|Home | About | Journals | Submit | Contact Us | Français|
The emergence and widespread occurrence of antiviral drug-resistant seasonal human influenza A viruses, especially oseltamivir-resistant A/H1N1 virus, are major concerns. To understand the genetic background of antiviral drug-resistant A/H1N1 viruses, we performed full genome sequencing of prepandemic A/H1N1 strains. Seasonal influenza A/H1N1 viruses, including antiviral-susceptible viruses, amantadine-resistant viruses, and oseltamivir-resistant viruses, obtained from several areas in Japan during the 2007-2008 and 2008-2009 influenza seasons were analyzed. Sequencing of the full genomes of these viruses was performed, and the phylogenetic relationships among the sequences of each individual genome segment were inferred. Reference genome sequences from the Influenza Virus Resource database were included to determine the closest ancestor for each segment. Phylogenetic analysis revealed that the oseltamivir-resistant strain evolved from a reassortant oseltamivir-susceptible strain (clade 2B) which circulated in the 2007-2008 season by acquiring the H275Y resistance-conferring mutation in the NA gene. The oseltamivir-resistant lineage (corresponding to the Northern European resistant lineage) represented 100% of the H1N1 isolates from the 2008-2009 season and further acquired at least one mutation in each of the polymerase basic protein 2 (PB2), polymerase basic protein 1 (PB1), hemagglutinin (HA), and neuraminidase (NA) genes. Therefore, a reassortment event involving two distinct oseltamivir-susceptible lineages, followed by the H275Y substitution in the NA gene and other mutations elsewhere in the genome, contributed to the emergence of the oseltamivir-resistant lineage. In contrast, amantadine-resistant viruses from the 2007-2008 season distinctly clustered in clade 2C and were characterized by extensive amino acid substitutions across their genomes, suggesting that a fitness gap among its genetic components might have driven these mutations to maintain it in the population.
Seasonal outbreaks of influenza cause substantial morbidity and mortality and significant economic losses each year (33). Periodically, new strains emerge in humans and cause pandemics that pose a great threat to human health (31). Vaccines are very important for the prevention of infection with influenza virus, but antiviral drugs remain essential for treatment as well as prophylaxis. Two classes of antiviral drugs with activity against the influenza virus are available: the M2 ion channel blockers, or adamantanes (rimantadine and amantadine), and the neuraminidase inhibitors (NAIs; oseltamivir and zanamivir) (22, 39).
The rapid surge in amantadine-resistant influenza A/H3N2 viruses since the 2003-2004 season and among H1N1 viruses in the 2005-2006 season is a great concern to the medical and public health communities (3, 8, 29, 38). Remarkably, while amantadine-resistant A/H3N2 viruses swiftly replaced susceptible viruses and have become dominant since then, amantadine-resistant A/H1N1 viruses could outcompete susceptible viruses during only two successive seasons (2006-2007 and 2007-2008) and retreated during the 2008-2009 season (2, 5, 29, 34). Nonetheless, an oseltamivir-resistant A/H1N1 strain, referred to as the Northern Europe lineage, emerged in the 2007-2008 season and eventually prevailed in Europe (comprising 68% of A/H1N1 viruses collected) and the southern hemisphere and later became predominant in many countries, including Japan, during the 2008-2009 season (9, 13, 21, 41).
Antiviral resistance is conferred by a single amino acid substitution in the target protein. Almost all amantadine-resistant viruses of both the A/H1N1 and the A/H3N2 subtypes have a serine-to-asparagine mutation at position 31 (S31N) of the M2 ion channel protein (14, 29, 30, 32), and the oseltamivir-resistant A/H1N1 strain has a histidine-to-tyrosine mutation at position 275 (H275Y, N1 numbering) of the neuraminidase (NA) protein (9, 13, 21). While mutations in target proteins are usually selected by drug pressure, drug selection alone does not seem to be the sole driving force for the establishment of an efficiently replicating and transmissible strain (9, 19, 32). This notion is supported by the fact that a high proportion of oseltamivir-resistant strains, namely, strains of the Northern Europe lineage, was first observed in Europe, where the level of NAI consumption is generally low (19, 21, 26), while Japan, which has been using more oseltamivir than the rest of the world, detected high proportions of resistant viruses 1 year later (35, 40). It is important to note that oseltamivir-resistant viruses were detected earlier but remained sporadic and could not prevail like the Northern Europe resistant lineage (18). Thus, compensating mutations occurring elsewhere in the genome were suggested to improve the fitness and transmissibility of resistant viruses (9, 21).
