Abstract

1. Introduction

Although influenza A virus is commonly known as the cause of annual influenza epidemics in humans, wild aquatic birds represent the natural reservoir of this virus. Influenza A viruses are divided into subtypes based on antigenicity of their hemagglutinin (HA) and neuraminidase (NA) proteins. During circulation in poultry, viruses of subtypes H5 and H7 may acquire multiple basic amino acids at the HA cleavage site, rendering these viruses highly pathogenic with a mortality of up to 100% in poultry.

Occasionally, avian influenza A viruses cross the species barrier into humans. A pandemic may arise if such viruses are able to spread efficiently from human to human, causing worldwide morbidity and mortality. The introduction of a virus of the H1N1 subtype in the human population in 1918 is probably the best-known example of a pandemic. An estimated 50 million people died as a result of this so-called ‘Spanish influenza’ pandemic [1].

In recent years, the increase in outbreaks of highly pathogenic avian influenza (HPAI) in poultry and occasional transmission of these viruses to humans has caused great concern for the emergence of a new influenza A virus pandemic. Several outbreaks of low pathogenic avian influenza (LPAI) and HPAI H7 viruses in poultry have resulted in transmission of these viruses to humans (Table 1). Most of these human infections were in one or a few persons and resulted in conjunctivitis and, occasionally, mild respiratory symptoms [2–7]. However, an outbreak of HPAI H7N7 virus in the Netherlands in 2003 resulted in infection of 89 humans. Although most individuals developed conjunctivitis or mild respiratory symptoms, one patient died as a result of severe pneumonia and related complications [8, 9].

The first recorded direct transmissions of avian influenza viruses to humans leading to deaths occurred in 1997, when 18 people were infected during an outbreak of HPAI H5N1 in poultry markets in Hong Kong, six of whom died [10]. Since 1997, HPAI H5N1 viruses have continued to circulate in Asia and reemerged in Hong Kong in 2003. Subsequently, the HPAI H5N1 viruses have surfaced in poultry and wild birds across the Eastern hemisphere, also causing disease in several terrestrial carnivores [11–15] and humans [16, 17]. In humans, ~ 400 cases of HPAI H5N1 infection were detected to date. Patients infected with HPAI H5N1 often suffer from severe pneumonia, progressing to acute respiratory distress syndrome, with a fatal outcome in ~60% of infections [17, 18]. In patients infected with HPAI H5N1, virus cannot only be detected in the respiratory tract, but also in blood, cerebrospinal fluid and viscera [17, 19, 20]. Below, known determinants of pathogenicity of HPAI viruses in mammals are summarized.

2.1 Polymerase complex

The influenza A virus ribonucleoprotein complex contains 4 proteins necessary for virus replication: PB2, PB1, PA and NP. Several substitutions in the polymerase complex proteins have been described, mainly in PB2, that are important for adaptation of avian viruses to replication in mammalian hosts. A glutamic acid (E) to lysine (K) substitution at position 627 of PB2 was first described as a determinant of host range in vitro [21]; 627E is typically found in avian viruses, while human viruses generally have 627K. In mouse models for infection with HPAI H5N1 and HPAI H7N7 viruses, this substitution was shown to be the main determinant of pathogenicity. The PB2 E627K substitution was the main reason why a HPAI H5N1 virus isolated from a patient with a fatal infection in 1997 caused a lethal systemic infection in mice, whereas a virus isolated from a patient with relatively mild disease caused a non-lethal respiratory infection [22]. Due to this same E627K substitution, the virus isolated from a fatal case of infection with HPAI H7N7 virus in the Netherlands in 2003 was intrinsically more pathogenic in mice than a virus isolated from a patient with conjunctivitis [23]. Other indications of the importance of the E627K substitution in PB2 are that adaptation of A/equine/London/1416/73 (H7N7) to mice resulted, amongst others, in a E627K substitution in PB2, leading to a 1000x increased virulence of this virus [24]; that a HPAI H7N1 virus with 627K in PB2, isolated in Italy during the 1999–2000 outbreak, was more pathogenic in mice than viruses lacking this substitution [25]; that a E627K substitution in PB2 resulted in increased replication of viruses in the upper respiratory tract of mice [26] and that avian viruses without the E627K substitution acquire it spontaneously upon a single passage in mice [26–28]. There was no direct correlation between the presence of 627K in PB2 and the outcome of disease in HPAI H5N1 virus infected humans, but in this cohort almost all viruses which did not contain the E627K substitution harbored another substitution (D701N) in the polymerase proteins that has been implicated in adaptation to replication in mammals [19].

