Skip to main content
NCBI home page
Journal of Neurophysiology logoLink to Journal of Neurophysiology
. 2020 Jan 8;123(2):707–717. doi: 10.1152/jn.00570.2019

Endocannabinoid degradation inhibitors ameliorate neuronal and synaptic alterations following traumatic brain injury

Elizabeth A Fucich 1,2, Zachary F Stielper 1,2, Heather L Cancienne 3, Scott Edwards 1,2, Nicholas W Gilpin 1,2, Patricia E Molina 1,2, Jason W Middleton 2,3,✉
PMCID: PMC7052644  PMID: 31913777

Abstract

Our previous work showed that lateral fluid percussion injury to the sensorimotor cortex (SMC) of anesthetized rats increased neuronal synaptic hyperexcitability in layer 5 (L5) neurons in ex vivo brain slices 10 days postinjury. Furthermore, endocannabinoid (EC) degradation inhibition via intraperitoneal JZL184 injection 30 min postinjury attenuated synaptic hyperexcitability. This study tested the hypothesis that traumatic brain injury (TBI) induces synaptic and intrinsic neuronal alterations of L5 SMC pyramidal neurons and that these alterations are significantly attenuated by in vivo post-TBI treatment with EC degradation inhibitors. We tested the effects of systemically administered EC degradation enzyme inhibitors (JZL184, MJN110, URB597, or JZL195) with differential selectivity for fatty acid amide hydrolase (FAAH) and monoacylglycerol lipase (MAGL) on electrophysiological parameters in SMC neurons of TBI- and sham-treated rats 10 days post-TBI. We recorded intrinsic neuronal properties, including resting membrane voltage, input resistance, spike threshold, spiking responses to current input, voltage “sag” (rebound response to hyperpolarization-activated inward current), and burst firing. We also measured the frequency and amplitude of spontaneous excitatory postsynaptic currents. We then used the aggregate parameter sets (intrinsic + synaptic properties) to apply a machine learning classification algorithm to quantitatively compare neural population responses from each experimental group. Collectively, our electrophysiological and computational results indicate that sham neurons are the most distinguishable from TBI neurons. Administration of EC degradation inhibitors post-TBI exerted varying degrees of rescue, approximating the neuronal phenotype of sham neurons, with neurons from TBI/JZL195 (a dual MAGL/FAAH inhibitor) being most similar to neurons from sham rats.

NEW & NOTEWORTHY This study elucidates neuronal properties altered by traumatic brain injury (TBI) in layer 5 of sensorimotor cortex, which may be implicated in post-TBI circuit dysfunction. We compared effects of systemic administration of four different endocannabinoid degradation inhibitors within a clinically relevant window postinjury. Electrophysiological measures and using a machine learning classification algorithm collectively suggest that pharmacological inhibitors targeting both monoacylglycerol lipase and fatty acid amide hydrolase (e.g., JZL195) may be most efficacious in attenuating TBI-induced neuronal dysfunction at site of injury.

Keywords: electrophysiology, fatty acid amide hydrolase, machine learning classification, monoacylglycerol lipase, sensorimotor cortex

INTRODUCTION

Traumatic brain injury (TBI) in humans produces diffuse neuroinflammation and synaptic dysfunction that persists after resolution of clinical manifestations (Korn et al. 2005; Kumar and Loane 2012; Ramlackhansingh et al. 2011; Roche et al. 2004). These TBI effects on the brain may lead to short- and long-term neurobehavioral dysfunction associated with brain injury, including pain, anxiety, depression, cognitive deficits, seizures, and neurodegeneration (Brettschneider et al. 2012; Bryant et al. 2010; Eikelenboom et al. 2010; Hibbard et al. 1998; Nampiaparampil 2008; Ofek and Defrin 2007). However, the mechanisms underlying persistent TBI-related neuropathology remain poorly understood.

Recent work from various brain injury models demonstrates synaptic dysfunction in the neocortex at the site of injury (e.g., as reviewed in Carron et al. 2016). Electrophysiological studies have reported initial increases in excitability in layer 5 (L5) cortical neurons following cortical compression in rat barrel cortex (Ding et al. 2011), after mild fluid percussion injury in mouse sensorimotor cortex (SMC; Greer et al. 2012; Hånell et al. 2015), and following weight drop (diffuse) TBI in rat sensory cortex (Johnstone et al. 2013). Although the latter group found that L5 activity was decreased or unchanged at later time points after weight drop TBI (Allitt et al. 2016a, 2016b; Alwis et al. 2012), other studies suggest that TBI-induced increases in excitability are sustained or progressively increase over 2 wk following mild-to-moderate lateral fluid percussion injury (Mayeux et al. 2017) and severe controlled cortical impact (Yang et al. 2010) in rat SMC. Overall, lasting functional changes in L5 pyramidal neurons may lead to altered activity of cortical projections to subcortical regions.

Endocannabinoids (ECs) are synthesized and released on demand (e.g., in response to injury) and can promote synaptic homeostasis and reduce inflammation following an insult, suggesting that pharmacological inhibition of EC degradative enzymes may have therapeutic utility to ameliorate deleterious sequelae after brain injury (Cravatt et al. 1996; Dinh et al. 2002; Xu and Chen 2015). We previously tested systemic post-TBI administration of inhibitors of monoacylglycerol lipase (MAGL) and fatty acid amide hydrolase (FAAH), the enzymes that rapidly degrade the two most abundant ECs in the brain: 2-arachidonoyl glycerol (2-AG) and N-arachidonoylethanolamine (anandamide, AEA), respectively. Our prior work showed that acute treatment 30 min postinjury with JZL184, a selective MAGL inhibitor, was more effective than URB597, a selective FAAH inhibitor, in attenuating TBI-induced neuroinflammation, blood-brain-barrier (BBB) disruption, and acute neurobehavioral and neurological impairment in rats 1 day later (Katz et al. 2015). Subsequent studies using JZL184 also showed attenuated neuronal hyperexcitability at the site of injury and improved neurological and neurobehavioral impairment up to 14 days after TBI (Fucich et al. 2019; Mayeux et al. 2017). Other compounds have since been derived that provide higher potency and selectivity for rat MAGL in vivo (e.g., MJN110) (Chang et al. 2012; Niphakis et al. 2013). Additionally, the dual inhibitor JZL195 has since been developed, which targets both MAGL and FAAH. However, it remains to be determined which pharmacological strategy is most effective in attenuating long-term TBI-induced functional deficits, including alterations in neuronal and synaptic excitability.

This study aimed to characterize intrinsic properties and synaptic transmission in L5 SMC pyramidal neurons of rats euthanized 10–11 days after TBI plus treatment with inhibitors of MAGL or FAAH or both relative to vehicle-treated TBI rats and vehicle-treated sham control rats. We hypothesized that L5 SMC neurons from TBI rats would be characterized by altered synaptic and intrinsic properties relative to sham rats and that post-TBI EC degradation inhibition would attenuate post-TBI neuronal changes. Because of our previous results showing more robust effects after acute treatment with JZL184 vs. URB597, we hypothesized that either more potent and selective MAGL inhibition (via the drug MJN110) or dual inhibition of both MAGL and FAAH (via the drug JZL195) would produce the most robust attenuation of TBI-induced neuronal alterations compared with JZL184 or URB597. Our secondary goal was to assess whether these raw electrophysiological data parameters were optimal for distinguishing and classifying electrophysiological properties, and also whether data transformed using principal component analysis (PCA) were optimal for this purpose. We hypothesized that, using a machine learning algorithm, we would be able to classify L5 SMC neurons based on these electrophysiological parameters and that the algorithm would more often classify TBI neurons from drug-treated animals as sham-like neurons than TBI-like neurons.

