Stockpiled Avian Influenza A(H7N9) Vaccines Induce Robust, Nonneutralizing Functional Antibodies Against Antigenically Drifted Fifth-Wave A(H7N9) Viruses

Abstract

Human infections caused by avian influenza A(H7N9) viruses have raised concerns of a pandemic. The capability of the current stockpiled A(H7N9) vaccines to induce cross-protective, nonneutralizing functional antibodies against antigenically drifted A(H7N9) viruses has not been evaluated before. Here we show that vaccination with either MF59- or AS03-adjuvanted inactivated A(H7N9) vaccines elicited robust, cross-reactive antibody-dependent cell-mediated cytotoxicity–mediating and neuraminidase-inhibiting functional antibodies against the antigenically drifted A(H7N9) viruses that emerged recently during the fifth-wave outbreak in China, including a highly pathogenic A(H7N9) human isolate. Such cross-reactive humoral immunity may provide vital first-line defense against fatal outcomes in case of an A(H7N9) pandemic.

Since the first human infection with avian influenza A(H7N9) virus was reported in March 2013, there have been a total of 6 “waves” of the A(H7N9) epidemic in mainland China so far, with 1567 laboratory-confirmed human cases and a case fatality rate of 39% [1]. Whereas only low pathogenic avian influenza viruses were detected during the first 4 waves of Asian A(H7N9) infection, the fifth wave in 2017 has been the most severe. Genetically, the circulating A(H7N9) viruses have diverged into 2 lineages, Pearl River Delta (PRD) lineage and Yangtze River Delta (YRD) lineage [2]. Antigenically, virus isolates from the YRD lineage have drifted away from the PRD lineage, as evident in analysis using ferret antisera [2]. Moreover, highly pathogenic avian influenza viruses (HPAIs) emerged during the fifth-wave A(H7N9) epidemic that were capable of transmitting and causing lethal infection in ferrets tested without prior adaption [3]. Together, current evidence suggests a high pandemic potential of the new emerging Asian A(H7N9) viruses.

Previous studies have established that inactivated prepandemic H7 vaccines were poorly immunogenic in humans [4]. A 2-dose regimen of adjuvanted vaccine is generally required to induce detectable neutralizing antibody responses to homologous H7 viruses. In the United States, A(H7N9) vaccines have been produced and stockpiled based on the 2 viruses that circulated during the first-wave A(H7N9) epidemic in 2013, that is, A/Shanghai/2/2013 (SH2) and A/Anhui/1/2013 (AH1) [2]. Comprehensive clinical trials have demonstrated that 2 doses of MF59- and/or AS03-adjuvanted, inactivated A(H7N9) vaccines were able to induce robust neutralizing antibody titers to the homologous A(H7N9) vaccine viruses in the majority of study participants [5, 6]. However, the cross-protective potential of the A(H7N9) vaccines measured by traditional hemagglutination inhibition and neutralization assays against antigenically drifted A(H7N9) viruses, such as those circulated during the fifth-wave H7N9 outbreak in China, appeared to be low (Levine et al, manuscript in preparation).

Development of strain-matched pandemic vaccines in case of an A(H7N9) pandemic poses extraordinary challenges due to unpredictability of hemagglutinin (HA) drift and time constraints of vaccine production. The likelihood of an HA antigenic mismatch between the stockpiled vaccine and newly emerging A(H7N9) pandemic strains is fairly high. Therefore, we evaluated the levels of cross-reactive functional antibodies that are not measured in traditional neutralization assays. We used sera from individuals immunized with the current stockpiled A(H7N9) vaccines based on the first-wave A(H7N9) virus and examined antibody-dependent cell-mediated cytotoxicity (ADCC) and neuraminidase inhibition (NAI) antibody responses induced against newly emerged fifth-wave A(H7N9) viruses that caused the recent severe outbreak in China.

