FluB-RAM and FluB-RANS: Genome re-arrangement as safe and efficacious live attenuated influenza B virus vaccines

Influenza B virus (IBV) is considered a major respiratory pathogen responsible for seasonal respiratory disease in humans, particularly severe in children and the elderly. Seasonal influenza vaccination is considered the most efficient strategy to prevent and control IBV infections. Live attenuated influenza virus vaccines (LAIVs) are thought to induce both humoral and cellular immune responses by mimicking a natural infection, but their effectiveness have recently come into question. Thus, the opportunity exists to find alternative approaches to improve overall influenza vaccine effectiveness. Two alternative IBV backbones were developed with re-arranged genomes, re-arranged M (FluB-RAM) and a re-arranged NS (FluB-RANS). Both re-arranged viruses showed temperature sensitivity in vitro compared to the WT type B/Bris strain, were genetically stable over multiple passages in embryonated chicken eggs and were attenuated in vivo in mice. In a primeboost regime in naïve mice, both re-arranged viruses induced antibodies against HA with hemagglutination inhibition titers considered of protective value. In addition, antibodies against NA and NP were readily detected with potential protective value. Upon lethal IBV challenge, mice previously vaccinated with either FluB-RAM or FluB-RANS were completely protected against clinical disease and mortality. In conclusion, genome re-arrangement renders efficacious LAIV candidates to protect mice against IBV.

Although vaccination is the most effective strategy to ameliorate the impact of influenza infections, the incidence of IBV shows an increasing trend. This is in part due to vaccine mismatch in trivalent vaccine formulations that contain only one IBV strain from one of the lineages [15][16][17][18][19][20][21][22]. These observations underscore the importance of including both IBV lineages in seasonal vaccine formulations as it is the case in several of the most recent FDA-approved quadrivalent vaccines [23]. However, additional efforts are warranted in order to improve vaccine protection against IBV. Live attenuated vaccine platforms have been among the most explored over the years (reviewed in [24]). In addition to the coldadapted LAIVs developed in the 60's that form the basis of the current LAIVs approved for human use, alternative LAIV approaches have been developed that include modifications and deletions to the NS1 gene segment, generation of M2 deficient viruses, alternative virus backbones with temperature sensitive phenotypes, among others [25][26][27][28][29][30][31]. We have previously shown that genome re-arrangement is a suitable strategy for the development of influenza A virus LAIVs [32]. In the present study, we expanded these studies into IBV and produced two distinct genome re-arrangements in the backbone of the B/Brisbane/60/2008 strain (Victoria lineage). The FluB-RAM re-arrangement involved producing a chimeric segment 1 that encodes PB1 and BM2, and a series of mutations in segment 7 to completely abrogate expression of BM2 from the latter. The FluB-RANS re-arrangement used a similar strategy whereby NS2 was cloned downstream of PB1 and segment 8 contains multiple mutations that precludes NS2 expression. Safety and efficacy of the FluB-RAM and FluB-RANS viruses was evaluated in DBA/2J mice [27,31,33,34]. Both vaccine candidates were immunogenic and effectively protected mice against homologous lethal IBV challenge.

Materials and Methods
Cells and eggs. Madin Darby canine kidney (MDCK) cells and 293T cells (ATCC CRL-3216) were used for reverse genetics of virus strains. Specific pathogen free embryonated chicken eggs (ECEs) used for virus propagation and stock titration were obtained from Charles Rivers (Wilmington, MA).

Rescue of FluB-RAM and FluB-RANS viruses with re-arranged genomes.
Recombinant viruses were rescued by reverse genetics as previously described [35]. We employed a 6+2 method whereby 6 plasmids containing 6 cDNA copies of the wild type gene segments from the B/Brisbane/60/2008 were mixed with the corresponding pair of plasmids (either pSCG-PB1BM2 and pSCG-BM1-∆M2, or pSCG-PB1BNS2 and pSCG-BNS1-∆NEP) to produce the B/Bris re-arranged M (FluB-RAM) and B/Bris re-arranged NS (FluB-RANS) viruses, respectively. The identity of the rescued viruses was confirmed by Sanger sequencing (Psomagen). The recombinant viruses were propagated and titrated in 11-dayold SPF ECEs incubated at 33°C for 48 h. Virus stocks were stored at -80°C until further use. These stocks constitute the first passage in ECEs (E1).

