BY DR. ALLEN FRANCES
"For every complex problem there is an answer that is clear, simple, and wrong." H. L. Mencken
The nature/nurture debate on the causes of mental disorder generates no end of silly controversy by proponents on both sides.
The biological reductionists act like the secret of psychiatric disorders is written in the genetic code. They are "mindless"- dismissing the crucial role of environment in how our brain develops and of psychology and social context in how it functions.
The environmental reductionists go to the opposite "brainless" extreme- argueing that a psychiatric disorder is a direct reflection of life stresses and dismissing the crucial role of biological vulnerability, particularly for the severe mental disorders.
Both sides fail to appreciate the complexity of interaction among biology, psychology, and social setting. The brain is the most complicated thing in the known universe. It contains more than 100 billion neurons (equal to the number of stars in our galaxy), each firing dozens to hundreds of times a second, and connected to each other by 240 trillion synapses. There is no way that our 20,000 genes could in any simple top/down fashion instruct the intricate wiring and firing of so many connections. The miracle is that this complex system usually works as well as it does.
Experience must play the crucial mediating role in facilitating appropriate adaptation. Neurons that fire together, wire together. It is equally silly to deny the role of biology in providing the hardware as it is to deny the role of experience in helping to shape the software.
Having a close relative with a psychiatric disorder is usually the most predictable risk factor for developing that disorder yourself, but the relationship is not inevitable. Even identical twins (who have identical genes) as often as not do not develop the same psychiatric disorder.
The intense fifty year search to figure out how heredity works has produced many hyped claims, false starts, blind alleys, and failed replications. This is a confusing minefield of contradictory findings, difficult for the non-expert to interpret.
Luckily we have the perfect guide. Steve Dubovsky MD is Professor and Chair of the Department of Psychiatry at the State University of New York at Buffalo and Adjoint Professor of Psychiatry and Medicine at the University of Colorado.
Dr Dubovsky writes: "You can hardly watch TV or open a magazine these days without coming across advertisements for practitioners with the latest scientific and most 'personalized' methods to diagnose and treat whatever might ail you. Among these modern wonders is the claim that genetic testing, can tell you exactly which medication will work best for your condition. Such promises are encouraged by an escalating number of studies demonstrating associations between versions (alleles) of various genes and illness subtypes, as well as increased availability of genetic testing.
Some of the information emerging from genetic studies currently has clinical applications in limited areas of medicine, a certain amount is deceptive, and most, while promising, is not yet ready for prime time. How can we tell these categories apart?
First some basic principles. Genes do not cause illnesses; they make proteins, usually by making messenger RNA. Because proteins have discrete functions, different proteins coded by different genes frequently interact with each other to produce complex manifestations called phenotypes.
Different genotypes (patterns of genes) produce different phenotypes. A few medical illnesses are phenotypes that are the downstream result of an aberrant allele of a single gene that produces a malfunctioning protein. For example, Huntington's disease is caused by an abnormal protein called huntingtin produced by a single defective gene. If you have the gene, you will get the illness. Not one of all the psychiatric disorders is in this category of simple gene causation.
Even though you are born with all the genes you will ever have, their expression varies over the lifetime and under different circumstances.
Many different interacting processes- environmental factors, experience, inner states, illnesses and medications- vary whether a gene does or does not get expressed. Because of 'epigenetics', the person's genotype (summary of alleles of various genes) does not inevitably predict phenotype.
This brings us to risk assessment based on genetics. About 10% of cases of breast cancer have a familial pattern that suggests the influence of a mutation of one of two genes, called BRCA1 and BRCA2. Although defective versions of these tumor suppressor genes (which produce proteins that block proliferation of cancer cells) convey a significantly increased risk of breast and ovarian cancer, only a small minority of women who have breast cancer have these genes. The breast is the simplest organ in the body, but we still don't understand much about the diseases associated with it.
It is therefore no surprise that is so difficult to sort out any consistent pattern for psychiatric disorders that are based on the most complicated interaction between a ridiculously complicated organ and a ridiculously complicated environment.
Every time a new study reports that a particular gene is found more frequently in people with a particular psychiatric disorder than normal subjects, it seems that the gene must be the cause of that disorder, or at least a reliable marker. And then another disorder turns out to have the same marker. And another. Or the study doesn't replicate at all.
What happened? For one thing, even though most major psychiatric disorders have a genetic component, that component is the sum of hundreds if not thousands of genes, each with a small effect, not to mention epigenetic influences on the expression of those genes.
