The Lancet 2004; 363:582-583 February 2004
DOI:10.1016/S0140-6736(04)15624-9
The inexact science of influenza prediction Maria Zambon a
See Articles
See Research Letters
Predicting the clinical outcome of zoonotic infection with influenza A remains an inexact and observational science. This week's Lancet contains clinical descriptions of two separate avian-to-human transmission events in 2003. Malik Peiris and colleagues report H5N1 in Hong Kong SAR China in a small family cluster with high mortality, respiratory illness, and multisystem failure caused by viruses similar to those associated with fatal cases in Vietnam and Thailand in 2004. Marion Koopmans and colleagues report H7N7 in the Netherlands in a large outbreak with 83 cases, mostly with conjunctivitis, a few mild respiratory illnesses, and one death. These contrasting clinical scenarios are a salutary reminder that knowledge of viral genome sequences, available within days of identification of virus, is necessary but not sufficient to predict clinical outcome, since the course of any infection is a balance between host response and virus replication.
Transmission and pathogenesis are two separate but linked characteristics of an infectious disease. Fitness gain of an organism to improve transmissibility may result in loss of some pathogenic features. Viruses which contain RNA, such as influenza A, are flexible and can mutate rapidly; when this property is allied with a segmented genome that allows recombination easily in nature, a formidable display of genetic diversity emerges. Influenza A subtypes are classified on the basis of viral haemagglutinin (HA) and neuraminidase (NA) surface-proteins. Only a few restricted subtypes?H1N1, H1N2, and H3N2?circulate in mammalian hosts, whereas all subtypes circulate freely in the natural hosts, wild birds, mixing and exchanging gene segments and allowing rapid mingling of genotypes, although rarely in association with disease. H5 and H7 subtypes can cause serious systemic illness and death in domestic poultry, in which it is assumed they are acquired from contact with wild birds.
Both the H5N1 and the H7N7 virus described in the two papers contain fully avian gene segments and receptors specific for α2,3 galactosidase rather than the human receptor α2,6 galactosidase, both have a multibasic cleavage site, and both are associated with severe disease in poultry flocks. However, in human beings one virus has a mortality ratio of 50% or higher, whereas the other has a ratio of less than 5%, perhaps with a different spectrum of disease including conjunctivitis, an illness not seen with H5N1.
The lack of a specific human receptor is evidently not a barrier to cross-species transmission of influenza A virus but it may contribute to the lack of human-to-human transmissibility. Available data suggest that adaptation of an avian receptor to a more human receptor would not require many mutational events in different subtypes1,2 and that when influenza A viruses do cross into human beings, adaptation is rapid.3 A key feature signalling significant adaptation of avian influenza virus to a human host would be the acquisition of the ability to transmit from human being to human being; this adaptation might simply involve adaptive changes in the receptor properties or might also require improved replication efficiencies, perhaps involving key mutations in the viral polymerase genes, increasing the viral load in the respiratory tract, and altered interaction with the host immune system involving changes in other viral genes. This adaptation might be achieved by sequential mutation in an avian virus genome or by mixing segments of an avian virus with segments from a virus already adapted to human beings.
Study of H5 and H7 subtypes in avian models has given rise to some of our understanding of influenza pathogenesis, notably the importance of a sequence of aminoacids at the junction between two parts of HA1 and HA2, known as the multibasic cleavage site. The correlation between this sequence motif and severe illness and death in birds leads to the inference that, if this sequence motif is present in an influenza virus in mammals, it will also be associated with severe illness. There is some justification for this view, derived from the use of reverse genetic systems in which H5 influenza viruses are engineered to remove the multibasic site with a resulting reduction in pathogenicity in a mouse model4 and from the available data on H5 infection in human beings.5,6 However, the genome of the 1918 influenza H1N1 Spanish influenza, which was associated with over 20 million deaths worldwide, did lack a multibasic cleavage site,7 indicating that different viral subtypes can behave differently in human beings despite common genetic motifs and that we lack information about key mutations associated with mammalian pathogenicity. Differences in clinical outcome could relate to differences between the H7N7 genome and the H5N1 genome, rather than the common sequences. Determining human cytokine responses in H5N1 compared with H3N2 infection may be important in pinpointing key viral-host interactions, as shown, by Peiris and colleagues' paper in this issue and previously.8
The presence of a multibasic cleavage site in the avian model of disease is sufficient to extend viral replication from the gastrointestinal tract of birds to replication in many different organs, which is considered to be the basis of pathogenicity in birds. Yet, in the 1997 H5 human cases and the H5N1 cases described today, there is no evidence of replication of virus outside the respiratory tract, although there may be evidence of multisystem failure.5,6 Peiris and colleagues advance cautious evidence which is consistent with earlier findings by the same group8 and work in the pig model9 which suggests that pathogenesis in human beings may be related to aberrant immune responses which in turn may correlate with the sequences of other non-structural viral genes. Pathogenicity in human beings might be dependent on motifs in several different viral genes.