To address this point, in the study described here, we performed full genome sequencing analysis of seasonal human influenza A/H1N1 viruses isolated in Japan during two influenza seasons, 2007-2008 and 2008-2009, to determine the genesis of antiviral drug-resistant viruses.
Twenty-three clinical influenza A/H1N1 isolates obtained from different regions in Japan during the 2007-2008 and 2008-2009 seasons were selected to represent major phylogenetic clusters of the hemagglutinin (HA) and NA trees of previously analyzed viruses. These viruses were previously characterized and tested for their susceptibilities to the antiviral drugs amantadine and oseltamivir (2).
RNA was extracted from 100 μl of the virus culture supernatants by using a commercial kit (Extragen II; Kainos, Japan), followed by reverse transcription with primer Uni12 (16) to make cDNA. PCR with primers with M13 overhangs on the 5′ ends was performed as described elsewhere (10). The PCR products were purified with an MSB Spin PCRapace kit (Invitek GmbH, Germany). Cycle sequencing was performed with a BigDye Terminator (version 3.1; Applied Biosystems) cycle sequencing kit, according to the manufacturer's instructions. The labeled products were then analyzed with an ABI Prism 3100 genetic analyzer (Applied Biosystems). The sequences were edited and assembled with SeqMan Pro software, included in the DNASTAR Lasergene package (Bioinformatics Pioneer DNASTAR, Inc.).
Full genome sequences available for human influenza A/H1N1 viruses recovered from 2003 to 2008 were downloaded from the Influenza Virus Resource database (1). A total of 346 full genome sets were available, with sequences from 2007 constituting the majority (294 sets). A phylogenetic tree for each segment was constructed, followed by cluster analysis, which was performed with TreeDyn software (7), to determine the topology of each virus in the phylogenies of each of the eight segments (data not shown). On the basis of the preliminarily constructed phylogenies, a representative sample of 50 viruses from major branches was selected for inclusion in the final analysis, along with the 23 viruses sequenced in this study. The final data set included genes encoding polymerase basic protein 2 (PB2; 2,280 nucleotides [nt]), polymerase basic protein 1 (PB1; 2,274 nt), polymerase protein (PA) (2,151 nt), hemagglutinin (HA) (1,698 nt), nucleoprotein (NP) (1,497 nt), neuraminidase (NA) (1,413 nt), matrix protein (M) (982 nt), and nonstructural protein (NS) (838 nt). Sequence alignments were generated with BioEdit (version 7.0) software (12). The phylogenetic history was inferred for each individual segment by the neighbor-joining method with bootstrap analysis (n = 1,000) by use of the MEGA (version 4.0) program (36). A/New Caledonia/20/1999, the vaccine strain recommended by WHO for use in the 2000 to 2007 influenza seasons and for which the full genome was available from the Influenza Virus Resource database (1), was employed as an outgroup root for the trees. Clades were based on the HA tree, as described in the report of Hauge et al. (13).
The genome sequences obtained in this study were deposited in the GenBank database under accession numbers CY043382 to CY043562.
The full genomes of the 23 influenza A/H1N1 isolates collected in Japan during the 2007-2008 and 2008-2009 influenza seasons were sequenced (Table (Table1).1). Five isolates from the 2007-2008 season were amantadine resistant and had the S31N mutation in the M2 protein, 6 isolates from the 2007-2008 season were amantadine and oseltamivir susceptible (antiviral drug susceptible), and 12 isolates from the 2007-2008 and 2008-2009 seasons were oseltamivir resistant and had the H275Y (N1 numbering) mutation. An additional 50 full genome sets for influenza viruses isolated in the United States, Australia, New Zealand, Japan, Nicaragua, and United Kingdom during the 2003 to 2008 seasons were included in the analysis; of the isolates in these samples, one isolate (recovered in the United States during the 2006-2007 season) was amantadine resistant and had the S31N mutation and one (recovered in the United Kingdom during the 2007-2008 season) was oseltamivir resistant and had the H275Y mutation.
