Mucosal Immunity against Neuraminidase Prevents Influenza B Virus Transmission in Guinea Pigs

Recently, the protective effect of anti-neuraminidase immunity has been highlighted by several studies in humans and animal models. However, so far the role that anti-neuraminidase immunity plays in inhibition of virus transmission has not been explored. In addition, neuraminidase has been ignored as an antigen for influenza virus vaccines. We show here that neuraminidase-based vaccines can inhibit the transmission of influenza virus. Therefore, neuraminidase should be considered as an antigen for improved influenza virus vaccines that not only protect individuals from disease but also inhibit further spread of the virus in the population.

sites, leading to antigenic drift that renders the previous year's vaccines ineffective and forces the annual reformulation and readministration of influenza vaccines (2,3). Recently, it has been shown that natural infection, but not vaccination, results in the generation of antibody responses that target the second surface glycoprotein, the immunosubdominant neuraminidase (NA) (4)(5)(6). The influenza virus NA is subject to reduced immune pressure which might contribute to slower amino acid changes in antigenic sites (7,8). As such, anti-NA antibodies have been shown to broadly bind heterologous NAs and can protect mice from lethal influenza A virus challenge (4,9). Hence, antibodies that target the NA have the potential to be broadly protective within influenza virus subtypes.
The viral NA is a tetrameric type II transmembrane protein that has receptordestroying activity, which is integral to the viral life cycle. Upon initial entry to the host, viruses encounter mucosal barriers that prevent the viral HA from attaching to sialic acid receptors found on the host cell surface (10)(11)(12). Mucosal fluids naturally express decoy receptors, such as heavily glycosylated mucins, that can neutralize virus by binding the HA before the virus has had the chance to reach the host cell (10,13). When this occurs, the viral NA can release virus particles that have become bound to these mucosal barriers by cleaving terminal sialic acid from decoy receptor glycans, allowing the HA to reach and bind the sialic acid receptors required for attachment to host cells (10,11). Following a successful replication cycle, virus particles begin to bud from the host cell membrane. During this process, the HA of newly formed virus particles can become reattached to sialic acids still expressed on the surface of the host cell, preventing the release of newly formed viral particles (14). However, the viral NA cleaves terminal sialic acids from the host cell, allowing the release of new viral particles (15). Finally, once viruses are released from infected cells, viral particles can form aggregates through interactions between sialylated HAs or other glycoproteins present in the mucus fluid. Individual viral particles can be released from these aggregates by the enzymatic activity of the NA (14,16,17).
Evidently, the NA has integral roles in the viral life cycle, from the point of first attachment to the final dispersal of nascent viral particles. Antibodies that target the NA may inhibit viral egress from infected cells and subsequent virus transmission. Indeed, vaccination regimes that induce anti-NA antibodies have been shown to induce broad protection in mice (18)(19)(20)(21)(22)(23), and the presence of anti-NA immunity has been shown to significantly reduce virus shedding and morbidity in humans (24)(25)(26)(27)(28). Although preliminary studies successfully explored the use of NAs as vaccine candidates (21-23, 25, 29-31), the role of anti-NA antibodies in preventing influenza virus transmission is relatively undefined. Given that influenza B viruses predominantly circulate in human hosts, there is a potential to eradicate this pathogen by using vaccination regimes that severely impact transmission. Hence, in this study, we assessed anti-NA immunity, induced by a recombinant NA protein vaccine candidate, for its ability to broadly inhibit influenza B virus transmission in the guinea pig model. Therefore, our work addresses all three research areas of the National Institute of Allergy and Infectious Disease (NIAID)'s strategic plan toward a universal influenza virus vaccine (32).

