α₂-Adrenergic receptor and isoflurane modulation of presynaptic Ca²⁺ influx and exocytosis in hippocampal neurons

Abstract

Background

Evidence indicates that the anesthetic-sparing effects of α₂-adrenergic receptor (AR) agonists involve α_2A-AR heteroreceptors on non-adrenergic neurons. Since volatile anesthetics inhibit neurotransmitter release by reducing synaptic vesicle (SV) exocytosis, we hypothesized that α₂-AR agonists inhibit non-adrenergic SV exocytosis and thereby potentiate presynaptic inhibition of exocytosis by isoflurane.

Methods

Quantitative imaging of fluorescent biosensors of action potential (AP) evoked SV exocytosis (synaptophysin-pHlourin) and Ca²⁺ influx (GCaMP6) were used to characterize presynaptic actions of the clinically used α₂-AR agonists dexmedetomidine and clonidine, and their interaction with isoflurane, in cultured rat hippocampal neurons.

Results

Dexmedetomidine (0.1 μM, n = 10) or clonidine (0.5 μM, n = 8) inhibited AP-evoked exocytosis (54 ± 5% and 59 ± 8% of control, respectively; p < 0.001). Effects on exocytosis were blocked by the subtype-nonselective α₂-AR antagonist atipamezole or the α_2A-AR selective antagonist BRL 44408, but not by the α_2C-AR selective antagonist JP 1302. Dexmedetomidine inhibited exocytosis and presynaptic Ca²⁺ influx without affecting Ca²⁺ coupling to exocytosis, consistent with an effect upstream of Ca²⁺-exocytosis coupling. Exocytosis coupled to both N-type and P/Q-type Ca²⁺ channels was inhibited by dexmedetomidine or clonidine. Dexmedetomidine potentiated inhibition of exocytosis by 0.7 mM isoflurane (to 42 ± 5%, compared to 63 ± 8% for isoflurane alone; p < 0.05).

Conclusions

Hippocampal SV exocytosis is inhibited by α_2A-AR activation in proportion to reduced Ca²⁺ entry. These effects are additive with those of isoflurane, consistent with a role for α_2A-AR presynaptic heteroreceptor inhibition of non-adrenergic synaptic transmission in the anesthetic-sparing effects of α_2A-AR agonists.

Introduction

General anesthesia is a reversible drug-induced state of neurological unresponsiveness characterized by amnesia, unconsciousness and immobility in response to painful stimuli. The molecular and cellular mechanisms that produce these key pharmacological features are poorly understood.¹ All general anesthetics modulate synaptic transmission and neuronal excitability, altering the balance between excitation and inhibition and reducing connectivity in central nervous system networks.^2,3 The principal molecular targets underlying these cellular and network effects include both ligand-gated and voltage-gated ion channels.^1,4 Dexmedetomidine (DEX) and clonidine (CLO) are not general anesthetics themselves, but produce sedative-hypnotic and anesthetic-sparing effects through activation of G protein-coupled α_2A-adrenergic receptors (α_2A-ARs).^5,6 The downstream targets coupled to α_2A-AR activation that contribute to their anesthetic-sparing effects are incompletely characterized.

A well-described effect of α₂-AR agonists is suppression of norepinephrine release from noradrenergic locus coeruleus (LC) neurons through inhibitory autoreceptor activation.⁷ This mechanism was originally suggested to underlie the sedative action of DEX.⁸ However, genetic analysis of the functional roles of α₂-AR subtypes in adrenergic and non-adrenergic cells indicates that the sedative-hypnotic effects of α₂-AR agonists are mediated not by presynaptic α_2A-AR autoreceptors but rather by α_2A-AR heteroreceptors on non-adrenergic neurons.⁹ Moreover, the cellular locations and actions of these critical non-adrenergic neuronal α_2A-ARs responsible for the sedative and anesthetic-sparing actions of α₂-AR agonists are unknown.¹⁰

Volatile anesthetics are known to inhibit the release of multiple neurotransmitters through direct presynaptic mechanisms, including more potent inhibition of the release of glutamate, the principal excitatory neurotransmitter in the central nervous system, compared to other neurotransmitters.^11–14 Since α₂-AR agonists reduce requirements for general anesthetics,¹⁵ we hypothesized that they also affect non-adrenergic synaptic transmission through presynaptic effects on evoked neurotransmitter release. Reduced excitatory transmission resulting in alteration of the balance between neuronal excitation and inhibition has been implicated in the effects volatile anesthetics,^1,16 and provides a plausible mechanism for the well known pharmacological interaction underlying the anesthetic-sparing effects of α₂-AR agonists.

