Schizophrenia Research Forum - A Catalyst for Creative Thinking

Altered Gene Expression Prioritizes CNVs in Autism

25 June 2012. Altered gene expression may help incriminate specific copy number variants (CNVs)—deleted or duplicated segments of DNA—in brain disorders, according to a new study published online June 21 in the American Journal of Human Genetics. Led by Daniel Geschwind of the University of California, Los Angeles, the study uses gene expression data to narrow in on those CNVs most likely to be pathogenic in autism, and a similar integrative approach may highlight CNVs involved in schizophrenia.

In the past five years, CNVs have been implicated in the pathogenesis of a number of illnesses, including schizophrenia and autism spectrum disorders (ASDs) (see SRF related news story and SRF news story; Sebat et al., 2007). All of us carry CNVs, however, so parsing the "innocent bystanders" from their disease-causing cousins and determining the functional consequences of the detrimental ones are important. In the new study, Geschwind’s group presents a new method that can shed light on these issues in autism, using genomewide transcriptional profiling to examine gene expression changes associated with CNVs.

Prioritizing problem CNVs
Drawing from the Simons Simplex Collection, first author Rui Luo and colleagues used lymphoblast cell lines (LCLs) from individuals with autism and their unaffected siblings in 244 families to analyze gene expression with microarrays. Outlier genes (defined as those expressed at levels at least three standard deviations away from the mean) were identified for the autism and sibling cohorts separately. Though the probands and unaffected siblings had similar numbers of outlier genes, pathway analysis revealed that, for the probands only, these outliers were enriched for neural-related genes, including those involved in neuropeptide signaling, synaptogenesis, and cell adhesion.

The researchers then combined these findings with recently published CNV data (Sanders et al., 2011), resulting in 330 samples with both genotype and expression data. This revealed that 10.7 percent of CNVs contained one or more outlier genes, and likewise, these genes were more likely to be present in CNVs than in other parts of the genome. In addition, duplications were associated with increased gene expression and deletions with decreased expression 92 percent of the time. These observations square with the commonly assumed, but rarely verified, impact of CNVs on gene expression.

These expression changes clustered in genes hit by rare de novo CNVs, which arise spontaneously and are considered the most pathogenic: after controlling for the potential confound of CNV size, Luo and colleagues found that rare de novo CNVs were associated with a significantly higher proportion of outlier genes than both rare inherited CNVs and common CNVs. When the researchers examined small CNVs and rare, non-recurrent CNVs of unknown significance to autism, they found an overrepresentation of outlier genes, thus highlighting new genomic locales to follow up, including deletions at 3q27, 3p13, and 3p26, and duplications at 2p15. In addition, the researchers homed in on microdeletions and microduplications of 16p11.2, a CNV region previously implicated in both autism and schizophrenia (see SRF related news story), and found marked gene expression changes. Interestingly, altered gene levels in CNVs within 16p11.2 were significantly associated with head size, an autism-related phenotype. As acknowledged by the authors, the CNVs couldn’t explain all of the outlier genes, and thus future studies are needed to uncover the mechanisms behind the additional gene changes in autism.

Blood versus brain
Due to limited availability of ASD brain tissue, Geschwind’s team isolated RNA from blood-derived LCLs, which express many, but not all, of the same genes as the central nervous system (Sullivan et al., 2006). LCLs are attractive candidates for this type of analysis because they are not subject to many of the limitations of postmortem brain tissue, including small sample size, and are widely available.

What about the use of LCLs in schizophrenia? In fact, they have been used to examine gene expression changes in schizophrenia, with mixed results. Some have reported differentially expressed genes in LCLs from cases versus controls, including an enrichment of those expressed in the brain (see SRF related news story; Vawter et al., 2004), while others have reported no differences using microarrays (Matigian et al., 2008). Studies have also reported altered mRNA and protein levels of single gene products in LCLs from schizophrenia subjects (Cheng et al., 2012; Morikawa et al., 2010).

The authors note that “analysis of peripheral-blood gene expression can provide a useful and direct assessment of the functional consequences of chromosomal structural variation in a neuropsychiatric condition,” adding that the two approaches are “more powerful together than alone.” The application of this method to schizophrenia in the future may help to shed light on the role of CNVs in the illness.—Allison A. Curley.

Luo R, Sanders SJ, Tian Y, Voineagu I, Huang N, Chu SH, Klei L, Cai C, Ou J, Lowe JK, Hurles ME, Devlin B, State MW, Geschwind DH. Genome-wide Transcriptome Profiling Reveals the Functional Impact of Rare De Novo and Recurrent CNVs in Autism Spectrum Disorders. Am J Hum Genet. 2012 Jun 21. Abstract

Comments on News and Primary Papers
Comment by:  Karoly Mirnics, SRF Advisor
Submitted 16 July 2012
Posted 16 July 2012

This is another excellent genomics study from the Geschwind laboratory, challenging us to think in the context of gene networks (rather than single genes). We always knew that genome deletion, duplications, and mutations will have an effect on development and cellular function only if they ultimately affect gene expression, but rarely has this been proven so eloquently as in this study. Knowing a genetic alteration is not sufficient—the consequences are what matter—and establishing the relevance/causality of mutations and CNVs vis-à-vis a disease process is quite challenging. Combining genetics and genomics can help to achieve this, especially if the expression studies can be performed on peripheral tissues from living patients. Still, even this approach has limitations: the brain has a very different expression profile from peripheral tissues—and the real effect of altered genetic sequence cannot be evaluated if a gene is uniquely expressed in the brain. Furthermore, we know that in schizophrenia, expression events in the periphery and the CNS are only marginally overlapping, and a number of neurotransmitters and phenotype-specific functional proteins are not expressed outside of the brain. In these cases, inhibitory postsynaptic currents (IPSCs) might be very beneficial to study the cause-effect relationship and in evaluating the significance of the genetic alterations. However, the approach of the Geschwind lab is perfectly suited for evaluation of multifunctional, generic developmental gene networks, where the genes are expressed and play a similar role across various tissues. It is also important to expand these studies to many patients across a variety of human brain disorders; otherwise, due to the staggering number of genetic variations, we might not be able to recognize the common patterns that are essential for understanding mental disorders.

View all comments by Karoly Mirnics