Identifying Pathways Leading to Prefrontal GABA-ergic Interneuron Dysfunction in Schizophrenia

Deficits in several cognitive domains, including working memory and other aspects of executive function, represent core features of schizophrenia (1, 2). These cognitive deficits often precede the onset of frank psychosis, correlate more strongly with long-term functional outcome than do positive symptoms, and occur in attenuated form in first-degree relatives of individuals with schizophrenia. Unfortunately, current antipsychotic treatments are largely ineffective for these deficits, which are a major cause of disability in schizophrenia. Thus, it is critically important to understand the basis of these cognitive deficits in order to devise novel ways of treating them.

Functional imaging studies suggest that cognitive deficits in schizophrenia reflect abnormal activity in the prefrontal cortex (3), and several observations suggest that this prefrontal dysfunction may result from abnormalities in prefrontal GABA-ergic interneurons (4, 5). Studies of postmortem brain tissue have consistently found reductions in prefrontal GAD67, an enzyme that produces the inhibitory neurotransmitter GABA. Reductions in prefrontal GAD67 are particularly prominent in a class of GABA-ergic interneurons that express the calcium-binding protein parvalbumin. Other markers of parvalbumin interneurons are also disrupted in the prefrontal cortex in schizophrenia, suggesting that the dysfunction of prefrontal parvalbumin interneurons may be a key contributor to cognitive deficits in schizophrenia.

In a report in this issue, Kimoto et al. (6) aimed to identify factors upstream of reductions in prefrontal GAD67 in order to understand potential causes of prefrontal dysfunction in schizophrenia. They focused on the immediate early gene Zif268, a transcription factor that is rapidly and transiently activated by neuronal activity. Thus, Zif268 can bridge changes in the activity of a neuron to changes in its function. The GAD67 promoter contains a Zif268 binding site, and Zif268 activation is known to increase GAD67 expression (7–9). Thus, the authors hypothesized that decreases in Zif268 may drive the reductions in GAD67 that are believed to contribute to prefrontal interneuron hypofunction in schizophrenia. To test their hypothesis, the authors quantified GAD67 and Zif268 mRNA levels within the dorsolateral prefrontal cortex in postmortem brain tissue from individuals with schizophrenia and 62 matched nonpsychiatric comparison subjects.

Kimoto et al. found that Zif268 levels were significantly lower in the schizophrenia subjects, by 32%. Furthermore, in the schizophrenia group, Zif268 levels correlated with those of GAD67, consistent with the hypothesis that reductions in Zif268 drive those in GAD67. The authors also performed several critical controls. First, they ascertained that Zif268 levels were reduced even in subjects with schizophrenia who were not taking sodium valproate or benzodiazepines at the time of death. This is important because Zif268 levels reflect neuronal activity—thus, drugs that reduce neuronal activity could introduce confounding decreases in Zif268 levels. Second, they determined that in monkeys, Zif268 levels do not change as a result of chronic exposure to antipsychotics. Third, the authors examined three other immediate early genes. None of them correlated with GAD67 levels, suggesting a specific relationship between Zif268 and GAD67. Fourth, they found that Zif268 levels did not differ as a function of several other variables related to substance use, nicotine use, death by suicide, or use of antidepressants or antipsychotics.

An important question is whether reductions in Zif268 occur specifically within parvalbumin interneurons. The discussion describes a sophisticated approach to address this question using laser microdissection to identify individual parvalbumin interneurons from a subset of the subjects studied by Kimoto et al. (10), and microarrays to measure mRNA levels within identified parvalbumin interneurons. Indeed, in nonpsychiatric comparison subjects, Zif268 was highly expressed in parvalbumin interneurons, consistent with the idea that reduced Zif268 affects these neurons. Furthermore, Zif268 mRNA levels appeared to be lower for parvalbumin interneurons from subjects with schizophrenia, although this reduction did not reach statistical significance, likely because of the limited power of the microarray analysis. One might have expected to find especially large reductions in Zif268 within parvalbumin interneurons from subjects with schizophrenia. However, as the authors note, during laser microdissection, parvalbumin interneurons were identified by the presence of perineuronal nets, which surround parvalbumin interneurons but are abnormal in schizophrenia. Thus, the parvalbumin interneurons sampled by laser microdissection may be the healthiest ones, whereas other parvalbumin interneurons from subjects with schizophrenia may have exhibited much larger reductions in Zif268, but may not have been identified and sampled because of disruptions in their perineuronal nets.

