Influenza virus in wild birds and mammals other than man

Abstract

Influenza virus is an RNA virus that exists as different types and subtypes. Influenza A virus strains are known to cause disease in several bird and mammalian species. Wild birds are believed to constitute the natural reservoir for influenza A virus. Influenza A virus has the ability to change through antigenic drift and recombination allowing for the emergence of new strains and subtype combinations. In man influenza A virus causes yearly seasonal epidemics and, at irregular intervals, pandemic outbreaks have had a devastating impact on mankind. For example, the Spanish influenza pandemic of 1918 is thought to have killed more than 50 million people. Influenza A virus is an important cause of disease in poultry, where virus strains of the H5 and H7 subtypes may change into forms that are highly pathogenic. These virus strains may transmit directly to man and multiple other species. This has been the case in the ongoing outbreak that started in South-east Asia in 2003. All known subtypes of influenza A virus have been isolated from wild birds living in aquatic environments, mainly dabbling ducks. This review focuses on what is known about the pathogenicity and spread of influenza A virus in species other than man, with particular emphasis on the wild bird reservoir.

• influenza A virus
• avian influenza
• wild birds
• ducks
• domestic birds
• mammals

History

The earliest written reports of a disease that might have been caused by influenza virus were made by both Hippocrates and Livy in 412 BC. Over the centuries there have been numerous accounts of epidemics and pandemics (worldwide epidemics caused by a new influenza virus variant which did not circulate before the outbreak) that may have been caused by influenza virus. The Russian flu pandemic of 1889–1892 is the first one for which we have good enough documentation to be sure that it was caused by influenza virus [1]. An overview of recent pandemics of influenza A virus is provided in Table I.

Table I.  A summary of influenza A virus pandemics and the time periods when different subtypes have circulated in the human population during the last centuries [2], [15], [37], [38].

Pandemic name	Time period	Subtype	Estimated mortality (initial outbreak)
Russian flu	1889–1892	H2N2?	6 million
	1898–1900?	H3N2 or H3N8?	0.5 million
Spanish flu	1918–1957	H1N1	>50 million
Asian flu	1957–1968	H2N2	4 million
Hong Kong flu	1968–	H3N2	2 million
Russian (red) flu	1977–	H1N1	?

During the years 1918–1920 one of the most devastating disease outbreaks in world history took place. It would become known as the Spanish flu pandemic and left the world in horror as it caused the death of perhaps as many as 50–100 million people [2]. It beats even the Great Plague in the number of people killed. It is not known where this pandemic started, although British army camps in northern France during the World War I have been suggested [3]. The first documented clinical cases were found in the United States and an alternative theory states that recruits traveling to the war brought the disease over to Europe. The Spanish flu spread from continent to continent and returned in three major waves during the following years with increasing virulence. This horrific event sparked research into the field that eventually led to the discovery of the viral culprit. In 1933 Smith and co-workers discovered a filterable substance that caused influenza-like respiratory disease in humans which could be transmitted between ferrets and rendered them immune to re-infection [4]. Smith and Stuart-Harris were later able to fulfill Koch's postulates by isolating the influenza virus from one of the researcher's throats when he developed influenza-like illness after he had accidentally been sneezed upon by one of the infected ferrets [1]. The results were published in The Lancet in 1936 [5]. By analyzing exhumed remains from of an individual buried in the arctic, where there had been permafrost since the outbreak, and by sampling tissues from victims stored in formalin, Taubenberger and Reid et al. [6–14] were able to recover enough RNA to determine the subtype of the Spanish Influenza pandemic virus as H1N1.

The world experienced two more severe pandemics during the 20th century. Although less devastating than the Spanish flu, they still caused high morbidity and mortality, with death tolls reaching 6 million worldwide [15]. During the years 1957–1958 there was a pandemic named the Asian flu with the H2N2 subtype, and between 1968 and 1970 the H3N2 subtype caused a pandemic known as the Hong Kong flu. After each pandemic the previously circulating strain disappeared for unknown reasons. Between 1977 and 1978 a very mild pandemic mainly affecting young people swept the world as the H1N1 subtype returned; possibly released by mistake during live vaccine trials in the Far East [16]. This strain currently co-circulates in humans with the H3N2 subtype from the Hong Kong pandemic of 1968. An overview of the different subtypes of influenza A virus that have caused pandemics and circulated in the human during the last century is given in Table I. Based on historical patterns, pandemics can be expected to occur on average three to four times each century, but there is still no way to predict when the next pandemic will hit the world. Considering the high population density and modes of travel in today's world, a new pandemic could have devastating consequences.

Influenza A virus not only causes disease in man but also in animals. In 1878, the disease was first identified in animals in Italy by Eduardo Perroncito. He described an initially mild disease in domestic birds that after a while became highly pathogenic, killing virtually all the birds in the area. In 1901 two other Italian scientists, Centanni and Savonuzzi, identified ‘Fowl plague’, as it was then called, to be a viral disease, but it was not until 1955 that influenza virus was identified as the causative agent. Between the years 1959 and 1999, 18 outbreaks of avian influenza with high mortality (HPAI) were reported in domestic poultry around the world. These outbreaks had devastating economic consequences for the affected countries. Millions of raised birds died from the disease or were culled to stop the outbreaks [17]. In recent years the frequency of outbreaks in domestic birds has increased [18].

The first documented outbreak of HPAI in the wild bird population was in 1961, when an outbreak in Common Terns (Sterna hirundo) killed about 1600 birds in South Africa [19]. This outbreak put the focus on wild birds as a possible reservoir for influenza A virus. When screening wild birds in search of Newcastle disease virus (known to be spread by wild birds), during an outbreak of Newcastle disease in poultry in California in 1974, Slemons et al. [20] revealed that low pathogenic influenza A virus (LPAI) could be isolated from wild birds. Further screening soon revealed that species living in aquatic environments such as ducks, gulls, geese, and shorebirds, harbored low pathogenic influenza A virus strains of many different subtypes and probably acted as a reservoir for these strains [21]. It has since been shown that all known influenza strains infecting humans and other mammals probably originates from wild bird strains [6], [22].

Research has shown that low pathogenic influenza A virus strains may, after circulation in poultry populations, sometimes mutate into highly pathogenic influenza virus strains [23]. During an epizootic in Italy between 1999 and 2001, the H7N1 virus, initially of low pathogenicity, mutated within 9 months to a highly pathogenic form [24]. More than 13 million birds died or were destroyed.