It is useful to know the spectrum of disease associated with infection. Serum surveys showed no evidence of extensive penetration of H5N1 in 1997 into an exposed human population suggesting very few mild or asymp-tomatic illnesses.10 This situation is in contrast to the situation described for H7N7 viruses in the Netherlands by Koopmans and colleagues, where significant numbers of people were infected but with few cases of respiratory illness or death. Serological data from studies of cases and contacts in the Netherlands will indicate whether there have been further infections associated with mild or subclinical infections. The first description of H7N7 conjunctivitis in human beings occurred in individuals who acquired H7N7 transmitted from seals being swabbed to investigate the cause of fatal illness, emphasising the contribution of the host to outcome of viral infection;11 other human cases have been described since.12 Perhaps the particular combination of H7 HA and N7 NA make this virus subtype particularly adept at targeting human conjunctival cells, rather than other cells in the respiratory tract.
Taken together, the two papers in today's Lancet emphasise that detailed experimental work is required to understand the virus-host cell interactions which give rise to the spectrum of illness when an avian influenza A virus jumps the species barrier into a mammalian host. Observations for one viral subtype might not hold for another. Different mammalian hosts have different diseases with the same virus, indicating the necessity for caution in extrapolating data from animal models.
Careful clinical observations will be essential in future zoonotic infections of influenza A to add to our understanding of disease, but may be overlooked in the maelstrom of ensuring the immediate public-health responses to the threat of avian influenza. Reducing the probability of adaptation of H5 or H7 to human beings by reducing the potential for recombination with viruses adapted to human beings underlies the current control policies of depopulation of infected poultry and intensive surveillance of human influenza cases in south-east Asia, which were successfully applied in the Netherlands in 2003. Influenza A zoonosis remains a rare event, but one which history teaches us we must take seriously to avoid a repeat of the 1918 pandemic of influenza.
I have no conflict of interest to declare.
<!--start simple-tail=-->References
1. Gamblin SJ, Haire LF, Russell RJ, et al. The structure and receptor binding propeorties of the 1918 influenza hemagglutinin. Science 2004;10.1126 /science 1093155.
2. Harvey R, Martin AC, Zambon M, Barclay WS. Restrictions to the adaptation of influenza A virus H5 hemagglutinin to the human host. J Virol 2004; 78: 502-507. MEDLINE | CrossRef
3. Matrosovich M, Tuzikov A, Bovin N, et al. Early alterations of the receptor-binding properties of H1, H2, and H3 avian influenza virus hemagglutinins after their introduction into mammals. J Virol 2000; 74: 8502-8512. MEDLINE | CrossRef
4. Hatta M, Gao P, Halfman P, Kawaoka Y. Molecular basis for high virulence of H5N1 influenza A viruses. Science 2001; 293: 1840-1842. MEDLINE | CrossRef
5. Yuen KY, Chan PK, Peiris M, et al. Clinical features and rapid viral diagnosis of human disease associated with avian influenza A H5N1 virus. Lancet 1998; 351: 467-471. Abstract | Full Text | PDF (174 KB) | MEDLINE | CrossRef
6. To KF, Chan PK, Chan KF, et al. Pathology of fatal human infection associated with avian influenza A H5N1 virus. J Med Virol 2001; 63: 242-246. MEDLINE | CrossRef
7. Taubenberger JK, Reid AH, Krafft AE, Bijwaard KE, Fanning TG. Initial genetic characterisation of the 1918 Spanish influenza virus. Science 1997; 275: 1793-1796. MEDLINE | CrossRef
8. Cheung CY, Poon LLM, Lau ASY, et al. Induction of proinflammatory cytokines in human macrophages by influenza A (H5N1) viruses: a mechanism for the unusual severity of human disease. Lancet 2002; 360: 1831-1837. Abstract | Full Text | PDF (115 KB) | MEDLINE | CrossRef
9. Seo SH, Hoffmann E, Webster RG. Lethal H5N1 influenza viruses escape host antiviral cytokine responses. Nat Med 2002; 8: 950-954. MEDLINE | CrossRef
10. Bridges C, Katz JM, Seto WH, et al. Risk of influenza A (H5N1) infection among health care workers exposed to patients with influenza A (H5N1), Hong Kong. J Infect Dis 2000; 181: 344-348. MEDLINE | CrossRef
11. Webster RG, Geraci J, Petursson G. Conjunctivitis in human beings caused by influenza A virus of seals. N Engl J Med 1981; 304: 911. MEDLINE
12. Kurtz J, Manvell R, Banks J. Avian influenza virus isolated from a woman with conjunctivitis. Lancet 1996; 348: 901-902. Full Text | PDF (52 KB) | MEDLINE | CrossRef
DOI:10.1016/S0140-6736(04)15624-9
The inexact science of influenza prediction Maria Zambon a
See Articles
See Research Letters
Predicting the clinical outcome of zoonotic infection with influenza A remains an inexact and observational science. This week's Lancet contains clinical descriptions of two separate avian-to-human transmission events in 2003. Malik Peiris and colleagues report H5N1 in Hong Kong SAR China in a small family cluster with high mortality, respiratory illness, and multisystem failure caused by viruses similar to those associated with fatal cases in Vietnam and Thailand in 2004. Marion Koopmans and colleagues report H7N7 in the Netherlands in a large outbreak with 83 cases, mostly with conjunctivitis, a few mild respiratory illnesses, and one death. These contrasting clinical scenarios are a salutary reminder that knowledge of viral genome sequences, available within days of identification of virus, is necessary but not sufficient to predict clinical outcome, since the course of any infection is a balance between host response and virus replication.
