Deep in the Pacific Northwest rainforest, a rough-skinned newt flashes its brightly colored belly at an oncoming attacker. For those predators that ignore the warning, a deadly neurotoxin ensures that the bite will be the last they take.
That is, unless the predator is a red-sided garter snake.
But even as this slithering reptile's nervous system has evolved to resist the chemical warfare, the newt, in response, has evolved ever more potent toxins in an attempt to defend itself against its only successful predator.
Such striking examples of joint evolution are not limited to the rainforest floor or the Savannah grasslands.
Even the human body, when confronted with a lingering viral infection or an encroaching tumor, becomes an ecosystem in which the struggle for evolutionary advantage is fought.
It's the dynamics of such microevolution in human ecosystems that intrigue Dr. Dominik Wodarz, a mathematical biologist in the Public Health Sciences Division.
Like the ecologists who study predator-prey interactions between creatures in the wild, Wodarz designs mathematical models to predict the adaptive changes that occur, for example, in viruses and tumors as they attempt to evade a watchful immune system.
In the case of ecosystem struggles between human immune cells and viruses like HIV, the virus that causes AIDS, such models can predict the most opportune times during the course of infection to administer antiviral drug therapy.
Wodarz, who joined the center in July, believes that only a combination of experimental and theoretical approaches can yield answers to such complex problems as the human immune response to HIV infection.
"Many immune-system factors interact during a viral infection," he said. "To understand the outcome, you can't design experiments without mathematical equations."
But the mathematician and the experimental biologist, he said, must work in lock step.
"It goes both ways," he said. "An experiment first gives rise to a problem for which a mathematical model can be developed. The model then leads to predictions that can be tested experimentally."
Among the factors that influence outcome of virus-host interactions are rates of viral multiplication, the numbers of immune cells involved in the response and drugs that either inhibit the virus or boost the immune system.
With subtle adjustment of one or more of these parameters, the model allows researchers to set up an elaborate theoretical game of cat and mouse that may reveal the ideal conditions under which the immune system gains the advantage.
Applying this approach to predict the outcome of HIV infection is a particularly challenging problem. Antiviral drugs, even potent cocktails of combination therapy, cannot completely eradicate every trace of the virus from its host. In addition, viral genes mutate rapidly, potentially giving rise to drug-resistant variants.
These factors, Wodarz said, mean that the time at which drug therapy is initiated can profoundly affect the ability of the immune system to dominate the interaction successfully. He and colleagues have developed mathematical models to predict drug regimens that would stimulate immune system memory cells' response to the virus should the viral load, or amount of virus in the blood, become elevated.
"Our models suggest that if drug therapy is begun early during the course of an infection, the rate of viral replication is inhibited, allowing the immune system to gain the upper hand," he said.
Experiments have since revealed that such a therapy regimen enables animals to mount a successful immune response upon subsequent re-exposure to the virus.
Wodarz's approach also has proved useful for study of other important pathogens that exhibit genetic variability - for example, the virus that causes the common cold. Collaborating with a pharmaceutical company, he has developed models to facilitate evaluation of drugs that might reduce symptoms of rhinovirus infection.
His interest in human ecosystem interactions is not limited to infectious diseases. Wodarz also developed mathematical models for studying cancer progression and treatment. For example, models can be used to describe response to chemotherapy, to evaluate immune system response to cancer and to evaluate the dynamics of novel therapies such as viruses engineered to attack cancer cells.
Like his research interests, his educational background is a blend of mathematics and ecology.
A native of Germany, Wodarz earned a Ph.D. in mathematical biology at the University of Oxford, where he began his investigations on virus-immune system modeling.
He completed his undergraduate degree in biology at Imperial College in London, with an emphasis on the dynamics of host-parasite dynamics. Most recently, he was a member of the theoretical biology program at the Institute for Advanced Study in Princeton, N.J.
With the many potential applications for mathematical models, Wodarz hopes to engage in collaborations with center laboratory researchers who study the dynamics of HIV infection or tumor progression.
Already, he has begun projects with Dr. Michael Emerman, a Human Biology Division investigator who studies HIV multiplication, and Dr. Larry Loeb, University of Washington professor of pathology and head of the Fred Hutchinson/UW program in genomic instability, who studies the relationship between mutations and human cancer.
Although his research interests focus on predator-prey dynamics, Wodarz expects that such collaborations will reflect a different ecological phenomenon: mutualism.
"You can learn the most by a combination of the theoretical and the experimental," he said. "The more new and interesting problems I can apply my models to, the better."