H5N1-kannoilla tehtyjä rokotuksia tutkittu niiden IgG vasteesta

 https://journals.asm.org/doi/10.1128/mBio.00449-21

 DOI: https://doi.org/10.1128/mbio.00449

Vaccines

Research Article

6 July 2021

Share on

Broadly Reactive IgG Responses to Heterologous H5 Prime-Boost Influenza Vaccination Are Shaped by Antigenic Relatedness to Priming Strains

Authors: Jiong Wang, Dongmei Li, et al. 

Abstract

Prime-boost vaccinations of humans with different H5 strains have generated broadly protective antibody levels. However, the effect of an individual's H5 exposure history on antibody responses to subsequent H5 vaccination is poorly understood. To investigate this, we analyzed the IgG responses to H5 influenza A/Indonesia/5/2005 (Ind05) virus vaccination in three cohorts: (i) a doubly primed group that had received two H5 virus vaccinations, namely, against influenza A/Vietnam/203/2004 (Vie04) virus 5 years prior and A/Hong Kong/156/1997 (HK97) 11 years prior to the Ind05 vaccination; (ii) a singly primed group that had received a vaccination against Vie04 virus 5 years prior to the Ind05 vaccination; and (iii) an H5-naive group that received two doses of the Ind05 vaccine 28 days apart. Hemagglutinin (HA)-reactive IgG levels were estimated by a multiplex assay against an HA panel that included 21 H5 strains and 9 other strains representing the H1, H3, H7, and H9 subtypes. Relative HA antibody landscapes were generated to quantitatively analyze the magnitude and breadth of antibody binding after vaccination. We found that short-interval priming and boosting with the Ind05 vaccine in the naive group generated a low anti-H5 response. Both primed groups generated robust antibody responses reactive to a broad range of H5 strains after receiving a booster injection of Ind05 vaccine; IgG antibody levels persisted longer in subjects who had been doubly primed years ago. Notably, the IgG responses were strongest against the first priming H5 strain, which reflects influenza virus immune imprinting. Finally, the broad anti-H5 IgG response was stronger against strains having a small antigenic distance from the initial priming strain. IMPORTANCE The antigenic shift and draft of hemagglutinin (HA) in influenza viruses is accepted as one of the major reasons for immune evasion. The analysis of B cell immune responses to influenza infection and vaccination is complicated by the impact of exposure history and antibody cross-reactions between antigenically similar influenza strains. To assist in such analyses, the influenza "antibody landscape" method has been used to analyze and visualize the relationship of antibody-mediated immunity to antigenic distances between influenza strains. In this study, we describe a "relative antibody landscape" method that calculates the antigenic distance between the vaccine influenza strain and other H5 strains and uses this relative antigenic distance to plot the anti-H5 IgG levels postvaccination. This new method quantitatively estimates and visualizes the correlation between the humoral response to a particular influenza strain and the antigenic distance from other strains. Our findings demonstrate the effect of a subject's H5 exposure history on H5 vaccine responses quantified by the relative antibody landscape method.

Keywords: H5 monovalent influenza vaccine (MIV); HA imprinting; hemagglutinin (HA) antigenic distance; influenza virus antibody landscape; original antigenic sin (OAS).

INTRODUCTION

A number of highly pathogenic avian influenza (HPAI) A viruses, such as the H5, H7, and H9 strains, pose a significant threat to cause human pandemics as a result of their fast mutation rate and high pathogenicity (1, 2). To date, there is no evidence of sustained human-to-human transmission of these strains, despite repeated documentation that humans can contract these viruses from infected poultry (3). The first known human H5N1 infection was reported in 1997 during a poultry H5 outbreak in Hong Kong (4). From 2003 to January 2015, a total of 694 laboratory-confirmed human H5 cases were reported across 16 countries, and 58% of those people died as a result (5). Vaccination against future pandemic strains is the most viable path toward mitigating potential outbreaks. However, current H5 nonadjuvanted monovalent influenza vaccine (MIV) formulations are poorly immunogenic (6–10) and generally require a prime and boost strategy in order to achieve protective levels of immunity (11, 12). Interestingly, boosting with nonadjuvanted MIV, even in subjects who had been primed several years prior, led to robust and broad antibody responses to variant H5 MIVs (11). Such prime and boost strategies also appear to be needed for recent RNA vaccines (13) to other non-influenza virus vaccines, and understanding the immunobiology of this phenomenon remains highly relevant.

