In 1994, a third of the lions native to the Serengeti ecosystem of Tanzania and Kenya were killed off by an epidemic of canine distemper virus (CDV). For the biologists of the Serengeti Lion Project, who’d been following the lion population of the Serengeti for going on three decades, the death toll was a shock.
In the hopes of understanding what happened, and perhaps mitigating future outbreaks, they set out to answer two big sets of questions. One was about the basic biology of the measles-like disease. What was it? Why was it so lethal? From what similar strains did it evolve? The other set of questions revolved around broader patterns of lion ecology and infectious disease epidemiology. How did the lions interact with each other? And how did those patterns of interaction—known as “contact patterns”—help spread the disease?
For help answering this second set of questions, they turned to Lauren Ancel Meyers, associate professor of integrative biology and one of the pioneers in the field of mathematical epidemiology.
“What attracted me to the system was the richness of the data available,” says Meyers, who’s been recognized for her work modeling the spread of human diseases like HIV, H1N1 (swine flu) and SARS. “For decades, the Serengeti Lion Project researchers have closely monitored lions throughout the ecosystem. They have named each lion and can uniquely identify them by their whisker spot patterns. On an almost daily basis, researchers observe groups of lions and record exactly what they are doing and who they are with. As a result, we have many years of detailed information about behavioral interactions among these amazing animals.”
Add in the information that the Lion Project has gleaned from lions fitted with radio collars and GPS tags, and the data they were able to collect during the actual spread of the CDV epidemic, and the result has been a goldmine for Meyers and her colleagues. With it, they can estimate how often lion prides come into contact with each other, how often these contacts are of the type that can facilitate disease transmission (e.g. fighting, mating, eating from the same carcass), and how groups of lions (prides and nomadic coalitions) move over the course of weeks, months, and even years.
Dr. Lauren Ancel Meyers is a pioneer in the field of mathematical epidemiology. She's been recognized for her work on, among other diseases, HIV, H1N1 (swine flu), and SARS.
The resulting model, in fact, is one of the most realistic ever created for disease transmission in wildlife species, and is far more precise than anything that’s been generated for humans (who aren’t particularly receptive to being followed around and observed by scientists in their natural habitat).
In a paper published in 2009, Meyers and her colleagues used documented cases of CDV infections to demonstrate that the epidemic in lions was likely the result of numerous, sporadic introductions from other carnivorous species like hyenas and jackals.
“Basically,” says Meyers, “the actual 1994 CDV epidemic jumped around too much in time and space to have been spread by direct lion-to-lion alone. Although lions do, in principle, interact enough to fuel an epidemic on their own, such an outbreak would exhibit a much more continuous wave of transmission than was actually observed.”
In a more recent paper, Meyers looked at the role played in the process of disease transmission by lion “nomads.” The question, says Meyers, is whether diseases like CDV are likely to be spread primarily through contacts between neighboring prides, or whether groups of nomadic lions, who aren’t attached to a pride, are acting as the critical agents of transmission to prides throughout the ecosytsem.
“Going into our study, we assumed that nomadic lions would be important vectors for spreading disease throughout the population,” says Meyers. “We knew that prides are highly territorial and only rarely interact with other prides, because it's costly to have aggressive confrontations. Nomads, on the other hand, move over much wider swaths of the habitat and may have physical encounters with multiple prides along the way. In our mental cartoon of the system, we imagined that nomads could pick up disease from an infected pride and traverse ecosystem to infect distant prides several days or weeks later.”
Surprisingly, as Meyers and her collaborators fed the data into their model, they found that although the day-to-day interactions of lion prides are mostly internal, prides do bump into each other often enough, and the contact is sustained enough, for disease to be spread directly through pride-to-pride interactions. In reality, then, the nomads are probably only marginal players.
“Prides that are adjacent to each other encounter each other fairly frequently,” says Meyers, “and even prides that are separated by the territories of other prides occasionally come into contact with each other. It doesn't take a whole lot of those contacts to fuel the spread of a disease through the ecosystem. While the nomads do, in fact, move much more extensively than prides, the speed at which they can carry diseases across large distances is not significantly faster than the speed at which diseases transmit directly among prides themselves.”
The value of such insights, says Meyers, is primarily lion-specific. Lion communities have such a unique structure that it wouldn’t be easy to extrapolate from their disease transmission patterns to other species. What might be of more general use, however, are the statistical and mathematical tools that Meyers and her colleagues have developed to construct the lion model.
“This approach to modeling infectious diseases is still at an early stage,” she says, “and we’re addressing general challenges to building predictive models. How much data does one need, for instance, to build a good model? How much detail does one need to include in model to get reliable predictions about epidemics? What kinds of biases does one have to account for and how can one best address them? As we get more data from diverse human and animal populations in the years to come, we will continue to answer such technical challenges and develop even better predictive models for forecasting and controlling disease outbreaks.”
Comments