MATERIALS AND METHODS

Animals

Male Wistar rats (Charles River Laboratories, Raleigh, NC) weighing 200–225 g on arrival were pair-housed in a temperature- and humidity-controlled room with a 12:12-h reverse light-dark cycle and ad libitum access to food and water. All animal procedures and experiments were approved by the Institutional Animal Care and Use Committee of the Louisiana State University Health Sciences Center and were in accordance with the guidelines of the National Institutes of Health.

Traumatic Brain Injury via Lateral Fluid Percussion

Animals received a 5-mm-diameter craniotomy above the left SMC (from bregma: anteroposterior −2 mm, mediolateral −3 mm) before undergoing TBI via lateral fluid percussion (fluid percussion injury device, model 01-B; Custom Design and Fabrication, Virginia Commonwealth University) as previously described (Teng and Molina 2014). Animals in the sham group were anesthetized and received craniotomy but were not subjected to TBI (surgical controls). Animals received an injury of ~2 atm of pressure, which produces a mild-to-moderate TBI, defined by brief systemic physiological alterations in the absence of structural damage to the brain (Dixon et al. 1987), and a low mortality rate (7.7% in the present study). Following surgery, topical lidocaine was applied to the incision site, and animals were singly housed and allowed to recover in their home cages until euthanized. The experimental approach and physiological responses elicited by this TBI model are described elsewhere (Teng and Molina 2014).

EC Degradative Enzyme Inhibition

As in our previous studies (Fucich et al. 2019; Katz et al. 2015; Mayeux et al. 2017) JZL184 (item no. 13158; Cayman Chemical, Ann Arbor, MI) was used to selectively inhibit MAGL, the enzyme that degrades 2-AG. JZL184 has displayed IC50 values of ~260 nM for rat MAGL and was the first inhibitor to demonstrate high selectivity (>100 fold) for MAGL over FAAH (Chang et al. 2012; Long et al. 2009a). MJN110 (product no. SML0872; Sigma-Aldrich, St. Louis, MO) was used to more potently and selectively inhibit MAGL, as it has demonstrated IC50 values of <100 nM in rat brain with >10,000-fold selectivity over FAAH (Niphakis et al. 2013). URB597 (item no. 10046; Cayman Chemical, Ann Arbor, MI) was used to selectively inhibit FAAH, the enzyme that degrades AEA. URB597 has displayed IC50 values of 4.6 nM for rat brain FAAH and demonstrates high selectivity (>7,500 fold) for FAAH over MAGL (Kathuria et al. 2003). JZL195 (item no. 13668; Cayman Chemical, Ann Arbor, MI) was used to dually inhibit MAGL and FAAH. JZL195 has displayed IC50 values of 100 nM for MAGL and 12 nM for FAAH in rat brain (Long et al. 2009b). JZL184 (16 mg/kg), MJN110 (20 mg/kg), URB597 (0.3 mg/kg), JZL195 (16 mg/kg), or vehicle (1:1:18 solution of alcohol, emulphor, and saline) was administered by intraperitoneal injection 30 min after TBI procedures at doses that elicit maximal enzyme inhibition and behavioral effects in vivo (Kathuria et al. 2003; Long et al. 2009a, 2009b; Niphakis et al. 2013).

Brain Slice Preparation

In preparation for electrophysiological experiments, animals were intracardially perfused with an N-methyl-d-glucamine (NMDG)-based artificial cerebrospinal fluid (ACSF) under deep isoflurane anesthesia 10–11 days post-TBI. Brains were removed and sliced (Leica VT1200-S vibratome; Leica, Buffalo Grove, IL) in room temperature ACSF containing (in mM) 92 NMDG, 2.5 KCl, 1.25 NaH2PO4, 30 NaHCO3, 20 HEPES, 25 glucose, 2 thiourea, 5 Na-ascorbate, 3 Na-pyruvate, 0.5 CaCl2, and 10 MgSO4; pH was titrated to 7.3–7.4 with concentrated hydrochloric acid (Ting et al. 2014). Coronal slices containing sensorimotor cortex underlying the site of injury were then transferred to a holding chamber (containing the same NMDG solution as during slicing) in a water heat bath kept at 37°C. After 12 min, slices were removed from the holding chamber in the heat bath and allowed to rest at room temperature in a chamber containing holding solution for 45 min before electrophysiological recordings. The holding solution consisted of (in mM) 92 NaCl, 2.5 KCl, 1.25 NaH2PO4, 30 NaHCO3, 20 HEPES, 25 glucose, 2 thiourea, 5 Na-ascorbate, 3 Na-pyruvate, 2 CaCl2, and 2 MgSO4.

Electrophysiology

At the time of recording, slices were transferred to slice chamber in the electrophysiology rig in a recording ACSF kept at 30–32°C by an in-line heater. The recording solution consisted of (in mM) 119 NaCl, 2.5 KCl, 1.25 NaH2PO4, 24 NaHCO3, 12.5 glucose, 2 CaCl2, and 2 MgSO4. Whole cell recordings were performed using borosilicate glass micropipettes (3–7 MΩ) filled with internal solution containing (in mM) 130 K-gluconate, 10 HEPES, 10 Na2-phosphocreatine, 4 MgCl2, 4 Na2-ATP, 0.4 Na-GTP, 3 ascorbic acid, and 0.2 EGTA (pH 7.25, 290–295 mosM). Electrical signals were amplified and digitized by a Multiclamp 700B amplifier (Molecular Devices, San Jose, CA). Recordings were sampled at 10 kHz and low-pass filtered with a cutoff of 4 kHz. Signals were further filtered offline in MATLAB with a cutoff of 2 kHz. Recordings were obtained from L5 pyramidal neurons at or near the center of the craniotomy site, most likely in the shoulder, limb, and trunk representation of primary SMC (Paxinos and Watson 2007). Nissl staining of the cortex within the injury domain did not reveal apparent visible differences in cell density (Fig. 1, A and B). However, moderate levels of cell death have been observed in the cortex at the site of injury, whereas there was relatively little cell death away from the site of injury (Ooigawa et al. 2006). To avoid this potential confound, we limited our electrophysiological recordings to neurons well within the circumference of the surgical craniotomy (Fig. 1C).

Fig. 1.

Fig. 1.

Open in a new tab

A: schematic of a coronal slice indicating typical recording regions. B: Nissl stain of a sample coronal slice from a Sham (top) and traumatic brain-injured (TBI) rat (bottom). C: 2-dimensional box plot showing the maximum/minimum, interquartile range, and mean of rostral-caudal and medial-lateral recording positions with respect to the perimeter of the craniotomy (circle). L, lateral; R, rostral.

Synaptic properties.

To collect data regarding pre- and postsynaptic alterations resulting from TBI and inhibitor treatment, spontaneous excitatory postsynaptic currents (sEPSCs) were recorded. When whole cell recording configuration was established, the voltage was clamped at −70 mV and 5 min of spontaneous postsynaptic current were recorded. Spontaneous postsynaptic events were detected in these recordings by thresholding rapid excursions in inward current; the average event amplitude and mean frequency over the 5-min recording period were quantified.