The pathogenicity of a variant of A/seal/Massachusetts/1/80 (H7N7), SC35M that is highly pathogenic to mice, was determined by a combination of two substitutions in PB2, D701N and S714R. These substitutions resulted in an increased polymerase activity in human cells [29]. A possible explanation for the increased replication efficiency may be that the D701N substitution in PB2 and a N319K substitution in NP resulted in a more efficient binding of these proteins to importin α1 and an increased nuclear localization of PB2, NP and importin α1 in human, but not avian cells [30]. The D701N substitution in PB2 was also found in HPAI H5N1 viruses isolated from healthy ducks in southern China from 1999 onwards. The viruses with 701N in PB2 were more pathogenic in mice than those with 701D in PB2 [27].

Although the effect was not mapped to individual substitutions, PB2 and PB1 contributed to the pathogenicity of A/Vietnam/1203/04 as compared to A/Chicken/Vietnam/C58/04 in mouse and ferret models. In both models, the chicken isolate was non-lethal whereas the human isolate caused considerable mortality [31]. In human cells, the polymerase activity of A/Vietnam/1203/04 was significantly higher than that of the chicken virus polymerase complex, again indicating that increased virus replication is related to pathogenicity [31].

Sequence comparison of avian and human influenza A viruses have identified 52 species-associated positions in the influenza A virus genome. Of these 52 positions, 35 were in the polymerase complex [32]. In similar genome comparisons, 32 amino acid positions were described that distinguish human and avian influenza viruses, of which 26 were in PB2, PA and NP [33]. Thus, substitutions in the polymerase complex proteins may facilitate influenza virus adaptation to mammals, and contribute to pathogenicity.

2.2 PB1-F2

The recently discovered PB1-F2 protein, transcribed from an alternative reading frame in the PB1 gene, induces apoptosis in infected cells. In mice, PB1-F2 contributed to the pathogenicity of HPAI H5N1 viruses isolated in 1997. All viruses that were highly pathogenic to mice had 66S in PB1-F2, whereas less virulent viruses had 66N. The N66S substitution resulted in more severe infection, higher virus titers and higher production of the cytokines IFN-γ and TNF-α in the lungs of infected mice [34]. Interestingly, 66S was also present in PB1-F2 of the 1918 Spanish influenza virus.

2.3 Hemagglutinin

One obvious determinant in HA of HPAI viruses is the multi basic cleavage site that renders these viruses highly pathogenic in chickens by facilitating systemic replication [35]. A HPAI H5N1 strain isolated in Hong Kong in 1997 that was pathogenic in mice, was highly attenuated upon replacement of the multi basic cleavage site with that of a low pathogenic virus [22]. Thus, cleavage of HA is not only a determinant of pathogenicity in poultry, but also in mammals.

HA also plays a role in viral adaptation through its’ function as the receptor binding protein. Whereas human influenza A viruses preferentially use α-2,6-linked sialic acids (SA) as a receptor, avian viruses prefer α-2,3-linked SA [36]. In virus attachment studies, human H3N2 viruses attach abundantly to the human trachea, in contrast to HPAI H5N1 virus. In the lower respiratory tract of humans, HPAI H5N1 virus attached predominantly to type II pneumocytes, alveolar macrophages and nonciliated cuboidal epithelial cells in terminal bronchioles [37]. The attachment pattern of H5N1 virus to these cells was in concordance with the diffuse alveolar damage observed in human cases of HPAI H5N1 virus infection [37]. In agreement with this attachment pattern of HPAI H5N1 virus, it was shown that α-2,3-linked SA are mainly present on nonciliated cuboidal bronchiolar cells and type II pneumocytes and are virtually absent in the human upper respiratory tract [38].

Immunohistochemical studies on lung samples from a patient infected with HPAI H5N1 virus showed that type II pneumocytes were predominantly infected with this virus [39]. However, replication of HPAI H5N1 virus was shown in human ex vivo tissues originating from the upper respiratory tract [40].

