Harnessing cortical plasticity via gabapentinoid administration promotes recovery after stroke

Abstract

Stroke causes devastating sensory-motor deficits and long-term disability due to disruption of descending motor pathways. Restoration of these functions enables independent living and therefore represents a high priority for those afflicted by stroke. Here, we report that daily administration of gabapentin, a clinically approved drug already used to treat various neurological disorders, promotes structural and functional plasticity of the corticospinal pathway after photothrombotic cortical stroke in adult mice. We found that gabapentin administration had no effects on vascular occlusion, haemodynamic changes nor survival of corticospinal neurons within the ipsilateral sensory-motor cortex in the acute stages of stroke. Instead, using a combination of tract tracing, electrical stimulation and functional connectivity mapping, we demonstrated that corticospinal axons originating from the contralateral side of the brain in mice administered gabapentin extend numerous collaterals, form new synaptic contacts and better integrate within spinal circuits that control forelimb muscles. Not only does gabapentin daily administration promote neuroplasticity, but it also dampens maladaptive plasticity by reducing the excitability of spinal motor circuitry. In turn, mice administered gabapentin starting 1 h or 1 day after stroke recovered skilled upper extremity function. Functional recovery persists even after stopping the treatment at 6 weeks following a stroke. Finally, chemogenetic silencing of cortical projections originating from the contralateral side of the brain transiently abrogated recovery in mice administered gabapentin, further supporting the conclusion that gabapentin-dependent reorganization of spared cortical pathways drives functional recovery after stroke. These observations highlight the strong potential for repurposing gabapentinoids as a promising treatment strategy for stroke repair.

stroke, corticospinal tract, gabapentinoids, structural plasticity, functional recovery

Introduction

Stroke remains a leading cause of mortality and long-term disability that imposes an enormous economic and emotional burden on society.¹ As clinically approved treatment options for acute ischaemic stroke mainly focus on the restoration of blood flow and rehabilitation, there is an additional need to develop novel and safe strategies that promote structural and functional rearrangement of neuronal circuits in the brain and spinal cord of afflicted individuals.

In the adult brain, a narrow, time-limited window of spontaneous neuroplasticity opens following a stroke.² Compensatory mechanisms including axonal sprouting enable partial recovery in the subacute phase after stroke. Recovery can be further augmented by rehabilitative training in the acute/subacute³ or chronic phases.⁴ Given that neurons in the adult CNS lose the ability to grow and regenerate as they mature,⁵ spontaneous neuroplasticity and recovery after CNS injury remain limited in adulthood. By manipulating the poor intrinsic growth state of adult neurons and overcoming extrinsic non-neuronal impediments after stroke, experimental strategies successfully trigger robust regrowth of axonal tracts, such as the corticospinal pathway that controls voluntary limb movement and sensory processing.^6–8 Of note, some sprouting axons can functionally integrate into short- and long-range neuronal circuits, effectively contributing to neurological recovery after stroke.^7,9 A critical point to consider is that several of these molecular strategies may not be practical for immediate clinical translation and require further development.

Using a systematic and unbiased approach, we recently discovered that the α2δ2 subunit of voltage-gated calcium channels acts as a developmental switch that suppresses axon growth and regeneration in the adult CNS.¹⁰ An α2δ2 pharmacological blockade via administration of the clinically approved gabapentinoids [e.g. gabapentin (GBP) and pregabalin], drugs used to treat neurological disorders,¹¹ enables axon sprouting and regeneration of sensory ascending and descending corticospinal pathways after spinal cord injury in mice.^10,12 To our knowledge, the effectiveness of the same treatment strategy in promoting structural and functional neuroplasticity after stroke has not yet been investigated.

Here, we show that chronic GBP administration promoted collateral sprouting of the corticospinal pathway from the uninjured hemisphere after photothrombotic cortical stroke in adult mice. Using a combination of tract tracing, electrical stimulation and functional connectivity mapping, we demonstrated that corticospinal collaterals originating from the contralateral side of the brain formed new synaptic contacts and integrated within the brainstem and spinal circuits that control forelimb muscles in mice administered GBP. GBP daily administration also dampened maladaptive plasticity by reducing the excitability of spinal motor circuitry after stroke. Not only did we find that these mice can recover upper extremity function, but we also discovered that chemogenetic silencing of cortical projections from the contralateral side of the brain transiently abrogated recovery. Importantly, functional recovery persisted even after stopping the treatment at 6 weeks following a stroke. This further indicates that GBP-dependent neuroplasticity can drive recovery after stroke in adult mice. Thus, targeting α2δ2 with a readily translatable pharmacological strategy clearly aids structural and functional repair of neuronal circuits after stroke.

Materials and methods

Animals

All animal experiments were performed following protocols approved by the Institutional Animal Care and Use Committee at The Ohio State University. Adult (7–8-week-old) female and male C57BL/6J mice (stock no. 000664, The Jackson Laboratory) were used for all experiments, except those specifying green fluorescent protein (GFP)-M and BAC Aldh1l1-eGFP mice. GFP-M (stock no. 007788; RRID: IMSR_JAX 007788) mice expressing GFP under the control of the Thy1 promoter were purchased from The Jackson Laboratory. Aldh1l1-eGFP mice expressing GFP under the control of Aldh1l1 promoter were kindly provided by Dr Min Zhou (The Ohio State University). Mice were randomly assigned to experimental groups. Experimenters were blind to group assignment and experimental conditions.

Antibodies

The following antibodies were used: rabbit polyclonal anti-βIII tubulin (Tuj1) (immunoblot, IB 1: 20 000, T2200, RRID: AB_262133, Sigma-Aldrich), mouse monoclonal anti-NeuN (immunohistochemistry, IHC 1: 500, MAB377, RRID: AB_2298772, Millipore), rabbit polyclonal anti-α2δ2 (IHC 1: 500, IB 1: 1000, ACC-102, RRID: AB_11124467, Alomone Labs), rabbit monoclonal anti-cFos (9F6) (IHC 1: 1000, 2250S, RRID: AB_2247211, Cell Signaling Technology), guinea pig polyclonal anti-VGLUT1 (IHC 1: 500, 135304, RRID: AB_887878, Synaptic Systems), guinea pig polyclonal anti-glutamate transporter subtype 1 (GLT1) (IHC 1: 1000, IB 1: 10 000, AB1783, RRID: AB_90949, Millipore), rabbit polyclonal anti-glutamate/aspartate transporter (GLAST) (IHC 1: 1000, IB 1: 10 000, ab416, RRID: AB_304334, Abcam) and rabbit monoclonal anti-GAPDH (IB 1: 5000, 2118, RRID: AB_561053, Cell Signaling Technology).

