Chemogenetics-mediated acute inhibition of excitatory neuronal activity improves stroke outcome

Abstract

Background and Purpose

Ischemic stroke significantly perturbs neuronal homeostasis leading to a cascade of pathologic events causing brain damage. In this study, we assessed acute stroke outcome after chemogenetic inhibition of forebrain excitatory neuronal activity.

Methods

We generated hM4Di-TG transgenic mice expressing the inhibitory hM4Di, a Designer Receptors Exclusively Activated by Designer Drugs (DREADD)-based chemogenetic receptor, in forebrain excitatory neurons. Clozapine-N-oxide (CNO) was used to activate hM4Di DREADD. Ischemic stroke was induced by transient occlusion of the middle cerebral artery. Neurologic function and infarct volumes were evaluated. Excitatory neuronal suppression in the hM4Di-TG mouse forebrain was assessed electrophysiologically in vitro and in vivo, based on evoked synaptic responses, and in vivo based on occurrence of potassium-induced cortical spreading depolarizations.

Results

Detailed characterization of hM4Di-TG mice confirmed that evoked synaptic responses in both in vitro hippocampal slices and in vivo motor cortex were significantly reduced after CNO-mediated activation of the inhibitory hM4Di DREADD. Further, CNO treatment had no obvious effects on physiology and motor function in either control or hM4Di-TG mice. Importantly, hM4Di-TG mice treated with CNO at either 10 minutes before ischemia or 30 minutes after reperfusion exhibited significantly improved neurologic function and smaller infarct volumes compared to CNO-treated control mice. Mechanistically, we showed that potassium-induced cortical spreading depression episodes were inhibited, including frequency and duration of DC shift, in CNO-treated hM4Di-TG mice.

Conclusions

Our data demonstrate that acute inhibition of a subset of excitatory neurons after ischemic stroke can prevent brain injury and improve functional outcome. This study, together with the previous work in optogenetic neuronal modulation during the chronic phase of stroke, supports the notion that targeting neuronal activity is a promising strategy in stroke therapy.

Introduction

Ischemic stroke causes marked restriction of energy supply to the affected brain, which results in membrane depolarization, glutamate release, and ionic imbalance, and thereby activates a cascade of events that cause rapid neuronal cell death in the ischemic core and potentially progressive cell death in the penumbra.¹ In this acute phase after stroke, perturbed neuronal activity is believed to be a major component contributing to infarct formation and expansion.^{2, 3}

Ischemic stroke leads to an increase in glutamate release, which is believed to cause excitotoxicity in neurons via excessive depolarization of the postsynaptic membrane.⁴ Further, peri-infarct depolarizations (PIDs) are electrophysiologic events that manifest as cortical waves of neuronal and glial depolarization that propagate from the marginal, penumbral region of cortical infarcts into normally perfused tissue, similar in most characteristics to potassium-induced cortical spreading depolarizations (CSDs).^{1, 5, 6} Importantly, the frequency and duration of PIDs correlate closely with the final infarct volume, suggesting that PIDs play a crucial role in infarct expansion.^7-9 Notably, by investigating both CSDs and spontaneous PIDs in the ischemic penumbra, Hinzman et al showed that excitotoxicity and CSDs/PIDs are not necessarily independent events, and in fact, excitotoxity may be a consequence of these spreading depolarizations (SDs).¹⁰ Collectively, secondary brain injury initiated during the acute phase of stroke is crucially linked to changes in neuronal activity that contribute to excitotoxicity and PIDs.

Investigations on the effects of induced changes in activity of specific neuronal cell types on stroke outcome, either acutely or during recovery, are now possible due in large part to technological advances that permit sophisticated manipulation of neuronal activity.^{11, 12} For example, recent studies have used optogenetics to study specific subsets of neurons in the mouse brain post stroke, and have shown that repeated stimulation of excitatory neurons in the ipsilesional primary motor cortex improves functional recovery.¹³ Similarly, inhibition of striatal GABAergic neurons increases vascular density in the penumbra and improves long-term function, while activation of these neurons worsens stroke outcome.^{14, 15} In these studies, optogenetic manipulation of neuronal activity was initiated several days after stroke, and corresponding data demonstrated that targeted stimulation of neuronal activity improves recovery during the chronic phase of stroke. However, no study had yet attempted to precisely modulate activity of a specific neuronal subset during the acute phase of stroke. Interestingly, it has been shown that sensory stimulation to evoke cortical activity during the acute phase of ischemia is protective if delivered by whisker deflection, but is detrimental by forepaw stimulation,^{8, 16} suggesting distinct involvement of activity of different neuronal subsets in defining stroke pathogenesis. As the first step toward clarifying the roles of different neurons in the acute stroke phase, we selectively manipulated excitatory neurons in the present study. We hypothesized that selective inhibition of forebrain excitatory neurons during the acute phase of stroke improves stroke outcome.

