Two of the nation’s leading experts in neuropsychopharmacology discuss the opportunities and challenges facing scientists as they search for better ways to treat disorders of brain function.
Edward M. Scolnick, M.D., a senior associate member of the Broad Institute of MIT and Harvard, is president emeritus of Merck Research Laboratories. Joseph T. Coyle, M.D., is Eben S. Draper Professor of Psychiatry and Neuroscience, and former chair of Psychiatry at the Harvard Medical School. They were interviewed by Lens magazine in 2003.
What are some of the challenges in developing new drugs for psychiatric disorders?
Coyle: The major challenge that we face for psychiatric disorders is that there is no obvious brain pathology that leaps out at you, unlike neurological disorders where there’s some very clear pathology and clear targets. And so the approach has historically relied on serendipity — that is, developing novel compounds and looking for behavioral effects. That’s really a very inefficient way to attack these problems. Even the development of SSRIs (serotonin-specific reuptake inhibitors, a new class of anti-depressants that includes Prozac and Paxil) was not linked to a detailed understanding of the pathophysiology of depression.
I think what has taken place over the last 15 years is the development of more powerful strategies to help us out. One strategy is brain imaging, both structural and functional brain imaging, which has allowed us to identify areas of the brain and circuits that do not appear to be working properly. These approaches give us a bit of a functional pathologic signature. The second strategy is genetics and the contributions of the Human Genome Project.
How useful are animal models for developing drugs for cognitive disorders?
Scolnick: They have been useful in many fields of medicine. Finding the genetic predispositions or the causative genes for any number of human diseases has allowed scientists over the last 10 or 20 years to model those diseases in genetically altered mice and rats. It’s been done in the atherosclerosis field, the cancer field; it’s been done somewhat in the field of inflammation, the obesity field. It’s quite remarkable how well and how much information has come from it, even though in some cases, like the lipid field many years ago people didn’t think this was a worthwhile thing to do at all.
Finding risk genes for cognitive disorders will allow scientists over the next decade to put these genes in animals or alter these genes, and then ask, “What can you learn from the behavior of a mouse or rat with a mutated or altered human risk gene that’s been associated with one of these illnesses? Will that become a useful way to test for more drugs?”
It’s probably impossible, given the difference in the brains of a rodent and a human, to model it perfectly, but I think useful information will come. It will be one of the ways the field progresses.
Coyle: I think in our field, the best example has been the work in Alzheimer’s disease. Mice don’t develop the pathology of Alzheimer’s disease because the gene that encodes the protein that creates amyloid deposits in human beings has a different amino acid sequence in the mouse, and so amyloid deposits are not formed naturally in rodents. Scientists have created an animal model for Alzheimer’s disease by inserting into the mouse genome the mutation in the human gene that has been linked to the inherited form of Alzheimer’s disease. A similar approach has been used with another gene, the mutant human presenilin gene, which increases the risk for Alzheimer’s disease and acts on the amyloid protein.
And so we now have mice that develop the pathology of Alzheimer’s disease. Major drug companies are developing drugs that will interfere with the generation of the amyloid or enhance its clearance. So, while these mice may not be behaviorally perfect models of Alzheimer’s disease, they certainly are powerful tools for teasing apart the pathologic pathway and providing drug targets.
One point I’d like to emphasize is that many of these risk genes ultimately may exert their effects by disturbing the development of the brain. And the abnormal behavior that is seen in the animal, the mouse, when it’s mature, may not simply reflect abnormal neuronal function, but the disruption of developmental processes that ultimately caused this behavioral manifestation. Several of the risk genes that have been identified or implicated in schizophrenia are genes that encode proteins that play a very important role in brain development.
How do you think our increased understanding of the genetics of brain disorders will ultimately improve drug development?
Scolnick: Through the emerging risk genes. But we’re only really starting to identify them because they required the human genome sequence being there in order to really make progress in the field. So it’s much too early for that to pay off with practical new therapeutics.
Coyle: I would agree. It’s a lot more complicated than disorders that follow classical Mendelian genetics. Autism, schizophrenia, the mood disorders involve, or likely involve, what is known as complex genetics in which there will be multiple genes of small effects that in combination result in the observed disease, and so that presents real challenges.
It turns out that because of the complex genetics, it’s quite possible that in different populations we’ll see certain risk genes that we don’t see in other populations. So it’s going to take a while, but I have to say that I’m much more optimistic than I was 10 years ago.
Why are you more optimistic?
Coyle: Because we have the human genome pretty much mapped, our ability to find these risk genes has been very powerfully enhanced. Right now is going to be kind of the grunt work of identifying them. If we can use the Alzheimer’s story as sort of a template, I think that once one or two are identified in a specific disorder, we’ll get a handle on a pathway that could be quite revealing.
Scolnick: I think finding these risk genes was a literally impossible problem before the human genome was mapped. I think it was just too hard. The technology wasn’t there and the information to do the studies wasn’t there. I don’t think anyone ever would have found them.
What are some of the ethical issues that impede hypothesis-testing in the psychiatric clinic?
Scolnick: I don’t think they’re special to this field. Clinical research studies require clear informed consent forms, and protocols need to be evaluated in advance by institutional review boards. The special problem in studying brain or behavioral diseases is whether the patient is competent to understand the form and sign it or whether someone else representing the patient needs to do it. I think that’s the more unique part of this field.