Intrinsic properties.

Because intrinsic subthreshold neural properties also contribute to integration of input synaptic signals and neural excitability, injury and/or treatment effects on intrinsic neural properties were assessed; specifically, input resistance, sag fraction, and resting membrane potential (RMP) were measured. To estimate input resistance, holding current was recorded in voltage-clamp mode with membrane voltage clamped at −70 mV. A −5-mV step was applied to the command voltage, and the change in holding current (ΔI) was measured. The input resistance was estimated using Ohm’s law, R = ΔV/ΔI. Rebound response to hyperpolarization-activated inward current, known as the voltage sag, was calculated as ΔV2/ΔV1, where ΔV1 is the difference between prestep baseline voltage (−70 mV) and the steady-state voltage during step current injection, and ΔV2 is the difference between current-step steady-state voltage and the minimum voltage during current injection. Voltage sag (Dembrow et al. 2010; Joshi et al. 2015) is reflective of subthreshold current mediated by hyperpolarization-activated cyclic nucleotide-gated (HCN) channels, also known as Ih. RMP was recorded in current-clamp mode with injected current set at 0 pA.

Action potential generation properties.

Properties related to action potential generation were also examined, including spike threshold, firing rate, and firing rate-input gain. The spike threshold is the most depolarized membrane voltage level before the action potential excursion. In this study, it was defined as the voltage at which the second derivative (i.e., curvature) of the voltage was at its local maximal value. Input current steps of varying current amplitudes were delivered, and the membrane voltage response to these current-step inputs was recorded. Action potentials during each current step were counted and used to construct a firing rate-input (FI) curve. The FI curve gain was quantified by calculating the slope of the curve between 200 and 400 pA.

Burst firing properties.

Finally, burst firing properties of neurons were examined, a hallmark of some classes of pyramidal neurons (Connors and Gutnick 1990). Bursting is the phenomenon characterized by a neuronal emission of a sequence (2 or more) of action potentials or spikes with short interspike intervals (ISIs), followed by a long ISI. In this study, neurons that exhibited burst firing had burst ISIs that were typically < 50 ms; additionally, these bursts were followed by intervals < 90 ms. An ISI threshold of 60 ms was chosen such that the burst size was quantified by the number of sequential ISIs < 60 ms, in response to a current step, followed by an interval > 60 ms.

Computational Analysis

All numerical analysis and statistics were performed in MATLAB (MathWorks, Natick, MA).

Principal components analysis.

In addition to determining differences in individual electrophysiological parameters among experimental groups, a secondary objective of this study was to determine if principal components analysis (PCA) leads to better distinction between vehicle-treated sham (Sham/Veh) and vehicle-treated TBI (TBI/Veh) and between TBI/Veh and TBI/inhibitor data sets. The aggregate electrophysiological parameter set comprised nine metrics relating to intrinsic and synaptic properties of neurons. PCA performs a linear transformation of multidimensional data sets to a new space where the variability of the original data set is mapped onto a lower dimensional space, often making it easier to discern systematic patterns (i.e., group clustering) in complex data sets (Jolliffe 2002). Principal components of increasing order explain diminishing proportions of the variance of the original data set, i.e., principal component 1 (PC1) explains the most variance, while the Nth principal component (PCN, in an N-dimensional space) explains the least. Principal component analysis was applied to transform the data into principal components PC1–PC9. The MATLAB program “pca” in the Statistics and Machine Learning Toolbox was used. Before the pca transformation was applied, each parameter (e.g., input resistance or resting membrane potential) was normalized by its standard deviation such that all normalized parameters had a standard deviation of 1, because of the nature and range of values from the different parameters.

Machine learning binary classification.

To explore if changes in multiple parameters in individual neurons are coordinated in some manner so as to make the collective physiological property set between Sham/Veh and TBI/Veh more distinguishable, a machine learning algorithm was employed to classifying multidimensional data sets. Machine learning algorithms are useful for prediction and decision-making from complex data sets. Specifically, a support vector machine (SVM) algorithm was used, which finds a line (or hyperplane) in a two-dimensional (or multidimensional) data space. This line is the optimal separation boundary between two data sets and is found using information from key data points near the overlap of the two data sets, also known as support vectors (Christianini and Shawe-Taylor 2000). The MATLAB program “fitcsvm” was used, with a linear kernel function to train the model on Sham/Veh and TBI/Veh electrophysiological data sets. An optimal linear classification boundary was obtained and used to classify and label neurons as either “Sham” or “TBI.” For the purpose of examining if the drug-treated neuron data sets were more similar to Sham/Veh or TBI/Veh, the trained SVM model was used to determine the proportion of neurons from drug-treated TBI rats classified as “Sham.”

Accuracy index.

The prior analysis compared how similar each electrophysiological data set from rats that received EC degradation enzyme inhibitors would be to Sham/Veh or TBI/Veh neurons. An alternative scenario was also considered where EC degradation inhibitors would prevent electrophysiological parameter sets from becoming “TBI”-like but potentially cause changes that make them different from “sham”-like properties in different directions. To assess this potential outcome, an accuracy index was used to quantify how distinguishable data sets between all groups were. This approach allows for optimal SVM boundary calculation for each comparison. The accuracy index (AI) is the number of correct positive classifications normalized by the size of the entire data set size. When data sets A and B are examined, the accuracy index is (% of A correctly classified + % of B correctly classified)/(total no. in A + total no. in B). Large values of the accuracy index (close to 1) indicate that the two data sets of comparison are dissimilar (more distinguishable), whereas low values (close to 0) indicate that the two data sets of comparison are very similar (less distinguishable).

Statistical Analysis

All data are means ± SE. Statistical differences were determined by one-way analysis of variance (ANOVA) as indicated in each figure legend. Tests for multiple comparisons using Tukey’s test were utilized when appropriate, as indicated in the text. Statistical significance was set at P < 0.05.

RESULTS

Synaptic Properties Following TBI and Administration of EC Degradation Inhibitors

Figure 2, A and B, shows representative traces of sEPSC currents and the average sEPSC event waveforms of these traces, respectively. A one-way ANOVA revealed a main effect of injury/treatment on sEPSC amplitude (F5,107 = 4.06, P = 0.002; Fig. 2C). Post hoc pairwise comparisons revealed significant differences in sEPSC amplitudes for the following comparisons: Sham/Veh (15.2 ± 1.2 pA) vs. TBI/Veh (27.5 ± 3.3 pA, P < 0.01), TBI/Veh vs. TBI/JZL184 (19.0 ± 2.0 pA, P < 0.05), TBI/Veh vs. TBI/JZL195 (17.9 ± 2.0 pA, P < 0.05), and TBI/Veh vs. TBI/URB597 (17.8 ± 1.6 pA, P = 0.05). A one-way ANOVA revealed no main effect of injury/treatment on sEPSC frequencies (F5,107 = 1.36, P = 0.243; Fig. 2D).

Fig. 2.

Fig. 2.

Open in a new tab

A and B: representative traces of spontaneous excitatory postsynaptic current (sEPSC) recordings (A) and average sEPSC event waveforms (B) from vehicle-treated sham (Sham/Veh; black) and vehicle-treated traumatic brain-injured (TBI/Veh; red) neurons. C and D: mean sEPSC amplitude (C) and frequency (D) of neuronal recordings from Sham/Veh, TBI/Veh, and inhibitor-treated TBI (TBI/JZL184, TBI/JZL195, TBI/URB597, and TBI/MJN110) groups. ANOVA revealed a main effect of injury/treatment on amplitude (F5,107 = 4.06, P = 0.002) and no main effect of injury/treatment on frequency (F5,107 = 1.36, P = 0.243). *P < 0.05; **P < 0.01.