Stability of FluB-RAM and FluB-RANS viruses through serial passages in ECEs.
Serial passages were performed in 11-day-old SPF embryonated chicken eggs as follows: Serial 10fold dilutions from FluB-RAM and FluB-RANS E1 viruses were prepared in 1X phosphate buffered saline (PBS) and 100 µL from each dilution were inoculated into each of five ECEs through the allantoic cavity to generate E2. The inoculated ECEs were incubated at 33°C for 48 h. Allantoic fluids were then tested for hemagglutination activity by the hemagglutination (HA) assay. Fluids collected from the previous to the last dilution with 5/5 embryos positive for HA activity were pooled together and used to prepare 10fold dilutions to inoculate the next set of embryos. The same procedure was repeated until 5 passages had been completed, generating E6. Aliquots from each passage were stored at -80°C until needed. RNA was extracted from fluids collected at each passage and from the original virus stock using the MagNA Pure LC Total Nucleic Acid Isolation Kit (Roche, San Francisco, CA). The PB1, M and/or NS gene segments were amplified by RT-PCR using SuperScript III One-Step RT-PCR System with Platinum Taq DNA Polymerase (ThermoFisher Scientific). Sanger sequencing (Psomagen) was then performed from the resulting RT-PCR products to confirm the re-arrangement at the PB1 gene segments and the presence of the introduced mutations within the M and NS gene segments, respectively. Multi-segment RT-PCR (using the same RT-PCR system) was performed as previously described [36] for full genome sequencing using next generation sequencing (NGS) as follows: Amplicon libraries were prepared using the Nextera XT DNA library preparation kit (Illumina, San Diego, CA) following the manufacturer's protocol. Barcoded libraries were multiplexed and sequenced on the high-throughput Illumina MiSeq NGS platform (Illumina) in a paired-end 150-nucleotide run format. De novo genome assembly was performed as described previously [37].
Virus Growth Kinetics. MDCK cells were seeded in 6-well plates and incubated overnight at 37°C, under 5% CO2. The next day, cells were inoculated with 0.01 MOI of either the B/Bris WT, FluB-RAM, or FluB-RANS virus contained in 500 µL, each in triplicate wells. Three set of plates were prepared for each virus. Inoculated cells were incubated for 1 h at 35ºC/5% CO2 with gentle rocking of the plates every 15 min. Subsequently, the virus inoculum was removed, and the cells were washed twice with 1X PBS and replenished with 2mL of fresh Opti-MEM (Gibco, ThermoFisher Scientific) supplemented with 1X antibiotic/antimycotic solution (Gibco, ThermoFisher Scientific) and 1µg/mL of L-1-tosylamido-2-phenylethyl chloromethyl ketone (TPCK) treated-Trypsin. Plates were set to incubate at either 33, 35 or 37°C, 5% CO2. Supernatants (200 µL) were collected at 0, 12, 24, 48, 72 and 96 hrs. post-inoculation (hpi) and stored at -80°C until processed. The amount of virus present in the collected samples was titrated by TCID50 in MDCK cells determining virus presence by HA assay. Virus titers were calculated using the Reed and Muench protocol [38] and plotted as the mean TCID50/mL ± SD.
Mouse Studies. Male and female DBA/2J mice (5 weeks old) were purchased from Jackson's Laboratories (Bar Harbor, ME) and raised until 7 weeks of age. Mice were housed in negative pressure caging in the Davison Life Sciences Complex, University of Georgia and were provided food and water ad libitum for the duration of the experiment.
