Manipulative Microbes

Manipulative Microbes

Manipulative Microbes

The invisible invaders that influence guts, brains, and decision making

by Teresa Lee

Credit: Amy Orsborn

We all know how complicated our behavior can be. Despite many psychiatric and biological advances in the past century, humans have never completely understood how external or internal forces influence our behavior—our upbringing, diet, parents, and genes have all been implicated in making us do what we do. Recent studies into the millions of microbes that share our bodies may add several (thousand) more culprits to the list. And even though microbes live inside us animals, only some of them are on our side.

Perfectly tailored


Take the aberrant behavior of an infected ant in the Thai jungle. Normally, a worker ant forages for materials along chemically-determined trails in its colony’s territory. But when spores of the Cordyceps fungus infiltrate an ant’s exoskeleton, the fungus begins to take root, feeding on the ant’s nonessential organs. After a few days, Cordyceps filaments will have grown into the ant’s brain, driving the ant to commit its last act. It will stagger from its nest, climb a nearby shrub, and clamp its mandibles onto a leaf’s vein. This zombie-like behavior is very precise: most infected ants will bite the underside of a leaf about 25 centimeters from the ground on the northwestern side of a plant—a fortunate thing for Cordyceps, because the conditions in this environment are perfect for the formation and release of spores. The fungus will spread through the ant’s corpse to fortify its new home by strengthening the ant’s exoskeleton and producing antibiotics to keep other microbes at bay. Within a few days, a stalk will erupt from the ant’s head to rain spores upon the rest of the ant’s former colony members.

This gruesome interaction is so sophisticated that it seems precisely tailored for these two organisms. And indeed it is: parasites often alter host behavior in ways that might be benign in other organisms, but are particularly detrimental to the host—usually causing the animal’s death. Ants infected by roundworms parade in a way that makes them more likely to be eaten by birds. Infection by wasp larvae alters the patterns a spider will spin into its web. Crickets are driven to watery deaths after becoming host to hairworms. The rabies virus increases aggression and salivation in dogs and other mammals before paralyzing their jaws. In all of these examples, the behavioral changes induced in the host ensure the continued survival of the parasite.

Michael Eisen, a professor of molecular and cell biology at UC Berkeley, is fascinated by the ubiquity of microbial manipulation. “We see examples of parasite manipulation across all taxa—insects, shellfish, even llamas can be infected. But what’s still lacking is an understanding of how microbes can accomplish some pretty sophisticated things. Some of these parasites are only one single cell,” he said. Everyone agrees that behavior manipulation is both fascinating and gruesome, but Eisen believes the most interesting part of these interactions is how finely-tuned each microbe’s response must be to its host. Along with two graduate students in his lab, Eisen uses molecular techniques to uncover the biological underpinnings of microbial manipulation. One of the lab’s projects uses familiar laboratory organisms, while the other addresses the role of Toxoplasma gondii, a parasite that’s received quite a bit of media attention for its putative role in human behavior.

Budding yeast cells are very sticky, latching onto anything that moves their culture. Here, an adult fruit fly is covered in yeast cells expressing green fluorescent protein, after exposure to the cultures in its food plates. Credit: Kelly Schiabor

Do yeast cells lure fruit flies?

The fruit fly Drosophila melanogaster and budding yeast Saccharomyces cerevisiae are two of the most widely studied organisms in biology—they are easy to maintain in the lab, reproduce quickly, and are eminently suited for genetic manipulation. Eisen and his graduate student Kelly Schiabor believe this scientific popularity will make it easier to discover the mechanisms behind behavioral manipulation. In nature, fruit flies are often found near spoiled or fermenting fruit covered in yeast, which is mysterious because they prefer to eat fresh fruit. Schiabor has shown that laboratory flies have a similar preference for yeast cultures taken late in the fermentation cycle. These late cultures have few nutrients left for either yeast or fly larvae, so there is no apparent reason for flies to be attracted to these late cultures.