Intrinsic Properties Following TBI and Administration of EC Degradation Inhibitors

Figure 3A shows a representative change in holding current used to calculate input resistance. A one-way ANOVA revealed a near-significant effect of injury/treatment on input resistance (F5,109 = 2.24, P = 0.055; Fig. 3B). Pairwise comparisons revealed significant differences between TBI/Veh (48.1 ± 3.7 MΩ) vs. TBI/JZL184 (62.6 ± 5.7 MΩ, P < 0.05) and TBI/Veh vs. TBI/JZL195 (62.9 ± 4.4 MΩ, P < 0.05). Figure 3C shows the characteristic rebound voltage excursion seen in some neurons following hyperpolarization caused by negative current steps, known as voltage sag. A one-way ANOVA revealed no effect of injury/treatment on sag (F5,109 = 1.16, P = 0.3333; Fig. 3D). A one-way ANOVA also revealed no effect of injury/treatment on RMP (F5,109 = 0.84, P = 0.5233; Fig. 3E).

Fig. 3.

Fig. 3.

Open in a new tab

A: representative step responses in voltage-clamp recording mode used to calculate input resistance. B: mean input resistance (Rinput) of neuronal recordings from vehicle-treated sham (Sham/Veh), vehicle-treated traumatic brain-injured (TBI/Veh), and inhibitor-treated TBI (TBI/JZL184, TBI/JZL195, TBI/URB597, and TBI/MJN110) groups. One-way ANOVA reveals a near-significant main effect of injury/treatment on Rinput (F5,109 = 2.24, P = 0.055). ΔI, change in holding current; ΔV, change in voltage. C: representative step responses in current-clamp recording mode used to calculate the sag fraction ΔV2/ΔV1, where ΔV1 is the difference between prestep baseline voltage (−70 mV) and steady-state voltage during step current injection, and ΔV2 is the difference between current-step steady-state voltage and the minimum voltage during current injection. D and E: mean sag fraction (D) and resting membrane potential (RMP; E) of neuronal recordings from Sham/Veh, TBI/Veh, TBI/JZL184, TBI/JZL195, TBI/URB597, and TBI/MJN110 groups. One-way ANOVA revealed no main effect of injury/treatment on sag (F5,109 = 1.16, P = 0.333) or RMP (F5,109 = 0.84, P = 0.523). *P < 0.05.

Action Potential Generation Properties Following TBI and Administration of EC Degradation Inhibitors

Figure 4A shows representative traces used to calculate spike threshold. A one-way ANOVA revealed a main effect of injury/treatment on spike threshold (F5,107 = 4.24, P = 0.002; Fig. 4B). Post hoc pairwise comparisons revealed significant differences in spike threshold for the following comparisons: TBI/JZL184 (−42.2 ± 1.3 mV) vs. TBI/URB597 (−34.0 ± 2.0 mV, P < 0.01) and TBI/URB597 vs. TBI/MJN110 (−39.6 ± 1.1 mV, P < 0.05). Figure 4C shows membrane voltage responses to current-step inputs. Figure 4D shows the FI curve constructed from the number of action potentials during each current step. The slope of this curve between 200 and 400 pA was used to calculate FI gain. A one-way ANOVA revealed a main effect of injury/treatment on FI gain (F5,107 = 2.75, P = 0.022; Fig. 4E). Post hoc pairwise comparisons revealed a significant difference in FI gain for the following comparisons: TBI/Veh (0.026 ± 0.008 Hz/pA) vs. TBI/JZL195 (0.049 ± 0.006 Hz/pA, P < 0.01), and TBI/URB597 (0.033 ± 0.006 Hz/pA) vs TBI/MJN110 (0.048 ± 0.05 Hz/pA, P < 0.05). A suprathreshold current level, 400 pA, elicited firing in most (97/115) sampled neurons in the entire data set. A one-way ANOVA revealed a near-significant effect of injury/treatment on firing rate (F5,109 = 2.29, P = 0.0503; Fig. 4F).

Fig. 4.

Fig. 4.

Open in a new tab

A: representative spike initiation (first spike above firing threshold) used to estimate spike threshold. B: average spike threshold of neuronal recordings from vehicle-treated sham (Sham/Veh), vehicle-treated traumatic brain-injured (TBI/Veh), and inhibitor-treated TBI (TBI/JZL184, TBI/JZL195, TBI/URB597, and TBI/MJN110) groups. One-way ANOVA revealed a main effect of injury/treatment on spike threshold (F5,107 = 4.24, P = 0.002). C: representative voltage responses (top traces) to current step inputs (bottom trace) in current-clamp recording mode used to estimate firing rate-input (FI) curves. D: representative average FI curve (from traces in C) used to extract the gain (blue) and firing level (gray). E and F: average FI gain (E) and firing rate (at 400 pA; F) of neuronal recordings from Sham/Veh, TBI/Veh, TBI/JZL184, TBI/JZL195, TBI/URB597, and TBI/MJN110 groups. One-way ANOVA revealed a main effect of injury/treatment on FI gain (F5,107 = 2.75, P = 0.022) and a near-significant effect of injury/treatment on firing rate at 400 pA (F5,107 = 2.29, P = 0.0503). *P < 0.05, **P < 0.01.

Burst Firing Properties Following TBI and Administration of EC Degradation Inhibitors

Figure 5A shows representative burst firing patterns that were used for quantifying burst size. A one-way ANOVA revealed a main effect of injury/treatment on burst size (F5,107 = 4.88, P = 0.0005; Fig. 5B). Post hoc pairwise comparisons revealed a significant difference in burst size for the following comparisons: Sham/Veh (1.4 ± 0.5) vs. TBI/MJN110 (4.9 ± 0.8, P < 0.01), TBI/Veh (1.8 ± 0.6) vs. TBI/MJN110 (P < 0.01), and TBI/URB597 (1.3 ± 0.2) vs. TBI/MJN110 (P < 0.05).

Fig. 5.

Fig. 5.

Open in a new tab

A: representative voltage responses recorded at 50–100 pA above current threshold used to quantify bursting. B: average burst size of neuronal recordings from vehicle-treated sham (Sham/Veh), vehicle-treated traumatic brain-injured (TBI/Veh), and inhibitor-treated TBI (TBI/JZL184, TBI/JZL195, TBI/URB597, and TBI/MJN110) groups. Interspike intervals (ISIs) below 60 ms were considered for membership in response bursts; sequences of sequential ISIs < 60 ms were used to calculate burst size. One-way ANOVA revealed a main effect of injury/treatment on burst size (F5,107 = 4.88, P = 0.0005). *P < 0.05, **P < 0.01.

