To Catch a Virus. John Booss
it “flu” (56). The objections were economic and arose from a fear that the swine-flu connection would turn away the public from the consumption of pork. The epizootic in pigs was massive, with millions becoming sick and thousands dying. Recognizing the annual recurrence and enormity of the epizootics with such devastating economic consequences, Shope and Lewis began in earnest their investigations in 1928.
What they uncovered, reported in 1931 in a series of three articles in the Journal of Experimental Medicine, was quite remarkable (36, 54, 55). They discovered two agents working synergistically to produce the disease, with neither producing severe disease on its own. The first organism, very much like the Pfeiffer bacillus, was named Haemophilus influenzae suis and failed to induce experimental disease by itself. The second organism, a newly recognized filterable virus, induced a milder disease in experimental pigs than seen on pig farms. The more severe disease in swine resulted when the two agents were administered simultaneously. It later emerged in the report by Wilson Smith, C. H. Andrewes, and Patrick Playfair Laidlaw that the human and swine influenza viruses were antigenically closely related (57). As Shope put it, “. . . despite the failure of human investigators of the 1918 influenza pandemic to discover the cause of the outbreak, Mother Nature, using swine as her experimental animals, had done so” (56). Shope reported that he and Laidlaw independently reached the conclusion of the “. . . likelihood that swine had indeed acquired their infection from man in 1918. . . .”
Prior to his work with Smith and Andrewes in which the virus of human influenza was isolated in ferrets, Laidlaw had collaborated with S. W. Dunkin on the experimental study of canine distemper in ferrets (13–15). The work was conducted at the National Institute for Medical Research Farm Laboratories, Mill Hill, London, United Kingdom, where thorough procedures were employed to prevent exogenous infection of experimental animals. The reasons to use the ferret were that it could “be confined in a small space with ease and comfort” and that keepers claimed that ferrets were very susceptible to dog distemper, which could wipe out an entire breeding colony. Hence, special buildings, cages, personnel practices, and experimental procedures to study ferrets were well established at Mill Hill by the 1920s. In developing the ferret model under such controlled conditions, Laidlaw and Dunkin successfully showed experimental transmission of canine distemper virus to the ferret with overt expression of clinical and pathological features of disease. They demonstrated the disease to be caused by a filterable virus, not by the bacterium Bacillus bronchisepticus (Bordetella bronchiseptica), which they characterized as a secondary invader (15). When influenza appeared again in London in 1933, “The ferret obviously was the animal to test for susceptibility to influenza. . . .” (6). Laidlaw’s team was the right group to perform the studies to isolate a filterable agent.
Isolation of human influenza virus in ferrets, reported in the 8 July 1933 issue of The Lancet by Smith, Andrewes, and Laidlaw, was a signal event in the history of human influenza (57). The article described work that was fastidious in the care to exclude exogenous infection and elegant in the clarity of its results. It also demonstrated the experimental serendipity of susceptibility to various viral infections by different species of animals, so-called “species specificity.” Although the possibility of influenza being a viral disease had been raised in about 1914 (31), ambiguous results had been obtained in humans, and unsuccessful attempts were made in other species. Smith et al. commented that “The filtrates, proved to be bacteriologically sterile, were used in attempts to infect many different species. All such attempts were entirely unsuccessful until the ferret was used . . .” (57).
The report in 1933, which the investigators termed “a preliminary communication,” detailed a number of critical parameters for experimentation. These included the source and nature of the inoculum, throat washings from people sick with influenza. Experimental manipulation and important clinical observations in ferrets included intranasal instillation, the biphasic clinical course, the nasal histopathology in infected ferrets, and serial passage. Finally, the characteristics of the agent were documented, including filterability, the absence of bacterial growth, and the neutralization of the virus by serum taken from people who had recovered from clinical influenza and from ferrets that had recovered from experimental infection. Smith and colleagues also studied the relationship to the virus of swine influenza received from Richard Shope; they found “a close antigenic relationship” (57). However, “[i]n striking contrast to swine influenza,” there was no synergistic role for H. influenzae suis in the production of experimental disease in ferrets.
The Mill Hill investigators’ report on the successful use of the ferret as an animal model received prompt confirmation from investigators on other continents. For example, T. Francis, at the Rockefeller Institute in New York City, working with sputum obtained from patients in a 1934 influenza epidemic in Puerto Rico, transmitted the disease to ferrets (18). F. M. Burnet reported the experimental transmission of influenza to ferrets from a 1935 epidemic in Melbourne, Australia (4).
The successful isolation of human influenza virus in the ferret in several laboratories triggered the exploration of other biological systems. Andrewes et al. reported the successful transmission of ferret-passed virus to mice (1), which was also reported in the following month by Francis (18). Reports of the successful cultivation of the virus in minced chicken embryo soon emerged (19, 58). Smith concluded that the egg membrane technique was “unsuitable for the study of this virus” (58). However, further studies would show that this biological system, the embryonated egg of chickens, was remarkably productive for the understanding of human influenza infection.
There are parallels between the yellow fever story and the study of the cause of human influenza. Noguchi claimed to have isolated a bacterial cause of yellow fever, Leptospira icteroides, in guinea pigs. The bacterium, while causing symptoms in guinea pigs similar to those of yellow fever, turned out not to be the cause of yellow fever, and the guinea pig turned out not to be a susceptible host of yellow fever. As noted above, Pfeiffer in 1893 reported the isolation of a gram-negative bacterium as the cause of human influenza (48). Haemophilus influenzae, or Pfeiffer’s bacillus, as it came to be known, was for some time thought to be the cause of influenza (42). Richard Shope and Paul Lewis found Haemophilus influenzae suis to be associated with swine influenza (36, 54, 55). It was to be shown that the ferret was the model of choice