The phylogenetic history of each of the genome segments is shown in Fig. Fig.1.1. Viruses from 2006 to 2009 were assigned to two major clades, clades 1 and 2 (subclades 2A, 2B, and 2C), on the basis of the tree created with the HA gene and previously reported classifications (21). Clade 1 accommodated the majority of the viruses from the 2006-2007 season which possessed the antiviral drug-susceptible genotype, and the viruses in clade 1 were most closely related to A/New Caledonia/20/1999-like virus (the vaccine strain recommended by WHO for use in the 2000 to 2007 seasons) in the HA phylogeny. Subclade 2A harbored antiviral drug-susceptible viruses from the 2006-2007 season, including A/Solomon Islands/3/2006-like virus (the vaccine strain for the 2007-2008 season). Subclade 2B could be further distinguished into two clusters, clusters 2B.I (representing the Hawaiian lineage) and 2B.II (representing the Northern Europe lineage). Subclade 2B.I harbored antiviral drug-susceptible viruses from the 2007-2008 season, including A/Brisbane/59/2007-like virus (the vaccine strain for the 2008-2009 season), in addition to two oseltamivir-resistant viruses with the H275Y mutation in their NA genes. Subclade 2B.II exclusively accommodated oseltamivir-resistant viruses from the 2007-2008 and 2008-2009 seasons that possessed the H275Y and D354G mutations in their NA genes. This lineage represented the major oseltamivir-resistant lineage that circulated in Europe during the 2007-2008 season (21) and in Japan and other countries during the 2008-2009 season. Subclade 2C accommodated amantadine-resistant viruses from the 2006-2007 and 2007-2008 seasons, and the viruses of subclade 2C were more closely related to A/Solomon Islands/3/2006-like viruses of subclade 2A.
In the phylogenies of all of the segments, amantadine-resistant subclade 2C was associated with the amantadine-susceptible A/Canterbury/106/2004 virus (the only virus from 2004 for which the full genome was available in the database) and was closely related to subclade 2A. On the other hand, subclade 2B had two different phylogenetic patterns. In the phylogenies of the PB2, PA, HA, and NA segments, subclade 2B was genetically related to subclades 2A and 2C, while in the phylogenies of the PB1, NP, M, and NS segments, it was most closely related to clade 1. Oseltamivir-resistant viruses, corresponding to the European resistant lineage, formed a distinct cluster (cluster 2B.II) within subclade 2B in the phylogenies of all segments except M and NS, in which oseltamivir-resistant and -susceptible viruses intermingled. This pattern strongly suggests that the oseltamivir-resistant strain has emerged in a reassortment event involving four segments each from a distinct oseltamivir-susceptible lineage, followed by the H275Y resistance-conferring mutation in the NA gene.
To gain better insight into the evolution of the amantadine-resistant A/H1N1 viruses and the oseltamivir-resistant A/H1N1 viruses, the amino acid sequence(s) of each segment was deduced. About 52 amino acid differences were found between subclade 2C amantadine-resistant viruses and subclade 2B amantadine-susceptible viruses (Fig. (Fig.1).1). These changes were not uniformly distributed among the 11 viral proteins. Of these mutations, 40 amino acid changes were detected in the subclade 2C background, and the remaining changes occurred in the subclade 2B background. The main oseltamivir-resistant strain (subclade 2B.II) had further mutations that consistently distinguished it from the oseltamivir-susceptible and -resistant viruses of subclades 2B.I and 2C: for PB2, the H453S mutation; for PB1, the N642S mutation (only viruses from the 2008-2009 season); for HA, the A193T mutation (only viruses from the 2008-2009 season); and for NA, the D354G mutation. The A193T mutation in HA (belongs to the receptor binding domain [RBD] and to the Sb antigenic site) was commonly detected among amantadine-resistant viruses and the major oseltamivir-resistant lineage, lineage 2B.II. All the A/H1N1 viruses sequenced in this study as well as those obtained from the database encoded for a truncated PB1-F2 protein of 57 amino acid residues.
Influenza A/H1N1 virus contributes to major epidemic strains and sometimes pandemic strains, such as the strain responsible for the current 2009 H1N1 pandemic. Genomic reassortment events between different strains of influenza A/H1N1 virus were associated with major antigenic variations and with the emergence of new strains (24, 25). Here we describe the results of a full genome sequence analysis of seasonal amantadine-resistant and oseltamivir-resistant influenza A/H1N1 viruses isolated during the 2007-2008 and 2008-2009 seasons in Japan. We show that amantadine-resistant A/H1N1 viruses originated from a strain that dates back to 2004, with extensive amino acid mutations occurring in the virus proteins. On the other hand, oseltamivir-resistant viruses were found to have evolved from a recent reassortant oseltamivir-susceptible strain during the 2007-2008 season by acquiring the H275Y resistance-conferring mutation in NA, in addition to minor amino acid modifications elsewhere in the genome.