Influenza A virus also infects and causes epizootics in mammalian species such as horses, pigs, seals, whales, ferrets, and mink [21], which will be discussed in more detail later in the text.

It was long thought that the disease outbreaks of influenza A virus in poultry were of no concern to humans, even though there had been reports of people suffering from conjunctivitis after being in contact with animals sick with influenza A virus, or when working with highly pathogenic influenza A isolates in the laboratory [25]. Influenza A virus was not considered to be a zoonotic disease agent of any importance until transmission from birds to humans occurred in Hong Kong in 1997. The outbreak in Hong Kong was caused by a highly pathogenic H5N1 strain that caused severe respiratory disease in 18 humans. Six of the infected people died of the disease [26]. The cases of human infection coincided with an epizootic of HPAI in Hong Kong's poultry population, caused by the same strain of influenza A virus. Investigation of the outbreak determined that close contact with live infected poultry was the source of human infection and that the virus had been acquired directly by humans from birds. Rapid destruction of Hong Kong's entire poultry population of 1.5 million birds reduced opportunities for further direct transmission to humans.

However, the world would soon again experience outbreaks of HPAI in poultry transmitting virus to and causing severe disease in humans. In 2003 an outbreak of highly pathogenic influenza A virus of the H7N7 subtype, caused the death of 1 veterinarian and mild illness in 88 other humans in the Netherlands. More than 30 million birds where killed at the cost of several million euros [170].

The worst outbreak of HPAI in modern times is currently plaguing the world. Starting in February 2003, the outbreak has, as of 16 May 2007, led to the death or destruction of more than 100 million birds and has caused verified disease in 206 humans, of which 185 have died [27] in 10 countries. The outbreak of the same H5N1 subtype that caused disease in Hong Kong in 1997 started in South-east Asia and subsequently spread to most parts of Eurasia and several countries in Africa. The virus has also transmitted to species that were previously not known to be susceptible to infection such as domestic cats, leopards, and tigers [28–30]. The H5N1 strains causing the current outbreak are thought to be related to the H5N1 strain that caused the outbreak in Hong Kong in 1997 [31].

Originally, these highly pathogenic strains are believed to have been derived from low pathogenic wild bird influenza A virus strains that after infecting poultry subsequently mutated into highly pathogenic variants. These highly pathogenic strains were then transmitted back into the wild bird population on several occasions. As a result, wild birds not normally seriously affected by influenza virus infection by low pathogenic influenza A virus strains have become ill or died. There are major differences in how highly pathogenic virus strains affect different species. For example, the currently circulating highly pathogenic H5N1 virus caused high mortality in wild birds at Qinghai lake in China in an outbreak that primarily affected Bar-headed Geese (Anser indicus) [32], [33], killing several thousand of this endangered species and highlighting the risk posed to endangered species. In contrast, some duck species have been shown to survive infection by the same strain without showing clinical signs [34]. Therefore ducks may have acted as long-distance disease-transmitting vectors [35] of this virus. After the outbreak at Qinghai Lake, the H5N1 virus was found among wild birds in many countries in Asia, Europe, and Africa, indicated by increased mortality in certain bird species such as swans and raptors. In Sweden, the H5N1 virus caused deaths primarily among Tufted Ducks (Aythia fuligula). Thus, wild birds have been carrying the virus between different geographic areas and as a consequence there have been outbreaks in domestic poultry in many countries. However, domestic poultry trade and poultry handling in conjunction with agricultural practices have probably been responsible for much of the transmission and the persistence of outbreaks [36].

The ongoing outbreak that started in South-east Asia has sparked fears of an imminent pandemic in humans. The possibility that humans, if concurrently infected with human and avian influenza strains, could serve as a ‘mixing vessel’ for the emergence of a novel subtype that has the ability to be easily transmitted from person to person increases as more humans become infected. If this were to happen it would probably mark the start of an influenza pandemic.

Structure and function

Influenza virus is a negative sense single-stranded RNA virus of the Orthomyxoviridae family. The genetic material is stored in 8 segments of viral RNA (7 for type C) coding for 11 different proteins (Table II) [39]. It exists in three types (A, B, and C) classified by antigenic differences in viral nucleoprotein and matrix proteins. Types A and B are epidemiologically important while type C only causes a mild and rarely diagnosed disease. Type B is relatively stable, while type A is highly unstable and has the ability to mutate, recombine, and exchange genetic information. Type C has been found to circulate in humans and pigs while type B has been found in humans and seals. Influenza virus type A (influenza A virus) infects and circulates in many animal species reported elsewhere in this text and is of primary interest to research because of its pandemic potential.

Table II.  Influenza A virus components and their main function.

Genome segment	Proteins coded	Main function
1	PB2	Subunit of RNA polymerase
2	PB1	Subunit of RNA polymerase
	PB1-F2	Induces cell death
3	PA	Subunit of RNA polymerase
4	HA	Viral attachment and membrane fusion
5	NP	Major structural component
6	NA	Release of new virions by preventing aggregation
7	M1	Facilitating migration of viral RNP in cell
	M2	Ion channel involved in uncoating of virus
8	NS1	Post-transcription modulation, interferon antagonist
	NS2 or NEP	Mediates nuclear export of vRNAs

The following text only refers to influenza type A virus. Influenza A virus can be further divided into different subtypes on the basis of antigenic differences of the primary antigenic structure, the hemagglutinin (HA), and the secondary antigenic structure, the neuraminidase (NA). The different subtypes of these antigenic structures do not cross-react serologically and thus immunity to one subtype does not provide protection for any other subtype. The subtype is determined by the particular combination of the two antigenic structures present in a particular influenza A virus. To date, 16 different HA subtypes and 9 different NA subtypes have been found in nature [40] and they can exist in many different combinations. Influenza virus strains are described by the influenza type, the host of origin, the strain number, the year of isolation, and finally the subtype (e.g. A/Mallard/Sweden/105/02 (H7N7)).

The enveloped influenza A virus contains eight gene segments, each with a ribonucleoprotein (RNP) complex made up of the polymerase A (PA), B1 (PB1), and B2 (PB2) proteins. The gene segments are encapsidated by the nucleoprotein (NP). The encapsidated particles are surrounded by a membrane derived from the plasma membrane of the infected cell. The matrix protein (M1) lines the inner surface of the membrane. Three proteins are embedded in the membrane and protrude from the surface: the HA protrudes as a trimer of identical subunits; the NA protrudes as a tetramer of identical subunits and the matrix (M2) is made up like a tetrameric membrane channel [1].

