Abstract
Excitotoxicity and the concurrent loss of inhibition are well-defined mechanisms driving acute elevation in excitatory/inhibitory (E/I) balance and neuronal cell death following an ischemic insult to the brain. Despite the high prevalence of long-term disability in survivors of global cerebral ischemia (GCI) as a consequence of cardiac arrest, it remains unclear whether E/I imbalance persists beyond the acute phase and negatively affects functional recovery. We previously demonstrated sustained impairment of long-term potentiation (LTP) in hippocampal CA1 neurons correlating with deficits in learning and memory tasks in a murine model of cardiac arrest/cardiopulmonary resuscitation (CA/CPR). Here, we use CA/CPR and an in vitro ischemia model to elucidate mechanisms by which E/I imbalance contributes to ongoing hippocampal dysfunction in male mice. We reveal increased postsynaptic GABAA receptor (GABAAR) clustering and function in the CA1 region of the hippocampus that reduces the E/I ratio. Importantly, reduced GABAAR clustering observed in the first 24 h rebounds to an elevation of GABAergic clustering by 3 d postischemia. This increase in GABAergic inhibition required activation of the Ca2+-permeable ion channel transient receptor potential melastatin-2 (TRPM2), previously implicated in persistent LTP and memory deficits following CA/CPR. Furthermore, we find Ca2+-signaling, likely downstream of TRPM2 activation, upregulates Ca2+/calmodulin-dependent protein kinase II (CaMKII) activity, thereby driving the elevation of postsynaptic inhibitory function. Thus, we propose a novel mechanism by which inhibitory synaptic strength is upregulated in the context of ischemia and identify TRPM2 and CaMKII as potential pharmacological targets to restore perturbed synaptic plasticity and ameliorate cognitive function.
Significance Statement
Excitatory/inhibitory (E/I) imbalance drives long-term disability in numerous central nervous system disorders, including cerebral ischemia. Previous studies indicated ischemia-induced hippocampal synaptic plasticity deficits contribute to long-term cognitive impairment, yet the mechanisms underlying hippocampal dysfunction are poorly defined. Here, we combine in vivo and in vitro approaches to demonstrate elevated GABAA receptor clustering and function contribute to a reduction in hippocampal E/I balance and deficits in long-term potentiation at delayed timepoints following ischemia. We further identify ongoing activation of the TRPM2 ion channel and Ca2+-dependent kinase, CaMKII, are required for the ischemia-induced enhancement of GABAergic synaptic inhibition, highlighting promising new targets to improve postischemic long-term functional recovery.
Introduction
Hippocampal dysfunction occurs as a consequence of brain injury, leading to impairments in learning and memory. Notably, cerebral ischemia has profound effects on hippocampal function due to the high metabolic requirements of the hippocampus (Petito et al., 1987; Schmidt-Kastner and Freund, 1991; Neumann et al., 2013). Ischemic insults trigger numerous pathways including excitotoxicity, which drives neuronal demise via excessive glutamate release and resultant activation of AMPA- and NMDA-type glutamate receptors. Thus, considerable research efforts have concentrated on mitigating excitotoxic effects through NMDA and/or AMPA receptor blockade (Turski et al., 1998; Walters et al., 2005; Kostandy, 2012; Wu and Tymianski, 2018). While this neuroprotective approach effectively reduces neuronal cell death in animal models, clinical trials have shown limited improvement in long-term functional recovery in patients likely due to the narrow therapeutic window required for these therapies (Cheng et al., 2004; Wahlgren and Ahmed, 2004; Katz et al., 2022).
Neurorestoration is an alternative strategy that aims to restore neural circuits perturbed by neuronal injury, by enhancing synaptic function and plasticity in surviving neurons, allowing for treatment within a broader therapeutic timeframe (Azad et al., 2016; Escobar et al., 2019). Indeed, this strategy can be reliably tested in a murine model of global cerebral ischemia (GCI) and cardiac arrest/cardiopulmonary resuscitation (CA/CPR). Despite the completion of cell death processes within days of the initial insult, we observed sustained impairments of long-term potentiation (LTP) in surviving hippocampal neurons, correlating with learning and memory deficits post-CA/CPR (Orfila et al., 2014; Dietz et al., 2020). Further, we demonstrated that delayed inhibition of the Ca2+-permeable, transient receptor potential melastatin-2 (TRPM2) ion channel reverses impairments in hippocampal LTP and hippocampal-dependent behavioral tasks (Dietz et al., 2020, 2021), implicating TRPM2 as a potential target for neurorestorative therapy. However, the precise molecular mechanisms underlying hippocampal dysfunction and the contribution of ongoing TRPM2 activity to LTP impairment remain elusive.
Extensive studies have emphasized the importance of maintaining a balance between excitatory and inhibitory (E/I) signaling for optimal neuronal function and induction of LTP mechanisms (Smith and Kittler, 2010; Vogels et al., 2011). Specifically, GABAergic synaptic inhibition plays a critical role in E/I balance by regulating circuit and neuronal excitability (Chiu et al., 2019). Further, prior work indicates that GABAergic inhibitory signaling directly impacts excitatory LTP (Steele and Mauk, 1999; Leao et al., 2012; Williams and Holtmaat, 2019; Udakis et al., 2020). In the context of cerebral ischemia, excitotoxic insult promotes rapid declustering of synaptic GABAARs, subsequent elimination of inhibitory synapses, and neuronal cell death (Smith et al., 2012; Mele et al., 2014; Costa et al., 2016). Conversely, at delayed timepoints following the completion of cell death processes, there is accumulating evidence of excessive GABAergic function through overactivation of synaptic (phasic; Hiu et al., 2016) and extrasynaptic (tonic) GABAARs (Clarkson et al., 2010; Carmichael, 2012; Orfila et al., 2019). However, the precise upstream mechanisms driving this enhancement and whether these pathways contribute to sustained synaptic plasticity and cognitive deficits remain poorly defined. In light of the established role of GABAergic inhibition in modulating excitatory synaptic plasticity, we tested the effect of ischemia on synaptic GABAergic inhibition and whether TRPM2 reverses LTP impairments via modulation of GABAA receptor (GABAAR) function.
Here, we present data demonstrating a shift in the balance of E/I signaling in the hippocampus after CA/CPR. At acute timepoints, we found a decrease in GABAAR clustering. However, in the postacute phase following the completion of cell death processes, we observed a sustained increase in postsynaptic inhibition, leading to a reduction in E/I balance. Using a combination of in vitro and in vivo approaches, we rigorously identified the TRPM2 ion channel as a mediator of augmented inhibitory function. We also provide compelling evidence that TRPM2 and Ca2+/calmodulin-dependent protein kinase II (CaMKII) activations are necessary for the sustained enhancement of postsynaptic GABAergic inhibition, highlighting potential therapeutic targets to improve functional recovery following brain ischemia.
Materials and Methods
Key resources
For details of key resources/reagents used in the study, see Table 1.
Key resources table
Experimental model and subject details
Animals
All studies conformed to the requirements of the National Institutes of Health Guide for the Care and Use of Laboratory Animals and were approved by the Institutional Animal Care and Use subcommittee of the University of Colorado, Denver AMC. C57BL/6 mice were bred in house in the Animal Resource Center at the University of Colorado Anschutz Medical Campus and monitored regularly for health. Mice were weaned between postnatal days 21 and 28 (P21 and P28) and housed in microisolator cages on a 14:10 light/dark cycle with water and chow available ad libitum.
