Inhibition of Monoacylglycerol Lipase Prevents Chronic Traumatic Encephalopathy-like Neuropathology in a Mouse Model of Repetitive Mild Closed Head Injury

Abstract

Emerging evidence suggests that the risk of developing chronic traumatic encephalopathy (CTE), a progressive neurodegenerative disease, is significantly increased in military personnel and contact sports players who have been exposed to repetitive trauma brain injury (TBI). Unfortunately there are no effective medications currently available for prevention and treatment of CTE. Here we demonstrate that inhibition of monoacylglycerol lipase (MAGL), the key enzyme that metabolizes the endocannabinoid 2-arachidonoylglycerol (2-AG) in the brain, significantly reduced CTE-like neuropathologic changes in a mouse model of repetitive mild closed head injury (rmCHI). Inhibition of 2-AG metabolism promoted neurologic recovery following rmCHI and reduced proinflammatory cytokines, astroglial reactivity, expression of amyloid precursor protein and the enzymes that make Aβ, as well as formation of Aβ. Importantly, neurodegeneration, TDP-43 protein aggregation, and tau phosphorylation, which are the neuropathologic hallmarks of CTE, were significantly suppressed by MAGL inactivation. Furthermore, alterations in expression of glutamate receptor subunits and impairments in basal synaptic transmission, long-term synaptic plasticity, and spatial learning and memory were recovered by inhibition of 2-AG metabolism in animals exposed to rmCHI. Our results suggest that MAGL inhibition, which boosts 2-AG and reduces 2-AG metabolites prostaglandins in the brain, may lead to a new therapy for CTE.

Keywords

Alzheimer's disease 2-arachidonoylglycerol chronic traumatic encephalopathy monoacylglycerol lipase neurodegeneration trauma brain injury

Introduction

Chronic traumatic encephalopathy (CTE) is a long-term progressive neurodegenerative disease characterized by persistent neuroinflammation, neurodegeneration, β-amyloid (Aβ) formation, phosphorylated tauopathy, transactive response DNA-binding protein 43 (TDP-43) proteinopathy, myelinated axonopathy, marked brain atrophy, memory loss, and dementia.^{1, 2, 3, 4, 5, 6, 7, 8, 9, 10} Latest studies reveal that the risk of developing CTE is significantly increased in athletes and military personnel who have been exposed to repetitive mild traumatic brain injury (rmTBI).^{7, 11} The features of the clinical symptoms and neuropathologic changes in CTE are similar to those seen in other neurodegenerative diseases, including Alzheimer's disease (AD).^{2, 6, 8, 10} Unfortunately there are no effective medications currently available for prevention and treatment of CTE.^{1, 6} Therefore, in the interest of public health, there is a great need to discover and develop novel and efficacious therapeutic interventions for preventing development of CTE or halting disease progression.

The acute brain damage after traumatic brain injury TBI results from primary injury, which is the result of the external mechanical force leading to contusion, laceration, and diffuse injuries, and from secondary injury immediately followed by primary injury, which is associated with a complex cascade of molecular, cellular and immune responses, resulting in neuroinflammation, excitotoxicity, oxidative stress, disruption of calcium homeostasis, mitochondrial dysfunction, neuronal injury, and death.^{4, 12} In the responses to secondary damage, the inflammatory response associated with other processes likely plays a key role in causing neuropathology following TBI. Inflammation has been recognized to be one of the important hallmarks in TBI. Proinflammatory markers such as cytokines interleukin (IL)-1β, IL-6, and tumor necrosis factor alpha (TNFα), and chemokines released from reactivated astroglial cells and infiltrated leukocytes in the brain and cerebrospinal fluid are robustly elevated after TBI.^{12, 13} The extent of neuroinflammatory response in TBI seems to be closely correlated with the outcome.¹³ Although the primary injury immediately following TBI is not preventable, the secondary injury provides a window for interventions to prevent further brain damage after the primary injury. Appropriate and timely intervention during this critical window following the primary injury after TBI may significantly reduce secondary brain damage and eventually prevent occurrence of CTE. Thus, resolving neuroinflammation immediately following TBI may be the key to prevent further brain damage, neuropathologic changes and occurrence of CTE.