Vaccine efficacy. Mice from the vaccine safety study (n=8/group, ½ females) were challenged i.n. with a lethal dose (10 7 EID50/mouse) of the B/Brisbane/60/2008 PB2-F406Y (B/Bris/ F406Y) strain [27] contained in 50 µL. A subset of mice in the mock group (n=8, ½ female) remained unchallenged and served as negative controls. Mice were monitored twice daily to record clinical signs and mortality for up to 14 days post-challenge (dpc). Body weight was recorded for up to 12 dpc. At 14 dpc, survivors were anesthetized, terminally bled to collect sera, and subsequently humanely euthanized (Fig. 2).
Hemagglutination Inhibition (HI) assay. Sera were prepared from whole blood collected at 20 dpb (n=4/group, except for FluB-RAM) and 14 dpc (n=8/group) by centrifugation at 1000 x g for 15 min at room temperature. The sera were treated with receptor destroying enzyme (RDE) and the HI assay was performed in V-bottomed microtiter plates, using 4 hemagglutination units (HAU) of viral antigen per 25 µL, as recommended by the OIE [1], using a suspension of turkey red blood cells (0.5%). HI titers were plotted using Prism v9 (GraphPad, San Diego, CA). The limit of detection was at dilution 1/10, samples with undetectable titers were assigned a dilution value of 1/8 for statistical purposes.
Microarray for IgG and IgA determination. Sera collected at 20 dpb and 14 dpc, and nasal washed collected at 14 dpc were analyzed through protein microarrays to determine anti-HA, -NA and -NP IgG and IgA levels from multiple Victoria-and Yamagata-like IBVs (Table 1). Purified IBV protein antigens were purchased from Sino Biological (Wayne, PA) ( Table 1). Microarrays were carried out as described elsewhere [39]. Results are expressed as the group mean fluorescence intensity (MFI) ± SD. The higher the MFI, the more Abs bound to a particular antigen. MFI were plotted using Prism v9 (GraphPad).  (Fig 1-A). Segment 1 is further modified with the inclusion of a linker peptide sequence (G4S) and the Tav 2A protease sequence between the PB1 and BM2 ORFs. The strategy leads to a chimeric polymerase PB1 subunit protein carrying the G4S linker and the Tav 2A protein sequences and the BM2 protein but with N-terminal proline. Segment 7 is mutagenized to eliminate the codon for the first methionine in the BM2 ORF, the inclusion of an additional stop codon in the BM1 ORF and two early stop codons in the BM2 ORF resulting in complete obliteration of BM2 expression from its cognate segment. In the second strategy, FluB-RANS, (Fig 1 In order to quickly visualize whether the viruses contained the corresponding re-arranged PB1 gene segments, RT-PCR targeting the region containing the BM2 or BNS2 insertions was performed (Fig 1-B). The RT-PCR showed that the FluB-RAM and FluB-RANS viruses carry PB1 segments with the expected size changes (402 and 444 base pairs, respectively). The sizes of the amplified fragments were consistent with those of the positive control reverse genetics plasmids used to generate the corresponding viruses (Fig 1-B prevent the expression of BM2 or BNS2, respectively (Table 2). To further evaluate genome re-arrangement stability and the mutations introduced in the M and NS gene segments, five serial passages from an E1 stock were performed in ECEs as described above. Segments 1 and 7 from the FluB-RAM virus and 1 and 8 from the FluB-RANS virus from each passage were amplified by RT-PCR and sequenced by Sanger (Table 2). In addition, NGS was performed on the last passage virus and compared to the original stock virus from passage 1. Both Sanger sequencing and NGS confirmed the presence of the BM2 or BNS2 downstream of the PB1 gene segment in either FluB-RAM or FluB-RANS, respectively. Sequencing results also confirmed the maintenance of the mutations introduced in either the M or NS gene segment from the corresponding virus. These results highlight the stability of the two genome re-arrangement strategies introduced in the B/Bris genome.  FluB-RAM FluB-RANS B/Bris wt as the mean TCID50/mL ± SD. Samples with undetected virus titers were assigned the limit of detection value (0.699 TCID50/mL). Data analysis and graphs were prepared using Prism v9. Curves were analyzed using multiple t-tests followed by the Holm-Sidak method to correct for multiple comparisons. Significant differences from the WT B/Bris are denotated by stars (*). * = p < 0.05, ** = p < 0.01, *** = p < 0.001, and **** = p < 0.0001. In order to determine the growth of the re-arranged viruses at different temperatures, MDCK cells were infected with either B/Bris WT, FluB-RAM, or FluB-RANS at 0.01 MOI. The growth kinetics for each virus was assessed at 33°C, 35°C, and 37°C for up to 96 hpi (Fig 1-C). Compared to the B/Bris WT virus, both FluB-RAM and FluB-RANS showed significant lower replication at all three temperatures. Of note, the replication of the FluB-RANS virus was lower than that of the FluB-RAM virus at either 33°C or 35°C and were almost undetectable at 37°C compared to the B/Bris WT and FluB-RAM viruses. These results demonstrate that both FluB-RAM and FluB-RANS are

FluB-RAM and FluB-RANS viruses show differences in attenuation.
