One would have to be very removed from access to the media not to be aware of the 2009 outbreak of swine flu (influenza). Although it presently appears that this outbreak is barely (if at all) more dangerous than are the seasonal outbreaks of ‘traditional’ human flu (that may yet change), it raises questions about both our preparedness against pandemics (a word that implies only a widespread distribution; of itself it says nothing about virulence) and our understanding of the means by which such strains arise, transmit, infect and display virulence in humans.

The term ‘zoonotic infection’ (‘zoonosis’) describes any disease that is capable of passing from animals to humans. The WHO definition at Defra implies that the animal should be a vertebrate. Zoonoses are exceedingly common, and their causative transmissible agents include viral, bacterial, fungal, parasitic and other organisms. However, our attention is focused on influenza Type A strains. These are RNA viruses of the familiy orthomyxoviridae and are typically characterized by their surface antigens from the haemagglutinin (H) and neuraminidase (N) molecules that they contain. The bird flu strain that has achieved recent notoriety is H5N1, and reminds us to be prepared, while the 2009 pandemic of swine flu is H1N1. Strains that are more or less avirulent in one host may be highly virulent in another. Thus there is the opportunity for a non-human species such as a bird or a pig to act as a carrier (by being infected but showing mild or even no symptoms, much as the infamous Typhoid Mary). The number of amino acid changes necessary to turn a harmless strain into one with considerable virulence and transmissibility can be rather small. (This mirrors the situation in polio, where the live, oral (Sabin) virus vaccines are rather similar in sequence to the original infective strain.) Certainly the chief human influenza pandemics (named after their presumed origins – Spanish flu 1918, Asian flu 1957, Hong Kong flu 1968) killed millions, so this is a highly significant disease to watch out for. The Hong Kong flu contains avian sequences and it is believed to have had pigs as an intermediate host. Influenza viruses encode just 8 proteins via 8 segments of negative-sense RNA, which are named the haemagglutinin (HA), neuraminidase (NA), matrix (M), nucleoprotein (NP), and nonstructural (NS) genes, together with 3 polymerase-encoding segments named PB1, PB2, and PA. There are also non-coding regions.

The particular scientific (and indeed socioeconomic) issue is that, especially when held in reservoirs of agricultural livestock, co-infection allows a recombination or reassortment of the genetic material of the virus, with unknown consequences on its ease of transmission between, infectivity of and virulence in its different host species. This recombination/reassortment may be potentially highly effective in generating novel diversity, as is known in both directed enzyme evolution and evolutionary computing. This is maybe what has happened in the recent ‘Mexican’ swine flu outbreak in N America, the viruses possibly containing sequences from 2 ‘pig’ strain, a ‘bird’ strain and a ‘human’ strain. There is evidence that domestication and association with humans may select for human pathogenicity, that precursors of pathogenic strains of the H5N1 strain are endemic in domesticated fowl in Asia, that the evolution and geographical spread of these strains may be considerable, that they may reside in such reservoirs for decades, and that there is selection for human receptor specificity. Despite the fact that they are both derivative of the influenza A virus, the topic “avian/bird flu/influenza” (4743 hits in Web of Knowledge) has been studied considerably more than has “pig/swine/porcine flu/influenza” (1470 hits), suggesting the need to remedy that.

There is of course some knowledge (e.g. Baigent & McCauley, 2003; Neumann & Kawaoka, 2006; Cinatl et al., 2007, 2007a; Abdel-Ghafar et al., 2008; Miotto et al, 2008; Srinivasan et al., 2008; Noton et al., 2009; Van Hoeven et al., 2009) of the relationship between the overall virus sequences and their pathogenicity, but the problem is that the possible number of sequences is simply enormous (and many are available). There is evidence, for instance, that just 2 amino acid changes in the H5 haemagglutinin sequence changed the binding specificity for its haemagglutinin sialic acid receptors for cell entry from that of ‘avian’ flu (i.e. susceptible hosts as birds) to that of humans. We commented in a recent article that the number of even just 30-mer nucleic acid sequences (430 or ~1018) was such that even with the smallest standard 5 μm microarray features, a 29 km2 ‘micro’array would be needed to assay them all in vitro; full-length sequences of Influenza A viruses in the current NCBI Influenza Virus database are 1489-1584 nucleotides long! Fortunately visualization tools are available.

