A New Dynamic With an eye toward host-pathogen interactions, can a Penn State center predict and prevent the next pandemic? By Brendan Borrell

Beaver Pond is a half-hour from the campus of Pennsylvania State University in State College. In late winter, ice still lines its shores, but a swift-flowing stream keeps the surface from freezing over. Red-spotted newts make their home here, swimming in the shallows and swallowing large snails and insect larvae.

During the breeding season, tiny leeches climb aboard the newts, sucking their blood, and possibly transmitting Icthyophonus, a fungus-like pathogen that hides in the newt's muscle. Newts have other parasites, too. Tom Raffel, a postdoc at the Center for Infectious Disease Dynamics (CIDD) at Pennsylvania State University, has documented more than 20 different parasites in Pennsylvania newts. Two are new to science.

In the past, a scientist might single out a pathogen, map its life cycle, and describe the consequences for its victims. Although pathogens represent more than half of all life on earth, only a small fraction have ever been studied. So, a new approach to infectious disease is taking root both around the world and here on the shores of Beaver Pond. Raffel doesn't study newts, or leeches, or Icthyophonus. He studies the Beaver Pond community and the myriad interactions within. Just a few miles away at CIDD, researchers are looking at human pathogens, too - measles, influenza, and Escherichia coli among others - and trying to understand the communities of these pathogens within cities and within hosts, piecing together the way these interactions evolve over time.

Despite advances in vaccine strategies and drug treatments, many scientists worry that not enough is being done to suppress, let alone anticipate, the next pandemic. Scientists at CIDD are taking principles of population biology, community ecology, and evolution and wedding them to epidemiology, immunology, and genomics. This approach could help optimize vaccination strategies, design eradication programs, halt incipient pandemics, and it could identify potential zoonoses before they've infected humans. In the three short years that CIDD has been around, it's become a hotbed of interdisciplinary collaboration with 12 faculty members from departments around the Penn State campus.

Eye of Science / Photo Researchers, Inc.

Daniel Falush, an evolutionary geneticist at Oxford University, describes one effect CIDD has had in the United Kingdom: "There was a great sucking sound because these famous British scientists were disappearing to Penn State." Actually, Ottar Bj￸rnstad, a Norwegian mathematical ecologist, was the first to make the move to State College in 2001. At that time, Peter Hudson was at the University of Stirling in Scotland but was displeased with their new president, who he says wasn't supportive of biology. When Penn State invited him for a visit, he loved the atmosphere, and it didn't hurt that his friend Bj￸rnstad had already scoped out the local pubs.

Three days after Hudson arrived, he was at lunch with the dean and the head of the biology department and they asked him, "Now that you're here what are you going to do?" In that first meeting, he mentioned the names of epidemiologist Bryan Grenfell, then at Cambridge, and evolutionary biologist Eddie Holmes at Oxford. "I said, ?If we could get these guys we could become world leaders.'" The university offered Hudson and Bj￸rnstad half a million dollars to set up CIDD along with the salaries and packages required to attract new faculty. The center now supports two postdocs, and pooled grant support funds four more. At least twice a year, CIDD hosts a workshop on infectious disease topics. One of their first, how wildlife can dilute disease risk for humans, led to a 2006 publication in PLoS Medicine.¹

"Our vision really is to have a systems approach to disease," says Hudson. "Issues that go from intracellular interactions between viruses and cells right the way through to pandemics, something we call the protein-to-pandemic link."

Pathogens don't just interact, they evolve. In the 1960s, optimistic researchers had declared victory over infectious disease. It was time to move on to heart disease, cancer, and psychiatric disorders. That bubble burst in 1981 with the recognition of HIV. The disease had leapt from nonhuman primates to humans and continued to evolve and diversify in ways scientists still don't understand.

Bryan Grenfell began his career on this first wave of disappointment. He worked with Roy Anderson, an early pioneer of disease ecology at Imperial College London, to build mathematical models of the population dynamics of the brown stomach worm, a parasite of cattle. Over the years, his models became more complex as he added a spatial dimension and started crunching a measles dataset that spans fifty years. He spends most of his days in front of two 24-inch Apple monitors, on which he can scrutinize the output from enormous simulations piped in from a campus facility where a 10-teraflop computer bank crunches his data.

"These are the observed dynamics of measles in 60 cities in England and Wales," Grenfell says, pulling up a map with colored circles growing and shrinking as the 20th century flies by. "The areas of the circles are proportional to the number of cases," he says. He points to a box in the top right-hand corner, where two lines trace the total number of measles cases at any given time. "As you can see, London is in yellow there, and the rest of the country is in white. You can see the epidemic patterns and how synchronized they are." Grenfell explains this as a matter of fact, but this observation - that England and Wales follow the same oscillatory patterns as Baltimore and New York - was not at all obvious to earlier researchers.

A traditional Fourier analysis was not up to task because it could not account for spatial and temporal variation in the measles data set. Instead, Grenfell and his colleagues used wavelet time-series analysis, which allowed him to extract the hierarchical waves of infection that move from large cities into small towns. Robert May at Oxford says, "It was Bryan who thought, Ah-ha! Look at it at the level of the city, and you'll see the clear signal." Grenfell's work went further in that it helped demonstrate the importance of metapopulation theory - an ecological theory that breaks populations into subgroups with varying degrees of connectedness - to epidemiology.

