Sequencing the nuclear RNA of more than 100,000 individual postmortem brain cells from people with and without autism spectrum disorder indicates the types of genes dysregulated in the condition and the types of cells in which such dysregulation occurs. The results, reported in Science today (May 16), help narrow the focus of future ASD studies to the most likely molecular and cellular anomalies, say researchers.
“It’s using the latest technology, it’s looking at the single cell level, and it validates and extends previous observations,” says autism researcher Daniel Geschwind of the University of California, Los Angeles, who was not involved in the research. “It takes the previous work and brings it to a level of resolution that we didn’t have before.”
“This was an experiment that needed to be done,” adds geneticist Stephan Sanders of the University of California, San Francisco, writing in an email to The Scientist. “At the tissue level, it broadly replicates previous data in autism. Then, [it] provides a first look at which cell types are responsible for the differences.”
ASD, which currently affects somewhere around 1 in 60 children in the United States, includes a broad range of conditions that are characterized by an impaired ability to communicate and interact socially. The heterogeneous nature of ASD has made studies of its molecular pathology difficult. Nevertheless, gene expression studies carried out on postmortem brain tissue from ASD patients have pointed to commonly affected pathways, including synapse function, says Dmitry Velmeshev, an author of the study and postdoc in the lab of neurologist Arnold Kriegstein, also an author.
Previous studies had further implied a tendency for a decreased expression of neuronal genes and an increased expression of glial genes in ASD, writes Sanders, who was not involved with the work but provided comments on the manuscript. However, it “was unclear whether these changes represented differences in the proportion of cells (e.g. relatively fewer neurons) or dysregulation of genes in cells,” he explains.
That’s because “what people have done is bulk gene expression studies, meaning there was no direct way to look at specific cell types,” explains Velmeshev. Now, using newly developed sequencing technology called single-nucleus RNA sequencing, Velmeshev and colleagues have examined gene expression in more than 50,000 individual cell nuclei from the cortices of 15 ASD patient brains and more than 50,000 nuclei from the brains of 16 control subjects.
Following automated isolation of single nuclei into droplets, barcoding of the RNA, and subsequent sequencing, the data were automatically assigned to individual cells (based on their barcodes) allowing cell types to be determined based on their expression of key marker genes. In this way, the team was able to detect the cell types in which gene expression was significantly different between patients and controls.
They found that neurons in the upper layers of the cortex tended to exhibit the most gene dysregulation in ASD patient samples, and that genes involved in synapse function and transcription were among the most affected. Microglial cells, which function in immunity of the brain, also showed significant gene dysregulation among people with ASD, with microglial activation genes and developmental transcription factor genes being the most affected.
“It’s an exciting advance and appears to be a very sound and important piece of work . . . describing in a very detailed and elegant way the pathways that are impacted,” says ASD researcher Dennis Wall of Stanford University School of Medicine who was not involved with the research. “It helps to highlight areas of promise to go after.”
While Joseph Buxbaum of the Icahn School of Medicine at Mount Sinai in New York agrees that “it’s an elegant study,” he is concerned about the sample size. Yes, there may be more than 50,000 nuclei, but these come from just 15 patients, he says. This is “no fault of the authors,” he continues, but rather reflects “a big problem in the field—the fundamental issue of very limited brain samples.”
“I completely trust that everything [in the paper] is true,” he says, “But it may not be universal.”
Velmeshev agrees that “replication in a larger dataset will definitely be important.” He adds, “when we compared bulk gene expression data from our samples to what has been [previously] generated . . . there was a significant number of ASD-associated transcriptional changes common between datasets.” A finding that “is very encouraging,” says Kriegstein.
D. Velmeshev et al., “Single-cell genomics identifies cell type–specific molecular changes in autism.” Science, 364:685–89, 2019.