Alternative Strategy for a Quadrivalent Live Attenuated Influenza Virus Vaccine

Seasonal influenza viruses infect 1 billion people worldwide and are associated with ∼500,000 deaths annually. In addition, the never-ending emergence of zoonotic influenza viruses associated with lethal human infections and of pandemic concern calls for the development of better vaccines and/or vaccination strategies against influenza virus. Regardless of the strategy, novel influenza virus vaccines must aim at providing protection against both seasonal influenza A and B viruses. In this study, we tested an alternative quadrivalent live attenuated influenza virus vaccine (QIV) formulation whose individual components have been previously shown to provide protection. We demonstrate in proof-of principle studies in mice that the QIV provides effective protection against lethal challenge with either influenza A or B virus.

deaths around the world every year (2). For the United States, influenza virus infections result in an average economic impact of $87 billion due to prophylactic, therapeutic, and hospitalization costs, as well as missed school days or workdays (3)(4)(5). IBVs contribute less to seasonal epidemics than IAVs, but they are a significant disease burden in the pediatric and elderly populations (6)(7)(8). In addition, IAVs have been associated with pandemic episodes when novel strains from animal reservoirs are introduced into humans (9). Public health concerns are aggravated by the inherent ability of influenza viruses to develop resistance to available antivirals (10). Thus, vaccination remains the best choice to protect humans against influenza virus infections, and it is considered the first line of defense against influenza virus-associated illness.
Vaccines against seasonal influenza viruses are manufactured to confer protection against circulating IAV and IBV strains. Current vaccines rely primarily on antibody responses to the virus' hemagglutinin (HA) surface protein. However, HA undergoes antigenic drift, requiring regular vaccine updates in order to antigenically match them to the currently circulating strains (11,12). Seasonal influenza virus vaccines have traditionally contained three influenza virus strains, two IAV strains (A/H1N1 and A/H3N2) and one IBV from either the B/Yamagata or the B/Victoria antigenic lineage (11). In recent years, however, the two IBV lineages have shown not only seasonal variations, but also significant differences in prevalence in different countries. Thus, the U.S. Food and Drug Administration (FDA) has approved quadrivalent vaccines that incorporate both IBV antigenic lineages, in addition to the two IAV strains (13,14). To date, the FDA has approved three types of influenza virus vaccines for human use: inactivated influenza virus vaccine (IIV), recombinant influenza protein (RIP) vaccine, and live attenuated influenza virus vaccine (LAIV) (15,16). The most widely used influenza virus vaccine is IIV, which can elicit protective humoral immunity by inducing the production of neutralizing antibodies that target epitopes on the virus HA (and, to a lesser extent, the neuraminidase [NA] surface protein). The RIP vaccine, similar to IIV, can induce neutralizing antibodies that target the HA protein (17). In contrast, LAIVs can elicit not only humoral but also cell-mediated immunity, which is considered more cross-protective, since it targets more conserved viral epitopes (18,19). LAIVs for human use were independently obtained by serial passage in eggs at low temperatures, resulting in cold-adapted (ca) and temperature-sensitive (ts) mutations that resulted in an attenuated (att) phenotype in vivo. In the United States, the only FDA-approved LAIV consists of viruses carrying a series of mutations in the internal gene segments of the master donor viruses (MDVs) (A/Ann Arbor/6/60 [H2N2] [MDV-A] and B/Ann Arbor/1/66 [MDV-B]) (20)(21)(22).
Early work in our laboratory demonstrated that a combination of an HA tag at the C terminus of PB1 and ts mutations in PB2 (S265) and PB1 (E391, G581, and T661) showed remarkable stability over multiple passages in IAV backbones of avian, human, and swine origin. This strategy-referred to as IAV att-is safe and effective, resulting in protection against avian-and mammalian-origin influenza virus strains in a variety of avian and mammalian species (23,24). We recently demonstrated that an IBV att strain modified with a similar strategy (PB1 with mutations G580, A660, and a C-terminal HA tag) resulted in a stable virus over multiple passages in tissue culture and eggs and with an att phenotype in vivo. A single intranasal (i.n.) dose of the IBV att protected mice against lethal challenge with antigenically matched and mismatched IBV strains (25). In this study, we used the IAV att and IBV att backbones to generate a quadrivalent LAIV (QIV) containing H3N2 and H1N1 IAV antigenic subtypes and two antigenic IBV lineages. We used the QIV to immunize mice intranasally in a prime-boost regime. We show that 2 doses of the QIV elicited discernible hemagglutination inhibition (HI) responses against the 4 viruses in the vaccine. We also observed that the QIV protected mice against lethal challenge with either IAV H1N1 or IBV.