METHODS

Human Serum Samples

Two panels of paired human serum samples were analyzed. The first panel was collected from 30 healthy adults (median age, 30 years [range, 21–58 years]) who received 2 doses of 7.5 μg HA per dose of MF59-adjuvanted, inactivated SH2 split vaccine with a 21-day interval [5] (ClinicalTrials.gov identifier: NCT01938742). The second panel was sampled from 30 healthy adults (median age, 40 years [range, 19–59 years]) who were immunized with 2 doses of 3.75 μg HA per dose of AS03-adjuvanted, inactivated SH2 split vaccine 21 days apart [6] (ClinicalTrials.gov identifier: NCT01942265). Sera were subsets randomly selected from strong responders based on immunogenicity evaluation of the original studies. Paired sera were collected on day 0 (pre-) and day 42–52 (post-) after the second dose of the vaccination.

Use of the sera was approved by Centers for Disease Control and Prevention, National Center for Immunization and Respiratory Diseases research determination review.

ADCC Natural Killer Cell Activation Assay

Influenza HA-specific ADCC-mediating antibodies were quantified using an ADCC natural killer (NK) cell activation assay described in detail previously [7]. Two hundred nanograms per well of full-length, trimeric, recombinant A(H7N9) HA proteins were used as antigens. Sera were heat-inactivated and serially diluted starting with 1:10 dilution. The ADCC antibody titers were defined as the highest serum dilution that achieved 3% threshold of CD107a⁺ NK cells in the assay as designated previously. A titer <40 was arbitrarily considered negative, whereas ≥40 was considered positive.

Enzyme-Linked Lectin Assay

Influenza NAI antibodies were quantified using enzyme-linked lectin assay as described previously [8]. Sera were heat-inactivated and serially diluted starting with 1:10 dilution. Two reassortant influenza viruses expressing a mismatched H6 antigen and an N9 antigen either from AH1 virus or from A/Guangdong/17SF003/2016 A(H7N9) virus (GD17SF) were used as NA antigen sources. The quantity of the reassortant viruses used was calibrated so that either of the viruses generated an optical density reading between 3.2 and 3.6 in the assay. The results were expressed as the highest serum dilution that achieved 50% of NAI in the assay.

RESULTS

HA-Specific ADCC-Mediating Antibodies

We first examined the ability of the MF59- and AS03-adjuvanted, inactivated A(H7N9) vaccine to elicit HA-specific ADCC-mediating antibodies. The titers of the ADCC antibodies to both the homologous and the heterologous H7 antigens of the 3 fifth-wave Asian A(H7N9) viruses tested were marginal before vaccination (between 31.7 and 42.9 pre-geometric mean titer [GMT]) (Table 1). There is no significant difference in baseline ADCC titers between the MF59- and AS03-adjuvanted groups (P > .05). Vaccination with MF59- and/or AS03-adjuvanted SH2 vaccine elicited strong ADCC-mediating antibody responses to the homologous H7 antigen (post-GMT: 508.0 in the MF59 group and 678.1 in the AS03 group). Among the subjects who received MF59-adjuvanted vaccine, 93.3% achieved seroconversion, eg, 4-fold or higher ADCC antibody titer rise postvaccination. Notably, strong cross-reactive ADCC antibodies were observed as well. The postvaccination ADCC antibodies to one PRD lineage A(H7N9) virus (HK61) and 2 YRD lineage A(H7N9) viruses (HK125 and GD17S) ranged between 359.2 and 485.1, with a seroconversion rate equal or similar to vaccine antigen (SH2). Similar results were observed in those subjects who received AS03-adjuvanted SH2 vaccine. Note that overall, the postvaccination ADCC antibody titers and the seroconversion rates were higher in the AS03-adjuvanted vaccine group compared with the MF59-adjuvanted vaccine group. However, the difference did not reach statistical significance (P > .05).