The safety and immunogenicity of the FluB-RAM and FluB-RANS viruses was tested in DBA/2J mice, a small animal model susceptible to influenza B viruses without further adaptation [27].
DBA/2J mice (7-week-old, male and female) were inoculated with 10 6 EID50/mouse i.n. following a prime/boost strategy 20 days apart with the corresponding re-arranged virus (Fig 2-A). As a control, a group of mice were inoculated with the B/Bris WT virus (10 6 EID50/mouse i.n.). Prime vaccination with the FluB-RANS resulted in neither clinical signs nor body weight changes in both male and female mice (Figs. 2-B). In contrast, male mice primed with the FluB-RAM virus showed an average of ~10% body weight loss between 7-9 dpv, but started to recover from 10 dpc onwards, whereas female mice showed a slight drop in body weight (<5%) on 7 dpv and quickly  (Fig 3-A top). Analysis of all the anti-Victoria HA responses combined and comparison between groups confirmed that the FluB-RANS vaccine induced higher responses than the FluB-RAM vaccine (p<0.0001) (Fig 3-A bottom). When looking at the anti-Yamagata responses, FluB-RANS showed numerically higher anti-HA IgG responses that the other vaccine groups; however, none of those were statistically significant (p>0.05) due to the high variability between samples within the group (Fig 3-B top). When the responses against all the Yamagata lineage HAs were combined, the FluB-RANS group had significantly higher IgG response compared to the other vaccine groups (p<0.0001 and p<0.0001, respectively), (Fig 3-B bottom). In contrast, anti-HA IgA responses were numerically higher for both IBV lineages in samples from the FluB att group, but not significantly different than the other groups (p>0.05) (Fig 3-C and -D). Combining the responses against all the Victoria or the Yamagata HA antigens, FluB att induced significantly higher IgA responses than FluB-RAM (p<0.0001 and p<0.0001, respectively) and FluB-RANS (p=0.0021 and p=0.0097, respectively) against both lineages (Figs 3-C and 3-D bottom). Interestingly, and despite showing the least attenuation, FluB-RAM samples showed the lowest levels of anti-HA IgG and IgA responses among the three vaccine groups (Fig 3). to compare responses between groups was performed using 2way ANOVA followed by a Tukey's test for multiple comparisons. Significant differences between groups are denotated by stars (*). * = p < 0.05, ** = p < 0.01, *** = p < 0.001, and **** = p < 0.0001.
Differences in serological responses against NA and NP were also observed (Fig 4). Anti-NA IgG and IgA showed a trend towards higher responses in samples from the FluB att group, although most of them were not statistically significant (p>0.05). However, the FluB att vaccine induced significantly higher anti-B/Phuket/3073/2013 IgG response than the other two groups (p=0.005 and p=0.0016, respectively). It was noted that the NA antigen derived from the B/Phuket/3073/2013 provided more reliable signals with low background noise (Fig 4-A and -B). In contrast, the NA antigen derived from B/Brisbane/60/2008 reacted poorly in the array when probing for IgG responses and it provided high background signal when probing for IgA responses. The trend of anti-NP IgG responses was also numerically higher in samples from the FluB att group (Fig 4-C).
