Study maps two opposing molecular patterns across genetically diverse autism models

Autism involves more than 1,200 risk genes, yet researchers have long wondered whether this genetic diversity produces fundamentally different brain changes or whether different mutations ultimately converge on shared biological pathways. A team led by Professor Kim Eunjoon at the Institute for Basic Science set out to answer that question by analyzing gene activity across the brains of mice carrying different autism-risk mutations.

The researchers examined RNA sequencing data from the prefrontal cortex of 17 genetically engineered mouse lines, each carrying mutations in genes involved in synaptic communication, gene regulation, or cell signaling. They included both male and female mice, and some were exposed to fluoxetine or lithium during early development. When they analyzed more than 1,000 brain transcriptomes using multiple complementary methods, a clear pattern emerged: the genetically diverse mice fell into two distinct molecular groups with opposing characteristics.

In Group 1, genes responsible for communication between nerve cells showed reduced activity, while genes controlling chromatin structure and RNA processing showed increased activity. Group 2 displayed the inverse pattern. This distinction held up across several different analytical approaches, suggesting the two groups reflected genuine recurring patterns rather than artifacts of a single method. As Dr. Bae Mihyun, a co-corresponding author, explained, the finding suggested that many different genetic mutations converge into a limited number of molecular brain states, offering a framework for understanding autism through shared biology rather than individual genes alone.

Yet the grouping proved more fluid than a simple genetic sorting. In seven of the 17 mouse lines, males and females carrying the identical mutation fell into different molecular groups. The patterns also shifted with development—when researchers examined four representative lines at a later developmental stage, some retained their original group assignments while others switched. The two-group distinction was much less pronounced in the hippocampus than in the prefrontal cortex. These findings indicated that the molecular effects of an autism-risk mutation depend on sex, developmental stage, and brain region, making the groups context-dependent patterns of gene activity rather than fixed categories tied to specific genes.

To understand the cellular basis of these patterns, the team analyzed gene expression in approximately one million individual cell nuclei from 205 mice. Group 1 showed broader changes in the relative proportions of certain neuronal and glial cell populations, while both groups displayed distinct patterns of gene activity across multiple cell types. This suggested the opposing molecular patterns arose from coordinated changes across different types of brain cells rather than from a single neuronal population.

The researchers then tested how the two groups responded to fluoxetine and lithium, drugs previously investigated in selected autism mouse models. In Group 1, both drugs shifted certain gene sets more consistently toward the expression patterns seen in control mice. Group 2 showed more variable responses, with effects differing depending on the gene set and cell type. Neither drug reversed the observed differences in brain-cell proportions; their effects concentrated instead on particular gene-expression programs in selected neuronal populations. This suggested that molecular grouping could help researchers understand why different mouse models respond differently to experimental drugs, though the study did not establish whether changes in gene expression translated into behavioral improvements.

To test whether the findings extended beyond mice, the researchers examined three additional mouse lines and analyzed previously collected prefrontal-cortex data from 40 autistic individuals and 17 neurotypical controls. The additional mouse lines could be provisionally assigned to the two molecular groups. In the human data, researchers also identified two subgroups with opposing patterns of synaptic gene activity. However, the human patterns were not identical to those in mice. Changes involving immune-related pathways were more prominent in the human samples, while some other molecular differences were less pronounced. The available human data also did not allow researchers to link subgroup membership to specific autism-risk mutations.

The researchers emphasized that these findings remain preliminary. The study does not establish two clinical types of autism or provide a way to predict an individual's symptoms, support needs, or medication response. Instead of asking which gene is mutated, Director Kim noted, they asked whether different mutations produce common molecular patterns in the brain—a perspective that revealed surprising convergence across genetically distinct forms of autism. Future research combining transcriptomic analysis with behavioral measurements, brain-circuit studies, and additional experimental treatments will be needed to determine whether these molecular patterns can guide therapy development and evaluation.