Policy ForumMedicine

What Are the Right Targets for Psychopharmacology?

See allHide authors and affiliations

Science  17 Jan 2003:
Vol. 299, Issue 5605, pp. 350-351
DOI: 10.1126/science.1077141

Conventional wisdom demands that those who would make medicines—whether in a pharmaceutical company or clinical department—have a validated drug target. Ideally, this is a drug-modifiable protein or nucleic acid that is either directly involved in or able to modify the pathophysiology of a disease. Also required, but sometimes taken for granted, is a “clinical target,” a well-defined risk state, illness, or symptom complex for which the treatment is meant and that can be monitored to test whether the drugs are effective.

Sophisticated approaches—including functional genomics and proteomics—are being harnessed to identify new molecular drug targets. Even as they facilitate the search for targets, these approaches can complicate the definition of diseases. In some cases, techniques such as DNA microarrays are leading to a redefinition of the boundaries of clinical entities. This is most evident in cancer, where disease tissue is readily available for analysis. Patterns of gene expression reveal that what looked like a single disease with the crude tool of histological stains may actually be two or more different cancers that may require different treatment approaches (1). As the complexity of genetic and nongenetic susceptibility factors that contribute to common diseases becomes clearer, many diseases may turn out to be clusters of distinct conditions that bear a strong family resemblance to each other, although differing in details of pathophysiology, ages of onset, course, or treatment response.

Because the major mental illnesses—brain diseases primarily affecting behavior, perception, emotion, and cognition—are common and contribute inordinately to worldwide disability (2, 3), the public health benefit of improving current treatments is undisputed. It is well known that there is a severe drought affecting drug targets in psychopharmacology. Put simply, the major targets of currently available medicines (for example, monoamine neurotransmitter receptors and their transporter proteins) were discovered by investigating the mechanisms of action of existing, serendipitously discovered drugs. Over the past 50 years, there can be little doubt that the successive generations of medicines developed on the basis of these prototypes have substantially reduced the burden of suffering associated with mental illness. At the same time, persistent disabilities experienced by many patients treated with even our best current medicines suggest that we need to do better.

As in many other medical illnesses, the molecular targets currently used for drug development in psychopharmacology have not been convincingly shown to play a role in pathophysiology. Indeed, pathophysiology remains poorly understood for all of the common mental disorders. What is less well recognized is that the boundaries of the disease entities themselves, and therefore the “clinical targets,” are also potentially problematic. Currently, major psychiatric illnesses are defined descriptively and indeed can be diagnosed reliably at the level of clinical signs and symptoms. Like many other genetically complex disorders, the etiology of mental disorders is likely to prove heterogeneous. Unlike many other complex disorders such as hypertension or type II diabetes mellitus, however, there are no objective tests to draw boundaries around a particular clinical state. This being the case, scientific progress generally, and treatment development in particular, can be constrained by the limitations of our diagnostic constructs.

A translational bottleneck.

Citations per year based on PUBMED.

In the absence of objective tests for mental disorders, consensus classification systems were developed of necessity. These are represented by the Diagnostic and Statistical Manual of Mental Disorders, 4th Edition (DSM-IV) (4) and the International Classification of Diseases, 10th Edition (ICD-10) (5). These frameworks for diagnostic classification serve a vital role in facilitating medical communication and clinical care. The strength of these diagnostic manuals, however, lies in their reliability (i.e., different observers arrive at the same diagnosis for the same person), not validity (i.e., defining “natural kinds” of disease). Important clinical features of a putative disorder (for example, psychotic symptoms occurring in the course of a major mood disorder) do not always cosegregate in affected families. On the other hand, some disorders that have been split on a phenomenological basis (for example, schizotypal personality disorder, schizoaffective disorder, and schizophrenia) may share genetic risk factors (6-8). In the case of personality disorders, an important contributor to disability and health expenditures, the majority of patients qualify for multiple DSM-IV diagnoses, suggesting the need for reassessment of the dimensions of behavioral and affective dysregulation that define these conditions. The point is not to criticize the clinical manuals; they have been critical tools supporting progress in clinical care. Rather, it is important to now note that future progress in genetics, neuroscience, and treatment development will require moving beyond current consensus systems of diagnosis (9, 10).

Although objective diagnoses should ultimately come from genetics and neuroscience, the genetic complexity and the early state of the neuroscience of mental disorders means that a satisfactory understanding of the molecular basis for mental disorders is still many years in the future (9, 10). Because major psychiatric syndromes may eventually be understood as families of related disorders that are individually distinguished by specific combinations of genetic and nongenetic susceptibility factors, an etiologically based classification will require iterative redefinitions of phenotype-genotype relationships (11). The problem is that individuals suffering from mental disorders today simply cannot wait for decades. How can we move treatment forward in the interim?