α_2A-ARs are expressed widely in neurons throughout the central nervous system,^17,18 primarily at presynaptic rather than postsynaptic sites,^9,19 consistent with a role for presynaptic α_2A-ARs on non-adrenergic neurons in their neuropharmacological effects. Suppression of both excitatory and inhibitory neurotransmission by α₂-AR agonists has been shown by electrophysiological recordings in brain slices,^20,21 but since neurotransmitter release was not measured directly, these synaptic effects could be mediated postsynaptically or indirectly through intrinsic noradrenergic afferents rather than by direct presynaptic actions on heterosynaptic α_2A-ARs. We therefore studied the effects and α₂-AR receptor subtype specificity of the clinically used α₂-AR agonists DEX and CLO and their pharmacodynamic interaction with isoflurane on action potential (AP)-evoked synaptic vesicle (SV) exocytosis and presynaptic Ca²⁺ influx in cultured rat hippocampal neurons using quantitative biosensor fluorescence live-cell imaging approaches^22–24.

Materials and Methods

Reagents and solutions

Dexmedetomidine (DEX), clonidine (CLO), atipamezole, BRL 44408, and JP 1302 were purchased from Tocris Bioscience (Bristol, UK); ω-conotoxin GIVA and ω-agatoxin IVA from Alomone Labs (Jerusalem, Israel); bafilomycin A1 from Calbiochem (San Diego, CA), and isoflurane from Abbott (Chicago, IL). All other reagents were purchased from Sigma-Aldrich (St. Louis MO). The synaptophysin-pHluorin (syn-pH) construct was kindly provided by Yongling Zhu (Northwestern University, Chicago, IL), and the GCaMP6 construct was kindly provided by Loren L. Looger (Janelia Farm Research Campus, Howard Hughes Medical Institute, Ashburn VA).²⁴

Isoflurane-saturated stock solutions (~12 mM) were prepared and diluted daily into gas-tight glass syringes, from which a sample was taken for determination of aqueous isoflurane concentration. Solutions were perfused focally onto imaged cells via a 150-μm diameter perfusion pipette using polytetrafluoroethylene tubing to minimize isoflurane loss. Concentrations used corresponded to 1–3 times the minimum alveolar concentration (MAC) in rat corrected to 30°C (0.35 mM).²⁵ Perfusate samples were taken at the tip of the perfusion manifold to determine delivered isoflurane concentrations, and reflected ~10% loss from the syringe to the pipette tip. Isoflurane concentrations were determined by extraction into n-heptane (1:1 v/v) followed by analysis using a Shimadzu GC-2010 Plus gas chromatograph (Tokyo, Japan) with external standard calibration.²⁶

Hippocampal neuron culture and transfection

Experiments were approved by the Weill Cornell Medical College Institutional Animal Care and Use Committee and conformed to NIH Guidelines for the Care and Use of Animals. Hippocampal CA3–CA1 regions were dissected from neonatal Sprague-Dawley rats (1–3 days old, ♂ and ♀), and cells were dissociated and plated as described.²² Neurons were transfected on DIV 7–8 with the syn-pH or GCaMP6 construct using the Ca₂PO₄ method, and imaging experiments were conducted on DIV 14–21.²³ Randomization was not used to assign experimental conditions as the pharmacologcal approaches used required specific sequential applications of drugs, therefore the experimenter was not blinded to the conditions. Rat pups born from at least three different parents were used for each set of experiments.

Synaptic vesicle exocytosis

Synaptophysin-pHlourin (syn-pH), a fusion protein of the engineered pH-sensitive green fluorescent protein pHluorin fused to the lumenal N-terminal tail of the synaptic vesicular protein synaptophysin, was used as an optical biosensor of SV exocytosis.^22,27 Changes in fluorescence (ΔF) during electrical stimulation of AP firing reflect alkalization of pHlourin due to exocytosis, while changes during the post-stimulus period reflect re-acidification following endocytosis.²² The transfection method yielded only several transfected cells per dish due to the low transfection efficiency such that boutons from single cells could be identified and imaged without interference from other neurons.

Live-cell imaging was performed at 30.0 ± 0.2°C with continuous superperfusion at 0.27 ml min⁻¹ with Tyrode’s solution containing (in mM) 119 NaCl, 2.5 KCl, 2 CaCl₂, 2 MgCl₂, 25 HEPES buffered to pH 7.4, and 30 glucose, with 10 μM 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX) and 50 μM D,L-2-amino-5-phosphonovaleric acid (AP5) included to block excitatory synaptic transmission and recurrent excitation. Fluorescence images were acquired with an Andor iXon1 camera (model DU-897E-BV; South Windsor, CT) with a solid-state diode pumped 488 nm laser shuttered using acousto-optic modulation. Data were acquired at 10 or 100 Hz by integrating for 30 or 9.74 ms in frame transfer mode and restricting imaging to a sub-area of the CCD chip. The ΔF for exocytosis in response to AP trains was defined as the difference between the average of 2–10 frames before and after the stimulus.

