Genomics of Neurodevelopmental Diseases
Guy Rouleau, MD, PhD, FRCPC, OQ, Director, Montreal Neurological Institute and Hospital
The brain is a complex organ for which we know very little. This complexity has hampered the application of genomic approaches to the study of this organ, and the diseases that affect it. In particular the complex anatomy, the limited access to high quality tissue, the plasticity, and the fact that humans show rather unique abilities (ex. speech) make classical approaches difficult. In addition, diagnoses are based on diagnostic criteria that involve subjective judgments, defining syndromes more than specific diseases. Nonetheless, numerous epidemiological studies have clearly implicated genetic factors in the etiology most brain diseases, suggesting that genomic strategies may prove fruitful, if they can overcome the difficulties inherent to the field. With improving technology, it has become possible to explore a deep resequencing approach to the identification of the genetic factors underlying brain disease. We first chose to study ion channels in episodic brain diseases. Specifically, we sequenced 150 brain expressed ion channel genes in 368 individuals with either epilepsy, migraine, bipolar disorder, Tourette syndrome or essential tremor. Among other finding, we identified a gene predisposing to migraine. We next chose to perform a deep resequencing study of schizophrenia (SCZ), autism (AUT) and mental retardation (MR), all common, devastating and poorly treated brain disorders. Converging evidence suggests that genetically disrupted synaptogenesis and plasticity during development may in part underlie their pathogenesis. We hypothesized that a significant fraction of SCZ, AUT and MR cases are a result of de novo mutations in many different genes involved in synapse formation and function. To identify genetic factors predisposing to MR, AUT and SCZ we adopted a two-step strategy: direct re-sequencing of genes encoding proteins acting at the synapse, followed by functional validation of variants in zebrafish, drosophila or mouse hippocampal cell models. We first established a list of 5,079 synaptic and potentially synaptic genes, and prioritized them using a ranking system. We next sequenced on one strand DNA samples from 96 MR, 143 SCZ and 142 AUT unrelated probands. In the first 402 genes screened we identified 15 de novo variants in SCZ and AUT, which is twice what was expected based on the known de novo mutation rate. In addition, there was an excess of de novo point mutations that are protein-disruptive and highly deleterious in ASD and SCZ. Some of the genes for which de novo variants were found were tested for functional effects using the different animal and cell models, confirming the role of several new genes in these diseases (ex. SHANK3, IL1RAPL1). Next we switched from PCR and Sanger sequencing to exome capture and sequencing using Illumina high throughput sequencing. Using this method we confirmed that there is a higher than expected de novo mutation rate in SCZ. Together, these data suggest that de novo mutations may explain part of the missing heritability in SCZ and AUT.