Photothrombotic stroke

Adult mice were anaesthetized with a mixture of ketamine (100 mg/kg body weight) and xylazine (10 mg/kg body weight). After injecting Rose Bengal [intraperitoneal (i.p.), 30 mg/kg in 0.9% saline, Sigma], anaesthetized mice were transferred to a stereotaxic frame (RWD). A midline incision was made, and the surface of the skull was dried using sterile cotton swabs. Five minutes after injection of Rose Bengal, a cold LED light source (KL2500 LED, Schott) was used to illuminate the sensory-motor cortex via a flexible fibre optic cable (155.100, Schott). To restrict the area of the skull that was illuminated, a mask with a small aperture (2 mm wide and 4 mm long) was applied on the intact skull. The mask was centred using the following anterior–posterior (AP) coordinates from bregma in mm: AP 1.0/1.3, 0.5/1.3, 0/1.3, −0.5/1.3. The LED light source was tuned to consistently deliver 20.5 klux to a light meter placed 120 mm away from the fibre optic output (via a 3D printed adapter). After 15 min of illumination, light exposure was stopped and the wound was sutured. The procedure for the sham operation was the same without illumination of the skull. The mice were then placed on soft bedding in their home cage on a warming surface held at 37°C until they were awake and alert. Beginning 1 h or 24 h after stroke, GBP (46 mg/kg body weight, PHR1049, CAS: 60142-96-3, Sigma) or the corresponding volume of vehicle (0.9% saline, B. Braun) was administered (intraperitoneal injections, three times per day for the first week, two times per day until the end of the study). In the cohort of mice administered vehicle or GBP beginning 24 h after stroke, the treatment was discontinued at 43 days after stroke and behavioural data were collected for two additional weeks. Solutions were stored at room temperature and replaced every 3–4 days.

TTC staining and measurement of infarct area

Brains were dissected 24 h after stroke or sham operation. Using a mouse brain slicer (68713, RWD), 2-mm coronal sections were cut. For analysis of the infarct area, sections between bregma 4.0 to 1.0 mm ± 0.5 mm were immediately immersed in 2% TTC (2,3,5-triphenyltetrazolium hydrochloride, T8877, Sigma) in 0.9% saline at 37°C for 10 min and imaged using a digital camera. The infarct area of each brain was calculated as the percentage of the total area of the brain section (ImageJ). Sections were then transferred to a 4% paraformaldehyde (PFA) solution for immersion fixation.

Vasculature labelling and 3D imaging

The brain vasculature was traced as reported elsewhere.^13,14 Briefly, mice were transcardially perfused with 4% PFA in PBS (pH 7.4). Mice were then perfused with 5 ml of 0.05% albumin-tetramethylrhodamine isothiocyanate bovine (A2289, Sigma) in 2% gelatin from porcine skin (G1890, Sigma). At the time of injection, the temperature of the gel solution was maintained at 45°C. After clamping the heart, mice were placed on ice to lower their body temperature and to allow for gel formation. The portions of the unsectioned brains (ipsilateral and contralateral sensory-motor cortex, 2–3 mm thick) were cleared using the advanced CUBIC protocol¹⁵ and imaged in 3D using a confocal microscope (C2 plus, Nikon).

Laser speckle imaging

To monitor changes in cerebral blood perfusion before and after stroke, mice were anaesthetized and transferred to a stereotaxic frame. After performing a midline incision to expose the skull, mice were positioned under the laser speckle contrast imaging system (RFLSI II, RWD), and the surface of the skull was illuminated with a 784-nm laser (60 mW). Blood flow was recorded for 10 s (2.8×, 2048 × 2048, 10 frames at 10 Hz with 2 ms exposure time). Mice were then subjected to stroke as described before, and blood perfusion was recorded again 10 min after stroke. In another cohort of mice (vehicle versus GBP comparison), laser speckle imaging was performed 24 h after stroke. A region of interest was centred at the injury site and data analysis was performed using dedicated software (RWD Life Science).

Immunohistochemistry

Mice were transcardially perfused with 4% PFA in PBS (pH 7.4). Neuronal tissues were post-fixed at 4°C in 4% PFA overnight, and subsequently dehydrated in 10, 20 and 30% sucrose. Tissues were embedded in optimum cutting temperature (OCT) compound (Tissue-Tek), frozen, sectioned (HM525 NX, Thermo Fisher Scientific) and mounted on slides. Slides were warmed at 37°C for at least 30 min and OCT was washed away with PBS. Sections were then blocked at room temperature with 2.5% bovine serum albumin (A3059, Sigma-Aldrich) in PBS with 0.1% triton X100 for 1 h and incubated overnight at 4°C with a primary antibody. After washing three times with PBS, cryosections were incubated with Alexa Fluor-conjugated secondary antibodies (1: 400, Life Technologies). When necessary, sections were counterstained with 4′,6-diamidino-2-phenylindole (DAPI) (1: 10 000, D9542, Sigma-Aldrich). Images were taken using a confocal (C2 plus, Nikon) or epifluorescence microscope (Axio Observer Z1, Zeiss). Linear fluorescence intensity was calculated using Fiji (version 2.3.0/1.53f). Colocalization was measured using Fiji software. At least three independent replicates for each condition were analysed.

Immunoblot analysis

Murine sensory-motor cortices were dissected at 7 days after the operation (sham and stroke) and lysed on ice in RIPA buffer (0.5 M Tris-HCl pH 7.4, 1.5 M NaCl, 2.5% deoxycholic acid, 10% NP-40, 10 mM EDTA) containing phosphatase and protease inhibitors (Sigma-Aldrich). The homogenates were then centrifuged, and the supernatant was collected. Using Bradford reagent (Bio-Rad), the protein concentration of the lysate was determined and a portion of the lysate (20 μg total protein) was then fractionated by sodium dodecyl-sulfate–polyacrylamide gel electrophoresis. The separated proteins were transferred to a 0.2-μm nitrocellulose membrane (Bio-Rad) that was stained to confirm equal loading and transfer of the samples with Ponceau S (P7170, Millipore Sigma). After blocking at room temperature with 5% non-fat milk (1706404, Bio-Rad) in Tris-buffered saline with 0.1% Tween 20 detergent for 1 h, the membrane was probed with rabbit polyclonal anti-α2δ2 (Alomone Labs), guinea pig polyclonal anti-GLT1 (Millipore) or rabbit polyclonal anti-GLAST (Abcam). Rabbit polyclonal anti-Tuj1 (Sigma-Aldrich) or rabbit monoclonal anti-GAPDH (Cell Signaling Technology) antibodies were used as the loading control. Densitometry analysis was performed using ImageJ (NIH). After background subtraction, the intensity of α2δ2, GLT1 and GLAST bands was measured and normalized to the loading control (e.g. Tuj1 or GAPDH). Three to five biological replicates for each experimental condition were analysed.