The emergence of an amantadine-resistant influenza A/H3N2 virus strain worldwide in the 2003-2004 season has widely limited the utility of amantadine against influenza virus infections (3, 14, 28, 29). Despite that, the amantadine-resistant A/H1N1 virus strain still emerged worldwide during the following influenza season (8, 30). Our analysis revealed that this strain (subclade 2C) evolved independently of the major amantadine-susceptible strain (clade 1), predominantly detected in the 2006-2007 season, and that its closest genetic ancestry dates back to 2004.
In Japan, amantadine-resistant A/H1N1 viruses were first detected during the 2006-2007 season and were found to occur at a prevalence of 64.2% (30). They continued to circulate during the 2007-2008 season at a rate of 62.5% but virtually subsided and were replaced in the 2008-2009 season by the antigenically drifted amantadine-susceptible A/Brisbane/59/2007-like viruses (subclade 2B viruses in this study). The successful emergence and continued circulation of the amantadine-resistant A/H1N1 strain containing the S31N mutation in M2, despite the absence of amantadine usage in many countries, suggests an association of the S31N mutation with mutations in other genome segments that favor its immune evasion or its replication. In this study, extensive mutations were observed in the genetic background of the amantadine-resistant viruses. The HA alone had 7 to 8 amino acid changes, 3 of which (R192M/T, A193T, and T197K) belonged to receptor binding and antigenic sites (4, 43). These changes might have altered the antigenicity of amantadine-resistant viruses and contributed to their ability to compete with susceptible ones. Nevertheless, this lineage was replaced by the emergence of an antigenically variant amantadine-susceptible A/Brisbane/59/2007-like strain, and a consistent drop in the prevalence of amantadine-resistant A/H1N1 during the 2008-2009 season was reported worldwide (5, 42).
On the other hand, the phylogenetic topology of subclade 2B, constituted of oseltamivir-susceptible and oseltamivir-resistant viruses, strongly suggested that the subclade 2B viruses are a product of a four segment plus four segment (PB2, PA, HA, NA + PB1, NP, M, NS) reassortment event involving two independent lineages. Subclade 2B accommodated A/Brisban/59/2007-like virus, a strain that has antigenically drifted from the A/Solomon Islands/3/2006 strain (subclade 2A), explaining the ability of the viruses of subclade 2B to prevail.
Oseltamivir-resistant A/H1N1 viruses emerged in the background of oseltamivir-susceptible subclade 2B viruses, as revealed by the close clustering of these viruses in the phylogenies of the eight segments, by acquiring the H275Y resistance-conferring mutation in NA. Neuraminidase inhibitors, including oseltamivir, were designed to closely resemble the natural sialic substrate, and thus, the notion was that NAI-resistant mutants would unlikely be able to retain normal enzyme activity (37). It was consistently shown that the viability and pathogenicity of A/H1N1 viruses with the H275Y mutation were severely compromised both in vitro and in vivo compared with the viability and pathogenicity of the corresponding wild type (15, 17), but contrasting data later emerged implying that the fitness of resistant viruses is determined instead by their genetic background (44). Furthermore, the recent emergence and rapid spread of the A/H1N1 viruses with the H275Y mutation in several regions of the world demonstrate that recent resistant viruses retain significant pathogenicity and transmissibility (9, 11, 13, 21, 40, 41). Data from in vitro analysis also show that oseltamivir-resistant viruses possess growth characteristics similar to those of oseltamivir-susceptible viruses and amantadine-resistant viruses (data not shown). Thus, at least in vitro, the drug resistance genotype does not affect the replication of these viruses. It is possible that oseltamivir-resistant viruses have benefited from the replication and transmissibility fitness available in the genetic elements of susceptible viruses and have further acquired slight genetic modifications that allowed them to predominate over the susceptible viruses. This should be further investigated in an animal model, namely, the ferret model, which would allow a better understanding of virus-host interactions, as well as the efficiency of transmissibility that might have contributed to the rapid increase in oseltamivir-resistant viruses, to be obtained.