Schiabor suspects that yeast may secrete chemicals to attract flies under false pretenses. S. cerevisiae cells are particularly sticky. The only hope they have for leaving a nutrient-poor piece of fruit are as stowaways on other animals. But what makes flies land on fruit when its nutrients have already been depleted? Schiabor explains that “these microbes can’t do too much – they’re limited to what chemicals they produce.” And it just so happens that late-fermentation cultures “tend to give off many aromatic, fruity smells.” Flies are lured by the promise behind the scent. Once they land, they realize it’s not a hospitable environment for larvae. By this point, the flies will be covered in yeast cells, which they can deposit elsewhere in the rest of their travels. The timing of this attraction is particularly interesting: yeast cultures tend to secrete these molecules just before the culture is about to run out of food in its current environment.

When the Cordyceps fungus first infects an ant, it sends filaments to feed on the ant’s organs. After infiltrating the ant’s brain, this parasitic fungus creates a zombie ant that leaves its colony to climb a shrub and clamp onto a leaf. Here, Cordyceps has spread to cover the ant’s exoskeleton and extended a stalk from the ant’s head, which can release spores onto the forest floor below. Credit: Pennsylvania State University

Since many molecular details are understood in both flies and yeast, Schiabor believes they represent a perfect system for understanding how behavior can be manipulated. “Behavior is one of the last frontiers,” she said. “I do believe that there’s biology behind behavior that’s the same from individual to individual.” But there are many other factors that can affect even the simplest behavioral circuit—the variation can seem endless and therefore hopeless to untangle. This is what excites Schiabor about her system. “If we can figure out how or why a yeast cell can manipulate a fly, think of what that might mean for our understanding of interactions between other species.”

Making mice forget their fears

One of the challenges Schiabor faces is finding out whether yeast cells intentionally attract flies. Do they produce fruity-smelling molecules because they want the flies to investigate their culture, or are these simply by-products of their natural processes? Evidence of potential manipulation becomes more apparent when the desires of the parasite are directly at odds with those of its host. Wendy Ingram, another graduate student in Eisen’s group, is studying a system in which this is exactly what happens. Toxoplasma gondii (Toxo for short) is a parasite that reproduces in the gut of its primary host, the cat. Infected cats shed fertilized Toxo eggs in their feces, which can infect mammals or birds that accidentally eat an egg. Toxo can spend years in a host, hidden in a latent stage, but it will eventually head to the brain of its host. There, it enters neurons and waits. Toxo can only sexually reproduce in cat intestines, so it needs to wait until its host is eaten by a cat. But it is not at the mercy of fate—by lodging itself in the brain, Toxo may be able to encourage the premature predation of its host.

Click to enlarge. Design: Amy Orsborn, photos: CDC (1), John Boothroyd (2), Carolina Caffaro (3), Yale Rosen (4)

This seems like a long shot, but researchers have discovered this actually does occur when Toxo infects rodents. “In theory,” said Ingram, “Toxo can get a rodent to walk right up to a cat and essentially deliver its next meal.” Uninfected rats have a strong and innate fear of cat smells. When rats are infected with Toxo, they seem to lose their natural caution around predators. In the 1990s, Professor Joanne Webster, an epidemiologist at Imperial College London, was the first to scientifically characterize this behavior. She found that not only did infected rats lose their natural aversion to cat odor, but they may even be attracted to the smell of cat urine.

Ingram studies the effects of Toxo infection in mice, which have the benefit of a sequenced genome in addition to well-understood behavioral assays. She first demonstrated that infected mice act the same as infected rats. Using urine from an assortment of felines (bobcat, tiger, mountain lion) and other animals (rabbit, hyena), she quantified the different behaviors of infected and uninfected mice. In a cage with a cup of any type of feline urine at one end, an uninfected mouse spends most of its time cowering at the farthest edge of the cage. But infected mice will roam over the entire cage, completely disregarding the presence of cat urine. This is the same response that all mice—infected and uninfected—have to rabbit or hyena urine, as these animals do not generally eat mice.

