Toxoplasma, Cat Piss and Mouse Brains: my lab’s first paper on microbial manipulation of animal behavior

All animals live in a microbe rich environment, with immense numbers of bacteria, archaea, fungi and other eukaryotic microbes living in, on and around them. For some of these microbes, the association is transitory and unimportant, but many make animals their permanent home, or interact with them in ways that are vital for their survival. Many members of an animal’s “microbiome” are affected by, and often become dependent on, aspects of the animal’s behavior. And, as microbes will do, some – and we believe many – of these microbes have evolved specific ways to manipulate the behavior of their animal neighbors to their advantage.

My lab has begun to study several such systems, seeking to discover the molecular mechanisms that underly these fascinating microbial adaptations – none of the several dozen cases in which microbial manipulation of animal behavior has been documented are understood in molecular detail.

One of these systems involves the eukaryotic parasite Toxoplasma gondii which reproduces clonally in most (if not all) warm blooded animals, but – for unknown reasons – only reproduces sexually in the digestive system of cats. Thus, in order to complete the Toxo lifecycle, an infected animal has to be eaten by a cat. This creates a conflict of interest between Toxo, who wants its host to be eaten by a cat, and the host, who would rather NOT be eaten by a cat. Indeed this “I don’t want to be eaten by a cat” effect is so strong, that many animals have evolved an innate fear of all things cat – especially their smells.

For example, if you take a laboratory mouse and put him (for a variety of reasons we usually do these experiments with males) in a box with a bowl or water they will largely ignore it. Swap out the water and put in something that the mouse has no reason to fear – like rabbit urine – and they still more or less ignore it. Swap that out and put in cat urine, and it’s a whole different ball game – the mouse spend most of its time on the other side of the cage.

Amazingly, it seems that rodents infected with Toxo lose this innate fear of cats – possibly as a result some property Toxo has evolved to increase the likelihood that it will end up in a cat’s tummy. Several papers have come out on the topic in recent years (from Robert Sapolsky’s lab at Stanford as well as others), but the molecular mechanism is unknown.

A graduate student – Wendy Ingram – became obsessed with this phenomena and is pursuing it as a joint project between my lab and that of Ellen Robey (an immunologist who studies the host response to Toxo infection and the ways the parasite evades immune surveillance).  Wendy has begun a bunch of experiments to examine this phenomenon – she is interested, in particular, in the role the immune system might play in mediating this response. Her first wave of experiments is done, and we have posted a preprint of a paper describing them on the arxiv.

Wendy first showed that the behavioral effect is robust, and is general across Toxo (previous experiments had used only one of the three major North American variants of Toxo – Wendy showed the same effect in the other two subtypes. But more interestingly, Wendy found that the effect was strong and persists for months in an attenuated Toxo strain that – unlike the other strains we and others have examined – is not detectable in the brains of infected animals after a few weeks. This would seem to refute – or at least make less likely – models in which the behavior effects is the result of direct physical action of parasites on specific parts of the brain. It’s just a start in trying to dissect a complicated phenomenon, but Wendy has a whole slew of followup experiments under way or in planning that should shed more light on what aspects of the innate fear response are being overridden and what, if any, role the immune system is playing.

As always, we welcome your thoughts and comments on the paper, released here as part of our commitment to make preprints of all of our lab’s papers available as soon as (if not before) we’re ready to submit them to a journal, and to make them available here for open peer review.


This entry was posted in EisenLab preprints, microbial manipulation of animal behavior, My lab. Bookmark the permalink. Both comments and trackbacks are currently closed.


  1. Joe Pickrell
    Posted April 3, 2013 at 6:42 am | Permalink

    Any variation across mouse strains in aversion to cat urine or response to Toxoplasma? I.e. is this something you can do genetics on?

    • Wendy
      Posted April 24, 2013 at 12:57 pm | Permalink

      Different mouse strains range in susceptibility to infection which unfortunately complicates the interpretation of innate fear and loss of fear following infection.

  2. DrugMonkey
    Posted April 3, 2013 at 7:09 am | Permalink

    This stuff is absolutely fascinating. When do you get to the part explaining those people who own a dozen cats?

    • Ileana
      Posted April 26, 2013 at 8:02 pm | Permalink

      I couldn’t quite tell from this post or from the National Geographic article that led me here, whether there’s an antibody test for all strains of T. gondii. If there is, the correlation, if any, between the parasite and the phenomenon of cat hoarding should be easy to study. If there isn’t and we have to examine brain tissue, then we can only study cat hoarders after death.

      • Ileana
        Posted April 26, 2013 at 8:12 pm | Permalink

        Of course this would only show correlation and not causation.

  3. Posted April 3, 2013 at 10:27 am | Permalink

    Serotonin is also at play in this system.