Action potentials (APs) were evoked by stimulation with 1-ms current pulses yielding fields of ~10 V cm⁻¹ using platinum-iridium electrodes. Cells were allowed to rest for ~60 s between 1–20 AP trains and at least 5 min between 10 Hz 200–600 AP trains. Experiments were followed by a maximally depleting stimulus (1200 APs at 10 Hz) in the presence of the v-ATPase inhibitor bafilomycin A1 (0.5 μM), which prevents SV re-acidification following exocytosis, to determine total recycling pool (TRP) size, and then by perfusion with 50 mM NH₄Cl (substituted for 50 mM NaCl and buffered to pH 7.4) to define the total pool (TP) by alkalization of all vesicles. Fluorescence measurements are expressed as a fraction of the TP (Figure 1).²³

Figure 1. α₂-Adrenergic receptor agonists inhibit action potential-evoked synaptic vesicle exocytosis from hippocampal neurons.

A, Schematic diagram of protocol to test the effects of α₂-adrenergic receptor (AR) agonists on synaptic vesicle (SV) exocytosis. Filled boxes indicate electrical stimulation at 10 Hz for 60 s (600 action potentials (APs)), followed by 3 cycles of 600 AP stimuli in the presence of the α₂-AR agonists 0.1 μM dexmedetomidine (DEX) or 0.5 μM clonidine (CLO) with at least 5 min rest between each stimulation. B, Representative fluorescence images of synaptophysin-pHluorin (syn-pH) expressing boutons at rest (upper panels) and after 600 APs (lower panels) for control (left) and DEX-treated neurons (right). Scale bar, 10 μm. C, Representative traces (gray) and average (red) of syn-pH fluorescence to determine total recycling pool (TRP) size and total pool (TP) obtained from 26 boutons analyzed from a single neuron were stimulated continuously at 10 Hz (bar indicates electrical stimulation) in the presence of 0.5 μM bafilomycin A1 to prevent SV re-acidification. The plateau in fluorescence reflects TRP. Vesicle alkalization with 50 mM NH₄Cl revealed the size of the TP. D, Time series of fluorescence changes, shown every 2.5 s, for 600 APs in the absence (Control) or presence of 0.1 μM DEX (following 2, 7, and 15 min drug exposure). Fluorescence intensities were normalized to the subsequent NH₄Cl response (TP). Bar indicates electrical stimulation. E, Mean values of peak syn-pH response amplitude at 10 s (100 APs) and 60 s (600 APs) of stimulation (boxed areas in D), normalized to control values for each time point. Data are expressed as mean ± SEM. ***p < 0.001 by one-way repeated measures ANOVA followed by Tukey’s multiple comparison test. F, Time series of fluorescence changes, shown every 5 s, for 600 APs in the absence or presence of 0.5 μM CLO (following 7 min exposure). G, Mean values of peak syn-pH response amplitude at 10 s (100 AP) and 60 s (600 AP) of stimulation (boxed areas in F), normalized to control values for each time point. Data are expressed as mean ± SEM. **p < 0.01, ***p < 0.001 compared to respective control by two-tailed paired t-test.

Calcium measurements

GCaMP6 was used to measure intracellular Ca²⁺ ([Ca²⁺]_i) in hippocampal neuron boutons stimulated by field-potential-generated APs in the presence of 2 mM extracellular Ca²⁺ ([Ca²⁺]_e).²⁴ GCaMP6 peak fluorescence (ΔF) for each stimulus was determined by averaging the two highest points post-stimulation and subtracting the average baseline of 10 points prior to AP stimulation.

Image and Statistical Analysis

Boutons were selected for analysis by demonstrating their responsiveness to test applications of NH₄Cl. Peak amplitude at 100 APs was selected at 10 s in the middle of a 10 Hz 20 s stimulation. Bafilomycin A1 and NH₄Cl effects were analyzed as mean plateau values. Images were analyzed in ImageJ (http://rsb.info.nih.gov/ij) with a custom plug-in (http://rsb.info.nih.gov/ij/plugins/time-series.html). Silent boutons, defined as those where the response to 100 APs was smaller than the standard deviation of the baseline before stimulation (ΔF₁₀₀ – σ ≤ 0), were excluded from analysis; < 10% of boutons did not respond to stimulation and were excluded from analysis. Based on previous studies,^13,14,22,23 sample sizes of ≥ 5 were used. Data are shown as mean ± s.e.m. ANOVA with Tukey or Bonferroni post hoc tests, two-tailed Student’s t-test (p < 0.05), and 95% confidence intervals were used for testing statistical significance. GraphPad Prism 6 (GraphPad Software, Inc., San Diego, CA) was used for statistical analysis.