All gene segments code for a single protein except the M, NS, and PB1 genes that code for two different proteins (M1 and M2 and NS1, NS2 – also called nuclear export protein, NEP, and PB1-F2, respectively), generated by RNA splicing. The NS1 protein is not present in the virion but only expressed in the infected cell. PB1-F2 is not expressed in all isolates [41].

Infection of a host cell is initiated when the HA attaches to sialic acids containing glycoprotein and glycolipid receptors on the host cell surface. Human influenza A virus strains preferentially bind sialic acid residues by an α2,6 linkage, while avian and equine virus strains preferentially bind to sialic acids by an α2,3 linkage [42], [43]. The HA is synthesized as a precursor HA₀ that needs to be cleaved by host proteases to create two subunits, HA₁ and HA₂, that are linked only by a single disulfide bond, in order to successfully infect cells [44].

After attaching, the influenza A virus is taken up by the cell via receptor-mediated endocytosis. Catalyzed by the M2 protein channels in the viral envelope, hydrogen ions flow through these channels into the viral endosome dissociating the M1 proteins. This dissociation is necessary for subsequent migration of the viral RNP to the nucleus. Blocking the function of M2 with drugs such as amantadine and rimantadine can inhibit infection.

The lower pH of the endosome induces conformational changes in the HA that facilitates fusion of the viral and host endosome cell membranes releasing viral RNP into the cell cytoplasm. The viral RNP then enters the nucleus where replication transcription is started by the viral RNA polymerase complex [1] [47].

The viral proteins are formed before new viral RNA. New nucleocapsids are formed when newly synthesized viral RNA binds to NP, possibly with the help of the M1 [45]. M1 promotes migration out of the nucleus [46] together with NEP. The NS1 protein down-regulates the host cell's antiviral interferon production by sequestering viral genomic RNA from host cell intracellular receptors [38].

At the cell membrane, the M1 covered nucleocapsids are enclosed by an envelope containing the viral surface glycoproteins [46]. Budding may be facilitated by interactions between the M1, M2, HA, and NA [47]. The NA is involved in the release and spread of mature virions as it prevents aggregation at the surface of the infected cell by enzymatic removal of sialic acids from the virus surface [48]. It may also have a role in the initial process of viral entry [49]. The use of neuraminidase inhibitors such as oseltamivir and zanamivir inhibits the release of new virions, causing them to aggregate on the cell surface and thereby preventing the infection of new cells.

Genetic variability

Influenza A virus has been called a master of metamorphosis [50]. It uses two main ways to elude host defenses: antigenic drift and antigenic shift.

Antigenic drift is characterized by slight changes in the antigenic structures of the virus that renders neutralizing antibodies, produced by the host immune system during the encounter of previous strains by infection or vaccination, obsolete. These antigenic changes take place frequently as influenza A virus lacks mechanisms for the ‘proofreading’ and repair of errors that occur during replication [51]. These errors result in variant strains that may become dominant and replace older strains, if they are better at evading the host immune response. Dysfunctional strains will die out but many variants will not, and several subpopulations may co-exist in an infected individual [50], [52]. Antigenic drift has been shown to take place in humans and other mammals at rapid rates and is the reason why influenza vaccination with updated strain content must be given yearly to provide protection from disease. Phylogenetic analyses have revealed that antigenic drift of influenza A virus does not take place to the same extent in wild birds. The strains that exist in wild birds appear to be in a relative evolutionary stasis, as shown by the fact that strains collected from wild birds more than 80 years apart were almost identical [21], [53]. This might imply that an optimal host–parasite relation has been reached and that any change in the genetic make up will result in less functional strains that will eventually become extinct.

The segmented genome of the influenza A virus allows for another potent way of evading host defenses. If a single cell is infected by two influenza A virus strains of different subtypes, genetic reassortment can take place. The reassortment can result in the creation of a novel combination of HA and NA and thus a subtype different from both the parental virus strains. This is known as antigenic shift. If an HA or NA subtype that has previously not been circulating in a population is created, for which there is no herd immunity, a pandemic spread might be the result. Antigenic shift that gives rise to a new subtype infecting humans has long been thought to require the involvement of pigs, since pigs are susceptible to infection with both avian and mammalian virus strains. They could thus serve as a ‘mixing vessel’ for the reassortment of genetic material from human and avian virus strains, resulting in the emergence of a novel subtype that has retained its ability to transmit easily between humans but includes a new antigenic subtype of the HA or NA structures. More recent research has shown that pigs may not be the only possible mixing vessel, since recent events have shown that for at least some influenza A virus subtypes circulating in bird populations, humans themselves could serve as the mixing vessel [50].

Host specificity

There are many factors determining whether a species can act as a host for an infection – the most obvious being sufficient contact between the host and the pathogen for infection to occur. This makes some species more likely to become infected than others. It is illustrated by the fact that since transmission is thought to mainly occur via water, aquatic bird species (species that live in an aquatic environment) are found to be infected by low pathogenic influenza A virus strains more often than non-aquatic species [54].

Successful attachment to a host cell is the next required step for infection to occur. As mentioned earlier, avian influenza A virus strains preferentially use α2,3-linked sialic acids as cellular receptor while human influenza A virus strains preferentially bind to α2,6-linked sialic acids. Before reaching and binding to the epithelial cells, however, there are other host barriers such as mucus and alveolar macrophages to pass [55]. For example, in the secretions protecting the eyes and respiratory tract, different mucins containing sialic acids are present that specifically bind and clear virus before they reach the epithelial cells. These mucins express different sialic acid linkages in different species and also in different organ systems of the same species. Humans are better at clearing avian influenza A virus from the respiratory tract than from the eye, since the mucins of the respiratory tracts are rich in α2,3-linked sialic acids while the secretions of the eye are rich in α2,6 sialic acid linkages. The situation is reversed in chimpanzees, since their respiratory tract secretions are rich in α2,6-linked sialic acids that make them partly resistant to infection by human influenza A virus [56]. At the cellular receptor binding level there are major differences in sialic acid linkage content between species, as well as differences within organ systems, and even between cells of the same organ systems. Taken together these differences may determine where, if at all, infection of the host may occur. Ducks that express α2,3-linked sialic acids in the intestinal system are primarily affected by infection of the cells lining the intestinal tract [57]. Within the human body α2,3-linked sialic acids have been found to be predominant in the eye. In the respiratory system, α2,6 sialic acid linkages are predominant in the upper part, whilst α2,3 linkages are present in the lower part, where influenza A virus has been shown to bind preferentially to pneumocytes type II [58], [59]. Although not predominant, α2,3 receptors are also found on ciliated cells of the upper respiratory tract [60]. Human influenza virus targets non-ciliated cells that express α2,6-linked sialic acids [60]. This might explain why there is only limited transmission of avian influenza virus to humans, why conjunctivitis has been a common symptom [61], and why the respiratory infections in humans are rare but severe when they occur [62].