Endocannabinoids are endogenous lipid mediators involved in a variety of physiologic, pharmacologic, and pathologic processes. 2-Arachidonoylglycerol (2-AG), the most abundant endogenous cannabinoid and a full agonist for both CB1/2 receptors, is primarily produced from diacylglycerol by diacylglycerol lipase (α and β) and largely hydrolyzed by monoacylglycerol lipase (MAGL) in the brain.^{14, 15} Earlier studies show that the levels of 2-AG are significantly increased in the brain following closed head injury (CHI) and that administration of 2-AG attenuates TBI-induced neuropathology,^{16, 17, 18} indicating that 2-AG likely is an endogenous signaling mediator ‘on demand’ that maintains brain homeostasis against harmful insults. Indeed, 2-AG possesses significant anti-inflammatory and neuroprotective properties in response to proinflammatory, excitotoxic, and mechanic insults.^{16, 17, 18, 19, 20, 21} The anti-inflammatory and neuroprotective behaviors of 2-AG suggest that manipulation of 2-AG signaling may provide novel therapeutic interventions for the prevention or reduction of TBI-induced neuropathology.²² As 2-AG is a very unstable fatty acid and is rapidly metabolized by the enzymes upon its release, inhibition of 2-AG metabolism by MAGL inactivation will significantly elevate endogenous 2-AG in the brain.^{14, 15} We demonstrate in this report that inhibition of MAGL by JZL184, which boosts 2-AG levels and decreases its metabolites arachidonic acid (AA) and prostaglandins, significantly reduced CTE-like neuropathologic changes in a mouse model of repetitive mild closed head injury (rmCHI). Inhibition of 2-AG metabolism reduced neuroinflammation, neurodegeneration, TDP-43 protein aggregation and tau phosphorylation, and improved synaptic and cognitive function in animals exposed to rmCHI. Our results suggest that MAGL may be a new therapeutic target for preventing CTE or halting its progression following repetitive mild TBI.

Materials and methods

Animals

Male C57BL/6 (Jackson Laboratory, Bar Harbor, ME, USA) at ages of 6 to 10 weeks were used in the present study. All animal studies were performed in compliance with the US Department of Health and Human Services Guide for the Care and Use of Laboratory Animals, and the care and use of the animals reported in this study were approved by the Institutional Animal Care and Use Committee of Louisiana State University Health Sciences Center. Mice were intraperitoneally injected with vehicle or 4-nitrophenyl-4-[bis (1,3-benzodioxol-5-yl)(hydroxy)methyl]piperidine-1-carboxylate (JZL184, 10 mg/kg) 30 minutes after each TBI and then once a day for 4 consecutive days (total 7 injections). JZL184 was prepared and dissolved in a vehicle containing Tween 80 (10%), dimethylsulfoxide (10%) and saline (80%) as described previously.¹⁹

Repetitively Mild Closed Head Injury

A mouse model of rmCHI was used as described previously with modification.^{5, 23, 24} Repetitive brain injuries were induced using an electromagnetic controlled stereotaxic impact device (Impact One Stereotaxic Impactor, Leica Biosystem, Buffalo Grove, IL, USA). Mice were placed in a stereotaxic frame after intraperitoneal anesthesia with Avertin (200 mg/kg body weight). The skull was exposed by a midline skin incision. A 3-mm blunt metal impactor tip was positioned at 1.8 mm caudal to bregma and 2.0 mm left of midline. The injury was triggered by the electromagnetic device driving the tip to the exposed skull at a strike velocity of 3.0 m/second and depth 2.2 mm with a dwell time of 100 milliseconds. After impact, the skin was sutured and the mice were allowed to recover from anesthesia on a warming pad and then returned to their home cages. A second and third identical closed-skull TBI procedure was performed on days 2 and 3 after the original injury. For sham injuries, the same procedures and anesthesia were performed except that no hit was delivered. Animals with skull fractures after impact (<2% of incidence of skull fracturing) were excluded from the study.

Hippocampal Slice Preparation

Hippocampal slices were prepared from mice as described previously.^{19, 25} Briefly, after decapitation, brains were rapidly removed and placed in cold oxygenated (95% O₂, 5% CO₂) artificial cerebrospinal fluid (ACSF) containing (mM) 125.0 NaCl, 2.5 KCl, 1.0 MgCl₂, 25.0 NaHCO₃, 1.25 NaH₂PO₄, 2.0 CaCl₂, 25.0 glucose, 3 pyruvic acid and 1 ascorbic acid. Slices were cut at a thickness of 350 to 400 μm and transferred to a holding chamber in an incubator containing ACSF at 36 °C for 0.5 to 1 hour, and maintained in an incubator containing oxygenated ACSF at room temperature (∼22 to 24 °C) for >1.5 hours before recordings. Slices were then transferred to a recording chamber, where they were continuously perfused with 95% O₂ and 5% CO₂-saturated standard ACSF at ∼32 to 34 °C.