Transmission and pathogenesis are two separate but linked characteristics of an infectious disease. Fitness gain of an organism to improve transmissibility may result in loss of some pathogenic features. Viruses which contain RNA, such as influenza A, are flexible and can mutate rapidly; when this property is allied with a segmented genome that allows recombination easily in nature, a formidable display of genetic diversity emerges. Influenza A subtypes are classified on the basis of viral haemagglutinin (HA) and neuraminidase (NA) surface-proteins. Only a few restricted subtypes?H1N1, H1N2, and H3N2?circulate in mammalian hosts, whereas all subtypes circulate freely in the natural hosts, wild birds, mixing and exchanging gene segments and allowing rapid mingling of genotypes, although rarely in association with disease. H5 and H7 subtypes can cause serious systemic illness and death in domestic poultry, in which it is assumed they are acquired from contact with wild birds.
Both the H5N1 and the H7N7 virus described in the two papers contain fully avian gene segments and receptors specific for α2,3 galactosidase rather than the human receptor α2,6 galactosidase, both have a multibasic cleavage site, and both are associated with severe disease in poultry flocks. However, in human beings one virus has a mortality ratio of 50% or higher, whereas the other has a ratio of less than 5%, perhaps with a different spectrum of disease including conjunctivitis, an illness not seen with H5N1.
The lack of a specific human receptor is evidently not a barrier to cross-species transmission of influenza A virus but it may contribute to the lack of human-to-human transmissibility. Available data suggest that adaptation of an avian receptor to a more human receptor would not require many mutational events in different subtypes1,2 and that when influenza A viruses do cross into human beings, adaptation is rapid.3 A key feature signalling significant adaptation of avian influenza virus to a human host would be the acquisition of the ability to transmit from human being to human being; this adaptation might simply involve adaptive changes in the receptor properties or might also require improved replication efficiencies, perhaps involving key mutations in the viral polymerase genes, increasing the viral load in the respiratory tract, and altered interaction with the host immune system involving changes in other viral genes. This adaptation might be achieved by sequential mutation in an avian virus genome or by mixing segments of an avian virus with segments from a virus already adapted to human beings.
Study of H5 and H7 subtypes in avian models has given rise to some of our understanding of influenza pathogenesis, notably the importance of a sequence of aminoacids at the junction between two parts of HA1 and HA2, known as the multibasic cleavage site. The correlation between this sequence motif and severe illness and death in birds leads to the inference that, if this sequence motif is present in an influenza virus in mammals, it will also be associated with severe illness. There is some justification for this view, derived from the use of reverse genetic systems in which H5 influenza viruses are engineered to remove the multibasic site with a resulting reduction in pathogenicity in a mouse model4 and from the available data on H5 infection in human beings.5,6 However, the genome of the 1918 influenza H1N1 Spanish influenza, which was associated with over 20 million deaths worldwide, did lack a multibasic cleavage site,7 indicating that different viral subtypes can behave differently in human beings despite common genetic motifs and that we lack information about key mutations associated with mammalian pathogenicity. Differences in clinical outcome could relate to differences between the H7N7 genome and the H5N1 genome, rather than the common sequences. Determining human cytokine responses in H5N1 compared with H3N2 infection may be important in pinpointing key viral-host interactions, as shown, by Peiris and colleagues' paper in this issue and previously.8
The presence of a multibasic cleavage site in the avian model of disease is sufficient to extend viral replication from the gastrointestinal tract of birds to replication in many different organs, which is considered to be the basis of pathogenicity in birds. Yet, in the 1997 H5 human cases and the H5N1 cases described today, there is no evidence of replication of virus outside the respiratory tract, although there may be evidence of multisystem failure.5,6 Peiris and colleagues advance cautious evidence which is consistent with earlier findings by the same group8 and work in the pig model9 which suggests that pathogenesis in human beings may be related to aberrant immune responses which in turn may correlate with the sequences of other non-structural viral genes. Pathogenicity in human beings might be dependent on motifs in several different viral genes.