RESULTS
Intranasal vaccination with B/Malaysia/2506/2004 NA prevents homologous transmission from vaccinated donor guinea pigs to naive recipients. Preliminary studies have indicated that NA recombinant protein vaccination via the intranasal (i.n.) route leads to reduced weight loss in influenza virus-infected mice compared to mice vaccinated via the intramuscular (i.m.) route (20). In this study, we assessed the roles of systemic and mucosal immunity in preventing influenza B virus transmission in the guinea pig model of influenza virus transmission. Although guinea pigs do not display any signs of disease, they are a good model for assessing influenza virus transmission, including in the context of vaccination (33,34 (Fig. 1A, B, and E). In the airborne transmission model, the transmission of B/Malaysia/2506/2004 virus was observed in all guinea pig pairs in which the donor had been vaccinated i.m. with an irrelevant protein (Fig. 1A). However, only 50% of donors vaccinated i.m. with the B/Malaysia/2506/2004 NA were able to transmit the homologous virus to their naive recipient partner (Fig. 1B).  immunity against a heterologous influenza B virus strain. It has been previously shown that vaccination with recombinant influenza B NA results in broadly protective heterologous antibody responses against both B/Yamagata/16/1988-like and B/Victoria/2/ 1987-like influenza B viruses in the mouse model (20). This finding was further corroborated by the existence of highly cross-reactive murine and human monoclonal antibodies that bind to conserved epitopes (4,9). Given these data, we also wanted to assess whether i.n. vaccination with B/Malaysia/2506/2004 NA could induce broadly inhibitory anti-NA antibody responses capable of preventing heterologous influenza B virus transmission in the guinea pig model. To assess this, we collected prechallenge serum from NA-vaccinated guinea pigs and measured their antibody responses toward a panel of purified influenza B viruses. We found that guinea pigs vaccinated with B/ Malaysia

DISCUSSION
In this study, we show that a vaccination regime, designed to induce mucosal immunity to the influenza virus NA, adversely affects the efficiency of subsequent influenza B virus infection and transmission in the guinea pig model. We demonstrate that prior NA vaccination of a virus-infected donor guinea pig via the i.n. route can seemingly prevent onward transmission of a homologous influenza B virus (0/3 transmission events with B/Malaysia/2506/2004 and 1/3 with B/Florida/04/2006). Interestingly, vaccination via the i.m. route reduced subsequent viral titers in infected donor animals but only partially prevented onward virus transmission to a naive recipient. Furthermore, we identified that i.n. vaccination with the NA protein induces mucosal antibody responses with increased magnitude and NA inhibitory activity compared to those induced by i.m. vaccination. Not surprisingly, serum IgG responses were slightly higher (although not significantly) in i.m.-vaccinated guinea pigs compared to those vaccinated via the i.n. route, lending support to the hypothesis that mucosal rather than systemic immunity is beneficial in preventing influenza virus transmission. Our data also demonstrate that antibodies toward the viral NA are responsible, at least in part, for the integral role that mucosal immunity plays in virus transmission from a vaccinated, infected donor onward to a naive, susceptible recipient.
Although we found that a homologous B/Malaysia/2506/2004 NA vaccination, fol-  found that viral titers were reduced in the naive recipients ( Fig. 1B and Fig. 4C). This may be due to the lower viral input transmitted from the vaccinated donors resulting in less viral replication and earlier clearance of infection or-more likely-a shifted peak in virus shedding that was missed with the chosen sampling time points. Our finding that NA-vaccinated donor guinea pigs are unable to transmit influenza B virus despite relatively robust replication in the nasal cavity raises the question about the mechanism underlying transmission inhibition. One possibility is that even a slight reduction in nasal titers can adversely impact transmissibility of influenza viruses, which is a plausible though unexciting explanation. However, virus titers in animals receiving i.m. or i.n. vaccination were similar, and only i.n. vaccination blocked transmission effectively (Fig. 1). A second explanation is that antibodies inhibiting the enzymatic activity of the viral NA alters virus transmissibility. It has been reported that in the absence of NA activity, influenza viruses tend to aggregate (14), which may negatively impact transmissibility. In addition, virus might be trapped by decoy receptors on natural defense proteins like mucin, also leading to aggregation and reduced ability to transmit. Finally, it is possible that virus is efficiently released from an infected host and transmitted to a susceptible one, but that the virus is coated by anti-NA antibody that impairs its ability to initiate a new infection, becoming trapped by mucins in the respiratory tract of the recipient. Regardless of the mechanism of action, however, mucosal immunity against the viral NA seems to be key in preventing efficient interhost transmission; our study and others (34,35) suggest that serum antibody levels are less important in this specific regard.
In our studies, we also performed transmission experiments in which naive, influenza B virus-infected donor guinea pigs were paired with recipients previously vaccinated i.n. with the homologous NA. Interestingly, we observed inefficient transmission to one of three vaccinated recipients in the airborne model and efficient transmission to three of three vaccinated recipients in the contact model. However, in both models, the vaccinated guinea pigs that were infected by transmission demonstrated very low nasal wash virus titers and a short duration of shedding.