RESULTS
Generation of a QIV. In this study, we sought to evaluate whether a QIV with the two laboratory att backbones was safe, immunogenic, and protective. Thus, reverse genetics was used to generate the following viruses: Ty/04 att (H3N2), Ca/04 att (H1N1), B/Bris att (Victoria lineage), and B/Wisc att (Yamagata lineage). The Ty/04 att, Ca/04 att, and B/Bris att viruses have been previously described (23)(24)(25), while the B/Wis att virus was prepared for this study (Table 1). Like the B/Bris att virus, the B/Wis att virus displayed a ts phenotype in vitro that is typically associated with attenuation in vivo (data not shown). The QIV formulation was prepared with doses containing equal amounts (2 ϫ 10 7 50% tissue culture infectious doses [TCID 50 ]/ml) of each att virus.
Safety and immunogenicity of QIV. DBA/2J mice were randomly distributed into two different groups and received either the QIV or phosphate-buffered saline (PBS) (mock vaccination) by intranasal inoculation. The mice were boosted at 21 days postvaccination (dpv) following the same strategy. No clinical signs of disease, body weight loss, or mortality were observed after prime or boost. HI titers were evaluated at 20 dpv and 20 days postboost (dpb) from 4 mice/group (Fig. 1). Mice in the QIV group showed low but discernible HI titers at 20 dpv against the four HA components in the vaccine (HI titers of 20 to 160 against Ca/04, 20 to 80 against B/Bris, 20 to 160 against Ty04, and 10 to 40 against B/Wisc). After boost, HI titers increased and were all within the range considered to be predictive of protection against clinical disease (HI titers of 160 to 1,280 against Ca/04, 160 to 640 against B/Bris, 320 to 640 against Ty/04, and 40 to 160 against B/Wisc). As expected, mice in the PBS group showed HI titers below the limit of detection. These results indicated that the QIV was safe and immunogenic in mice.
Protection efficacy for QIV in mice. In order to assess the protective effects of the QIV regime, mice were challenge with 10 7 TCID 50 per mouse of either the Ca/04 or B/Bris virus by the intranasal route (Fig. 2). No signs of disease and minor weight loss were observed in the QIV-vaccinated mice within the first 2 days after challenge with the Ca/04 virus (QIV-Ca/04), but the mice quickly recovered from day 3 onward. All of the QIV-Ca/04 mice (8 out 8) survived the challenge by 14 days postchallenge (dpc). In contrast, the PBS-Ca/04 mice showed rapid body weight loss and weakened after challenge and succumbed or had to be euthanized by 6 dpc. Despite not showing overt clinical signs of disease, the mice in the QIV-B/Bris group showed substantial weight loss (10 to 15%) within the first 3 days after challenge and slowly recovered. By 10 dpc, the mice in the QIV-B/Bris group showed average body weights similar to those of unchallenged control mice. By 14 dpc, all the mice (8 out of 8) in the QIV-B/Bris group had survived the challenge. By comparison, the body weights of mice in the PBS-B/Bris group declined rapidly, and by day 7, 7 out of the 8 mice had succumbed to the challenge. It must be noted that one mouse in the PBS-B/Bris group appeared not to have been infected properly because it showed no body weight loss during the follow-up at 14 days postchallenge (and thus, it was not considered for the calculation of body weight changes in the group). The same mouse is shown as a survivor in the survival curve (an asterisk is used to denote the outlier in Fig. 2E and F). The protective effects of the QIV were also reflected by the lack of detectable virus in lung homogenates and nasal turbinates (NTs) collected at 5 dpc (Fig. 2). Conversely, significant virus titers were observed in lung and NT homogenates obtained from mice in the PBS-Ca/04 and PBS-B/Bris control groups. As expected, HI titers at 21 dpc against the challenge virus were increased in each group but remained unchanged (or slightly lower than at 20 dpb) against other viruses in the vaccine (Fig. 1). These observations suggest that the QIV protected the mice against lethal IAV H1N1 or IBV (Victoria lineage) challenge. Virus-specific IgA and IgG antibody responses in mice after challenge. To further explore virus-specific antibody responses, we evaluated IgA and IgG levels against the 4 viruses in the vaccine. Enzyme-linked immune absorbent assays (ELISAs) using whole-virus-coated plates were performed to detect IgA and IgG antibodies in samples from lungs and NTs collected at 5 dpc and sera, nasal washes (NW), and  (Fig. 3). Based on OD 450 values, IgA responses against the Ca/04 and B/Bris challenge viruses were high in NTs and lung homogenates at 5 dpc and remained high in BALFs and NW at 21 dpc compared to PBS-PBS control group samples (note that mock-vaccinated mice did not survive