It is useful to know the spectrum of disease associated with infection. Serum surveys showed no evidence of extensive penetration of H5N1 in 1997 into an exposed human population suggesting very few mild or asymp-tomatic illnesses.10 This situation is in contrast to the situation described for H7N7 viruses in the Netherlands by Koopmans and colleagues, where significant numbers of people were infected but with few cases of respiratory illness or death. Serological data from studies of cases and contacts in the Netherlands will indicate whether there have been further infections associated with mild or subclinical infections. The first description of H7N7 conjunctivitis in human beings occurred in individuals who acquired H7N7 transmitted from seals being swabbed to investigate the cause of fatal illness, emphasising the contribution of the host to outcome of viral infection;11 other human cases have been described since.12 Perhaps the particular combination of H7 HA and N7 NA make this virus subtype particularly adept at targeting human conjunctival cells, rather than other cells in the respiratory tract.
Taken together, the two papers in today's Lancet emphasise that detailed experimental work is required to understand the virus-host cell interactions which give rise to the spectrum of illness when an avian influenza A virus jumps the species barrier into a mammalian host. Observations for one viral subtype might not hold for another. Different mammalian hosts have different diseases with the same virus, indicating the necessity for caution in extrapolating data from animal models.
Careful clinical observations will be essential in future zoonotic infections of influenza A to add to our understanding of disease, but may be overlooked in the maelstrom of ensuring the immediate public-health responses to the threat of avian influenza. Reducing the probability of adaptation of H5 or H7 to human beings by reducing the potential for recombination with viruses adapted to human beings underlies the current control policies of depopulation of infected poultry and intensive surveillance of human influenza cases in south-east Asia, which were successfully applied in the Netherlands in 2003. Influenza A zoonosis remains a rare event, but one which history teaches us we must take seriously to avoid a repeat of the 1918 pandemic of influenza.
I have no conflict of interest to declare.
<!--start simple-tail=-->References
1. Gamblin SJ, Haire LF, Russell RJ, et al. The structure and receptor binding propeorties of the 1918 influenza hemagglutinin. Science 2004;10.1126 /science 1093155.
2. Harvey R, Martin AC, Zambon M, Barclay WS. Restrictions to the adaptation of influenza A virus H5 hemagglutinin to the human host. J Virol 2004; 78: 502-507. MEDLINE | CrossRef
3. Matrosovich M, Tuzikov A, Bovin N, et al. Early alterations of the receptor-binding properties of H1, H2, and H3 avian influenza virus hemagglutinins after their introduction into mammals. J Virol 2000; 74: 8502-8512. MEDLINE | CrossRef
4. Hatta M, Gao P, Halfman P, Kawaoka Y. Molecular basis for high virulence of H5N1 influenza A viruses. Science 2001; 293: 1840-1842. MEDLINE | CrossRef
5. Yuen KY, Chan PK, Peiris M, et al. Clinical features and rapid viral diagnosis of human disease associated with avian influenza A H5N1 virus. Lancet 1998; 351: 467-471. Abstract | Full Text | PDF (174 KB) | MEDLINE | CrossRef
6. To KF, Chan PK, Chan KF, et al. Pathology of fatal human infection associated with avian influenza A H5N1 virus. J Med Virol 2001; 63: 242-246. MEDLINE | CrossRef
7. Taubenberger JK, Reid AH, Krafft AE, Bijwaard KE, Fanning TG. Initial genetic characterisation of the 1918 Spanish influenza virus. Science 1997; 275: 1793-1796. MEDLINE | CrossRef
8. Cheung CY, Poon LLM, Lau ASY, et al. Induction of proinflammatory cytokines in human macrophages by influenza A (H5N1) viruses: a mechanism for the unusual severity of human disease. Lancet 2002; 360: 1831-1837. Abstract | Full Text | PDF (115 KB) | MEDLINE | CrossRef
9. Seo SH, Hoffmann E, Webster RG. Lethal H5N1 influenza viruses escape host antiviral cytokine responses. Nat Med 2002; 8: 950-954. MEDLINE | CrossRef
10. Bridges C, Katz JM, Seto WH, et al. Risk of influenza A (H5N1) infection among health care workers exposed to patients with influenza A (H5N1), Hong Kong. J Infect Dis 2000; 181: 344-348. MEDLINE | CrossRef
11. Webster RG, Geraci J, Petursson G. Conjunctivitis in human beings caused by influenza A virus of seals. N Engl J Med 1981; 304: 911. MEDLINE
12. Kurtz J, Manvell R, Banks J. Avian influenza virus isolated from a woman with conjunctivitis. Lancet 1996; 348: 901-902. Full Text | PDF (52 KB) | MEDLINE | CrossRef
Comment