Results

α₂-AR agonists inhibit synaptic vesicle exocytosis

The effects of α₂-AR agonists on SV exocytosis were studied using live-cell imaging of cultured rat hippocampal neurons. Exocytosis evoked by 60 s of field electrical stimulation at a frequency of 10 Hz across 25–40 boutons led to a rapid increase in fluorescence due to exocytosis and externalization of syn-pH (Fig. 1B). Blocking re-acidification with bafilomycin A1 led to a plateau in fluorescence at ~200 s identifying the total recycling pool (TRP) size (Fig. 1C). Signals were normalized at each bouton to the total pool (TP) obtained by rapid alkalization of the entire labeled vesicle pool using NH₄Cl (Fig. 1C), thus correcting signals for variations in syn-pH expression levels. The time course of the fluorescence change during a train of 600 APs averaged over a population of individual boutons from a single transfected neuron is shown in Fig. 1D. Fluorescence reached a peak that decayed after the stimulus period due to endocytosis and SV re-acidification (Fig. 1D).

AP-evoked exocytosis was inhibited by the clinically used sedative α₂-AR agonists DEX (0.1 μM; Fig. 1B, D, E) or CLO (0.5 μM; Fig. 1F, G), concentrations known to be effective in vitro.^21,28 The more selective α₂-AR agonist DEX applied for 15 min reduced peak exocytosis elicited by 100 APs to 54 ± 5% of control (95% CI [0.42, 0.65], Fig. 1E), with no significant difference in effect between 2 min and 15 min of DEX application indicating a rapid onset of inhibition. The less selective partial α₂-AR agonist CLO inhibited peak SV exocytosis to 59 ± 8% of control (95% CI [0.40, 0.78], Fig. 1G). The degree of inhibition of SV exocytosis was less at the end of 60 s of stimulation (600 AP) compared to 10 s of stimulation (100 AP) for DEX (95% CI [−0.21, −0.01], p = 0.04 by two-tailed paired t-test, n = 10) or CLO (95% CI [−0.36, −0.05], p = 0.016 by two-tailed paired t-test, n = 8), indicating that inhibition can be partially overcome by a longer stimulation period (Fig. 1E, G). During prolonged stimulation in the presence of bafilomycin A1, neither DEX (0.59 ± 0.03, 95% CI [0.51, 0.67], p = 0.95, n = 10) nor CLO (0.61 ± 0.03, 95% CI [0.54, 0.68], p = 0.76, n = 8) affected TRP size as a fraction of TP compared to control (0.60 ± 0.03, 95% CI [0.53, 0.66], n = 11, by two-tailed unpaired t-test; data not shown).

Inhibition of exocytosis is mediated by the α_2A-AR subtype

There are three major α₂-AR receptor subtypes: α_2A, α_2B, and α_2C.29 The specific subtype mediating inhibition of SV exocytosis by DEX and CLO was examined using subtype-selective antagonists. DEX (0.1 μM) reduced peak exocytosis elicited by 100 AP to 57 ± 6% of control (95% CI [0.43, 0.71], n = 9, Fig. 2A). The nonselective α₂-AR antagonist atipamezole (APZ) abolished inhibition of SV exocytosis by DEX, indicating a specific α₂-AR receptor-mediated mechanism (n = 7, Fig. 2B). Treatment with the α_2A-AR selective antagonist BRL 44408 (1 μM) also blocked inhibition by DEX (n = 8, Fig. 2C), while the α_2C-AR selective antagonist JP 1302 (3 μM) had no effect on inhibition of exocytosis by DEX (n = 7, Fig. 2D). Similar receptor subtype selectivity was observed for CLO (data not shown). These findings indicate that DEX and CLO inhibit SV exocytosis exclusively through interaction with of α_2A-ARs.

Figure 2. Effects of α₂-adrenergic receptor antagonists on dexmedetomidine inhibition of synaptic vesicle exocytosis.

Top, Schematic diagram of protocol to test the effect of dexmedetomidine (DEX) on synaptophysin-pHluorin (syn-pH) fluorescence in the presence of α₂-adrenergic receptor (AR) antagonists. Filled boxes indicate electrical stimulation at 10 Hz for 20 sec. A, Top, Time series of fluorescence changes in the absence (Control) or presence of DEX applied for 7 min, shown every 2.5 s, normalized to total pool (TP) before and after stimulation (horizontal bar). Bottom, Mean effect of 0.1 μM DEX on synaptic vesicle (SV) exocytosis at 10 sec of stimulation (DEX 57 ± 6% of control). Data are expressed as mean ± SEM. **p < 0.01 by two-tailed paired t-test (n = 9). B–D, Effects of α₂-AR antagonists on action potential (AP)-evoked SV exocytosis at 10 s of stimulation (boxes). Top, Fluorescence changes with time before and after stimulation. Fluorescence intensities normalized to TP, with data shown every 2.5 s. Bottom, ΔF at 10 s of stimulation. Data are expressed as mean ± SEM. ***p < 0.001; ns, not significant by one-way repeated measures ANOVA followed by Tukey’s multiple comparison test.

The effects of DEX and CLO on exocytosis evoked by increasing numbers of APs were also investigated (Fig. 3A). Peak exocytosis increased incrementally from 1 to 20 AP stimuli. The degree of inhibition by DEX was comparable across this range of stimuli (57 – 60% of control, n = 7, Fig. 3B), with inhibition to 60 ± 6% of control at 20 APs (95% CI [0.35, 0.80]). Similar results were observed for 0.5 μM CLO (58 – 65% of control, n = 9, data not shown).