Combining the many genetic tools available to mouse researchers with some newly developed genetic assays in Toxo, Ingram hopes to uncover how Toxo can negate a mouse’s potent aversion to predators. “There are clearly things about biology that Toxo understands better than we do,” she said. Toxo has already proven to be a wily parasite—it can trick the immune system to slip past its host’s blood-brain barrier, which is usually impenetrable to most infecting agents. Last year, Webster and Glenn McConkey, a parasitologist at the University of Leeds, discovered that Toxo has several genes that affect the production of dopamine, a common neurotransmitter involved in signaling pleasure, motivation, and fear. Toxo may also be able to inject molecules into the nuclei of its host’s cells, which could allow it to hijack whether certain genes are turned on or off in those cells.

Does Toxo affect us?

Credit: Amy Orsborn

It’s natural to wonder whether humans are subject to this eerie manipulation. Many of us have been exposed to Toxo—in some countries, more than half the population has been infected, primarily from eating undercooked meat. Although humans are not a natural part of Toxo’s life cycle, we share many genes and molecular pathways with rodents. This makes them useful models for human disease, but it may also leave us susceptible to accidental manipulation by the parasite. If Toxo does affect dopamine production, it could very well play a role in our behavior. Dopamine signaling is involved in many processes, including the powerful neural circuits of fear, pleasure, anxiety, and arousal.

Jaroslav Flegr, an evolutionary biologist at Charles University in Prague, believes he’s found evidence of Toxo infection affecting human behavior. He conducted personality studies linking infection in humans to an unlikely variety of behavioral traits—introversion in men, image-consciousness in women, aggression, likelihood of committing suicide, and even attractiveness to the other sex. In 2011, he published a separate study that linked brain damage in schizophrenia patients to Toxo infection. Using MRI machines to scan patients’ brains, he found that those infected with Toxo had more reduction in brain tissue than uninfected patients. Non-schizophrenics had no significant difference in brain tissue, regardless of their infection status. Of course, many environmental and genetic factors have been associated with schizophrenia, which complicates any role that infection might play in understanding the causes of this disorder.

Another of Flegr’s studies from 2002 found that people infected by Toxo are at least two times more likely to be involved in a traffic accident. When I asked Ingram about the possibility of Toxo manipulating human behavior, she rolled her eyes and smiled. “What people don’t mention is that Flegr’s study didn’t consider whether the infected person was driving the car, or just a passenger,” she pointed out. Most of Flegr’s studies have only correlated Toxo infection with behavioral changes and correlation, as the mantra goes, does not equal causation. Ingram notes that many of the behaviors associated with infection might actually make someone more likely to be infected by Toxo, rather than being caused by infection itself. “We are so far from understanding how behavior works in infected rodents, which are a pretty well-understood system, let alone even beginning to comprehend what happens in infected humans.” She believes that her work on the mechanism behind Toxo’s manipulation of mouse behavior may help to uncover whether it affects human behavior, but doesn’t think the changes will be as straightforward as Flegr’s studies claim. “I respect the fact that there are microbes that want to use us for their own benefit,” she said. It is very possible that this may involve changes in our behavior. “The interactions between microbes, our immune system, and our brain are so complex. There’s a lot going on we don’t even realize.”

What lies within

The guts of human infants are colonized by microbes during birth, but the community of microbes can change drastically over time. Shown is a graphical representation of the microbial community in the gut of an infant girl studied by the Banfield group. Each square represents a sample taken on the specified day, with colors denoting the twenty most dominant bacterial groups. Researchers noticed periods of relative stability in the microbial community, where one or two groups would be present at much higher numbers than other groups. These were punctuated by abrupt transitions, in which another bacterial group would assume dominance. It’s interesting to note that changes to the infant’s diet after days 9 and 14 were followed by changes to the population of her gut microbiome. Credit: Amy Orsborn