Retrograde labelling of supraspinal sites

Four days after stroke, a cervical (C)3–5 laminectomy was performed in GFP-M adult mice, and Fluoro-Gold tracer (1%, Fluorochrome) was injected (0.5 μl/spot at 100 nl/min, four spots) into the dorsal corticospinal tract to retrogradely label corticospinal neurons in the cortex. To label vestibulospinal, reticulospinal, rubrospinal, pontine reticular nuclei, 5-HT and reticular formation neurons, 1% Fluoro-Gold tracer was injected into the C3-5 spinal cord (0.5 mm from the midline at a depth of 0.8 mm from the surface, 0.5 μl/spot at 100 nl/min, one spot at midpoint of each segment, three spots in total) of a separate cohort of mice. Three days after injection, the mice were perfused, and the brains were dissected and sequentially dehydrated in 10, 20 and 30% sucrose in PBS. Tissues were then embedded in OCT compound, frozen, sectioned (20 to 30-μm thick) and mounted on slides. Slides were then processed following standard immunohistochemistry protocols (see the previous immunohistochemistry section). The expression of α2δ2 in Fluoro-Gold positive corticospinal neurons was calculated using Zen Blue software (Zeiss). A minimum of three to five independent biological replicates (two or more sections/mouse) were analysed per condition.

Assessment of GBP stability

To assess the stability of GBP solutions in 0.9% saline at room temperature, solutions consisting of 50 mg of GBP in 12.6 ml of solution were solubilized at staggered starting times to provide the appropriate time points for analysis. Each sample was then divided into three replicates prior to incubation and incubated for the appropriate time period at room temperature. Following incubation, samples were diluted 10 000-fold to fall within the established calibration range of 1 ng/ml to 1 µg/ml (1 µl into a final volume of 10 ml) and transferred to a deep well 96-well microplate with sealing mat (Axygen) for analysis. Samples were analysed using a Thermo Accucore Vanquish C18+ column on a Thermo Vanquish ultra-high performance liquid chromatography with Thermo TSQ Quantiva mass spectrometer. Separation of GBP used a gradient elution from 10% B to 98% B over 3.25 min followed by a 3.5-min re-equilibration into starting conditions with mobile phase A consisting of 0.1% formic acid in water and mobile phase B consisting of 0.1% formic acid in methanol. Samples were introduced into the mass spectrometer through a heated electrospray ionization source in positive ion mode with a capillary voltage of 3.5 kV and a vaporizer temperature of 350°C with the following gas flow rates in arbitrary units: sheath gas −50, auxiliary gas −10 and sweep gas −1. GBP was quantified in multiple reaction monitoring mode by following the ion transition from 172 to 154 Da with a collision energy of 10.3 V while also monitoring the ion transition of 172 to 137 Da with a collision energy of 16 V to ensure selectivity.

In vivo multichannel recording of spontaneous firing

Seven days after stroke and sham operation, the mice were anaesthetized with a mixture of ketamine (100 mg/kg body weight) and xylazine (10 mg/kg body weight). A craniotomy was performed to expose the sensory-motor cortex on the contralateral side of the brain. A 32-channel silicon electrode array (Buzsaki 32-A32, NeuroNexus Technologies) connected to a stereotaxic frame was inserted 500–600 μm deep into the forelimb sensory-motor cortex. The four shanks were inserted parallel to the midline (1.3 mm). Spontaneous neuronal firing was recorded at a 25-kHz sampling rate and low-pass filtered at 10 kHz using the SmartBox acquisition system (NeuroNexus Technologies). Recording data were analysed using Igor Pro (v.8, WaveMetrics) as described previously.¹² Briefly, the recordings were filtered at 300–3000 Hz offline and passed through a box smoothing procedure (n = 5 points), and the spike detection threshold was set as five times the standard deviation of the baseline signals. The spiking refractory period was set as 1.5 ms, and the spikes were sorted using k-means clustering methods following a principal component analysis.

BDA tracing

Two weeks after stroke and sham operation, the intact corticospinal tract was traced with 10% BDA (biotinylated dextran amine, 10 000 MW, D1956, Life Technologies) with AP coordinates from bregma in mm: AP 1.0/1.3, 0.5/1.3, 0/1.3, −0.5/1.3, all at a depth of 0.6 mm from the surface, 500 nl/injection site, 50 nl/min. The mice were perfused with 4% PFA in PBS 2 weeks after BDA injection. Spinal cords and medulla oblongata were post-fixed in 4% PFA in PBS overnight at 4°C. The following day, they were immersed in 10, 20 and 30% sucrose in PBS, and cryosections were prepared. After quenching endogenous peroxidase with 0.3% H₂O₂ in PBS for 30 min, 20-μm thick coronal sections were incubated for 2 h with streptavidin–horseradish peroxidase conjugate (1: 200 in 2% Triton PBS, NEL7500001EA, Perkin Elmer). The TSA Cyanine 3 system (SAT704A001EA, Perkin Elmer) was then used for immunofluorescence amplification of the BDA signal. The procedures to count corticospinal fibres at the medullary level and the quantification of sprouting axons were the same as described elsewhere.¹⁶ To quantify sprouting axons within the brainstem, regions of interest were applied based on the Allen mouse brain atlas¹⁷ (post-natal Day 56; www.brain-map.org). After background subtraction, the mean grey values were measured and data were normalized to the number of BDA-labelled axons at the medullary region. Immunostaining was performed following standard procedure. Putative synapses along BDA-labelled corticospinal axons were identified as BDA and VGLUT1 overlayed puncta. High power confocal images were deconvolved using Nikon NIS Elements software (Nikon). The average density of BDA/VGLUT1 puncta was shown as the contour plot of the 2D histogram.

Behavioural testing

All researchers who tested or analysed the behavioural data were blinded to treatment groups.

Activity Box-Mice were placed in activity boxes (Columbus Instruments), and spontaneous activity was recorded in vertical and horizontal planes for 10 min using Fusion software (v.6.4 r1194, Omnitech Electronics). Activity was recorded before the injury to establish baselines and after injury at regular intervals until the study’s end point.

Cylinder test: Mice were placed in a 500-ml clear glass beaker with a small amount of clean bedding on the bottom. Animals were allowed to explore, and the first 10 paw placements on the sides of the beaker were recorded as left, right or both paws. The percentage of each paw placed was averaged to calculate forelimb use asymmetry. Mice were tested before the injury to establish baselines and after injury at regular intervals until the study end point.

