Psychiatric Genomics
Psychiatric Genomics Emerges at DukeThis article originally appeared in the February 2005 issue of GenomeLife.
As Chair of Duke's Department of Psychiatry and its 370-plus faculty members, Ranga Krishnan readily concedes that the application of genetics and genomics to psychiatry has so far yielded mixed results. "In many psychiatric genomics studies, even if you find a gene you often end up with a 'So what?' situation because you can't do much [with that knowledge]." But, he says, there are signs that the state of the art has evolved to the point where the combination of genomics and psychiatry can begin to produce clinically useful information. And Duke is right in the thick of those efforts.
For Krishnan, the key to solving the molecular basis of psychiatric diseases is to study large populations where the results will be more likely to hold up under scrutiny. One of Duke's advantages in that regard is the vast number of psychiatric patients who are seen at Duke University Hospital or other Duke facilities. "We have something like 150,000 patient visits per year," he notes. "We gather information on every patient who comes through our system as part of our electronic record-keeping. And the data can all be made anonymous to protect patient confidentiality."
This is your Brain on Drugs
By building such an infrastructure, Krishnan says, Duke Psychiatry will be among the first institutions in a position to use pharmacogenomics to administer medications to its patients based on their particular genomes (genotypes). David Goldstein, newly appointed Director of the IGSP's new Center for Population Genomics & Pharmacogenomics, says the potential for pharmacogenomics in psychiatry is huge.
"Many antidepressant and antipsychotic drugs are metabolized by enzymes whose activities vary dramatically among people. This means that some patients are being given much larger effective doses and others much smaller. The 'one dose fits all genotypes' assumption behind current prescribing practices is neither justified nor necessary."
Recent findings from James B. Duke Professor of Cell Biology Marc Caron's lab (published in the January 6 issue of Neuron) suggest the potential power of one genotype to affect the way psychiatric patients respond to drugs. The Caron team identified a mutation in a gene that appears to account for depression in a subset of patients. What distinguishes this work, says Caron, is that there is a clear biochemical disruption in patients carrying the mutated gene (called TPH2): they cannot produce enough serotonin, a neurotransmitter widely implicated in depression and other psychiatric disorders. "This is the first direct, tangible evidence that a gene involved in regulating serotonin may well be involved in depression," he says. Further, it implies that those carrying the mutation are unlikely to respond to drugs such as Paxil, Prozac or Zoloft, which work by keeping existing serotonin around.
Goldstein believes the TPH2 data are tantalizing, but urges caution, given that the sample size was relatively small, the mutation was also found in some control subjects (albeit ones with psychiatric histories), and the effect in humans has not been directly proven yet. "As promising as this is, I think the next step is a larger retrospective study," he says. "If that holds, then I think we could get very excited and start to think about evaluating this on a prospective basis."
Farther Afield
Krishnan notes that some Duke investigators are carrying out population studies far from campus. Professor of Medical Psychiatry Kathleen Welsh-Bohmer, for example, heads up a large-scale effort in Cache County, Utah (The Cache County Memory and Aging Study) designed to elucidate genetic and environmental risk factors to the elderly for developing dementia. And in western North Carolina, the Department of Child & Adolescent Psychiatry's Adrian Angold and Jane Costello have been following some 1400 children since the early 1990s in order to learn about the evolution of psychopathology (The Great Smoky Mountains Study); the youngest of these subjects are now 21. The researchers have been measuring a wide range of psychiatric conditions (phenotypes), including attention-deficit hyperactivity disorder (ADHD), anxiety disorders, panic disorder, depression, and eating disorders, among others.
Angold cautions that genomic studies on this population have yet to commence so he can't predict the outcome, but he believes it is particularly well suited to identifying gene-environment interactions. "We have a large sample, we have lots of measures of the environment, and we also have a pretty good handle on the phenotypes we're studying because we've measured people at multiple points in time." The latter point is crucial, according to Krishnan. "It's clear from the early data that if we look at children at different time points, the same illness will look different," he says. "That's going to be true for almost all psychiatric disorders and may explain why there's been a lot of difficulty in understanding the biology of those disorders."
Associate Research Professor of Psychiatry Brenda Plassman is using the oft-utilized tool of twin studies to get at the genetic basis of Alzheimer's disease and other dementias. As Director of Collaborative Twin Studies of Alzheimer's Disease, she is following a cohort of male twins who served in World War II. In addition to trying to understand the etiology of Alzheimer's, her group is interested in the role of major depression as both a precursor and risk factor for dementia. She relishes the prospect of finally having actual genes (like the one identified by the Caron lab) to study in her population and shares Angold's belief in the importance of gene-environment studies. "Given the same genetic risk for developing depression [such as the risk shared by identical twins], why do some individuals express the depressive phenotype while others do not? Combining genomics and twin studies of depression or dementia can help answer this question by sorting out differences in gene function and by identifying gene-environment interactions."
Paddling Upstream
Because psychiatric phenotypes such as depression can be so difficult to classify and change over time, some researchers have begun to look at more easily measured characteristics. Redford Williams is focusing on risk factors that are thought to be intermediates along the path between genes and psychiatric disease just as high cholesterol is a risk factor for heart disease. Williams, Director of Duke's Center for Behavioral Medicine, is looking for genes that make people more susceptible to the effects of stress. His group has identified a variant in the serotonin transporter gene that causes people to have much larger spikes in blood pressure when angry than those who lack the variant. For Williams, the difficulty in finding genes for psychiatric disorders is analogous to the difficulty in identifying genes that predispose to cardiovascular disease.
"Finding genes that are associated with having a heart attack is a pretty tough thing to do because chances are that any one gene will only account for a small amount of that risk. But if we come back upstream in the pathway, it's a lot easier to find genes involved in blood pressure reactivity to stress-we have a much better chance of finding genes responsible for these pre-disease [phenotypes]."
From Description to Prescription
While genomics may not have had a direct bearing on psychiatric treatments just yet, it has nevertheless had a palpable effect on patients. According to Leonard Handelsman, Medical Director of the Duke Addictions Program, "Genomics has changed how patients understand their illnesses. The discussion between doctor and patient has moved from notions of family background and heritability to genomics. It's making it easier for people to think about having disorders rather than 'flaws.' The language of genes makes the subject more specific."
Beyond changing doctor-patient language, Geoff Ginsburg, Director of the IGSP's Center for Genomic Medicine, is convinced that the potential for the genome sciences to help patients is real. "The phenotypes of psychiatric illnesses have been among the most challenging to define. But by bringing molecular approaches and pharmacogenomics to this area, we should be entering a new era of precision that will be of great benefit to patients. I think what's going on here at Duke will help take psychiatry from a descriptive to a prescriptive clinical science."