Electrophysiological Recordings

Field excitatory postsynaptic potential recordings at hippocampal Schaffer-collateral synapses in response to stimuli at a frequency of 0.05 Hz were made using an Axoclamp-2B patch-clamp amplifier (Molecular Devices, Sunnyvale, CA, USA) in bridge mode. Recording pipettes were pulled from borosilicate glass with a micropipette puller (Sutter Instrument, Novato, CA, USA), filled with artificial ACSF (2 MΩ–4 MΩ). Long-term potentiation (LTP) at CA3–CA1 synapses was induced by high-frequency stimulation, as described previously.^{19, 25} The input—output function was tested before recording LTP, and the baseline stimulation strength was set to provide field excitatory postsynaptic potential with an amplitude of ∼30% from the subthreshold maximum derived from the input—output function.

Neurological Severity Score

The neurological severity score (NSS), ratings from 0 (normal) to 10 (severely impaired) as described previously,²⁶ was assessed to determine motor and reflex function, flexion of limbs, and resistance to position changes 2 hours after each impact and 5 days after three impacts.

Western Blots

Western blot assay was conducted to determine expressions of glutamate receptor subunits GluA1, GluA2, GluN1, GluN2A, and GluN2B, amyloid precursor protein (APP), β-site amyloid precursor protein cleaving enzyme 1 (BACE1), A Disintegrin and metalloproteinase domain-containing protein 10 (ADAM-10), Nicastrin (NCT), TDP-43, tau and p-tau, p-GSK3β, cdk5, p35, IL-1β, IL-6, and TNFα in cortical and hippocampal tissues from mice treated with vehicle or JZL184, as described previously.^{19, 20, 25} Cortical and hippocampal tissues were extracted and immediately homogenized in RIPA lysis buffer and protease inhibitors, and incubated on ice for 30 minutes, then centrifuged for 10 minutes at 10,000 r.p.m. at 4 °C. Supernatants were fractionated on 4 to 15% SDS-PAGE gels (Bio-Rad, Hercules, CA, USA) and transferred onto polyvinylidenedifluoride membranes (Bio-Rad). The membrane was incubated with specific antibodies at 4 °C overnight. The blots were washed and incubated with a secondary antibody (goat anti-rabbit 1:2,000, Life Tech, Grand Island, NY, USA) at room temperature for 1 hour. Proteins were visualized by enhanced chemiluminescence (GE Healthcare Biosciences, Pittsburgh, PA, USA). The densities of specific bands were quantified by densitometry using FUJIFILM Multi Gauge software (version 3.0). Band densities were normalized to the total amount of protein loaded in each well as determined by mouse anti-β-actin (1:4,000, Sigma, St Louis, MO, USA), as described previously.^{19, 20, 25}

Reverse Transcription and Real-Time PCR

Total RNA was prepared from harvested tissue with the RNeasy Mini Kit (Qiagen) and treated with RNase-free DNase (Qiagen) according to the manufacturer's instructions. The RNA concentration was measured by a spectrophotometer (DU 640; Beckman, Brea, CA, USA). RNA integrity was verified by electrophoresis in a 1% agarose gel.

