Adjuvanted recombinant hemagglutinin H7 vaccine to highly pathogenic influenza A(H7N9) elicits high and sustained antibody responses in healthy adults

An unprecedented number of human infections with avian influenza A(H7N9) in the fifth epidemic wave during the winter of 2016–2017 in China and their antigenic divergence from the viruses that emerged in 2013 prompted development of updated vaccines for pandemic preparedness. We report on the findings of a clinical study in healthy adults designed to evaluate the safety and immunogenicity of three dose levels of recombinant influenza vaccine derived from highly pathogenic A/Guangdong/17SF003/2016 (H7N9) virus adjuvanted with AS03 or MF59 oil-in water emulsions. Most of the six study groups meet the FDA CBER-specified vaccine licensure criterion of 70% seroprotection rate (SPR) for hemagglutination inhibition antibodies to the homologous virus. A substantial proportion of subjects show high cross-reactivity to antigenically distinct heterologous A(H7N9) viruses from the first epidemic wave of 2013. These results provide critical information to develop a pandemic response strategy and support regulatory requirements for vaccination under Emergency Use Authorization.


INTRODUCTION
Zoonotic infections with a novel Asian lineage avian-origin influenza A(H7N9) virus were first reported in China in March 2013 and caused severe, often fatal, lower respiratory tract disease in humans 1,2 . Since then, influenza A(H7N9) virus was found to be circulating in poultry in China, and until 2017, human epidemics have been reported annually between fall and early spring 3 . During the fifth epidemic in the winter of 2016-2017, an unprecedented number of human influenza A(H7N9) cases were identified. Genetic and antigenic analyses indicated that more than 90% of the circulating H7N9 viruses belonged to a new group designated Yangtze River Delta lineage, antigenically distinct from the previously dominant Pearl River Delta lineage 4 . The new group had reduced cross-reactivity with antibodies raised to existing candidate vaccine viruses (CVVs) made in 2013, prompting the World Health Organization (WHO) to update the pandemic influenza vaccine recommendations 5 . Although the influenza A(H7N9) viruses that emerged in 2013 were characterized by having low pathogenicity in chickens (low pathogenic avian influenza or LPAI), some of the Yangtze River Delta lineage viruses emerging in late 2016 were highly pathogenic for poultry (highly pathogenic avian influenza; HPAI) and caused several human infections 3,6 . Furthermore, mutations in some A(H7N9) viruses detected in humans resulted in reduced susceptibility to influenza antiviral drugs, increasing the potential public health impact of these viruses 7 . Consequently, global health authorities convened by the WHO recommended development of CVVs based on a LPAI A/Hong Kong/125/2017-like virus and a HPAI A/ Guangdong/17SF003/2016-like virus 5 , hereafter referred to as A/ HK/2017 and A/GD/2016, respectively.
The United States (US) Department of Health and Human Services (HHS) continuously monitors pandemic risk and prepares to respond to the threat of novel influenza virus outbreaks in the US. To this end, the Biomedical Advanced Research and Development Authority (BARDA), within the Office of the Assistant Secretary for Preparedness and Response (ASPR), has established and maintains the National Pre-Pandemic Influenza Vaccine Stockpile (NPIVS) comprising adjuvants and pre-pandemic bulk antigens from avian influenza viruses determined to pose a significant risk for a pandemic. In addition, to shorten timelines to make influenza vaccines available to immunize the US population, HHS has supported expansion of domestic vaccine manufacturing capacity, including the use of adjuvanted vaccines, and the licensure of pre-pandemic and seasonal egg-based, cell-based, and recombinant influenza vaccines. While egg-and cell-based novel influenza vaccine antigens have been clinically evaluated with adjuvants AS03 and MF59 to support their deployment for pandemic mitigation under Emergency Use Authorization by the FDA [8][9][10][11][12][13][14][15][16][17] , safety and dose-sparing evidence of these adjuvants with recombinant H7 hemagglutinin (HA) antigens is lacking. The present study closed this gap by evaluating the safety and immunogenicity of adjuvanted influenza H7 vaccine derived from A/GD/2016 utilizing recombinant protein technology that can be used to respond quickly to a pandemic influenza virus. By formulating the recombinant protein vaccine with adjuvants from the NPIVS, the results from this study provide critical insights for the US Government to develop a response strategy for a pandemic emergency.

