Influenza A(H7N9) Virus Evolution: Which Genetic Mutations Are Antigenically Important?

Integrated genetic and antigenic methods to characterize seasonal influenza viruses are increasingly being used to support surveillance, vaccine strain selection, and estimates of vaccine effectiveness [1, 2]. Advances in these methods and increased characterization of circulating viruses has also allowed identification of single-site mutations in the surface hemagglutinin (HA) and neuraminidase (NA) proteins that contribute to the influenza virus’ ability to escape the immunity of human populations [3, 4]. At the same time, ongoing global surveillance of seasonal human influenza viruses and emerging avian influenza viruses has meticulously documented the evolutionary trajectories of HA and NA. In this issue of The Journal of Infectious Diseases, Ning et al study mutations in the HA of influenza A(H7N9) viruses that have accumulated over the last 5 years, with the goal of identifying changes that might affect the virus’s antigenicity [5]. Influenza A(H7N9) virus is an avian virus that was first detected in humans in 2013 [6]. It has since resulted in at least 1625 infections and 623 deaths across 6 epidemic waves in China [7]. The frequency of infections and resulting deaths make this virus a serious pandemic threat; however, the virus still does not transmit efficiently from human to human [8]. The vast majority of infected individuals had contact with poultry, and other than a few cases linked to travel, infections have been limited to China. Notably, the 2016–2017 influenza A(H7N9) epidemic wave was the largest since this virus emerged, accounting for about one third of all infections and deaths to date. Mass immunization of Chinese poultry with a bivalent influenza A(H7N9) and A(H5N1) vaccine began in September 2017, and subsequently, only 3 human infections have been detected in the 2017–2018 sixth wave [7]. While A(H7N9) viruses evolve from year to year, like A(H3N2) and A(H1N1) viruses, most of the evolution happens in avian host species. Early isolates in 2013 demonstrated the circulation of genetically similar viruses from a common source [9]. Since then, influenza A(H7N9) viruses have diversified into a large number of clades that are dispersed to varying degrees across China [8]. The selection pressures driving the virus’s evolution in birds are unclear, particularly given that vaccines had not been used until recently. Most viruses contain G186V and Q226L/I substitutions in the HA gene, which mediate increased affinity for human sialic acid receptors [10]. Many also have substitutions in the PB2 protein, which is known to increase pathogenicity in mammalian hosts [11]. Concerningly, the fifth wave saw the emergence of viruses with a polybasic motif in HA characteristic of highly pathogenic avian influenza viruses. The ongoing evolution of influenza A(H7N9) virus and the large number of cases in the fifth wave have put human vaccine development and strain selection front and center. With the goal of understanding how the evolution may complicate strain selection for vaccines, Ning et al assessed the antigenic effects of specific influenza A(H7N9) virus mutations [5]. Using publicly available HA sequences of influenza A (H7N9) viruses isolated from humans, they identified 53 single amino acid substitutions relative to the ancestral A/Anhui/1/2013 virus that were found at high frequencies (>5%) or were located in a putative antigenic or receptor binding site. They then generated a panel of A/Anhui-derived pseudoviruses with these single amino acid substitutions and assessed the ability of sera from A/ Anhui-vaccinated guinea pigs to neutralize each one. They identified 3 amino acid substitutions—A143V, A143T, and R148K— that resulted in a 4-fold reduction in neutralization by A/Anhui antisera. Viruses containing both the A143V and R148K substitutions had a 10-fold reduction in susceptibility to neutralization. Importantly, this pairing of amino acid substitutions defined the major clade of influenza A(H7N9) viruses isolated in 2016–2017, representing about three quarters of human infections during the fifth wave. The A143V and R148K mutations both appear to have arisen in early 2014 at the root of a clade that dominated the fifth wave [12]. This explains their high frequency in the authors’ study but not what drove their emergence. Given their presence in antigenic sites, the authors frame the problem as one of antigenic drift. E D I T O R I A L C O M M E N T A R Y