Emory Researchers Offer New Approach For Testing Potential HIV Vaccines In Animals

In the Emory study, the researchers, using computer simulations, developed an experimental design in which animals are repeatedly exposed to low doses of HIV (similar to how humans are exposed and infected). The belief that experiments involving realistically low challenge doses would require large numbers of animals has so far prevented the development of such trials, the researchers say.

Through computer simulations and statistical analysis of their virtual experiments, the Emory researchers showed that such trials would require far fewer animals than previously thought. Their research was published in the July 19 issue of the Public Library of Science Medicine. The full journal article can be found here.

"We demonstrate that using low doses and challenging repeatedly -- which also is more realistic because humans are typically exposed repeatedly to HIV -- represents a very promising design," says Roland Regoes, a postdoctoral researcher in Emory's biology department and lead author of the study.

Trials in animal models have long played an essential role in evaluating the effectiveness of potential HIV vaccines and treatments. When assessing vaccine efficacy in animal models, the animals are first given the potential vaccination. They are then "challenged" (or infected) with the virus or pathogen against which the vaccine should give protection. In simian models of HIV infection, Simian Immunodeficiency Virus (SIV), closely related to HIV, is used to challenge macaques. The trials are usually conducted with very high challenge doses of the virus that result in certain infection.

"Developing a vaccine against HIV is one of the major goals of AIDS research," he says. "Our work suggests how to improve the animal models in which possible vaccines are assessed before they are tested clinically in humans. By using lower doses and challenging repeatedly, preclinical trials would be more similar to epidemiological phase III trials in humans. This allows us to optimize vaccines in preclinical trials with respect to what really matters epidemiologically."

Using a standard statistical power analysis, the researchers simulated low challenge dose experiments more than 100,000 times. The outcome of these virtual experiments was statistically analyzed in the same way a real experiment would be.

Infection with low doses has to-date not been performed very often, Regoes says, adding that a handful of research groups have already started trials using low-dose models.

"It will be interesting to study HIV infection following challenges with low virus doses, and I am certain we will be surprised in many ways by results of these studies. I believe that low-dose challenge experiments represent an interesting alternative for the preclinical assessment of vaccines," Regoes says. In addition to improving the design of preclinical trials, low challenge dose experiments may also allow AIDS researchers to investigate "immunological correlates of protection," such as how many antibodies or T cells are needed to prevent infection, he says.

Regoes conducted the study with Ira Longini of Emory's Rollins School of Public Health, Silvija Staprans of the Emory School of Medicine and the Emory Vaccine Center, and Mark Feinberg, a researcher with Merck. They made several recommendations in the PLoS paper on how to design preclinical studies with low doses. Some of these suggestions are related to specific aspects of conducting such trials, such as how much and how often subjects should be exposed to HIV. Other recommendations aim to make preclinical trials more realistic in other important ways, such as the route of infection or the HIV strain to use.