In humans, several observations have been made regarding the impact of anti-NA antibody titers on virus infection and disease. In influenza A virus (H3N2) challenge studies, individuals with high anti-NA antibody titers were permissive to infection with the homologous influenza virus, but they had reduced nasal virus shedding and illness after infection (25,27). Similar observations were made in a more recent human challenge study with pandemic H1N1 virus (28). In addition, NI titers have been shown to correlate with protection from natural infection with H1N1 and H3N2 infection, respectively, in two independent studies (24,26). Anti-NA antibodies might also have played a role in protecting H2N2 experienced individuals during the H3N2 pandemic in 1968 (27). Thus, in humans, anti-NA antibodies might confer a degree of protection from natural infection with influenza A viruses, while in challenge studies they do not confer immunity from infection but do reduce viral shedding and disease. While no comparable studies exist for influenza B, the influenza A data in humans echo our findings in the guinea pig model using influenza B virus. When NA-vaccinated guinea pigs were directly inoculated with influenza B virus, they showed reduced nasal wash virus titers, similar to that which has been shown in human challenge studies. When NA-vaccinated guinea pigs were the recipients, paired with naive, infected donors (analogous to natural infection in humans), we observed only partial inhibition of transmission. However, a reduction in the magnitude and duration of virus replication in NA vaccinated recipients that became infected by transmission was observed, a finding reflected in human studies. What remains unknown in humans is the extent to which anti-NA antibodies affect the contagiousness of influenza virus-infected individuals, i.e., are humans with robust anti-NA antibody titers less likely to transmit influenza viruses on to other people? This question, while difficult to answer with human studies, is very important from a public health perspective. If anti-NA immunity impairs the ability of infected individuals to infect others, then vaccines inducing high titers of anti-NA antibodies would be valuable for control of influenza B on a population scale. Our data from the guinea pig model suggest that vaccinated animals (those with mucosal antibodies inhibiting the enzymatic function of the viral NA) are poor onward transmitters to susceptible partner animals, thus breaking the transmission chain.
Potentially, there are many ways to translate our findings into vaccines and vaccination regimens that induce robust anti-NA immunity. The easiest solution might be NA-only vaccines that are given in addition to the current vaccines or admixture of recombinant NA to inactivated vaccines (to avoid an additional shot) (3,20). The use of mucosally administered inactivated vaccines is currently being explored, and recombinant NA could be added to those vaccines as well (36). In addition, next-generation influenza virus vaccines, e.g., based on RNA or DNA vaccines or viral vectors, could also include NA as an antigen (37,38).
In conclusion, our data suggest that anti-NA immunity can significantly contribute to reduction of influenza B virus shedding. Importantly, mucosal anti-NA immunity blocks efficient transmission of influenza B viruses in the guinea pig model. In theory, vaccines that impair the transmission efficiency of influenza B viruses among humans could, with sufficient herd immunity, drastically limit their spread. This is an especially attractive proposition for influenza B viruses, which lack an animal reservoir, because it opens up the possibility of their eradication from the human population. These data argue for further research into and development of influenza virus vaccines inducing robust mucosal anti-NA antibody responses.

MATERIALS AND METHODS
Ethics statement. All animal procedures in this study were performed in accordance with the Institutional Animal Care and Use Committee (IACUC) guidelines and have been approved by the Icahn School of Medicine at Mount Sinai IACUC.
Recombinant proteins. Soluble HA and NA proteins containing a T4 foldon trimerization domain or a vasodilator stimulated phosphoprotein tetramerization domain, respectively, were generated using the baculovirus expression system as previously described (9,20,39). The HA and NA recombinant proteins expressed C-terminal and N-terminal hexahistidine tags, respectively, for purification purposes.
Guinea pig vaccination. Five-to six-week-old female guinea pigs were purchased from Charles River Laboratory and randomly assigned to different vaccination groups. Guinea pigs were either primed i.m. pairs were set up in cages that precluded physical contact but allowed lateral airflow (airborne transmission) or were cocaged (contact transmission). On days 2, 4, 6, 8, and 10 from the initial donor challenge, nasal washes were collected from anaesthetized donor and recipient guinea pigs. Blood was collected via venipuncture of the lateral saphenous vein at the prime, challenge, and euthanasia time points for the determination of serum antibody responses and NA inhibition (NI). Nasal washes were also collected at prime and challenge time points for the determination of antibody responses and NI. On days 18 to 21 postchallenge, guinea pigs were anaesthetized with ketamine (44 mg/kg) and xylazine (5 mg/kg) and terminally bled.
For B/Malaysia/2506/2004 transmission experiments, guinea pigs were housed at 20°C and 20% relative humidity (RH). The B/Florida/04/2006 transmission experiments occurred at 5°C and 20% RH. These environmental settings have been previously shown to be optimal for these influenza B virus strains to transmit efficiently among guinea pigs (41).