It would take comparisons of millions of subjects with different diagnoses to show that a particular subtype is associated with a particular constellation of genes. And the finding would likely apply only to a very small percentage of people with the disorder.
In addition, there is a great deal of symptomatic overlap between psychiatric diagnoses (e.g., patients with schizophrenia, bipolar disorder, major depression, delusional disorder, personality disorders and dementia can all experience psychosis). The genetic factors may be more related to discrete symptoms than to overall disorders.
The conclusion? There is not yet any genetic test for any of the psychiatric disorders.
The next question is whether genetics can predict the effects of psychiatric meds.
The body's metabolic machinery for handling, breaking down, and eliminating xenobiotics (foreign material, usually from plants) consists of enzymes. These are proteins that move material in and out of cells (transporters) and receptors. Medications, many of which resemble plants, are affected by these enzymes, some of which have different levels of activity depending on the genes that code for them.
For example, one of the best studied metabolizing enzyme, cytochrome P450 2D6, has 4 major phenotypes depending on whether the genes inherited from mother and father convey great activity, moderate activity, or no activity. Even this apparent simple situation is complicated by the fact that it is possible to have multiple copies of the same functional or nonfunctional gene. The situation is complicated even further by the fact that most medications are metabolized by multiple enzymes, so if activity of one pathway is low, another pathway will hypertrophy to eliminate its substrate normally. Even if genotypes could predict the actual level of a medication in the blood, there is no clearly demonstrated relationship between blood level and clinical effect or side effects for most psychiatric medications.
A number of studies have attempted to take into consideration the actions of networks of enzymes and other proteins coded by multiple genes in predicting which medications will be best tolerated and/or most effective for a given patient. Of the independently funded studies, the Genome-Based Therapeutic Drugs for Depression (GENDEP), the Sequenced Treatment Alternatives to Relieve Depression (STAR*D) study, and the Munich Antidepressant Response Study, which involved a total of 2641 depressed patients, as well as the Clinical Antipsychotic Trials of Intervention Effectiveness (CATIE) study involving 738 schizophrenia patients, found no combination of genetic markers that predicted treatment response.
In not surprising contrast, two studies supported by the manufacturer of a composite report of genotypes done on a total of only 271 patients found a positive result. Unlike the independently funded studies, patients were not assigned to receive genetic monitoring or no monitoring, and treatment and assessments were not blinded to the use of monitoring to guide treatment. As a result, the finding that patients whose antidepressant treatment was guided by genetic testing seemed to have a better outcome could well be due to patients with monitoring feeling better about getting a new approach or evaluators knowing whether monitoring was used. The company did a third study of 51 depressed patients who were randomly assigned to monitoring or no monitoring and had blinded assessments, but there were not significant differences between patients whose medications were or were not chosen according to genotype. In a fourth study from the same company, only 97 patients were studied with equally inconclusive results.
The Bottom Line: The rush to apply each new genetic finding to the clinic before it can be examined critically in studies that take into account the complexity of human neurobiology and experience is an example of the aphorism that the faster you go, the longer it takes to get where you are going. Right now, genetic studies give us an early insight into interacting dimensions of illness that are influenced not just by genes, but by interactions of genes with regulatory components, experience, and the actual illness and its treatment.
The fact that we cannot yet directly translate the influence of genetic factors into practice does not mean that this research is not helpful. When we learn how to consider specific features that are more closely tied to genetic influences, and how to assess the expression and interactions of multiple genes, and when studies are conducted that are designed to compare outcomes in different and very large populations, we will be better able to start to apply gene network findings to predicting aspects of treatment outcome.
But no matter what emerges in further research, we will never be able to do without the expertise and experience of clinicians and the depth of human experience. Anything less than that would be demeaning for patients and boring for clinicians."
Thanks so much, Dr Dubovsky, for clarifying the murk that surrounds genetic testing.
Decoding the genome has been one of mankind's greatest intellectual achievements- an affirmation of all that is best in human nature. Selling the genome before its time for greedy commercial gain is an example of intellectual dishonesty that reflects a less exalted part of human nature.
The work of understanding psychiatric disorder will require decades, not years. Most seeming break-throughs will turn out to be busts. There will be no home runs, no walks, many strike-outs, and only occasional singles. Progress will be steady, but frustratingly slow. In the meantime, the good news is that we already have very effective treatments, if only they were much more accessible and applied more specifically to those who really need them.
Leave a comment