Cardiac arrest/cardiopulmonary resuscitation
Male mice that were approximately 8–12 weeks old were subjected to either cardiac arrest and cardiopulmonary resuscitation (CA/CPR) or sham procedures as described previously (Deng et al., 2017; Dietz et al., 2020). Briefly, mice were anesthetized with 3% isoflurane. Mice were intubated and connected to a mouse ventilator set to 160 breaths per minute. Cardiac function was monitored via electrocardiography, and pericranial temperature was maintained at 37.5°C ± 0.2°C using a water-filled coil. Asystolic cardiac arrest was induced by KCl injection via a jugular catheter. CPR began 6 min after induction of cardiac arrest, by slow injection of 0.5–1.0 ml of epinephrine (16 μg epinephrine/ml, 0.9% saline), chest compressions at a rate of ∼300 min−1, and ventilation with 100% oxygen. If the return of spontaneous circulation could not be achieved within 3 min of CPR, resuscitation was terminated and the mouse was excluded from the study. After surgical procedures, mice were housed individually in microisolator cages on heating pads. Postsurgical care included daily saline injections (1 ml) and moist chow for 72 h. Investigators performed all experiments blind to the surgical procedure of the animal, with a separate investigator generating the code.
Dissociated hippocampal cultures
Primary hippocampal cultures were prepared as described previously (Crosby et al., 2019; Rajgor et al., 2020; Garcia et al., 2021). Briefly, the hippocampi from neonatal rat pups (P0–P1) were dissected and dissociated in papain. The isolated neurons were then seeded in MEM supplemented with 10% FBS and penicillin/streptomycin. Cells were plated at a density of 150,000–200,000 cells per 18 mm, #1.5 glass coverslip coated with poly-d-lysine. The MEM was replaced with Neurobasal (NB) media (GIBCO) supplemented with B27 (GIBCO) and 2 mM GlutaMAX 24 h after plating. The media was refreshed every 5 d before removing half of the existing media and replacing it with fresh NB media. To restrict the growth of actively dividing cells, mitotic inhibitors (uridine fluoro deoxyuridine) were introduced on Day 5. The cultures were maintained at 37°C with 5% CO2 for a period of 13–14 d before conducting OGD experiments.
Method details
Hippocampal slice preparation
Hippocampal slices were prepared during 7 d postsurgical procedures. Mice were anesthetized with 3% isoflurane in an oxygen-enriched chamber and then transcardially perfused with oxygenated ice-cold artificial cerebral spinal fluid (aCSF) containing the following (in mM): 126 NaCl, 25 NaHCO3, 12 glucose, 2.5 KCl, 2.4 CaCl2, 1.3 NaHPO4, and 1.2 MgCl2. Horizontal slices (300µM) were cut in aCSF supplemented with 9 mM MgSO4 and continuous oxygenation using a Vibratome 1200 (Leica) and transferred to a holding chamber containing aCSF warmed to 33°C. After 30 min, slices recovered for an additional 30 min at room temperature (RT), prior to electrophysiology recordings.
Whole-cell patch-clamp electrophysiology
CA1 pyramidal neurons were visualized using infrared digital interference contrast optics under 63× magnification and patch-clamped in whole-cell configuration. Borosilicate glass recording electrodes were pulled with either a Narishige or Sutter P-97 Flaming/Brown electrode puller (Sutter Instrument) with a resistance of 2.5–4.0 MΩ. The recording solution was exchanged at a flow rate of approximately 2 ml per minute at RT. Responses were amplified and filtered at 5 kHz (MultiClamp 700B) and digitized at 20 kHz (Digidata 1440A and Clampex 10.7). No series resistance compensation nor junction potential corrections were performed. Access resistance was monitored by delivering a −5 mV voltage step, and any experiments in which access resistance was above 30 MΩ or changed >20% between the onset and completion of the experiment were not used for experimental analyses.
Excitatory/inhibitory balance experiments were conducted by recording excitatory postsynaptic currents (EPSCs) at −70 mV and inhibitory postsynaptic currents (IPSCs) at 0 mV from the same cell with an internal solution containing the following (in mM): 135 CsMeSO4, 10 HEPES, 10 BAPTA, 5 Qx314, 4 Na2ATP, 4 MgCl2, and 0.3 NaGTP, at pH 7.25 with 1 M CsOH. Responses were evoked using a monopolar electrode in the stratum radiatum to stimulate Schaffer collateral-commissural (SCC) apical dendrites using a constant current source (Digitimer). The episodic mode was used to record EPSCs and IPSCs responses evoked every 20 s for a minimum of 2 min each. Spontaneous inhibitory postsynaptic currents (sIPSCs) were recorded at −60 mV in gap-free mode, with an internal solution containing the following (in mM): 140 CsCl, 10 NaCl, 10 HEPES, 5 EGTA, 0.5 CaCl2, 2 MgATP, and 5 Qx314 in aSCF containing 10 µM DNQX to block AMPA currents (Banks et al., 1998). The EGTA and CaCl2 in the internal solution were replaced with 10 mM BAPTA to examine the role of Ca2+ signaling in postsynaptic neuron sIPSCs following CA/CPR. To determine the role of CaMKII activity in postsynaptic neurons sIPSCs following CA/CPR, the internal solution was supplemented with 5 µM tatCN19o (Barcomb et al., 2015; a gift from KU Bayer). To assess the role of TRPM2 activation in inhibitory function, 2 µM tatM2NX (Cruz-Torres et al., 2020; Dietz et al., 2020) was bath applied for at least 20 min prior to and during recording. For clotrimazole (CTZ) experiments, the baseline was recorded for 3 min, and clotrimazole (20 µM; Verma et al., 2012) was perfused onto the slice for 12 min. Access resistance was monitored every 3 min. Recordings that exhibited >20% drift in access resistance were discarded. The final 3 min of clotrimazole recordings were used for statistical comparison to baseline.
Field electrophysiology
Hippocampal slices were placed in a heat-controlled interface chamber perfused with aCSF at a rate of 1.5 ml/min at 32°C. Responses were evoked using an insulated tungsten bipolar stimulating electrode placed in stratum radiatum to stimulate Schaffer collateral-commissural (SCC) and recorded with a glass electrode containing 150 mM NaCl placed in the distal dendrites of CA1 pyramidal cell layer. Analog field excitatory postsynaptic potentials (fEPSPs) were amplified (1,000×) and filtered through a preamplifier (Grass Model P511) 0.03 Hz to 1.0 kHz, digitized at 10 kHz and stored on a computer for later offline analysis (Datawave Technologies). The derivative (dV/dT) of the initial fEPSP slope was measured. The fEPSPs were adjusted to 50% of the maximum slope and test pulses were evoked every 20 s. Paired-pulse responses were recorded using a 50 ms interpulse interval (20 Hz) and expressed as a ratio of the slopes of the second pulse over the first pulse. Picrotoxin (5 mM; Costa and Grybko, 2005) was applied for at least 20 min during the acquisition of a stable fEPSP baseline. Following the baseline recording, picrotoxin was washed out with normal aCSF, and theta burst stimulation (TBS) was delivered, which included a train of four pulses delivered at 100 Hz in 30 ms bursts repeated 10 times with 200 ms interburst intervals. Following TBS, the fEPSP was recorded for 60 min. The averaged 10 min slope from 50 to 60 min after TBS was divided by the average of the 10 min baseline (set to 100%) prior to TBS to determine the amount of potentiation. For time course graphs, normalized fEPSP slope values were averaged and plotted as the percent change from baseline.
Immunohistochemistry
Mice were anesthetized and transcardially perfused with ice-cold PBS followed by 4% paraformaldehyde (PFA). Whole brains were removed and postfixed in 4% PFA at 4°C overnight. After 24 h, brains were transferred to a glycerol and Sorenson's buffer cryoprotection solution for long-term storage. Frozen coronal sections were made using a sliding microtome, and slices were placed in a cryostorage solution containing phosphate buffer, ethylene glycol, polyvinylpyrrolidone, and sucrose and stored at 4°C until staining was performed. Free-floating sections were washed for 15 min (three times) in PBS at RT then blocked and permeabilized (5% BSA, 5% NGS, 0.5% Triton X-100, and 1× PBS) at RT for 5–6 h on a rocker. Slices were incubated with gephyrin (1:500 Synaptic Systems, mouse, 147011) and VGAT (1:1,000 Synaptic Systems, rabbit, 131003) antibodies in permeabilization solution overnight at 4°C on a rocker. Slices are washed for 20 min in PBS (four times) and incubated with appropriate secondary antibodies (1:1,000 Thermo Fisher Scientific, Alexa Fluor 488 and 568) for 2 h in a blocking solution. Prior to mounting with ProLong Gold, slices were washed for 20 min in PBS (four times).