It was previously thought that pigs (known to express both α2,3- and α2,6-linked sialic acids in their respiratory epithelium) were unique in their potential to act as a mixing vessel host species, where pandemic virus strains could arise by the recombination of avian and human influenza virus strains infecting the same cell [57], [63]. However, the finding that both humans [60] and chickens [64] harbor the different receptor types in different cells indicates that theoretically this could happen in other animal hosts as well. Further research has shown that although avian influenza A virus strains preferentially bind α2,3-linked sialic acids, a further refinement of specificity exists that differs between avian species. The refinement is based on recognition of differences in the inner part of the oligosaccharide receptor [65].

Successful attachment to a cell does not necessarily imply that infection can occur, since the virus must also be able to enter the cell and cause it to replicate genetic material. In this process the internal genes of the virus are the determinants. It has been shown that there are host-specific lineages of all the different internal genes indicating species adaptation and optimization of each gene [66]. Some of these differences have been analyzed in detail and found to be important. For example, the PB2 gene of the virus polymerase complex plays a major role. Research has shown that in avian influenza A virus the amino acid residue 627 of the PB2 protein differs from mammalian virus strains in that avian virus strains have glutamic acid at this site whereas mammalian strains have a lysine and that this is of major importance for host range restriction [67]. This difference has been associated with optimal replication at different temperatures. Human influenza virus strains replicate in an environment of about 33°C in the trachea, while avian strains are adapted to replication in the intestinal tracts of birds at a temperature close to 41°C [68]. It has also been shown experimentally that a change from glutamic acid to lysine at this site results in increased virulence for mice [69] and this change has been seen in H5N1 virus strains and H7N7 virus strains that have caused severe disease in humans [61], [69].

Once successful replication has taken place, the newly constructed virus must be released from the surface of the infected cell to invade new cells and, like the hemagglutinin, the neuraminidase of avian virus strains preferentially operates by cleaving the sialic acids that are α2,3 -linked, while human neuraminidases prefer α2,6-linked sialic acids [66].

Even if replication and release of new virions has been successful, there are still factors that determine whether or not the infection will remain localized in the organ of entry and if it will prevail. First of all, the immunity of the host must be dealt with. In order to hinder the host's innate immunity to produce interferons that put infected cells into an antiviral state, the NS1 polypeptide of the influenza A virus sequesters double-stranded RNA away from certain protein kinases and suppresses host cell post-transcriptional processing of mRNA so that the infected cell remains undetected [38], [55], [66], [70]. Research has shown that there are host-specific differences in the NS1 gene of different strains. For example, the human NS1 gene is not optimal when introduced into mouse strains [16].

The factors determining the ability of influenza A virus to produce systemic rather than localized infection in mammals are not fully understood [55]. In poultry, however, the ability of the virus to invade other organs depends on whether or not the HA of the virus can be cleaved by ubiquitous extracellular proteases or only by specific proteases that are restricted to the respiratory and gastrointestinal tract. This ability is determined by the amino acid sequence at the cleavage site. Poultry strains can thus be characterized as highly pathogenic or low pathogenic based on the presence of multiple basic amino acids at the cleavage site. These highly pathogenic influenza A virus strains are thought to arise from low pathogenic virus strains. How this occurs is not well understood [71], [72].

Persistence and modes of transmission

The influenza A virus, like other viruses, cannot replicate outside a host cell. In order to infect new individuals it needs to persist for some time in the environment. It seems that influenza A virus is well adapted to persist in water. Under experimental conditions, avian influenza A virus strains stored in distilled water at 28°C could remain infective for 100 days, at 17°C for 200 days, and possibly for as long as 1000 days at 4°C [73]. However, under natural conditions, persistence of active virus is limited by the effects of pH, salinity, UV radiation, and the presence of biologically active material such as degrading enzymes, bacteria, and other microorganisms.

Human influenza A virus strains are stable at a pH from neutral to 8.5. Infectivity decreases rapidly below pH 6.0. Avian influenza A virus strains exhibit more stability than human influenza A virus strains and can persist and remain active at pH 4.0, whereas human isolates do not persist at pH below 5.0 [74]. Infectivity is inversely related to salt content of water for avian influenza A virus [75].

In open air, human strains of influenza A virus can spread via aerosol. The persistence and infectivity of these strains in open air is promoted by low humidity. Aerosols containing influenza A virus may remain infective for up to 24 hours or more at low humidity but only for an hour at high humidity [76], [77]. Other limiting factors in open air are UV radiation and wind. Some strains of influenza A virus can also be spread via fomites on hard surfaces such as stainless steel, where it can survive for up to 2 days [78].

Thus, many factors determine the suitability of different environments for persistence, infectivity, and transmission of influenza A virus.

Influenza A virus in man introduced by other species

The influenza A virus strains that currently circulate in humans are all believed to have wild bird progenitors. However, they may have been introduced via other species such as pigs. Experimentally, avian influenza virus strains from wild birds do not replicate well in humans [79] and human strains do not replicate well in waterfowl [80]. Until the outbreak in Hong Kong in 1997, the occurrence of transmission of avian strains was believed to be a rare event, only causing conjunctivitis in the few affected cases [81]. However, a serological survey in rural China suggests that infection with avian subtypes has not been uncommon in people who have had close contact with domestic ducks and poultry [82]. In recent years, outbreaks of highly pathogenic strains that have evolved in poultry have occurred rather frequently. The symptoms of disease and the disease pattern in humans have been variable depending on the strain, ranging from mild symptoms of conjunctivitis or influenza-like symptoms to the severe respiratory disease caused in most cases of the current H5N1 outbreak. So far, the development of a highly pathogenic strain in domestic birds has been a prerequisite for human infection, but there is increasing evidence that direct infection may occur [59]. Low pathogenic H9N2 virus has been isolated in two children [83] with mild influenza symptoms. H9N2 virus strains are considered to be even more likely than H5N1 to become the cause of a pandemic, since the strains that circulate in domestic chicken and ducks worldwide [84] have already acquired receptor specificity to prefer α2,6-linked sialic acids, which are found on human cells [85].

