by Steven E. Hyman, M.D.
Provost, Harvard University
Professor of Neurobiology, Harvard Medical School
Since its formation in 1992, the Dana Alliance for Brain Initiatives has sought to keep the public informed about cutting-edge research in neuroscience. Given the large and growing number of people afflicted with brain disorders, public understanding of the research becomes more crucial each year. Among the important findings of the past year presented in this, the sixteenth in the Alliance’s annual progress reports on brain research, I want to highlight promising outcomes in the genetics of neuropsychiatric disorders. I do this not only because of the significance of the results, but also because they arrive after years of frustration.
Neuroscientists face daunting challenges in attempting to understand the processes that go awry in the brain to cause autism, schizophrenia, bipolar disorder, obsessive-compulsive disorder, depression and other neuropsychiatric disorders. The difficulty derives in part from the fact that these disorders affect the highlevel integrated functioning of the human brain, impairing complex domains such as social cognition, control of behavior and mood regulation. Consequently, it has been very difficult to model these disorders convincingly in animals. In addition, rather than exhibiting readily identifiable pathology such as cell death, the symptoms of neuropsychiatric disorders often reflect abnormal activity in brain circuits. In psychiatry, for example, the science has matured to the point that depression is no longer seen in excessively simplistic terms as the deficiency of one or two neurotransmitters; rather, it is viewed as the result of regulatory mechanisms in the brain gone awry.
Non-invasive neuroimaging provides some assistance, by enabling researchers to observe the brain at work while performing tasks that might help distinguish between health and disorder or probe specific impairments in conditions such as autism or schizophrenia. With imaging, it has been possible to study, for example, how people with autism process social information and to study how the impairment of “working memory,” which is the ability to hold information “online” to guide thought and behavior, affects people with schizophrenia. A more recent development, deep brain stimulation (DBS), described in this report, promises to give us new information on the workings and possible infirmities of brain circuits that regulate mood and aspects of cognition. By delivering electrical current through fine electrodes implanted deep in the brain, DBS can activate or inhibit specific brain circuits and thus regulate mood or diminish unwanted, intrusive thoughts (obsessions). The combination of results from neuroimaging and DBS has given us exciting new hypotheses about the circuits that regulate mood in the brain and those that might malfunction, for example, to cause depression or obsessive-compulsive disorders. Nonetheless, these are only pieces in an extraordinarily complex puzzle.
Scientists have long hoped that the identification of the precise genetic variations that contribute to neuropsychiatric disorders would provide tools that neurobiologists could use to decipher the disease processes. Though finding such genes has proved difficult (for reasons that I will describe below), nevertheless, after years of frustration, results reported during the last year have begun to identify genes involved in autism and, with a bit less certainty, in schizophrenia and bipolar disorder.
An important distinction is also relevant to this research: the distinction between genetics and epigenetics, both of which produced significant new findings in psychiatric research in 2009. Briefly, epigenetics is the study of how the action of genes may change once an organism’s underlying DNA blueprint is laid down. Here’s what that means: DNA, the genetic material, is bound inside the nucleus of the cell by a large diversity of proteins. If uncoiled, DNA molecules would stretch far beyond the boundaries of individual cells, but they do not, because they are held in coiled and folded conformations by histones and other proteins. As a kind of spool for the DNA, histones are the most important in holding its structure. If a gene is to be expressed— that is, to be read out to produce either an RNA or a protein product—it must be accessible to the transcriptional machinery that does the reading, and histones help make that possible, too. We have long known that both DNA and histones undergo chemical modification (e.g., by adding or removing methyl groups or other chemical groups) and that these modifications make it more or less likely that a nearby gene will be expressed. The modification of DNA and histones is what is meant by epigenetic regulation of the genome.
