Neurodivergence meets neural diversity: Shedding light on transcriptional dysregulation in autism spectrum disorder
Neurodivergent (adjective). “Differing in mental or neurological function from what is considered typical or normal […] Not neurotypical.” (Oxford Languages)
It’s rare that a single word can encompass so many people. Yet, neurodivergent is a term used to reflect individuals with dyslexia, ADHD, and, of course, autism spectrum disorder (ASD). It’s estimated that roughly 1% of the global population is on the ASD spectrum, while in the United States roughly 1 in 44 children are diagnosed (1).
Although ASD is common, it remains complicated to pinpoint its molecular and cellular origins. This is in large part because the brain itself—particularly in disorders—is a complicated tangle of where, what, and how. Where are ASD-associated changes taking place? What is changing? And how can the answer to these questions help better shape our understanding and future studies?
Answering each of those questions is critical to understanding the whole of ASD. For Autism Awareness Month, we’re spotlighting the incredible amount of work done in a recent article by researchers seeking to use both bulk and single-nuclei RNA-seq (snRNA-seq) to provide a more unified picture of the molecular neurodiversity of individuals with ASD (2).
The Where: The importance of neuroanatomy in ASD
“We observe consistent transcriptomic signatures of ASD across all 11 cortical regions profiled, with highly concordant effect-size changes across each region.” –Gandal et al. (2)
To analyze transcriptomic changes as a function of brain region in ASD, the researchers collected samples from 54 neurotypical (NT) controls and 49 patients with ASD across multiple brain regions, resulting in a staggering 725 samples processed with bulk RNA-seq.
Starting with an analysis of the entire cortex, the researchers found a large number of not just differentially expressed genes but also transcript isoforms, suggesting substantial involvement of transcriptional isoform generation in ASD. Their analysis yielded 4,200 individual genes and 9,474 transcripts differentially expressed in ASD versus NT patients.
Region-specific ASD transcriptional changes were also observed in each of the 11 cortical regions the group analyzed. Intriguingly, the largest changes by far—both in effect size and in transcriptional dysregulation—were in the primary visual cortex (BA17), where 3,264 genes were differentially expressed.
Researchers next sorted these widespread transcriptional changes into “modules” of differentially expressed genes and found that 38 modules were disrupted in ASD. Some of these modules were cortex-wide, including one module (GeneM5) centered on synaptic function with multiple ASD risk genes, another focused around reactive astrocytes, and a third on blood–brain barrier–associated genes. The group noted this is consistent with previous work in the frontotemporal cortex, and that the current findings show these changes encompass the entire cortex.
Focusing on “regional” modules, BA17 once again came to the forefront with four ASD-associated modules that were unique to this region. Unlike region-wide modules, several of these modules focused on oligodendrocytes, inhibitory neurons, or development and long intergenic non-coding RNAs. Also of interest is that multiple ASD-associated dysregulated transcripts showed not just regional specificity, but that the magnitude of gene expression changes followed an anatomical gradient (e.g., increasing expression from anterior to posterior). This effect was most notable in, again, BA17.
The What: The cellular basis of transcriptomic dysregulation
“Notably, we see that cell-type–specific transcriptomic dysregulation contributes substantially to the changes observed with bulk RNA-seq, whereas cell-type proportion contributions are subtle.” –Gandal et al. (2)
There are multiple ways that gene expression changes can come about: the proportion of specific cell type(s) may be altered. Disorders and/or disease can give rise to subpopulations of cells with novel transcriptomic signatures. Or, gene expression changes may be driven within existing cell types.
Though bulk RNA-seq can offer valuable insights, it is unable to parse the cellular basis of transcriptomic changes seen in bulk sequencing data. To address this limitation, the researchers turned to snRNA-seq, using the Chromium Single Cell 3’ Gene Expression assay to identify the specific cell type(s) comprising these gene expression signatures.
Using 250,000 nuclei across three regions (frontal, parietal, and occipital cortex) from 6 ASD and 6 NT patients, the group first asked whether altered cell-type composition was responsible. While their data showed 26 clusters that confirmed the presence of all major cortical cell types (as well as some region-specific cell subtypes), there were no significantly different proportions of cells in ASD versus NT patients.
They next turned to whether cell type–specific gene expression differences could drive the changes they saw in the bulk data. They were able to recapitulate the anterior-to-posterior gradient of ASD-associated gene expression changes in individual cell types, with a far greater (3 to 4 times more) number of differentially expressed genes seen in occipital and parietal versus prefrontal cortices.
Interestingly, the majority of differential gene expression was driven by excitatory neuronal subtypes, which also reflected the anatomic gradient with both more genes and a greater magnitude of differential expression observed in posterior regions. However, >90% of the modules showed substantial cell type–specific enrichment of differentially expressed genes, with both cell-type and/or regional specificity.
For example, while one module focused on heat-shock isoforms (a fairly general cell process) with genes upregulated across almost all cell types, the previously mentioned GeneM5 demonstrated a particular emphasis on downregulated genes in multiple excitatory neuronal and oligodendrocyte cells. Further emphasizing the posterior connection, one ASD-linked, immune-focused module showed both enrichment in major glial cell types and further enrichment in posterior versus anterior regions.
The How: Advancing our understanding of ASD
“As we seek to gain a complete understanding of ASD neural pathology, future approaches that integrate different sources of biological data […] to determine how ASD risk genes affect the brain will be essential.” –Gandal et al. (2)
This study represents a truly impressive amount of work, not only building on previous findings in the field—such as transcriptomic dysregulation in the prefrontal cortex—but expanding it to the entire cortex. It highlights the importance of the where in the brain (both whole cortex and in individual regions), the what of specific cell types that drive differential gene expression in ASD, and the how of combining them to gain a more complete understanding of ASD molecular and neural pathology.
Finally, as the authors note, it’s critical to integrate different sources of biological data: something their research did an admirable job underscoring. It lays the groundwork for moving forward and building our understanding of neurodivergence, but there’s still much more to do.
Interested in taking a deeper look at the technologies used in this article? Check out Chromium Single Cell 3’ Gene Expression, or take a different approach and discover how you can combine single cell and spatial analyses with subcellular resolution with Xenium In Situ.
- Autism. Who.int. https://www.who.int/news-room/fact-sheets/detail/autism-spectrum-disorders
- Gandal MJ, et al. Broad transcriptomic dysregulation occurs across the cerebral cortex in ASD. Nature 611: 532–539 (2022). doi: 10.1038/s41586-022-05377-7