Researchers urge caution in studies of mice and microbes

Over the past several years, researchers have begun to investigate links between the trillions of bacteria that inhabit our digestive tract and autism. These studies reflect the undeniable presence of gastrointestinal symptoms in autism, as well as a growing appreciation for the so-called microbiome and its role in overall health.

A new review raises caution, however, about using mice to model the human microbiome. Published in this month’s issue of Disease Models & Mechanisms, it serves as an important reminder for autism researchers to interpret such studies with skepticism.

One reason for this caution comes down to anatomy. Compared with people, mice have a proportionally larger large intestine and cecum — the pouch at the beginning of the large intestine where bacteria ferment undigestible plant material. These differences reflect different diets: Although mice, like humans, are omnivores, they eat a greater proportion of plants.

The two organisms also have different distributions of goblet cells and Paneth cells in the gut wall. These cells help to coordinate immune responses in the gut and so are likely to influence the composition of the microbial community.

The types of resident bacteria also vary between people and mice. Both microbiomes are dominated by two major bacterial groups: bacteroidetes and firmicutes. But their precise composition is difficult to compare because researchers have usually used different methods to study them.

There are at least 79 subgroups, or genera, of bacteria that exist in both mice and people. However, the relative abundance of genera may differ between the two. For example, Prevotella appears to be abundant in the human gut but rare in the mouse; the opposite is true of Lactobacillus.

Despite these differences, mouse models have helped to uncover the interrelationships between diet, the microbiome and obesity in people. Dietary manipulations in mice produce microbiome changes similar to those seen in people, for example.

On the other hand, researchers have struggled to create mouse models of inflammatory bowel disease that accurately reflect the human condition. Some microbiome changes are similar between the two organisms but others are puzzlingly inconsistent. For example, a bacterium called Akkermansia is scarce in people with inflammatory bowel disease but more abundant than normal in one mouse model of the disease.

No research tool is perfect, and the idea that mice do not always accurately model a human condition is hardly news to autism researchers. After all, the field continues to debate which mouse behaviors best correspond to the social and communication deficits seen in autism. Some researchers go so far as to question whether it’s even possible to model autism in an animal that strays so significantly from people in terms of social behavior and cognition.

Still, I think there’s great potential for mouse microbiome studies of autism because there are so many mouse models of the disorder. Researchers could compare the microbiomes of different mouse models to look for subsets that have similar changes. They could also manipulate the mouse microbiome to tease out the relationships between gut bacteria, gastrointestinal symptoms and autism traits — even if this practice only provides clues about those relationships in people.

Finally, researchers could test whether drugs being developed for specific forms of autism affect the microbiome and whether such changes are key to a drug’s effectiveness. An imperfect model can nevertheless be a perfect opportunity.