Most of the attention paid to epigenetics was focused on early development. Every cell in our bodies begins with the same genome,but some become liver cells; others become one of many thousands of different types of neurons; and still others take on the humble but necessary task of making fingernails. The stable patterns of gene expression that give rise to the myriad cell types of the body are in great part the result of epigenetic modifications within cells. A surprise that crystallized in the recent past is that epigenetic regulation can be induced by stress, other types of life experience and both therapeutic and abused drugs. The implications of these discoveries are now a matter of very exciting research, and one possibility is that epigenetic mechanisms may provide new avenues for designing treatments.
Genetics provides tools for biological investigation in many ways. At the simplest level, a version of a gene that predisposes to a disease, such as autism, can be compared with a different version of the same gene that does not. If that particular genetic variant strongly influences the symptoms of autism (i.e., the variant’s degree of “penetrance”), researchers might insert it into a mouse genome to observe its effect on brain development, brain function and behavior in the engineered mice.
Despite the substantial influence of genes, attempts to identify the variants that confer risk of neuropsychiatric illness have proved frustrating for the last two decades. This is largely because the concept of heritability that we investigate is an aggregate measure—it lumps together the totality of the effects of genetic influence. If we “look under the hood,” we find that a very large number of different genes affecting brain function can contribute to specific disorders, and that, in different families, an individual might have symptoms of schizophrenia as a result of different combinations of genes. This scientifically messy situation, which is described as “genetic complexity,” is typical of most common human illnesses. Harmful mutations that act alone to cause a serious illness tend to decrease reproductive fitness, and thus they often get weeded out of the gene pool. As a result, disorders caused by a single harmful high-penetrance mutation tend to be rare. If, instead, illness is caused by the interaction of a large number of gene variants that are not harmful, except in unlucky combinations, the risk-conferring gene variants will remain in the population. This is described as the “common disease, common variant” hypothesis. Alternatively, disorders that appear genetically complex can result from diverse high-penetrance mutations, each individually rare, acting in different extended families. Thus, for example, a disorder causing vision loss, retinitis pigmentosa, is a single-gene disorder in each extended family, but there are a very large number of individual mutations that lead to retinal degeneration, ultimately by converging molecular mechanisms.
Modern genomic technologies have begun to solve the genetics of diseases resulting from both types of complexity, but the common neuropsychiatric disorders exhibit one further level of difficulty to geneticists. Unlike most other complex diseases, such as type 2 diabetes mellitus or inflammatory bowel disease, no objective medical tests exist to narrow the study populations of neuropsychiatric disorders. Geneticists—indeed, all scientists studying these disorders—must therefore rely on clinical observation, with all its inherent imprecision, to make diagnoses.
After years of taking one step forward and one step back, research efforts in 2009 yielded notable progress in the search for genetic variants that contribute to autism and promising results for those involved in schizophrenia and bipolar disorder. I readily confess that this work is far from complete and, in most cases, still some distance from providing the tools that neurobiologists need to interrogate the brain. However, the sense of progress and promise is palpableand exciting.
The Human Genome Project and other large-scale efforts in genomics have provided new information and technologies relevant to understanding disease risk. Although I am not a human geneticist, I have long been concerned with the question of how to exploit the high heritability of neuropsychiatric disorders to provide tools for neurobiology. A decade ago, when I was director of the National Institute of Mental Health (NIMH), the reality of genetic complexity and the challenges it posed were coming squarely into view. Traditional linkage studies, based on markers taken from lowresolution maps of the genome, were not yielding reproducible results. The question was how to spend federal resources in a way that would maximize the possibility of long-term success. I began programs to collect and store DNA samples and extensive phenotype information from large numbers of individuals and families affected by schizophrenia, bipolar disorder, early onset depression and later, in collaboration with family groups, autism. I also instituted a “sharing policy” so that these resources could be available broadly to researchers as the technologies improved. The sharing policy was not initially popular with the entire community of investigators, but it is now widely accepted. As the psychiatric research community matured, it was widely recognized that pooling of samples would be necessary to generate studies large enough to yield reproducible results. Not surprisingly, many of the recent successes have come from large international collaborations.
