In May of 2000, heavy rains pummeled Walkerton, Ontario, a town of around five thousand people located a hundred miles northwest of Toronto. Excrement from cattle containing the bacteria E. coli and campylobacter washed into the municipal water supply. An epidemic of bacterial dysentery ensued; nearly half the residents became ill and seven people died.
Two years later, Canadian researchers began surveying the exposed population. One question they asked was: Since the outbreak, have you been told for the first time by a doctor or a health-care professional that you’ve developed depression, anxiety disorder, panic disorder, or post-traumatic-stress disorder? Residents who had diarrhea after being exposed to the contaminated water were more likely to answer yes than those who hadn’t had any gastrointestinal symptoms.
This association could simply represent a natural response to a serious illness. It’s often the case that if you get sick one year, you get depressed the next. Or, maybe, people who suffered gastrointestinal distress were inclined to respond affirmatively about any symptom—particularly given their pending lawsuits against the city. The researchers dealt with this potential bias by inserting a control question on ear-buzzing, for which the rates of the two groups were the same. This allowed the researchers to raise a third possibility: Could the bacteria itself cause depression?
A lot of public and scientific attention has been paid recently to the idea that the microbiome—the collection of bacteria, viruses, fungi, and other microbes that share our bodies, outnumbering our own cells ten to one—can cause diseases widely conceptualized as non-communicable. According to well-designed, peer-reviewed studies on rodents and humans, the microbiome appears to be a major contributor to obesity, diabetes, atherosclerosis, malnutrition, hypertension, asthma, rheumatoid arthritis, colon cancer, ulcers, inflammatory bowel disease, lymphoma, liver cancer, psoriasis, and even ear wax. We are, in many ways, a result of the organisms that live inside us. (Michael Specter wrote a feature on the microbiome in the magazine last year.)
But the link between the microbiome and how we feel and behave seems far more tenuous, if only because diseases of the mind are influenced by so many factors, and often elude clear biological pathways. A number of elegant studies, however, suggest that the microbiome may have as many implications for our brains and behavior as it does for more easily defined diseases.
At a recent National Institutes of Health conference on the topic, Ted Dinan, an avuncular, scholarly psychiatrist from Cork, Ireland, explained one way that bacteria in our gut could alter our behavior. Many organisms are capable of making neurotransmitters such as norepinephrine; nerves need to communicate with each other, and neurotransmitters serve as the key facilitators of this communication. Many people are familiar with the neurotransmitter serotonin, for instance, because it is targeted by widely used antidepressants, like Prozac. What many don’t realize, however, is that gut bacteria are actually the body’s major producer of serotonin.
Dinan’s own work relies heavily on observations of mice. Mice don’t have the same personality variations as people, but they do exhibit fear, anxiety, risk-taking, and depression, among other traits. Dinan and his colleagues, for example, have performed a so-called forced-swim test, where mice are made to swim along a cylinder. They will pause periodically and become immobile; this immobility is considered a sign of hopelessness. Reporting in the Proceedings of the National Academy of Science, Dinan and colleagues found that feeding mice lactobacillus, a bacteria commonly found in yogurt, resulted in shorter immobility times, suggesting less hopelessness. When they severed the vagus nerve, which transmits signals from the gut to the brain, the mice became hopeless again.
Other studies that used mice raised in germ-free facilities and fed only sterilized food and water have convincingly shown that microbes in the bowel affect the emotional states, and consequent behaviors, of the specimens. A commentary by Elizabeth Pennisi published earlier this year in Science noted that germ-free mice are hyperactive, and that this behavior was reversed if the mice were colonized with specific bacteria introduced at a critical age.
Microbes, it seems, also make mice fat, and recent data suggest that this association depends, at least in part, on the microbiome’s influence on the animals’ behavior. The genes of the innate immune system create a defense similar to the Coast Guard, vigilantly protecting the body from microbial invasion; if these immune-system genes are altered, the microbes that survive in the colon also change. Matam Vijay-Kumar, a scientist at Emory who studies the immune system, made the surprising observation that one such gene-altered mouse developed obesity and diabetes. When he gave similarly altered mice antibiotics, their metabolic abnormalities disappeared. And when he placed normal, germ-free mice in the same cage as mice with the altered innate immune system, the previously healthy mice became obese. It seemed that the tendency to become obese and diabetic was literally contagious.
Why would the microbiome have this effect? There are several possible underlying mechanisms: the bacteria might make cells resistant to insulin, for instance, or they could extract calories from undigested food, thus adding to the organism’s weight. Intriguingly, however, the link between bacteria and weight-gain was largely mediated by behavior—the mice with the altered microbiome ate significantly more. We are a long way from treating obesity with a dose of penicillin, but the possibility that our microbiomes can determine how hungry we feel, or how poorly we cope, certainly warrants further investigation.
As intriguing as this early evidence may be, targeting the microbiome in order to treat mental illness or change stubborn behaviors will be complicated by the fact that the role of bacteria in human behavior may begin long before any signs of illness appear. We are host to a hundred trillion or so bacteria, and every portion of the bowel represents a different microbial ecosystem. Organisms that are present when we’re two months old may have shaped our brain, but they have long since disappeared when we hit twenty or forty or sixty. Indeed, while a recent summary in the journal JAMA Pediatrics suggests that bowel bacteria may provide insight into “autism, schizophrenia and anxiety,” the authors also emphasize the role that timing plays in the microbiome’s influence over the developing brain. Designing interventions, then, will depend not only on identifying the bacterial links to diseases but on understanding when such an influence occurs.
Identifying which bacteria are critical and which are bystanders, which strains affect which behaviors, by which pathways, in whom and when, is a formidable task. But we now have reason to believe that it is not impossible. The Walkerton epidemic is a milestone, because it provides human epidemiological data to bear on the association between the microbiome and personality change. And, it has biological plausibility because of earlier basic science research in mice. Fifteen years ago, Mark Lyte and his colleagues, a microbiology team from the Minneapolis Medical Research Foundation, studied the effect of infecting mice with campylobacter, one of the bacteria implicated in the Walkerton epidemic. The dose of bacteria was high enough to be detected in the intestine, but not so high that the mice developed overt illness. You probably won’t be surprised to learn that the campylobacter-infected mice exhibited more anxiety when navigating a maze than the control mice.
Because such mass bacterial exposures are rare, gathering human evidence of the mind-microbial connection will take time and ingenuity. It’s too soon to know whether a cleanse might contribute to your insomnia, or whether swallowing dirt will keep you lean. But next time you are served a delicious meal and fall asleep before doing the dishes, it wouldn’t hurt to try a new excuse: “The E. coli made me do it.”
Correction: Walkerton is northwest of Toronto, not northeast.
James T. Rosenbaum is a professor of ophthalmology, medicine, and cell biology at Oregon Health & Science University, where he holds the Edward E. Rosenbaum Chair for inflammation research, and the chief of ophthalmology at the Legacy Devers Eye Clinic, in Portland, Oregon, where he holds the Chenoweth Chair.
Photograph: Zoran Kolundzija