Immunocytochemistry
Coverslips containing neuronal cultures were fixed in a 4% PFA solution consisting of 4% sucrose, 1× PBS, and 50 mM HEPES (pH 7.4) for 5 min at RT. After fixation, the cells were blocked in a solution containing 5% BSA, 2% normal goat serum (NGS), and 1× PBS at RT for 30 min. Staining for surface GABAAR-γ2 subunit (1:500, Synaptic Systems, guinea pig, 224004) was performed under nonpermeabilized conditions in the blocking solution for 1 h at RT. Following the primary antibody incubation, coverslips were washed three times for 5 min each with 1× PBS. Subsequently, permeabilization was carried out using 0.5% NP-40 for 2 min, followed by blocking at RT for 30 min. Staining for gephyrin (1:600, Synaptic Systems, mouse, 3B11 clone, 147111) and VGAT (1:1,000, Synaptic Systems, rabbit, 131003) or MAP2 (1:1,000, Synaptic Systems, mouse, 188011) was performed in the blocking solution for 1 h, following by three 5 min washes with PBS. The coverslips were then incubated with appropriate secondary antibodies (1:1,000, Thermo Fisher Scientific, Alexa Fluor 488, 568, and 547). Coverslips were washed three times for 5 min and were then mounted on microscope slides using ProLong Gold mounting media (Thermo Fisher Scientific). For experiments requiring nonpermeabilizing conditions, the 0.5% NP-40 step was omitted.
Oxygen glucose deprivation (OGD) in neuronal culture
OGD was induced in DIV13–15 hippocampal neuronal cultures using a HEPES-buffered solution. The OGD-HEPES solution contained the following (in mM): 25 HEPES (pH 7.4), 140 NaCl, 5 KCl, 2 CaCl2, 1 MgCl2, and 10 sucrose (or supplemented with 10 mM glucose for control conditions). Prior to OGD treatment, the OGD-HEPES solution was placed in an anaerobic workstation at 37°C with a controlled atmosphere of 95% N2 and 5% CO2 (Bugbox Plus, Baker Co) for 24 h to allow for deoxygenation. Neuronal cultures were washed twice and incubated with OGD-HEPES solution in the anoxic chamber for 20 min. Reoxygenation was then initiated by replacing the OGD-HEPES solution with glucose-containing conditioned media and returning coverslips to an aerobic incubator. After 96 h in aerobic conditions, coverslips were fixed for immunocytochemistry. For the treatment conditions, coverslips were treated with various inhibitors for 1 h prior to fixation. The following inhibitors and concentrations were used (in µM): 20 CTZ, 2 tatM2NX, 2 tatScr, 5 tatCN19o, and 5 KN93. Control neurons were incubated at 37°C, 5% CO2 with control-HEPES solution for 20 min and returned to conditioned media before fixation at the 96 h timepoint.
Protein fractionation and Western immunoblotting
Mice were anesthetized, followed by rapid decapitation and brain removal. The whole hippocampus was dissected and flash frozen and stored at −80°C until membrane fractionation was performed. Membrane fraction preparation was performed as described previously (Deng et al., 2017). The whole hippocampus was homogenized in ice-cold homogenization buffer containing the following (in mM) 10 Tris (pH7.4), 320 sucrose, 1 µM EDTA, and 1 µM EGTA, with phosphatase and protease inhibitors (Thermo Fisher Scientific) using glass homogenizers and a drill fitted with a pestle. Homogenates were then transferred to 1.5 ml Eppendorf tubes and centrifuged for 10 min at 1,000 × g at 4°C. The supernatant was collected and placed in a clean Eppendorf tube and centrifuged at 10,000 × g at 4°C. The supernatant was then removed, and the resulting pellet (P2) containing the membrane fraction was resuspended in 60 µl Neuronal Protein Extraction Reagent (Thermo Fisher Scientific). Protein concentrations were quantified using Pierce BCA Protein Assay Kit (Thermo Fisher Scientific) before diluting each sample in 5× SDS loading buffer to equal concentrations and heating to 95°C for 5 min. Protein separation was achieved by SDS-PAGE. Western immunoblotting was performed as described previously (Cruz-Torres et al., 2020). Proteins were transferred to PVDF membranes and using the following primary antibodies: CaMKII (1:1,000; CaMKII antibody, BD Transduction Laboratories, catalog #611293), phosphoCaMKII T286 (1:1,000; PhosphoSolutions, catalog #p1005-286), vinculin (1:1,000; Cell Signaling Technology, catalog #13901), and species appropriate secondaries (anti-mouse 1:5,000, Jackson ImmunoResearch; anti-rabbit, 1:5,000; Thermo Fisher Scientific).
Quantitative real-time PCR
For measurement of GABAAR subunit transcripts, CA1 isolates were harvested 7 d following sham and CA/CPR surgeries. RNA isolations and PCR were performed as previously described (Dietz et al., 2020). Briefly, per the manufacturer's instructions, RNA was isolated using the RNAqueous-4 PCR kit (Ambion). Approximately 0.5 mg of tissue was lysed in lysis buffer and total RNA was isolated and eluted from a column with 50 µl RNase-free elution buffer and further treated with Turbo DNase (Ambion). RNA (500 ng) was reverse transcribed to single-stranded cDNA using the iScript cDNA Synthesis Kit (Bio-Rad). Real-time PCR reactions using ssoFast PCR mastermix (Bio-Rad) were performed on the Bio-Rad CFX connect detection system and performed in triplicate using 50 ng of cDNA. Taqman (Thermo Fisher Scientific) primers were used to detect GABRG2, GABRB3, GABRA1, and 18 s transcripts. Cycle parameters used were 95°C for 10 min followed by 40 cycles of 95°C for 15 s and 60°C for 30 s. Relative expression levels were calculated using ΔΔCT as the ratio of the target gene to the housekeeping gene 18 s.
Image acquisition and data analysis
Confocal microscopy
Confocal images were acquired on a Zeiss Axio Observer Z1 upright microscope equipped with a Yokogawa CSU-X1 spinning disk unit; a 63× oil immersion objective using 2× digital zoom (Plan Apo/1.4 NA); an Evolve 512 EM-CCD camera (PhotoMetrics) with 16-bit range; and SlideBook 6.0. Alternatively, neurons were imaged using an Olympus FV1000 laser scanning confocal microscope, 60× oil immersion objective with 2× digital zoom, and Fluoview software (Olympus FluoView, FV10-ASW). Images on both microscopes were attained at 0.3 µm intervals (4 µm Z-stack projection). Cluster analysis was performed using ImageJ (NIH) by selecting regions of interest (ROIs) to differentiate between dendritic and somatic compartments. A user-based threshold was determined by sampling several images per condition across all conditions, and clusters were defined with a minimum size of 0.05 µm2. For IHC experiments, the CA1 hippocampus was identified by VGAT staining of the pyramidal cell layer (PCL). ROIs captured both the PCL and stratum radiatum in the same frame, and different user-based thresholds were used for the cell bodies and dendrites. Density was calculated by dividing the number of clusters by the ROI area (per µm2). A minimum of five animals per condition were utilized, with two slices per animal analyzed. Analysis was performed blind to surgical condition. For the in vitro experiments, ROIs were delineated by tracing along dendrites. The density of clusters was calculated by measuring the number of clusters divided by the length of the delineated dendrites (per 10 µm). A total of 30–36 neurons were analyzed per condition from three independent hippocampal preparations.
