"Smallpox vaccine is the best vaccine ever devised by man, which is ironic since it's basically unchanged since Jenner's time," said Jonathan Yewdell, chief of the cellular biology section in the Laboratory of Viral Diseases at the National Institute of Allergy and Infectious Diseases. "It's the only vaccine to have successfully eradicated a disease, though polio may be close."

Edward Jenner carried out the first purposeful vaccination program in the late 18th century, when he used cowpox as a vaccination to protect against smallpox. However, the eradication of smallpox ultimately has not lessened the need for a smallpox vaccine. Instead, with the disease a likely candidate for bioterrorist activities, research into smallpox vaccines has increased over the past few years. And while the smallpox vaccine might be the best vaccine ever devised by man, it still can have dangerous side effects, particularly in immunocompromised people.

Whatever gave smallpox its name, it was not the size of its genome, which is one of the largest known for a virus. And that supersized genome poses a problem for researchers trying to map antigenic determinants - for practical purposes, the epitopes - for vaccine development. Several strategies exist for winnowing all the possible epitopes down to a manageable number.

One is to exclude proteins expressed late in the viral life cycle on the theory that the immune system is likely to recognize early expressed proteins, and the earlier it gets a crack at the virus, the better its fighting chance. Another is to rely on sequence comparisons of different vaccinia virus strains (a more benign relative of the smallpox-causing variola virus that is used for vaccine production) and focus on highly conserved protein sequences.

Now, research published in a paper titled "Identification of poxvirus CD8 T-cell determinants to enable rational design and characterization of smallpox vaccines" in the Jan. 3, 2005, issue of the Journal of Experimental Medicine suggests that both strategies might not lead to the strongest possible vaccines. The scientists, who are from the NIAID; the Queensland Institute of Medical Research in Herston, Australia; Australian National University in Canberra; and the La Jolla Institute for Allergy and Immunology in San Diego were investigating the contributions of cytotoxic "killer" T cells to the immune response after vaccination.

"Most vaccines that work are based on antibodies," Yewdell told BioWorld Today. "If you can make a neutralizing antibody response, that's best because it can work on the virus before the virus can get up much steam." T cells, in contrast, can work only after several rounds of viral replication.

However, it is not always possible to elicit an antibody response. Also, when attenuated live viruses are used for vaccination, as is the case for the current smallpox vaccine, the exact mechanisms of their effectiveness, and to what extent that effectiveness is based on antibody vs. T-cell responses, often are unclear.

"People would be surprised by how little we really know about what works in vaccines," Yewdell said.

For a given virus, different T cells respond to a number of epitopes. There usually is one response stronger than the rest, and that dominant response plus responses to a small number of subdominant epitopes account for a fairly large proportion of the total response.

Late Bloomer Elicits Killer T-Cell Response

To elicit the most potent response possible, the scientists wanted to identify both dominant and subdominant T-cell responses to the vaccinia virus. They first generated an expression library of the vaccinia proteome, inserting the gene for each protein into a separate plasmid and expressing the proteins in cell culture. Those cells then were introduced to killer T cells isolated from mice infected with vaccinia virus. Five of the proteins induced an IFN-? immune response in the killer T cells; at least one of those proteins is expressed late in the viral-replication cycle, meaning that screens based on expression timing would have failed to identify that epitope. The expression timing of two others are unknown. Through a combination of in silico predictions and empirical testing, the scientists then identified the specific epitopes responsible. In follow-up in vivo experiments, the five epitopes were shown to account for about half of the total killer T-cell response to vaccinia virus.

The scientists then used synthetic peptides to test whether vaccine candidates based on different vaccinia strains could induce T-cell responses to all five epitopes. WR is the current standard vaccine, which is based on live attenuated vaccinia virus, and was used in the experiments that first identified the epitopes. The scientists compared WR to a vaccine known as Dryvax, which is based on a different strain of vaccinia virus, and modified vaccinia virus Ankara (MVA), a candidate replacement smallpox vaccine based on a third strain.

The scientists found that while Dryvax-immunized animals responded to all five epitopes they had identified using T cells isolated from WR-immunized mice, MVA-vaccinated mice failed to mount an immune response to two of the epitopes, even though the epitopes themselves were contained in the MVA gene sequence. In fact, the response to each of the four subdominant epitopes was significantly different between MVA- and WR-immunized animals. Only the response to the dominant epitope remained relatively unchanged. Control experiments showed that this was not merely due to the fact that MVA is a weakened vaccine.

Unexpectedly, the method of vaccination also affected the immunodominance hierarchy. Vaccinia-based vaccines usually are administered by scratching the skin, rather than through injection. When the scientists compared responses to the epitopes after skin administration vs. injection of the vaccine, the dominance hierarchy was strongly affected. Skin administration led to a stronger response to the immunodominant epitope and weaker responses to the four subdominant ones. In many mice vaccinated by skin administration, responses to two of the subdominant epitopes were extremely weak, meaning that they effectively mounted a T-cell response only to three of the epitopes.

"That's probably the finding with the greatest practical application, that the route of administration could affect the quality of the response," Yewdell said. "And again, you wouldn't care if all the [killer T cells] could all do the same thing. But maybe they can't."