Horizontal ladder: Mice were trained to run one way across a ladder (regularly spaced rungs) to an enriched cage on the opposite end. Once trained, mice ran across the ladder to the enriched cage while being recorded with a video camera for analysis. The mice were then tested before the injury to establish baselines and at regular intervals after injury until the study end point. The percentage of correctly placed steps was determined through frame-by-frame analysis of the video recordings.

cFos activity mapping in the ventral spinal cord

At 28 days, stroke mice administered vehicle or GBP were anaesthetized with a ketamine/xylazine mixture. The contralateral (uninjured side) sensory-motor cortex was electrically stimulated (300 μA, 0.5-ms biphasic pulses at 5 Hz for 15 min) by inserting a tungsten concentric bipolar electrode (TM33CCINS, World Precision Instruments, depth 500 µm, AP coordinates from bregma in mm: AP 0.25/1.3) connected to an isolated pulse stimulator (A-M Systems Model 2100). The animals were perfused 1 h after the end of the stimulation, and frozen coronal sections (20 µm) of the spinal cord were prepared. Before the blocking step, antigen-retrieval consisting of 2–3 min incubation of the tissue sections in citric acid-based antigen-unmasking solution (H-3300, Vector Laboratories) was performed at 95–100°C. The sections were labelled for cFos and NeuN and the average density of cFos+ neurons was shown as the contour plot of the 2D histogram.

H-reflex electrophysiological recording

M and H wave electrophysiological recordings were performed using a clinical electrodiagnostic system (Cadwell) under ketamine (20 mg/kg body weight) and xylazine anaesthesia (2.5 mg/kg body weight). To prevent corneal dryness and irritation, a petroleum-based eye lubricant (Dechra) was applied. Body temperature was maintained using a thermostatically controlled far infrared heating pad (Kent Scientific). Two 28-gauge monopolar recording electrodes (Natus Neurology) were inserted subcutaneously at the dorsal forelimb paw: one (G1) at the mid-carpal region and the other (G2) at the medial metacarpal-phalangeal joint. A disposably surface adhesive disc electrode was placed on the surface of the skin of the tail (Natus Neurology) as the common reference (G0). For stimulation, another pair of 28-gauge monopolar electrodes were inserted subcutaneously at the axilla for stimulation of the brachial plexus. A constant current stimulator was used to deliver pulses (1.0 ms duration). The stimulation intensity was adjusted to record the maximum H response and then further increased to ensure supramaximal stimulation to record the maximum M response. High and low frequency filter settings were set at 10 kHz and 10 Hz, respectively. M and H wave amplitudes were measured peak-to-peak.

Chemogenetics

Four weeks after stroke, intact corticospinal axons were transduced by injecting AAV2-hSyn-hM4D(Gi)-mCherry (1-2e13 GC/ml, 50475, Addgene) into the contralateral forelimb sensory-motor cortex. Two weeks after adeno-associated virus (AAV) injection (AP 1.0/1.3, 0.5/1.3, 0/1.3, –0.5/1.3, all at a depth of 0.6 mm from the surface, 500 nl/injection site, 50 nl/min), the mice underwent horizontal ladder and cylinder behavioural tests starting 15 min after vehicle injection (0.9% saline). The following day, the mice were administered clozapine N-oxide (CNO) (1 mg/kg, i.p.) (4936, CAS: 34233-69-7, Tocris Bioscience) to transiently silence corticospinal projections, and the behavioural procedures were repeated. Vehicle and CNO were administered in equal amounts. The alternation of vehicle and CNO administration was repeated for two consecutive sessions for 1 week, and the results were averaged between sessions. In a separate cohort of naïve mice, we transduced corticospinal axons with non-specific viral particles expressing enhanced GFP (eGFP) (AAV1-eGFP, ≥1e13 GC/ml, 105 530, Addgene) to test the extent to which CNO, when administered intraperitoneally at 1 mg/kg, causes changes in mouse behaviour in the absence of the hM4D(Gi) receptor. Two weeks after AAV-eGFP injection into the right sensory-motor cortex (AP 1.0/1.3, 0.5/1.3, 0/1.3, –0.5/1.3, all at a depth of 0.6 mm from the surface, 500 nl/injection site, 50 nl/min), the same behavioural protocols described previously were repeated, and the results were averaged between sessions. Behavioural analysis was carried out by investigators blinded to the treatment and experimental condition.

Statistical analysis

Statistical analysis was performed using SAS (SAS 9.4; SAS Institute) as follows: signed rank test, paired (Fig. 1F), Kruskal–Wallis test (Supplementary Fig. 1B), Wilcoxon rank sum test (Figs 2B, 3E, G and 5F and I and Supplementary Figs 1E, G and 3B), mixed model (Fig. 2E and Supplementary Figs 1H, 2B and H), two-sample Kolmogorov–Smirnov test (Fig. 2F), linear regression model (Figs 3C and 5J and K and Supplementary Figs 5C, 6I and J), mixed model with repeated measures using compound symmetry covariance structure and controlled on baseline values (Figs 5A, B and 6B and C), mixed model with repeated measures using compound symmetry covariance structure (Fig. 5C and Supplementary Figs 3D, E and 6B–D). For all analyses performed, significance was defined as P < 0.05. The exact values of n and the definition of measures are shown in the corresponding figure legends.

Figure 1

Vascular and haemodynamic changes within the sensory-motor cortex in the acute stages of stroke. (A) Schematic of photothrombotic stroke in mice. (B) Representative photographs of the brain following TTC staining 1 day after operation. Scale bar = 2 mm. (C) Quantification of B. Violin plot with median (n = 5). (D) Automated tile scanning of the vasculature in the cleared sensory-motor cortex 1 day after operation (DPO) (D = dorsal, V = ventral). Scale bar = 250 μm. (E) Laser speckle imaging before stroke and 10 min after offset of light. The dashed box indicates the infarct area (R = rostral, C = caudal). Scale bar = 1 mm. (F) Quantification of E. Aligned dot plot (signed rank test, paired *P < 0.05; n = 7).

Figure 2

A reduction in α2δ2 expression coincides with changes in electrophysiological properties of corticospinal neurons after stroke. (A) Immunoblot shows α2δ2 expression in the contralateral sensory-motor cortex 7 days after operation. Under reducing conditions, the α2δ2 antibody recognizes two bands at 130 and 105 kDa. Tuj1 is used as the loading control. (B) Quantification of A. Mean and SEM. Biological replicates originated from two independent experiments with sham #1, 2, 3 and stroke #1, 2, 3 from experiment no. 1 and sham #4, 5 and stroke #4 from experiment no. 2. Immunoblots were processed in parallel. Data normalized using loading control (Wilcoxon rank sum test *P < 0.05, sham n = 5 and stroke n = 4). (C) Schematic of retrograde labelling of corticospinal neurons in the brain. (D) Representative fluorescence images of retrogradely labelled corticospinal neurons (arrows) on the contralateral side of the brain 7 days after operation. Scale bar = 50 μm. (E) Quantification of D. Mean and SEM (mixed model type III test of fixed effects *P < 0.05; sham n = 5 and stroke n = 5; 55–133 neurons/animal, 440–461 neurons/group in total). (F) Differential distribution of firing frequency for all recorded single units (two-sample Kolmogorov–Smirnov test *P < 0.05; sham n = 5 and stroke n = 5; 353–355 single units/experimental group). (G) Raster plots show spontaneous firing within layer V of the sensory-motor cortex 7 days after operation. Bottom: histograms of firing events. Inset: spiking waveform of the single unit; the coloured lines show the average waveform and grey lines show all recorded waveforms.