Anti-NP IgA serum responses were low, except the serum from one female in the FluB-RAM group, which clearly show reactivity well above background (Fig 4-D). Interestingly, samples from

FluB-RAM FluB-RANS FluB att PBS
Protection efficacy of the re-arranged viruses was tested using a lethal challenge dose of 10 7 EID50/mouse of B/Bris/PB2 F406Y strain, administered i.n. [27] 3 weeks after boost. Mice in the three vaccine groups (FluB att data included for comparison) were fully protected as no signs of disease and no mortality were observed (Fig 5-A and -B). In contrast, PBS-vaccinated/challenged mice showed severe body weight loss. Only one female (out of 8) and none of the male mice survived in the PBS-vaccinated/challenged group, consistent with previous studies [27,31]. Body weight values were graphed as the group mean ± SD. Survival data were analyzed using the Log-rank test. HI titers are represented as the group mean ± SEM. Data analysis and graphs were prepared using Prism v9. HI titers were compared between groups though a 2Way ANOVA followed by a Tukey's test for multiple comparisons. Significant differences between group are denotated by stars (*). * = p < 0.05, ** = p < 0.01, *** = p < 0.001, and **** = p < 0.0001.

Qualitative differences in humoral and mucosal responses among different vaccine groups at 14 dpc.
HI responses at 14 dpc were similar among the three vaccine groups with a mean antibody titer increase of about 1 Log2 compared to post-boost HI titers, particularly in samples from the re-arranged vaccine groups (Fig 5-C). No statistically significant differences were observed between vaccine groups with trends like those observed post-boost. With respect to the re-arranged vaccine groups, the data showed better responses in female mice than in male mice. Further analyses of serum and nasal wash samples collected at 14 dpc revealed recall IgG and IgA anti-HA responses against both the Victoria-and Yamagata-lineage antigens (Fig 6). As expected, the reactivity of serum samples from all vaccine groups against Victoria lineage HA antigens was 1.5-2-fold higher than to those of the Yamagata lineage (Fig 6-A and -B). Interestingly, the HA1 antigen derived from B/Hong Kong/05/1972 (before the split of the two IBV lineages) reacted well with samples from all groups (Fig 6-A), whereas the HA1 antigen from B/Florida/4/2006 and B/Utah/02/2007 (Yamagata lineage) show the lowest reaction with the serum samples (Fig 6-B). Of note, the fulllength HA of B/Florida/4/2006 reacted well with samples from all three vaccine groups (Fig 6-B). The pattern of anti-HA Victoria lineage serum IgA was similar among all vaccine groups where differences in reactivity could be attributed to the different antigens in the array (Fig 6-C). Postchallenge serum IgA responses against Yamagata-lineage HA antigens showed reactivity patterns attributed also to the different antigens but trending towards better reactivity in samples from the FluB-RANS group (Fig 6-D). FluB-RAM FluB-RANS FluB att PBS Significant differences between groups are denotated by stars (*). * = p < 0.05, ** = p < 0.01, *** = p < 0.001, and **** = p < 0.0001.  Perhaps the most striking differences in IgG and IgA profiles were observed in the NW samples (Fig 7). The trend of serum IgG, but not IgA, anti-HA responses at 20 dpb (Fig 3) translated similarly in the NW samples for both IgG and IgA responses (Fig 7). Thus, a trend of higher IgG and IgA anti-HA responses were observed for samples of the FluB-RANS group, whereas those from the FluB-RAM and FluB att had the lowest of such responses and were like each other. The FluB-RANS group displayed significantly higher general anti-Victoria IgG readings than both FluB-RAM and FluB att (p<0.0001) (Fig 7-A bottom); and higher anti-Yamagata IgG response than FluB att (p=0039). IgA responses detected in the NW material appeared to be more robust than the IgG responses. When comparing groups within the same HA antigen, the FluB-RANS group had significantly higher anti-  ANOVA followed by a Tukey's test for multiple comparisons. Significant differences between groups are denotated by stars (*). * = p < 0.05, ** = p < 0.01, *** = p < 0.001, and **** = p < 0.0001.

