularly when using NP antigen. Comparators immunized with “RepRNA + c-di-AMP” showed low evidence for specific cell proliferation (low percentage of FSC^high cells when resimulated in vitro with E2 and HA at day 3 and 5, whereas an enhanced proliferation of cells was observed in two of five pigs following NP restimulation at day 3) (Figure 3C, left lower panel). No significant improvement was seen for the group “[RepRNA/PEI] + c-di-AMP”. In contrast, cells derived from draining LNs of pigs immunized with VRP gave a clear enriched FSC^high population, most notable at day 5 (E2 (***), HA (**) and NP (**)). Altogether, these results show a clear distinction between pigs immunized with naked RepRNA or polyplexes (weak FSC^high cell enrichment at day 3) and pigs immunized with VRPs (specific FSC^high cell enrichment occurring in the draining LNs at days 3-5 p.i.). This increased FSC^high cell population strongly suggested a specific T- and B-cell response against influenza virus antigens.

In order to confirm the above results, we assessed cell proliferation by reduction in the CellTrace signal from preloaded cells (due to cell division distributing the dye to daughter cells). Figure 3D displays examples of the dot plots gated on FSC^high cells from draining LNs, following NP restimulation. Cells from comparators receiving “RepRNA + c-di-AMP” lacked clear evidence for specific proliferation – the percentage of CellTrace^low was low, remaining stable (pig #1542) or only slightly increased (pigs #1548 and #1552) between days 3 and 5. In contrast, cells from animals receiving “[RepRNA/PEI] + c-di-AMP” showed proliferation – the percentage of CellTrace^low increased between day 3 and 5. Cells from animals immunised with VRP showed a higher proliferation – the percentage of CellTrace^low was high, with strong increases between day 3 and 5. Altogether, these results established that RepRNA must be correctly packaged, either in PEI-based polyplex or in VRPs, for efficient delivery in vivo leading to intracellular release of the RepRNA for translation of the encoded antigens to promote specific cell proliferation in secondary lymphoid organs. Of note, VRP were clearly more efficient at genome delivery than the PEI-based polyplexes since the latter required 10³ times more RNA than the VRP for comparable immune induction (see materials and methods).

Considering the apparent advantage of VRP delivery leading to RepRNA translation of the encoded influenza virus antigens inducing specific immune responses, the question arose as to the compartmentalisation of these responses. Firstly, CD4⁺ T cell proliferation was measured using in vitro restimulation with E2, HA, and NP recombinant proteins, using again the CellTrace signal from preloaded cells as readout. By day 5, a moderate proliferation of CD4⁺ T cells was observed in unstimulated wells from certain comparators, up to seven rounds of cell division. This background proliferation was slightly increased after in vitro restimulation with NP. In comparison, peripheral blood cells from pigs receiving polyplexes or VRPs also showed a slight background proliferation of CD4⁺ T cells in unstimulated wells. As with the comparators, this proliferation increased in response to influenza virus NP, more substantially for “[RepRNA/PEI] + c-di-AMP” group (Figure 4A, Peripheral blood).

The above results with LN cells from the “RepRNA + c-di-AMP” group proved to be relatively minor compared to results obtained with cells from pigs receiving polyplexes and VRPs. Notably for the latter, CD4⁺ T cells displayed a very strong proliferation in response to the E2, HA and NP antigens. This was observed as a very high percentage of division peaks D5 – D7 for E2 (57.3%), HA (58.2%), and NP (56.5%), compared to background level for the unstimulated “-” (17.9%) (Figure 4A, Draining LNs). Dot plots for these results are shown in Figure 4B for three animals from each group. Altogether, these results demonstrate that immunization with RepRNA packaged in PEI-based polyplex or VRP promotes translation of encoded antigens, likely by DCs, in pigs.

In addition to the above assessment of CD4⁺ T cell responses, the same analysis was applied to cytotoxic CD8⁺ T cells. Again, PBMCs and cells isolated from draining LNs of most pigs immunized with “RepRNA + c-di-AMP” failed to display a sufficient FSC^high population for accurate analysis (only one out of five pigs). Restimulation of its CD8⁺ T cells with E2 and NP induced a weak cell proliferation with peaks D5 – D7 for E2 (14.9%) and for NP (10.8%) compared to background, unstimulated levels (“-”, 5.3%) (Figure 5A, top panel, “Peripheral blood”). However, cells derived from draining LNs failed to induce significant anti-E2, anti-HA or anti-NP response, when compared to background proliferation in unstimulated wells (“-”.) (Figure 5A, top panel “Draining LNs” and Figure 5B). This supports the notion that specific proliferation in response to the encoded influenza virus antigens was at best moderate for the unique responder animal in this vaccination group.

Like the above comparator animals, no clear anti-E2, anti-HA, or anti-NP CD8⁺ T cell response responses were detectable in peripheral blood samples of pigs immunized with “[RepRNA/PEI] + c-di-AMP” (Figure 5A, middle panel, “Peripheral blood”). However, cells derived from draining LNs showed clear CD8⁺ immune responses against influenza antigens, and particularly for NP (increased frequency at day 5 of peaks D5 – D7 (Figure 5A, middle panel, “Draining LNs”). This is also illustrated in the dot plots shown in Figure 5B. Peaks of low CellTrace intensity were easily detected, even if at lower percentage than those measured with CD4⁺ T cells.

