Gene Expression Differences Detected in Toddlers with Autism Interview with:

Michael Lombardo, PhD Assistant professor of Psychology the University of Cyprus 

Dr. Lombardo

Michael Lombardo, PhD
Assistant professor of Psychology
the University of Cyprus What is the background for this study?

Response: Autism is a diagnostic label we give to children with difficulties in the areas of social-communication and restricted, repetitive stereotyped behaviors and interests. The diagnosis is made based on observations about behavior and is a consensus label, meaning that clinicians can show high degrees of agreement that a given set of behaviors is ‘autism’. But aside from the diagnostic label, there is a fair degree of heterogeneity within patients that have the diagnosis. One way in which patients are heterogeneous is with regard to early language development. Some toddlers with autism are minimally verbal, while at the other end, many toddlers with autism develop language typically. An important question to answer is whether that kind of difference in language development indicates a subtype with different underlying biology.

To examine this question, we first split toddlers with autism into two subtypes defined by their language outcome at 4 years of age. Some toddlers were classified as poor language outcome, because their language performance was 1 standard deviation below typical norms. Other toddlers with autism had relatively good language outcome, as their language performance by 4 years of age was within 1 standard deviation of typical norms.

We also measured the biology behind these two autism subtypes. First we used functional magnetic resonance imaging (fMRI), which is a non-invasive method to look at blood oxygenation response that changes according to a task. Blood oxygenation changes are an indirect measure of neural activity. We used fMRI during natural sleep at around 29 months of age while the toddlers were played language stimuli through headphones to elicit neural responses to speech. Second, we measured molecular aspects of biology, by taking blood samples, isolating leukocyte cells, and then quantifying gene expression for all protein coding genes in the genome, at around the same time as the fMRI scan. What are the main findings?

Response: We found that widespread and coordinated patterns of gene expression activity in leukocytes was associated with widespread neural responses to speech across the brain. This gene expression-fMRI association however, was different in typically developing toddlers and toddlers with autism and poor or good language outcome. This result shows us that there are molecular mechanisms identifiable in leukocyte cells that are associated with in vivo neural response to speech in these toddlers. However, the way those mechanisms are linked to neural response to speech is different across the groups, and that is telling that the biology behind autism with poor language outcome is different compared to typically developing toddlers or toddlers with autism and good language outcome.

Since the cells used to measure gene expression aren’t brain cells, we looked to see if the important genes identified by the analysis were indeed genes that can have impact on brain tissue. Using another data resource of gene expression patterns across tissues, we were able to show that many of the important genes we found are those that are indeed expressed in the brain, but are also expressed in many other tissues. These genes are called broadly expressed genes. The fact that broadly expressed genes can be measured in non-neural tissues, and show an association with neural response measured with fMRI is interesting, given that we know that these genes do indeed also have impact on neural tissue.

We also looked to see if the important genes we found were involved in component processes for language, such as vocal learning. We re-analyzed a gene expression dataset examining vocal learning in a species of song birds that like humans, exhibits vocal learning abilities. Many of our important genes linked to fMRI response to speech are the same genes that in the brains of song birds, are important for vocal learning. Vocal learning isn’t all there is to human language, obviously, so we also looked at sets of genes that are specifically upregulated in the cortex of humans versus our nearest non-human primate neighbors, chimpanzees, who do not use language. Here again, many of the genes we find linked to fMRI response to speech in humans are those genes that are upregulated specifically in the human brain – likely indicating that some of these important genes are overlapping with genes that have changed during human evolution and likely enable us to develop and use language to communicate. And then finally, we looked at evidence relating to genes involved in autism and which are dysregulated in the cortex of actual autistic patients. Here again, we find that many of our important genes linked to fMRI response to speech are the same genes we know are highly linked to autism, and when looking directly at gene expression in cortical tissue of autistic patients, those genes are downregulated in expression. Many genes linked to autism are also highly expressed during prenatal periods of development, and we also found that our important genes linked to fMRI response to speech are genes of importance to autism that also are highly active during prenatal brain development. What should readers take away from your report? 

Response: Biology behind different neural development in toddlers with autism who exhibit poor language development is likely different from individuals with the same diagnostic label, autism, but whom show good language development. What recommendations do you have for future research as a result of this work?

Response: We’d like to capitalize on this method of looking at gene expression in blood and link it to in vivo neural phenotypes. If one wants to look at molecular biology linked to brain processes in living individuals, this is notoriously difficult, because brain tissue is largely inaccessible to molecular assays in living patients. Therefore, any way to get a look at molecular mechanisms linked to brain processes in living patients that gets around this issue, likely has large potential. One could imagine that perhaps we could use this method to evaluate treatment responses in living patients, to see if the treatment is affecting biology linked to functional brain processes. More generally, we don’t know much about the molecular mechanisms linked to neural phenotypes measured with a method like fMRI, so this method could likely open up possibilities to begin discovering how peripheral samples like blood can carry information relevant to help us understand the mechanisms likely linked to neuroimaging phenotypes in living individuals as they develop and change over the lifespan. 


Michael V. Lombardo, Tiziano Pramparo, Vahid Gazestani, Varun Warrier, Richard A. I. Bethlehem, Cynthia Carter Barnes, Linda Lopez, Nathan E. Lewis, Lisa Eyler, Karen Pierce, Eric Courchesne. Large-scale associations between the leukocyte transcriptome and BOLD responses to speech differ in autism early language outcome subtypes. Nature Neuroscience, 2018; 21 (12): 1680 DOI: 10.1038/s41593-018-0281-3

Nov 27, 2018 @ 3:22 pm 

The information on is provided for educational purposes only, and is in no way intended to diagnose, cure, or treat any medical or other condition. Always seek the advice of your physician or other qualified health and ask your doctor any questions you may have regarding a medical condition. In addition to all other limitations and disclaimers in this agreement, service provider and its third party providers disclaim any liability or loss in connection with the content provided on this website.


Leave a Reply

Your email address will not be published.

This site uses Akismet to reduce spam. Learn how your comment data is processed.