Here, we used a Designer Receptors Exclusively Activated by Designer Drugs (DREADDs)-based chemogenetic approach to non-invasively inhibit forebrain excitatory neurons, and then examined the effects on stroke outcome. DREADDs belong to a class of chemogenetically engineered proteins that are based on G protein-coupled receptors. DREADDs can be activated or silenced by small molecule actuators such as clozapine-N-oxide (CNO). The hM4Di DREADD is commonly used to inhibit neuronal activity.¹² It has been proposed that inhibition of neuronal activity by hM4Di activation is mediated by 2 mechanisms: 1) through induction of post-synaptic membrane hyperpolarization by inwardly rectifying potassium channels, and 2) through inhibition of the pre-synaptic release of neuronal transmitters.¹² In the current study, we used a transgenic mouse line in which the transgene hM4Di was expressed widely in forebrain excitatory neurons. We provide the first direct evidence that targeted inhibition of excitatory neurons in the acute phase can significantly improve stroke outcome, which could be attributed partly to its suppressive effects on the incidence of ischemia-induced spreading depolarizations.

Methods

Animals.

Animal experiments were approved by the Duke University Animal Care and Use Committee. All studies were conducted in accordance with the United States Public Health Service’s Policy on Humane Care and Use of Laboratory Animals. Mouse lines R26-LSL-hM4Di-DREADD (#026219) and Emx1^Cre/Cre (#005628; C57B1/6 background) were purchased from The Jackson Laboratory (Maine, USA). The R26-LSL-hM4Di-DREADD mouse line has been backcrossed to C57Bl/6 for more than 6 generations in the Jackson Laboratory and for 2 more generations in our laboratory. We generated R26-LSL-hM4Di-DREADD;Emx1-Cre (hM4Di-TG) mice to restrict inhibitory hM4Di DREADD expression predominantly to forebrain excitatory neurons and Emx1-Cre mice for controls. Genotyping was performed by PCR analysis using mouse-tail genomic DNA samples and the following primers: the primer set 1 for the hM4Di gene (5’-TCATAGCGATTG TGGGATGA-3’ and 5’-CGAAGTTATTAGGTCCCTCGAC-3’) and the primer set 2 for the Cre gene (5’-GGTCGATGCAACGAGTGATGAGG-3’ and 5’- GCCAGATTACGTATATCCTGGCAG-3’). The online tool Quickcalcs (http://www.graphpad.com/quickcalcs/) was used to randomly assign animals to groups.

Animal surgery.

Animal surgery was performed on male mice (2-4 months old). A mouse stroke model of transient middle cerebral artery occlusion (tMCAO) was conducted as described previously.¹⁷ Briefly, mice were first subjected to 5% isoflurane in 30% O₂ balanced with N₂O, and then orally intubated and mechanically ventilated with1.5% isoflurane. Throughout the entire procedure, rectal temperature was maintained at 36.5°C ± 0.2°C by a heating pad and a heating lamp, and regional cerebral blood flow (rCBF) was monitored in the territory of the MCA using laser Doppler flowmetry (Moor Instruments, Devon, UK). Focal brain ischemia was induced by inserting a 6-0 monofilament (Doccol Corporation, Sharon, MA, USA) into the right internal carotid artery and anterior cerebral artery via the external carotid artery. After 45 minutes of ischemia, the filament was removed, and the skin incision was sutured. Mice were then supplemented with 0.5 mL saline and returned to their home cages. Animals with intracranial hemorrhage and those that did not show a reduction in rCBF > 80% during MCAO and a recovery of rCBF > 70% after 5 minutes reperfusion, were excluded. Animals excluded for analysis are listed in Table S1.