Mutations in HA that affect pathogenicity have been described, although the effect of these substitutions on receptor binding was not always shown. An I227S substitution in HA of a HPAI H5N1 virus isolated from a fatal case in Hong Kong in 1997 increased the virulence of this virus in mice [22]. The removal of a potential glycosylation site at position 154 of HA of the 1997 HPAI H5N1 viruses increased the virulence of these viruses [41]. HA of the HPAI H7N7 virus isolated from the fatal case in the Netherlands in 2003 contained 3 substitutions compared to a virus isolated from a patient with conjunctivitis, including a substitution introducing a potential glycosylation site near the receptor binding site. This HA caused increased virus titers in the lungs and distribution of virus to different organs in infected mice. In virus attachment studies, it was shown that there are subtle differences in attachment to the human lower respiratory tract of the virus isolated from the fatal case and a virus from a conjunctivitis case, that may explain differences in pathogenicity [23].

Adaptation of HA to enable efficient binding to α-2,6-linked SA is currently considered one of the most important prerequisites for efficient human-to-human transmission and thus for the emergence of a new pandemic. Although several HPAI H5N1 viruses with mutations that increase the affinity of HA for α-2,6-linked SA have been isolated, a HA with sole preference for human receptors has not yet been identified [42–46].

2.4 NS1

Influenza A virus replication induces an innate immune response, that can be modulated by NS1. Virus pathogenicity may in part be determined by NS1. In contrast to earlier HPAI H5 strains, the HPAI H5N1 strains isolated in Hong Kong in 1997 did not lead to increased production of the proinflammatory cytokine TGF-β in serum of infected mice [47]. A glutamic acid at position 92 of NS1 rendered 1997 HPAI H5N1 viruses insensitive to IFN-α, IFN-γ and TNF-α in vitro and was a determinant of pathogenicity in infected pigs [48]. HPAI H5N1 viruses that are pathogenic to mice induced much higher levels of inflammatory cytokines in mouse lungs [49]. Higher levels of proinflammatory cytokines were produced in human monocyte-derived macrophages infected with 1997 HPAI H5N1 virus, than in those infected with a human H1N1 or H3N2 virus [50]. In human primary bronchial and alveolar epithelial cells, higher levels of proinflammatory cytokines were induced by HPAI H5N1 viruses from 1997 and 2004 than by a human H1N1 virus [51]. NS1 of 1997 HPAI H5N1 viruses did not bind efficiently to CPSF30, resulting in higher levels of IFN-β mRNA and inhibition of viral replication. HPAI H5N1 viruses isolated after 1997 do bind to CPSF30, resulting in more efficient virus replication in vitro, due to a L103F and I106M substitution in NS1 [52]. This lack of NS1 binding to CPSF30 could be partially overcome by the internal genes of this virus [52]. In HPAI H5N1 viruses isolated from ducks in 1997, a P42S substitution in NS1 led to a 10000x more virulent virus in mice and lower IFNα/β levels in vitro [53]. Finally, NS of A/Vietnam/1203/04 was shown to be an important determinant of pathogenicity in ferrets [31].

In patients infected with HPAI H5N1 in Vietnam in 2004 and 2005, cytokine levels in peripheral blood were elevated, especially in patients with a fatal outcome of disease. Whether this was related to NS1 or merely due to enhanced virus replication is not known [19]. Likewise, it is unclear at present whether and to what extent the increased cytokine levels contribute to disease pathogenesis.

3. Conclusions

Adaptation of the polymerase complex proteins, NP, PB1-F2, HA and NS1 are all important features in the pathogenicity of avian influenza viruses in humans. Although the summary above focused on HPAI H5 and H7 viruses, remarkable similarities were observed in pathogenicity studies using the 1918 Spanish influenza virus. It was shown that the 1918 virus polymerase complex (including PB2 E627K) [54, 55], PB1-F2 (with 66N) [34], HA [55–57], NA [55, 57] and possibly NS1 [58] contributed to the pathogenicity of this virus. From these data, two important conclusions can be drawn. First, multiple genes determine the pathogenicity of influenza A viruses (Fig.1). Second, common mechanisms exist across influenza A virus strains and subtypes, through which influenza viruses adapt to mammals and gain or loose pathogenicity. While the determinants of avian influenza virus pathogenicity in mammals are thus becoming clear, the determinants for transmission between mammals remain largely unknown, requiring more attention in the future.

Open in a separate window

Figure 1

Determinants of adaptation and pathogenicity of avian influenza viruses in mammals. The functions of eight viral proteins, that are discussed in the main text as important determinants of pathogenicity, are indicated.

Acknowledgements

Footnotes

References