Science, May 30, 2003
In 1859 a rancher decided to introduce European rabbits into Australia so that he could have something to hunt. Before long the rabbits had exploded across the continent, eating so much vegetation that they began to cause serious soil erosion. In the 1950s scientists deployed a biological counteroffensive, myxoma virus, a pathogen from South America. It didn’t eliminate the rabbits, but it did provide grist for an ongoing debate about virulence.
Although myxoma was almost 100% deadly at first, its virulence soon dropped significantly. This fit with a widely held view at the time that all parasites evolve into milder forms as they adapt to the environment. Evolution favors mildness, the argument went, because it allows host and pathogen to enjoy a peaceful coexistence. “If you found a virulent association, you assumed it was recent,” says evolutionary biologist Bruce Levin of Emory University in Atlanta, Georgia. “It was even used as a way of dating parasite-host associations.”
In hindsight, this conventional wisdom seems naïve. Myxoma became milder, but its decline stopped after a few years and it still remains lethal to many of its rabbit hosts. When evolutionary biologists took a hard look in the 1980s, they realized that a pathogen can evolve to become harmless, more deadly, or anything in between, depending on the forces guiding natural selection. Such forces can also pull the pathogen in opposite directions at the same time, creating an evolutionary tradeoff.
Experts began to model the competing forces, and some confidently suggested that the models could serve public health, through “virulence management.” For example, by making it more difficult for a cholera-causing bacterium to be transmitted from one individual to the next, they argued, health programs could take advantage of an evolutionary balance to favor benign strains. Deadly organisms, whose toxic machinery is somewhat burdensome, might compete poorly and yield to milder strains. The notion that virulence could be managed in this way has grown increasingly popular; last year Cambridge University Press published an entire book on the topic.
Now, two prominent evolutionary biologists—James Bull of the University of Texas, Austin, and Dieter Ebert of the University of Fribourg, Switzerland—have raised a provocative dissent. They directly challenge the assumptions that underlie virulence management, arguing that the models used to forecast pathogens’ behavior collapse in the face of the true complexity of diseases, and that the quest to manage the virulence of widespread infestations is doomed to failure. Yet even the skeptics agree that the idea might be useful in planning vaccination campaigns (see sidebar).
Born to be mild
The concept of virulence management emerged from a particular tradeoff in evolution: the one between how fast a pathogen breeds and how easily it can infect new hosts. Evolution favors parasites that can produce lots of offspring, the argument goes, but producing offspring takes a toll on the host. The more the pathogen feeds on its host, spews out toxins, or otherwise causes damage, the more likely it is to kill. If it kills its host before its offspring can get to a new host, all its efforts are for naught. Some biologists predicted that a parasite’s virulence would evolve out of the balance between its competing needs to breed and spread. Changing the balance might change a disease’s deadliness. As it becomes easier for a parasite to infect new hosts, pathogens can afford to evolve into deadlier forms. As transmission gets rarer, gentler strains should take over.
Many evolutionary biologists were charmed by the elegance of the tradeoff model. Bull himself found support for it in experiments with a virus that infects bacteria. “I confess that when I first heard the ideas I bought them hook, line, and sinker,” says Bull. “A lot of people in the field did. The appeal of the original proposal was that you could take this simple concept of a tradeoff and you could apply it to any infectious disease.”
As the tradeoff model gained strength, some evolutionary biologists thought it could become the basis for fighting human diseases. Paul Ewald, a biologist now at the University of Louisville, Kentucky, argued that altering the transmission of diseases could make them less dangerous. Putting up mosquito nets over beds and windows, for example, would make it harder for mosquitoes to carry malaria parasites from an infected person to a new host. The virulent form of malaria, which makes its hosts bedridden and often kills them, would be put at a disadvantage. “You should be able to have the milder strains favored strongly by natural selection,” Ewald predicts.
But some biologists—Bull among them—started to have misgivings. Bull and his colleagues ran new experiments in which bacteria-dwelling viruses evolved under more realistic conditions. They found that faster transmission made the viruses more harmful, as the tradeoff model had predicted, but the difference was so slight that Bull “became much less impressed with the results,” he now says. “I began to think, hmm, this doesn’t work too well.”
Bull last year discovered a kindred spirit in Ebert, who has independently carried out some of the most important experiments on the evolution of virulence. “They’re two giants of the field,” says Michael Hochberg of the University of Montpellier, France. Although Ebert’s results seemed to support the tradeoff model, he had grown disenchanted as well. “We had similar frustrations,” says Ebert. The two decided to attack the tradeoff model in the January issue of Trends in Microbiology. “We’re not pioneers here,” says Bull. “We know lots of other people who feel the same way.”