The iScript cDNA synthesis kit (Bio-Rad) was used for the reverse transcription reaction. We used 1 μg total RNA, with 4 μL 5 × iscript reaction mix and 1 μL iscript reverse transcriptase. The total volume was 20 μL. Samples were incubated for 5 minutes at 25 °C. All samples were then heated to 42 °C for 30 minutes, and reactions were stopped by heating to 85 °C for 5 minutes. Real-time reverse transcriptase-PCR-specific primers for IL-1β, IL-6, TNFα, vimentin (Vim), and glyceraldehyde-3-phosphate-dehydrogenase were selected using Beacon Designer Software (Bio-Rad) and synthesized by IDT (Coralville, IA, USA). They are listed as follows (name: forward primer, reverse primer (amplicon size), genebank number): IL-1β: 5′-TGGAGAGTGTGGATCCCAAGCAAT-3′, 5′-TGTCCTGACCACTGTTGTTTCCCA-3′ (180 bp), NM_008361.3; IL-6: 5′-TCTCTGGGAAATCGTGGAAATG-3′, 5′-ACTCCAGGTAGCTATGGTACTC-3′ (203 bp), NM_031168.1; TNFα: 5′-GTCTACTGAACTTCGGGGTGA-3′, 5′-CACTTGGTGGTTTGCTACGAC-3′, (142 bp), NM_013693; Vim: 5′-AGATGGCTCGTCACCTTCGTGAAT-3′, 5′-TTGAGTGGGTGTCAACCAGAGGAA-3′ (192 bp), NM_011701; glyceraldehyde-3-phosphate-dehydrogenase: 5′-ACCACAGTCCATGCCATCAC-3′, 5′-ACCTTGCCCACAGCCTTG-3′ (134 bp), M32599. The PCR amplification of each product was further assessed using 10-fold dilutions of mouse brain cDNA library as a template and was found to be linear over five orders of magnitude and at >95% efficiency. All the PCR products were verified by sequencing. The reactions were set up in duplicate in total volumes of 25 μL containing 12.5 μL 2 × iQSYBR green Supermix (Bio-Rad) and 5 μL template (1:10 dilution from RT product) with a final concentration of 400 nM of the primer. The PCR cycle was as follows: 95 °C/3 minutes, 45 cycles of 95 °C/30 seconds, 58 °C/45 seconds, and 95 °C/1 minute, and the melt-curve analysis was performed at the end of each experiment to verify that a single product per primer pair was amplified. Furthermore, the sizes of the amplified DNA fragments were verified by gel electrophoresis on a 3% agarose gel. The amplification and analysis were performed using an iCycler iQ Multicolor Real-Time PCR Detection System (Bio-Rad). Samples were compared using the relative CT method. The fold increase or decrease was determined relative to naive or sham controls after normalizing to a housekeeping gene using 2^−ΔΔCT, where ΔCT is (gene of interest CT)−(glyceraldehyde-3-phosphate-dehydrogenase CT), and ΔΔCT is (ΔCT treated)−(ΔCT control), as described previously.^{19, 20, 25}

Immunohistochemistry

Immunohistochemical analyses were performed to determine Aβ, TDP-43, phosphorylated tau, Iba1, and glial fibrillary acidic protein in coronal brain sections as described previously.^{19, 25} Animals were anesthetized with ketamine/xylazine (200/10 mg/kg) and subsequently transcardially perfused with phosphate-buffered saline followed by 4% paraformaldehyde in phosphate buffer. The brains were quickly removed from the skulls and fixed in 4% paraformaldehyde overnight, and then transferred into the phosphate-buffered saline containing 30% sucrose until they sank to the bottom of the small glass jars. Cryostat sectioning was made on a freezing Vibratome at 40 μm and series sections (10 to 12 slices) were collected in 0.1 mol/L phosphate buffer. Free floating sections were immunostained using the specific antibodies listed above, followed by incubation with the corresponding fluorescent-labeled secondary antibody. 4′-6-diamidino-2-phenylindole (DAPI), a fluorescent stain that binds strongly to DNA, was used to detect cell nuclei in the sections. The sections were then mounted on slides for immunofluorescence detection using a Zeiss (New York, NY, USA) confocal or deconvolution microscope. The numbers and the area of total Aβ, and the fluorescence intensity (in arbitrary densitometric units) of glial fibrillary acidic protein and Iba1 in the hippocampal area in each image were analyzed and quantified using SlideBook 5.0 and NIH Image J software as described previously.^{19, 25}

Histochemistry

Degenerated neurons were detected using Fluoro-Jade C (FJC), which is an anionic dye that specifically stains the soma and neurites of degenerating neurons and thus is unique as a neurodegenerative marker. Cryostat-cut sections were incubated in the solution with FJC (0.0001% solution, EMD Millipore, Temecula, CA, USA) and 4′-6-diamidino-2-phenylindole (0.5 μg/mL) for 10 minutes, followed by 3 × 1-minute wash with distilled water. The slices were dried naturally at room temperature without light. The images were taken using a Zeiss deconvolution microscope with SlideBook 5.5 software, as described previously.^{19, 25}