Study population
A total of 366 subjects were enrolled and randomly assigned to each treatment group ( Fig. 1 and Table 1). Subjects received two doses of monovalent recombinant influenza H7 vaccine derived from A/GD/2016 at three antigen dose levels (3.75, 7.5, and 15 µg) adjuvanted with AS03 or MF59 administered 28 days apart. The mean subject age was 35.9 years (range: 18-49 years), 55.2% female (range: 48.4-61.7%), 79.5% white (range: 70.0-86.9%), and 16.4% Black or African American (range: 8.2-28.3%). Treatment compliance was high for both vaccine administrations. All subjects who received at least one vaccine dose were considered part of the safety population, while 339 subjects who received both vaccine doses and completed the primary immunogenicity endpoint at day 50, comprised the immunogenicity per protocol population.

Safety and tolerability
The adjuvanted recombinant H7 vaccines were well-tolerated regardless of antigen dose (Fig. 2 and Table 2), and there was no evidence of increased frequencies of subjects experiencing adverse events (AEs) with increasing dose. In general, the frequencies and distribution of local and systemic AEs were similar between adjuvants. Most solicited local and systemic AEs were considered mild to moderate severity. The most common solicited AE for local reactogenicity was injection site pain (Fig. 2). There were no AEs of special interest or potentially immunemediated medical conditions (PIMMCs) reported. There were no serious AEs (SAEs) related to vaccination as defined by 21 CFR 312.32(a), and there were no AEs or SAEs that led to study withdrawal. Finally, no significant differences were observed in changes from baseline for clinical laboratory results or vital signs for any of the study groups (data not shown).

Immunogenicity
The purpose of this study was to inform the US Department of Health and Human Services (HHS) Pandemic Influenza Plan preparedness and response strategy. Therefore, any comparisons between different treatment groups were outside the scope and intent of this study and not performed. The primary immunogenicity objective of the study was to determine seroprotection rates (SPRs) in healthy adults following two doses of adjuvanted recombinant H7 vaccine based on serum hemagglutination inhibition (HAI) antibody titers on day 50, defined as an HAI antibody titer ≥1:40 against two representative A(H7N9) influenza viruses which emerged during the fifth epidemic and cocirculated in China: the HPAI A/GD/2016 virus used in the vaccine and LPAI A/HK/2017 virus.