Experimental design and statistical analysis
All analyses were conducted blind to condition. The experimenter was additionally blind to condition when performing in vivo studies. A number of animals and cells are indicated in the figure legend. All data in the figures are presented as mean ± SEM. Statistical significance was determined using appropriate tests indicated in the figure legends. Statistical details for every dataset are provided in tables throughout the Results section. A p-value ≤0.05 was used to declare significance. All statistical analyses were performed on GraphPad Prism v9.4.
Results
E/I balance is disrupted following CA/CPR
To assess the effect of GCI on E/I balance, we employed patch-clamp electrophysiology to record evoked GABA (0 mV) and AMPA (−70 mV) responses from CA1 pyramidal neurons 7 d following CA/CPR or sham surgery (Fig. 1A), a postacute timepoint at which cell death processes have subsided and perturbations in synaptic function are reflective of the surviving network. The ratio of IPSC (GABA) to EPSC (AMPA) amplitude was greater in CA/CPR compared with sham, indicating increased synaptic inhibitory function relative to excitatory function (Fig. 1B; Table 2). This increase in inhibitory function following CA/CPR was observed with no difference in release probability in the CA/CPR group compared with sham as measured by the paired-pulse ratio of IPSCs (Table 3). No differences in EPSC release probability were also observed (Table 3). This suggests a potential enhancement of postsynaptic inhibitory function without a presynaptic effect, leading to an overall reduction in the E/I ratio. To determine the effect of increased inhibition on LTP deficits following CA/CPR, we performed extracellular field recordings in the CA1-Schaffer collateral pathway and recorded fEPSPs (Fig. 1C). We then treated hippocampal sections from CA/CPR mice with picrotoxin (5 mM; Costa and Grybko, 2005), a GABAAR pore blocker, to test the effect of GABAergic inhibition on LTP. Following theta burst stimulation (TBS), CA/CPR slices without treatment exhibited impaired LTP, consistent with previous reports (Fig. 1D; Table 2; Orfila et al., 2014, 2018; Dietz et al., 2020). However, treatment with picrotoxin restored LTP to sham levels (Fig. 1D,E; Table 2). No differences in the input–output curve (Table 4) and paired-pulse ratio (Table 4) were observed in the field recordings, suggesting enhanced postsynaptic GABAAR function contributes to LTP deficits at postacute timepoints following CA/CPR.
Excitatory/inhibitory balance is disrupted following CA/CPR. A, Shown are representative traces for evoked EPSCs (solid) and IPSCs (dotted) recorded from the same CA1 hippocampal neurons of CA/CPR (blue) and sham (gray/black) mice. The AMPA eEPSCs are recorded with the neuron held at −70 mV, and GABA eIPSCs are recorded at 0 mV. B, Quantification of the absolute value of the ratio of GABA eIPSC amplitude to the AMPA eEPSC amplitudes; n = 11–15 cells/3–5 animals per condition; unpaired t test. C, Representative fEPSP traces from the stratum radiatum region of the CA1 hippocampus in CA/CPR slices without treatment (black) and CA/CPR slices (blue) treated with 5 µM picrotoxin during baseline. D, LTP data presented as a percentage of the baseline, where the baseline is set at 100%. E, Normalized slope of fEPSPs after theta burst stimulus; n = 4–6 slices/4–5 animals per condition; one-way ANOVA, Tukey's post hoc test. Values represent mean ± SEM. *p < 0.05, **p < 0.01.
Statistical details for Figure 1
Summary of passive membrane properties and paired-pulse ratio for whole-cell evoked E/I experiments, related to Figures 1, A and B, and 2, G and H
Summary of fEPSP input–output slope and paired-pulse ratio for field electrophysiology experiments, related to Figure 1, C–E
GABA sIPSC amplitude is increased following CA/CPR: inhibition of the TRPM2 ion channel rapidly restores sIPSC amplitude and E/I balance post-CA/CPR
To determine whether the increase in inhibitory function is pre- or postsynaptic, we recorded spontaneous IPSCs (sIPSCs) from CA1 neurons in CA/CPR and sham-operated mice and assessed amplitude, frequency, and decay kinetics (Fig. 2A; Table 5). The cumulative frequency of sIPSCs amplitude was shifted rightward in CA/CPR mice compared with sham-operated mice, indicating an overall increase in amplitude (Fig. 2B). Consistent with this, CA/CPR mice exhibited increased mean sIPSC amplitude compared with sham (Fig. 2C1; Table 6). No differences in mean frequency (Fig. 2C2; Table 6) or tau decay (Fig. 2C3; Table 6) were detected, suggesting the increase in inhibitory function is primarily postsynaptic.
GABA sIPSC amplitude is increased following CA/CPR. TRPM2 ion channel inhibition rapidly restores sIPSC amplitude and E/I balance post-CA/CPR. A, Representative traces from sham (top, left), sham with 2 µM tatM2NX bath applied (bottom, left), CA/CPR (top, right), CA/CPR with bath applied tatM2NX (bottom right) of whole-cell voltage-clamp recordings of sIPSC events recorded from CA1 pyramidal neurons in acute hippocampal slices using CsCl based internal solution holding at −60 mV. B, Cumulative frequency distribution of sIPSC amplitude from sham cells (black, solid), sham + tatM2NX (black, dotted), CA/CPR (pink, solid), CA/CPR + tatM2NX (pink, dotted). C, Mean amplitude (1), frequency (2), and tau decay (3) kinetics were measured from sIPSC events in different cells from sham and CA/CPR operated mice with and without bath application of tatM2NX; n = 18–24 cells/8–10 animals per condition; one-way ANOVA, Tukey's post hoc test. D, Representative traces from sham (top, left), sham after clotrimazole (CTZ) from the same cell (bottom, left), CA/CPR (top, right), CA/CPR after CTZ from the same cell (bottom, right). E, Cumulative frequency distribution of sIPSC amplitude from CA/CPR (black, dotted) following treatment with CTZ (20 mM) in the same cell (green, dotted). F, Mean amplitude (1), frequency (2), and tau decay (3) kinetics were measured from sIPSC events in the same cell before and after bath application of CTZ from sham and CA/CPR operated mice; n = 21–22 cells/9–11 animals per condition; two-way ANOVA with repeated measures, Sidak's post hoc test. G, Shown are representative traces for evoked EPSCs (solid) and IPSCs (dotted) recorded from the same CA1 hippocampal neurons of CA/CPR (blue) and CA/CPR with CTZ treatment (gray) mice. The AMPA eEPSCs are recorded with the neuron held at −70 mV, and GABA eIPSCs are recorded at 0 mV. H, Quantification of the absolute value of the ratio of GABA eIPSC amplitude to the AMPA eEPSC amplitudes; n = 11–15 cells/3–5 animals per condition; unpaired t test. Sham and CA/CPR data from Figure 1B was used to compare the CA/CPR + CTZ group. Values represent mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.
Statistical details for Figure 2
Given the potential link between elevated GABAergic inhibition and LTP impairments in our experiments, we next asked whether TRPM2 inhibition, previously shown to restore LTP after CA/CPR (Dietz et al., 2020, 2021), potentially acts through postsynaptic GABAergic changes. To test this, we measured sIPSCs from CA1 neurons and bath applied tatM2NX (2 mM), a potent and specific TRPM2 inhibitor (Cruz-Torres et al., 2020), for at least 20 min prior to recording. CA/CPR slices treated with tatM2NX showed a reduction in sIPSC amplitude compared with CA/CPR slices without treatment, shifting the cumulative frequency of sIPSCs amplitude leftward (Fig. 2B). The mean amplitude was also reduced in tatM2NX-treated CA/CPR slices compared with slices without treatment (Fig. 2C1; Table 6). No differences in mean frequency (Fig. 2C2; Table 6) or tau decay (Fig. 2C3; Table 6) were detected across all conditions.