Coordinated Changes in Multiple Properties Following TBI and Administration of EC Degradation Inhibitors

First, a smaller electrophysiological parameter set was used to more easily visualize potential separability of electrophysiological data sets and examine their effectiveness for use in SVM binary classification. sEPSC amplitude and FI gain (the parameter with the next largest relative difference between Sham/Veh and TBI/Veh neurons) were chosen to apply the SVM classification algorithm. The SVM algorithm was first trained on these data sets to generate a classification boundary resulting in largely separable data point clusters (Fig. 6A; note: these data have been normalized to have a standard deviation of 1). We quantified the percentage of neurons in each group that were classified as “Sham”: Sham/Veh and TBI/Veh neurons were classified as “Sham” 76.5% and 15.8% of the time, respectively (Fig. 6H). The corollary is that neurons from TBI/Veh rats were correctly classified as “TBI” 84.2% of the time. SVM binary classification was then applied to the data sets from TBI animals treated with EC degradation enzyme inhibitors (Fig. 6, B and C) using the classification boundary derived from the comparison of Sham/Veh and TBI/Veh neurons. The spread of data points from all TBI/drug data sets spanned both sides of the classification boundary. To identify how similar each data set was to the Sham/Veh data set, the percentage of data points on the Sham/Veh side of the classification boundary was calculated, resulting in these percentages: 62% for JZL184, 75% for JZL195, 72% for URB597, and 45% for MJN110 neurons, respectively (Fig. 6H).

Fig. 6.

Fig. 6.

Open in a new tab

A and E: electrophysiological parameters (Ephys; A) or principal components 1 and 2 (PC1 and PC2; E) from vehicle-treated sham (Sham/Veh) and vehicle-treated traumatic brain injury (TBI/Veh) data sets used to train a support machine vector (SVM) algorithm to determine binary classification boundary. Neuronal classification of TBI/JZL184, TBI/JZL195 (B and F), TBI/URB597, and TBI/MJN110 data sets (C and G) was performed using these boundaries. D: percentage of variance (Var) explained as a function of the PC number. H–K: percentage of neurons classified as “Sham/Veh” using either the first 2 elements of (H and I) or the full Ephys (J) or PC (K) data set. Amp, amplitude; a.u., arbitrary units; ePSC, excitatory postsynaptic current; FI, firing rate-input.

A secondary objective of this study was to examine if principal components analysis (PCA) leads to better distinction between Sham/Veh and TBI/Veh data sets than achieved using the raw electrophysiological parameters. The percentage of variance explained by each principal component is shown in Fig. 6D. Applying SVM to PC1 and PC2 of Sham/Veh and TBI/Veh resulted in a separation boundary that produced highly accurate data separation (Fig. 6E). Sham/Veh and TBI/Veh neurons were classified as “Sham” 88.2% and 73.7% of the time, respectively (Fig. 6I). The algorithm was then applied, with a classification boundary trained on Sham/Veh and TBI/Veh neurons, to the PC1 and PC2 data sets from TBI animals that received EC degradation enzyme inhibitors (Fig. 6, F and G). A similar classification profile (compared with the electrophysiological parameter space) was seen when PC1 and PC2 were analyzed from the EC degradation inhibitor data sets (Fig. 6I).

The same SVM binary classification algorithm applied to the full set of electrophysiological parameters (i.e., binary classification boundary was derived using: sEPSC amplitude, sEPSC frequency, input resistance, sag, RMP, spike threshold, FI magnitude and gain, and burst size) showed that the proportions of JZL184, JZL195, URB597, and MJN110 neurons classified as Sham/Veh neurons using the full electrophysiological data set are 52%, 80%, 72%, and 75%, respectively (Fig. 6J). Similar qualitative proportions are observed when the same analysis is applied to the PCA data set: JZL184, JZL195, URB597, and MJN110 neurons are classified as Sham/Veh neurons using the full electrophysiological data set 62%, 85%, 72%, and 75% of the time, respectively (Fig. 6K). The fact that the percentage of variance explained drops off slowly as a function of component number and the fact that there was no single dominant contribution of electrophysiological parameters to any one PC may be consistent with the PC data set not outperforming the original electrophysiological data set in binary classification.

Quantified Distinguishability of Parameter Sets Among Sham-, TBI-, and TBI-EC Degradation Inhibitor-Treated Groups

To quantify how distinguishable data sets between all groups were, accuracy indices were calculated. Figure 7 shows matrices of accuracy indices for pairwise data set comparisons in the four cases used in Fig. 6: sEPSC amplitude/FI gain, PC1/PC2, full electrophysiological data set, and full PC data set. Using sEPSC amplitude and FI gain (Fig. 7A), we found that Sham/Veh neurons are most distinguishable from TBI/Veh neurons (AI = 0.81) and least distinguishable from TBI/URB597 neurons (AI = 0.54). With these parameters, TBI/Veh neurons are also highly distinguishable from TBI/URB597 neurons (AI = 0.78). Using the full electrophysiological data set (Fig. 7B), we found that Sham/Veh neurons are the most distinguishable from TBI/Veh neurons (AI = 0.82) and the least distinguishable from TBI/URB597 neurons (AI = 0.72). With these parameters, TBI/Veh neurons are the most distinguishable from TBI/JZL195 neurons (AI = 0.83). Using PC1 and PC2, we found that Sham/Veh neurons are most distinguishable from TBI/Veh neurons (AI = 0.82) and least distinguishable from TBI/URB597 neurons (AI = 0.60). With these parameters, TBI/Veh neurons are also highly distinguishable from TBI/URB597 neurons (AI = 0.79). Using the full principal component data set (i.e., PC1–PC9; Fig. 7B), we found that Sham/Veh neurons are the most distinguishable from TBI/Veh neurons (AI = 0.81) and the least distinguishable from TBI/URB597 neurons (AI = 0.71). With these parameters, TBI/Veh neurons are also highly distinguishable from TBI/JZL195 neurons (AI = 0.79).

Fig. 7.

Fig. 7.

Open in a new tab

A: accuracy index (AI) matrix for the electrophysiological (Ephys) parameters: spontaneous excitatory postsynaptic current (sEPSC) amplitude and firing rate-input (FI) gain. B: AI matrix of the full Ephys data sets. C: AI matrix using principal components 1 and 2 (PC1 and PC2). D: AI matrix using the principal component (PC) data sets. With the use of full parameter data sets (Ephys, B; or PC, D), Sham/Veh neurons are most distinguishable from TBI/Veh neurons (AI = 0.82 or 0.81) and least distinguishable from TBI/URB597 (AI = 0.72 or 0.71), and TBI/Veh neurons are most distinguishable from TBI/JZL195 neurons (AI = 0.83 or 0.79).

DISCUSSION

This study tested the hypothesis that L5 SMC neurons from TBI rats would be characterized by altered synaptic and intrinsic neuronal properties 10–11 days postinjury, and that post-TBI EC degradation inhibition would attenuate these TBI-related alterations. TBI altered several electrophysiological parameters, including sEPSC amplitude, input resistance, FI gain, spike threshold, and burst size. TBI/Veh neurons exhibited increased sEPSC amplitude (but not frequency) compared with Sham/Veh neurons, suggesting increased synaptic excitability following TBI (Isaacson and Walmsley 1995). EC degradation inhibitors JZL184, JZL195, and URB97, but not MJN110, significantly attenuated the TBI-induced increase in sEPSC amplitude. Sham/Veh and TBI/Veh neurons exhibited trending differences in several additional parameters, in particular, decreases in input resistance, FI gain, and FI level at 400 pA. Post-TBI treatment with JZL184 and JZL195 increased input resistance compared with that of TBI/Veh neurons, and JZL195 increased FI gain compared with that of TBI/Veh neurons. Together, the findings suggest that these neurons are functionally altered by TBI; for example, signal-to-noise ratio of integration of incoming inputs may be decreased (smaller FI gain with increase background synaptic noise).