Figure 3. Effect of dexmedetomidine on synaptic vesicle exocytosis and presynaptic Ca²⁺ influx in response to increasing stimuli.

A, Representative synaptophysin-pHluorin (syn-pH) fluorescence responses of a single cell relative to total pool (TP) size evoked by 1, 3, 5, 10, 15 and 20 APs in 2 mM [Ca²⁺]_e for control (left traces) or in the presence of 0.1 μM dexmedetomidine (DEX) applied for 7 min after control recording (right traces). Bars on top indicate duration of 20 Hz stimuli. Each trace was averaged from 1–5 trials of 25–40 boutons. Scale bar = 1% TP, 1 s. B, Inhibition of synaptic vesicle (SV) exocytosis by DEX. Top, Peak syn-pH response as a function of action potential (AP) number in the absence (Control) or presence of DEX (0.1 μM) Data are expressed as mean ± SEM. *p < 0.05, **p < 0.01 compared to control by two-way repeated measures ANOVA with Bonferroni post hoc test (n = 7). Bottom, Mean peak responses of syn-pH normalized to each control response (no significant differences by one-way ANOVA). C, Representative GCaMP6 fluorescence responses elicited by 1, 3, 5, 10, 15 and 20 APs in 2 mM [Ca²⁺]_e for control (left traces) or in the presence of 0.1 μM DEX (right traces) applied for 7 min after control recording. Bars on top indicate duration of 20 Hz AP stimuli. Each trace was averaged from 1–5 trials of 25~40 boutons. Scale bar = 100% ΔF/F₀, 1 s. D, Inhibition of presynaptic Ca²⁺ influx by DEX. Top, Peaks of GCaMP6 response as a function of AP number (1, 3, 5, 10, 15 and 20 APs) in the absence (Control) or presence of DEX (0.1 μM). Data are expressed as mean ± SEM. *p < 0.01, ****p < 0.0001 compared to control by two-way repeated measures ANOVA with Bonferroni post hoc test (n = 6). Bottom, Mean peak responses of GCaMP6 normalized to control for each AP number (no significant difference by one-way ANOVA).

α₂-AR agonists inhibit exocytosis by reducing Ca²⁺ influx

Increases in intracellular Ca²⁺ ([Ca²⁺]_i) were measured in hippocampal boutons using the optogenetic fluorescent reporter GCaMP6 as an indicator of presynaptic Ca²⁺ influx.²⁴ Peak [Ca²⁺]_i increased incrementally from 1 to 20 AP stimuli (Fig. 3C). The degree of inhibition by DEX was comparable across this range of stimuli (55 – 62% of control, n = 6, Fig. 3D), with 62 ± 3% of control at 20 APs (95% CI [0.48, 0.78]), comparable to the degree of inhibition of SV exocytosis. Similar results were observed for 0.5 μM CLO (63 – 74% of control, n = 8, data not shown).

Correlation between AP-evoked SV exocytosis and Ca²⁺ influx was measured over a range of AP stimuli to quantify the efficiency of exocytosis at different degrees of Ca²⁺ influx (Ca²⁺-exocytosis coupling). The relationship between paired exocytosis-Ca²⁺ influx data for DEX overlapped control data (Fig. 4), confirming that the effect of DEX on SV exocytosis is directly proportional to reduction in Ca²⁺ influx with no measurable effect on the Ca²⁺-sensitivity of exocytosis.

Figure 4. Dexmedetomidine reduces Ca²⁺ influx and synaptic vesicle exocytosis without affecting the Ca²⁺ sensitivity of exocytosis.

Exocytosis plotted as a function of Ca²⁺ influx in the absence (control) or presence of 0.1 μM dexmedetomidine (DEX) combining data for synaptic vesicle (SV) exocytosis (from Fig. 3B) with data for Ca²⁺ influx (from Fig. 3D) for 1–20 AP stimuli. Data are fitted to a linear model (exocytosis = 0.0107 ± 0.0014[Ca²⁺]_i – 0.00389±0.00458).

