Share this article on:

Study of Gene Expression in Children's Brains Yields Early Autism Clues

Samson, Kurt

doi: 10.1097/01.NT.0000450197.79627.c0
Back to Top | Article Outline




Two new papers provide detail three-dimensional blueprints of genic expression in postmortem tissue — one, from children with autism, and the other, from the intact brains of fetuses from the developing brain at mid-pregnancy.

Researchers have found patches of abnormal neurons throughout different layers of the cortex in postmortem tissue samples of brains from children with autism, according to a study by the Autism Center of Excellence at the University of California, San Diego (UCSD) and the Allen Institute for Brain Science in Seattle.

Focal patches of abnormal laminar cytoarchitecture and cortical disorganization of neurons, but not glia, were found in prefrontal and temporal cortical tissue samples from 10 out of 11 autistic children between the ages of 2 and 15 years, but in only one unaffected child. The samples were from children who had died from various causes.

Co-author Ed Lein, PhD, a developmental neurobiology investigator at the Allen Institute co-authored the paper in the March 27 New England Journal of Medicine.



Offering more insight on the developing brain, the Allen Institute for Brain Science unveiled on April 2 the BrainSpan Atlas of the Developing Human Brain — a comprehensive 3-D atlas of gene expression in the prenatal brain. A description of the atlas was published in Nature.

The NEJM study was led by Eric Courchesne, PhD, professor of neurosciences and director of the UCSD autism center, and first author Rich Stoner, PhD, a postdoctoral fellow at the center, with Dr. Lein as a co-author.

As with the BrainSpan Atlas, the team used RNA in situ hybridization, and a panel of layer- and cell-type–specific molecular markers to phenotype cortical microstructure, laser capture micro dissection, and transcription analysis.

A total of 25 genetic markers were evaluated, including markers for neurons and glia along with genes that have been implicated in the risk of autism — in prefrontal, temporal, and occipital neocortical tissue.

“The markers were selected based on a few criteria,” Dr. Stoner explained. “Most were selected for their ability to label cells of a certain cell-type or location, such as specific cortical layers, while others were selected for their previous link with autism, such as CNTNAP2 [contactin associated protein-like 2] and FOXP2 [forkhead box P2].”

The researchers discovered key genetic markers were absent in brain cells in multiple layers of the cortex in 91 percent of the autism samples compared with just 9 percent of control specimens. Samples from autistic children were dappled with patches of abnormal neurons throughout each of six cortical layers.

“Importantly, many of the markers with aberrant expression in the patch regions were laminar markers. It is more appropriate to think of these markers as indicators of pathology rather than trying to link specific markers to a mechanism,” Dr. Stoner told Neurology Today.

There was heterogeneity between cases regarding cell types that were most abnormal in the patches and the layers that were most affected by the pathological features. No cortical layer was spared, but the clearest signs of abnormal expression were observed in layers 4 and 5.

Signs of disorganization were localized in focal patches that were 5–7 mm in length and encompassed multiple cortical layers, including the frontal and temporal lobes.

The specific locations of the patches may underlie the expression and severity of various autism symptoms, Dr. Stoner told Neurology Today in a telephone interview.



“The goal of our research was to find brain-wide changes in gene expression, but we never expected to find differences in focal regions,” he said.

“Development of the brain during pregnancy involves creating a six-layered cortex, and we discovered focal patches of disrupted expression in one or more of these layers in the majority of samples from children with autism,” he explained. “That these defects occur in patches, rather than across the entirety of cortex, gives us a lot of evidence about the nature of autism and its origin in utero.”

Many questions remain however about the significance of the findings and the mechanisms involved, he continued. “We do not know what makes the cells in the patch regions abnormal, how they developed, or how they relate to behavior. However, the fact that we found these patches in only the frontal and temporal areas does match up with autism symptoms.”

Dr. Lein told Neurology Today that the most surprising finding was the discovery of similar early developmental pathology across nearly all of the autistic brains, especially given the diversity of symptoms in patients with autism.

“The fact that we were able to find these patches is remarkable, given that the cortex is roughly the size of the surface of a basketball and we only examined pieces of tissue the size of a pencil eraser,” he said “This suggests that these abnormalities are quite pervasive across the surface of the cortex.”