Although we lack conclusive evidence that Toxo infection affects our behavior, there are still many other likely manipulators that live much closer to home. Humans are host to a huge collection of bacteria and other microbes that are completely integrated into our body’s functions: our microbiome. We like to think of ourselves as ourselves, but we could accurately be described as a mobile scaffold for our microbiome, which technically outnumbers us. Collectively, our microbiome has ten times more cells than our bodies, with perhaps a hundred times more genes than exist in our genome. In the past few decades, researchers have come to believe the relationship between our microbiome and our bodies doesn’t just benefit the microbes: it is mutually beneficial, and vital to our health. The bacteria in our intestines make our digestion more efficient, help develop our immune systems, manufacture vitamins, assist in fat storage, and prevent the growth of pathogenic bacteria. Recent work also suggests that our gut microbiome can influence us in more subtle ways, affecting neural development, normal brain function, and even our emotional state.

It’s easy to dismiss how essential microbes are to our health. They’re too small to see, and many of the bacteria that live on our bodies have not been cultured or observed in a laboratory. We pay much more attention to the pathogens that harm us than to the beneficial microbes that help us.

To better understand our coexistence with this community of microbes, the National Institutes of Health launched the Human Microbiome Project, a five-year effort to characterize the microorganisms in healthy and diseased humans. Jillian Banfield, a professor of earth and planetary science who studies geological microbiomes (see “Germ Warfare,” 38), has become interested in the process of bacterial colonization of the human gut. Unlike our genomes, which we inherit from our parents, we do not come with a predetermined gut microbiome. As infants, we accumulate bacteria from our mothers during birth and acquire more diversity in the first year of our lives. Our gut bacteria play an important role in the development of our immune system, helping it learn which bacteria are friendly and which are harmful.

In the past decade, drastic advances in DNA sequencing have allowed us to take a closer look at our microbiome. Because its diversity is so complex, it would be impractical to examine bacteria on a species-by-species basis. Banfield’s group takes a metagenomic approach to circumvent this difficulty. Metagenomics allows researchers to identify bacterial species in a sample without knowing beforehand which species were present. Researchers can sequence DNA isolated directly from its environment—in this case, infant feces—instead of relying on bacterial cultures grown in a lab. Itai Sharon, a postdoctoral researcher in the Banfield lab, explained how this works. “We isolated the genomes of the entire bacterial population from infant guts and sequenced the same gene—16S ribosomal RNA, which is present in all species—to see how many differences exist in the population.” By comparing these DNA sequences to the ribosomal RNA of known species, the researchers can infer how many types of bacteria are present in the microbial population. Using the number and extent of differences in DNA sequence, they can also identify whether different strains of the same species are present. This is an important distinction: while one strain of bacteria might be helpful, another strain of the same species could be pathogenic.

The Banfield group has now examined the microbiome of two infants, taking fecal samples daily during the first few weeks of the infants’ lives. The interactions between different microbes have proven to be more interdependent and dynamic than they previously thought. A strain of bacteria will achieve dominance in the population only to cede it to a different strain a few days later. Their work, along with other studies reported by the Human Microbiome Project, has enumerated the sheer diversity of our microbial passengers. Our guts carry many beneficial strains, but researchers also found a significant amount of pathogenic bacteria in healthy human guts. “Our bodies are just another complex ecological system,” Sharon says. As in any ecological environment, he believes that the diversity in our microbiomes will be important for our overall health, rather than which species is dominant at a given time. The Banfield group and others involved in the Human Microbiome Project would like to classify which bacterial species make up a “normal” microbiome in healthy humans. Knowing what a healthy microbiome looks like may help us understand how variation in bacterial populations could lead to poor health, both physical and mental.

How deep does this interaction go?