CNO administration.

CNO (Cayman Chemical, Ann Arbor, MI, USA) was initially dissolved in DMSO (10 mg/mL), and then diluted with saline (5% DMSO). Animals were treated with CNO at the dose of 2 mg/kg by intraperitoneal (ip) injection.

Physiologic measurements.

Control and hM4Di-TG mice were intubated, mechanically ventilated, and maintained with anesthesia (1.5% isoflurane), as described above. During the entire procedure, rectal temperature was maintained at 37.0 ± 0.2°C by a heating pad. Microcatheters (PE-10) were inserted into the right femoral artery for monitoring blood pressure and into the right femoral vein for drug administration. Blood pressure (via a pressure transducer), heart rate (via ECG recording), and rCBF (via laser Doppler flowmetry) were continuously recorded using the data acquisition system PowerLab (ADInstruments, Denver, CO, USA), and analyzed using LabChart software (ADInstruments).

Behavioral tests.

All evaluations were performed by observers who were blinded to genotype and group assignment. For stroke experiments, all final behavioral tests were conducted on day 1 after tMCAO.

Neurologic scores.

A 48-point scoring system was used to evaluate neurologic deficits, as detailed previously.¹⁸ Briefly, this comprehensive scoring system assesses general status, motor deficits, and sensory deficits. The final score for each animal is the sum of the scores, with 0 = no deficit and 48 = maximal deficit.

Rotarod test.

Mice were placed on an accelerating rotating rod (4-40 rpm; ENV-577M, Med Associates Inc, St Albans, VT, USA), and the latency to fall from the rod was recorded.¹⁹ All mice were trained for 3 days before surgery.

Open field test.

Mice were placed in an open field (50×50×50 cm; CleverSys Inc, Reston, VA, USA), and allowed to move freely.²⁰ Spontaneous locomotor activity was recorded for 10 minutes or 60 minutes with a 20-second delay. Topscan software (CleverSys) was used to calculate the distance traveled and time spent in the center.

Tight rope test.

Mice were placed in the middle of a rope (60-cm long) above the platform, and the time lapse to reach the end of the rope was recorded and analyzed. The maximum testing time was 60 seconds.²¹

Infarct volume.

On postoperative day 3, mice were euthanized. Brains were harvested and stained with 2,3,5-triphenyltetrazolium chloride (TTC), and infarct volumes were measured according to our previous method.¹⁷ Briefly, the area of infarct of each section (1 mm) was measured by subtracting the non-infarcted area in the ipsilateral hemisphere from the total area of the contralateral hemisphere, and then the final infarct volume was calculated by summing the infarct areas in all sections and multiplying by the section thickness.

Immunohistochemistry and Nissl staining.

Mice were deeply anesthetized with isoflurane, and transcardially perfused with saline followed by 4% paraformaldehyde, and fixed brains were then collected. Immunohistochemistry was performed as described.²² Briefly, immune staining with anti-GFP (A11122; Invitrogen, Carlsbad, CA, USA) and NeuN (MAB377; Millipore, Burlington, MA, USA) was performed on frozen sections (25 μm) using a free-floating staining method. For Nissl staining, paraffin-embedded sagittal brain sections (5 μm) were stained with 0.1% cresyl violet solution.²³ Whole-section images were generated by stitching multiple images together, using an Axio Observer Z1 microscope (Carl Zeiss Microscopy, Thornwood, NY, USA).

Slice preparation and in vitro electrophysiology.

Acute hippocampal slices were prepared from control and hM4Di-TG mice (2-3 months old). Mice were anesthetized with isoflurane and decapitated. Brains were rapidly removed, placed into ice-cold artificial cerebrospinal fluid (ACSF; pH 7.4) containing (in mM) 124 NaCl, 3 KCl, 1.25 NaH₂PO₄, 24 NaHCO₃, 2 CaCl₂, 2 MgSO₄, and 10 glucose, and aerated with 95% O₂/5% CO₂. Coronal hippocampal slices (400 μm) were prepared using a Vibratome 1000 plus (Leica Biosystems, Buffalo Grove, IL, USA), and allowed to recover in an oxygenated chamber at room temperature for at least 1 hour. Slices were kept in a submerged slice chamber (Warner Instruments, Hamden, CT, USA) and superfused continuously (7 mL/min) with oxygenated ACSF maintained at 36°C. Note that with this fast perfusion rate in superfused submerged slices, tissue concentrations of exogenously applied compounds (eg, CNO) reach stable levels rapidly (ie, < 60 sec) and completely within the tissue.