    Aside from the well-discussed dopamine production, there is also a lesser known effect – the interferon-gamma mediated expression of indoleamine 2-3 dioxygenase (IDO) PMID 7897249 and tryptophanyl-tRNA synthetase P.NAR which act in concert to deplete the mammalian cell of free tryptophan (yet simultaneously saving it for protein synthesis), This is known to be effective in starving toxoplasma PMID 15206721., a Trp auxotroph, and inhibiting its growth. Trp depletion is also invoved in inhibiting trypanosomiasis PMID 10721096

    This effect may have a downstream effect on lowering serotonin, by Trp substrate depletion. Or the strange downstream metabolites of IDO – Trp breakdown products – may be involved as they accumulate, possibly inhibiting serotonin receptors.

    In any event, there are a lot of small molecule changes at play, even at early infection, that will need biochemical assays to fully untangle.

    Try repeating with uninfected rats fed a low Trp diet, and see how they respond, or trigger the interferon-gamma response without that particular parasite present.

    • Alex
      Posted April 4, 2013 at 6:17 am | Permalink

      I don’t think the purpose of this work was to describe any kind molecular mechanism from the host or the parasites perspective. Instead it was to provide data on a new model of toxoplasma invasion – one that links modulation of the host fear response with it being independent of cyst formation in the brain.

      Although you bring up an interesting point, it is one that would warrant another thesis. The authors are suggestive of molecular mechanisms of host manipulation, however, that is a postulate and not the main focus.

  4. Posted April 5, 2013 at 11:22 am | Permalink


    biggest question for me is whether this is a general loss of fear or an olfactory processing deficit. Are the animals braver, less scared, or anosmic..? How do they perform on an elevated T-maze? Do they have problems finding food? Are mothers good at finding displaced pups?

    figure 1B suggests that even infected animals stick to the corners of the box — they may still be anxious. It’d be neat if T. gondii is specifically screwing up the animal’s ability to smell predatory odors. Suggests brain region-specific mode of action.

    • Richard H
      Posted April 7, 2013 at 10:30 am | Permalink

      Vyas et al (2007) make some strong claims about specificity and aversion being turned into attraction but their n’s are low (as low as 5 in some parts of the studies reported), some effects are only just noticeable even though statistically significant and they switch between species and sexes without apparent regard

      Gonzalez et al (2007) say that rats spend more time on open arm of plus maze and more time investigating in social interaction tests – ergo less anxious

      Anosmia, generalised reduction in fear – a combination of both? – Interesting work but definitely a situation where Occam’s razor needs to be applied with great care …

      Gonzalez, L. E., Rojnik, B., Urrea, F., Urdaneta, H., Petrosino, P., Colasante, C., … & Hernandez, L. (2007). Toxoplasma gondii infection lower anxiety as measured in the plus-maze and social interaction tests in rats: A behavioral analysis. Behavioural brain research, 177(1), 70-79.

      Vyas, A., Kim, S. K., Giacomini, N., Boothroyd, J. C., & Sapolsky, R. M. (2007). Behavioral changes induced by Toxoplasma infection of rodents are highly specific to aversion of cat odors. Proceedings of the National Academy of Sciences, 104(15), 6442-6447.

  5. Richard H
    Posted April 7, 2013 at 9:38 am | Permalink

    Just wondering how you discount the possibility that the animals sense of smell is being disrupted – since that effect alone could account for the finding that infected mice no longer find bobcat urine aversive – i.e. they just can’t smell it any more?

    • Ed Yong
      Posted April 10, 2013 at 7:14 am | Permalink

      No time to track down the ref for this but I’m pretty sure that they still maintain aversion to dog piss. It’s a specific aversion to cat piss.

      • Posted April 10, 2013 at 7:40 am | Permalink

        We get inconsistent results for urine of other carnivores – like mink – though we haven’t done enough trials to be sure of what we’re seeing. Our next set of experiments will look at this more completely, and will include comparisons of several cats (e.g. lions, tigers) as well as non-feline members of the carnovora (e.g. bears). With lions and tigers and bears – lions and tigers and bears – we should have a clear answer.

    • Posted April 10, 2013 at 7:38 am | Permalink

      We’ve done experiments (as have others) showing that their senses of smell and taste are intact. For example, we do something called a “hidden cookie test” (don’t you love that name?) in which we test the time it takes for them to find a cookie buried in bedding. The infected mice do just as well as the uninfected ones.

  6. Matt
    Posted April 16, 2013 at 9:07 pm | Permalink

    Fascinating! Have you tried to isolate the fear-inducing constituent in cat pee? I smell a Science paper, a great tool compound, and a lifetime of humane rat deterrent royalties for you if successful.

    Now for a favorite example of potentially olfactory behavioral control in humans, for your NIH proposals:

    • Ileana
      Posted April 26, 2013 at 8:07 pm | Permalink

      Are you suggesting cat urine (or some constituent of it) as a rodent deterrent? It would obviously only be useful (and marketable) if it didn’t also bring with it the smell.

  7. Ajai Vyas
    Posted June 8, 2014 at 6:56 pm | Permalink

    Very nice paper. It says “attenuated Toxo strain” cleared from mice brain after a few weeks. Was parasite still present in other immune-privileged organs like muscles or testes?

One Trackback