We then wanted to determine whether TRPM2 inhibition can rapidly reverse the CA/CPR-induced increase in amplitude within the same cell. We patched CA1 neurons from sham and CA/CPR mice and measured sIPSC amplitude, frequency, and tau decay before and after bath application of clotrimazole (CTZ, 20 mM), a fast-acting TRPM2 pore blocker (Table 5; Hill et al., 2004). CA/CPR slices treated with CTZ shifted the cumulative frequency of sIPSC amplitude leftward (Fig. 2E), indicating CTZ treatment restored sIPSC amplitude to sham levels. TRPM2 inhibition by CTZ rapidly reduced the CA/CPR-induced increase in sIPSC mean amplitude (Fig. 2F1; Table 6). Surprisingly, we observed a reduction in mean frequency in CA/CPR slices following bath application of CTZ (Fig. 2F2; Table 6). This was likely an off-target effect as the more specific TRPM2 inhibitor, tatM2NX, had no effect on frequency (Fig. 2C2; Table 6). However, consistent with the tatM2NX data, no differences were detected in tau decay across all conditions (Fig. 2F3; Table 6). These data suggest the increase in GABAergic function observed following CA/CPR is primarily postsynaptic and can be rapidly reversed following TRPM2 ion channel inhibition.
We next tested whether TRPM2 blockade would also reverse the CA/CPR-induced increase in the GABA:AMPA ratio shown in Figure 1B. To assess this, we again recorded evoked GABA (0 mV) and AMPA (−70 mV) responses from CA1 pyramidal neurons 7 d following CA/CPR and bath applied CTZ (20 mM; Fig. 2G). As predicted, CTZ treatment restored the GABA:AMPA ratio to sham levels (Fig. 2H; Table 6). Taken together, these findings indicate ongoing TRPM2 activity increases postsynaptic GABAergic function, resulting in disrupted E/I balance after CA/CPR.
CA/CPR induces a persistent increase in the clustering and density of gephyrin
We next investigated the mechanisms underlying the increase in inhibitory synaptic function driven by TRPM2 following CA/CPR. The increase in sIPSC amplitude suggests more postsynaptic receptor clustering at the synapse and a larger postsynaptic domain. To test this, we performed immunohistochemistry 7 d following sham and CA/CPR procedures and stained for the postsynaptic inhibitory scaffold protein, gephyrin, and the presynaptic inhibitory synapse marker, vesicular GABA transporter (VGAT; Fig. 3A,B). There was no difference in VGAT cluster area nor density between CA/CPR and sham in either the stratum radiatum (Fig. 3C; Table 7) or the stratum pyramidale (Fig. 3D; Table 7). In contrast, we observed an increase in gephyrin cluster density in the stratum radiatum (Fig. 3C; Table 7). In the stratum pyramidale, both gephyrin cluster area and density were increased in CA/CPR mice compared with sham (Fig. 3D; Table 7). These changes in postsynaptic GABAergic components were not due to altered gephyrin protein expression (Fig. 4A,B; Table 8) or altered transcription of various GABAAR subunits (Fig. 4C,E; Table 8). These findings align well with our electrophysiology data, suggesting CA/CPR enhances postsynaptic GABAergic function through increased receptor density, which can be rapidly reversed without affecting presynaptic inhibitory function.
CA/CPR induces a persistent increase in gephyrin clustering and density. A, B, Shown are representative images of the CA1 region of the hippocampus from (A) sham-operated and (B) CA/CPR mice. Gephyrin staining is shown in green and VGAT is shown in magenta. C, Quantification of the cluster area and density of both gephyrin and VGAT in the stratum radiatum; n = 6 animals per group; unpaired t test. D, Quantification of the cluster area and density of both gephyrin and VGAT in the stratum pyramidale; n = 6 animals per group; unpaired t test. Values represent mean ± SEM. *p < 0.05; **p < 0.01.
Protein and mRNA expression of postsynaptic GABAergic components are unaltered 7 d following CA/CPR. A, Western immunoblot measuring total levels of gephyrin normalized to vinculin loading control in the P2 fraction from whole hippocampi. B, Quantification of gephyrin levels normalized to mean sham; n = 4 animals per condition; unpaired t test. C–E, Levels of mRNA transcript in CA1 hippocampal isolates as measured by quantitative qPCR for genes encoding the GABAAR-γ2 (C), GABAAR-b3 (D), GABAAR-a1 (E) subunits; n = 8 animals per condition; unpaired t test. Values represent mean ± SEM.
Statistical details for Figure 3
Statistical details for Figure 4
OGD induces a persistent increase in the clustering and density of postsynaptic GABAergic proteins
To directly visualize GABAARs and interrogate the mechanisms contributing to the enhancement of synaptic GABAAR clustering, we extended our studies to a well-established in vitro model of GCI (Arancibia-Carcamo et al., 2009; Smith et al., 2017; Garcia et al., 2021). We exposed dissociated hippocampal neurons to 20 min oxygen–glucose deprivation (OGD) followed by reoxygenation and fixed the neurons at varying timepoints to mimic the delayed 7 d timepoint we analyzed post-CA/CPR (Fig. 5A). To assess inhibitory synaptic size and postsynaptic receptor clustering, we immunostained for gephyrin, surface GABAARs (γ2 subunit), and VGAT and evaluated cluster area and density of these markers in dendrites of pyramidal neurons using confocal microscopy (Fig. 5B). We observed a decrease in cluster area and density of all synaptic GABAergic components 24 h post-OGD, consistent with prior work showing an acute loss of GABAergic synapses (Garcia et al., 2021). However, by 48 h, these measurements were no different from control levels. By 72 and 96 h, both the cluster area (Fig. 5C,E; Table 9) and density (Fig. 5F,H; Table 9) increased for all synaptic GABAergic components. The increase in the presynaptic marker, VGAT, was unexpected given the in vivo CA/CPR data showed no changes in VGAT clustering (Fig. 5C,D; Table 7) or a presynaptic functional effect (Table 6). Despite this discrepancy, both postsynaptic markers exhibited an acute reduction in clustering followed by an increase in the chronic phase, consistent with the in vivo results shown prior (Fig. 3). To ensure the nonpermeabilization protocol used for these experiments reliably labeled the surface GABAARs, we stained for all three GABAergic markers under nonpermeabilizing (Fig. 6A) and permeabilizing conditions (Fig. 6B). As expected, we observed robust GABAAR staining under nonpermeabilizing conditions, while immunostaining for gephyrin and VGAT is notably absent due to lack of antibody access to cytoplasmic epitopes without permeabilization (Fig. 6A). Staining for all three GABAergic markers is present under permeabilizing conditions (Fig. 6B), suggesting the nonpermeabilization protocol used here for surface GABAAR staining is reliable. Finally, to ensure cell membrane integrity was not compromised by the OGD stimulus, we stained for the somatodendritic marker, MAP2, and found no evidence of membrane instability after 20 min OGD compared with the control condition (Fig. 6C). The GABAergic synaptic markers used for our study were also sufficient to delineate dendrites as we observed a similar increase in GABAAR cluster area and density using MAP2 as a guide for dendritic analysis (Fig. 6D; Table 10). Thus, we utilized this optimized in vitro protocol, specifically the 96 h timepoint, to investigate the mechanism of TRPM2-mediated postsynaptic GABAergic enhancement.
OGD induces a persistent increase in the clustering and density of postsynaptic GABAergic proteins. A, Cartoon illustrating timeline for in vitro OGD experiments. Dissociated hippocampal neurons were reoxygenated following 20 min OGD exposure. Neurons were fixed 24, 48, 72, and 96 h following reoxygenation. B, Representative confocal images of dendritic segments from pyramidal neurons stained for gephyrin (green), GABAAR-γ2 subunit (cyan), and VGAT (magenta). C–E, Quantification of cluster area for (C) gephyrin, (D) surface GABAAR-γ2, and (E) VGAT; n = 30–36 neurons per condition; one-way ANOVA, Dunnett's post hoc. F–H, Quantification of cluster density for (F) gephyrin, (G) surface GABAAR-γ2, and (H) VGAT; n = 30–36 neurons per condition; one-way ANOVA, Dunnett's post hoc. Values represent mean ± SEM. *p < 0.05; **p < 0.01, ***p < 0.001, ****p < 0.0001.