As powerful as this approach was, Grenfell, too, had been leaving evolution out of the equation. But by 2004, molecular evolutionary methods were sophisticated enough that they could go hand in hand with epidemiological work. Population-level processes leave their traces in sequence data; evolution bears the imprint of epidemics past. Grenfell and a number of collaborators, including Eddie Holmes, wrote a paper in Science outlining an approach to unify the fields, coining the expression phylodynamics (see "Evolving Epidemiology").

"The challenge," says disease ecologist Nicholas Grassly at Imperial College, "is getting the data sufficient to apply those phylodynamic approaches." Indeed, the method is data-intensive, requiring fine-scale knowledge of molecular evolution and its phenotypic consequences. Katia Koelle, a former postdoc in Grenfell's lab, recently used phylodynamics to explain the punctuated nature of antigenic change in influenza A. For other diseases, the question is whether phylodynamics can provide practical information on eradication strategies that they can't obtain from standard mathematical epidemiological approaches. "It's on the brink," says Grassly, "Data [are] becoming available."

A modest collection of stuffed animal parasites sits on a shelf in Hudson's lab. Ebola and athlete's foot are not the huggable pathogens that Hudson studies, but the cartoonish representations epitomize his big picture approach to disease. "I'm really a naturalist at heart," he says. Somewhere along the way, he realized that parasite communities can regulate wildlife communities and vice versa. In 1985, he proposed a radical hypothesis, which says that parasites cause cyclic oscillations in animal populations, specifically the fluctuations of the red grouse, a managed game bird in Scotland. Thirteen years later, he convinced the last of his skeptics by treating birds with an anthelminthic drug. Their populations stopped cycling. It "was the first time that had ever been done," he says.

Each week, members of Hudson's lab meet for a journal group. Today, they are discussing an article in PLOS Biology that reconsiders the basic reproductive number, R0, of malaria, a parameter that correlates with its transmission rate.² Ronald Ross, one of the first to use mathematical models in epidemiology, first introduced the concept of R0 in 1897. If R0 is low (less than one), then it may be possible to wipe out the disease, but as it increases eradication becomes increasingly unlikely.

Still, the work of mathematicians was neglected by public health officials for the better part of the century. As May puts it, epidemiologists in Africa were deemed about as useful as CAT scanners, seen as overkill in a continent lacking basic human services. Halfdan Mahler, head of the World Health Organization from 1948 to 1990, dreamed of barefoot doctors in every village.

Hudson grows animated as the group discusses the paper, which demonstrates that Ross was only partly right. The basic reproductive number is remarkably heterogeneous: It can vary anywhere between 1 and 3,000. The students complain that the paper is impractical, and one student suggests that it really just comes down to finding breeding mosquitoes and wiping them out on a case-by-case basis. Hudson wonders out loud: "Do we really need science to address these questions or is it just common sense?"

He talks about a privately run park he visited in South Africa in which operators dumped animals of different types into fenced-in tobacco fields. "To begin with, of course, they bought too many leopards and cheetahs and they ate all the gazelles and things. Then, they all started dying off, and the elephants increased quite dramatically and they suddenly discovered they were knee-deep in elephant [dung]," he says. "The only thing they needed to introduce was dung beetles. Then, everything went off. We didn't need ecology to tell them how to do it. They just put it all in the same pot and stirred it up."

His comments hint at the challenges CIDD will face as they attempt to scale up from proteins to pandemics. Uncertainties at the small scale become magnified in the ecological realm, and the true currency of disease ecology is not a measurable parameter like R0, but rather, stochasticity and heterogeneity. Research of the type that CIDD is promoting may never conquer infectious disease, but it might assist in controlling it.

Later that afternoon Hudson is heading to a late meeting. He's been scheming for a way to hire three new faculty members from the United Kingdom and Australia, and he counts on his fingers the number of E-mails he has to send to gain support from the administration. Bj￸rnstad, on his way to the Allen Street Grill, asks Hudson to join him, but to no avail. Hudson has two more meetings planned that night at his house. With a grin, he adds that he likes talking about enormous gobs of money. He's been doing that quite a bit lately. Recently he claimed space in a proposed biology building that will bring together CIDD faculty currently dispersed on campus.

Allen Street Grill is one of several pubs crammed onto the corner directly across from campus. In the wood-paneled room upstairs, Bj￸rnstad greets Grenfell and Stephan Schuster. Grenfell and Bj￸rnstad recently improved the measles model by including a "gravity" term, which factors in both the distance and size of cities in propagating an epidemic. Schuster, another CIDD member, has just developed a program, MEGAN, which can be used in performing a metagenomic analysis on a slab of beef, taking advantage of the shotgun approach to classify every microbe contained within. In his spare time, he's been sequencing the woolly mammoth.

As drinks are poured, the conversation eventually shifts from work to a recent announcement of a futures market for avian influenza. The Robert Wood Johnson Foundation put up roughly $250,000 to create a trading market in which authorities might wager on their predictions for pandemic scenarios. In a bar full of experts, even lighthearted prophesy takes on a foreboding air. Postdoc Jamie Lloyd-Smith remarks, "I'm bullish about bird flu."