Virus titers. Virus titers were determined by plaque assay on MDCK cell monolayers. Virus stocks and nasal washes were diluted 10-fold in PBS and incubated on MDCK cells for 1 h before the addition of an agarose overlay containing a final concentration of 0.64% agarose (Oxoid), 1ϫ minimum essential medium (MEM) (10% 10ϫ minimal essential medium [Gibco], 2 mM L-glutamine, 0.1% of sodium bicarbonate [wt/vol; Gibco], 10 mM 4-HEPES, 100 U/ml penicillin-100 g/ml streptomycin, and 0.2% bovine serum albumin (BSA), 1 g/ml TPCK (tolylsulfonyl phenylalanyl chloromethyl ketone)-treated trypsin, and 0.1% (wt/vol) DEAE (diethylaminoethyl)-dextran was added to the cells. The cells were then incubated for 72 h at 33°C, and visible plaques were counted after fixing them with 3.7% formaldehyde and visualization with a crystal violet counterstain (Sigma-Aldrich). All virus titers are presented as the log 10 PFU/ml. The limit of detection for these assays was 10 PFU/ml. The area under the curve (AUC) values for virus titers over the time course of infection were calculated using the log PFU/ml, and the AUC value was calculated using GraphPad Prism 7 software.
Enzyme linked immunosorbent assay. Immulon plates (Immulon 4HBX; Thermo Scientific) were coated with 2 g/ml of purified virus (50 l/well) in coating buffer (KPL coating solution; Sera Care) at 4°C overnight. The following day, the plates were washed three times with PBS containing 0.1% Tween 20 (PBS-T) and blocked in blocking solution (3% goat serum and 0.5% milk in PBS-T) for 1 h at room temperature. After blocking, prediluted serum was added to the first well to a final concentration of 1:100 in blocking solution. For assessing antibody responses in the nasal washes, a neat nasal wash was added to the first well. Serum and nasal washes were then serially diluted and incubated at room temperature for 2 h. Plates were then washed three times with PBS-T before the addition of donkey anti-guinea pig IgG-horseradish peroxidase (IgG-HRP; EMD Millipore) in blocking solution for 1 h at room temperature. Plates were then washed four times with PBS-T with shaking and then the O-phenylenediamine dihydrochloride (OPD) substrate (SigmaFast OPD; Sigma-Aldrich) was added. After 10 min of incubation at room temperature, the reaction was stopped by adding 50 l of 3 M HCl to the mixture. The optical density (OD) was measured at 490 nm on a Synergy 4 plate reader (BioTek). A cutoff value of the average of the OD values of blank wells plus three standard deviations was established for each plate and used for calculating the AUC values of the nasal washes and the sera, which were the readout for this assay.
Enzyme-linked lectin assay to determine NI. Immulon plates (Immulon 4HBX; Thermo Scientific) were coated with 50 g/l of fetuin (150 l/well) in coating buffer (KPL coating solution; Sera Care) at 4°C overnight. The following day, the plates were washed six times with PBS-T and blocked in blocking solution (5% BSA in PBS) for 2 h at room temperature. While the plates were being blocked, virus was serially diluted 1:2 on a separate 96-well plate in PBS containing 1% BSA. After this blocking step, 100 l of serially diluted virus was transferred to the fetuin-coated plates, and the plates were incubated at 33°C for 2 h. The plates were then washed, and a secondary solution of peanut agglutinin (PNA) conjugated to HRP (5 g/ml) was added to the plates. After a 2-h incubation in the dark, the plates were washed six times and developed with 100 l of SigmaFast OPD. After developing for 7 min, 50 l of 3 M HCl was added, and the OD was measured at 490 nm on a Synergy 4 plate reader (BioTek).
To perform NI assays, Immulon plates were coated and blocked as described above. While the plates were being blocked, guinea pig serum and nasal washes were serially diluted 1:2 in a separate 96-well plate in PBS. A starting dilution of 1:50 was used for serum, while we started with a neat nasal wash. A 75-l portion of virus diluted to a 2ϫ 50% effective concentration was added to each well of the serially diluted serum plate, and the plates were shaken at room temperature for 2 h. Before the shaking incubation time expired, the fetuin-coated plates were washed six times with PBS-T, and 100 l of the virus/serum mixture was transferred to the fetuin-coated plates, followed by incubation for 2 h at 33°C. The remainder of the NI assay was performed as described above. The values obtained from the plate reader were divided by the average of the virus-only well subtracted from the no-virus control well and multiplied by 100 to obtain the NA activity. The percent inhibition was calculated by subtracting the NA activity from 100.
Statistics. Grouped data (except in Fig. 2C and D) are expressed as individual dot plots, and means are represented as geometric mean (GEM). Data in Fig. 2C and D are presented as grouped data, and means are expressed as standard errors of the mean (SEM). Statistical differences between two groups were determined by Student paired t test. Statistical differences between three or more groups were determined by two-way analysis of variance, followed by a Bonferroni multiple-comparison test. All statistical analyses were performed using GraphPad Prism 7 for Windows. In all cases, probability levels of Ͻ0.05 were indicative of statistical significance (*, P Ͻ 0.05; ***, P Ͻ 0.001).