Data availability

The data that support the findings of this study are available from the corresponding authors upon reasonable request.

Results

GBP administration does not exert beneficial effects on acute stroke pathophysiology

We first determined vascular and haemodynamic changes within the sensory-motor cortex in the acute stages of stroke. To do this, we used a photothrombotic stroke model¹⁸ (Fig. 1A) as it allows us to target the murine sensory-motor cortex with anatomical precision, thereby causing behavioural impairments in a reliable fashion. Using TTC staining, we assessed the reproducibility of photochemical cortical lesions and quantified tissue infarct size 24 h after stroke (Fig. 1B and C). Of note, photothrombotic damage affected all layers of the cortex but not the striatum. We then verified tissue damage and vascular occlusion down to the capillary level using a modified version of the vessel lumen staining protocol developed elsewhere^13,14 in combination with brain clearing and 3D imaging methods^15,19 (Fig. 1D). Penetrating arterioles and venules, as well as deep microvessels were occluded, as shown by the lack of fluorescent gel perfusate at the lesion site. Next, we used laser speckle imaging to corroborate haemodynamic changes within the targeted region on the ipsilateral side of the brain before and 10 min after stroke (Fig. 1E and F). The disruption of blood supply to neurons causes rapid neuronal cell loss after stroke.²⁰ We took advantage of transgenic mice with sparse GFP expression in different classes of adult neurons, including corticospinal neurons,²¹ to assess corticospinal cell death 24 h after stroke. We found that most corticospinal neurons on the ipsilateral side of the brain died quickly (Supplementary Fig. 1A and B). Neuroprotective strategies are necessary to enable neuron survival and functional recovery after stroke. Gabapentinoids have been shown to exert neuroprotection in experimental models of spinal cord injury and spinal ischaemia reperfusion.^22–24 In turn, we assessed whether systemic administration of GBP (46 mg/kg, three times per day) starting 1 h after stroke exerts any beneficial effect on vasculature structure and haemodynamic changes as well as survival of corticospinal neurons. When dissolved in 0.9% saline, GBP is stable for several days at room temperature (Supplementary Fig. 1C). Overall, we found no acute changes in blood perfusion, vascular occlusion, infarct size and layer V neuron survival by comparing mice administered vehicle (0.9% saline) and GBP (Supplementary Fig. 1D–H). Together, these data support the notion that photothrombotic stroke causes detrimental changes in haemodynamics that lead to cell death and suggest that early GBP administration does not benefit acute stroke pathophysiology.

α2δ2 subunit undergoes injury-dependent downregulation after stroke

Neuronal networks on both sides of the brain must undergo structural and functional reorganization to drive spontaneous recovery after stroke.²⁵ Experimental evidence suggests that induction of axonal sprouting from the contralateral side enhances recovery of skilled functions after stroke.^6,26 Given that α2δ2 negatively regulates sprouting and regeneration of adult neurons,^10,12 we asked whether α2δ2 expression is altered in corticospinal neurons that undergo structural changes on the contralateral side of the brain. We subjected adult mice to stroke and, 7 days later, collected sensory-motor cortices. Our immunoblot analysis showed that α2δ2 expression on the contralateral side of the brain was downregulated (Fig. 2A and B). To validate this finding, we injected the retrograde tracer Fluoro-Gold into the cervical spinal cord of adult mice to target corticospinal projections 4 days after stroke (Fig. 2C) and, 3 days later, dissected the brains and stained sagittal sections of the contralateral hemisphere containing retrogradely labelled corticospinal neurons with a polyclonal antibody that recognizes murine α2δ2. Our analysis of the forelimb sensory-motor cortex confirmed that expression of α2δ2 decreased in corticospinal neurons 7 days after stroke (Fig. 2D and E). No changes in α2δ2 expression were found in other supraspinal sites (vestibulospinal, reticulospinal, rubrospinal, pontine reticular nuclei, reticular formation and 5-HT neurons) that project to the cervical spinal cord (Supplementary Fig. 2A and B). Given that α2δ2 positively regulates synaptic transmission^27,28 and therefore neuronal network excitability, we determined whether a reduction in α2δ2 expression coincides with changes in spontaneous firing of adult corticospinal neurons on the contralateral side of the brain. To do so, we recorded single units within layer V in vivo using multichannel electrode arrays at 7 days after stroke (Supplementary Fig. 2C–E). Interestingly, we found a decreased frequency of neuronal spiking activity on the contralateral sensory-motor cortex after stroke when compared to the control condition (Fig. 2F and G). Next, we tested whether cortical astrocytes on the contralateral side of the brain exhibit changes associated with reduced neuron excitability on the contralateral side of the brain. Our immunoblot analysis showed decreased expression of the glutamate transporters GLT1 and GLAST at 7 days after stroke (Supplementary Fig. 2F). Using immunohistochemistry, we confirmed reduced GLT1 expression in cortical astrocytes around the soma of retrogradely labelled corticospinal neurons on the contralateral side of the brain (Supplementary Fig. 2G and H). Altogether, these data indicate that neurophysiological dynamics may be associated with spontaneous reorganization of neuronal circuits and reductions in astrocyte glutamate transporters.