They claim that most of the support for the model comes from extreme, unnatural conditions. Myxoma, a veritable poster child for virulence, was not in any sort of equilibrium with its host; an imported pathogen, it was matched against a vulnerable population that had never been exposed to it before. As for experiments, the more natural their conditions, the fuzzier their results became. “You always need additional explanations to keep the tradeoff model working,” says Ebert.
Ebert and Bull also point out that many real-world diseases fail to support the model. The Spanish flu epidemic of 1918 broke out in the cramped, foul conditions of World War I in which transmission was easy, killing millions, Ebert and Bull acknowledge. But they wonder why foul, cramped conditions have never triggered another flu epidemic since then. And even when the tradeoff model is relevant, its practical value may be remote. “If I told you I can do something about malaria, but it will take me 10,000 years, you’d tell me to forget about it,” says Ebert.
Ebert and Bull’s challenge has been seconded by other researchers. “It’s long overdue,” says Levin. Marc Lipsitch of Harvard University adds, “They’re quite correct, and that’s why I don’t work in the field anymore.”
The mobility factor
The original architects of the tradeoff model are not impressed by Ebert and Bull’s arguments, however. “Nothing very new in this,” says Roy Anderson of Imperial College in London. Ewald calls it “very sad and dangerous.”
Ewald complains that the critics are trying to demolish a straw man. He says that they leave out a crucial component of his work, for example, the mode by which a disease infects new hosts. If hosts become so sick they can’t move, a parasite can only infect other people who come close, unless a vector such as a mosquito can transport it. This factor is crucial in Ewald’s explanation of Spanish flu. He doesn’t ascribe the deadliness of the epidemic simply to cramped conditions. “That wasn’t my argument,” says Ewald. “My argument was that at the Western Front you had conditions in which people who were completely immobilized could contact hundreds or thousands of people.” Sick soldiers were moved on stretchers to triage areas, then to makeshift hospitals, then onto crowded trains. In these conditions, a flu virus could devastate its host but still infect vast numbers of people. “My argument was that we wouldn’t see a 1918 pandemic arise unless we duplicated this situation which occurred on the Western Front,” says Ewald.
Nor does Ewald think critics have addressed his evidence on cholera. Vibrio cholerae makes people sick by releasing a toxin that triggers diarrhea. As a result, competing organisms get flushed out of the bowels while V. cholerae clings to the intestinal lining. It can then release its offspring into the diarrhea to infect new hosts. The bacteria reach those new hosts by two routes. Untreated sewage or runoff from laundered sheets can contaminate drinking water. Alternatively, an infected person can transmit the bacteria while handling food, shaking hands, or engaging in other social interactions—which generally require a host healthy enough to get out of bed.
Ewald argues that in places with poor sanitation, cholera can make hosts deathly ill but still find new hosts. As a result, it will evolve to high virulence. On the other hand, in places with protected water supplies, that route is cut off. The bacteria’s only option for survival is to let the host move around, which translates into reduced virulence.
Ewald’s observations of the cholera outbreak that struck South America in the 1990s support the hypothesis. In countries with poor sanitation, such as Ecuador, the outbreak was far deadlier than in countries with clean water, such as Chile. Ewald also measured the toxins produced by strains of cholera from different countries. He found that toxin production in Chile dwindled through the 1990s. “In Ecuador, it’s the mirror image of Chile,” says Ewald. “Over a 6-year period, you have only the virulent strains winning out.”
Other experts agree that the critics have not yet made their case. Andrew Read of the University of Edinburgh, U.K., says of Bull and Ebert, “I think they’re reacting to a quite old view. … There was a lot of optimism flung around in the late 1980s and early 1990s. I think the last 10 years have given everybody a feel for the complexity involved and the lack of data, so that nobody that I know of is making wildly optimistic statements. So they’re tilting at a caricature.”
Read nevertheless concedes that evolutionary biologists are a long way from becoming virulence managers: “We don’t know enough about any one disease to be enacting anything now.” Even Ewald grants that it will be a tough hike. Just demonstrating that a change in the transmission of a pathogen can make it less harmful to humans would take a colossal study of thousands of people. In some cases the scale would require “billion-dollar experiments.” For now, Ewald suggests, we may have no choice but to continue studying “natural experiments” to see whether virulent pathogens behave as the theorists have predicted.
Copyright 2003 American Association for the Advancement of Science. Reprinted with permission.