MIAMI BEACH— Understanding the pathophysiology and pathogenesis may lie behind treatment for, and eventual prevention of, schizophrenia. Achieving this understanding is complicated because “what is missing with schizophrenia is a clinical pathological correlation—that is, the association of a clinical syndrome with a pathological entity,” said David A. Lewis, MD, Professor in the Departments of Psychiatry and Neuroscience and Associate Director for Basic Research at Western Psychiatric Institute and Clinic at the University of Pittsburgh.

Dr. Lewis’ strategy has been to begin with one aspect of the clinical syndrome, such as working memory, and to try to understand the underlying pathogenesis and pathophysiology of this component. According to Dr. Lewis, cognitive deficits, impairments of attention and executive function, and certain types of memory deficits appear to represent core features of the illness.

BEST PREDICTOR: COGNITIVE IMPAIRMENT

Although these cognitive deficits, albeit in mild forms, are present in individuals who are at genetic risk for illness and do not manifest in a clinical sense, they are present during the premorbid and prodromal phases of the illness long before psychosis becomes evident, he said. In contrast to psychosis, which can wax and wane over the life of an individual, cognitive deficits are persistent. And perhaps most importantly, said Dr. Lewis, the degree of cognitive impairment is the best predictor of long-term functional outcome in individuals with schizophrenia. Some of these disturbances in cognition, such as impairments in working memory, appear to reflect dysfunction of the prefrontal cortex, he noted at the 57th Annual Meeting of the American Academy of Neurology.

As Dr. Lewis recounted, patients with schizophrenia are unable to increase the activation of the dorsolateral prefrontal cortex at higher levels of working memory load and show poor performance on memory tasks. The impaired activation of the prefrontal cortex and the poor behavioral performance appear to be “at least somewhat specific to” the clinical symptoms of schizophrenia in that they are not present in individuals with depression or other forms of psychosis.

The starting point in detecting the disease process was attempting to understand the pathological entity associated with working memory impairment.

UNDEREXPRESSION OF GAD67

Among genes that show altered expression in the prefrontal cortex of individuals with schizophrenia, disturbances in a group of genes that regulate gamma-aminobutyric acid (GABA) neurotransmission were of particular interest because of primate studies demonstrating that GABA neurotransmission in the prefrontal cortex is a critical mediator of working memory function. This led Dr. Lewis’ team to suspect that “it was possible that a disruption or alteration in gene expression in GABA neurons might somehow be an underlying neurobiological substrate for the working memory disturbances seen in schizophrenia.” Additional research examined these genes in more detail—in particular the genes encoding the 67-kilodalton isoform of glutamic acid decarboxylase (GAD67), an enzyme responsible for the synthesis of GABA, and a GABA membrane transporter (GAT1) that is responsible for the reuptake of synaptically released GABA into the axon terminal.

GAD67 was especially interesting to the investigators, as research had demonstrated this gene was underexpressed in the prefrontal cortex of those with schizophrenia. In fact, Dr. Lewis pointed out, this alteration appears to be the most robust and most replicated observation in postmortem studies of schizophrenia. “Of course, the most immediate question that comes to mind is, if we are looking at the brains of individuals who have died with schizophrenia, how do we know this alteration in GAD67 expression represents the disease process and not the treatment of the illness?” he acknowledged. Dr. Lewis outlined the three lines of convergent evidence suggesting that it reflects the disease process and not the treatment: First, this observation is present in individuals with schizophrenia who are not receiving treatment at the time of death. Second, it is not present in individuals who were treated with antipsychotics for psychotic depression. Third, these changes in gene expression are not present in primates that were exposed for more than one year to atypical or typical antipsychotics in a manner, and with serum levels and side effects, that mimics those used in clinical practice.

“If we really want to understand the circuitry of the prefrontal cortex, we need to know if this deficit in GAD67 expression is something that is common to all GABA neurons in the cortex, or might it be restricted to a subset?” Dr. Lewis said. This question is important because different classes of GABA neurons play different roles in controlling the activity of pyramidal cells, the major excitatory neurons of the cortex, he pointed out.

Referring to a study by Volk et al, Dr. Lewis noted that patients with schizophrenia showed a decrease in the density of neurons that express GAD67, especially in the middle cortical layers, but no difference in the levels of expression per neuron. These findings, in concert with studies demonstrating that the total number of neurons in the prefrontal cortex is not altered in individuals with schizophrenia, suggest that 70% to 75% of GABA neurons in the prefrontal cortex are normal, at least in terms of GAD67 expression. But, he said, a subpopulation is so altered that GAD67 is no longer detectable.