It has been generally accepted that immunological protection against influenza virus infection is due predominately to antibodies directed against the viral surface hemagglutinin (HA) protein, which is thus the major target of most influenza vaccines (14). A specific language has evolved to describe the potential confounding effects of such exposure on the development of subsequent immunity to influenza. HA imprinting is the initial exposure to an influenza virus strain, first described for childhood H1 influenza, which emerging evidence suggests may protect from subsequent H5 infection (2). However, when a person is sequentially exposed to two related virus strains, they tend to elicit an immune response dominated by antibodies against the first strain to which they were exposed (15, 16). This is true even following a secondary infection or vaccination. This phenomenon has been variously referred to as “original antigenic sin” (OAS), HA seniority, or a negative antigenic interaction (17–19). Thus, the immune response to a new influenza viral infection or vaccination is at least partially shaped by preexisting influenza immunity. Because there is still antigenic overlap between even mostly dissimilar influenza strains, it is critical to understand the antibody responses against antigenically similar virus stains for vaccine development, especially within the context of OAS.

The HA protein is composed of two domains, the highly plastic globular HA1 head domain and the conserved HA2 stalk domain. The hypervariable head domain is believed to be immunodominant, and virus infection or/and vaccination elicits strain-specific neutralizing antibodies primarily targeting this domain, resulting in limited cross-reactivity to divergent virus strains that vary significantly in their HA1 head domain sequences (20). In contrast, antibodies targeting the conserved HA2 stalk domain have been shown to broadly cross-react with multiple influenza viral strains (21). The viruses themselves can be categorized based on the phylogenetic distances of HA sequences. Ten clades of H5 HA (clades 0 to 9) have been identified within the H5N1 virus subtype (22). H5N1 viruses from clades 0, 1, 2, and 7 have the capacity to infect humans (23). These scatter into three distinct antigenic clusters, as determined by antigenic cartography generated by analyzing neutralizing serum antibody levels elicited in mice vaccinated against single influenza virus strains (1). An effective H5 influenza vaccine would ideally induce broad cross-reactivity against all three H5 clades. However, as discussed above, HA imprinting or OAS may impede the generation of broadly cross-reactive H5N1 antibodies if the prime and boost H5N1 vaccine strains reside in different antigenic clusters.

To address this issue, we reanalyzed serum samples from a previous H5 human vaccine study (DMID 08-0059) (24) using our mPlex-Flu multiplex assay (25) to measure the anti-HA IgG antibodies against all 10 clades (subclades) of H5 influenza virus. During this study (Fig. 1), longitudinal samples were collected prior to and after vaccination with an inactivated influenza A/Indonesia/5/2005 (Ind05) MIV from (i) subjects who had received two prime H5 MIV vaccinations (A/Hong Kong/156/1997 [HK97] in 1997 to 1998 and A/Vietnam/1203/2004 [Vie04] in 2005 to 2006 [the doubly primed long-interval boost {DL-boost} group]), (ii) subjects who had received only one prime Vie04 vaccination in 2005 to 2006 (long-interval boost [L-boost] group), and (iii) subjects in an H5 influenza virus-naive group, who were also given the Ind05 booster 28 days after the prime event (short-interval boost [S-boost] group). The mPlex-Flu assay (25) enables us to simultaneously evaluate the magnitude and breadth of the IgG repertoire directed against HAs from 21 H5 influenza virus strains and 9 other influenza A virus (IAV) strains (H1, H3, H7, H9). We also introduce a novel multiple-dimensional data analysis method named relative antibody landscapes, which enables quantitative analysis of antibody responses to antigenically similar influenza virus strains related to vaccine strains. The relative antibody landscapes method enables analysis of antibody-mediated immunity to a spectrum of HAs after H5 vaccine priming and boosting. This report demonstrates that as the relative antigenic distance between the original priming strain and the new H5 boosting vaccine strain becomes smaller (i.e., the strains are more antigenically similar), the greater the increase in the anti-HA IgG response to the original H5 MIV strain. Thus, in a vaccine response, the original HA imprinting influences vaccine responses occurring significantly later. We discuss the relevance of these findings to the development of influenza vaccines designed to induce broad antibody-mediated protection.