Click to enlarge. Credit: Journal of Cell Biology (left), Amy Orsborn (right)

Ads for yogurt or kombucha products tout the benefits of probiotic supplements, so it’s fairly common knowledge that the contents of our gut microbiome factor into daily health. But mammals depend on a healthy microbiome for normal development long before they begin eating solid foods. A study performed in infant mice at the Karolinska Institute in Sweden has demonstrated that mammalian brain development requires exposure to gut bacteria at crucial stages. Mice raised in completely sterile conditions are more active and react differently to environmental stimuli than control mice, which researchers exposed to specific gut microbes. These differences dwindle when germ-free mouse pups are exposed to gut bacteria, but only if they receive this exposure within a certain period in their lives. By the time germ-free mice reach adulthood, exposure to gut bacteria cannot make them act more like the control mice. It seems that microbial interaction has been important for so much of mammalian evolution that even the proper wiring of a mouse brain has come to rely on bacterial colonization of the gut.

Gut bacteria can also affect the brains of fully developed animals. In 2011, researchers led by Professor John Cyran at University College Cork in Ireland found that mice that ate Lactobacillus—a common bacteria found in yogurt—displayed fewer behaviors associated with anxiety and depression. As opposed to the artificially reared germ-free mice (like those in the Karolinska study), these mice were normal healthy adults that came with a full complement of their own gut microbiome. Lactobacillus ingestion caused these mice to have lower levels of stress hormones in stressful situations and higher levels of GABA, a neurotransmitter involved in anxiety and stress responses. But how can bacteria residing in the gut generate relatively immediate changes in brain chemistry? Cryan’s group suspected that a likely conduit was the vagus nerve, a meandering nerve that carries information to the brain from the intestine and other internal organs. And in the intestines, only a thin layer of epithelial cells separates nerve endings from bacteria and whatever molecules they produce. As they theorized, when Cyran’s group severed the vagus nerve in mice, these animals no longer displayed the anxiety-reducing effects of eating Lactobacillus.

If bacteria are involved in determining our mental state, any changes to our microbiomes would have profound ramifications. Pathogens make their presence felt, often in unpleasant ways, and we have developed effective antibiotics to fight them. But these antibiotics also decimate thousands of helpful bacteria in our bodies. One of the infants studied in the Banfield lab took a round of antibiotics during sample collection. There were substantial changes in the composition of her gut microbiome, which is similar to what has been observed in adults. “After antibiotics, your microbiome never returns to the way it was before,” Sharon explained. “We can’t say yet whether it is necessarily better or worse than before antibiotics, but it is definitely changed forever.” Probiotics claim to replenish our gut microbiome, but these consist of only a few bacterial species, which cannot replace the previously thriving and complex ecosystem of our microbiome. One of the hopes of the Human Microbiome Project is to understand how to harness our microbial ecosystem to fight disease. This approach might be able to target pathogens and reduce the side effects of broad-spectrum antibiotics, which are usually a crude solution to specific and identifiable problems. Since beginning this research, Sharon has assumed a different outlook on the role that bacteria play in our lives. “I’m not really worried anymore when my kids play in dirt,” he said.

“Not complete puppets”

In nature, many examples of behavior modification by microbes are nefarious, causing zombification or death of the host animal. For humans, the data don’t support these drastic effects, but research into microbes may uncover some mysterious influences behind our behavior. We like to think that we control our environment, but in truth, we are as much at the mercy of our environment as any other creature. Although it’s easy to forget, our environment consists of untold numbers of microbes. It seems like a stretch to blame our behaviors on their influence, but it isn’t implausible to think they affect our brain chemistry. “At the end of the day, we’re made up of a bunch of cells, responding in pretty predictable ways,” said Schiabor. Signaling pathways are often shared between living things, and molecules produced by bacteria can, and do, interact with our cells.

Every animal exists in an ecological context, and humans are no exception: we have evolved along with our microbiome, and these bacteria have become essential to our functioning, while eking out a cozier place for themselves. The environment shapes our development as much as our genes do, and as the Human Microbiome Project makes clear, the influence of the environment might be more proximal than we previously thought. Maybe it helps to think of it as Ingram does: “We’re not complete puppets, but we do have some strings that microbes can use to pull us in different directions.”

Type to Search

See all results