Validation of immunostaining in oxygen–glucose deprivation experiments. A, Representative low magnification image of CA1 pyramidal neuron (left) and high magnification images of dendritic segments (right) following gephyrin (green), GABAAR-γ2 subunit (cyan), and VGAT (magenta) staining under nonpermeabilizing conditions 96 h following OGD. B, Representative low magnification image of CA1 pyramidal neuron (left) and high magnification images of dendritic segments (right) following gephyrin (green), GABAAR-γ2 subunit (cyan), and VGAT (magenta) staining under permeabilizing conditions 96 h following OGD. C, Low magnification representative images of CA1 pyramidal neurons immunostained for MAP2 (blue), GABAAR-γ2 subunit (cyan), and VGAT (magenta) following control (top) and 96 h-OGD (bottom) treatment. D, Subset analysis of cluster area (left) and cluster density (right) GABAAR-γ2 subunit are increased 96 h following OGD using MAP2 as a guide for ROI analysis; n = 9–12 neurons per condition, unpaired t test. Values represent mean ± SEM. *p < 0.05; ***p < 0.001.
Statistical details for Figure 5
Statistical details for Figure 6
The TRPM2 ion channel mediates the OGD-induced increase in clustering of postsynaptic GABAergic components
We next investigated whether TRPM2 activation affects the density of postsynaptic GABAergic proteins following OGD. We subjected hippocampal neurons to 20 min OGD reoxygenation and treated with either tatM2NX or CTZ 1 h prior to fixation (Fig. 7A). We again immunostained for gephyrin, surface GABAAR-γ2, and VGAT and imaged dendritic segments of pyramidal neurons (Fig. 7B). Treatment with tatM2NX, a potent and specific TRPM2 inhibitor, reduced the OGD-induced increase in gephyrin (Fig. 7C; Table 11) and surface GABAAR-γ2 subunit cluster area and density (Fig. 7D; Table 11) as well as the presynaptic marker, VGAT (Fig. 7E; Table 11). We then performed similar experiments using the noncompetitive TRPM2 blocker, CTZ. We found CTZ treatment also significantly reduces the OGD effect on gephyrin (Fig. 7F; Table 11) and surface GABAAR-γ2 (Fig. 7G; Table 11) cluster area and density, with a reduction in the OGD-induced increase in VGAT cluster density (Fig. 7H; Table 11). Using two pharmacological inhibitors of TRPM2, our data strongly support that the persistent OGD-induced increase in postsynaptic GABAAR density is TRPM2-dependent, consistent with the in vivo functional data which implicates TRPM2 activity in the increase in GABAergic sIPSC amplitude post-CA/CPR (Fig. 2).
The TRPM2 ion channel mediates the OGD-induced increase in the clustering of postsynaptic GABAergic proteins. A, Cartoon illustrating experimental timeline for neurons subjected to OGD reoxygenation and treated with tatM2NX (2 mM) or CTZ (20 mM) 1 h prior to fixation. B, Representative confocal images of dendritic segments from pyramidal neurons immunostained for gephyrin (green), GABAAR-γ2 subunit (cyan), and VGAT (magenta). C–E, Quantification of cluster area (left) and cluster density (right) following treatment with tatM2NX for (C) gephyrin, (D) surface GABAAR-γ2, and (E) VGAT; n = 30–36 neurons per condition; one-way ANOVA, Tukey's post hoc. F–H, Quantification of cluster area (left) and cluster density (right) following treatment with CTZ for (F) gephyrin, (G) surface GABAAR-γ2, and (H) VGAT, n = 30–36 neurons per condition; one-way ANOVA, Tukey's post hoc. Values represent mean ± SEM. *p < 0.05; **p < 0.01, ***p < 0.001, ****p < 0.0001.
Statistical details for Figure 7
Ca2+ signaling and CaMKII activity contribute to CA/CPR-induced increase in GABA sIPSC amplitude
While these data implicate TRPM2 in the ischemia-induced enhancement of postsynaptic inhibitory function, the downstream signaling required to regulate GABAAR density remains yet to be determined. We hypothesized that TRPM2-mediated Ca2+ influx may modulate postsynaptic GABAAR function. To examine this, we performed patch-clamp recordings with a high concentration of BAPTA (10 mM), a selective Ca2+ buffer, in the patch pipette and recorded GABA sIPSCs from the same cell before and after CTZ treatment (Fig. 8A). Our results showed Ca2+ chelation in the postsynaptic neuron occluded the CA/CPR-induced enhancement of sIPSC mean amplitude, as there was no longer a difference between CA/CPR and sham conditions (Fig. 8B1; Table 12). Further, we did not observe a change in amplitude following CTZ treatment in cells recorded with BAPTA in the patch pipette (Fig. 8B1; Table 12). We also found no differences between the CA/CPR and sham conditions and the CTZ treatment in the frequency (Fig. 8B2; Table 12) or the tau decay (Fig. 8B3; Table 12) of sIPSCs when BAPTA was present. Given the lack of additive effect of BAPTA and CTZ on sIPSC amplitude, this suggests that postsynaptic Ca2+ signaling mediates the observed TRPM2-dependent increase in inhibition.
Ca2+ signaling and CaMKII activity contribute to CA/CPR-induced increase in the amplitude of GABA sIPSCs. A, Representative traces from sham (far left), sham after CTZ from the same cell (center left), CA/CPR (center right), and CA/CPR after CTZ from the same cell (far right). BAPTA (10 mM) was included in the patch pipette across all conditions. B, Mean amplitude (1), frequency (2), and tau decay (3) kinetics were measured from sIPSC events in the same cell before and after bath application of CTZ from sham and CA/CPR operated mice with 10 mM BAPTA included in patch pipette; n = 10 cells/4–7 animals per condition; two-way ANOVA with repeated measures, Sidak's post hoc test. C, Western immunoblot measuring total levels of CaMKII normalized to vinculin loading control in the membrane fraction of whole hippocampi from sham and CA/CPR operated mice; n = 5 animals per condition; unpaired t test. D, Western immunoblot measuring levels of T286 phosphorylation of CaMKII normalized to total CaMKII in the membrane fraction of whole hippocampi from sham and CA/CPR operated mice; n = 4 animals per condition; unpaired t test. E, Representative traces from sham (far left), sham after CTZ from the same cell (center left), CA/CPR (center right), and CA/CPR after CTZ from the same cell (far right). TatCN19o (5 mM) was included in the patch pipette across all conditions. F, Mean amplitude (1), frequency (2), and tau decay (3) kinetics were measured from sIPSC events in the same cell before and after bath application of CTZ from sham and CA/CPR operated mice with 5 mM tatCN19o included in patch pipette; n = 9 cells/45 animals per condition; two-way ANOVA with repeated measures, Sidak's post hoc test. Values represent mean ± SEM. *p < 0.05.
Statistical details for Figure 8
CaMKII is a Ca2+-dependent kinase that has been previously shown to be required for activity-dependent postsynaptic GABAAR potentiation and clustering (Petrini et al., 2014; Chiu et al., 2018; Cook et al., 2021, 2022). We therefore hypothesized that TRPM2-mediated Ca2+ signaling might activate CaMKII to enhance postsynaptic GABAergic function. To test this, we used western immunoblotting to detect total protein levels of CaMKII in the membrane fractions of the whole hippocampus from CA/CPR and sham mice 7 d following the surgeries. Our data show that CaMKII levels are significantly reduced in the membrane fraction of whole hippocampi (Fig. 8C; Table 12). Interestingly, we observed relatively higher T286 phosphorylation of CaMKII in the synaptic membrane fraction obtained from CA/CPR mice compared with sham (Fig. 8D; Table 12), indicating a sustained increased synaptic CaMKII activity following GCI.