While N-type voltage gated Ca²⁺ channels (VGCCs) are closely coupled to depolarization-evoked release of norepinephrine in sympathetic neurons,^30,31 SV exocytosis in non-adrenergic neurons involves contributions from both N- and P/Q-type VGCCs.^32–35 We used VGCC subtype-specific peptide neurotoxins to evaluate the roles of N- and P/Q-type VGCCs in the inhibitory effects of DEX on hippocampal SV exocytosis (Fig. 5). The specific N-type VGCC blocker ω–conotoxin GIVA inhibited exocytosis to 62 ± 9% of control (95% CI [0.40, 0.83]) following 10 s of 10 Hz stimulation at 2 mM extracellular Ca²⁺ ([Ca²⁺]_e). Treatment with DEX further inhibited exocytosis to 47 ± 10% of control (95% CI [0.22, 0.72]), consistent with an effect of DEX on exocytosis mediated by P/Q-type channels (n = 7, Fig. 5A). The specific P/Q-type VGCC toxin ω-agatoxin IVA inhibited SV exocytosis to 53 ± 10% of control (95% CI [0.29, 0.78]). Treatment with DEX further inhibited exocytosis to 33 ± 7% of control (95% CI [0.17, 0.49]), consistent with an effect of DEX on exocytosis mediated by uninhibited N-type channels as well (n = 8, Fig. 5B). Thus DEX inhibits exocytosis mediated by the two major presynaptic VGCC subtypes known to be coupled to SV exocytosis in the hippocampus.^32–35 Similar results were observed for 0.5 μM CLO (data not shown).

Figure 5. Voltage-gated Ca²⁺ channel subtypes contributing to modulation of exocytosis by dexmedetomidine.

A, Effect of 0.1 μM dexmedetomidine (DEX) on exocytosis in the presence of the selective N-type voltage-gated Ca²⁺ channel (VGCC) antagonist ω-conotoxin GIVA (Cono, 1 μM). Top, Schematic diagram of protocol. Middle, Time series of fluorescence changes with stimulation at 10 Hz for 20 sec. Fluorescence was normalized to TP with data shown every 2.5 s. Bottom, Mean amplitudes of synaptophysin-pHluorin (syn-pHy) responses at 10 s of 10 Hz stimulation normalized to control (n = 7). B, Effect of 0.1 μM DEX in the presence of the specific P/Q-type VGCC antagonist ω-agatoxin IVA (Aga, 0.4 μM). Top, Schematic diagram of protocol. Middle, Time series of fluorescence change with 20 s stimulation at 10 Hz. Fluorescence was normalized to TP with data shown every 2.5 s. Bottom, Mean amplitudes of syn-pH responses at 10 s AP of 10 Hz stimulation normalized to control (n = 8). Data are expressed as mean ± SEM. *p < 0.05, ***p < 0.001 by one-way repeated measures ANOVA followed by Tukey’s multiple comparison test.

Presynaptic interaction between dexmedetomidine and isoflurane

Clinical features of α₂-AR agonists include their ability to produce sedation and to increase the potency of general anesthetics (anesthetic-sparing effect).^5,6 Since volatile anesthetics such as isoflurane also inhibit SV exocytosis in hippocampal neurons,^1,13,14 we studied the interaction between isoflurane (ISO) and DEX in suppressing exocytosis (Fig. 6). ISO inhibited exocytosis in a concentration-dependent manner (n = 5, Fig. 6A): 2 MAC ISO inhibited exocytosis to 63 ± 8 % of control (95% CI [0.45, 0.82]); addition of 0.1 μM DEX further inhibited exocytosis to 42 ± 5% of control (95% CI [0.31, 0.53], n = 7, Fig. 6B), greater than the effect of DEX alone (54 ± 5%, 95% CI [0.42, 0.65]; Fig. 1E). The effect of ISO did not involve activation of α₂-ARs as it was not prevented by the subtype nonselective α₂-AR antagonist atipamezole, which also had no effect on exocytosis alone (n = 7, Fig. 6C).

Figure 6. Dexmedetomidine potentiates isoflurane inhibition of action potential-evoked synaptic vesicle exocytosis.

A, Top, Schematic diagram of protocol. Filled boxes indicate 20 s of 10 Hz electrical stimulation with sequential exposure to 1 minimum alveolar concentration (MAC, 0.35 mM), 2 MAC or 3 MAC isoflurane (ISO) for 5 min. Middle, Time series of synaptophysin-pHluorin (syn-pH) fluorescence changes, shown every 2.5 s, normalized to total pool (TP) before and after 20 s stimulation at 10 Hz. Bottom, Mean effect of ISO on synaptic vesicle (SV) exocytosis at 10 s of stimulation (box) normalized to control (n = 5). B, Effects of 0.1 μM dexmedetomidine (DEX) in the presence of 2 MAC ISO on action potential (AP)-evoked SV exocytosis at 20 s of 10 Hz stimulation. Top, Schematic diagram of protocol. Middle, Time series of syn-pH fluorescence changes, shown every 2.5 s, normalized to TP before and after 20 s stimulation at 10 Hz. Bottom, Mean effect of ISO and ISO + DEX on SV exocytosis at 10 s of stimulation (box) normalized to control (n = 7). C, Effects of the nonselective α₂-adrenergic receptor (AR) antagonist atipamezole (APZ; 1 μM) on 2 MAC ISO inhibition of AP-evoked SV exocytosis at 10 s of 10 Hz stimulation. Top, Schematic diagram of protocol. Middle, Time series of fluorescence changes, shown every 2.5 s, normalized to TP before and after 20 s stimulation at 10 Hz. Bottom, Mean effect of APZ and APZ + ISO on SV exocytosis at 10 s of stimulation (box) normalized to control (n = 7). Data are expressed as mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, ns, not significant by one-way repeated measures ANOVA followed by Tukey’s multiple comparison test.

