See Allison Curley's snapshots from the conference.
January 28, 2014. Inhibitory interneurons, especially the 25 percent that contain the calcium-binding protein parvalbumin (PV), have emerged as key characters in the ever unfolding schizophrenia story, although their exact role in the illness—and the extent to which they are primary drivers of symptoms or more downstream consequences of other pathological changes—remains to be uncovered (see SRF related news report). A large number of posters at the Neuroscience 2013 meeting held in San Diego explored the role of GABA neuron dysfunction in the impaired oscillations and cognitive deficits of schizophrenia. What follows is a small sample of the work presented on this topic.
At the first poster session of the meeting on Saturday afternoon, November 9, Kristen Delevich of Cold Spring Harbor Laboratory, New York, presented her study of prefrontal cortex circuit functioning in a DISC1 heterozygous mouse, a model of the chromosomal translocation associated with mental illness in humans (see SRF related news report). Using optogenetics, Delevich observed a lower paired pulse ratio of currents evoked from PV—but not somatostatin (SST)-containing interneurons—suggestive of a reduction in the GABA release probability in PV interneurons. As a consequence of this reduced release probability, pyramidal cells in the prefrontal cortex received less feed-forward inhibition from the mediodorsal thalamus. This reduced inhibitory input upset the pyramidal cells' normal inhibitory-to-excitatory balance, providing a possible circuit basis for the impaired executive function that is found in both DISC1 mutant mice and schizophrenia. (For more DISC1 news from the conference, see SRF related conference report.)
Going after GAD67
Although the GABA release probability of SST cells was not affected in the DISC1 mutants, on Sunday afternoon Brad Rocco of the University of Pittsburgh, Pennsylvania, presented immunofluorescence data from schizophrenia postmortem tissue suggesting that these cells are far from healthy. SST degrades very quickly after death, so Rocco and colleagues used the process of elimination to find them. By examining axon terminals that contained the GABA-synthesizing enzyme GAD67, but lacked PV or the other GABA maker GAD65 (which their prior research has shown marks cannabinoid 1 basket cells but not SST terminals), the researchers enriched for SST terminals. Rocco observed a 70 percent reduction in the density of the terminals, as well as lower levels of the GAD67 protein present in them. Given the role of these cells in synaptic integration, the researchers suggest that a deficit in these cells in schizophrenia may lead to improper input/output of large groups of neurons.
On Monday afternoon, Sivan Subburaju, McLean Hospital, Belmont, Massachusetts, described an extension of earlier findings on a network of genes associated with GAD67 regulation that are differentially expressed in schizophrenia and bipolar disorder. Using the HiB5 hippocampal cell culture model, the researchers previously showed that the inhibition of two of these genes—epigenetic regulators HDAC1 and Daxx—produced elevated GAD67 mRNA (Subburaju and Benes, 2012). To examine whether inhibition of the two genes in particular can influence other genes involved in the regulation of GAD67, Subburaju and colleagues knocked them down using lentiviral vectors carrying small hairpin RNAi sequences. The knockdown HDAC1 increased HDAC2 and RunX2 and decreased GluR6, while silencing of Daxx increased RunX2 and decreased Pax5, GluR6, and GluR7 expression. The authors concluded that genes in the GAD67 regulatory network have complex interactions with each other that may influence GABA neuron function.
On Tuesday afternoon, Lichao Chen, Harvard University, Boston, Massachusetts, described the use of RNA interference (RNAi) and the Cre-lox recombination system to examine whether the widely reported reduction in GAD67 mRNA is a cause of schizophrenia symptoms or a compensatory response to other cortical abnormalities. In adult mice, knockdown of GAD67 expression specifically in PV neurons in the prefrontal cortex (PFC) produced a significant reduction in the number of PV neurons that had detectable levels of GAD67, but did not affect the power or frequency of kainate-induced γ oscillations in vitro. In contrast, knockdown in adolescent mice showed a trend toward a reduction in γ band power, suggesting lower GAD67 may be a cause of the altered γ oscillations observed in schizophrenia, but a reduction in GAD67 prior to adulthood is needed to produce the effect.
The path to impaired cognition
On Saturday afternoon, Kathleen Cho of the University of California, San Francisco, used a mouse model to link the deficits in PV cell signaling to the cognitive difficulties of schizophrenia through impaired γ oscillations. She reported that mice heterozygous for both Dlx5 and Dlx6 (two transcription factors that are crucial to PV neuron development) exhibit behavioral and EEG abnormalities that are consistent with schizophrenia: a reduction in γ band power during a social exploration task that depends on the PFC, as well as an increase in baseline γ power. Dlx5/6 heterozygous mice also show impairments on some aspects of a cognitive flexibility task, and these deficits are accompanied by disturbed oscillatory activity.