beyond 7 dpc, except for 1 mouse in the PBS-B/Bris group). As expected, virus-specific IgA levels were also high in fecal samples (not shown) but moderate in serum samples collected at 21 dpc. Similarly, IgG responses were relatively high against the challenge viruses in samples from QIV-vaccinated mice collected after challenge (Fig. 4). Typically, virus-specific IgG was consistently detected well above background in samples diluted between 1:100 and 1:1,000 (up to 1:100,000 in serum) from QIV-vaccinated/challenged mice but not in similarly diluted samples from mock-vaccinated/challenged mice. Ty/04-and B/Wisc-specific IgA and IgG absorbance readings provided some evidence of the relative levels of QIV-induced responses. However, it must be noted that the samples were collected after challenge with either Ca/04 or B/Bris virus, and the whole-virus ELISAs do not discriminate between the potential postchallenge stimula-

DISCUSSION
Influenza virus vaccine effectiveness ranged from 19% to 48% during the past three influenza seasons (26,27).  in other markets where the same vaccine is approved). Although such issues appear to have been resolved for the 2017-2018 season and alternative LAIVs approved for other markets (based on the A/Leningrad and B/USSR MDV strains) do not seem to have faced similar issues, there is a need for the development of novel LAIV strategies that will likely impact the future development of influenza virus vaccines. In general, weakened (attenuated) forms of viral pathogens offer the longest-lasting protection because they produce multidimensional responses by stimulating the humoral and cellular arms of the immune system. The IAV att component in the QIV studied here, based on the internal gene segment constellation of the Ty/04 strain, contains mutations found in the A/Ann Arbor MDV-A (PB2 S265 and PB1 E391, G581, and T661). These mutations can confer an att phenotype on some, but not all, IAV strains, as we have previously shown (23,24,28,29). However, the incorporation of the HA epitope tag at the C terminus of PB1, along with the ts mutations (ts ϩ HA tag ϭ att) resulted in attenuation in vivo. The IAV att strains that we have produced so far are remarkably stable and show optimal safety and protective profiles when used as monovalent vaccines. The most widely used MDV-As (A/Leningrad/and A/Ann Arbor) were isolated nearly 60 years ago. Over this period, IAVs have continued to evolve and reassort and have accumulated mutations on relevant cytotoxic T lymphocyte (CTL) epitopes that differ from those in the MDV-A internal backbones (30). Altering the gene constellation of MDV-A would require further preclinical and clinical testing. Thus, a major advantage of our strategy is the possibility to impart the att phenotype to virtually any IAV strain. We favored the Ty/04 att backbone because it carries the potential to act as a universal att backbone for use in avian and mammalian species, including humans. To be used in humans, however, a complementary IBV att strategy is highly desirable. For that reason, we incorporated an IBV PB1 gene segment carrying the G580 and A660 mutations and the C-terminal HA tag and established that it resulted in attenuation of the B/Bris strain. We further demonstrated that the B/Bris att strain was stable, safe, and immunogenic and protected mice against lethal IBV challenge. The modifications of the IBV att are unique to our virus and are not shared with the B/Ann Arbor or B/USSR MDV-B.
In this study, we evaluated whether a QIV version of the att strategy was safe, immunogenic, and protective against lethal influenza virus challenge in a mouse model. We used DBA/2J mice because they are more susceptible to IAV and IBV strains and develop anti-influenza humoral responses comparable to those of other mouse strains (31,32). We opted for a prime-boost strategy in order to determine whether we could further stimulate humoral responses against all the components of the vaccine. Such a 2-dose strategy (1 month apart) is recommended for children 2 to 8 years old who receive the LAIV in the United States. It must be noted that, except for the Ca/04 att strain, the viral components in the QIV do not carry mouse-adapted mutations. Thus, it was remarkable to realize that a second dose of the QIV stimulated HI responses to all components in the vaccine to within the protective range (HI titers of Ն40) (Fig. 1). Differences in the levels of HI responses, particularly the B/Wisc att, could be due to a number of factors, including replication fitness differences, competition between IAV and/or IBV strains, and antigenic/ immunogenic differences. Further studies, including refinements in virus particle versus infectious dose, will be needed but are beyond the scope of the present report.