Discussion

Noradrenergic signaling plays important roles in controlling the endogenous sleep-awake cycle and general anesthesia,³⁶ but the mechanisms involved in the interaction between general anesthetics and the anesthetic-sparing effects of α₂-AR agonists are unclear. Potentiation of general anesthesia (anesthetic sparing) is a characteristic neuropharmacological effect of α_2A-AR agonists.^5,6,37 A well known action of α_2A-AR agonists is their modulation of norepinephrine release through presynaptic autoreceptors (receptors for the same transmitter released by the neuron), but elegant genetic studies targeting cell specific α₂-AR receptor expression indicate that classical presynaptic noradrenergic neuron autoreceptor effects are not involved in this anesthetic-sparing action.^9,10 Here we show that the α₂-AR agonists DEX and CLO inhibit SV exocytosis and Ca²⁺ entry in non-adrenergic hippocampal neurons by a heteroreceptor (receptors for a transmitter not released by the neuron) α_2A-AR–mediated effect. Moreover, this mechanism is additive with inhibition of SV exocytosis by the volatile anesthetic isoflurane (for summary of possible mechanisms see Fig. 7).

Figure 7.

Presynaptic mechanisms relevant to the effects of α₂-adrenergic receptor agonists on synaptic vesicle exocytosis. Inhibition by the volatile anesthetic isoflurane (ISO) is due primarily to inhibition of voltage-gated Na⁺ channels (Na_v) rather than direct inhibition of voltage-gated Ca²⁺ channels (Ca_v) or SNARE proteins. Effects of α_2A-adrenergic receptor (AR) agonists such as dexmedetomidine (DEX) and clonidine (CLO) are mediated by G-protein-coupled receptors, specifically through the Gα_i2 isoform. The relevant downstream targets of G-protein activation by α_2A-AR are unknown; plausible candidates include activation of K⁺ channels, inhibition of Ca_v, and/or by reduced action potential (AP)-induced depolarization through inhibition of Na_v. Use of the fluorescent biosensor synaptophysin-pHlourin (syn-pH; green circles) to measure exocytosis is also shown; increased fluorescence is induced by the increase in pH following synaptic vesicle fusion (right).

Advances in fast microscopic imaging and sensitive fluorescent biosensors allowed us to determine the effects of α₂-AR agonists on both AP-evoked SV exocytosis^22,23,27 and Ca²⁺ influx²⁴ in intact central nervous system neurons without interference from noradrenergic innervation. We observed that the clinically used α₂-AR agonists DEX and CLO inhibited SV exocytosis from hippocampal neurons through activation of α_2A-ARs to reduce Ca²⁺ influx mediated by both N-type and P/Q-type VGCCs. Moreover, the highly selective α₂-AR agonist DEX potentiated the presynaptic effects of isoflurane to reduce SV exocytosis. This provides a putative presynaptic target for the anesthetic-sparing properties of α₂-AR agonists,⁶ a clinically relevant pharmacological interaction that allows reduced dosing of general anesthetics to mitigate their dangerous side-effect profiles.³⁷

Contrary to our findings, conventional electrophysiological studies have not detected DEX or medetomidine effects on basal neurotransmission of Schaffer collaterals in rat hippocampus.^38,39 Field excitatory postsynaptic potentials (fEPSPs) are the sum of individual signals from many cells, including inhibitory GABArgic interneurons that have extensive arbors that innervate pyramidal cells. Individual GABArgic interneurons can powerfully inhibit thousands of excitatory pyramidal neurons. Activity of postsynaptic neurons also contributes to fEPSPs such that fEPSPs are affected both by inhibitory signals from GABArgic interneurons and by postsynaptic mechanisms, possibly explaining the insensitivity of the fEPSP to the effects of DEX. In contrast, the method we used selectively reports activity from presynaptic boutons of only a single cell, most commonly nonGABAergic neurons as detected by vGAT immunoreactivity.¹⁴ Although we did not routinely determine neuronal phonotype at the time of this study, we did screen 29 of the 96 neurons described here using vGAT-Oyster labeling, and all 29 were negative (i.e. not GABAergic and thus assumed to be glutamatergic; data not shown). Further studies of neurons of defined transmitter phenotype will be necessary to determine whether there are transmitter-specific differences in α_2A-AR modulation of SV exocytosis.

