Editors' ChoiceVaccines

COBRA Insures Against Flu

See allHide authors and affiliations

Science Translational Medicine  04 Apr 2012:
Vol. 4, Issue 128, pp. 128ec57
DOI: 10.1126/scitranslmed.3004064

The threat of a novel pandemic flu continues to hover darkly over each flu season. The highly pathogenic H5N1 circulates among birds, with relatively few reported cases of spread to humans, and because the virus is not easily transmitted among people, the absolute number of patient deaths has been low. However, the mortality among humans who have been infected with H5N1 is high—a startling 60%. Ten clades (or families) of the pathogenic H5N1 avian influenza viruses continue to spread among waterfowl, and four of these (0, 1, 2, and 7) also have caused infections in humans. A major unmet need is a flu vaccine that provides cross-clade protection. Now, Giles and colleagues use computational algorithms to demonstrate how this goal might be achieved.

Structural and functional division of flu viruses does not stop at clades, which are further divided into strains. Influenza strains evolve over time, and genetic changes may produce a virus with the high mortality of H5N1 that also spreads efficiently among humans. But because we cannot predict how influenza genes will mutate over time, ideal vaccines would cover more than one clade, and the broad immunity conferred by these multitalented vaccines would protect vaccine recipients from serious disease no matter which clades spread to from fowl to humans.

Giles and colleagues used a new method called COBRA—computationally optimized broadly reactive antigen—to generate immune-reactive viral antigens on the basis of global H5N1 surveillance and genome sequencing. Computer algorithms analyzed the genome sequence information to select antigens with characteristics that represent a wide variety of H5N1 clades. These antigens then were used to produce a vaccine that that was expected to display a wide breadth of cross-clade coverage.

Traditional vaccine-antigen selection methods zero in on the most prevalent antigens and are subject to selection bias—antigen selection invariably depends on the sample of viruses contained within the selection pool. The authors tested the cross-clade computer-designed vaccine in a macaque model that had not been used previously for clade 2 vaccine testing and demonstrated broader protection (that is, higher antibody production with less inflammation in the lungs) compared with that elicited by a single H5N1 strain–containing vaccine. Although the optimized broadly reactive antigens were chosen from clade 2 viral genomic sequences, the resulting vaccine also recognized viruses from clades 1 and 7 (but not clades 0 and 1). These results suggest that the COBRA method can aid in the design of craftier vaccines with increased clade coverage and, consequently, higher effectiveness. Such vaccines may provide physicians with versatile weapons for fighting a future flu pandemic.

B. M. Giles et al., A computationally-optimized hemagglutinin VLP vaccine elicits broadly-reactive antibodies that protect non-human primates from H5N1 infection. J. Infect. Dis. 23 March 2012 (10.1093/infdis/jis232). [PubMed]

Stay Connected to Science Translational Medicine

Navigate This Article