© Lisa Kyle

While many scientists continue to chip away at the mysterious upstream genetic and environmental factors that cause schizophrenia, neuroscientist David Lewis takes a different approach. Instead of focusing on upstream causes, Lewis's lab at the University of Pittsburgh seeks to uncover conserved downstream abnormalities, including clinical features and protein defects that are shared by most, if not all, individuals with schizophrenia.

There is a great deal of symptomatic variation among schizophrenic individuals, but Lewis argues that some symptoms are conserved in all patients. "Somehow, all of these different processes must converge to produce a smaller number of pathological entities," he explains. "There must be something about what we call schizophrenia, as heterogeneous as it is, that we still recognize reliably by a particular set of signs and symptoms."

Lewis and his colleagues have identified two conserved aspects of the disease. One is cognitive impairment in the form of disturbances in executive function and working memory, which seem to occur in all individuals with schizophrenia. The second is an abnormality in the expression of the gene for glutamic acid decarboxylase-67 (GAD67), which synthesizes gamma-aminobutyric acid, or GABA. "What seems to be most conserved at the level of pathology - the most commonly found and widely replicated in postmortem studies of schizophrenia - is a decreased level of the gene product for [this] enzyme," he says.¹

While no one really knows what causes the GAD67 abnormalities - they might be due to variants in the gene for GAD67, problems with gene expression, or both - the ultimate cause may not really matter, says Lewis, since drugs might be able to regulate neuronal activity by targeting these downstream problems directly.

His lab has discovered that GAD67 abnormalities affect only a subset of neurons: those that produce parvalbumin and somatostatin.² Of note, these cells are also involved in certain types of brain oscillations, characterized by the simultaneous firing of thousands of neurons at specific frequencies. Parvalbumin neurons are involved in gamma oscillations (50 Hz), which are implicated in working memory, whereas somatostatin neurons participate in theta oscillations (4-8 Hz), which are thought to be involved in memory, navigation, and in the coupling of sensory information with motor output.

Lewis believes that disturbances in GABA neuron transmission impair these oscillations and are the substrate for cognitive impairments in schizophrenia, but he admits that there's "still a lot of biology to sort out." To this end, his lab is performing a number of experiments to tease out the link. First, the lab is studying how oscillations change during adolescence, a period when working memory function is known to improve. They are also trying to understand, using animal models, the relative roles of different GABA neurons and GABA-A receptor types in regulating oscillations. Finally, they are working to compare levels of extracellular GABA with oscillation activity. "We're trying to test hypotheses that will home in more on this relationship," he explains.

Lewis's team is also testing, in Phase II clinical trials, a drug developed by Merck that is based on the idea that normalizing GABA function will improve cognitive function. When GABA levels decrease in the brain, transmission at a subset of GABA-A receptors containing alpha-2 subunits also decreases. Lewis's new drug selectively boosts GABA signals at these receptors only when neurons normally release the neurotransmitter, since continually boosting the signal at all GABA-A receptors interferes with unrelated GABA-related brain processes and with the timing of cellular outputs involved in neural oscillations. Benzodiazepines, drugs that increase the affinity of GABA for all GABA-A receptor subtypes, have been shown in clinical trials to improve symptoms of psychosis in patients with schizophrenia, but not without a plethora of side effects, including excessive sedation.

Lewis hopes that his more selective benzodiazepine-like drug will improve cognitive symptoms without these side effects. It is "a very promising strategy for reducing cognitive impairments that have not responded to available antipsychotic medications," writes John Krystal, a clinical pharmacologist and psychiatrist at Yale University, in an E-mail. "The testing of this mechanism for the treatment of schizophrenia is something that many people, myself included, have been awaiting." Data on the drug's efficacy should be available sometime next year, according to Lewis.

Regardless of whether this particular drug makes it to market, Lewis says the ultimate goal is to revolutionize how we think about treating schizophrenia. "We need to move beyond the serendipitous approach to treatment development that we"ve had for the last 50 years," he says. "The really exciting thing about schizophrenia research is [that] we're now seeing a number of examples arising where people are going after novel molecular targets for drug treatment based upon an understanding of the illness."