To further examine the role of elevated CaMKII activity on postsynaptic GABAergic function following CA/CPR, we used patch-clamp electrophysiology to record sIPSCs while inhibiting postsynaptic CaMKII using tatCN19o (5 mM) in the patch pipette (Fig. 8E). Our results show that inhibition of postsynaptic CaMKII activity occluded the CA/CPR-induced increase in sIPSC amplitude, as indicated by equivalent sIPSC amplitudes observed in tatCN19o treated cells and sham conditions (Fig. 8F1; Table 12). Additionally, CTZ no longer impacted the mean amplitude of sIPSCs in the CA/CPR mice recorded with tatCN19o (Fig. 8F1; Table 12). Similar to the effects observed with Ca2+ chelation, we did not observe a change between the CA/CPR and sham conditions and the CTZ treatment in the frequency (Fig. 8F2; Table 12) and tau decay (Fig. 8F3; Table 12) of sIPSCs when the CaMKII inhibitor was present. These results indicate that CaMKII-dependent increases in postsynaptic inhibitory function could be downstream of Ca2+ entry via TRPM2 channel activity.
Ca2+-dependent CaMKII activity contributes to increased clustering of postsynaptic GABAergic components via a TRPM2-dependent pathway
We next investigated whether TRPM2- and CaMKII-mediated enhancement of postsynaptic sIPSC amplitude was due to an increase in the density of postsynaptic GABAARs. To assess this, we used the 96 h timepoint following OGD reoxygenation in dissociated rat hippocampal neurons and treated cells with vehicle (control), tatCN19o alone, or combined tatCN19o and tatM2NX 1 h prior to fixation (Fig. 9A) and immunostained for GABAergic synaptic markers (Fig. 9B). We found that treatment with tatCN19o reduced the cluster area and density of the GABAAR-γ2 (Fig. 9D; Table 13) and gephyrin (Fig. 9C; Table 13) following OGD compared with the OGD vehicle control. Additionally, we observed a significant decrease in the cluster area and density of the postsynaptic components in the combined tatM2NX and tatCN19o condition compared with the control (veh) condition following OGD; however, there was no additional reduction in the cluster area and density of the postsynaptic inhibitory synaptic proteins in the combined treatment condition compared with the tatCN19o alone (Fig. 9C,D; Table 13). We found no differences across all conditions when measuring the VGAT cluster area, but we did observe a decrease in VGAT cluster density in the combined condition (Fig. 9E; Table 13). These results suggest that TRPM2 and CaMKII likely converge on the same pathway to regulate postsynaptic GABAAR density.
Ca2+-dependent CaMKII activity contributes to the increased clustering of postsynaptic GABAergic proteins via a TRPM2-dependent pathway. A, Cartoon illustrating experimental timeline for neurons subjected to OGD reoxygenation and treated with tatCN19o (5 mM), KN93 (5 mM), or combined treatment with tatM2NX (2 mM) 1 h prior to fixation. B, Representative confocal images of dendritic segments from pyramidal neurons stained for gephyrin (green), GABAAR-γ2 subunit (cyan), and VGAT (magenta). C–E, Quantification of cluster area (left) and density (right) following treatment with tatCN19o or tatCN19o + tatM2NX for (C) gephyrin, (D) surface GABAAR-γ2, and (E) VGAT; n = 30–36 neurons per condition; one-way ANOVA, Tukey's post hoc. F–H, Quantification of cluster area (left) and density (right) following treatment with KN93 or KN93 + tatM2NX for (E) gephyrin, (F) surface GABAAR-γ2, and (G) VGAT; n = 30–36 neurons per condition; one-way ANOVA, Tukey's post hoc. Values represent mean ± SEM. *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001.
Statistical details for Figure 9
The tatCN19o peptide was shown to block both the Ca2+-independent autonomous and Ca2+-stimulated CaMKII activity (Vest et al., 2010). Based on our functional data showing that Ca2+ signaling was required for the increase in sIPSC amplitude and is reversible with Ca2+ chelation (Fig. 9), we hypothesized that the Ca2+-stimulated CaMKII activity contributes to elevated GABAAR clustering. To test this, we treated neurons 96 h following OGD reoxygenation with KN93 (5 mM), a small molecule CaMKII inhibitor known to preferentially block Ca2+-stimulated CaMKII activity (Fig. 9A; Vest et al., 2010). We found that KN93 significantly reduced the cluster area and density of both postsynaptic GABAergic proteins following OGD compared with the OGD vehicle control condition (Fig. 9F,G; Table 13). This reduction in cluster area of postsynaptic components persisted in the combined KN93 and tatM2NX condition, showing no differences compared with KN93 treatment alone following OGD (Fig. 9F,G; Table 13). While there was a significant increase in the presynaptic marker, VGAT, cluster area, and density in the vehicle OGD condition compared with the control no OGD group, neither KN93 nor the combined KN93 and tatM2NX treatment reduced the VGAT cluster size and density (Fig. 9H; Table 13). Altogether, these data suggest that the TRPM2 channel and Ca2+-stimulated CaMKII activity converge on the same molecular pathway to regulate postsynaptic GABAA receptor density at delayed timepoints following OGD.
Discussion
Here, we combine in vitro and in vivo approaches to elucidate the enhancement of postsynaptic GABAergic function in the hippocampus, accounting for the reduction in the E/I ratio and driving the LTP deficits following CA/CPR. Using an in vitro model that emulates the direct ischemic insult to the hippocampus observed in GCI, we found a shift from an acute reduction in GABAergic signaling to a sustained elevation in the clustering of GABAergic proteins at synaptic sites following OGD reoxygenation. To our knowledge, this is the first study to employ the OGD in vitro system to examine synaptic alterations days after ischemic insult following the completion of cell death processes. Using this novel in vitro paradigm and a murine model of GCI, we identified the TRPM2 ion channel as a potential mediator of E/I imbalance and GABAergic synaptic enhancement. TRPM2 inhibition effectively blocked the CA/CPR-induced increase in GABA:AMPA ratio and sIPSC amplitude and the OGD-induced effect on cluster area of postsynaptic GABAergic components. Furthermore, our data revealed that TRPM2 and Ca2+-CaMKII are both required for the elevated postsynaptic GABAergic function. Chelation of Ca2+ and CaMKII blockade both alleviated the CA/CPR effect on sIPSC amplitude and reversed the OGD-induced increase in cluster area of synaptic GABAergic proteins. Additionally, blockade of the TRPM2 channel showed no additive effect on the reduction in sIPSC amplitude or GABAAR clustering following CaMKII inhibition, suggesting that TRPM2 and CaMKII reside in the same pathway to regulate inhibitory synaptic function after ischemia.
The elevation of excitatory signaling with concurrent loss of GABAergic inhibitory synapses in the acute period is well-defined (Alicke and Schwartz-Bloom, 1995; Schwartz-Bloom and Sah, 2001; Mele et al., 2014). E/I balance and the role of GABAergic signaling beyond acute cell death has received relatively less attention; however, emerging work provides evidence of reduced E/I balance in the surviving network, primarily driven by enhanced inhibition through extrasynaptic (tonic) GABAARs (Carmichael, 2012; Joy and Carmichael, 2021). Excessive tonic inhibition has been demonstrated to hinder cortical and hippocampal recovery in animal models of focal ischemia (Clarkson et al., 2010; Lake et al., 2015; Orfila et al., 2019). Additionally, one study has shown phasic inhibition, achieved through rapid activation of synaptic GABAARs, is enhanced in the peri-infarct cortex following stroke, correlating with motor deficits (Hiu et al., 2016). In line with these findings, our study contributes to the existing literature by describing elevated postsynaptic phasic inhibition in the hippocampus following global insult and its potential role in impairing excitatory LTP. The question of whether excessive tonic inhibition also contributes to hippocampal synaptic plasticity deficits following GCI warrants further investigation. Altogether, our observation of elevated postsynaptic (phasic) GABAergic signaling, which can be pharmacologically interrupted/targeted without the unfavorable adverse effects of directly blocking GABAAR, offers a promising approach for therapeutic development.