It has taken far longer than I could have imagined to get to where we are now, and the NIMH DNA resources have proven important but far too small on their own. As I have described, it is clear that multiple pathways lead to illnesses such as autism and schizophrenia; some individuals are at risk because of an unlucky combination of a very large number of common genetic variants, and others may have rare harmful mutations. Perhaps not surprisingly, the discovery of common variants has led to a new controversy: since each variant has such a small individual influence, one can reasonably ask whether the results of whole-genome association studies can really inform biology. I believe that there will be a large biological payoff, but that it will require very clever scientists, including computer scientists, to show us how all this information comes together.
If the genetic clues converge on a limited number of pathways that can illuminate the biology of the illness, clues to treatment developments will follow. Unfortunately, there are no guarantees that this will happen. The discovery of rare high-penetrance mutations may initially be more useful to neurobiologists, because such mutations are more likely to produce substantial biological effects. Already researchers have produced genetic mouse models with human mutations that cause autism or disorders that include symptoms of autism, such as Rett syndrome. I do not want to over-promise on the rate of progress, but I think that, after decades of effort, we will have a new and important window into the biology of neuropsychiatric disorders. This will be complemented by new understandings of the role of epigenetics and by new avenues of research on the brain.
One such avenue described in this report brought good news to scientists pursuing treatments for disorders that do not respond to pharmacological therapies: encouraging results from deep brain stimulation (DBS) studies. DBS involves neurosurgery to implant thin electrodes deep in the brain, attaching them to a device implanted under the collarbone that delivers a steady electrical current to the neurons the electrodes are touching. Developed in the late 1980s and 1990s, DBS is now widely used to treat symptoms of treatment-resistant Parkinson’s disease and is approved by the FDA on a limited basis for obsessive-compulsive disorder, as well as some movement disorders. New studies are showing that some patients with depression who had not responded to medication, psychotherapy and electroconvulsive therapy (shock treatment) have experienced longlasting improvements with this therapy.
Beyond the value of genetics for psychiatric research, genetic advances in 2009 continued to improve the outlook for new therapies for diseases, such as multiple sclerosis, in which multiple genes are implicated. Certain populations, especially those of Northern European stock, are more susceptible to MS, but environmental pressures such as a virus may be needed as a catalyst. Once scientists are able to determine which genes are interacting with which environmental factors, new treatments will become possible. In 2009, scientists using a genome database identified several specific genes, all genes in the immune system, that contribute to susceptibility to MS.
Equally noteworthy in 2009 were findings about molecular processes providing insight into normal brain function. Among these, scientists studying the molecules involved in memory succeeded in selectively removing fearful memories in rats by using a protein, CREB, to identify the specific neurons carrying the memory, and then destroying those neurons. Other research explored enhancing memories by increasing the expression of an enzyme, PKM zeta, which helps turn short-term memories into long-term memories.
Against the backdrop of debilitating disorders and those that threaten people’s very identity, it is easy to forget that, for most people, the risk is that harm to the brain will come in the form of disabling brain or nervous system injury, by stroke, accidents such as car crashes and chronic disease. In the search for effective therapies and treatments for injury and disease, scientists have begun delving into the mystery of neuroprotection—the methods by which neurons protect themselves against injury. In 2009 scientists added several chemicals to the small list of known neuroprotectors. Sex hormones, especially estrogen, figure heavily in the reduction of neuron loss with aging, while vitamin D helps to minimize cognitive decline.
But it was genetics that took center stage in the advancement of neuroscience in 2009. Genes dictate how our brains develop, what diseases we may be susceptible to and even how well our neurons will survive into old age. Mutations and variants, which scientists are now able to identify through the use of the Humane Genome Project, hold clues to treating psychiatric illnesses, such as Parkinson’s, and to unlocking cures for diseases such as multiple sclerosis. In combination with the discoveries of the molecules that govern brain function, genetics promises to propel neuroscientists to greater understanding about how the brain works, and how we can both heal it and improve its function.