After 30 minutes of superfusion, the Schaffer collateral/commissural pathway was stimulated with a bipolar stimulating electrode inserted into the stratum radiatum of the hippocampal CA1 region, delivering single pulses (100 μs) every 30 seconds. Extracellular glass microelectrodes (3-7 MΩ) filled with 0.2 mol/L NaCl were placed into the CA1 stratum radiatum to record evoked field excitatory postsynaptic potentials (fEPSPs) using a Multiclamp 700B amplifier (10X gain). Data were digitized using pClamp software (Molecular Devices, San Jose, CA, USA). After plotting an input/output (I/O) curve to establish the stimulus required to achieve saturation values for fEPSP, the stimulus current was adjusted to evoke an fEPSP of ~60% of maximal amplitude. After a 20-minute baseline recording period, slices from control and hM4Di-TG mice were exposed to ACSF containing 5 μM CNO (< 0.02% DMSO) for 20 minutes. The percentage change in fEPSP amplitude was calculated as the fEPSP amplitude after CNO application divided by the average of the 20-minute fEPSP baseline recording before CNO X 100.

In vivo evoked field potential recordings.

Burr holes (~ 1 mm diameter) were performed over homologous regions of the motor cortex (1 mm anterior to bregma and 1.5 mm lateral to the midline on both side of the brain). To activate cortical pyramidal cells located in layer V by transcallosal stimulation, we placed a coated, custom-made stainless steel bipolar stimulating electrode into the right motor cortex 1.5 mm below the cortical surface. Extracellular glass microelectrodes (3-7 MΩ) filled with 0.2 mol/L NaCl were placed through a burr hole on the contralateral side 1.5 mm deep to record evoked fEPSPs. After plotting an I/O curve to establish the stimulus required to achieve saturation values for fEPSP, the stimulus current was adjusted to evoke an fEPSP of ~60% of maximal amplitude. Single pulses (100 μs) were applied every 30 seconds. After recording a stable baseline, CNO was administered.

In vivo SD recordings.

Control and hM4Di-TG mice (2-3 months old) were intubated for mechanical ventilation under isoflurane anesthesia (2.5% induction, 1% maintenance in 70% N₂/30% O₂), and the femoral artery was catheterized for mean arterial blood pressure (MABP) monitoring and blood gas sampling. Rectal temperature was monitored and maintained at 37.0°C ± 0.2°C. The head was then fixed in a stereotaxic frame. Two small burr holes (0.5 mm) were carefully drilled at the following coordinates: −1 mm lateral and −3 mm AP with respect to bregma, and −1 mm lateral and −5 mm posterior to bregma, for continuous extracellular potential (DC) recording and monitoring using glass microelectrodes. CSD was initiated more anteriorly with the application of KCl (300 μM) through a larger burr hole (~2 mm in diameter) placed anterior to the bregma (−1 mm laterally and +1 mm AP); the dura was left intact during this procedure. Two single-barrel glass microelectrodes filled with isotonic saline (2-3 MΩ) were lowered to a depth of 100 μm into the small burr holes (2 mm apart), and Ag/AgCl ground electrodes were placed under the skin. The amplified (X10) signal was digitized at 1 KHz and recorded with LabChart software (ADInstruments). Relative changes in local CBF were detected by laser Doppler (LD) flowmetry (Moor) with an LD probe (0.46 mm) positioned over the skull between the 2 recording electrodes. Multiple K+-induced CSDs were evoked and then recorded as the CSDs propagated sequentially to the electrodes and blood flow probe. To ensure the occurrence of CSD, we verified that CBF transients detected by the laser Doppler probe were associated with an extracellular potential shift of greater than 4 mV, characteristic of CSD.

Statistical Analysis.