Our previous studies compared the acute effects of post-TBI MAGL inhibition via JZL184 with FAAH inhibition via URB97 on neuroinflammation, blood-brain barrier (BBB) disruption, and neurological and neurobehavioral impairment and found that JZL184 was more effective in attenuating these TBI-induced changes (Katz et al. 2015). We therefore only tested the effects of JZL184 on longer term TBI-induced excitability and negative affective behavior in subsequent studies (Fucich et al. 2019; Mayeux et al. 2017). Although the previously tested drug JZL184 is relatively selective for MAGL over FAAH in rats, mice, and humans, JZL184 has some cross-reactivity with FAAH at high or chronic doses and is far less potent against rat MAGL compared with mouse or human MAGL (Chang et al. 2012; Long et al. 2009a), and there is debate regarding whether JZL184 is BBB penetrant or exerts its neurobehavioral effects via peripheral mechanisms (e.g., Kerr et al. 2013; Woodhams et al. 2012; but see Oleson et al. 2012; Seillier et al. 2014). For these reasons, we tested the effects of MJN110, a highly selective and more potent MAGL inhibitor that effectively elevates 2-AG levels in rat brain after systemic administration at the dose used in the present study (e.g., Sticht et al. 2016), and we hypothesized that MJN110 would produce more robust effects than either JZL184 or URB597 in attenuating TBI-induced excitability. Contrary to our hypothesis, MJN110 was the only drug treatment that did not significantly attenuate the TBI-induced increase in sEPSC amplitude, and in fact, neurons from TBI/MJN110 animals showed significantly higher burst size than neurons from Sham/Veh, TBI/Veh, and TBI/URB597 rats. It is not clear what caused this higher burst size, but it is possible that that MJN110 causes an overshoot past baseline and that this overshoot leads to increased bursting. Additionally, in vivo post-TBI URB597 treatment attenuated TBI-induced increases in sEPSC amplitude and produced significantly higher spike thresholds compared with those in neurons from TBI animals treated with the MAGL inhibitors JZL184 and MJN110, suggesting potential beneficial effects of FAAH inhibition on TBI-induced excitability, although URB597 did little to attenuate other electrophysiological changes produced by TBI such as decreased FI gain. Interestingly, one in vivo study reported that L5 neurons exhibited increased spontaneous activity (but diminished sensory-evoked responses) after TBI (Johnstone et al. 2014). This in vivo activity phenotype is consistent with our major findings, specifically, increased background activity could arise from increased spontaneous excitatory synaptic inputs, whereas decreased stimulus response robustness is consistent with the decreased firing gain we observed.

The changes in L5 cortical synaptic excitability in this and previous studies may have a functional role in altering post-TBI excitability in subcortical brain regions that receive inputs from these pyramidal cells. Because EC degradation inhibition ameliorates increased synaptic excitability in L5 cortical pyramidal neurons at the site of injury, as seen in this study and our previous work (Mayeux et al. 2017), and because it also attenuates acute and long-term behavioral impairment (Fucich et al. 2019; Katz et al. 2015; Mayeux et al. 2017), it is possible that attenuations in post-TBI behavioral impairment by EC degradation inhibition are mediated by rescue of TBI-induced dysfunction of SMC circuit function. Alternative mechanisms by which inhibiting EC degradation may attenuate behavioral changes include suppression of inflammation and decreased BBB leak, and these effects may contribute to EC degradation inhibitor effects on post-TBI neuronal excitability (Shlosberg et al. 2010). Further investigation is needed to understand the mechanisms by which treatment with these inhibitors improve the neural and behavioral sequelae following TBI.

The present study also tested the hypothesis that with the use of a machine learning algorithm, L5 SMC neurons could be accurately classified as either Sham-like or TBI-like neurons based on either raw electrophysiological parameters measured or data transformed using PCA, and that the algorithm would more often classify TBI/drug neurons as Sham-like rather than TBI-like neurons. Principal components have long been used to cluster and classify extracellular waveform recordings for cell-type identification in in vivo electrophysiology (Lewicki 1998) and can be used to identify distinct functionally correlated neuronal assemblies (Lopes-dos-Santos et al. 2011). Our results showed that the proportion of Sham/Veh neurons correctly identified is higher when principal components rather than raw electrophysiological parameters are used. Regardless, in both coordinate systems, JZL195 resulted in the highest proportion of neurons classified as Sham/Veh neurons compared with other EC degradation enzyme inhibitors (80% TBI/JZL195 neurons classified as Sham/Veh using the full set of electrophysiological parameters; 85% using the PCA data set). Accuracy indices for the analyzed data sets also revealed that Sham/Veh neurons were least distinguishable from TBI/URB597 neurons and that TBI/Veh neurons were consistently the most distinguishable from TBI/JZL195 neurons, providing further support that post-TBI JZL195 treatment results in electrophysiological profiles that are least TBI-like, thus putatively offering the most neuroprotection, with regard to physiological properties, from the effects of TBI.

Collectively, these electrophysiological and computational data suggest that dual inhibition of both MAGL and FAAH may be more beneficial for protecting against TBI-induced neuronal functional alterations at the site of injury than targeting MAGL alone, and they also highlight an important role for FAAH inhibition in post-TBI plasticity at site of injury.

GRANTS

Support for this study was provided by National Institutes of Health Grants R01 AA025792, T32 AA007577, F32 AA026779, and F30 AA026468.

DISCLOSURES

N.W.G. holds shares in Glauser Life Sciences, Inc. S.E. is on the Scientific Advisory Board of Halyard Health. These activities have no relation to any of the work presented in this paper. None of the other authors has any conflicts of interest, financial or otherwise, to disclose.

AUTHOR CONTRIBUTIONS

E.A.F., Z.F.S., S.E., N.W.G., P.E.M., and J.W.M. conceived and designed research; E.A.F., Z.F.S., H.L.C., and J.W.M. performed experiments; E.A.F. and J.W.M. analyzed data; E.A.F. and J.W.M. interpreted results of experiments; J.W.M. prepared figures; E.A.F. and J.W.M. drafted manuscript; E.A.F., Z.F.S., H.L.C., S.E., N.W.G., P.E.M., and J.W.M. edited and revised manuscript; E.A.F., Z.F.S., H.L.C., S.E., N.W.G., P.E.M., and J.W.M. approved final version of manuscript.

ACKNOWLEDGMENTS

Present address for H. L. Cancienne: San Diego State University Research Foundation, 5250 Campanile Dr., San Diego, CA 92182.