DISCUSSION
Several policy documents from agencies of the US government establish a preparedness goal of maintaining enough prepandemic antigen and adjuvants to rapidly formulate vaccine from the NPIVS to vaccinate 26 million persons as well as to expand the domestic vaccine manufacturing capacity to produce enough vaccine for the entire US population within 6 months of a pandemic declaration [19][20][21] . More recently, the HHS Pandemic Influenza Plan was updated to establish delivery of first doses of pandemic vaccine within 12 weeks of a pandemic declaration 22 . Because recombinant influenza vaccines can proceed on virus sequence alone and do not require prior biosafety assessment and WHO distribution of a candidate vaccine virus to begin manufacturing 23,24 , the platform is expected to accelerate the time to release of product and thus be critical for a timely and effective pandemic response.
American Indian or Alaska Native Native Hawaiian or Other Pacific Islander Subjects with more than one race category recorded on the case report form appear in the multiracial category. Early antigenic analyses of fifth wave A(H7N9) influenza viruses published by the WHO indicated that the HAI titers of ferret antisera to the A/SH/2013 were substantially lower than the homologous titers 5 . In contrast, ferret antisera to the newly designated fifth wave CVVs, especially to the HPAI A/GD/2016 virus, were broadly reactive with all contemporary H7N9 viruses, including LPAI and HPAI viruses collected in 2016-2017 from humans and birds. These studies in ferrets led us to expect that recombinant H7 vaccine produced from the HPAI virus sequence would elicit highly cross-reactive antibody responses to a majority of the fifth epidemic emergent viruses in subjects who receive two doses with adjuvant. Furthermore, vaccine formulation with adjuvant may be able to overcome significant antigenic distance as the viruses evolve in subsequent years. In the absence of such viruses, antigenically distant ancestor viruses can be used for this purpose in assays; i.e., viruses that circulated during the first H7N9 epidemics in 2013. Indeed, strong cross-reactive antibody responses were observed to the antigenically distant A/Shanghai/2/2013 (H7N9) first epidemic strain suggesting that antibodies elicited by this adjuvanted vaccine may provide adequate protection for viruses evolving antigenically in the coming years. In agreement with these findings, three of six groups achieved seroprotection against the heterologous but antigenically related A/Hong Kong/125/2017 (H7N9) from the fifth epidemic wave. Furthermore, SPRs persisted near or above 70% at 3 months and remained above baseline at 6 months post-vaccination. The crossreactive immune responses observed here and in other studies [27][28][29][30] underscore the importance of continual testing of vaccines for candidate influenza viruses with high risk of pandemic potential. Sera obtained from vaccines that administered the A/SH/2013 (H7N9) antigen did not highly cross-react to fifth epidemic H7N9 viruses (Levine et al., 2017, personal communication), however in this study we show that subjects receiving Recombinant H7 had relatively high cross-reactive antibody responses to the only distantly antigenically related A/SH/2013 (H7N9). This one-way loss of reactivity is not completely understood, but the purity of the recombinant vaccine antigen may have a role. The National Institutes of Health (NIH), National Institute of Allergy and Infectious Diseases (NIAID) has several vaccine studies in progress using egg-based H7N9 vaccine derived from the LPAI A/HK/2017 virus, and cross-reactivity data generated will be used to better inform the choice of antigen. More studies are certainly needed and may include vaccines manufactured using multiple platforms, e.g., recombinant, cell-based, and/or egg-based antigens.
Taken together, recombinant H7 was safe and well-tolerated, with AEs reported being expected reactions to influenza vaccination, and the safety and immunogenicity of the recombinant H7 is comparable to that of the H7N9 vaccines produced in eggs or cells in response to the emergence of these viruses in 2013. While the NPIVS has an important role in a pandemic influenza preparedness and response, it is not the sole solution due to the ever-present antigenic evolution of influenza viruses. Therefore, the US must continue to develop more effective seasonal influenza vaccines and maintain sustainable domestic manufacturing capacity to rapidly produce, release, and deliver several hundred million doses of strain-matched influenza virus vaccines and dose-sparing adjuvant during a pandemic response. In addition, all efforts must continue to develop a universal influenza vaccine with broad and long-lasting immunity to protect from seasonal and pandemic influenza viruses.

Study design
This was a double-blind, randomized, Phase 2 clinical study (Clinical Trials. gov identifier: NCT03283319; registration date: September 12, 2017) that assessed the safety and immunogenicity of two doses of monovalent influenza recombinant H7 vaccine at three antigen dose levels (3.75, 7.5, and 15 µg) adjuvanted with AS03 or MF59 administered 28 days apart. This study was conducted in healthy male and non-pregnant female adults aged 18-49 years. It was conducted at four clinical research sites in the US between October 2017 and November 2018 in accordance with Good Clinical Practice guidelines, Declaration of Helsinki, and all applicable regulations. All study-related documents were approved by FDA, BARDA, and an institutional review board. Written informed consent was obtained from all enrolled subjects.

Study vaccine
The vaccine used in this study was recombinant H7, a monovalent H7N9 recombinant HA derived from A/Guangdong/17SF003/2016 (H7N9) HPAI virus, manufactured under a licensed process in the baculovirus expression vector system by Protein Sciences Corporation (Meriden, CT) 31 . AS03 adjuvant (GlaxoSmithKline, GSK) and MF59 adjuvant (Seqirus) were provided by the BARDA-managed NPIVS. Vaccines and adjuvants used in this study passed all release tests and met all specifications for both drug substance and final formulated drug product, and data were submitted, reviewed, and are on file as part of the Investigational New Drug (IND) application to FDA for this study. Subjects were randomized to receive one of three antigen dose levels (3.75, 7.5, or 15 μg) mixed 1:1 with adjuvant at the time of vaccination and administered intramuscularly (IM) as a 0.5 mL dose.