Influenza A virus in mammals other than humans

Influenza A virus infects and on some occasions creates stable lineages in several mammalian species. This has been shown both experimentally [86] and in nature as described for the species below. Highly pathogenic virus strains, such as the currently circulating H5N1 virus that originates from South-east Asia, have been found to infect many species that had previously not been considered vulnerable. Thus, the range of species infectivity is heavily dependent on strain type.

Pigs

Influenza A virus is a common cause of respiratory disease in pigs. Pigs harbor pig-adapted strains such as the classic swine-like influenza virus H1N1, but are also susceptible to human and avian influenza virus strains [21]. This is partly due to the presence of both α2,3- and α2,6-linked sialic acids in the respiratory epithelium of pigs as discussed previously [57]. Avian virus strains have been found in pigs on a couple of occasions. For example, an H4N6 strain caused respiratory disease in pigs in Canada [87]. In further Canadian porcine surveys avian H1N1 and H3N3 were also isolated [88]. Avian H1N1 has been isolated in China [89], where it caused a severe outbreak in pigs in 1979–1980 [90] and has remained in the pig population since that time. Experimentally, even those avian strains that did not at first replicate in pigs could be made to replicate after reassortment with swine-like virus strains in pigs that were co-infected [91]. Several studies have reported human H3N2 strains in pigs after the antigenic shift in the human population in 1968 [22].

Studies have also shown that both avian-like and swine-like H1N1 strains circulate at the same time in pigs [92], as well as human-like H3N2 and avian-like H9N2 [93]. The different strains allow for reassortment. Such reassortment does take place in pigs, where it has been shown to create variants such as H1N7 and H1N2 [94], [95]. Reassortants between human-like H3N2 and avian-like HIN1 also occur [96].

Humans may be affected by strains transmitted by pigs. Direct transmission of swine-like H1N1 to humans occurs and has been fatal [97]. Reassortants between swine-like and human-like influenza A virus strains may also cause disease in humans. In 1992 such a reassortant was isolated from sick children in the Netherlands [98]. Fortunately, this virus was unable to spread between humans. These examples show that pigs are indeed a mixing vessel for human and pig virus strains as suggested [63], albeit possibly not the only one.

Horses

Influenza A virus strains in horses are thought to be of avian origin. Different subtypes have been found to infect horses and antigenic drift creates distinct lineages within the subtypes. At least two subtypes have created stable lineages: H7N7, H3N8 [99–102]. Some strains have been suggested to be recent introductions from wild birds [103].

Canines

Historically, canine species have not been found to carry influenza virus. However, in 2004 an outbreak in racing greyhounds was found to be caused by the H3N8 influenza subtype and is thought to be an equine influenza A virus variant that had adapted to spread in canines [104]. During the current outbreak in South-east Asia a recent investigation has isolated H5N1 influenza virus from dogs and has also found that antibodies to H5N1 are common in Thai dogs, suggesting that they have previously been infected [105].

Felines

Feline species were not considered particularly susceptible to influenza virus before the current outbreak of avian influenza H5N1 that started in South-east Asia in 2003. However, after captive tigers and leopards [28], [106] became ill and died after being fed infected chicken carcasses several investigations were performed. It was shown that there was not only direct transmission from the contaminated food but also probable transmission between tigers [29]. Experimental infection of domestic cats has shown that cats infected with the H5N1 highly pathogenic strain develop lethal systemic infection and excrete virus in both the respiratory and digestive tract secretions. The cats in the experiment could also infect each other [30], [107]. In Europe, cats have also been found to be infected by the H5N1 virus in areas where there have been outbreaks in wild birds [108].

Mink and ferrets

Mink and ferrets have been found to be susceptible to influenza A virus in experiments [109]. Mink have also been found to be naturally infected by avian influenza A virus of the subtype H10N4 during an outbreak in farmed mink in Blekinge, Sweden [110]. Further investigation revealed that the virus strain causing the outbreak in the Swedish mink was most likely of wild bird origin. Although the virus was very similar to avian virus strains, it was adapted to spread between mink [110–115].

Seals and whales

In the winter of 1979–1980 seals off the coast of eastern USA died of hemorrhagic pneumonia. The cause of disease was found to be influenza A virus of the subtype H7N7. The virus contained avian-like genes, but biologically behaved as a mammalian strain [116], [117]. During the autopsies and handling of experimentally infected seals, people handling the animals developed conjunctivitis. Influenza virus could be isolated from eye swabs of the affected peoples’ eyes [118]. In a subsequent outbreak among seals during the season 1982–1983, another even more avian-like virus was recovered from seals suffering from pneumonia. This virus belonged to the H4N5 subtype [119]. Further surveys of seals in the area have also found H3N3 virus strains to be present in seals [120].

Whales have been found to be infected with strains closely related to the H13 strains of gulls [121], [122].

Investigations of the receptor specificity of the sialic acids in whale and seal lungs showed the presence of α2,3-linked sialic acids [123] and only a weak association with α2,6-linked sialic acids [124]. This might explain why these seals and whales are susceptible to infection by avian virus strains. However, seals have also been shown to be infected by influenza B virus of human origin [125] and have been found to have antibodies against human influenza A virus of the H3 subtype [126], proving that they are susceptible to influenza A virus strains of mammalian origin.

Influenza A virus in domestic birds

Influenza A virus causes a wide spectrum of symptoms in birds, from mild illness to a highly contagious and fatal disease resulting in severe epidemics. The latter is known as highly pathogenic avian influenza (HPAI). This form is characterized by severe illness, rapid death, and a mortality in the affected populations that approaches 100% within 72 hours. Many different species of reared birds including chickens, turkeys, quail, and ostriches are susceptible to epidemics of rapidly fatal influenza [127].