Pharmacological blockade of α2δ2 promotes corticospinal plasticity after stroke

Although sprouting of the contralateral corticospinal pathway spontaneously occurs after stroke (Supplementary Fig. 3A and B) and CNS injury in mice,^29–31 recovery of complex sensory and motor behaviours remains limited in adulthood (Supplementary Fig. 3C–E). Similar results have been obtained in non-human primates.³² Hence, we asked whether chronic α2δ2 pharmacological blockade in vivo through GBP administration is sufficient to boost corticospinal collateral sprouting and enhance functional connectivity with brainstem target regions as well as across the denervated side of the spinal cord. We subjected adult mice to stroke and administered vehicle (0.9% saline) or GBP (46 mg/kg) starting 1 h after injury until the end of the study (Fig. 3A). Two weeks after stroke, we injected the anterograde tracer BDA into the contralateral sensory-motor cortex. Whereas control mice had limited sprouting (Fig. 3B), we found enhanced collateral sprouting of adult corticospinal axons in mice administered GBP (Fig. 3B–E). Not only was corticospinal sprouting increased in the cervical spinal cord of these mice, but also in the brainstem region (gigantocellular reticular nucleus) (Fig. 3F and G). Trans-midline corticospinal axons correctly integrate into spinal networks, as shown by the formation of new excitatory synaptic contacts (Fig. 4A). Whereas the murine dorsal spinal cord is responsible for sensory processing,³³ the intermediate and ventral spinal cord are critical for sensory-motor integration to fine-tune motor output and motor execution, respectively.³⁴ In mice administered GBP, we found that corticospinal collaterals preferentially targeted intermediate and ventral laminae, more specifically laminae VII, VIII and X (Fig. 4B). It is important to note that others reported similar changes in corticospinal innervation patterns by using different therapeutic interventions after stroke.⁸ In mice administered GBP, the number of excitatory puncta along BDA positive collaterals outnumbered those in the control (Fig. 4B). To map functional connectivity in these mice, we electrically stimulated the contralesional sensory-motor cortex 28 days after stroke to trigger immediate-early gene expression in post-synaptic spinal neurons (Supplementary Fig. 4A and B). Mice administered GBP had increased cFos immunoreactivity in laminae VIII, IX and X (Fig. 4C).

Figure 3

GBP administration promotes corticospinal plasticity after stroke. (A) Experimental scheme of B. (B) Representative fluorescence images of C7 spinal cord sections from adult mice 4 weeks after stroke (D = dorsal, V = ventral). The yellow arrows indicate corticospinal collaterals (bottom). Scale bar = 250 μm. (C) Quantification of B. Mean and SEM (type III test from linear regression model ***P < 0.001; vehicle n = 8 and GBP n = 9). (D) BDA-labelled corticospinal axons in the medullary region. Scale bar = 100 μm. (E) Quantification of D. Mean and SEM (Wilcoxon rank sum test; ns = not significant; vehicle n = 8 and GBP n = 9). (F) Representative fluorescence images of brainstem. GiV = gigantocellular reticular nucleus; Sp5O = spinal trigeminal nucleus; Py = pyramidal tract. Scale bar = 500 μm. (G) Quantification of F. Mean and SEM (Wilcoxon rank sum test, *P < 0.05, ns = not significant; vehicle n = 8 and GBP n = 9).

Figure 4

Mice administered GBP display enhanced functional connectivity. (A) Representative fluorescence images of C7 spinal cord sections from stroke mice administered GBP. Bottom: Orthogonal projections of the region in the main panel indicated by the arrow. Scale bar = 20 μm. (B) Mapping of excitatory presynaptic terminals on the contralateral side of the spinal cord 4 weeks after stroke (vehicle n = 9 and GBP n = 9). Roman numerals indicate spinal laminae. Scale bar = 200 μm. (C) cFos activity mapping in the ventral horn on the contralateral side of the C7 spinal cord. 2D histogram represents average density and yellow asterisks indicate the location of cFos+ neurons (vehicle n = 4 and GBP n = 4). Dashed line and Roman numerals indicate the border of the grey matter and spinal laminae, respectively. Scale bar = 200 μm.

Given that astrocytes play a crucial role in neurotransmitter clearance³⁵ and that gabapentinoids dampen excitatory neurotransmission,^10,36 we asked whether cortical astrocytes on the contralateral side of the brain undergo changes that parallel enhanced corticospinal plasticity after stroke in mice administered GBP. We subjected adult transgenic mice expressing eGFP under the control of the Aldh1l1 promoter that is selectively activated in adult cortical and spinal cord astrocytes³⁷ (Supplementary Video 1 and Supplementary Fig. 5A) to stroke and, 7 days later, collected sensory-motor cortices and fixed brain samples. Our immunoblot analysis showed that GLT1 and GLAST expression on the contralateral side of the brain was downregulated in mice administered GBP (Supplementary Fig. 5B and C). Next, we cleared the brain of Aldh1l1-eGFP mice and collected 3D imaging of cortical astrocytes (layer V) on the contralateral side of the brain at 7 days after stroke. Overall, we did not identify gross changes in either cellular architecture or the astrocytic network when the two experimental conditions were compared (Supplementary Fig. 5D).

Together, these data indicate that α2δ2 pharmacological blockade through systemic administration of GBP promotes structural corticospinal plasticity and functional connectivity after stroke in adulthood.

GBP administration enhances forelimb functional recovery after stroke

Next, we evaluated the extent to which enhanced corticospinal sprouting and connectivity on chronic GBP administration promote recovery of voluntary forelimb function after stroke in adult mice. After a stroke, degeneration of corticospinal projections and limited plasticity of the intact corticospinal pathway originating from the contralateral side of the brain hinders sensory-motor integration and neurological function as shown by behavioural deficits in forelimb function (Supplementary Fig. 3D and E). After collecting baseline measures, adult mice were randomly assigned to experimental groups and subjected to stroke. Beginning 1 h after injury, mice started receiving vehicle (0.9% saline) or GBP (46 mg/kg) administration. Compared to the control condition, mice administered GBP substantially recovered forelimb function as demonstrated by the increase in the percentage of correct forelimb steps at 42 days after stroke (Fig. 5A). We replicated these findings in a separate cohort of mice (Fig. 5B). In addition, chronic GBP administration restored forelimb symmetry in rearing behaviour (Fig. 5C) and dampened excitability of spinal circuits associated with maladaptive plasticity after stroke (Fig. 5D–F). No changes in general locomotor activity were found between the two experimental conditions (Supplementary Fig. 6A–D). To determine whether a causal link exists between enhanced cortical plasticity and forelimb recovery in mice administered GBP, we used a chemogenetic strategy.³⁸ Specifically, we injected AAV particles expressing the designer receptors exclusively activated by designer drugs receptor hM4Di into the contralateral forelimb sensory-motor cortex 4 weeks after stroke to chemogenetically silence cortical projections (Fig. 5G–I). Activation of hM4D signalling has been shown to cause a reduction in synaptic release probability and synaptic current amplitude.³⁹ Indeed, chemogenetic silencing of the contralateral cortical pathways via CNO (i.e. a pharmacologically inert metabolite of the atypical antipsychotic drug clozapine) administration temporarily abolished forelimb recovery in GBP-treated mice 6 weeks after stroke (Fig. 5J and K). We verified that CNO, when administered intraperitoneally at 1 mg/kg, was behaviourally inert in naïve mice injected with non-specific AAV particles expressing eGFP (Supplementary Video 2 and Supplementary Fig. 6E–J). Given that GBP administration beginning 1 h after stroke has a limited therapeutic window, we tested whether administering GBP beginning 24 h after stroke may be as effective in promoting recovery of forelimb function (Fig. 6A). Mice administered GBP (46 mg/kg) beginning 24 h after injury recovered forelimb function over 6 weeks (Fig. 6B and C), with the beneficial action on structural plasticity persisting long after treatment has been discontinued (Fig. 6B and C).

