Zolpidem reduces hippocampal neuronal activity in freely behaving mice: a large scale calcium imaging study with miniaturized fluorescence microscope

Abstract

Therapeutic drugs for cognitive and psychiatric disorders are often characterized by their molecular mechanism of action. Here we demonstrate a new approach to elucidate drug action on large-scale neuronal activity by tracking somatic calcium dynamics in hundreds of CA1 hippocampal neurons of pharmacologically manipulated behaving mice. We used an adeno-associated viral vector to express the calcium sensor GCaMP3 in CA1 pyramidal cells under control of the CaMKII promoter and a miniaturized microscope to observe cellular dynamics. We visualized these dynamics with and without a systemic administration of Zolpidem, a GABAA agonist that is the most commonly prescribed drug for the treatment of insomnia in the United States. Despite growing concerns about the potential adverse effects of Zolpidem on memory and cognition, it remained unclear whether Zolpidem alters neuronal activity in the hippocampus, a brain area critical for cognition and memory. Zolpidem, when delivered at a dose known to induce and prolong sleep, strongly suppressed CA1 calcium signaling. The rate of calcium transients after Zolpidem administration was significantly lower compared to vehicle treatment. To factor out the contribution of changes in locomotor or physiological conditions following Zolpidem treatment, we compared the cellular activity across comparable epochs matched by locomotor and physiological assessments. This analysis revealed significantly depressive effects of Zolpidem regardless of the animal's state. Individual hippocampal CA1 pyramidal cells differed in their responses to Zolpidem with the majority (∼ 65%) significantly decreasing the rate of calcium transients, and a small subset (3%) showing an unexpected and significant increase. By linking molecular mechanisms with the dynamics of neural circuitry and behavioral states, this approach has the potential to contribute substantially to the development of new therapeutics for the treatment of CNS disorders.

Figure 1. Experimental session and imaging data analysis procedure.

A: timeline of imaging sessions. We imaged in “interrupted regime” (“On”, red; “Off”, blue) for 45 minutes in vehicle and Zolpidem periods to collect 30 min of neuronal data in each period. B: Spatial locations of independent components (ICs) corresponding to the individual cells (“IC Spatial Filters”) identified by PCA/ICA cell sorting algorithm. C: Relative fluorescent changes (ΔF′(t)/F′0 = (F′(t) – F′0)/F′0, where F′0 is the projected mean intensity for all frames) for five representative cells (ICs, highlighted on panel B and illustrated by Video S1) are calculated and plotted across time. Ca²⁺ transients were identified by searching each trace for local maxima that had peak amplitude more than two standard deviations (st. dev., y-axis) from the baseline (defined as the median of the trace calculated across the entire session); an occurrence of a calcium transient is indicated as a tick mark.

Figure 2. Zolpidem decreased frequencies of calcium transients in CA1.

A: Calcium transients (indicated by tick marks) detected in individual cells (vertical axis) are plotted across time following vehicle (water, left) and Zolpidem (10 mg/kg, right) administration in a representative animal. B: Histogram of calcium transients (“Event Rate”) in the representative animal. C: Average rate of calcium transients (s.e.m. error bar) in all animals used in the study. Zolpidem decreased the frequency of calcium transients by 71% (from 0.7120 to 0.2087 events/min/cell, p<0.0001, Wilcoxon Signed Rank Test).

Figure 3. Decrease in locomotion was not sufficient to explain Zolpidem-induced decrease in neuronal activity.

A: Raster plot of calcium transients in individual cells (vertical axis) following vehicle (left) and Zolpidem (10 mg/kg, right) administration in an example animal with identified inactive (displacements <−.2 cm/min) periods (green shading). The corresponding speed (in cm/min) trace is plotted below the raster plot. B: Comparison of average frequencies of calcium transients (number of events/minute/cell) during active periods (black bars) and inactive periods (green bars) in the representative animal (s.e.m. error bar). C: Average rate of calcium transients (s.e.m. error bar) in all animals with identified inactive periods following both vehicle (left) and Zolpidem (right) administration.

Figure 4. Individual hippocampal CA1 pyramidal cells differed in their responses to Zolpidem.

A: Representative raster plot of calcium transients in individual cells (n = 195) that are color-coded depending on their response to Zolpidem (Mann-Whitney-Wilcoxon test, p<0.05 criterion of significance): significant decrease (blue); significant increase (red), non-significant change (black). B: Locations of individual cells identified in the same representative imaging session, layered atop a mean fluorescent image. C: Rate of calcium transients post-Zolpidem (“Zolpidem Event Rate”) vs post-vehicle (“Vehicle Event Rate”); each dot is an individual cell (n = 1275). The majority of individual neurons (65%) significantly lowered neuronal activity following Zolpidem administration; 32% of neurons did not show a significant change; and a small neuronal subset (∼3%) showed a significant increase.

Figure 5. Neuronal activity during Zolpidem-induced NREM sleep was lower than neuronal activity during physiological NREM.

A: Raster plots of calcium transients in 478 individual cells (vertical axis) during pre-treatment active wake periods (“Baseline”) and post-treatment NREM periods (“NREM”) in two imaging sessions (“Vehicle” and “Zolpidem”). B: Average frequencies of calcium transients (“Event Rate”: number of events/minute/cell) during pre-treatment active wake (in both Vehicle and Zolpidem sessions, black bars), physiological NREM (“Vehicle”, grey bar) and Zolpidem-induced NREM (“Zolpidem”, grey bar). The error bars are the s.e.m. for each condition across all cells. Zolpidem NREM neuronal activity was significantly lower than vehicle NREM neuronal activity (0.09 and 0.15 events/minute/cell, respectively, 40% change, p<0.004, WSR test).