Coyle: I would agree with that. I think a very special challenge would be autism, where the symptom onset is typically in the second year of life. There’s growing evidence from behavioral and educational intervention research that the earlier the intervention is brought to bear, the better the outcome. If we’re going to think about potential pharmacological interventions to treat autism, we’re going to be presented with special challenges about how do you do this in very young children.
The final thing I would say is that I don’t see these disorders as being easily parsed into disorders that simply respond to drugs and not to psychological interventions. I think what we’re finding more and more is that the combination of an appropriate psychological or psycho-educational intervention with the appropriate drug can result in much more robust responses.
Scolnick (to Coyle): Do you think that the system for conducting clinical trials in younger patients in the U.S. is optimally set up from an operational and training perspective, or do you think more attention or more training programs are needed?
Coyle: The National Institute of Mental Health funds a consortium of child and adolescent psychiatry clinical trial units, so that’s the good news. The bad news is that the clinician-scientist is an endangered species, and especially in the area of child and adolescent psychiatry, the number of individuals who are involved in research and have the knowledge and skill sets to do this research is really quite small. At the leading residency training programs, a significant portion of the M.D./Ph.D.s – about 20 percent – go into psychiatry. But that is not enough to carry the load. This is an area that really needs, I think, a sustained investment from NIH.
On the other hand, the proposed merger of the National Institute of Drug Abuse and the National Institute on Alcohol Abuse and Alcoholism is a positive sign. Genetic studies are revealing risk genes common to different types of substance abuse, and including problems with gambling. Furthermore, serious mental illnesses such as schizophrenia and bipolar disorder have a high prevalence of co-occurring substance abuse that adversely affects outcome. Anything we can do to increase the cross talk among these three institutes, the NIDA, NIAAA and NIMH, should be beneficial for developing more effective treatments.
Is it more difficult to develop drugs for children and adolescents?
Scolnick: Developing drugs for use in children is harder. Significantly more safety data are required. Characteristically any trials with a new medicine are started in children only after almost all of the animal safety is done, including carcinogenicity studies, and after significant information is available in adults on efficacy and safety. And then you have to start again to find the right dose (in children and adolescents), because the dose that you find in adults may or may not be the right dose based on the size of the younger people plus their metabolic state. It’s significantly more work to do that.
Are there concerns in treating children with these kinds of drugs, considering that their brains are still developing?
Coyle: Yes, and it cuts two ways. Up until about 15 years ago, because of the dominance of psychoanalytic thinking in the field, it was believed that children could not have depression. Unfortunately, children not only can have depression, they can commit suicide, suicide being the third most common cause of death in adolescents. Studies done characterizing depression in children have shown that at any point in time about 1-2% of children under the age of puberty would satisfy the diagnosis of major depressive disorder.
When depressed children were followed longitudinally, it turned out that they would spend somewhere around 70% of their time in a state of depression or minor depression, so when you think about the impact of being psychologically depressed, feeling bad about yourself, feeling pessimistic, perhaps feeling suicidal, from say the age of 8 until the age of 15, that’s half one’s life. So there’s clear evidence that not treating can markedly distort the developmental trajectory.
So, as I say, it cuts two ways. Yes, there is a very real concern that these drugs, which affect how neurons communicate with each other in the brain, could have some adverse effects on brain development. That is currently a topic of investigation and research supported by the National Institute of Mental Health. On the other hand, not intervening can result in a child with a condition being persistently symptomatic and having a very skewed developmental trajectory.
By studying genes that affect drug metabolism, scientists are beginning to understand better why certain individuals respond differently to medications. Will this field of study, called pharmacogenomics, contribute to a new era of “individualized medicine,” the tailoring of medical treatment to individuals with psychiatric disorders?
Scolnick: Again, I think it’s a long way off because we’re just starting to find the genes and then there are the genes that affect the metabolism of the drug, and those will be different in different people, so it’s a long way before we can do that. Ultimately, that’s what will happen in most of medicine over the next, I don’t know, 10 years, 50 years. It’s really hard to tell, but ultimately that’s the way it’s going to be.
Coyle: I agree that pharmacogenomics will have a substantial impact on psychiatric treatment. I think we’ll look back at this time with our DSM-IV (fourth edition of the Diagnostic and Statistical Manual), and see it as an incredibly naïve way of categorizing these disorders.
The manual is a catalog of mental disorders based on diagnostic characteristics that have been developed through epidemiological studies. For example, we’ve worked really hard to try to separate schizophrenia from bipolar disorder, both of which are characterized by psychosis. Now with these genetic studies, it looks like there may be risk genes unique to each disorder; some may be shared by both. Once we understand the genetics better, we’ll have a very different take on how to parse these disorders out, and therefore how to treat them.
How will that pharmacogenomics the economics of drug development? If there are no more “blockbuster drugs,” will it become economically difficult to develop new medications for niche markets?
Scolnick: No, no, that’s just not going to happen. Pharmacogenomics is going to improve the ability to find drugs, make better drugs, make safer drugs, do the clinical trials better. Trials are going to be cheaper and easier to do, and so if the big companies don’t do it, the littler companies will do it. It’s not going to impede anything.
Brain scans courtesy of Sohee Park, Vanderbilt University