REFERENCES

  1. Allitt BJ, Iva P, Yan EB, Rajan R. Hypo-excitation across all cortical laminae defines intermediate stages of cortical neuronal dysfunction in diffuse traumatic brain injury. Neuroscience 334: 290–308, 2016a. doi: 10.1016/j.neuroscience.2016.08.018. [DOI] [PubMed] [Google Scholar]
  2. Allitt BJ, Johnstone VP, Richards K, Yan EB, Rajan R. Progesterone exacerbates short-term effects of traumatic brain injury on supragranular responses in sensory cortex and overexcites infragranular responses in the long term. J Neurotrauma 33: 375–389, 2016b. doi: 10.1089/neu.2015.3946. [DOI] [PubMed] [Google Scholar]
  3. Alwis DS, Yan EB, Morganti-Kossmann MC, Rajan R. Sensory cortex underpinnings of traumatic brain injury deficits. PLoS One 7: e52169, 2012. doi: 10.1371/journal.pone.0052169. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Brettschneider J, Toledo JB, Van Deerlin VM, Elman L, McCluskey L, Lee VM, Trojanowski JQ. Microglial activation correlates with disease progression and upper motor neuron clinical symptoms in amyotrophic lateral sclerosis. PLoS One 7: e39216, 2012. doi: 10.1371/journal.pone.0039216. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Bryant RA, O’Donnell ML, Creamer M, McFarlane AC, Clark CR, Silove D. The psychiatric sequelae of traumatic injury. Am J Psychiatry 167: 312–320, 2010. doi: 10.1176/appi.ajp.2009.09050617. [DOI] [PubMed] [Google Scholar]
  6. Carron SF, Alwis DS, Rajan R. Traumatic brain injury and neuronal functionality changes in sensory cortex. Front Syst Neurosci 10: 47, 2016. doi: 10.3389/fnsys.2016.00047. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Chang JW, Niphakis MJ, Lum KM, Cognetta AB 3rd, Wang C, Matthews ML, Niessen S, Buczynski MW, Parsons LH, Cravatt BF. Highly selective inhibitors of monoacylglycerol lipase bearing a reactive group that is bioisosteric with endocannabinoid substrates. Chem Biol 19: 579–588, 2012. doi: 10.1016/j.chembiol.2012.03.009. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Christianini N, Shawe-Taylor J. An Introduction to Support Vector Machines and Other Kernel-Based Learning Methods. Cambridge, UK: Cambridge University Press, 2000. [Google Scholar]
  9. Connors BW, Gutnick MJ. Intrinsic firing patterns of diverse neocortical neurons. Trends Neurosci 13: 99–104, 1990. doi: 10.1016/0166-2236(90)90185-D. [DOI] [PubMed] [Google Scholar]
  10. Cravatt BF, Giang DK, Mayfield SP, Boger DL, Lerner RA, Gilula NB. Molecular characterization of an enzyme that degrades neuromodulatory fatty-acid amides. Nature 384: 83–87, 1996. doi: 10.1038/384083a0. [DOI] [PubMed] [Google Scholar]
  11. Dembrow NC, Chitwood RA, Johnston D. Projection-specific neuromodulation of medial prefrontal cortex neurons. J Neurosci 30: 16922–16937, 2010. doi: 10.1523/JNEUROSCI.3644-10.2010. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Ding MC, Wang Q, Lo EH, Stanley GB. Cortical excitation and inhibition following focal traumatic brain injury. J Neurosci 31: 14085–14094, 2011. doi: 10.1523/JNEUROSCI.3572-11.2011. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Dinh TP, Freund TF, Piomelli D. A role for monoglyceride lipase in 2-arachidonoylglycerol inactivation. Chem Phys Lipids 121: 149–158, 2002. doi: 10.1016/S0009-3084(02)00150-0. [DOI] [PubMed] [Google Scholar]
  14. Dixon CE, Lyeth BG, Povlishock JT, Findling RL, Hamm RJ, Marmarou A, Young HF, Hayes RL. A fluid percussion model of experimental brain injury in the rat. J Neurosurg 67: 110–119, 1987. doi: 10.3171/jns.1987.67.1.0110. [DOI] [PubMed] [Google Scholar]
  15. Eikelenboom P, van Exel E, Hoozemans JJ, Veerhuis R, Rozemuller AJ, van Gool WA. Neuroinflammation – an early event in both the history and pathogenesis of Alzheimer’s disease. Neurodegener Dis 7: 38–41, 2010. doi: 10.1159/000283480. [DOI] [PubMed] [Google Scholar]
  16. Fucich EA, Mayeux JP, McGinn MA, Gilpin NW, Edwards S, Molina PE. A novel role for the endocannabinoid system in ameliorating motivation for alcohol drinking and negative behavioral affect after traumatic brain injury in rats. J Neurotrauma 36: 1847–1855, 2019. doi: 10.1089/neu.2018.5854. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Greer JE, Povlishock JT, Jacobs KM. Electrophysiological abnormalities in both axotomized and nonaxotomized pyramidal neurons following mild traumatic brain injury. J Neurosci 32: 6682–6687, 2012. doi: 10.1523/JNEUROSCI.0881-12.2012. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Hånell A, Greer JE, Jacobs KM. Increased network excitability due to altered synaptic inputs to neocortical layer V intact and axotomized pyramidal neurons after mild traumatic brain injury. J Neurotrauma 32: 1590–1598, 2015. doi: 10.1089/neu.2014.3592. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Hibbard MR, Uysal S, Kepler K, Bogdany J, Silver J. Axis I psychopathology in individuals with traumatic brain injury. J Head Trauma Rehabil 13: 24–39, 1998. doi: 10.1097/00001199-199808000-00003. [DOI] [PubMed] [Google Scholar]
  20. Isaacson JS, Walmsley B. Counting quanta: direct measurements of transmitter release at a central synapse. Neuron 15: 875–884, 1995. doi: 10.1016/0896-6273(95)90178-7. [DOI] [PubMed] [Google Scholar]
  21. Johnstone VP, Yan EB, Alwis DS, Rajan R. Cortical hypoexcitation defines neuronal responses in the immediate aftermath of traumatic brain injury. PLoS One 8: e63454, 2013. doi: 10.1371/journal.pone.0063454. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Johnstone VP, Shultz SR, Yan EB, O’Brien TJ, Rajan R. The acute phase of mild traumatic brain injury is characterized by a distance-dependent neuronal hypoactivity. J Neurotrauma 31: 1881–1895, 2014. doi: 10.1089/neu.2014.3343. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Jolliffe IT. Principal Component Analysis. New York: Springer, 2002. [Google Scholar]
  24. Joshi A, Middleton JW, Anderson CT, Borges K, Suter BA, Shepherd GM, Tzounopoulos T. Cell-specific activity-dependent fractionation of layer 2/3→5B excitatory signaling in mouse auditory cortex. J Neurosci 35: 3112–3123, 2015. doi: 10.1523/JNEUROSCI.0836-14.2015. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Kathuria S, Gaetani S, Fegley D, Valiño F, Duranti A, Tontini A, Mor M, Tarzia G, La Rana G, Calignano A, Giustino A, Tattoli M, Palmery M, Cuomo V, Piomelli D. Modulation of anxiety through blockade of anandamide hydrolysis. Nat Med 9: 76–81, 2003. doi: 10.1038/nm803. [DOI] [PubMed] [Google Scholar]
