Somatostatin neuron contributions to cortical slow wave dysfunction in adult mice exposed to developmental ethanol

Abstract

Introduction: Transitions between sleep and waking and sleep-dependent cortical oscillations are heavily dependent on GABAergic neurons. Importantly, GABAergic neurons are especially sensitive to developmental ethanol exposure, suggesting a potential unique vulnerability of sleep circuits to early ethanol. In fact, developmental ethanol exposure can produce long-lasting impairments in sleep, including increased sleep fragmentation and decreased delta wave amplitude. Here, we assessed the efficacy of optogenetic manipulations of somatostatin (SST) GABAergic neurons in the neocortex of adult mice exposed to saline or ethanol on P7, to modulate cortical slow-wave physiology.

Methods: SST-cre × Ai32 mice, which selectively express channel rhodopsin in SST neurons, were exposed to ethanol or saline on P7. This line expressed similar developmental ethanol induced loss of SST cortical neurons and sleep impairments as C57BL/6By mice. As adults, optical fibers were implanted targeting the prefrontal cortex (PFC) and telemetry electrodes were implanted in the neocortex to monitor slow-wave activity and sleep-wake states.

Results: Optical stimulation of PFC SST neurons evoked slow-wave potentials and long-latency single-unit excitation in saline treated mice but not in ethanol mice. Closed-loop optogenetic stimulation of PFC SST neuron activation on spontaneous slow-waves enhanced cortical delta oscillations, and this manipulation was more effective in saline mice than P7 ethanol mice.

Discussion: Together, these results suggest that SST cortical neurons may contribute to slow-wave impairment after developmental ethanol.

Keywords: GABA; closed loop optogenetics; cortical interneurons; fetal alcohol spectrum disorder (FASD); prefrontal cortex; slow wave sleep (SWS); somatostatin.

FIGURE 1

Slow-wave sleep in adult SST cre × Ai32 mice exposed to ethanol at P7 was impaired compared to saline controls. (A) Representative cortical LFP recording with raw LFP shown in bottom trace and time-frequency pseudocolor spectrograph shown in center. Top trace shows r.m.s. delta amplitude and 1 standard deviation above mean delta amplitude (red line) extracted from a 24 h recording. Periods of delta amplitude above the red line were classified as slow-wave sleep (horizontal blue markers) and coincided with behavioral inactivity as previously described (Wilson et al., 2016; Lewin et al., 2018). Slow-wave changes in ethanol mice compared to saline (n = 8/group) included (B) shorter SWS bout durations and (C) increased sleep-wake transitions, which together constitute sleep fragmentation. (D) In addition, delta amplitude oscillations were modified to display significantly fewer high amplitude waves. Horizontal lines highlight regions of significant post hoc comparisons between treatment groups (p < 0.05). (E) Stereological cell counts were used to evaluate the number of somatostatin cells in neocortex, in sections double labeled with ant-GFP and anti-somatostatin antibodies (n = 5 mice/group). In 85% of the SST-Cre cells, SST-immunolabeling (arrowheads) was found, but it was not confirmed in the remaining cells (arrow). We did not find SST-immunolabeled cells that lacked SST-Cre. Scale bar = 10 um. (F,G) P7 Ethanol exposure significantly reduced SST cell count as assessed in both GFP cells and SST immunolabeled cells. (H) Separation of stereological cell counts by anterior-posterior location indicated that the ethanol-induced reduction of SST cells is similar throughout the neocortex. Measured neuron densities were separated by their rostral-caudal section number. In brains that had more than 9 sections, the final small end sections were combined. This line graph provides a qualitative overview of local differences, as the stereological strategy was designed to sample the whole cortex, and data from each section includes an average of only 24 sampling sites from each brain. (I) The location of the neocortex (white lines) that was sampled for stereological estimates of SST neuron number and density is shown on a reconstruction of one brain made from block-face images taken during sectioning. Every 12th 50 uM thick coronal section through the neocortex was sampled. Asterisks signify significant difference between ethanol and saline conditions in all panels.

