When and Where Autism Genes Act in the Developing Brain

Hundreds of genes have been linked to autism, but knowing which ones matter most has been a challenge.

Researchers at the University of Aberdeen combined data from thousands of developing human brain cells to track when and where these genes are active. The study shows that many autism-linked genes act through the brain’s extracellular matrix (ECM).

Early brain development and the foundations of autism

Autism is widely understood as a developmental condition, but the biological changes linked to it begin far earlier than behavior or diagnosis. Many neurodevelopmental conditions, including autism, intellectual disability and ADHD, trace back to early brain development during the first and second trimesters of pregnancy.

Large genetic studies have identified hundreds of autism risk genes; however, understanding where, when and how these genes act during brain development has proven difficult.

Most research has focused mainly on neurons, yet brain development depends on the coordination between many cell types.

“Human brain development is an incredibly complex process. Stem cells need to know what to become and where to go, and many different cell types have to coordinate their behavior. This coordination happens within the dynamic environment of the ECM, which guides and supports these developing cells,” said senior author Dr. Eunchai Kang, a lecturer at the University of Aberdeen.

Often described as scaffolding, the ECM is better thought of as an active environment that shapes how cells divide, move and connect. It also controls how signals pass between cells as the brain forms.

Links between the ECM and brain disorders have already been identified, and mutations in ECM-related genes such as RELN, laminins and collagens can lead to severe brain malformations.

This raises an important question for autism research: could smaller changes in these same systems contribute to more subtle developmental differences?

Until now, the field has lacked a clear answer. Most studies looked at single ECM components, bulk tissue samples or animal models.

“While we know the ECM plays a crucial role in guiding these processes, its specific contributions and how changes in these pathways may relate to neurodevelopmental conditions remain underexplored,” said Kang.

The new study set out to build a detailed, human-focused map of ECM gene activity during early brain development.

Mapping autism-linked genes across developing brain cells

Kang and the team carried out a large analysis of single-cell RNA sequencing data from 6 independent studies, covering brain samples from 37 human fetuses between 8–26 weeks of gestation. When merged, these datasets captured gene activity in more than 213,000 individual cells, including stem cells, neurons, glial cells, immune cells and blood vessel cells.

This is the first study to combine large human single-cell datasets to produce a cell-by-cell map of ECM gene activity during early brain development, allowing the team to compare how ECM-related genes behaved across cell types and over time.

“By pooling together data from different studies on the ECM, we have been able to create a detailed map of its activity and see how it changes over time,” said Kang. “Helping us understand what these changes may mean for the developing brain.”

Each major brain cell type had its own pattern of ECM gene activity. These patterns also changed as development progressed, and the greatest changes in ECM gene activity occurred during the early second trimester, a period already linked to autism-related brain changes.

Many ECM genes overlapped with known autism and neurodevelopmental disorder risk genes. These links were often specific to certain cell types and time windows, suggesting that when and where a gene is active may matter as much as the gene itself.

Non-neuronal cells, such as radial glia, astrocytes and endothelial cells, were also major contributors to the ECM, challenging the idea that neurons alone drive developmental risk.

The team also examined how ECM genes support communication between cells.

They identified signaling routes involving molecules such as midkine and pleiotrophin, which are active during early brain growth. They also highlighted LGALS3, a gene linked to autism risk, as a marker of a specific group of stem cells that later give rise to glial cells. This finding was supported using staining in human fetal brain tissue.

Although the analysis included genes linked to several neurodevelopmental conditions, many of the strongest patterns involved genes associated with autism.

What ECM changes may mean for autism research

This study provides a reference map for understanding how the ECM shapes early human brain development.

For autism research, rather than listing risk genes in isolation, it shows which cells use these genes and at what stages of development.

“Knowing which genes are active in specific cell types at different stages of early development gives us a clearer picture of how the brain is built,” explained co-author Dr. Daniel Berg, a lecturer at the University of Aberdeen.

The findings suggest that altered brain environments, rather than altered neurons, may play a part. Changes in how cells interact through the ECM could influence brain wiring long before symptoms appear, which helps explain why autism genetics has been hard to interpret using neuron-focused models alone.

However, the study is based entirely on gene activity data, so it cannot show cause and effect, and gene expression does not always reflect protein levels or how the ECM is physically organized. The work also focuses on the cortex, leaving open questions about other brain regions involved in autism.

Future work could involve studying ECM genes in human brain organoids and animal models, as well as combining gene data with protein and spatial mapping. “This will guide future experiments and help researchers study gene function in the right biological context,” added Berg.

“This knowledge provides an important foundation for understanding the pathways involved in developmental brain conditions. In the long term, it may also support efforts to develop more precise and targeted therapeutic approaches,” said Berg.

 

Reference: Gim DH, Assir MZK, Soper O, et al. Deciphering cell-type-and temporally specific matrisome expression signatures in human cortical development and neurodevelopmental disorders via scRNA-seq meta-analysis. Nat Commun. 2025;16(1):9907. doi: 10.1038/s41467-025-64381-3

 

This article is a rework of a press release issued by the University of Aberdeen. Material has been edited for length and content.