The main difference between infection with highly pathogenic virus strains and low pathogenic virus strains is systemic vs localized infection [128]. The hemagglutinin of strains causing HPAI can be cleaved by ubiquitous proteases and is thus not restricted to cells in the respiratory tract. All outbreaks of the highly pathogenic form have been caused by the subtypes H5 or H7. Influenza A virus of these subtypes may enter into domestic bird populations as low pathogenic strains that only cause mild disease. From these low pathogenic strains highly pathogenic strains then arise by mutation. This is probably due to the extremely high propagation rates in dense flocks. Birds that have already been infected and survived infection with the low pathogenic influenza A virus strain have protection against infection with the highly pathogenic variant [129]. Several mutations may add to the pathogenicity of strains causing HPAI but the accumulation of basic amino acids at the cleavage site is diagnostic for disease outbreaks.

Direct or indirect contact of domestic flocks with wild migratory waterfowl has been implicated as a cause of epizootics [21]. Spread between farms during an outbreak is most likely caused by the movement of people and the transport of goods [36]. Outbreaks of HPAI are often difficult to control since the virus can persist and remain active for some time in the environment and because it is highly transmissible. In areas with dense poultry populations and/or limited resources for surveillance and control, outbreaks have lasted for years. This has been the case in Mexico (1992–1995), Italy (1999–2000), and with the ongoing outbreaks in South-east Asia and Africa.

Influenza A virus in wild birds

The wild bird reservoir

To date the search for a reservoir of influenza A virus in nature has pointed towards wild birds as being pivotal in the transmission and persistence of the virus. It is thought that all influenza virus strains infecting mammalian species originate from wild birds [21]. The evidence for the existence of a wild bird reservoir is strong. Influenza A virus has been isolated from birds on all continents except Antarctica from where there is only serological evidence [130–132]. However, most studies have taken place in developed countries and the situation in Africa, South America, and parts of Asia is much less explored. All 16 HA and 9 NA subtypes in the majority of possible combinations have been isolated from avian species [133]. Most influenza virus strains found in wild birds have been defined as low pathogenic. Although low pathogenic influenza A virus strains have been isolated from more than 105 species from 26 different families of birds [54], [134], almost all isolates come from the families Anseriformes and to a lesser extent Charadriformes and Laridae. These families include birds such as ducks, geese, swans, waders, and gulls, which although they are different species evolutionarily speaking, share the trait of being adapted to life in an aquatic environment. In contrast, isolations of low pathogenic virus strains from pure land-dwelling birds are rare.

Indirect evidence that wild birds constitute a reservoir for influenza A virus comes from studies on viral evolution. These studies have shown that influenza A virus strains in wild ducks show only limited evolution over time. It has therefore been suggested that influenza A virus exists in an evolutionary stasis in the reservoir species [21], [135]. In support of this suggestion, analysis of strains recovered from wild ducks that had been preserved in museums since the early 20th century showed almost no antigenic drift when compared to modern avian strains [53]. This situation is very different from the situation when influenza A virus is introduced into mammals. Avian strains that are introduced into a new host species are evolving at high rates [136]. The low rate of viral evolution in ducks suggests that adaptation to the host has reached an optimum. New variants have no survival advantage and are thus not successfully sustained. It has also been found that co-infections of different influenza virus strains are detected less frequently in ducks than in other species, suggesting that host-adapted strains prevent co-infection by other strains [137]. Considering that the Anseriformes species have existed for millions of years, influenza virus has had ample time for this adaptation to take place [82].

There are many reasons why species connected to aquatic environments are good reservoir species for the virus. These are elaborated on in more detail later in the text, but the implication for the wild bird reservoir is mentioned here.

The populations of Anatidae, Charadriformes, and Laridae species in the world are large, the Mallard (Anas platyrhynchos) population alone is estimated to comprise 27 million birds [138]. Most populations present in different areas are also connected as these birds travel over large distances and congregate in large numbers at specific important staging sites during migrations [139]. Wild duck populations could therefore be hypothesized to support the perpetuation of even short-lived infections such as influenza, as there are enough susceptibles at any given time. The minimum population to support the perpetuation of the measles virus has been estimated to be 500 000 humans [140], and although the attack rate of influenza A virus in ducks might be lower, the population size of ducks is probably still large enough. Also, certain aspects of duck demography and ecology might make the minimum population needed even lower. First of all, the populations of ducks have high turn-over rates. In Mallards about one-third of the population is replaced each year, implying that this proportion is immunologically naïve [141]. Further, as ducks infected with low pathogenic influenza A virus strains do not appear to be severely affected by the disease, the infection does not seem to limit an infected bird's interaction with other birds or the environment. Nor does infection with low pathogenic strains seem to limit the capability for long-distance flights that could transport the virus long distances to new susceptible flocks. Since virus is shed by infected birds in high quantities into an environment where it can survive for an extended time, the feeding and social behavior of the reservoir species makes it likely that susceptible individuals are exposed to the pathogen.

Wild birds living in aquatic environments are thus a reservoir for influenza A virus in nature. Questions that remain to be answered are to what extent different species of birds are involved and indeed if wild birds are the only reservoir in nature?

Geographic constraints

Even though influenza A virus in the wild bird reservoir has been said to be in ‘evolutionary stasis’, there have been slight changes over time. Different genetic lineages of influenza A virus have evolved in bird populations that are separated by oceans. There are only a limited number of wild bird species that migrate across the Atlantic or the Pacific Ocean, resulting in limited interaction between the populations of Eurasia and the Americas. Avian influenza A virus strains of the Americas can thus be separated from those in the rest of the world [142–144]. There are also inter-regional differences in the prevalence of subtypes in different duck populations that use different migratory routes within continents [145], again probably as a result of limited interaction with the other populations. Lately, however, limited transmission of genes between the North American and Eurasian populations have been reported [143], [146–148], indicating that the interaction that does take place is sufficient for exchange to occur. Knowledge on the interaction and spread of pathogens between different fly-ways and continents may be crucial in estimating the spread of highly pathogenic variants between different areas of the world. Therefore, increased knowledge on this topic is urgently needed.

Species preference

The species preference of influenza A virus is likely to be determined by the mode of transmission, i.e. by the fecal–oral route via contaminated water. Species that feed in shallow calm waters, where influenza A virus is found in the highest concentrations, run the highest risk of becoming infected. In line with this argument species like ducks, geese, waders, and gulls have the highest prevalence of influenza A virus [54]. Scavenger species such as raptors that may feed on diseased birds are also likely to become infected, but they are unlikely to take part in efficient transmission as they do not dwell in water.

