A paper published this month in Stem Cell Reports by Le Belle et al. out of UCLA suggests that maternal inflammation during pregnancy, such as occurs during an acute illness or a chronic disease, may influence autism risk. The researchers used a mouse model, exposing pregnant mice to lipopolysaccharide (LPS), a molecule usually present on the surfaces of Gram-negative bacteria that elicits a strong immune response in animals, in order to study the effects on neuronal proliferation and mouse behavior. They combined this with an additional genetic model for autism risk, the Pten knockout mouse, in order to study the effects of a multi-hit scenario.

Illustration of the cell wall, containing membrane-bound LPS, of a gram-negative bacterium. Borrowed from here.

After exposing the mice to LPS nine days post-conception, which is the time period just following neural tube closure and the beginning of neural stem cell proliferation underlying the brain, the researchers reported finding a mild postnatal brain overgrowth. At birth, overall body size was smaller in the exposed pups, meanwhile the ratio between brain and body size was higher than in controls by about 15%. This finding remained consistent by postnatal day 20 (P20) at which point approximately 25% of exposed mice displayed megalencephaly, although this effect was lost by P60. Absolute brain size at birth however didn’t significantly vary between the groups. Interestingly, there were no sex differences that were apparent.

At the microscopic level, they observed increases in cortical thickness and area in the LPS-exposed mice but no obvious changes in the thickness of the corpus callosum. In addition, at P0, the number of proliferating Nestin-positive neural stem cells in the subventricular zone was significantly increased compared to control and when cultured, showed signs of prolonged self-renewal (multipotency). Numbers of microglia, the innate immune cells of the CNS, were also increased throughout all time periods sampled.

Brain weight of heterozygous Pten knockout pups at birth was only approximately 8% larger than that of their normal littermates. However, in Pten‘s exposed to LPS prenatally, brain weights were exponentially greater, at 34-58% that of controls. This highlights the potential effect of multiple hits on a developing system, suggesting that even for “genetic” cases of autism, these genes certainly influence but may not determine the development of the condition. Instead, they may be exponentially additive.

Illustration of a simplified version of the Akt pathway, involving Pten (circled in red) which acts as a negative regulator (suppressor) of Akt pathway activation and thus cell growth, proliferation, and differentiation. Borrowed from here.

On behavioral testing, all exposed animals displayed disturbances in vocalizations in infancy and socialization, anxiety, and repetitive grooming in adulthood, the classic tests used to diagnose “autism” in a mouse.

Because redox mechanisms are upregulated during cellular stress, including inflammation, the researchers went on to attempt to target and rescue exposed mice by antagonizing the reactive oxygen species (ROS)-generating enzyme, NOX, thereby depressing ROS signaling. NOX inhibition resulted in a reduction of brain overgrowth in LPS-exposed mice, and actually produced a brain undergrowth in some pups. Microglial numbers were also reduced, as was neural stem cell self-renewal. Finally, NOX inhibition rescued the excessive grooming displayed by older exposed mice, but it failed to rescue the vocalization deficit in infancy for reasons unknown. The researchers therefore hypothesized that most of the effects seen in this study were ROS-dependent.

On that point I partially agree: while the findings were clearly influenced by ROS availability, I would hesitate to focus primarily on reduction-oxidation mechanisms, even though redox is a hot topic in the autism/mitochondrial literature. Instead, because cells under stress react with a combination of inflammatory, redox, and other intermediate metabolites (a stress reaction known as “hormesis”), it would probably be better to view these results more broadly. Redox mechanisms are one important branch of hormesis which a cell uses to stimulate self-repair following an injury such as infection. Therefore, targeting and inhibiting other branches of hormesis may also produce similar results as those seen in this study.

Overall, however, it was an interesting study and gets at several important points:

1. that maternal inflammation during pregnancy, due perhaps to an acute infection such a rubella or even a chronic inflammatory process such as an autoimmune disorder, could influence the development of autism;
2. a multi-hit model of autism and other neurodevelopmental conditions may more accurately reflect the causes of a broader array of cases;
3. hormesis or the cellular stress reaction in general may positively influence the development of autism.