NAI Antibodies

We then determined the titers of NAI antibodies elicited by the 2 stockpiled, adjuvanted A(H7N9) vaccines. Due to limitations of the available H6N9 antigens, only A(H6N9) reassortant viruses generated from the N9 of AH1 and the HPAI DG17S A(H7N9) virus was tested in the current study. There were only background levels of preexisting, H7-specific NAI antibodies in both panels of serum samples tested before vaccination (pre-GMT range, 9.0–21.1) (Table 2). Vaccination with MF59-adjuvanted, inactivated SH2 vaccine led to a robust NAI antibody response to the homologous N9 antigen. The post-GMT titer from the 30 immunized individuals was 68.9, with an 83.3% seroconversion rate. Most importantly, robust NAI antibodies cross-reactive to the N9 antigen from the HPAI GD17S A(H7N9) virus were detected (post-GMT: 228.9 [95% confidence interval, 169.4–309.3]), with 93.3% achieving seroconversion. The AS03-adjuvanted vaccine group also mounted robust NAI antibody responses. Similar to the ADCC antibody responses (Table 1), NAI antibodies detected in the AS03-adjuvanted vaccine group appeared to be higher than those immunized with MF59-adjuvanted SH2 vaccine. However, the difference was statistically not significant (P > .05).

DISCUSSION

Strain-specific neutralizing antibodies are critical in prevention of influenza infection. Recently, we found that only moderate levels of cross-reactive neutralizing antibodies against antigenically drifted fifth-wave A(H7N9) viruses were induced following vaccination with MF59- and/or AS03-adjuvanted stockpiled A(H7N9) vaccines (Levine et al, manuscript in preparation). A recent study also showed that titers of cross-reactive HI antibodies to fifth-wave HPAI A(H7N9) viruses were low after either A(H7N7) prime followed by unadjuvanted A(H7N9) boost, or 2 doses of AS03-adjuvanted A(H7N9) vaccination based on 2013 antigens [9]. Increasing evidence suggests that nonneutralizing functional antibodies play an important role in the mitigation of severe A(H7N9) avian influenza virus infection in humans [10]. Therefore, we examined the capability of 2 adjuvanted A(H7N9) vaccine regimens to induce cross-reactive, nonneutralizing functional antibodies in the present study.

Our data revealed that immunization with MF59- and/or AS03-adjuvanted inactivated A(H7N9) vaccines resulted in induction of robust HA-specific ADCC-mediating antibodies against antigenically drifted A(H7N9) viruses from the fifth-wave A(H7N9) outbreak (Table 1). Several recent studies have shown that influenza-specific ADCC antibodies may directly contribute to immune protection against influenza clinical illness caused by the A(H7N9) viruses. It was observed that early onset of high titers of HA-specific, nonneutralizing antibodies was directly correlated with the clinical outcome of the patients with severe A(H7N9) influenza infection [10]. Passive transfer experiments in a mouse model have shown that H7-specific, nonneutralizing human monoclonal antibodies can provide substantial protection via FcγR-mediated mechanisms [11]. In light of these new findings, it is conceivable that the cross-reactive ADCC-mediating antibodies induced by MF59 and/or AS03 adjuvanted A(H7N9) vaccines based on a 2013 A(H7N9) virus may contribute to the recovery of severe A(H7N9) infections from recent fifth-wave A(H7N9) viruses, even in the presence of only low levels of cross-reactive neutralizing antibodies.

Contribution of NAI antibodies in protection against influenza illness has been well documented. It has been demonstrated in a mouse model that NAI antibodies were capable of providing a broad range of cross-protection against heterologous influenza viruses within the same NA subtype of the vaccine strain [12]. Moreover, results from recent clinical trials of influenza vaccines have demonstrated that NAI antibodies are a new, independent correlate of protection against influenza [13, 14]. In the present study, we observed that immunization with either of the adjuvanted A(H7N9) vaccines led to induction of robust, cross-reactive NAI antibodies against a highly pathogenic A(H7N9) virus isolated from a patient during the recent fifth-wave A(H7N9) epidemic in China. Together, our data strongly suggest that the cross-reactive NAI antibodies detected may represent a second major component of the nonneutralizing functional antibodies that could contribute to the reduction of disease severity from fifth-wave A(H7N9) infections in humans.

