Based on twin and family studies, we have long known that some mental disorders have a high degree of heritability, as great, or greater than most other common medical disorders. In recent years, NIMH-supported researchers have discovered several genes that are associated with autism spectrum disorder (ASD), schizophrenia, attention deficit hyperactivity disorder, and bipolar disorder. Most of these genes were discovered either through a candidate gene approach comparing cases and controls or by looking for linkage to genetic variation associated with occurrence of the disease in a family. However, the genomic variants discovered to date can explain only a small fraction of the genetic risk. So where are the missing genetic signals for mental disorders?
Three years ago, we thought the answer would come from whole genome association (WGA) studies. These studies were based on the identification of roughly 3 million points of common variation in the human genome. WGA studies used genotyping to measure hundreds of thousands of these single base differences between the genomes of those with and without a disorder. In what became known as the “common variant – common disease approach”, genomic risk was mapped for a number of prevalent disorders, including macular degeneration, inflammatory bowel disease, and diabetes.1 However, thus far, the search for common variations has yielded only a few success stories for mental disorders. These discoveries are notable, with potential implications for mechanisms of disease, but they still fall far short of explaining more than a small fraction of the genetic risk.
In the past two years, we have realized that people with serious mental disorders are more likely to have rare variations known as copy number variations (CNVs). CNVs are variations within the genome that results from deletions or duplications of genomic segments, sometimes involving millions of bases of DNA. CNVs are sometimes confined to a single family or develop “de novo” in just one individual. While each CNV is rare, there are so many of these structural variants (over 38,000 have been identified), that some scientists claim these CNVs contribute more variance than the millions of single base changes interrogated by WGA. 2 And unlike single base variations that usually have very modest effects on risk, some of these large mutations may be more penetrant, that is, more likely to cause disease.
Although these large CNVs are 10 times more common in people with schizophrenia or autism, 3 4 most of the known CNVs do not seem to be associated with any single neurodevelopmental disorder. Even within a single family, the same genetic lesion appears to be associated with different mental or developmental disorders. And even though these may be huge mutations, some CNVs by themselves have subtle effects unless there is a second insult such as a second mutation or an environmental influence.5 As we look closer, it’s clear that there are many different forms of variation in the genome and many potential genomic roads to mental disorder. So many, that some scientists have taken to quoting Tolstoy’s opening to Anna Karenina: “All happy familes are alike; each unhappy family is unhappy in its own way”.
How will we determine which of these various genetic signals contribute to mental disorders? We can leverage the insights that we have gained from linkage and WGA studies, which have pointed us toward several chromosomal regions for deeper analysis. Rare variants will be identified from candidate gene studies and full genome sequencing. At present, sequencing costs constrain full genome sequencing efforts, but technological advances are increasing the speed and decreasing the cost of sequencing by orders of magnitude. In the near future, full genome sequencing will be common practice. In the meantime, efforts can be focused on fully sequencing target genes, regions of interest, or the exome (the coding regions of the genome). And attending to Tolstoy’s insight, the search for genomic variation in pedigrees may be more instructive than looking for rare events in case control series.
Another important area of focus will be epigenomics--the mechanisms through which environmental and experiential influences interact with genes to control their function. Epigenetic changes describe alterations to DNA structure and packaging that do not affect the underlying sequence. For example, a rare CNV associated with ASD deletes the gene that codes for the oxytocin receptor. In many individuals with ASD who do not have this deletion, the gene is silenced by epigenomic modifications, essentially producing the same outcome as a gene deletion.6 By combining epigenomic studies with refined genomic sequence analyses, we will be one step closer to understanding mechanisms of pathophysiology.
In order to carry out these large scale genomic and epigenomic studies, it will require a concerted effort to share data and biospecimens. NIMH has spent over a decade building a resource for the research community—the NIMH Collaborative Center for Genetic Studies of Mental Disorders –a repository of DNA, cell cultures, and clinical data from tens of thousands of patients and controls. In the coming years, our goal is to further enrich the repository to include samples from parents and first degree relatives. We expect all NIMH-funded researchers to share data from genomics studies through this repository or other databases such as the NIH database of Genotypes and Phenotypes , or the National Database for Autism Research .
The journey to discover the genomic risk factors for mental disorders continues to be full of surprises, such as finding new sources of variation. The journey is also speeding up, with the advent of faster and cheaper sequencing tools. Clearly, genomics is only part of the cause for serious mental illness, but it is a part that is finite and tractable and, in the near future, offers our best portal to the pathophysiology of these complex disorders.
1Manolio TA et al. A HapMap harvest of insights into the genetics of common disease. J Clin Invest. 2008 May;118(5):1590-605.
2Zhang F et al. Copy number variation in human health, disease, and evolution. Annu Rev Genomics Hum Genet. 2009;10:451-81.
3Sebat J et al. Strong association of de novo copy number mutations with autism. Science. 2007 Apr 20;316(5823):445-9. Epub 2007 Mar 15.
4Walsh T et al. Rare structural variants disrupt multiple genes in neurodevelopmental pathways in schizophrenia. Science. 2008 Apr 25;320(5875):539-43. Epub 2008 Mar 27.
5Girirajan S et al. A recurrent 16p12.1 microdeletion supports a two-hit model for severe developmental delay. Nat Genet. 2010 Mar;42(3):203-9. Epub 2010 Feb 14.
6Gregory SG et al. Genomic and epigenetic evidence for oxytocin receptor deficiency in autism. BMC Med. 2009 Oct 22;7:62.