We have shown previously that isoflurane inhibits SV exocytosis in hippocampal neurons.^13,14 While the effects of DEX and isoflurane on exocytosis are additive, their molecular mechanisms are distinct since the isoflurane effect is insensitive to α₂-AR antagonism. The inhibitory effects of volatile anesthetics on hippocampal exocytosis are thus independent of α₂-AR coupled G protein signaling, but rather appear to involve primarily direct depression of presynaptic voltage-gated Na⁺ channels (Na_v) to reduce presynaptic excitability.^13,40 In contrast, the sedative and anesthetic-sparing properties of α₂-AR agonists are mediated by G protein-coupled receptors, specifically through the Gα_i2 isoform.⁴¹ The relevant G protein-regulated downstream targets for the presynaptic effects of α₂-AR agonists on Ca²⁺ influx and in turn SV exocytosis remain to be established. CLO can inhibit both N- and P/Q-type Ca²⁺ currents in mouse amygdala slices, consistent with our findings of effects on both pathways of Ca²⁺ entry.⁴² Moreover, a recent study shows that DEX inhibits Na_v1.8 currents in rat dorsal root ganglion neurons by increasing activation threshold and decreasing AP firing, suggesting that α_2A-AR agonists might also affect AP frequency and propagation.⁴³ These parallel pathways of inhibition result in additive effects of isoflurane and α₂-AR agonists on both Ca²⁺ influx and SV exocytosis, thus providing a plausible cellular mechanism for their pharmacodynamic anesthetic-sparing interactions in vivo.^9,10 Electrophysiological studies or optical measurements of action potential waveforms using microbial rhodopsin-based biosensors⁴⁴ should allow further studies to determine directly whether α_2A-AR agonists affect presynaptic AP properties.

The α₂-ARs were among the first presynaptic receptors identified by their role as autoreceptors coupled to inhibition of norepinephrine release.²⁹ Previous studies suggested that the sedative and anesthetic-sparing effects of selective α₂-AR agonists involved reductions in noradrenergic neurotransmission through autoreceptor activation.⁴⁵ However, more recent studies have implicated effects of α₂-AR agonists on non-adrenergic neurons in these actions.^9,46,47 Genetically engineered mice that express α_2A-ARs only in noradrenergic terminals show minimal neurological effects of the α₂-AR agonist medetomidine, including loss of righting reflex and anesthetic-sparing, in contrast to a strong hypnotic effect in wild-type littermates.⁹ In another study in vivo, acute knockdown of α_2A-AR expression in the locus coeruleus failed to affect DEX induced sedation.⁴⁶ These findings suggest that α_2A-ARs on non-adrenergic neurons mediate their sedative effects. Furthermore, dopamine-β-hydroxylase knockout mice that have no synaptic norepinephrine release show enhanced sensitivity to and delayed emergence from DEX-induced hypnosis compared to wild-type mice.⁴⁷ This further supports the concept that norepinephrine release from locus coeruleus neurons is not critical for DEX-induced hypnosis, but rather supports a role for direct α₂-AR agonist actions on non-adrenergic neurons in their sedative and anesthetic-sparing actions. Our findings demonstrate a presynaptic site of interaction between the volatile anesthetic isoflurane and the highly selective α₂-AR agonist DEX in reducing SV exocytosis by blocking Ca²⁺ influx in hippocampal neuron axon terminals, a mechanism previously implicated in the presynaptic anesthetic actions of volatile anesthetics.^13,14

The role of presynaptic α₂-AR agonist-mediated inhibition of neurotransmitter release from non-adrenergic neurons in specific anesthetic endpoints is compelling, but identification of the relevant neuronal networks involved will require further study. The hypnotic effects of α₂-AR agonists were initially proposed to be mediated by presynaptic α_2A-ARs on noradrenergic projections from the locus coeruleus to mimic endogenous sleep mechanisms.^45,48 While indirect neurophysiological studies in vivo continue to invoke this mechanism,⁴⁹ genetic studies provide strong evidence that non-adrenergic, not noradrenergic, α_2A-ARs mediate the sedative, hypnotic and anesthetic-sparing actions of α₂-AR agonists.^9,10,46,47 Plausible targets for the hypnotic and anesthetic-sparing effects of α₂-AR agonists include direct suppression of synaptic transmission in cortico-cortical³ and thalamo-cortical networks.^48,49 A recent study in vivo using the TetTag-hM₃D_q system to record and reactivate neuronal groups activated by DEX implicates the preoptic hypothalamus and neighboring dorsal structures in DEX-induced sedation.⁴⁶

Although our studies were conducted in hippocampal neurons, and are therefore most directly relevant to the amnestic properties of the α₂-AR agonist DEX,^50,51 the effects observed involve fundamental mechanisms regulating SV exocytosis in all neurons.⁵² Given the widespread expression of α_2A-AR receptors,^17,18 this pharmacodynamic interaction is likely to apply throughout the central nervous system. Identification of the locations of the specific non-adrenergic α_2A-ARs involved in the multiple neuropharmacological endpoints essential for enhancing general anesthetic effects is critical to development of more targeted agents with improved specificity and therefore safety. Modulation of synaptic transmission by α_2A-ARs might also contribute to their emerging organoprotective effects, further augmenting their clinical utility by both reducing anesthetic doses and preserving organ function.⁵³