Morris Water Maze

The classic Morris water maze test was used to determine spatial learning and memory, as described previously.^{19, 25} A circular water tank (diameter 120 cm) was filled with water and the water was made opaque with non-toxic white paint. A round platform (diameter 15 cm) was hidden 1 cm beneath the surface of the water at the center of a given quadrant of the water tank. Mice (grouped as naive, sham, TBI-vehicle, or TBI-JZL184) received training in the Morris water maze for 7 days and each session consisted of four trials. Animals that failed to find the platform hidden 1 cm beneath the surface of the water during the 7 days of training were excluded from the experiments. For each trial, the mouse was released from the wall of the tank and allowed to search, find, and stand on the platform for 10 seconds within the 60-second trial period. For each training session, the starting quadrant and sequence of the four quadrants from where the mouse was released into the water tank were randomly chosen so that it was different among the separate sessions for each animal and was different for individual animals. The mice in the water pool were recorded by a video-camera and the task performances, including swimming paths, speed, and time spent in each quadrant, were recorded using an EthoVision video tracking system (Noldus, Leesburg, VA, USA). A probe test was conducted 24 hours after the completion of the training. During the probe test, the platform was removed from the pool, and the task performances were recorded for 60 seconds.

Data Analysis

Data are presented as mean±s.e.m. Unless stated otherwise, analysis of variance followed by post-hoc tests was were used for statistical comparison when appropriate. Differences were considered significant when P<0.05.

Results

Monoacylglycerol Lipase Inhibition Promotes Neurological Recovery following Repetitive Mild Closed Head Injury

To investigate the consequences of repetitive TBI, we used a mouse model of mild CHI with three impacts at an interval of 24 hours, as described previously with modification.^{5, 23, 24} Animals were assigned into four groups: (a) naive control, (b) sham control, (c) TBI-vehicle, and (d) TBI-JZL184, a highly selective and potent MAGL inhibitor, which significantly elevates 2-AG levels in the brain.^{14, 15, 19, 21, 27} JZL184 (10 mg/kg) was injected 30 minutes after each TBI and then once a day for 4 consecutive days (Figure 1A). The neurologic severity score (NSS), which includes 10 tasks as described previously,²⁶ was assessed to determine the functional status of animals after rmCHI. We found that the NSS score was increased with the increased number of impacts and displayed a cumulative effect. The highest NSS score after three impacts was 3.8±0.3 (n=24 animals), as shown in Figure 1. This NSS score is within the range of the mild injury, as described previously.²⁶ However, the NSS scores were significantly reduced in impacted animals treated with JZL184 (Figure 1B), indicating that inhibition of 2-AG metabolism promotes neurologic recovery following repetitive TBI. Apparently, no positive NSS scores were detected 5 days after three impacts in both vehicle- or JZL184-treated TBI animals, suggesting that neurologic injuries are completely recovered 5 days after the impacts. In addition, no intracranial hemorrhage and convulsions were observed following rmCHI.

Figure 1.

Monoacylglycerol lipase inhibition promotes neurologic recovery following repetitive mild closed head injury. (A) Schematic illustration of the experimental protocol. Repetitive mild closed head injury (rmCHI) was induced using an electromagnetic controlled stereotaxic impact device. The black arrows represent the procedures (trauma brain injury (TBI) or sham), while the red arrows represent JZL184 injections. Neurologic severity score (NSS), proinflammatory cytokines, astroglial reactivation, Aβ, the enzymes synthesizing Aβ, TDP-43 aggregation, tau phosphorylation, glutamate receptor subunits, basal synaptic transmission, long-term potentiation (LTP), and spatial learning and memory were detected using quantitative PCR (qPCR), western blot (WB), immunohistochemistry (IHC), electrophysiologic recordings, and the Morris water maze (MWM). (B) Monoacylglycerol lipase inhibition improves neurologic recovery following rmTBI. JZL184 (10 mg/kg) or vehicle was injected intraperitoneally 30 minutes after each impact for 3 days. NSS was assessed 2 hours after each TBI and 5 days after 3 impacts. The data are means±s.e.m. ∗∗P<0.01 compared with naive or sham; ^#P<0.05, ^##P<0.01 compared with TBI-vehicle (analysis of variance with the Bonferroni post-hoc test, n=24 to 29 animals/group).