IDENTIFYING THE SUBSET

The question for Dr. Lewis’ investigative team then became, if there is a subset of GABA neurons that is selectively affected, can that subset be identified? One way of classifying GABA neurons, aside from their functional roles and synaptic connections, is by their expression of calcium-binding proteins. “So we were interested in comparing the GABA neurons that express parvalbumin, which provide synaptic input to the perisomatic region of pyramidal cells, versus GABA neurons, which express calretinin, which principally targets additional dendrites of pyramidal cells,” Dr. Lewis elaborated.

Parvalbumin, he continued, is principally expressed in GABA neurons in the middle cortical layers, whereas calretinin tends to be expressed in GABA neurons that are located more superficially. There was no difference in the expression of calretinin messenger RNA between patients with schizophrenia and controls but there was a significant decrease in parvalbumin messenger RNA expression, particularly in the middle cortical layers—the same layers that showed the deficits of GAD67 and GAT1 messenger RNA expression. Furthermore, analysis at the cellular level indicated that the density of neurons that expressed parvalbumin messenger RNA was not changed in schizophrenia. This confirmed to the researchers that these neurons are still present but underexpressing this gene. Additionally, Dr. Lewis and colleagues found that in dual-label studies nearly 50% of the parvalbumin-positive neurons in patients with schizophrenia lacked detectable GAD67—an observation that was never made in controls.

A subclass of parvalbumin-containing neurons that interested the researchers because of their distinctive functional properties was the chandelier neurons, which form inhibitory synapses exclusively on the axon-initial state of pyramidal cells, Dr. Lewis said. A study by Pierri et al found a significant decrease in detectability of GAT1 immunoreactive axon cartridges, especially in the middle cortical layers, “the same layers in which we observe the principal gene expression findings,” Dr. Lewis commented. The same study demonstrated that the deficit in GAT1-positive axon cartridges was found in about 80% of patients with schizophrenia. As Dr. Lewis asked, “What is actually going on at this inhibitory input to the axon-initial segment of the pyramidal cells? Is there too little, or too much, GABA? One would reason that if the more proximal problem is the deficit in the expression of GAD67, there will be a decrease in synthesis of GABA and too little GABA released synaptically. Alternatively, if the proximal problem is a deficit in GAT1, then the GABA that is released will remain at the synaptic level, and what we’ll have functionally is too much GABA.”

One way to distinguish between these two possibilities is to examine the GABA-A receptors that are postsynaptic to the input. Dr. Lewis’ reasoning was that too little GABA would cause the receptors to be up-regulated, and too much GABA would cause the receptors to be down-regulated. Dr. Lewis and his colleagues found that the GABA-A receptors containing α₂ subunits, which are predominantly found in axon-initial segments, were markedly increased in subjects with schizophrenia.

“What we would really like to have is a drug that would selectively activate GABA-A receptors that contain a2 subunit and do so only when GABA is normally released from chandelier neuron axon terminals,” Dr. Lewis allowed. Chronic activation of the receptors or a change in the firing rate of chandelier neurons should be avoided because “timing in this system is of utmost importance.” The ideal drug possesses properties currently available in some agents—that is, they are positive allosteric modulators capable of increasing chloride ion flow through GABA-A receptors only when GABA is present. However, Dr. Lewis continued, currently available benzodiazepines are active in GABA-A α₂ receptors, but also GABA-A receptors that contain α₁ subunits, which mediate sedation, and α₅ subunits, which change hippocampal function—and as a consequence have a substantial cognitive cost associated with them.

“What we need is a benzodiazepine-like drug, but [one] with selective activity at α₂ receptors, and we think that such a drug, given the upregulated state of GABA-A α₂ receptors at the axon-initial segment of pyramidal cells, could have substantial specificity in targeting those receptors,” Dr. Lewis said. There is now such a compound in phase II clinical trials.

—Heidi W. Moore

Suggested Reading
Lewis DA, Hashimoto T, Volk DW. Cortical inhibitory neurons and schizophrenia. Nat Rev Neurosci. 2005;6:312-324.
Lewis DA, Volk DW, Hashimoto T. Selective alterations in prefrontal cortical GABA neurotransmission in schizophrenia: a novel target for the treatment of working memory dysfunction. Psychopharmacology (Berl). 2004;174:143-150.
Pierri JN, Chaudry AS, Woo TU, Lewis DA. Alterations in chandelier neuron axon terminals in the prefrontal cortex of schizophrenic subjects. Am J Psychiatry. 1999;156:1709-1719.
Tamminga C, Hashimoto T, Volk DW, Lewis DA. GABA neurons in the human prefrontal cortex. Am J Psychiatry. 2004;161:1764.
Volk DW, Austin MC, Pierri JN, et al. Decreased glutamic acid decarboxylase67 messenger RNA expression in a subset of prefrontal cortical gamma-aminobutyric acid neurons in subjects with schizophrenia. Arch Gen Psychiatry. 2000;57:237-245.

Return to table of contents