Despite the substantial HA antigenic drift since the first wave of the A(H7N9) outbreak, we detected strong cross-reactive NAI antibody responses to fifth-wave viruses following vaccination using 2013 antigens (Table 2). N9 antigens between 2013 SH2 vaccine virus and 2017 HPAI GD17S virus share 98.2% amino acid sequence identity and only differ in 3 amino acid residues among 7 potential antigenic sites in the N9 antigens (Supplementary Table 1). Such discordant antigenic drift of HA and NA may be of particular importance in combating an A(H7N9) pandemic using stockpiled vaccines.

A detailed head-to-head comparison has revealed that compared to MF59, AS03-adjuvanted A(H7N9) vaccines elicited stronger neutralizing antibody responses to the vaccine virus [6]. In the current study, the sera used are only a subset selected from 2 different original studies. We only evaluated the antigen doses anticipated to be used for stockpiled A(H7N9) vaccines. The amount of the NA antigens is not controlled in influenza vaccine formulations generally, and here we did not measure the relative NA content in the 2 SH2 vaccine regimens used for the immunization. Nonetheless, our results reveal that either MF59 or AS03 adjuvant was efficient in stimulating cross-reactive ADCC and NAI antibody responses to the antigenically drifted fifth-wave A(H7N9) viruses.

Multiple immune mechanisms may contribute to recovery from severe A(H7N9) infection in humans, including CD8/CD4 T cells and neutralizing and nonneutralizing antibodies [10]. A recent study in a ferret model has shown that vaccination with an inactivated, low pathogenic A(H7N9) vaccine led to accelerated viral clearance in the upper respiratory tract and significantly reduced clinical illness in the immunized animals following challenge with a highly pathogenic A(H7N9) virus, despite low titers of cross-reactive neutralizing antibodies against the challenge virus [15]. In case of severe A(H7N9) infection in humans, we believe that cross-reactive, nonneutralizing functional antibodies, such as ADCC and NAI antibodies, may constitute key components of the immune response to prevent fatal outcome caused by new emerging A(H7N9) viruses. Despite the antigenic drift of HAs, the stockpiled A(H7N9) vaccines based on the 2013 A(H7N9) viruses with either of the adjuvants may provide partial protection against newly emerging A(H7N9) viruses through the induction of nonneutralizing, functional ADCC and NAI antibodies.

SUPPLEMENTARY DATA

Supplementary materials are available at The Journal of Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.

Notes

Acknowledgments. The authors thank Dr Chris Roberts (National Institutes of Health) for kindly providing the 2 panels of the human serum samples; Dr James Stevens (Influenza Division, Centers for Disease Control and Prevention [CDC]) for providing the recombinant hemagglutinin antigens; Dr Maryna Eichelberger (US Food and Drug Administration) for providing the reassortant A(H6N9) influenza viruses, and St Jude Children’s Research Hospital for providing plasmids that were used to generate these reassortant influenza viruses; and NantKwest Inc and Dr Kerry Campbell (Fox Chase Cancer Center) for providing the human natural killer cell line. The authors also thank industry partners from Seqirus USA and colleagues from the Biomedical Advanced Research and Development Authority, US Department of Health and Human Services, for critically reading the manuscript.

Disclaimer. The views expressed in this work solely represent those of the authors and do not reflect the official policy of the CDC. GlaxoSmithKline Biologicals SA was provided the opportunity to review a preliminary version of this manuscript for factual accuracy, but the authors are solely responsible for final content and interpretation.

Financial support. This work was supported by the CDC.

Potential conflicts of interest. Both authors: No reported conflicts of interest. Both authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.

References