MFI
Differences were also observed at 14 dpc in the pattern of IgG and IgA serum and NW reactivity against the NA and NP antigens on the array (Fig 8). Anti-NA and anti-NP IgG serum responses at 14 dpc were similar among vaccine groups (Fig 8-A) but were close to background against NA and low against NP in NW at 14 dpc (Fig 8-B). The pattern of IgA serum and NW against NA and NP at 14 dpc (Fig 8-C and -D) followed the patterns observed at 20 dpb despite their initial low signals (Fig 4-B and -D). Thus, samples from the FluB-RANS group tended to have the lowest IgA responses in both serum and NW samples, whereas those from the FluB-RAM group had the overall highest responses, particularly against NP; although, not statistically different (p>0.05). Overall, the pattern of anti-NA and anti-NP responses post challenge showed opposite trends with respect to the anti-HA responses at 14 dpc (Figs 6, 7 and 8).

Discussion
Influenza virus genome re-arrangement is a viable alternative for the development LAIV vaccines. We previously showed such potential within the background of a H9N2 virus carrying full-length H9 and H5 HA proteins while maintaining a full set of the remaining viral proteins [32]. Similar to the approach followed in this study, the NS2 ORF from the H9N2 IAV was inserted downstream the PB1 gene, whereas the NS segment was modified to carry the NS1 ORF and a prototypic H5 HA ORF with a modified monobasic cleavage site. The H9N2/H5 virus showed successful protection against lethal highly pathogenic avian influenza A/Vietnam/1203/04 (H5N1) in mice and ferrets [32]. The same strategy was used to generate H9N2 viruses successfully expressing enhanced green fluorescent protein (eGFP) and secreted Gaussia luciferase (GLuc), and a 2009 prototypic H1N1 virus expressing GLuc [32,40]. Based on these previous studies, we developed the FluB-RAM and FluB-RANS LAIV candidates, with the exception that these viruses do not express foreign antigens. The FluB-RAM and FluB-RANS viruses remained stable after six serial passages in ECEs as shown by RT-PCR and Sanger and NGS sequence analyses.
In vitro growth of FluB-RAM and FluB-RANS viruses was impaired under multiple temperature conditions in MDCK cells compared to the WT B/Bris strain. Of the two rearranged viruses, FluB-RANS grew at lower titers than FluB-RAM and its growth at 37°C was barely over the limit of detection at 72 hpi (Fig 1-C). However, both re-arranged viruses reached titers of a least 10 8 EID50/mL in ECEs that would make them suitable as vaccine candidates. The growth kinetics results were consistent with the observations during the in vivo safety assessment. Both FluB-RAM and FluB-RANS were attenuated in comparison to the B/Bris WT. The safety profile of FluB-RANS showed more attenuation than another LAIV candidate, FluB att with 4 amino acid mutations in PB1 (E48K, K391E, E580G and S660A), resulting in no noticeable signs of disease and no body weight changes. In contrast, the FluB-RAM virus induced some body weight loss, particularly in male mice with one of those having to be euthanized. Nevertheless, the clinical signs induced by FluB-RAM inoculation in male mice were significantly lower, whereas they were almost nonexistent in female mice compared to those observed with the B/Bris WT strain (Fig 2-B).