The primary functional outcome for the stroke experiments was infarct volume, which was used to determine group sizes. Data were analyzed using Prism 6 software (GraphPad). Statistical analysis was assessed by Mann-Whitney U test for neurologic scores and CSD frequency, 1-way analysis of variance (ANOVA) for multiple group comparisons, 2-way repeated measures (RM) ANOVA for physiologic parameters, and unpaired Student’s t-test for all other data. The post hoc Bonferroni test was used for pairwise comparisons. Data are presented as mean ± SEM, percentage, or median. The level of significance was set at p < 0.05.

Results

Characterization of hM4Di-TG mice.

To express the inhibitory hM4Di DREADD predominantly in forebrain excitatory neurons, we generated double transgenic hM4Di-TG mice in which Emx1-Cre was used to control spatial expression of the transgene hM4Di.^{22, 24, 25} Consistent with a previous report,²⁴ hM4Di-TG mice showed mCitrine reporter expression (detected by anti-GFP antibody) in the neurons in the forebrain including the hippocampus and cortex (Fig. 1A,C), reflecting the expression pattern of transgene hM4Di. Further, Nissl staining of the sagittal brain sections showed normal brain structure in both control (Emx1-Cre) and hM4Di-TG mice (Fig. S1). The hM4Di-TG mice were healthy and showed no overt phenotype. For the following experiments, CNO was chosen to activate hM4Di in the hM4Di-TG mouse brain due to its excellent blood-brain barrier permeability and widely validated in vivo efficacy.¹²

Figure 1. Characterization of hM4Di-TG mice.

(A) Expression pattern of the hM4Di DREADD in the hippocampus of the hM4Di-TG mouse brain. Brain sections were co-stained for mCitrine (GFP) and NeuN (neuronal marker). (B) Marked reduction in excitatory synaptic transmission after hM4Di DREADD receptor activation in hippocampal slices. Left panel: Representative fEPSP traces recorded in the stratum radiatum in hippocampal slices from control (top) and hM4Di-TG mice (bottom) during the baseline period (pre-CNO) and at the end of 20 minutes CNO application. Right panel: Mean time-course of fEPSP amplitudes during baseline (pre-CNO), CNO application, and recovery in control (n = 3) and hM4Di-TG mice (n = 6). (C) Expression pattern of the hM4Di DREADD in the cortex of the hM4Di-TG mouse brain. Coronal sections were co-stained for mCitrine and NeuN. Enlarged images of the boxed areas are shown at the bottom. (D) Representative traces of fEPSP recorded in the left motor cortex after homotopic contralateral cortical stimulation. The evoked response was suppressed in hM4Di-TG mouse brains after CNO treatment.

To validate the inhibitory effects of hM4Di activation on excitatory neurons in hM4Di-TG mice, we analyzed evoked fEPSPs both in vitro (hippocampal slices; Fig. 1B) and in vivo (homologous motor cortex; Fig. 1D). Application of 5 μM CNO to activate inhibitory hM4Di DREADD in vitro led to a rapid and persistent drop of the evoked fEPSP amplitude, which declined to 17.4% ± 2.0% of baseline values within 20 min (p < 0.001; 2-way RMANOVA post hoc Bonferroni, baseline vs 20 min CNO) and remained suppressed even after a 20-minute wash-out period, confirming the prolonged inhibitory effect. The amplitude of the presynaptic fiber-volley remained stable during the 60-minute recording period. In contrast, the fEPSP amplitude in hippocampal slices from control mice showed no significant change after CNO treatment (Fig. 1B). Of note, both the 10-90% rise time (2.2 ± 0.04ms vs 2.1 ± 0.07ms; control vs hM4Di-TG mice, p = 0.51) and half width (5.3 ± 0.2ms vs 5.7 ± 0.2ms; control vs hM4Di-TG mice, p = 0.28) for the baseline fEPSPs were similar between the genotypes. Next, we confirmed the inhibitory effects of hM4Di activation in vivo by measuring evoked fEPSPs from stimulating homologous, contralateral motor cortex (Fig. 1D). After recording a stable baseline (10 min) in hM4Di-TG mice, CNO administration resulted in a reproducible fEPSP suppression to 42 ± 15% of baseline values (n = 3) after 30 min, while in control mice the amplitude of the evoked fEPSPs remained stable over at least 60 min recording period after CNO injection.