“We didn't know what to expect, and were somewhat surprised to discover very focal disruptions of cortical architecture and gene regulation. The genetics of autism is very complicated, yet this finding was common to 10 of 11 cases examined. These focal disruptions correlated to areas associated with the symptoms of autism, the prefrontal and temporal cortices, but not in others such as the visual cortex. These data suggest an origin during the early formation of the neocortex, but do not yet provide a mechanism. We are a long way from being able to use this data as a diagnostic tool but it does tell us where to look.”

Back to Top | Article Outline


The BrainSpan Atlas, developed as part of a multi-institutional consortium working to create a map of the transcriptome across the entire course of human development, provides a detailed three-dimensional blueprint of genic expression during mid-pregnancy.

This first analysis of data from a variety of gene expression, anatomical, and imaging techniques was performed on donated, intact human prenatal brains. Two of the brains were from fetuses at 15 to 16 weeks and another 2 at 21 weeks after conception. In addition to studying autism, the map is expected to help researchers studying other neurodevelopmental disorders such as schizophrenia, according to Dr. Lein, who provides oversight for the creation of all of Allen's large-scale gene expression atlases.

Dr. Lein told Neurology Today in a telephone interview that many psychiatric disorders show altered gene activity in the cortex, suggesting harmful changes during development of this region.

“Knowing when and where genes are turned on and off and their expression in specific nuclei and cell types can provide a powerful way to identify genes tied to neurodevelopmental disorders,” Dr. Lein said. “Many neuropsychiatric diseases are likely the result of abnormal brain development during gestation, and an anatomically precise molecular atlas of the brain during this time period is a first step to understanding how the human brain develops normally and what can go wrong.”



The BrainSpan Atlas, which was funded by NIH's National Institute of Mental Health and the National Institute of Child Health and Human Development, is publicly available at the Allen Brain Atlas data portal.

Back to Top | Article Outline


Until recently, autism was relatively impenetrable from a neuroclinical standpoint, with diagnosis made entirely by behavioral evaluation, said John Mazziotta MD, PhD, FAAN, chair of neurology at the University of California, Los Angeles (UCLA) and director of UCLA's Brain Mapping Center.

“Autism is probably not a single disease but the result of a number of as yet unknown changes in the brain. Today diagnosis is made by applying a set of criteria ‘made by committee’ based on certain behaviors, and it takes a certain number of these criteria to make a diagnosis,” he told Neurology Today in a telephone interview.

“Like schizophrenia, we have been waiting for evidence of some unique brain pathology to explain the disorder, and this is the first paper using mapping technology to show developmental differences between normal and autistic brains in term of genetics and brain function. If a biomarker could be found, it would dramatically change the diagnostic landscape of autism.”

Moreover, further analysis of the 10 autistic samples in the study might reveal potential outlying processes and characteristics of individuals with the disorder, he added.

The UCLA Brain Mapping Center has collaborated with researchers in Europe, to develop automated processes for evaluating very thin slices of brain tissue for analysis and study. Previously this has been extremely laborious and costly, he said, but with automated systems it is possible to scan images of slices and evaluate cell patterns in multiple frames to see abnormal processes in different brain regions.

“I believe that other researchers will start using BrainSpan to look into other disorders, especially schizophrenia and dementia. We are continuing to look for dementia biomarkers that may be evident before any symptoms appear.”

In the future, PET tracers that bind to amyloid-beta, tau, and dopamine will be used to differentiate different types of dementia, he said, and researchers are developing techniques to differentiate Lewy body dementia from that caused by tau or amyloid, he said.

“This is pretty close on the horizon and I expect we will have something within the next few years.”

Back to Top | Article Outline


•. Miller JA, Ding S-L, Sunkin SM, et al. Transcription landscape of the prenatal human brain. Nature 2014;Volume:508:Pages:199–206.
    •. Oh SW, Harries JA, Ng L, et al. A mesoscale connectome of the mouse brain. Nature 2014;508:207–214.
      •. Stoner R, Chow ML, Boyle MO, et al. Patches of disorganization in the neocortex of children with autism. N Engl J Med 2014; 370:1209–1219.
        •. King MC, Gulsuner S, Walsh T, et al. Spatial and temporal mapping of de novo mutations in schizophrenia to a fetal prefrontal cortical network. Cell 2013;154:518–529.
          •. Habas PA, Kim K, Rousseau F, et al. Atlas-based segmentation of developing tissues in the human brain with quantitative validation in young fetuses. Hum Brain Mapp 2010;31:1348–1358.
            •. More about BrainSpan:
              © 2014 American Academy of Neurology