Commentary: More than Trial and Error
Veteran vaccine developer Stanley Plotkin reminds researchers that not all vaccines were developed without any notion of how they worked
By Stanley Plotkin
“Time present and time past are both perhaps present in time future.”
—T.S. Eliot
As new strategies for vaccine development proliferate, mainly based on genetic engineering, and as systems biology comes into its own, it has become cliché to say that prior vaccines were developed empirically, without any idea 
of the mechanisms leading to effective immunogens or the ways to produce them. Perhaps a member of the older generation may be forgiven for pointing out that this is a canard with respect to any vaccine developed since the mid-20th century, and is insulting to a large group of living and dead researchers.
It is true that Jenner had no clear idea of how vaccination protects when he used vaccinia virus to prevent smallpox. Rather, he developed this approach based on the observation that milkmaids were protected against the disease by prior exposure to cowpox, and on the analogy of vaccination to variolation, the centuries-old practice of minimizing the severity of a later smallpox infection by deliberate infection with dried smallpox scabs in order to provide immunity during subsequent exposure to natural smallpox. It is also true that Pasteur’s discovery of attenuation of the chicken cholera bacillus some 80 years later was serendipitous.
Nevertheless, Pasteur applied attenuation techniques to other organisms including the anthrax bacillus, and, after his famous experiment at Pouilly-le-Fort showed protection by the anthrax vaccine, he stated that it was now possible to generalize his methods and to make vaccines at will. Moreover, by the end of the 19th century, Salmon and Smith in the US and Roux and Chamberland in France developed methods to kill bacteria with the expressed purpose of rendering them harmless while guarding their immunogenicity. “Immunity,” they wrote, “is the result of the exposure of...the animal body to the chemical products of the growth of specific microbes which constitute the virus of contagious fevers.”(1)
The discovery of diphtheria and tetanus toxins by Behring and Kitasato, also at the end of the 19th century, confirmed that substances secreted by bacteria could cause disease. In addition, they demonstrated that injection of those substances elicited neutralizing factors in blood. Early in the 20th century, Paul Ehrlich’s idea of antibodies as substances generated by the host in response to foreign antigens became accepted, and when Ramon inactivated diphtheria and tetanus toxins by formol to produce toxoids, his purpose was to induce antibodies to the toxins without causing adverse reactions.
During the first half of the 20th century, two major vaccines were developed, BCG [Bacillus Calmette-Guérin] and yellow fever. In both cases, the idea was that passage of a pathogen in bacteriological media or in an unnatural host would weaken its virulence for the natural host, as had been demonstrated by Pasteur for rabies years before. Success was achieved by Calmette and Guérin, who developed the BCG vaccine, and by Max Theiler, who developed the yellow fever vaccine, although it should be noted that Theiler had several competitors who were also passaging yellow fever virus in animals. All were pursuing the idea that selection by passage in a foreign host or in vitro would favor attenuation. Although they and subsequent researchers knew nothing of genetic engineering, they did know that their methods were selecting for attenuated variants.
The discovery that viruses could be grown in cell culture allowed expanding the method of attenuation by serial passage in cell culture. Of the two polio vaccines that launched the mid-20th century explosion of vaccines, the oral vaccines of Koprowski and Sabin followed precisely that idea. They selected populations of poliovirus cloned by plaquing in cell culture and tested each one for monkey neurovirulence. Similarly, measles and mumps vaccines were created by passage in cell culture, with attenuation measured in humans. During the fiercely competitive activity to develop avirulent strains of polio, it was found that passage at suboptimal growth temperatures allowed more rapid attenuation, a principle followed in the development of rubella vaccine. Also, varicella vaccine could be created because passage in guinea pig instead of human cells exerted a profound selection for attenuation.
Meanwhile, Salk’s inactivated polio vaccine followed directly his work on inactivated influenza vaccine, the goal of both being to produce antibodies. In fact, the success of the polio vaccine was foretold by the prior demonstration that polio could be prevented by serum antibodies contained in pooled gamma globulin.
Development of the vaccines that have followed was also predominantly guided by the principle that the induction of antibodies either in the serum or on the mucosa is the way to prevent infection, as repeatedly demonstrated by prior experimentation. This principle extends to the vaccines recently produced by genetic engineering: hepatitis B and human papillomavirus. Rotavirus vaccines were developed by serial cell culture passage or by reassortment of RNA segments of the viral genome to induce limited replication in the intestine with the aim of inducing mucosal antibody and local cellular responses similar to those after natural infection. In contrast, the zoster vaccine induces T-cell responses that prevent the reactivation of varicella virus latent in neurons, following another established principle that once infection takes place, T-cell responses control replication.
My desire here is not to deny the hope we all have that future vaccine development will be informed by new strategies, including the identification by systems biology of genes whose products predict protection. However, hope for the future does not require denigration of the past. The main difficulty we face in the case of vaccination against HIV is our inability to identify naturally protective immune responses, particularly on the mucosa. However, studies in non-human primates and the recent moderately successful RV144 trial suggest that as in other diseases, antibody of the right specificity and functionality prevents infection, whereas cellular immunity controls replication.(2)
The writer, an emeritus professor of the University of Pennsylvania and the Wistar Institute, developed the rubella vaccine that is used throughout the world, is co-developer of the pentavalent rotavirus vaccine, and has worked extensively on the development and application of other vaccines including anthrax, oral polio, rabies, varicella, and cytomegalovirus. He is a consultant to all major vaccine manufacturers.
1. Salmon DE, Smith T. On a new method of producing immunity from contagious diseases. Amer. Vet. Rev. 10: 63-67, 1886
2. Plotkin SA. Correlates of protection induced by vaccination. Clin. Vaccine Immunol. 17: 1055-65, 2010.