Finally, to assess potential physiologic effects of CNO and CNO-activated hM4Di in the brain, we monitored heart rate, cerebral blood flow, and blood pressure before and after CNO treatment in control and hM4Di-TG mice. No notable changes in these parameters were observed (Fig. S2). However, to exclude potential effects of CNO that are independent of hM4Di, both control and hM4Di-TG mice were dosed with CNO in the following experiments.

Stroke outcome in hM4Di-TG mice after acute hM4Di activation.

After characterizing hM4Di-TG mice, we set out to test our hypothesis that acute activation of hM4Di to inhibit excitatory activity in glutamatergic neurons improves stroke outcome in hM4Di-TG mice. We first demonstrated that CNO treatment had no obvious effect on outcome of motor-coordination tests (ie, rotarod and tight rope) and open field test in wild-type mice (Fig. 2A-D). Then, we tested the effects of hM4Di activation on these tests using control and hM4Di-TG mice (Fig. 2E-H). No significant difference was identified in rotarod and tight rope tests between control and hM4Di-TG mice (Fig. 2E,F). Interestingly, in the open field test, while mice from both groups traveled a similar distance during the 60-minute period, hM4Di-TG mice appeared to spend less time in the center of the arena after CNO treatment, which may indicate a slight change in anxiety (Fig. 2G,H). Overall, the data demonstrated that neither CNO nor CNO-activated hM4Di in the brain had notable effects on behavioral testing (Fig. 2). Also, the body weight of hM4Di-TG mice and control mice was similar (Fig. S3A).

Figure 2. Effects of CNO and hM4Di DREADD activation on behavioral tests.

(A-D) CNO effect. Two groups of C57Bl/6 mice (n =10/group) received vehicle or CNO treatment. Twenty minutes later, mice were subjected to rotarod (A), tight rope (B), and open field (C,D; 10-minute period) tests. (E-H) hM4Di DREADD activation effect. Behavioral tests, including rotarod (A), tight rope (B), and open field (C, D; 60-minute period) tests, were used to evaluate control and hM4Di-TG (TG) mice before and after CNO treatment (n = 5-6/group). Data are presented as mean ± SEM. *, p < 0.05.

For the first stroke experiment, hM4Di-TG mice were dosed with CNO 10 minutes before tMCAO to inhibit excitatory neurons during and after ischemia. No effect of CNO treatment on cerebral blood flow response was observed during the tMCAO procedure in hM4Di-TG mice (Fig. S3B). However, compared to control mice, hM4Di-TG mice exhibited significantly better performance on rotarod and neurologic scoring, and demonstrated smaller infarct volumes (68.7 ± 1.7 mm³ vs 54.6 ± 2.2 mm³ [control vs hM4Di-TG], p < 0.001, t-test) after stroke (Fig. 3). We then delayed neuronal activity suppression until the early reperfusion period and assessed stroke outcome. Control and hM4Di-TG mice were subjected to 45 minutes tMCAO, and at 30 minutes after reperfusion, all mice were dosed with CNO. Again, hM4Di-TG traveled a greater distance in the open field test, performed significantly better on the rotarod test, had a trend for improved neurologic scores, and displayed significantly smaller infarct volumes (60.2 ± 3.0 mm³ vs 46.3 ± 3.7 mm³ [control vs hM4Di-TG], p < 0.05, t-test), compared to control mice (Fig. 4). Taken together, stroke outcome was significantly improved after either pre-stroke treatment or post-reperfusion treatment to inhibit excitatory neurons.

Figure 3. Stroke outcome after pre-treatment with CNO to activate chemogenetics-mediated suppression of neuronal activity in hM4Di-TG mice.

Control and hM4Di-TG (TG) mice (n = 8/group) were dosed with CNO at 10 minutes before 45 minutes of middle cerebral artery occlusion (MCAO). After 24 hours of reperfusion, mice were evaluated for rotarod (A) and neurologic scores (B). Infarct volumes were then determined by TTC staining on day 3 post surgery (C). Data are presented as median value or mean ± SEM. **, p < 0.01; ***, p < 0.001.