We would argue that clinical neuroscience has already provided the tools to test hypotheses in treatment research that are at least partly independent of current classifications. Schizophrenia illustrates the potential utility of reexamining current psychiatric diagnostic approaches to create new options for treatment development. Affecting almost 1% of the population worldwide, schizophrenia is a symptomatically complex disorder characterized by “positive” symptoms such as hallucinations, delusions, and disorganized thought; “negative” or “deficit” symptoms such as lack of motivation and blunted affect; and “cognitive impairments” in attention, working memory, and aspects of executive function. Delusions and hallucinations are the clinical targets for which all current schizophrenia therapeutics were developed and for which they are most efficacious. Although these symptoms define the core of the DSM and ICD diagnostic construct, it is the motivational and cognitive impairments of schizophrenia that are most associated with the profound long-term disability typically produced by the disease.

One proposal for new therapeutics in schizophrenia is to dissect the DSM entity into component symptom complexes and to develop treatments for these narrower “clinical targets,” rather than trying to develop a “monotherapy” for what is likely to be a heterogeneous disorder. The cognitive deficits in schizophrenia have been fairly well characterized (12-14), and rating scales already exist that could be used to assess response to therapies (15, 16). Cognitive abnormalities in schizophrenia have been an important focus of scientific investigation over the last decade, as illustrated in the figure on page 350. A large number of papers have been published on cognition in schizophrenia as quantified by a MEDLINE search, and a substantial body of data shows abnormalities in prefrontal cortical activation associated with abnormalities in executive function (17-19). Although the effects of drug intervention on these cognitive deficits has largely been studied in animal models, therapies aimed at cognition in schizophrenia might be developed through drugs that target D1 dopamine receptors, nicotinic acetylcholine receptors, or others (20-22).

Although basic studies far outnumber clinical trials in many disease states, regrettably, the figure also indicates what might be termed a “translational bottleneck” in moving advances in our basic understanding of the neural circuitry and neuropharmacology of cognition from the laboratory to the development of medicines targeting this devastating aspect of schizophrenia. During a decade in which publications relating to clinical trials of schizophrenia increased, human clinical trials specifically targeting cognition in this disorder remained virtually nonexistent.

One reason for this “translational bottleneck” may be that, to a substantial extent, the pharmaceutical industry and many academic investigators remain tightly focused on seeking treatments for consensus-defined psychiatric disorders using clinical endpoints that have remained unchanged for decades. The perception of limited opportunities for drug registration targeting new clinical endpoints may contribute to this trend. At the same time, in accepting clinical targets for registration, the policies of regulatory agencies, including the U.S. Food and Drug Administration (FDA), necessarily follow and reflect the collective scientific perspectives of industry and the academic community. As a result, with few exceptions (such as the recent FDA approval of symptomatic treatment of agitation in Alzheimer's disease), psychiatric drug development continues to focus on monotherapy of DSM or ICD entities.

The National Institute of Mental Health has initiated a program to identify new clinical targets poorly served by this strategy (23). The goals of this initiative are to facilitate a broad academic, industry, and regulatory consensus regarding promising new clinical targets for drug development; measures to define these targets as endpoints in human clinical trials; and appropriate experimental designs to evaluate efficacy and clinical significance. In pursuit of these aims, we will be holding six consensus-building conferences over the next 18 months. Reflecting the convergence of public health significance and scientific opportunity, these initial efforts will focus on measurement, methodological issues, and neural mechanisms involved in cognition in schizophrenia. At the same time cognition and schizophrenia can serve as “test cases” for government, academic, and industry collaboration that, if successful, might be applied to other clinical targets such as anhedonia in depression, avolition in schizophrenia, or impulsivity in personality disorders where neuroscience is beginning to yield clues about circuitry and putative mechanisms. If NIMH's convening activities can achieve broad consensus regarding clinical endpoints and experimental designs, we are hopeful that regulatory agencies will embrace these new clinical targets, providing academics and industry with powerful incentives to focus development programs.

We believe it likely that the psychopharmacology of the future will entail the dissection of complex individual clinical profiles into component dysfunctions more proximate to pathophysiology than our current global diagnostic concepts. With improved measurement tools, selected symptom complexes can be used as new endpoints in the development of pharmacological and psychosocial therapeutics. Another advantage of this narrower approach is that it might be possible to draw on evolving cognitive neuroscience and neuroimaging research to develop objective biomarkers for cognitive deficits in schizophrenia (and other symptom complexes) that could be used to monitor treatment outcome, even while biological markers for the overall syndrome of schizophrenia as defined in the DSM remain elusive.

Clearly, this approach will not obviate the need to pursue genetics and molecular biology to identify valid disease entities and pathophysiologically relevant molecular targets in the brain. Nonetheless, it would be shortsighted and, ultimately, tragic to permit a rigid adherence to a reliable, but provisional, diagnostic system and static set of clinical endpoints to inhibit the development of potentially important symptomatic therapies.

References and Notes

View Abstract

Stay Connected to Science

Navigate This Article