It is important to note that during the process of testing safety and efficacy of the different vaccines, we observed biological sex as a variable for susceptibility to IBV. In our experience, male mice were more prone than female mice to develop more significant signs of disease and mortality upon IBV infection (Figs 2-B, 5-A, and 5-B). In addition, female mice, but not male mice showed a biphasic curve of associated clinical signs after IBV challenge, with an initial phase of pronounced body weight loss, recovery close to initial body weight and then a second phase of mild body weight loss before a second recovery phase (Fig 5-A). Sex differences related to susceptibility to IAV have been extensively characterized [41,42]. However, previous studies have determined that female mice are more susceptible than males to IAV infection. In this regard, the differences in susceptibility of male versus female mice infected with IBV of this report follow the pattern observed in humans where biological males are more prone to hospitalization due to influenza than biological females. In addition, although non-statistically significant, antibody titers after boost vaccination and after challenge showed a trend towards higher responses in female than in male mice (Fig 2 and 5 and data not shown). These observations are consistent with previous studies assessing the response to vaccination in humans and mice that revealed higher antibody responses, higher B cell responses, higher cross-reactive antibodies, and higher CD4+ T cell numbers in females compared to males [41][42][43][44][45]. Thus, understanding sex as variable to study IBV susceptibility and vaccine responses is warranted but beyond the scope of this report.
Comparing the results from the present study to previous observations with the FluB att virus published elsewhere but part of the same study, the rearranged FluB-RAM and FluB-RANS viruses induced comparable HI antibody levels within 1 Log2 difference of each other, both before and after challenge (Fig 2 and 5). Further, qualitative differences in IgG and IgA responses were observed among different vaccine groups as explore using a protein microarray (Fig 3 and 4). It was interesting to observe that the most attenuated virus, FluB-RANS led to overall higher anti-HA serum IgG responses before challenge. In contrast, anti-HA serum IgG responses from the FluB-RAM were among the weakest before challenge despite the virus being the least attenuated. The pattern of anti-HA serum IgG in samples from the FluB att were intermediate between the two re-arranged vaccine groups. However, serum samples from the FluB att group were consistently higher for IgA against HA and for IgG and IgA against NA and NP antigens in the array. Further, overall IgG and IgA responses against NA and NP from the FluB-RANS group were among the weakest. Despite these qualitative differences, both re-arranged viruses protected mice against lethal challenge with the B/Bris/PB2-F406Y strain. NW antibody responses after challenge were of particular interest since they reflect recall antibody responses to the site that would most efficiently prevent infection. In NW samples from the FluB-RANS group, anti-HA IgG and IgA responses were particularly prominent, but anti-NA/NP IgA responses were the weakest compared to other groups. Interestingly, anti-NA/NP IgA responses were more prominent after challenge in serum and NW samples from the FluB-RAM group. Thus, we observed opposite patterns between anti-HA and anti-NA/NP responses for the FluB-RANS and FluB-RAM groups and intermediate patterns for the FluB att group. These observations are significant because they suggest that humoral responses against different IBV antigens are not equally impacted by the different LAIV backbones. Despite the relatively less attenuation of the FluB-RAM virus, it could be useful in dose sparing situations and/or in the presence of pre-existing immunity as complement boost vaccine. Previous studies have suggested that priming with a LAIV followed by a killed virus vaccine leads to more complete protective responses than prime-boost strategies using a single type of vaccine against the 2009 pandemic H1N1 virus. More relevant to this report, vaccination with a seasonal H1 LAIV (pre-2009 H1N1 antigen) followed by a boost with pandemic H1 LAIV led to more robust protective responses than either vaccine administered twice [46,47]. Thus, it is tempting to speculate that one or more LAIV platforms could be used in prime-boost approaches that would improve the protective response of currently approved vaccines against IBV.
Author Contributions: DRP, DSR and SCG conceptualized. SCG, DRP and DSR designed the experiments. SCG performed cloning for the generation of pSCG-PB1BM2, pSCG-PB1BNS2, pSCG-BM1-∆M2, and pSCG-BNS1-∆NEP reverse genetics plasmids, virus rescue, and growth kinetics. SCG and GG viral sequencing. SCG, CJC, JM performed in vivo experiments, sample collection. SCG performed sample processing. AJ, RN and HD performed influenza antigen microarray. SCG and DRP performed data analysis. SCG and DRP wrote the manuscript. All authors approved the final version of the manuscript.