Also for pigs immunized with “VRP”, no clear anti-E2, anti-HA or anti-NP responses were detectable in peripheral blood samples (Figure 5A, bottom panel, “Peripheral blood”). Again, cells derived from draining LNs showed clear CD8⁺ immune responses against influenza antigens, albeit at lower percentage for NP than those measured in pigs immunized with polyplexes (Figure 5A, “Draining LNs” and Figure 5B). In conclusion, the combined observations for CD4⁺ and CD8⁺ T cells in pigs receiving RepRNA packaged in PEI-based polyplexes or VRPs, confirmed that the T-lymphocytes had clearly responded to RepRNA encoded GOI (HA and NP antigens), which could only have arisen following RepRNA translation in vivo.

CD4⁺CD8⁺ double-positive (DP) lymphocytes proliferate in response to stimulation with recall viral antigen and include most likely memory/effector T cells [for a review see (43)]. For PBMCs, a moderate proliferation of DP T cells was observed in unstimulated wells from certain comparators. This background proliferation was slightly increased after NP restimulation. However, this clear anti-NP response was higher in pigs receiving polyplexes (peaks D5–7 = 37.8%), whereas no clear response were detectable in peripheral blood samples from pigs immunized with “VRP (Figure 6A, Peripheral blood).

Once again, the above results proved to be relatively minor compared to what observed with cells isolated from draining LNs. DP T cells derived from pigs receiving polyplexes showed clear DP T cell immune responses against influenza antigens (Figure 6B). But this was moderate compared to pigs immunized with VRPs, where DP T cells displayed a very strong proliferation in response to E2, HA and NP (percentage of division peaks D5 – D7: 40.0%, 44.1%, and 37.0%, respectively) (Figure 6A, Draining LNs). This is also illustrated in the dot plots shown in Figure 6B.

Given that DP T cells displayed strong proliferations in LN samples following restimulation, we next wondered whether the number of this specific subset could increase in animals receiving either polyplexes or VRPs. Expansion could be indicative of the induction of a memory T cell response (as shown in Figure 3A, large differences could be seen in the frequency of this T cell subset depending of the experimental condition). Effectively, a significant increase in size was observed following antigen restimulation for pigs receiving VRPs, whereas three out of five pigs receiving polyplexes displayed a clear enrichment of DP T cells (Figure 6C). Taken together, these results strongly suggest that our vaccine formulations can induce an immunological T cell memory against influenza antigens.

Considering the importance of the humoral response against influenza virus, specific B cell response against the RepRNA encoded influenza virus antigens was assessed. For PBMCs derived from pigs immunized with “RepRNA + c-di-AMP” that displayed FSC^high cells, a clear anti-NP response was observed (at day 5, peaks D5-7 = 8.2%). However, this specific anti-NP response was higher in pigs receiving the antigens using VRPs (peaks D5-7 = 11.7%) (Figure 7A “Peripheral blood”).

B cells derived from the LNs of comparator pigs showed no sign of activation of E2-, NP- and HA-specific B cells, respectively. Those readouts were also negligible in pigs receiving polyplexes, with the slight exception of NP restimulation. In contrast, very strong B cell proliferation was observed in pigs receiving VRPs in response to E2, HA and particularly NP (Figure 7A, “Draining LNs”). Figure 7B recapitulates the percentages of CellTrace^low division peaks D5 – D7 for E2 (15.7% ± 0.8), HA (17.5% ± 4.2), and NP (33.8% ± 1.5), compared to background, unstimulated levels (“-”, 7.3% ± 4.0) in pigs receiving polyplexes and VRPs.

The above results clearly demonstrate that the VRP-delivered RepRNA, and to a lesser extent PEI-delivered RepRNA, translated in vivo to provide the encoded influenza virus antigens for stimulating both T- and B-lymphocyte immune responses. Accordingly, we considered that in vivo induction of anti-HA and -NP specific antibodies would be possible.

However, immunization with the comparator “RepRNA + c-di-AMP” and “[RepRNA/PEI] + c-di-AMP” were overall poor at inducing detectable anti-HA or anti-NP antibody responses; any detectable IgG titers were close to non-specific background levels observed with pre-vaccination bleeds especially with the HA antigen (Figure 7C).

Compared to the previous groups, VRP immunization induced superior humoral responses against RepRNA-encoded influenza virus antigens, detectable in all animals (Figure 7C: “VRP”, anti-HA: p<0.001; anti-NP: p=0.047). Interestingly, immunization with the HA-encoding VRP induced higher titers than those observed following immunization with the NP-encoding VRP. Moreover, tight grouping of the titers from individuals was observed for anti-HA antibodies, relating also to higher titers from all individuals (titers IgG anti-HA: 8047.9 ± 597.3; anti-NP: 1014.7 ± 957.7).

It was then considered that the anti-NP response induced by the VRPs might be influenced by the influenza virus strain employed for the vaccination or analyses. In order to understand better why anti-NP IgG titers were more moderate than the anti-HA IgG titers following VRP immunization, we performed ELISA by coating with recombinant NP proteins from two different influenza virus strains: A/Brisbane/10/2007 (H3N2) and A/Thailand/1(KAN-1)/2004 (H5N1). This had no influence on the anti-NP readout compared with the anti-HA readout (data not shown).

Taken together, these results showed that VRP immunization induced strong anti-HA and anti-NP humoral responses. Overall, this confirms the requirement of RepRNA to be properly packaged for efficient translation of the GOI.