Although almost all possible subtype combinations have been found in the wild bird reservoir, some subtypes of HA have only been isolated from certain species. The H13 and H16 subtypes have almost exclusively been recovered from gulls [40], [149]. In addition, isolates from shorebirds and gulls do not always replicate in ducks. This indicates that there are species preferential lineages within the wild bird reservoir [150], although this species specificity may not be relevant for all genes of the influenza A virus [151].

Modes of transmission

Wild ducks readily get infected with avian strains of influenza A virus through the intake of contaminated food and water. The virus still remains active after passage through the low pH of the duck gizzard and produces infection in the cells lining the intestinal tract [82]. The infection may also affect cells in the lungs of ducks. During the period of infection, large amounts of virus of up to 10⁸ EID50 are shed in the duck feces for about 7 days [74] and sometimes for as long as 21 days [152]. Progeny virus is to a lesser extent also shed from the trachea. The fact that infected birds shed high amounts of influenza A virus via feces implies that birds living in aquatic environments will contaminate the water where they live. Influenza A virus of different subtypes has been isolated in concentrations of up to 10^2,8 EID50/ml of water from unconcentrated water in lakes where wild ducks congregate [153], [154]. Since the virus remains viable for some time in water, it permits transmission to other birds in the area that ingest contaminated water. Influenza A virus strains were even recovered from lake water for about a month after the birds in the lake had migrated south for the winter, indicating that lakes may be a source of infection for other birds for a long time. These data indicate that fecal–oral transmission via water is the most likely and efficient mode of transmission for low pathogenic strains in wild birds.

Highly pathogenic strains may transmit in other ways. Unpublished data from recent research presented by Fouchier at the FAO/OIE conference on avian influenza and wild birds, held in Rome on 30–31 May 2006, show that the highly pathogenic H5N1 strain causing the ongoing outbreak that started in South-east Asia is primarily excreted from the respiratory tract of infected birds.

Clinical picture and immunity

All birds are thought to be susceptible to infection with influenza A virus, although some species are more resistant to infection than others. Depending on the subtype, strain, host species, and individual, the disease caused by infection with influenza virus may range from non-pathogenic to lethal [155]. Even within the duck family there are many different species of wild and domestic ducks and they may also show different responses to infection with influenza A virus [128].

Infections by low pathogenic strains in ducks have traditionally been considered non-harmful, as there are no evident clinical signs of disease. This notion may not hold as more investigations into the effects of infection by these strains are studied. Clinical symptoms may be hard to detect. It is difficult to evaluate if the birds are completely free of symptoms or actually sick in a subtle way. The latter suggestion is supported by experiments in infected ducks that showed histopathologic signs of mild pneumonia even though no clinical signs of disease were evident [156].

Highly pathogenic strains behave differently and strains that are highly pathogenic for chicken may cause milder disease and different signs of disease in other species [157]. They may even show no signs of disease in some more resistant species of ducks [156], [158], [159] and gulls [160]. Due to the lack of evident effects on the birds’ health status, ducks and gulls may act as carriers of some highly pathogenic strains. If these strains persist in the wild bird reservoir, the eradication of such strains from domestic bird populations will be difficult due to the risk of re-introduction from wild birds.

Experimental infections of ducks have revealed that ducks only produce a short-lived low-level humoral immune response and a weak cell-mediated immune response as seen by suppressed T-cell function and enhanced macrophage phagocytic activity [155]. Studies have also shown that ducks may be re-infected with the same strain after only 2 months [152], indicating that the protection of acquired immunity is poor. However, in large-scale studies, juvenile ducks were found to be infected with influenza A virus more frequently than adult birds [21], indicative of some sort of acquired immunity or improved immune response. Ducks are readily co-infected by different strains and subtypes, paving the way for reassortment between different influenza A virus strains and creation of new variants [137].

Other bird species such as chicken, pheasant, turkey, and quail do show a humoral response with IgM and IgY production [128]. In poultry, previous infection with a low pathogenic strain provides protection from disease caused if the low pathogenic strain subsequently evolves into a highly pathogenic strain [129].

Enzootic cycle

A clear picture of the enzootic cycle of influenza A virus in wild birds does not exist. More research is needed to elucidate the different stages involved in the interaction between the virus, the host, and the environment. It seems likely, however, that nestlings and juvenile birds are exposed to and probably contract several infections with different influenza A virus strains present in the surrounding environment early in their life. During subsequent migration the birds may encounter even more strains as they mingle with other flocks of birds and are exposed to virus-contaminated environments.

As prevalence of influenza A virus has been shown to be higher in juvenile than in adult birds; the input of juvenile and thus immunologically naïve birds is most likely of key importance for upholding the number of susceptible birds in the population. As mentioned earlier, ducks and the other reservoir species have a high input of juveniles each year and few birds get old, ensuring a high number of susceptible individuals.

Another key issue of the enzootic cycle is the persistence of the virus in different environments as the infection primarily passes on from one bird to another through fecal contamination of the environment and the subsequent ingestion of infective material. The interaction between different host species and the environment must therefore be of crucial importance in the enzootic cycle, in addition to the interaction between different bird species and different flocks of the same species. These determinants are likely to be important for the distribution of virus strains around the globe and deserve more attention.

One of the enigmas of influenza A virus ecology is how so many subtypes can circulate in the wild bird populations and persist from year to year, when some of these subtypes are isolated rarely and since the prevalence in the studied bird populations differs greatly between studied species, place, and time of year. Studies in North America have shown that the prevalence in some species of ducks is highest in the fall, when up to 30% of the ducks may be infected [21]. These numbers dwindle to levels close to or below 1% during winter [161] (although, more recent studies suggest that this may not be true for all areas of North America [162]) and remain low during spring and summer. In other North American studies a reversed prevalence picture has been found in shorebirds that have high prevalence during the spring and low prevalence during the fall [150], [163]. While some subtypes are frequently isolated, others have been isolated only rarely in specific places or in specific species, such as the H13 and H16 subtypes, which have almost exclusively been isolated from gulls. A lot of these variations and peculiarities are likely to be explained in the future when more data on the prevalence of different subtypes in different species and at different locations are available, but to date, the explanations remain elusive. Some suggestions have, however, been put forward as to how different influenza A virus subtypes and strains are perpetuated.