Safety assessment
The primary safety endpoints were solicited local or systemic reactogenicity symptoms that occurred within 8 days of each vaccine administration. These reactogenicity symptoms consisted of the following: solicited local reactions at the injection site (erythema/redness, induration/swelling, and pain) and solicited systemic reactions (fever, myalgia, arthralgia, fatigue, headache, nausea, vomiting, diarrhea, and chills). The AEs grading scale followed guidelines established by the Division of AIDS, NIH (i.e., mild, Grade 1; moderate, Grade 2; severe, Grades 3, 4, and 5) 32 .
Secondary safety assessments included unsolicited AEs from the time of the first vaccination through 21 days after the second vaccination. Additionally, venous blood samples for routine clinical laboratory safety evaluations were obtained at screening and 8 days after each of the two vaccinations. Secondary safety assessments included serious adverse events (SAEs), medically attended adverse events (MAAEs), and PIMMCs through 13 months following the first vaccine administration.

Immunogenicity assessment
Serum was collected prior to vaccination on day 1 and on days 29 (prior to second dose), 50, 121, and 212. HAI and MN assays were performed by Southern Research (Birmingham, AL) using Good Laboratory Practice as previously reported 16 against antigenic variants including reassortant homologous A/Guangdong/17SF003/2016xPR8 (CBER-RG7C, H7N9, FDA, Silver Spring, MD) or heterologous A/Hong Kong/125/2017xPR8 (IDCDC-RG56B, H7N9, CDC, Atlanta, GA), and A/Shanghai/02/2013xPR8 (IDCDC-RG32A, H7N9, CDC). The HAI and MN assays were qualified prior to immunogenicity assessment to determine the limit of detection (LOD) and to define the lowest (lower limit of quantitation, LLOQ) and highest (upper limit of quantitation, ULOQ) amount of analyte that could be measured with acceptable precision and accuracy (Table 3). Together, the ULOQ and LLOQ define the range of quantification for the respective assays, and results outside these ranges may be uncertain. Positive controls in the serological assays included ferret antisera to A/Guangdong/17SF003/     Briefly, sera were adsorbed with horse red blood cells (HRBCs, Lampire Biological Laboratories, Pipersville, PA) and tested for non-specific agglutination. Agglutinated or partially agglutinated samples underwent RBC adsorption a second time. HRBC-adsorbed serum samples were treated with receptor-destroying enzyme (RDEII, Denka-Seiken, Tokyo, Japan) to inactivate non-specific serum inhibitors. Serial, 2-fold dilutions of the HRBC-adsorbed and heat-inactivated serum were incubated with 4 hemagglutination units of the appropriate influenza virus to allow antigen-antibody binding. An equal volume of 0.5% HRBCs was added to each well. HAI titers were determined as the reciprocal of the highest serum dilution that completely inhibited hemagglutination. Sera tested for MN activity were heat-inactivated to remove non-specific inhibitors. Serial, 2-fold dilutions of inactivated serum were incubated with 100 TCID 50 /50 µl to allow for antigen-antibody binding. MDCK cells (1.5 × 10 4 cells; London cell line; Sigma-Aldrich, St. Louis, MO) were added to each well and incubated overnight. Cells were fixed in 80% acetone-PBS. ELISAs were performed to determine HA-specific IgG endpoint titers, and MN titers were defined as the reciprocal of the highest dilution of serum that gave 50% neutralization.

Statistical analysis
SPRs were defined as the proportion of subjects achieving a serum HAI antibody titer ≥40 against homologous antigen, and the 95% exact confidence intervals were determined for each individual group. GMTs and back-transformed 95% CIs based on the t distribution were summarized.    LLOQ lower limit of quantitation, ULOQ upper limit of quantitation, LOD limit of detection, nd not done. Fig. 6 Serum hemagglutination inhibition antibody responses to each of the two fifth epidemic H7N9 strains tested were highly correlated. HAI antibody titers generated by AS03-adjuvanted recombinant H7 (a) and MF59-adjuvanted recombinant H7 (b) against HPAI and LPAI H7N9 viruses from the fifth epidemic.