Figure 5

Mice administered GBP beginning 1 h after stroke recover forelimb function. (A) Recovery of forelimb skilled function was assessed using the horizontal ladder rung walking test. Mean and SEM (mixed model with repeated measures using compound symmetry covariance structure and controlled on baseline values *P < 0.05; vehicle n = 11 and GBP n = 11). (B) Replication of the study shown in A. Mean and SEM (mixed model with repeated measures using compound symmetry covariance structure and controlled on baseline values *P < 0.05; vehicle n = 9 and GBP n = 9). (C) Recovery of forelimb symmetry was assessed using the cylinder test. Mean and SEM (mixed model with repeated measures using compound symmetry covariance structure *P < 0.05; vehicle n = 11 and GBP n = 11). (D) Schematic of H-reflex electrophysiological recording. (E) Representative M and H wave responses from forelimb muscles. (F) H_max/M_max ratio at 35 days after stroke. Mean and SEM (Wilcoxon rank sum test *P < 0.05; vehicle n = 7 and GBP n = 8). (G) Schematic of chemogenetic silencing. (H) hM4Di(Gi)-mCherry transduced corticospinal axons in the medullary region. Scale bar = 100 μm. (I) Quantification of H. Mean and SEM (Wilcoxon rank sum test; ns, not significant; vehicle n = 9 and GBP n = 9). Abrogation of recovery of (J) forelimb skilled walking and (K) forelimb symmetry in rearing after stroke on transient activation of hM4Di(Gi) in corticospinal neurons of mice administered GBP. Aligned dot plot (linear regression model *P < 0.05, ***P < 0.001; ns = not significant; vehicle n = 9 and GBP n = 9).

Figure 6

Mice administered GBP beginning 24 h after stroke recover forelimb function, even after treatment has been discontinued. (A) Experimental scheme of B and C. (B) Recovery of forelimb skilled function was assessed using the horizontal ladder rung walking test. Mean and SEM (mixed model with repeated measures using compound symmetry covariance structure and controlled on baseline values. For the 54 and 61 DPI comparison, time was used as categorical variable in the model *P < 0.05, **P < 0.01; vehicle n = 10 and GBP n = 10). (C) Recovery of forelimb symmetry was assessed using the cylinder test. Mean and SEM (mixed model with repeated measures using compound symmetry covariance structure. For the 54 and 61 DPI comparison, time was used as categorical variable in the model **P < 0.01, ***P < 0.001; vehicle n = 10 and GBP n = 10).

Altogether, these data provide direct evidence that GBP administration promotes recovery of forelimb function after stroke in adult mice and that recovery in mice administered GBP relies on the structural reorganization of the cortical pathways originating from the contralateral side of the brain.

Discussion

By targeting intrinsic neuronal and extrinsic non-neuronal mechanisms, previous studies have shown that promoting corticospinal neuroplasticity aids functional recovery after cortical stroke.^7,8 Our study indicates that α2δ2 pharmacological blockade through GBP administration, a readily translatable pharmacological strategy, promotes plasticity of the corticospinal pathway after photothrombotic stroke in adult mice. More importantly, structural rearrangement of the intact cortical pathways on the contralateral side of the brain drives functional recovery in mice administered GBP. The beneficial action of gabapentinoids in promoting neurological recovery after acute CNS trauma is gaining support.^12,40,41 As gabapentinoids are clinically approved drugs prescribed for a wide range of neurological disorders, our findings highlight the strong potential for repurposing GBP as a promising treatment strategy for stroke repair.

GBP administration does not affect acute stroke pathophysiology

Minimally invasive photothrombotic stroke models allow the study of focal ischaemic cortical infarction with high spatial precision. A clear advantage of this experimental model over other stroke protocols is that experimental parameters such as the intensity and exposure of a cold light source can easily be adjusted to consistently produce medium-to-large cortical lesions that cause sustained sensory-motor deficits. Ischaemic stroke leads to rapid neuronal cell death within the infarct core. Within minutes after interruption of blood flow and oxygen supply, spine loss and dendritic blebbing are observed in cortical neurons.⁴² Early detrimental changes in neuronal cytoarchitecture can be reversed by rapid restoration of blood flow.⁴² Recombinant tissue plasminogen activator, the only clinically approved thrombolytic agent to treat acute ischaemic stroke, has a narrow therapeutic window due to the risk of haemorrhagic conversion of an ischaemic stroke.⁴³ In addition, moderate recanalization rate and contraindications for thrombolysis further restrict the number of acute stroke individuals eligible to receive this treatment. Thrombectomy represents another treatment option for individuals with large vessel occlusion that present with a favourable penumbra region based on perfusion imaging methods.^44–46 Major pathways implicated in cell death after ischaemic stroke are excitotoxicity and ionic imbalance, oxidative and nitrosative stress, and apoptotic-like mechanisms.⁴⁷ Gabapentinoids are clinically approved drugs for the treatment of neurological disorders, including seizures and nerve pain.⁴⁸ They bind with high affinity and selectivity to the α2δ1/2 subunits of voltage-gated calcium channels.^11,49 Both GBP and pregabalin have been shown to exert neuroprotection in experimental models of spinal cord injury and spinal ischaemia reperfusion.^22–24 Our data indicate that when administered as early as 1 h after stroke, GBP failed to normalize cerebral haemodynamic changes or promote neuroprotection of corticospinal neurons 24 h after stroke in adult mice. Given that the peak plasma concentration for GBP occurs within 2–3 h after administration and that cell death takes place within minutes at the lesion core, it is unlikely GBP alone may be sufficient to counteract acute stroke pathophysiology and promote neuron survival. Whether chronic GBP administration alters neuron, glia and vascular cell crosstalk that actively participates in tissue repair mechanisms⁵⁰ remains to be tested.