In addition to the H275Y mutation, all oseltamivir-resistant viruses except two from the 2007-2008 season, had G354 in the NA gene, whereas susceptible viruses had a G354D mutation. The location of residue 354 at the top the neuraminidase tetramer and away from the enzyme binding site makes it unlikely that it compensates for the H275Y mutation (27). Interestingly, the main strain of oseltamivir-resistant viruses (clade 2B.II) which constituted the sole resistant strain in the 2008-2009 season in Japan and other countries (5, 42) acquired a mutation (A193T) in the 190 loop of the RBD of HA1 (43), in parallel with the H275Y resistance-conferring mutation in NA. Analysis of neuraminidase activity revealed that H275Y reduces the enzyme activity of N1 (44), which might have forced a parallel mutation in the RBD of the HA to retain a balance of activity.
The mutations observed in the polymerase complex (H453S in PB2 and N642S in PB1) of subclade 2B.II oseltamivir-resistant viruses might be also implicated in the improved overall fitness of these viruses in the human host. The former mutation (H453S in PB2) belongs to the nuclear localization signal of PB2 (amino acids 449 to 495), which binds to the α-importin (23). The latter mutation (N642S in PB1) is located within the PB2 binding site in PB1, which makes it interesting to study whether this mutation could compensate for the H127Q substitution, located in the PB1 binding site on the PB2 segment (20), that occurred in the background of subclade 2B viruses (both susceptible and resistant isolates).
It was recently suggested that the continued use of monotherapy should be reconsidered in favor of using a combination of an adamantane and a neuraminidase inhibitor (26). Such a strategy would be useful if widespread antiviral drug resistance had not yet occurred. However, now that resistance to the adamantanes and NAIs has already been established, the use of combined therapy might force the selection of viruses with double resistance. The cocirculation of amantadine-resistant and oseltamivir-resistant strains sets the ground for the mixing of these two lineages and, consequently, might result in a combination of the S31N resistance-conferring mutation in M2 and the H275Y resistance-conferring mutation in NA in one genetic background, as was recently reported in Hong Kong (6). Given the evolutionary nature of influenza virus, the possibility that such a strain might prevail is very likely. Rationalization of the usage of antivirals on the basis of epidemic type/subtype information in each area or limitation of the use of antivirals to severely compromised or high-risk patients, such as elderly individuals and children, until resistance subsides should be considered. On the other hand, the importance of vaccine selection should not be undermined, as a vaccine with a closer immunogenic match to the resistant strains would help eradicate resistant viruses and give a window for susceptible viruses to come back.
In conclusion, genomic modifications at the amino acid level and recombination events play important roles in producing transmissible antiviral-resistant influenza strains. Therefore, the continuation and strengthening of influenza surveillance programs and influenza genome projects so that they include viruses from undersurveyed areas are of great importance for the quick identification of such strains as they emerge.
This study was supported by Japan grants-in-aid for scientific research from Special Research of Health and Labor Sciences Research Grants (Ministry of Education, Culture, Sports, Science and Technology, Japan) and the Acute Respiratory Infections Panel, U.S.-Japan Cooperative Medical Science Program (U.S. Department of Health and Human Services and U.S. Department of State in the United States and the Ministry of Foreign Affairs; the Ministry of Health, Labor, and Welfare; and the Ministry of Education, Culture, Sports, Science, and Technology in Japan).
We thank Y. Kato and K. Kudo of the International Medical Center of Japan and K. Mori and T. Enami of the Ministry of Health, Welfare, and Labor. We also thank the clinicians who helped with sample collection: R. Sugai, S. Kimura, N. Sasaki, T. Kawashima, I. Sato, S. Hibi, S. Ikushima, F. Fujiwara, K. Tsunamoto, M. Hashida, R. Matsuda, M. Seguchi, N. Ishitani, H. Masaki, Y. Shirahige, S. Degawa, N. Aso, and K. Hoshino. We are grateful to A. Miyashita and R. Kuwano in the Department of Molecular Genetics, Bioresource Science Branch, Center for Bioresources, Brain Research Institute, Niigata University, for facilitating the DNA sequencing. We thank A. Watanabe for her excellent technical assistance and Y. Kato for intensive secretarial work.
Published ahead of print on 3 February 2010.