QIV-vaccinated mice were fully protected against lethal challenge (ϳ10,000 50% mouse lethal dose [MLD 50 ]), with the Ca/04 strain showing minor body weight losses within the first 2 days after challenge (Fig. 2). The Ca/04 strain used in this study carries mutations on the surface and internal gene segments that increase its virulence for mice (42). However, by 5 dpc, no evidence of Ca/04 replication was detected in the lungs or NTs of the QIV-vaccinated/challenge mice, unlike the mock-vaccinated/challenged mice. QIVvaccinated mice also survived challenge with the B/Bris strain carrying the PB2 F406Y mutation and by 5 dpc showed no evidence of virus replication in lungs and NTs. We have previously shown that the PB2 F406Y mutation increases the virulence for mice of the wild-type B/Bris strain (it also slightly increases the virulence of B/Wisc) (25). Unexpectedly, the B/Bris challenge dose (Յ10 50% lethal doses [LD 50 ]) was sufficient to cause significant body weight loss in the QIV-vaccinated/challenged mice compared to the PBS-PBS mice. It is possible that the IAV att strains had a competitive advantage over the IBV att strains in the QIV that affected humoral and cellular responses against the latter. Whether the effect is due to the mouse model or to true competitive-fitness advantages of one virus over the rest in the QIV remains to be determined. The potential for interference between type A and type B live attenuated influenza viruses has been previously described in the mouse model, although early studies did not contemplate the use of QIV formulations (33,34). As expected, by 21 dpc, HI titers were increased against the challenge viruses, but only in their respective groups, whereas HI titers against other viruses were slightly lower than at 20 dpb. It was also noted that the HI responses to the H3 HA proteins were limited to the Ty/04 strain with very limited cross-reactivity to other older or newer seasonal H3 strains (not shown).
LAIVs have been shown to heighten innate immune responses (35)(36)(37) and to stimulate cross-protective responses to heterologous or antigenically divergent strains (38,39). We explored anti-IAV and anti-IBV IgA and IgG responses using whole-virus preparations in an ELISA. NTs and lung homogenate samples collected at 5 dpc provided an indication of preexisting immunity against all the components of the QIV. Although direct comparisons are not possible because the 4 viruses in the QIV are different, a trend toward higher OD values for IgA and IgG against the Ca/04 and B/Bris strains than against the Ty/04 and B/Wisc strains was observed, consistent with the first two being the viruses used for challenge. It is expected that cross-protective antibodies and/or antibodies against conserved proteins are represented at some level in the ELISAs, and this appears to be the case for the IBV strains. However, it is important to emphasize that anti-influenza responses are skewed toward the HA and NA proteins. Thus, it is tempting to speculate that the IgA and IgG profiles observed are bona fide representations of antibody responses against each of the viruses in the vaccine (and cross-reactive IBV HA and/or NA antibodies).
In summary, we provide proof of principle of the potential of an alternative QIV as a safe and effective vaccine against IAV and IBV in mice. The strategy increases the arsenal of LAIV options against seasonal and pandemic influenza.  Table 1. The virus stocks were amplified in 9-day-old specific-pathogen-free (SPF) embryonated chicken eggs and were stored at Ϫ80°C. Generation of att viruses by reverse genetics. The Ty/04 att (H3N2) and B/Bris att virus vaccines have been previously described and are referenced in Table 1. The Ca/04 att (H1N1) virus carries the PB1 and PB2 segments from the Ty/04 att virus and 6 gene segments from the mouse-adapted Ca/04 virus. B/Wisc att carries the HA and NA surface gene segments of B/Wisc and the backbone of B/Bris att. The att viruses were rescued by reverse genetics using coculture of 293T and MDCK cells as previously described (40). The IAV att and IBV att rescued viruses (passage 0 [P0]) were expanded in fresh MDCK cells to produce P1 viruses. The P1 viruses were used in a second round of amplification in MDCK cells (P2) and in 9-day-old SPF eggs (E1). The IAV att and IBV att viruses were amplified in eggs at 35°C and 33°C, respectively. Virus stocks were titrated by TCID 50 using the Reed-Muench method (43). The virus stocks were sequenced by next-generation sequencing (NGS) using the MiSeq platform (Illumina, San Diego, CA). The NGS results showed integrity of att mutations and absence of spurious mutations. The QIV was formulated to contain equal doses (2 ϫ10 7 TCID 50 /ml) of Ty/04 att, Ca/04 att, B/Bris att, and B/Wisc att viruses.

