Viral attenuation by design

A less-virulent poliovirus generated by altering codon usage suggests a new strategy for vaccine development.

Although live attenuated viruses provide the most effective vaccination strategy ever devised, current research has largely abandoned classical methods for vaccine development in favor of 'high-tech' approaches such as protein subunit and recombinant DNA vaccines. A recent study¹ in Science describes a method for reducing viral virulence that may eventually lead to a new class of live attenuated vaccines. Manipulation of codon usage has been exploited previously to increase the expression of select antigens in DNA vaccines²; Coleman et al.¹ apply it to opposite effect—to decrease gene expression in poliovirus by introducing underrepresented codon pairs (Fig. 1).

Figure 1: Development of an attenuated vaccine by two classical methods and the proposed method of Coleman et al.¹.

Kim Caesar

(a) A virus from one species is used in another species. Edward Jenner used fluid from a milk maid's acquired case of cowpox to demonstrate the protective potential of the cowpox virus for human smallpox⁸. (b) Serial passage in chicken cells. Maurice Hilleman recovered mumps virus from a clinical case in his 6-year-old daughter and passed the recovered virus in embryonated eggs and then in cultured chicken cells to produce the Jeryl Lynn mumps vaccine⁹. (c) Manipulation of codon pair bias. Coleman et al.¹ used a computational algorithm to design a sequence with codon pair bias, an automated DNA synthesizer to construct the generated sequence, and transfected cultures to recover an attenuated poliovirus.

Live attenuated vaccines have saved millions of lives and days of sickness by eliciting long-lasting antibody and T-cell responses³. Examples include the highly successful smallpox, measles, mumps, rubella, chicken pox and oral polio vaccines. Unlike inactivated, subunit or recombinant protein vaccines (such as those against diphtheria, pertussis, tetanus, hepatitis B virus and flu), which require multiple doses and regular boosting, live attenuated vaccines provided as single doses elicit long-lasting protection. Moreover, they are the most cost-effective form of vaccination. Manufacture requires only the growth of the virus. Purification is minimal because the vaccines, which amplify as they vaccinate, are effective at low doses. Live attenuated vaccines have the additional benefit of not requiring adjuvants: replicating pathogens provide the reservoirs of antigen and 'danger signals' needed to mobilize the immune response that adjuvants provide for nonreplicating vaccines.

Despite the tremendous advantages of live attenuated vaccines, current vaccine research is centered on the use of crystal structures to design protein subunit vaccines⁴, of databases to develop inserts for recombinant viral and DNA vaccines⁵ and of cell-signaling microarrays to choose adjuvants⁶. Gone are the days when live attenuated vaccines were developed with biological systems, such as serial passage of a pathogen (mumps and measles vaccines), testing of virus derived from individual virions (poliovirus vaccine) or the use of related viruses from other species (smallpox vaccine). Coleman et al.¹ have provided a contemporary approach for the development of live attenuated vaccines by showing that a pathogen's virulence can be restricted simply by engineering its codon pair bias. Their approach, termed synthetic attenuated virus engineering (SAVE), involves substitution of synonymous codons to reduce the codon pair bias characteristic of the wild-type virus.

Specifically, the authors used recombinant DNA technology to produce a poliovirus in which the 5′-most coding sequence, for the capsid protein P1, was recoded with 631 synonymous mutations to give a codon pair bias ∼4.5-fold lower than that of wild-type poliovirus. The attenuated virus was further manipulated by recombining segments of wild-type polio sequences into the codon pair–biased sequence to generate two viable viruses with reduced translation and replication.

How well did manipulation of codon pair bias translate to vaccine efficacy in vivo? Compared with wild-type virus, the two attenuated viruses required tenfold higher doses of plaque-forming units to kill transgenic mice expressing the poliovirus receptor. In a pilot vaccination study, three high doses of the attenuated vaccines given at weekly intervals resulted in the death of 9 of 16 mice but protected the 7 survivors against a paralytic polio challenge. The codon pair bias selection resulted in two attenuated viruses that are only about tenfold attenuated in their 50% paralytic dose and, at the same time, at least tenfold attenuated in their ability to be manufactured¹. In contrast, the selection of individual infectious units of poliovirus for attenuated neurotropism resulted in an oral vaccine that is >500,000-fold attenuated in its 50% paralytic dose without any attenuation of its ability to be produced⁷. Thus, selection against neurotropism to create the current polio vaccine, which involves three key codons⁷, yielded a much more effective vaccine than the 631 mutations in capsid protein P1 (ref. 1).

Efforts to eradicate polio have suffered more from a lack of political will and popular fears than from limitations on the ability to manufacture and deliver the current oral vaccines. Although the use of SAVE to produce growth-attenuated polioviruses is unlikely to yield a vaccine that is superior to existing polio vaccines, this strategy may prove valuable in the case of other infectious diseases for which there are no effective vaccines. However, any candidate attenuated virus generated by this approach would require attenuation for disease prevention and not only for growth. For example, our current highly successful polio vaccines were selected for the following: “(1) maintaining high degree of infectivity in cell culture and the human intestinal tract, (2) inducing detectable levels of neutralizing antibody in a high proportion of susceptible (seronegative) recipients, (3) displaying low neurovirulence in monkeys, (4) demonstrating a lack of association with paralytic disease in humans, and (5) maintaining genetic stability after replication in the human host”⁷. Thus, the study of Coleman et al.¹ opens a new path to attenuation that despite its universality will have to be tailored to the pathogenesis of each agent to be useful for future vaccine development.