Finding new treatments for disease begins and ends with the patient, with significant steps in between, said David Lewis, M.D., in a lecture at APA's 2001 annual meeting in Honolulu in May.

The initial clinical observations of the patient must be converted into hypotheses about the disease processes. These can then be tested under controlled conditions in the lab to guide novel treatments for patients, said Lewis, professor and chair of the Department of Psychiatry at the University of Pittsburgh and medical director and director of research at the Western Psychiatric Institute and Clinic.

"To make that work requires a molecular target within the pathological entity that can be affected by a drug," he said. There must also be a biomarker to measure the function of the brain in a living human being. Finally, behavioral changes in the clinical syndrome must be measurable, too.

Lewis uses that approach while focusing on cognitive deficits that are core symptoms in schizophrenia—attention, executive function, and memory.

These are also among the early warning signs of the disorder, said Lewis. Young people who go on to develop schizophrenia score 1.5 to 2 standard deviations below the general population on measures of cognition as early as age 7.

Those cognitive deficits persist across the course of illness, predict long-term functional outcomes, and are often present in milder forms in relatives of people with schizophrenia.

Working memory is particularly impaired in patients with schizophrenia, said Lewis. His search for a hypothesis began with knowledge that GABA activity is needed for the proper working of the dorsolateral prefrontal cortex during working memory tasks. GABA neurons control the timing of activities there and GABA antagonists there disrupt working memory.

"You need the right amount of GABA at the right time for normal working memory," he said.

A form of glutamic acid decarboxylase (GAD₆₇) converts GABA into glutamate. Neurons in people with schizophrenia show an average of 15 percent lower expression of GAD₆₇ mRNA, compared with controls, although with much variance, said Lewis.

"But, are the lower GAD₆₇ mRNA levels a reflection of the disease process or a consequence of it?," he asked.

The former, he said. "They are a conserved feature associated with the diagnosis of schizophrenia that actually affects the disease process."

With too little GAD₆₇, there's too little GABA, he hypothesized, and that leads to a number of compensatory changes in the brain observed in schizophrenia, including a loss of gamma synchronicity, which contributes to working memory impairments.

Simply boosting GABA all the time won't affect the deficit, but selectively augmenting GABA neurotransmitters at the GABA_A-alpha₂ receptors while retaining the synchronicity might help the brain with compensatory responses, he suggested.

Drugs called positive allosteric modulators (such as benzodiazepines) bind to the alpha₂ subunits, but they also bind to alpha₁ subunits, with sedative effects, and to alpha₅ subunits, impairing hippocampal memory functions.

"So we want to find a GABA_A-alpha₂ positive allosteric modulator to enhance GABA neurotransmission onto pyramidal cells only by way of GABA_A-alpha₂ receptors, which would enhance the ability for GABA neurons to entrain pyramidal cells into gamma oscillations and improve working memory," he said.

Lewis and his colleagues tested an experimental drug with those properties, MK-077, in a randomized, double-blind, controlled trial of 15 subjects. (MK-077 was originally developed as a nonsedating antianxiety drug.) They found some improvement on several cognitive tests reflecting processes in the dorsolateral prefrontal cortex, which is known to be impaired in schizophrenia.

"Enhanced signaling through these GABA receptors may improve both the biological and behavioral measures of prefrontal cortical function in people with schizophrenia," he said.

The research also suggests that targeting these receptors in people showing prodromal cognitive impairments serves as a preventive intervention.

"But this is only one pathway," he said. Others must be explored as well.

Although the trial was small and needs replication, it does suggest the approach's value to finding drug targets.

"This translational neuroscience approach to understanding disease processes in schizophrenia could lead us from dependence on serendipity to a rational basis for novel, effective treatments," he stated.