  26. Katz PS, Sulzer JK, Impastato RA, Teng SX, Rogers EK, Molina PE. Endocannabinoid degradation inhibition improves neurobehavioral function, blood-brain barrier integrity, and neuroinflammation following mild traumatic brain injury. J Neurotrauma 32: 297–306, 2015. doi: 10.1089/neu.2014.3508. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Kerr DM, Harhen B, Okine BN, Egan LJ, Finn DP, Roche M. The monoacylglycerol lipase inhibitor JZL184 attenuates LPS-induced increases in cytokine expression in the rat frontal cortex and plasma: differential mechanisms of action. Br J Pharmacol 169: 808–819, 2013. doi: 10.1111/j.1476-5381.2012.02237.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Korn A, Golan H, Melamed I, Pascual-Marqui R, Friedman A. Focal cortical dysfunction and blood-brain barrier disruption in patients with postconcussion syndrome. J Clin Neurophysiol 22: 1–9, 2005. doi: 10.1097/01.WNP.0000150973.24324.A7. [DOI] [PubMed] [Google Scholar]
  29. Kumar A, Loane DJ. Neuroinflammation after traumatic brain injury: opportunities for therapeutic intervention. Brain Behav Immun 26: 1191–1201, 2012. doi: 10.1016/j.bbi.2012.06.008. [DOI] [PubMed] [Google Scholar]
  30. Lewicki MS. A review of methods for spike sorting: the detection and classification of neural action potentials. Network 9: R53–R78, 1998. doi: 10.1088/0954-898X_9_4_001. [DOI] [PubMed] [Google Scholar]
  31. Long JZ, Nomura DK, Cravatt BF. Characterization of monoacylglycerol lipase inhibition reveals differences in central and peripheral endocannabinoid metabolism. Chem Biol 16: 744–753, 2009a. doi: 10.1016/j.chembiol.2009.05.009. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Long JZ, Nomura DK, Vann RE, Walentiny DM, Booker L, Jin X, Burston JJ, Sim-Selley LJ, Lichtman AH, Wiley JL, Cravatt BF. Dual blockade of FAAH and MAGL identifies behavioral processes regulated by endocannabinoid crosstalk in vivo. Proc Natl Acad Sci USA 106: 20270–20275, 2009b. doi: 10.1073/pnas.0909411106. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Lopes-dos-Santos V, Conde-Ocazionez S, Nicolelis MAL, Ribeiro ST, Tort ABL. Neuronal assembly detection and cell membership specification by principal component analysis. PLoS One 6: e20996, 2011. doi: 10.1371/journal.pone.0020996. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Mayeux J, Katz P, Edwards S, Middleton JW, Molina PE. Inhibition of endocannabinoid degradation improves outcomes from mild traumatic brain injury: a mechanistic role for synaptic hyperexcitability. J Neurotrauma 34: 436–443, 2017. doi: 10.1089/neu.2016.4452. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Nampiaparampil DE. Prevalence of chronic pain after traumatic brain injury: a systematic review. JAMA 300: 711–719, 2008. doi: 10.1001/jama.300.6.711. [DOI] [PubMed] [Google Scholar]
  36. Niphakis MJ, Cognetta AB 3rd, Chang JW, Buczynski MW, Parsons LH, Byrne F, Burston JJ, Chapman V, Cravatt BF. Evaluation of NHS carbamates as a potent and selective class of endocannabinoid hydrolase inhibitors. ACS Chem Neurosci 4: 1322–1332, 2013. doi: 10.1021/cn400116z. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Ofek H, Defrin R. The characteristics of chronic central pain after traumatic brain injury. Pain 131: 330–340, 2007. doi: 10.1016/j.pain.2007.06.015. [DOI] [PubMed] [Google Scholar]
  38. Oleson EB, Beckert MV, Morra JT, Lansink CS, Cachope R, Abdullah RA, Loriaux AL, Schetters D, Pattij T, Roitman MF, Lichtman AH, Cheer JF. Endocannabinoids shape accumbal encoding of cue-motivated behavior via CB1 receptor activation in the ventral tegmentum. Neuron 73: 360–373, 2012. doi: 10.1016/j.neuron.2011.11.018. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Ooigawa H, Nawashiro H, Fukui S, Otani N, Osumi A, Toyooka T, Shima K. The fate of Nissl-stained dark neurons following traumatic brain injury in rats: difference between neocortex and hippocampus regarding survival rate. Acta Neuropathol 112: 471–481, 2006. doi: 10.1007/s00401-006-0108-2. [DOI] [PubMed] [Google Scholar]
  40. Paxinos G, Watson CR. The Rat Brain in Stereotaxic Coordinates (6th ed.). Boston, MA: Academic, 2007. [DOI] [PubMed] [Google Scholar]
  41. Ramlackhansingh AF, Brooks DJ, Greenwood RJ, Bose SK, Turkheimer FE, Kinnunen KM, Gentleman S, Heckemann RA, Gunanayagam K, Gelosa G, Sharp DJ. Inflammation after trauma: microglial activation and traumatic brain injury. Ann Neurol 70: 374–383, 2011. doi: 10.1002/ana.22455. [DOI] [PubMed] [Google Scholar]
  42. Roche RA, Dockree PM, Garavan H, Foxe JJ, Robertson IH, O’Mara SM. EEG alpha power changes reflect response inhibition deficits after traumatic brain injury (TBI) in humans. Neurosci Lett 362: 1–5, 2004. doi: 10.1016/j.neulet.2003.11.064. [DOI] [PubMed] [Google Scholar]
  43. Seillier A, Dominguez Aguilar D, Giuffrida A. The dual FAAH/MAGL inhibitor JZL195 has enhanced effects on endocannabinoid transmission and motor behavior in rats as compared to those of the MAGL inhibitor JZL184. Pharmacol Biochem Behav 124: 153–159, 2014. doi: 10.1016/j.pbb.2014.05.022. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Shlosberg D, Benifla M, Kaufer D, Friedman A. Blood-brain barrier breakdown as a therapeutic target in traumatic brain injury. Nat Rev Neurol 6: 393–403, 2010. doi: 10.1038/nrneurol.2010.74. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Sticht MA, Limebeer CL, Rafla BR, Abdullah RA, Poklis JL, Ho W, Niphakis MJ, Cravatt BF, Sharkey KA, Lichtman AH, Parker LA. Endocannabinoid regulation of nausea is mediated by 2-arachidonoylglycerol (2-AG) in the rat visceral insular cortex. Neuropharmacology 102: 92–102, 2016. doi: 10.1016/j.neuropharm.2015.10.039. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Teng SX, Molina PE. Acute alcohol intoxication prolongs neuroinflammation without exacerbating neurobehavioral dysfunction following mild traumatic brain injury. J Neurotrauma 31: 378–386, 2014. doi: 10.1089/neu.2013.3093. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Ting JT, Daigle TL, Chen Q, Feng G. Acute brain slice methods for adult and aging animals: application of targeted patch clamp analysis and optogenetics. In: Patch-Clamp Methods and Protocols, edited by Martina M, Taverna S. New York: Humana, 2014, vol. 1183, p. 221–242. Methods in Molecular Biology. doi: 10.1007/978-1-4939-1096-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Woodhams SG, Wong A, Barrett DA, Bennett AJ, Chapman V, Alexander SP. Spinal administration of the monoacylglycerol lipase inhibitor JZL184 produces robust inhibitory effects on nociceptive processing and the development of central sensitization in the rat. Br J Pharmacol 167: 1609–1619, 2012. doi: 10.1111/j.1476-5381.2012.02179.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Xu JY, Chen C. Endocannabinoids in synaptic plasticity and neuroprotection. Neuroscientist 21: 152–168, 2015. doi: 10.1177/1073858414524632. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Yang L, Afroz S, Michelson HB, Goodman JH, Valsamis HA, Ling DS. Spontaneous epileptiform activity in rat neocortex after controlled cortical impact injury. J Neurotrauma 27: 1541–1548, 2010. doi: 10.1089/neu.2009.1244. [DOI] [PubMed] [Google Scholar]

Articles from Journal of Neurophysiology are provided here courtesy of American Physiological Society

RESOURCES