Assuming that reductions in Zif268 and GAD67 occur together in the same neurons, the next question is whether decreases in Zif268 cause reductions in GAD67. If so, this poses the question of what causes reduced Zif268. The authors suggest that since Zif268 normally transduces changes in activity, reductions in Zif268 may reflect diminished excitatory input to GABA-ergic interneurons, possibly as a consequence of abnormalities in cortical excitatory neurons. Another possibility—not mutually exclusive with the first—is that the normal relationship between activity and Zif268 activity may be altered in schizophrenia, disrupting the biological “rheostat” that normally governs the output of GABA-ergic interneurons.

Another important question is that of when, exactly, reductions in Zif268 appear. This study found that Zif268 levels decrease with age in parallel in subjects with schizophrenia and in comparison subjects. However, because there were relatively few young subjects, the study may have missed a decrease in Zif268 levels that occurs early in the course of schizophrenia. This is important, because finding that Zif268 levels decrease early in the course of the illness, rather than prior to illness onset, could have important implications for therapeutic interventions aimed at Zif268.

What does the fact that Zif268 levels are lower tell us about the pathophysiology of schizophrenia? We are beginning to understand pathways that may connect deficits in GABA-ergic interneurons to cognitive deficits in schizophrenia. For example, parvalbumin interneurons, which comprise ∼40% of all GABA-ergic interneurons, generate synchronized, rhythmic (∼25–100 Hz) patterns of neuronal activity called “gamma oscillations” (11, 12). These oscillations may facilitate communication between brain regions and/or other aspects of information processing in neuronal circuits. Thus, parvalbumin interneuron dysfunction may disrupt gamma oscillations in ways that contribute to cognitive deficits—and indeed, many studies have found abnormal gamma oscillations in schizophrenia (13). The Kimoto et al. study may help elucidate the molecular pathways that lead to dysfunction of parvalbumin interneurons and other GABA-ergic interneurons. Interestingly, another study in this issue, by Paterson et al. (14), shows how a schizophrenia risk allele for neuregulin-1 leads to lower levels of a neuregulin isoform that is expressed during the early stages of prefrontal cortical development. Disruptions in neuregulin-1 signaling can elicit parvalbumin interneuron dysfunction and abnormal behaviors in mice (15). Other studies have shown how disruptions of glutamatergic input to GABA-ergic interneurons can elicit reductions in GAD67 as well as schizophrenia-like behavioral phenotypes in mice (16, 17). Thus, it will be important for future studies to determine whether Zif268 might link any of these pathological insults to the reductions in GAD67 expression and prefrontal interneuron dysfunction believed to play an important role in schizophrenia.

For decades, schizophrenia research has been dominated by pathophysiological hypotheses centered on particular neurotransmitter systems. More recently, we have begun to uncover a diverse set of genes implicated in schizophrenia. So far, neither of these approaches has yielded dramatic insights or novel therapies. For skeptics, studies like the one by Kimoto et al. might only add to the list of molecules and markers implicated in schizophrenia. Those with a more sanguine view might offer the following. Somewhere between the multitude of heterogeneous genes and the more limited set of reliable neuropathological findings associated with schizophrenia are cellular processes that represent points of convergence for many pathogenic insults and transform these insults into specific pathophysiological consequences, such as parvalbumin interneuron dysfunction. If deficits in GAD67 are actually relevant to the pathophysiology of schizophrenia, then identifying this role for Zif268 brings us one step closer to finding these critical processes that bridge the genes and the markers, and should represent a potent target for novel treatments.