Studies in mice show infection in pregnancy can affect neural development of babies. Credit: Steve Gschmeissner/SPL/Getty

A century ago, a largely forgotten, worldwide epidemic that would kill nearly a million people was beginning to take hold. Labelled as sleepy sickness — or more properly encephalitis lethargica — the disease caused a number of bizarre mental and physical symptoms and frequently left people in a catatonic state, sometimes for decades. (Oliver Sacks described his successful treatment of some of them in 1969, in the book Awakenings.) The cause has never been officially pinned down, but the most common suggestion is that some kind of infectious agent triggered an autoimmune response, which targeted and inflamed part of the brain.

The role of the immune system in mental disorders is subject to much important research at the moment. The onset of conditions from depression and psychosis to obsessive–compulsive disorder has been linked to the abrupt changes in biology and physiology that occur when the body responds to infection, especially in childhood. And some researchers have traced the possible chain of events back a generation. Studies have highlighted that pregnant women could react to infection in a way that influences their baby’s developing brain, which could lead to cognitive and neurodevelopmental problems in the child.

One consequence of this ‘maternal immune activation’ (MIA) in some women could be to increase the risk of autism in their children. And two papers published online this week in Nature (S.Kimet al.Naturehttp://dx.doi.org/10.1038/nature23910;2017 and Y.S. Yimet al.Naturehttp://dx.doi.org/10.1038/nature23909;2017) use animal models to examine how this might happen, as well as suggest some possible strategies to reduce the risk.

Kim et al. looked at the impact of MIA on the brains and behaviour of mice. They found that pregnant female animals exposed to circumstances similar to a viral infection have offspring that are more likely to show atypical behaviour, and they unpick some of the cellular and molecular mechanisms responsible. Some of their results confirm what scientists already suspected: pregnancy changes the female mouse’s immune response, specifically, by turning on the production of a protein called interleukin-17a. But the authors also conducted further experiments that give clues about the mechanisms at work.

It’s tempting to draw parallels with mechanisms that might increase the risk of autism in some people.

The types of bacteria in the mouse’s gut seem to be important. When the scientists used antibiotics to wipe out common gut microorganisms called segmented filamentous bacteria in female mice, this seemed to protect the animals’ babies from the impact of the simulated infection. The offspring of mice given the antibiotic treatment did not show the unusual behaviours, such as reduced sociability and repetitive actions. Segmented filamentous bacteria are known to encourage cells to produce more interleukin-17a, and an accompanying News & Views article (C. M.PowellNaturehttp://dx.doi.org/10.1038/nature24139;2017) discusses one obvious implication: some pregnant women could use diet or drugs to manipulate their gut micro­biome to reduce the risk of harm to their baby if an infection triggers their immune response. Much science still needs to be done before such a course could be recommended — not least further research to confirm and build on these results.

Yim et al. analysed the developing brain of mice born to mothers who showed MIA. They traced the abnormalities to a region called the dysgranular zone of the primary somato-sensory cortex (S1DZ). The authors genetically engineered the mice so that neurons in this region could be activated by light, and they showed that activation of S1DZ induced the same telltale atypical behaviours, even in mice that were born to mothers with no MIA.

It’s unusual to be able to demonstrate such a direct link between the activities of brain regions and specific behaviours — although plenty of work on mental disorders makes a strong theoretical case for linking particular conditions to over- and under-active brain zones and circuitry.

Encephalitis lethargica, for example, has been linked to changes in the deep regions of the basal ganglia, and the disease produces symptoms that are similar to those often seen in autism, including stereotyped and repetitive behaviours. Yim et al.’s study shows that the S1DZ region projects to one of those deep brain regions — the striatum — and that this connection helps to trigger repetitive actions in the animals. But S1DZ also connects to a separate, distinct, region in the cortex, and this is what seems to drive the changes in sociability.

Taking the two studies together, it’s tempting to draw parallels with mechanisms that might increase the risk of autism in some people and explain some of its symptoms. Scientists and others should be cautious about doing so — much can change when results from animal models are applied to human biology. But the studies do offer some intriguing leads.