Two Sunday afternoon posters from Tracie Paine’s group at Oberlin College, Ohio, also used a rodent model to probe the role of GABA transmission on specific aspects of cognitive function in schizophrenia. Work presented by Avery O’Hara found that blockade of GABAA receptors, but not GABA synthesis, impaired decision making on a gambling task in rats. Paine presented a second poster suggesting that blocking GABA synthesis also had no effect on attention as measured by the five-choice serial reaction time task. Together, these data suggest that the widely reported deficit in GAD67 in schizophrenia does not contribute to the decision making and attention deficits found in schizophrenia.
In the same session, Rachel So of the University of Connecticut in Storrs presented her work examining neural synchrony patterns and levels of GABA and glutamate/glutamine during a working memory task in schizophrenia. The researchers found that the strength of the synchronization between the dorsolateral PFC and the occipital lobe increased with task difficulty in the schizophrenia subjects but not controls, and that the synchronization was correlated with GABA (but not glutamate/glutamine) levels in the left dorsolateral prefrontal cortex (DLPFC). So and colleagues suggest that the altered neural synchrony patterns in schizophrenia could represent a compensatory strategy to preserve working memory performance.
The excitatory side
On Sunday afternoon, a poster by Tsung-Ung Woo of Harvard Medical School in Boston, Massachusetts, described the mRNA and microRNA expression profiling of neuronal subtypes in schizophrenia. Pyramidal neurons in the PFC of cases exhibited differentially expressed mRNAs in the transforming growth factor β and bone morphogenetic protein signaling pathways, as well as those involved in the cytoskeleton, extracellular matrix, apoptosis, and oxidative stress. PV neurons from the PFC in schizophrenia displayed differentially expressed genes in pathways such as Wnt, Notch, and prostaglandin E2 pathways, along with transcripts that control the cell cycle and apoptosis. Woo and colleagues also observed a predominant downregulation of microRNAs in the illness.
In addition to receiving inputs from other GABA neurons, PV cells also receive synaptic contacts from excitatory pyramidal cells. One hypothesis is that glutamate receptor hypofunction may be the cause of PV dysfunction in schizophrenia (Gonzalez-Burgos and Lewis, 2012), and several Neuroscience 2013 posters modeled this proposed reduction of excitatory inputs onto PV neurons. On Sunday afternoon, Martha Hvoslef-Eide, Cambridge University, U.K., presented a cognitive characterization of mice with a conditional knockout of NR1 glutamate receptors on nearly half of cortical interneurons early in development (see SRF related news report). She reported that NR1 knockouts were significantly impaired on the continuous performance task, a test of attention on which schizophrenia subjects perform poorly, suggesting that conditional NR1 knockouts are a valuable model for studying cognitive dysfunction in schizophrenia.
Recent in vitro data suggest that synaptic events onto PV neurons may not have a strong NMDA component during adulthood. However, the presence of significant tonic NMDA current in PV neurons suggests a role for extrasynaptic NMDA receptors. On the Tuesday morning session, Eastman Lewis presented his work testing this prediction, finding that functional NMDA receptors are present in adult rat PFC. Consistent with the presence of extrasynaptic NMDA receptors, inhibition of glutamate reuptake enhanced the NMDA currents in PV neurons. The results suggest that modulation of extrasynaptic NMDA currents could be used to restore the balance of cortical excitation and inhibition in schizophrenia.
A trio of posters on Tuesday afternoon from the laboratory of Terry Sejnowski at the Salk Institute in La Jolla, California, examined the effects of postnatal ablation of mGluR5 glutamate receptors in PV neurons. The group has previously shown that PV-specific mGluR5 knockout mice exhibit lower levels of GAD67 and fewer inhibitory synapses onto pyramidal cells than controls. Aaron Kappe examined the neural circuit abnormalities by recording epidural auditory event-related potentials. He reported an increased amplitude of P20 and P80 components in both sexes and a decrease in the amplitude of the N40 component in females. Stephanie Barnes examined the behavioral effects of mGluR5 ablation, finding deficits in recognition (but not spatial) memory and social function. The mice also showed a reduced sensitivity to PCP-induced behavioral abnormalities, hinting at NMDA receptor hypofunction. António Pinto-Duarte found that although the mice seemed to develop normally and lacked any obvious gross abnormalities, PV-specific mGluR5 knockouts showed a reduction in the number of PV synaptic contacts in multiple brain areas such as the hippocampus. Basal synaptic transmission was also impaired, suggesting a deficit in the maturation of PV cells. Together, these results suggest that abnormal mGluR5 modulation of PV neurons may play a role in schizophrenia, and provide evidence of their utility as a therapeutic target in the illness (see SRF related conference report).—Allison A. Curley.