2 August 2011. Upsetting the balance of excitatory and inhibitory signaling in the cortex can induce cognitive and social impairments in mice, according to a study published July 27 in Nature. Led by Karl Deisseroth of Stanford University, the researchers found that selective increases in excitation, but not in inhibition, rapidly induced brain and behavioral anomalies associated with autism and schizophrenia, which were reversed upon restoring the balance. This suggests that deficits in these disorders reflect faulty signaling that might be remedied.
The findings bolster the idea that the diverse genetic factors involved in these disorders result in similar phenotypes through a common neural circuit pathology. Brain signals consist of excitatory ones that boost activity between neurons, and inhibitory ones that rein it in, and together these govern how information flows through the brain. Multiple lines of evidence suggest that schizophrenia and autism are marked by excess excitation and/or impoverished inhibition: genetic studies of autism and schizophrenia finger defects in synaptic proteins that often result in increases in excitation or decreases in inhibition (Rubenstein, 2010, and see SRF related news story); brain imaging points to hyperexcitation in autism (Gomot et al., 2008); deficits in inhibition-related molecules are found in postmortem tissue in schizophrenia (Beneyto et al., 2011); and electroencephalography (EEG) reveals oscillations consistent with these deficits (see SRF hypothesis).
To move beyond circumstantial evidence, Deisseroth's team directly manipulated excitatory and inhibitory signaling in the cortex of awake, behaving mice. To do this, they engineered new optogenetic techniques (see SRF related news story) to selectively activate neurons for time periods long enough to observe an effect in freely moving animals, and to separately activate excitatory pyramidal cells and inhibitory interneurons with different wavelengths of light in the same animal in order to probe their relative contributions to behavior.
Toward excitation, away from social interaction
First authors Ofer Yizhar and Lief Fenno began by tweaking channel rhodopsin 2 (ChR2), a light-sensitive channel that triggers action potentials by admitting cations into the neurons in which it is expressed. They engineered two mutations in ChR2 that, once activated by a single light flash, allowed it to stably depolarize cells for 30 minutes. Next, they used an adeno-associated virus vector, combined with other genetic techniques, to express the channel in either excitatory pyramidal neurons or parvalbumin (PV)-containing inhibitory interneurons in the medial prefrontal cortex (mPFC). Extracellular recordings in anesthetized mice, whole-cell recordings in brain slices, and staining for cFOS in brain tissue all verified that this technique was working as it should: activating pyramidal cells with a flash of light increased neuron activity in the mPFC, whereas activating the interneurons suppressed it.
The light-absorbing pocket of C1V1, a new opsin activated by light wavelengths longer than typical opsins. Image credit: Yizhar et al.
Elevating the ratio of excitatory to inhibitory signaling (also termed E/I balance) by increasing excitation disrupted social and cognitive behaviors, but decreasing this ratio by boosting inhibition did not. Mice with an elevated E/I balance spent substantially less time exploring another mouse introduced into their cages compared to control animals. In contrast, those with a decreased E/I balance were no different from controls. In a fear-conditioning task, mice with an elevated E/I balance during conditioning did not respond to the conditioned cues 24 hours later, when their E/I balance was presumably normal. Afterwards, reconditioning them without manipulating their circuitry showed that they were capable of learning the cues. In contrast, mice with decreased E/I balance showed no cognitive impairments in this paradigm, behaving much like controls.
Mobility, specificity, and rhythmic activity
Looking closer at the elevated E/I balance effects, the researchers found that these deficits were not readily explained by motor impairments because mice in the elevated E/I balance condition explored an open field and a novel object normally. They also performed on an elevated plus maze normally, suggesting that the condition was not associated with increased anxiety.
These impairments may stem specifically from alterations in mPFC, a region implicated in social behavior and complex cognition. In another test of social behavior that lets a mouse choose to spend time in a chamber containing another mouse, in an empty chamber, or in the center chamber connecting the other two, control mice prefer to spend time with the other mouse. But with an elevated E/I balance in mPFC, mice lost this preference. Compared to their baseline behavior (prior to the light flash), these mice spent less time in the chamber containing another mouse and more time in the empty chamber. When an elevated E/I balance was induced in the visual cortex, this shift was not seen, which argues that social impairments do not result when the E/I balance of any bit of cortical circuitry is disrupted.
Because EEG anomalies have been reported for autism and schizophrenia, the researchers also assessed changes in rhythmic brain activity in these mice. Elevating excitation with a flash of light in the mPFC induced an increase in high-frequency oscillations, with a peak of 80 Hz. This change could be recorded in awake, behaving mice, which allowed the researchers to observe a concomitant decrease in social exploration (compared to the same mice before the light flash).
Restoring the balance
If tipping the E/I balance toward excitation causes these impairments, would restoring the balance by simultaneously boosting inhibitory neurons activity fix them? To address this question, the team developed two complementary opsins that could be selectively activated by two different wavelengths of light.
One of these channels was introduced into excitatory pyramidal neurons in the mPFC, and the other into the PV-interneurons—essentially giving the researchers a way to independently dial up the activity of these two cell types in the same animal. In the three chamber social task, activation of interneurons alone did not alter a mouse's baseline preference for spending time with the other mouse, whereas activation of excitatory neurons alone abolished this preference. When both excitatory neurons and inhibitory neurons were simultaneously activated, the preference re-emerged—but only just, as it was not as pronounced as at baseline.
Despite this, the result is a powerful one, because it suggests that rectifying an E/I imbalance may be a viable treatment strategy. Previous evidence pointing to an E/I imbalance in autism and schizophrenia could not easily discern whether such an imbalance directly contributed to the abnormal behaviors associated with these disorders, or was instead a consequence of some other causal factor. The ability of optogenetics to acutely change circuits can cut through this conundrum, and the new study argues for a causal role for an E/I imbalance. The remarkable tools developed in this study will allow future research to thoroughly probe the extent to which specific circuitry changes shape behavior, in health and in mental illness alike.—Michele Solis.
Ofer Yizhar, Lief E. Fenno, Matthias Prigge, Franziska Schneider, Thomas J. Davidson, Daniel J. O’Shea, Vikaas S. Sohal, Inbal Goshen, Joel Finkelstein, Jeanne T. Paz, Katja Stehfest, Roman Fudim, Charu Ramakrishnan, John R. Huguenard, Peter Hegemann & Karl Deisseroth. Neocortical excitation/inhibition balance in information processing and social dysfunction. Nature. 2011 July 28. Abstract