Figure 4. Stroke outcome after post-treatment with CNO to activate chemogenetics-mediated suppression of neuronal activity in hM4Di-TG mice.

Control and hM4Di-TG (TG) mice (n = 8-9/group) were subjected to 45 minutes of middle cerebral artery occlusion (MCAO). At 30 minutes reperfusion, all mice were dosed with CNO. After 24 hours of reperfusion, mice were evaluated for spontaneous locomotor activity (A; a 10-minute test period), rotarod (B), and neurologic scores (C). Infarct volumes were determined by TTC staining on day 3 post surgery (D). Data are presented as median value or mean ± SEM. *, p < 0.05; **, p < 0.01.

Potassium-induced CSDs in vivo after activating hM4Di in hM4Di-TG mouse brains.

We speculated that suppression of PIDs could be a mechanism underlying the protective effects observed above. As a first attempt to test this possibility, we used a simple and reproducible in vivo potassium-triggered CSD model, which results in frequent spontaneous SD responses over time. In this experiment (Fig. 5), we measured blood gas to ensure that systemic physiologic parameters were within normal limits for all experimental animals (MABP: 70-85 mmHg; PaO₂: 115 ± 2; PaCO₂: 34.0 ± 1.2, pH: 7.40 ± 0.02). Thirty minutes before KCl (300 μM) application, which induced CSD waves during our 60-minute recording, both hM4Di-TG and control mice received an intraperitoneal (ip) injection of CNO. Compared to control mice, CNO-treated hM4Di-TG mice showed significantly suppressed CSD activity, as indicated by a 36.5% ± 5.5% reduction in CSD frequency (p < 0.01; Mann-Whitney test; Fig. 5B), a prolonged CSD latency (55.01 ± 12.15 seconds vs 159.9 ± 33.7 seconds [control vs hM4Di-TG]; p < 0.05; t-test; Fig. 5C), and decreased CSD propagation rate (3.61 ± 0.18 mm/min vs 4.61 ± 0.36 mm/min [control vs hM4Di-TG]; p < 0.05; t-test; Fig. 5D). Although the amplitude of the negative DC shift was not significantly different between the 2 groups (15.6 ± 1.5 mV vs 12.6 ± 1.6 mV [control vs hM4Di-TG]), its duration was significantly shorter in hM4Di-TG mice (98.2 ± 11 seconds vs 49 ± 2 seconds [control vs hM4Di-TG]; p < 0.001; t-test; Fig. 5E). As expected, reproducible hyperemic responses were observed, with a lower CBF increase in hM4Di-TG mice (140% ± 8.3% vs 108.4% ± 15% [control vs hM4Di-TG]; p < 0.05; t-test; Fig. 5F).

Figure 5. Potassium-induced cortical spreading depolarization (CSD) after activation of hM4Di DREADD receptor in hM4Di-TG mice in vivo.

(A) Representative recordings of extracellular DC potentials in control and hM4Di-TG mice. Both control and hM4Di-TG mice received CNO treatment. After 30 minutes, continuous application of KCl (300 mM) via a cranial window anterior to the recording site was used to trigger SDs. (B-F) Characteristics of KCl-induced CSDs in control and hM4Di-TG mice. CSD frequency (B), CSD latency (C), CSD propagation rate (D), DC-shift duration (E), and CBF changes (F) were calculated during the 60-minute recording following KCl application. Data are presented as median value or mean ± SEM (n = 4-5/group). *, p < 0.05; **, p < 0.01; ***, p < 0.001.

Discussion

Here, for the first time, we applied a novel inhibitory chemogenetic tool in acute stroke research. Using transgenic hM4Di-TG mice, our data showed that acute suppression of neuronal activity limited to excitatory neurons is sufficient to improve stroke outcome. Further, we demonstrated that inhibition of excitatory neurons significantly reduced potassium-induced CSD activity m vivo, a clear parallel physiologic event to PIDs after stroke.