Neuroinflammation Following Repeated Mild Chronic Head Injury Is Attenuated by JZL184

Neuroinflammation has been recognized as one of the important hallmarks in TBI.^{12, 13} To determine whether proinflammatory cytokines are elevated and astroglial cells are activated in our animal model of rmCHI and whether these inflammatory responses can be suppressed by MAGL inactivation, we assessed messenger RNA expression of proinflammatory markers vimentin (Vim) and cytokines, including IL-1β, IL-6 and TNFα, in the ipsilateral brain using qPCR analysis 4 days after the first injury and glial reactivity 8 days after the first TBI. As seen in Figure 2A, expression of IL-1β, IL-6, TNFα, and Vim was robustly increased in TBI animals that received vehicle when compared with those in naive or sham controls. The expression of these proinflammatory markers was significantly decreased by administration of JZL184, indicating that inhibition of 2-AG metabolism is capable of suppressing neuroinflammation in animals that received rmCHI. This assumption was further supported by the results from immunoreactivities to Ib1a, a microglial marker, and glial fibrillary acidic protein, an astrocytic marker in TBI animals that received JZL184. As illustrated in Figure 2B, microglial activation in the cortex and astrocytic activation in the cortex, CA1, and dentate gyrus (DG) areas were significantly increased and these increases were inhibited by MAGL inactivation. These results suggest that inhibition of 2-AG metabolism may prevent further brain damage and neuropathologic changes following TBI by resolving neuroinflammation.

Figure 2.

Neuroinflammation following repetitive mild closed head injury (rmCHI) is prevented by inhibition of monoacylglycerol lipase. (A) Expression of proinflammatory markers vimentin (Vim) and cytokines, including interleukin (IL)-1β, IL-6, and tumor necrosis factor alpha (TNFα). Messenger RNA (mRNA) expression of cytokines in the ipsilateral brains from animals that received different treatments was analyzed using quantitative PCR 4 days after the first injury. The data are means±s.e.m. ∗∗P<0.01 compared with naive or sham; ^##P<0.01 compared with TBI-vehicle (analysis of variance with the Fisher's protected least significant difference (PLSD) test, n=3 animals/group). (B) Immunoreactivity of Iba1 (microglial marker) and (C) glial fibrillary acidic protein (GFAP) (astrocytic marker) in the ipsilateral cortex (CTX), hippocampal CA1, and dentate gyrus (DG) was determined using immunostaining analysis. Images were taken 8 days after the first TBI using a Zeiss deconvolution microscope. Scale bars: 40 μm. The data are means ±s.e.m. ∗∗P<0.01 compared with naive or sham; ^##P<0.01 compared with TBI-vehicle (n=4 to 6 animals/group). DAPI, 4′-6-diamidino-2-phenylindole; GFAP, glial fibrillary acidic protein.

Traumatic Brain Injury-Increased Formation of Aβ and Expression of the Enzymes Synthesizing Aβ are Suppressed by Inhibition of 2-Arachidonoylglycerol Metabolism

Accumulation and deposition of Aβ are the hallmarks of neuropathology in AD. Recent evidence shows that Aβ plaques are present in the brains of war veterans and sports players with CTE and that expression of APP and production of Aβ are increased following TBI.^{1, 4, 8} To determine whether expression of APP and production of Aβ are increased in our rmCHI model and whether the increases are suppressed by MAGL inactivation, we measured expression of APP and the enzymes that cleave APP, including β-site amyloid precursor protein cleaving enzyme 1 (BACE1), a Disintegrin and metalloproteinase domain-containing protein 10 (ADAM-10, α-secretase), and nicastrin (NCT, an important component of the γ secretase protein complex), in the brain 8 days after the first TBI. As shown in Figure 3A, expression of APP, BACE1, and NCT was significantly elevated in animals that received rmTBI when compared with those in naive or sham, while expression of ADAM-10 was not altered. The increases in APP, BACE1, and NCT following rmCHI were inhibited by JZL184. Since APP, BACE1, and NCT are critical for Aβ formation, increased expression of these molecules following rmCHI indicates an increased Aβ synthesis. Indeed, as shown in Figure 3B, 4G8 (total Aβ) immunoreactivity in the cortex was increased, but the increase was suppressed by JZL184.

Figure 3.

Inhibition of 2-arachidonoylglycerol metabolism reduces traumatic brain injury-elevated formation of Aβ and expression of the enzymes that synthesize Aβ. (A) Immunoblot analysis of amyloid precursor protein (APP), ADAM-10 (α-secretase), BACE1 (β-secretase), and nicastrin (NCT, a γ-secretase component) in the ipsilateral side of injured brains 8 days after the first TBI. The data are means±s.e.m. ∗P<0.05, ∗∗P<0.01, ∗∗∗P<0.001 compared with naive or sham; ^##P<0.01 compared with TBI-vehicle (n=3 animals/group). (B) Immunostaining of 4G8 (total Aβ) in the ipsilateral cortex 8 days after the first TBI. The data are means ±s.e.m. ∗∗P<0.01 compared with naive or sham; ^##P<0.01 compared with TBI-vehicle (n=4 animals/group). Scale bars: 40 μm.