Cells and viruses
Animal use and compliance. Animal studies were performed under animal biosafety level 2 (ABSL2) containment conditions and followed protocols approved by the Institutional Animal Care and Use Committees (IACUC) at the University of Georgia. Mice that lost Ն25% of their initial body weight (a score of 3 or higher on a 4-point scale of disease severity) were humanely euthanized.
Mouse experiments. Five-to 6-week-old female DBA/2J mice were purchased from Jackson Laboratories (Bar Harbor, ME). The mice were randomly distributed into 2 groups. The mice were anesthetized with isoflurane and subsequently inoculated via the i.n. route with 50 l of an inoculum containing either PBS (mock vaccinated; group 1; n ϭ 33) or QIV (10 6 TCID 50 /mouse; group 2; n ϭ 22). At 21 dpv, the mice in group 2 were boosted i.n. with QIV, whereas the control mice in group 1 received PBS. At 20 dpv and 20 dpb, 4 mice/group were bled from the submandibular vein, and serum samples were collected and used to measure neutralizing and virus-specific antibody responses. At 21 dpb, mice were challenged with an inoculum of 50 l containing either the Ca/04 virus (10 7 TCID 50 /mouse; ϳ10,000 MLD 50 ; n ϭ 11/group; groups QIV-Ca/04 and PBS-Ca/04) or the B/Bris virus (10 7 TCID 50 /mouse; ϳ10 MLD 50 ; n ϭ 11/group; groups QIV-B/Bris and PBS-B/Bris). A third group of mice from the mock-vaccinated group were mock challenged with PBS (n ϭ 11; group PBS-PBS). After challenge, the mice were monitored daily for clinical signs of disease, body weight loss, and mortality. At 5 dpc, 3 mice per group were sacrificed, and lungs and NT samples were collected to determine virus titers. At 21 dpc, all the mice were sacrificed, and serum samples, NW, and BALF samples were collected for evaluating virus-specific antibody responses.
HI assay. Serum samples collected at 20 dpv, 20 dpb, and 21 dpc were screened for the presence of neutralizing antibodies by the HI assay using the wild-type Ty/04, Ca/04, B/Bris, and B/Wisc strains (41). Briefly, the sera were treated with receptor-destroying enzyme at 37°C overnight and then heat inactivated at 56°C for 30 min. Then, the sera were diluted 1:10 with PBS and subsequently serially diluted 2-fold and mixed with 8 hemagglutination units (HAU) of virus in a 96-well plate and incubated at room temperature for 15 min. The HI activity was visualized by adding 0.5% turkey red blood cells to the virus-serum mixtures and further incubation at room temperature for 30 min before reading.
Detection of virus-specific IgA and IgG antibodies. NW samples, BALFs, and serum samples collected at 21 dpc, as well as NT and lung homogenates prepared at 5 dpc, were screened for the presence of virus-specific IgA and IgG antibodies using an ELISA. Whole-virus preparations were obtained by centrifuging 180 ml of virus (about 5,120 HAU/ml) at 24,000 rpm (ϳ105,000 ϫ g) for 2 h in a Beckman SW32 Ti swing bucket rotor and a Beckman Optima L-100 XP ultracentrifuge. The pelleted viruses were resuspended in 1.2 ml of PBS. The resuspended virus protein concentration was titrated with a Pierce bicinchoninic acid (BCA) protein assay kit (ThermoFisher, Rockford, IL) and diluted to 50 g/ml in coating buffer (0.1 M sodium carbonate, pH 9.6). The plates were coated with diluted virus (100 l/well) at 4°C overnight. Then, the plates were blocked with 10% skim milk in PBS at room temperature for 2 h and washed twice with PBS containing 0.05% Tween 20 (PBST). The plates were incubated with test samples diluted in 2% skim milk in PBS at room temperature for 2 h. After washing 3 times with PBST, the plates were incubated with a horseradish peroxidase (HRP)-conjugated anti-mouse IgA antibody (Bethyl Laboratories, Montgomery, TX) or anti-mouse IgG antibody (ThermoFisher, Rockford, IL) at room temperature for 1 h and then washed 3 times with PBST. Then, 100 l TMB solution (ThermoFisher, Vienna, Austria) was added to each well and incubated for 15 min at room temperature. The reaction was stopped by adding 3% H 2 SO 4 , and OD 450 values were measured.
Statistical analyses. All data analyses were performed using GraphPad Prism software version 7 (GraphPad Software Inc., San Diego, CA). For multiple comparisons, two-way analysis of variance (ANOVA) was performed. Differences in survival curves were analyzed using the log-rank test. A P value below 0.05 was considered significant.