There are some advantages to using a DREADD-based chemogenetic approach to manipulate neuronal activity in stroke research. First, DREADD-expressing neurons can be non-invasively controlled by a simple timed ip injection of CNO. Second, CNO has a rapid and long-lasting effect (hours) on target neurons, which may be required to produce notable effects on stroke outcome. Finally, neuronal activity can be modulated in a large or small brain area by controlling spatial expression of DREADDs. In the current study, the transgene inhibitory hM4Di DREADD was expressed widely in forebrain excitatory neurons in hM4Di-TG mice.²⁶ Using this mouse line, we confirmed the effectiveness of neuronal inhibition after hM4Di activation by CNO treatment, as evidenced by rapid suppression of evoked fEPSPs in both in vitro hippocampal slices and in vivo motor cortex. The dramatic suppression in the hippocampal slices appears to be inconsistent with a recent study describing a mild fEPSP suppression after CNO treatment.²⁷ However, this seeming discrepancy could be explained by several differences between the 2 studies, such as the expression levels of transgene hM4Di (ie, transgenic mice vs virus-mediated gene delivery), the promoter used (ie, Emx1 vs CamK2α), and the experimental condition (see Methods).

Importantly, using the hM4Di-TG mouse line, we provide the first direct evidence that selective inhibition of excitatory neurons during the acute phase of stroke can improve stroke outcome. The significant reduction in infarct size and improvement in neurologic function were observed with either pre-stroke treatment or post-reperfusion treatment. These results largely agree with previous findings that early suppression of glutamatergic input via pharmacologic approaches is beneficial for stroke recovery.⁴ However, our data more specifically indicate the critical and specific role of activity limited to excitatory neurons in brain damage during the hyper-acute phase of stroke. In future studies, we may use chemogenetic tools to manipulate activity of brain cells in a highly selective fashion, and thus clarify the roles of different cell subsets in stroke outcome with detailed spatial and temporal information. Of note, chemogenetic tools can modulate activity of not only neurons, but also astrocytes.²⁸

One mechanism underlying the neuroprotective effects observed in hM4Di-TG mice could be related to amelioration of stroke-induced PIDs. Since CSDs and PIDs share similar cellular mechanisms (intense depolarization) and spread across the cortex,^{6, 29} investigation of potassium-induced CSDs in vivo can be considered an appropriate surrogate measure for PID occurrence and propagation. Here, we show, for the first time, that targeted inhibition of excitatory neurons using a chemogenetic approach can significantly suppress the incidence of CSDs. In line with this finding, 2 recent studies showed that targeted depolarization of excitatory neurons can trigger CSDs in the healthy brain.^{30, 31} Together, these data indicate that neuronal activity is critically involved in CSD initiation and propagation. Therefore, it is likely that in our potassium-induced CSD experiments, the hM4Di-mediated decrease in neuronal excitability leads to a significant reduction in both CSD frequency and duration. This is a critical finding, because it is widely believed that stroke-induced SDs are a key contributor to infarct formation. Thus, a better understanding of how specific subsets of neurons contribute to CSD mechanisms would inform development of effective strategies for therapeutic manipulation of PIDs to improve stroke outcome. In this regard, like optogenetics, chemogenetic tools could be highly useful, as we have demonstrated here.

Our data also show that CNO treatment pre-ischemia vs post-reperfusion appeared to offer more robust neuroprotection. This could be explained by the finding that PIDs start to occur immediately after ischemia onset and then almost completely fade away during early reperfusion.³ Thus, while suppressing excitotoxity could underlie neuroprotection in both pre- and post-treated hM4Di-TG mice, these first, early PIDs would be optimally reduced only in pretreated hM4Di-TG mice. To further clarify the mechanisms, more experiments are warranted.

In summary, we demonstrate proof-of-concept experiments that acute inhibition of a subset of excitatory neurons after stroke can prevent brain injury and improve functional outcome, which may result from a reduction in PIDs and excitotoxity. This study, together with the optogenetics studies mentioned earlier, supports the notion that modulating neuronal activity is a promising strategy for the treatment of stroke.

Highlights.

• CNO treatment of hM4Di-TG mice suppresses evoked synaptic responses in vitro and in vivo.

• hM4Di-TG mice treated with CNO at 10 minutes before ischemia improves stroke outcome.

• hM4Di-TG mice treated with CNO at 30 minutes after reperfusion improves stroke outcome.

• hM4Di-mediated inhibition of excitatory neurons reduces K+-induced CSD activity in vivo.