A reduction in α2δ2 expression parallels with a decrease in spontaneous firing of intact corticospinal neurons after stroke

Among other promising approaches, recovery of forelimb function after stroke may be attained by promoting appropriate neuroplasticity of the corticospinal pathway originating from the contralateral side of the brain.^6,26 The corticospinal pathway is the most prominent descending fibre tract that controls voluntary movements in humans.⁵¹ As corticospinal neurons mature, however, they lose the ability to grow and regenerate axons.^12,16,52 In exploring the mechanisms underpinning axon growth and regeneration failure in adulthood, we recently discovered that Cacna2d2, the gene encoding the α2δ2 subunit of voltage-gated calcium channels, acts as an intrinsic neuronal switch inhibiting axon growth in the adult CNS. Our previous findings suggest that a reduction in α2δ2 is sufficient to promote axon outgrowth in both mouse sensory and cortical neurons.^10,12 Cortical remapping of sensory and motor representation contributes to the reorganization of cortical circuits after stroke.^53,54 Interestingly, we have now found that α2δ2 expression is downregulated in corticospinal neurons on the contralateral side of the brain but not in other supraspinal sites that project to the spinal cord 7 days after stroke. The transcriptional or post-transcriptional-dependent mechanisms underlying changes in α2δ2 expression are unknown and will be investigated in future studies. Our new data also indicate that a reduction in α2δ2 expression correlates with some degree of spontaneous corticospinal plasticity after stroke. Given that α2δ2 subunits positively regulate synaptic transmission and therefore network excitability,²⁸ it may not be surprising that a reduction in α2δ2 expression parallels a decrease in the frequency of neuronal spiking activity in layer V neurons on the contralateral side of the brain. In both experimental animal models and humans, brain areas surrounding cortical ischaemic lesions undergo changes in neuron excitability.⁵⁵ Others have reported increased excitability of rat layer II/III cortical neurons using acute brain slice preparations 7–10 days after photothrombotic lesion.⁵⁶ Such a discrepancy with our findings may be explained by the fact that different classes of cortical excitatory or inhibitory neurons are highly heterogeneous and may respond differently to the same type of injury. Among glial cells, astrocytes modulate neuronal network activity by controlling extracellular glutamate homeostasis.^57–59 Given that a single cortical astrocyte can enwrap four to eight neuronal soma and contact 300–600 neuronal dendrites in the mouse brain,⁶⁰ changes in astrocyte behaviour can affect neuronal circuit function under normal and pathophysiological conditions. Accordingly, we asked whether cortical astrocytes on the contralateral side of the brain undergo changes associated with reduced neuron excitability. Interestingly, we found reduced expression of the glutamate transporter GLT1 mainly expressed by cortical astrocytes at 7 days after stroke, suggesting that spontaneous reorganization of cortical circuits may be associated with neurophysiological dynamics and changes in astrocyte behaviour. A comprehensive and unbiased in vivo characterization of the changes in cortical neuron excitability and glial cell function after stroke will be an important direction for future investigations.

α2δ2 pharmacological blockade through GBP administration promotes recovery of forelimb function after stroke

After a stroke, degeneration of corticospinal projections and limited plasticity of the intact corticospinal pathway originating from the contralateral side of the brain hinder sensory-motor integration and neurological function as shown by behavioural deficits in forelimb function. Although corticospinal axons originating from the contralateral side of the brain spontaneously sprout after stroke, our data indicate that the number of collaterals is rather limited and may not be sufficient to drive the changes in network activity needed to control limb muscle after stroke. General ground locomotion spontaneously recovers, but the behavioural deficit in skilled forelimb function persists. This prompted us to test whether pharmacological blockade of α2δ2 through GBP administration is sufficient to boost corticospinal plasticity and promote functional recovery after ischaemic cortical infarction. Neuronal circuits display modular organization with short- and long-range connections within modules and nodes, respectively. Not only was neuroplasticity of the contralateral corticospinal pathway enhanced across the denervated side of the spinal cord, but also within brainstem regions in mice administered GBP. Our functional connectivity mapping suggests that corticospinal collaterals preferentially target the intermediate and ventral spinal cord hub regions. These regions play a key role in motor execution.³⁴ GBP daily administration also dampens maladaptive plasticity that often develops following stroke.⁶¹ As a result of increased excitability of spinal motor circuitry, H-reflex alterations have been reported following cortical infarct.^62,63 Accordingly, we recorded the H-reflex from forelimb muscles 35 days after stroke and discovered that mice administered GBP displayed a reduction in H_max/M_max ratio. We also found that mice administered GBP had decreased GLT1 and GLAST expression on the contralateral side of the brain at 7 days after stroke. Given that the genes encoding α2δ1/2 subunits are primarily expressed in neurons and not astrocytes in the mouse cortex,⁶⁴ it is unlikely that cortical astrocytes directly respond to the treatment. Rather, the expression of glutamate transporters in astrocytes may be regulated by changes in excitatory transmission and homeostatic needs after stroke.

Mice administered GBP beginning 1 h or 1 day after injury recovered skilled forelimb function over 6 weeks after photothrombotic stroke. It is also important to note that the beneficial action of GBP on cortical plasticity persisted long after treatment was discontinued. Interestingly, a retrospective study has found that stroke patients prescribed antiepileptic drugs including gabapentinoids exhibited no change in their outcomes during an intensive rehabilitation program.⁶⁵ When assessing functional outcomes in these patients, however, key parameters such as the specific contribution of individual drugs, gabapentinoids dosage and initiation of gabapentinoids administration were not considered. We and others have shown that the initiation of gabapentinoids administration is the key to maximizing any chance of recovery in the injured CNS.^10,12,40,41 We previously demonstrated that the maximal outcome is achieved when gabapentinoid treatment is delivered soon, rather than weeks, after injury.¹⁰ Along this line, a multicentre cohort study has found that early (versus late) administration of gabapentinoids improves functional recovery in spinal cord injury individuals.⁴¹

Of the different behavioural tests used to challenge forelimb function, we used the skilled rung walking test,⁶⁶ as this generated the most consistent and reliable behavioural measures. Forelimb recovery in mice administered GBP relies on the structural reorganization of the contralesional cortical pathways as chemogenetic silencing transiently abrogated recovery. One caveat of our chemogenetic strategy is that we cannot rule out the net contribution of enhanced brainstem or spinal cord connectivity in mice administered GBP. It is currently unknown whether GBP’s beneficial effects may be extended to other experimental models of stroke such as haemorrhagic and large infarct stroke. Whether cardiovascular rehabilitation and activity-based training may facilitate refinement and consolidation of functional connectivity in the injured brain and spinal cord of mice administered GBP remains to be tested. In conclusion, our study highlights the strong potential for repurposing GBP as a promising treatment strategy for ischaemic stroke repair.

Acknowledgements

We would like to thank Drs Michele Curcio and Raman Saggu for critical reading of the manuscript, Rochelle Rodrigo for technical assistance and all members of the laboratory for discussion.

Funding

This work was supported by the National Institute of Neurological Disorders and Stroke (grants R01NS110681 and R21NS109787) and Chronic Brain Injury Discovery Theme (grant 202001) at The Ohio State University, with additional support provided by NIH grant P30NS104177.

Competing interests

The authors report no competing interests. The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Supplementary material

Supplementary material is available at Brain online.

Abbreviations

 
• AP

  anterior–posterior

 
• BDA

  biotinylated dextran amine

 
• CNO

  clozapine N-oxide

 
• GBP

  gabapentin

 
• GFP

  green fluorescent protein

 
• GLAST

  glutamate/aspartate transporter

 
• GLT1

  glutamate transporter subtype 1