FIGURE 2

(A) Adult SSTcreXAi32 mice exposed to EtOH (n = 9) or saline (n = 5) at P7, received optical stimulation of SST neurons in prefrontal cortex. (B) Histological section showing typical prefrontal cortex optotrode location (including surface damage during histology). Associated atlas image below. White and Blue markers denote electrode (gray) attached to optical fiber (blue). PrL, prelimbic cortex; FrA, frontal association cortex, OFC, orbitofrontal cortex. Blue shows DAPI staining and green shows SST-GFP expression. (C) Histological and atlas section showing visual cortex electrode location (including surface damage during histology). RS, retrosplenial cortex, V1, primary visual cortex, V2, secondary visual cortex. (D) Prefrontal cortex optical stimulation evoked a robust, positive wave lasting 200–400 ms, which corresponds to a 2–5 Hz oscillation (delta frequency band) in P7 saline-treated adult mice and which traveled the anterior-posterior extent of the neocortex, producing a similar, though smaller wave in the visual cortex. The same stimulation in P7 ethanol-treated adult mice evoked an early field potential without the later evoked slow-wave. No evoked slow-wave was observed in the visual cortex of P7 ethanol-treated mice. Shown are means (solid line, n > 5 mice) and SEM. Blue mark indicates 473 nm, 50 ms flash in prefrontal cortex. Repeated measures ANOVA detects a significant difference between evoked waveforms in saline and ethanol treated mice at time points marked by green horizontal line. In PFC, the difference between saline and ethanol responses was primarily >200 ms post flash.

FIGURE 3

PFC unit recordings from adult SSTcreXAi32 mice exposed to P7 saline or ethanol (n = 4 saline mice, n = 47 units; 4 EtOH mice, n = 46 units). (A–D) Rasterplots and peri-stimulus time histogram examples of diverse unit responses to 50 ms, 473 nm light (vertical blue mark) in PFC of saline-treated mice. (A) A putative SST neuron in an ethanol exposed mouse excited during the light followed by suppression (relative to pre-stimulus activity). Horizontal line represents mean spontaneous activity with red shading representing ± 2 S.D. (B) A putative SST neuron in a saline-exposed mouse excited during the light and showing an excitatory rebound 300–400 ms later. (C) An ethanol-exposed non-SST neuron (i.e., no excitation to light) showing a delayed excitatory response 400–500 ms post flash. (D) A saline-exposed non-SST neuron displaying suppression 100–200 ms post-flash and an excitatory rebound at >400 ms. (E) Proportion of all units showing early (<200 ms) suppression and late (>200 ms) excitation in adults exposed to P7 saline or ethanol. There was a significant decrease in the probability of showing a late excitation response in the ethanol-treated mice (asterisk, p < 0.05). (F) Proportion of optogenetically identified putative SST neurons showing early (<200 ms) suppression and late (>200 ms) excitation in adults exposed to P7 saline or ethanol. There was a significant decrease in the probability of showing a late excitation response in the ethanol-treated mice (asterisk, p < 0.05) compared to saline controls.

FIGURE 4

Using adult SSTcreXAi32 mice exposed to EtOH (n = 6) or saline at P7 (n = 6) (A), closed loop optical stimulation of cortical SST GABAergic neurons in frontal cortex (B) with 50 ms flashes triggered on spontaneously occurring slow-waves, induced sustained slow wave activity in saline but not EtOH mice. Blue marks signify timing of light flashes. (C) A 50 ms delay in light activation reinforced slow-wave activity compared to no light, while a 0 ms delay was much less effective. Gray and blue marks signify timing of light flashes starting either 0 or 50 ms, respectively, post slow-wave sleep peak. (D) Over a prolonged period (30–60 min) of such stimulation, delta band oscillation amplitude was significantly enhanced in saline controls stimulated at the 50 ms delay but not at the 0 ms delay. Asterisks = sig. diff between delays. (E) In P7 EtOH treated adult mice, neither stimulation protocol was effective at enhancing delta oscillation amplitude. (F,G) Replotting the data shows the robust enhancement of delta amplitude in saline treated mice, but not in P7 EtOH treated mice, suggesting that stimulation of spared SST cortical neurons cannot compensate to improve sleep-related slow-wave amplitude. Asterisks = significant post-hoc test differences between groups.