Neurodegeneration following Repeated Mild Traumatic Brain Injury Is Attenuated by Inhibition of 2-Arachidonoylglycerol Metabolism

Neurodegeneration is one of the important neuropathologic changes in CTE. To determine whether neurodegeneration occurs in our rmCHI model and whether inhibition of 2-AG metabolism is capable of attenuating neurodegeneration, we detected FJC-positive neurons in the brain from animals exposed to rmCHI 8 days after the first TBI. As seen in Figure 4, rmCHI robustly increased the number of FJC-positive neurons in the cortex, CA1 and DG. As expected, the number of FJC-positive neurons was significantly reduced in TBI animals that received JZL184, suggesting MAGL inhibition decreases neurodegeneration after rmTBI.

Figure 4.

Neurodegeneration induced by repetitive mild closed head injury is attenuated by monoacylglycerol lipase inactivation. Fluoro-Jade C staining (FJC, a specific marker for degenerating neurons) in the ipsilateral CTX, CA1 and DG areas was made 8 days after the first injury. The data are means ±s.e.m. ∗∗P<0.01 compared with naive or sham; ^##P<0.01 compared with TBI-vehicle (n=4 animals/group). Scale bars: 40 μm. DAPI, 4′-6-diamidino-2-phenylindole.

TDP-43 Aggregation and tau Phosphorylation Are Inhibited by Monoacylglycerol Lipase Inactivation

Aggregation of TDP-43 and hyperphosphorylated tau are the two important hallmarks of CTE.^{1, 2, 4, 8} To determine whether TDP-43 aggregation and tau phosphorylation (p-tau) can be replicated in our rmCHI model and whether these neuropathologic changes can be suppressed by inhibition of 2-AG metabolism, we determined TDP-43, tau, cdk5 regulatory subunits p25/p35, and GSK3β 8 days after the first TBI. As shown in Figures 5A and 5C, TDP-43 and p-tau were robustly elevated in TBI animals. Repetitive mild CHI also increased p35/p25, p-GSK3β, but not cdk5. Tau is largely phosphorylated by GSK3β and cdk5/p25. Although expression of cdk5 was not changed after rmCHI, elevated p35/p25 and phosphorylation of GSK3β are sufficient to phosphorylate tau. TBI-induced increases in p35/p25, p-GSK3β, TDP-43 and p-tau were suppressed by JZL184. We noticed that there was a slight increase in p-tau in sham when compared with that in the native control 8 days after procedures (Figure 5A). This is likely an acute response to repeated surgical operations. To confirm this speculation, we assessed TDP-43 and p-tau 30 days after rmCHI. As shown in Figure 5B, p-tau was diminished in sham animals 30 days after surgical operations. However, in TBI animals, TDP-43 protein and p-tau were persistently elevated 30 days after rmTBI and the elevation was blocked by JZL184 (Figure 5B). These results suggest that TDP-43 protein was aggregated and tau protein was phosphorylated in our rmCHI model, and the changes in TDP-43 and p-tau as well as p35/p25 and p-GSK3β can be blocked by inhibition of MAGL. In addition, increased TDP-43 and p-tau after rmCHI and their suppression by JZL184 were confirmed by immunostaining, as seen in Figures 5C and 5D.

Figure 5.

Transactive response DNA-binding protein (TDP) aggregation and tau phosphorylation after repetitive mild closed head injury are diminished by inhibition of 2-AG metabolism. (A) Western blot analysis of brain cdk5, cdk5 regulatory subunits p25/p35, GSK3β phosphorylation (p- GSK3β), TDP-43, and tau phosphorylation (p-tau) 8 days after the first injury. The data are means ±s.e.m. ∗P<0.05; ∗∗P<0.01 compared with naive; ^##P<0.01 compared with TBI-vehicle (n=3 animals/group). (B) Western blot analysis of TDP-43 and p-tau 30 days after the first TBI. ∗∗P<0.01 compared with naive; ^##P<0.01 compared with TBI-vehicle (n=6 animals/group). (C, D) Immunostaining analysis of TDP-43 and p-tau in the brain 8 days after the first injury. Scale bars: 40 μm.