Thus, like most problems in biology, this is a multiobjective combinatorial optimization problem of systems biology, in which the exact sequence of each of the 8 proteins can affect each of the important steps (e.g. survival, transmission, infectivity, virulence) in each of several hosts (at least various species and strains of pigs, birds and humans, and even guinea pigs). Environmental conditions such as temperature and humidity also play a role. This complexity emphasises why, given the cost and specialised facilities needed for ‘wet’ experiments on virulent pathogens, we also need to use high-class informatics to integrate knowledge of all of these things; as with other fields, there is a huge literature out there.

One strategy for such problems is a ‘divide-and-conquer’ strategy (similar to the classical reductionist strategy of molecular biology, but best approached with an eye on the iteration between ideas and data). Thus, although these subsystems interact and are not independent, we need to begin to understand better, and to integrate, the relationships between sequence and the contribution to the various steps leading to and including ultimate virulence of each of the different components of the virus, including those that determine its host range. High-throughput methods will help considerably. As our knowledge of human genomic variation improves, together with its relationship to the sensitivity of individuals to flu viruses, we can also begin to understand why some people (and animals) seem far more sensitive or resistant than others (though this is of course a far larger problem involving the entire immunological response). For instance, highly pathogenic H5N1 viruses not only trigger the overproduction of proinflammatory cytokines but also are resistant to the antiviral effects of interferon and TNF-α. Given appropriate knowledge, there is also presumably scope for breeding poultry with greater resistance to both transmission and infection.

In terms of prevention in humans, the chief strategies are vaccination against the pertinent strains, though that was not an unqualified success in an earlier swine flu programme (in time monoclonal antibodies may work in infected patients), and the use of small molecule inhibitors of the neuraminidase such as ‘Tamiflu’ (Oseltamivir) and ‘Relenza’ (Zanamivir) that work by stopping the virus emerging from infected cells. It is not known (to me) how easily resistance to these drugs might emerge, but clearly it is likely to be selected for, and there are known examples, including during infection. Relenza has been much less used and is considered more effective, but is not orally bioavailable (I suspect this is because of a lack of those important drug carriers, again). Clearly novel inhibitors (e.g. Ogata et al., 2009) may be important if they are bioavailable and well tolerated. Vaccine production is still mainly performed in chicken eggs, and this can take quite some time (ca 6 months for the necessary large batches) and is not easily expanded. Alternative methods are needed, and there is encouraging progress with vaccinia (e.g. Poon et al., 2009; Kreijtz et al., 2009; Mayrhofer et al., 2009; Rimmelzwaan & Sutter, 2009) As I blogged last time, we are still uncertain as to what epitopes (and indeed adjuvants) are likely to lead to a good vaccine, though progress in avian flu is happening. Note that in some diseases, vaccination may even enhance susceptibility,and we do not yet properly understand why.

As well as improving our detailed biological and biotechnological understanding, readers will be interested in lower-tech mechanisms to prevent the emergence of virulent strains of these kinds of organism. Minimising contact between wild birds, poultry and pigs is definitely one, and avoiding excessive clonality in our commercial breeding and production stocks is another. Avoiding transmission via non-pharmaceutical interventions such as face masks and general hygiene is also being evaluated.

Fineberg has commented, “Dollars spent on key research areas, such as the development of a safe and effective vaccine and of rapid diagnostics and understanding the molecular foundation of virus transmission, infectivity, and virulence, are necessary to improve the capacity of the United States and the world to respond to a pandemic threat. These investments are best understood as a trade-off between present preparedness and stronger future capacity to prevent and cope with a pandemic.”

So, even if pigs cannot necessarily fly, humans, birds and the viruses of all these and other organisms certainly can; we need to be as well prepared for this as is possible.

Related posts (based on tags and chronology):