The first one suggests that influenza A virus is continuously circulating within the reservoir species. This requires that a sufficiently large number of individuals in the wild bird population are infected at any given time to pass on the specific strain or subtype to new individuals. Differences in prevalence of different subtypes may be explained by the immune status of different populations of birds. When herd immunity in a population of birds falls the birds will once again be susceptible to infection and the prevalence of the particular subtype causing infection will rise again [21].

A second suggestion claims that influenza A virus may survive and remain active for extended periods in the environment. The virus may, for example, persist frozen in lakes on the birds’ breeding grounds and be passed on to susceptible birds when they return from wintering in other areas. In support of this proposition, influenza A virus has been isolated from lake water several months after the birds in the lake had migrated south for the winter [154]. However, virus has not been isolated from lakes in the spring before the migratory birds return.

A third suggestion has been that different species that have high prevalence of influenza A virus during different seasons of the year and share the same habitat may interact and pass on the infection to each other [150]. It has been proposed that ducks and waders interact in such a way since it has been found in North American studies that the prevalence of influenza A virus infection in shorebirds is high during the spring and low in the fall, while in ducks it is high during the fall and low in the spring.

More research is needed to determine which of these theories, or a combination of them, holds true, but studies performed in Eurasia support the first theory. In Eurasia the perpetuation of influenza A virus in migrating dabbling ducks might be accomplished by the ducks themselves based on findings of influenza A virus prevalence of up to 9.5% in some spring months in Sweden [164], [165]. It is also supported by the 8% influenza A virus prevalence at breeding grounds in eastern Siberia [166] and a 4.1% influenza A virus prevalence among wintering mallards in Italy [167], [168]. Studies on migrating birds in Northern Europe have shown, in accordance with the North American data, that the prevalence is highest during the fall. This is likely due to the influx of juvenile immunologically naïve birds. However, the season in Northern Europe starts later, in September and last longer, until December, in comparison with the North American findings where the prevalence is high in August and September.

Spread by wild birds

Most low pathogenic influenza A virus strains do not seem to hinder birds from migrating. Thus, these virus strains may be carried over large distances by the birds either through non-stop long-distance flights or in a relay pattern where one bird carries the virus a short distance and another carries it further. Until the present outbreak of H5N1 that started in Asia it was not believed that wild birds could be infected with highly pathogenic virus strains and still perform long-distance migrations. As some species of ducks have been shown to be resistant to these strains this belief has had to be reviewed. The sudden outbreaks in wild birds in China and the subsequent appearance of H5N1 in wild birds in Russia, the Middle East, Europe, and Africa suggest that the transport of highly pathogenic strains by wild birds have occurred.

Primary introduction of influenza A virus into poultry and domestic animal holdings are likely due to fecal contamination by wild birds either directly by contamination of the holdings or indirectly through contaminated water supplies or feed. In Europe Mallards have been shown to carry influenza A virus strains that are almost identical to the highly pathogenic strains that have caused outbreaks in European poultry during the last 7 years [18]. In fact, the H5 and H7 subtypes that are prone to develop into highly pathogenic strains were commonly found in fecal samples from Mallards during a large survey in northern Europe [165]. Other studies have shown that holdings where wild birds and domestic birds share the same habitat due to agricultural practices are at the highest risk for outbreaks [36], suggesting that wild bird transmission is the most common route. Transmission to new holdings in an area where an outbreak has occurred may well be the result of spread by the movement of contaminated goods, animals, and people.

Further research needs to clarify to what extent wild birds are responsible for the transmission of influenza A virus strains around the globe and into domestic bird populations in relation to the smuggling and transport of infected animals and goods. By studying these underlying mechanisms accurate bio-security measures can be taken to hinder such transmission.

Concluding remarks

An increased understanding of the prevalence and transmission of influenza a viruses in different species may not only be valuable for the common goal of increasing the knowledge of pathogens in the world but is also of importance for the risk assessments, risk management, and vaccine development needed to limit spread of influenza virus strains to humans and domestic species, thereby averting another influenza pandemic. However, as we cannot rid the world of influenza virus due to its diverse reservoir in wild birds, the threat of a new pandemic will loom over us well into the future and outbreaks in poultry will continue. Although influenza virus strains are present in the wild bird populations and may transmit and cause outbreaks in other animals, the current situation of worldwide spread of a highly pathogenic virus is a man-made problem. The large populations of poultry needed to feed the human population have created an environment where different viruses have an unprecedented opportunity to evolve. In order to halt this development, vast changes in husbandry practices as well as the slaughtering and distribution processes are needed to minimize the contact between wild and domestic birds and the contact between sick birds and humans [36], [169]. These changes will have severe cultural and economic implications for many countries and will not take place unless richer countries assist the poorer ones. An alternative solution to protect us from new pandemics would be the development of a vaccine that can provide protection for any type of influenza virus. This has so far proven to be difficult to develop.

Abbreviations

Table

EID50	egg infectious dose 50
HA	hemagglutinin
HPAI	highly pathogenic avian influenza
IAV	influenza A virus
LPAI	low pathogenic avian influenza
M1	matrix protein
M2	membrane ion channel protein
NA	neuraminidase
NP	nucleoprotein
NS	non-structural protein
PA	polymerase A
PB1	polymerase B1
PB2	polymerase B2
RT-PCR	reverse transcriptase polymerase chain reaction
vRNP	viral ribonucleoprotein

Species index

Table

English name	Scientific name
Arctic Tern	Sterna paradisaea
Barnacle Goose	Branta leucopsis
Black-headed Gull	Larus ridibundus
Brent Goose	Branta bernicla
Gadwall	Anas strepera
Common Eider	Somateria mollissima
Common Pochard	Aythya farina
Common Shelduck	Tadorna tadorna
Common Tern	Sterna hirundo
Eurasian Teal	Anas crecca
Eurasian Wigeon	Anas penelope
Goldeneye	Bucephala clangula
Greylag Goose	Anser anser
Guillemot	Uria aalge
Long-tailed Duck	Clangula hyemalis
Mallard	Anas platyrhyncos
Mute Swan	Cygnus olor
Northern Pintail	Anas acuta
Red-breasted Merganser	Mergus serrator
Tufted Duck	Aythya fuligula