February 9, 2016
Emerging diseases have captured the public’s attention over the past several years – Ebola, H1N1 Flu, MERS, and now Zika virus. Such diseases pose global threats to humanity due to our globalized transportation system, which can move diseases across the world in a matter of hours. However, people and their domesticated animals have been migrating from place to place and carrying diseases with them for centuries.
One of the oldest such diseases of domesticated animals is bovine tuberculosis (BTB), a disease that mainly infects cattle, but can spread to many other mammals, including wildlife and humans. In the early 1900’s, cattle carrying bovine tuberculosis were brought to South Africa, which was then free of the disease. Around 1960, cattle transmitted bovine tuberculosis to some African buffalo in the southern part of Kruger National Park. Being in the same family as cattle, buffalo proved to be a good host for the disease and it began spreading from buffalo to buffalo.
Bovine tuberculosis is a relatively slow disease. Unlike some diseases that spread through a population in days or weeks, bovine tuberculosis (like human tuberculosis) doesn’t cause immediate sickness in the majority of individuals it infects. Instead, it lurks, causing symptoms and contagiousness many months or years after initial infection. And so the disease spread very slowly north through the park, reaching the northern boundary with Zimbabwe and Mozambique in 2006.
For the past quarter century, buffalo in Kruger National Park have been monitored and tested and studied to determine the impact of bovine tuberculosis on the buffalo population. It turns out that while buffalo do get sick and die from the disease, especially during droughts, the buffalo population itself doesn’t get smaller because of it. Enough new buffalo are born to accommodate those that die of bovine tuberculosis.
Then in 1995, park veterinarians got a shock: a lion tested positive for bovine tuberculosis. And soon tourists started to report sick, emaciated lions lying by the sides of roads. Park managers were worried.
The number of lions has been falling across Africa for decades, and they are currently classified as ‘vulnerable’ on the International Union for Conservation of Nature (IUCN) Red List of Threatened Species. There are now only a handful of reserves left in Africa that are big enough to contain self-sustaining lion populations, including Kruger National Park. What would become of Kruger’s lions now that they were catching bovine tuberculosis?
Answering this question is not easy. While tuberculosis has been studied in humans for decades, its effects on animals are not as well understood. And the impact on lions was completely unknown in 1995. It also wasn’t clear exactly how lions caught the disease. Did they get it from breathing infected air near buffalo? Did they get it from eating infected buffalo? Could they pass it from one lion to another? Veterinarians started tackling these questions immediately.
Meanwhile, there was the question of the entire lion population. How many lions were infected? Where were they? Would the lion population prove resilient like the buffalo population and maintain its size, or would the number of lions in the park fall? Wildlife biologists began investigating.
In 2009, a meeting was convened, organized by the Conservation Breeding Specialist Group (part of the IUCN) to bring together a diverse set of experts to share everything that was known about lions and bovine tuberculosis. Attendees also developed a simulation model, based on shared knowledge, to predict the future of Kruger’s lions. But there was a problem: the model consistently predicted a sharp decline in the lion population between the first confirmed lion bovine tuberculosis test in 1995 and the meeting – a decline that hadn’t been seen in real life.
After the meeting, I and several other attendees reconsidered the model. It had, in part, been based on scientifically supported data. But it had also relied on educated guesses by experts based on their observations and experience. We wondered whether it would be possible to model the lion population using fewer educated guesses.
We turned to a new modeling approach, called Approximate Bayesian Computation. In our new model, we were able to leave out all the information that we were unsure of, and only include data in which we had high confidence. In particular, we know a lot about lion social structure and the number and location of lions throughout the park. We also had information about buffalo infection levels over time. What we didn’t know were epidemiological variables about how bovine tuberculosis spread through the population and about its direct impacts on individual lions. We ran hundreds of thousands of simulations on the University of Minnesota Supercomputing Institute’s computers to figure out which epidemiological variables would match on-the-ground observations of Kruger’s lion population size and the prevalence of bovine tuberculosis in Kruger’s lion population over time.
The model provided two mathematical solutions. In one solution, lions become infected with bovine tuberculosis, but don’t necessarily show symptoms for many months or years. When they do become sick, they die fairly quickly. In the second solution, lions become sick quickly after becoming infected, but mostly don’t die from the disease. Because the second mathematical solution did not agree with veterinary findings, it was considered biologically unrealistic. This left the first solution, which specified epidemiological values that could be used in further modeling.
The parameterized model was run decades into the future and the results showed several things. First, the model showed that under current conditions, bovine tuberculosis will cause the lion population will drop only slightly, about 3%. It also implied that transmission of bovine tuberculosis between lions is mostly unimportant, and that lions typically get infected directly from buffalo. (Veterinary findings support this result and suggest that the usual route of infection is through eating infected buffalo rather than breathing in the disease, as is common for humans.) Further, the model suggests conditions that might make the lion population more susceptible to bovine tuberculosis. Coinfection with another introduced disease that suppresses the immune system is one concern. Another concern is drought, which may increase the transmission of disease from buffalo to lions.
The results of this study suggest that lions have a bit of a reprieve. Bovine tuberculosis is not an immediate threat to the population, although tourists will likely continue to find sick and emaciated lions by roadsides for the foreseeable future. Ultimately, for lions, the question is whether bovine tuberculosis can be eradicated from the buffalo population. The buffalo can’t be all quarantined or killed (as they might be if they were domestic livestock). Instead, the current hope is for a vaccine. Vaccine development is expensive, but could end up being key to protecting buffalo, lions, and other mammals from the disease. This would not only help Kruger’s conservation efforts, but would also help ensure that bovine tuberculosis doesn’t continue to spread further through wildlife in southern Africa or escape from wildlife back into the area’s cattle herds.
This is a plain language summary of the paper:
Kosmala, M., P. Miller, S. Ferreira, P. Funston, D. Keet, C. Packer. (2016). Estimating wildlife disease dynamics in complex systems using approximate Bayesian computation models. Ecological Applications, 26(1): 295-308.