In our study, we identify the TRPM2 ion channel as a novel mediator of inhibitory function and E/I balance following ischemia. Although TRPM2 has been extensively studied in the context of neuronal cell death, with acute inhibition of the channel conferring neuroprotection (Jia et al., 2011; Nakayama et al., 2013; Shimizu et al., 2013), recent evidence has highlighted its involvement in the postacute phase, indicating a role for the channel in mechanisms in the expression of synaptic plasticity and its promise as a neurorestorative therapeutic target (Xie et al., 2011; Dietz et al., 2020, 2021). Our prior study revealed that sustained TRPM2 activity following GCI has been found to contribute to ischemia-induced LTP impairment via a calcineurin-GSK3b-dependent pathway within glutamatergic synapses (Dietz et al., 2020). In contrast, our data obtained here is the first to suggest an additional role for TRPM2 in the regulation of inhibitory synaptic strength and reveal Ca2+ and CaMKII as the most plausible downstream mediators. Consistent with this, other work supports TRPM2 as the Ca2+ source for CaMKII activity in the context of cell cycle regulation (Wang et al., 2017; Cai et al., 2023), suggesting this may be a common mechanism across multiple cellular functions. Moreover, our BAPTA data indicate that continuous Ca2+ stimulus is required for enhanced sIPSC amplitude, a mechanism likely mediated by ongoing CaMKII activation downstream of TRPM2-Ca2+ influx (Fig. 8A,B). Notably, a prominent property of the TRPM2 channel is its ability to conduct substantial Ca2+ current due to its high Ca2+ permeability and prolonged open time of several hundred milliseconds (Perraud et al., 2001; Sano et al., 2001; Kraft et al., 2004). Nonetheless, more work directly linking TRPM2-Ca2+ influx upstream of CaMKII activation should be a priority for future investigation but remains technically challenging due to the lack of reliable antibodies directed at the channel (Dietz et al., 2020) and the poorly understood mechanism by which the channel is activated within the postsynaptic microdomain.
The Ca2+-activated signaling cascade downstream of TRPM2 ion channel activation is important to elucidate both for increased understanding of the mechanism of ischemia-induced alterations in E/I balance as well as future therapeutic development. Using two distinct pharmacological inhibitors of CaMKII, our data show that Ca2+-stimulated CaMKII activity is required for the elevation in postsynaptic GABAergic inhibition following GCI. Specifically, treatment with KN93, an inhibitor known to preferentially block Ca2+-dependent CaMKII (Vest et al., 2010), occluded the OGD effect on clustering (Fig. 9F,G). Consistent with this, we find that the tatCN19o peptide, which blocks both stimulated and autonomous CaMKII activity (Vest et al., 2010), also reduced the OGD effect on postsynaptic GABAergic clustering (Fig. 9C,D) and the CA/CPR-induced increase in sIPSC amplitude (Fig. 8F1). Thus, we cannot exclude the involvement of autonomous CaMKII in this mechanism. These data agree with the literature implicating CaMKII in inhibitory synaptic potentiation. Autonomous CaMKII (T286phosphorylated) constitutively localizes to inhibitory synapses (Marsden et al., 2010; Cook et al., 2022), and intracellular application of CaMKII was shown to enhance GABA IPSCs (Wang et al., 1995; Wei et al., 2004). Following its movement to inhibitory synapses, CaMKII was shown to phosphorylate b2 or b3 subunits of GABAARs, resulting in GABAAR synaptic aggregation and inhibitory potentiation (Petrini et al., 2014; Chiu et al., 2018). Exploring whether CaMKII moves to inhibitory synapses and directly phosphorylates GABAergic proteins to alter E/I balance after ischemia is an important avenue for future investigation. One caveat to targeting CaMKII to restore ischemia-induced E/I imbalance is its well-established role in mediating AMPA receptor currents (Kristensen et al., 2011). Thus, both excitatory and inhibitory signaling would likely be affected by CaMKII inhibition. To mitigate the effect of CaMKII activation on excitatory transmission, future work should consider precise targeting of its impact on inhibitory synapses, potentially through blockade of its potential GABAergic phosphorylation sites.
Accumulating evidence suggests inhibitory LTP and homeostatic scaling up share downstream signaling pathways to regulate synaptic strength (Vitureira and Goda, 2013; Galanis and Vlachos, 2020). Exocytosis and subsequent lateral diffusion of GABAARs from the extrasynaptic membrane is a well-documented mechanism for rapidly fine-tuning inhibitory synaptic strength (Luscher et al., 2011). Multiple findings here support the notion that GABAAR clustering is increased through lateral diffusion. First, our data demonstrate that sIPSC amplitude and GABAAR clustering can be restored rapidly (within minutes to 1 h). Second, we observe unaltered mRNA expression of GABAAR transcripts and protein levels of gephyrin post-CA/CPR, indicating no changes in the transcription and translation of GABAergic components at this timepoint. Lastly, our findings raise the intriguing possibility that increased inhibitory synaptic function represents a maladaptive homeostatic response to the acute loss of inhibition following excitotoxic insult. The OGD reoxygenation data hints at this possibility, revealing an initial acute reduction in the clustering of GABAergic components, followed by a steadily increasing and sustained increase in GABAergic synapses at later timepoints.
In summary, this study uncovers a novel mechanism whereby enhanced GABAergic inhibition impairs excitatory synaptic plasticity in the context of cerebral ischemia, revealing TRPM2 and CaMKII as targets for pharmacological intervention. This is particularly significant given the clinical challenges associated with direct GABAAR modulation. GABAAR antagonism induces epileptic seizures (Sperk et al., 2004), while GABAAR agonists largely failed in clinical trials as a treatment for acute ischemic stroke due to lack of efficacy and patients reporting numerous problematic side effects (Wahlgren et al., 2000; Lyden et al., 2002; Liu et al., 2018). Furthermore, our data shed light on the temporal changes in inhibitory synaptic function in response to ischemic insult. This observation may have a profound impact on the therapeutic window of some interventions, particularly GABA agonists, which have been shown to confer neuroprotection in animal models acutely (Liu et al., 2010, 2018) but, in light of this data, may be detrimental to functional recovery if administered at more chronic timepoints. Beyond ischemic injury, the pathway described here likely has broader implications considering that disruptions in synaptic accumulation of GABAARs are linked to numerous central nervous system disorders (Jacob et al., 2008; Mele et al., 2019). Further research into the precise molecular and cellular mechanisms underlying the influence of enhanced postsynaptic GABAergic function on excitatory LTP will be pivotal in developing targeted interventions to mitigate the long-term cognitive deficits associated with cerebral ischemia and various neurological diseases.
Footnotes
This work was supported by an NIH Predoctoral NRSA F31NS120422 and T32GM763540 (A.M.B.), R01NS046072 (N.Q. and P.S.H.), R01NS118786 (P.S.H), R01NS092645 (P.S.H.), and R01MH119154 (K.R.S.) R01MH128199 (K.R.S.) and an AHA Career Development Award (K.R.S.). J.D.G. was supported by Grant T32GM763540, an AHA Predoctoral Fellowship (19PRE34380542), and an NIH Blueprint Diversity Specialized Predoctoral to Postdoctoral Advancement in Neuroscience (D-SPAN) Award (FNS120640A). We thank Dr. Ulrich Bayer for the gift of tatCN19o peptide and Drs. Mark Dell’Acqua and Jacob Basak for critically reading the manuscript. Graphical abstract and OGD diagrams were created by BioRender.com.
↵*K.R.S., N.Q., and P.S.H. contributed equally to this work.
The authors declare no competing financial interests.
- Correspondence should be addressed to Paco S. Herson at paco.herson{at}osumc.edu or Nidia Quillinan at nidia.quillinan{at}cuanschutz.edu.