Impairments in Synaptic and Cognitive Function after Repeated Mild Traumatic Brain Injury Are Reversed by Monoacylglycerol Lipase Inactivation

Synaptic and memory impairments are the major consequences of TBI. To determine whether our rmCHI model displays deficits in synaptic and cognitive function, we measured basal synaptic transmission in terms of input—output function and long-term synaptic plasticity in terms of LTP at hippocampal CA3–CA1 synapses using electrophysiologic recordings and spatial learning and memory using the Morris water maze test 30 days after the first injury. As shown in Figure 6A and 6B, input—output function and LTP at hippocampal CA3–CA1 synapses were significantly impaired in TBI animals when compared with that in naive or sham controls. The deficits in basal synaptic transmission and plasticity induced by rmCHI were diminished in animals that received JZL184. LTP is one of the major cellular mechanisms that underlies learning and memory. Indeed, the Morris water maze test displayed impaired spatial learning and learning in TBI animals as indicated by the prolonged time required to find the platform, longer traveling distance, and reduced number of times crossing the target zone (the location where the platform was previously located). Interestingly, the impairments in behavioral performance in TBI animals were significantly alleviated by JZL184. These data provide important evidence that treatment with JZL184 for 7 days is sufficient to prevent the deleterious effects in synaptic and cognitive function after rmCHI, suggesting that inhibition of 2-AG metabolism would be a novel therapeutic intervention for preventing deficits in synaptic and cognitive function after rmTBI.

Figure 6.

Inhibition of 2-arachidonoylglycerol metabolism improves synaptic and cognitive function in traumatic brain injury (TBI) animals. (A) Input-output function recorded at hippocampal Shaffer-collateral synapses 30 days after the first injury. Stimulus intensity was normalized to the maximum intensity (4 to 6 mice/group). (B) Long-term potentiation at hippocampal Shaffer-collateral synapses 30 days after the first injury (6 to 8 animals/group). Scale bar: 0.3 mV/10 milliseconds. (C) Spatial learning and memory were detected using the Morris water maze test 30 days after the first injury (n=19 to 23 animals/group). The probe test was conducted 24 hours following 7 days of invisible training. The data are means±s.e.m. ∗∗P<0.01 compared with naive control and ^#P<0.05 compared with TBI-vehicle. fEPSP, field excitatory postsynaptic potential.

Traumatic Brain Injury-Altered Expression of Glutamate Receptor Subunits Is Attenuated by Inhibition of 2-Arachidonoylglycerol

Integrity of excitatory synaptic transmission and plasticity is largely dependent on expression and function of the glutamate receptors. Since hippocampal synaptic plasticity, and spatial learning and memory are impaired (Figure 6), we hypothesized that expression of glutamate receptor subunits in the hippocampus is likely altered in our animal model of rmCHI. To test this prediction, we detected expression of α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid and N-methyl-D-aspartate receptor subunits GluA1, GluA2, GluN1, GluN2A, and GluN2B 8 days after the first TBI. As shown in Figure 7A, expression of GluA1, GluA2, GluN2A, and GluN2B was significantly reduced in TBI animals, while expression of GluN1 was slightly increased. The TBI-reduced glutamate receptor subunits GluA1, GluN2A and GluN2B were reversed by JZL184. Although expression of GluA2 and GluN1was slightly recovered by JZL184, there were no statistical significances between TBI-Veh and TBI-JZL184. The expression of glutamate receptor subunits was persistently altered 30 days after the first TBI. As shown in Figure 7B, expression of GluA1, GluA2, GluN2A, and GluN2B was still repressed in TBI animals compared with those in naive controls. The reduced expression of these subunits almost returned to the control levels by JZL184 treatment. It was noticed that expression of GluN1 was further elevated 30 days after the first injury and not suppressed by JZL184. The results from this experiment suggest that improved synaptic and cognitive function in TBI animals by inhibition of 2-AG metabolism is likely associated with recovery of the glutamate receptor subunits GluA1, GluA2, GluN2A, and GluN2B.

Figure 7.

Alterations in expression of glutamate receptor subunits after repetitive mild closed head injury (rmCHI) are reversed by inhibition of monoacylglycerol lipase. (A) Immunoblot analysis of hippocampal expression of α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (GluA1 and GluA2) and N-methyl-D-aspartate (GluN1, GluN2A, and GluN2B) subunits in animals that received rmCHI. The analysis was performed 8 days after the first injury. ∗P<0.05, ∗∗P<0.01 compared with naive control and ^##P<0.01 compared with TBI-vehicle (n=3). (B) Immunoblot analysis of hippocampal expression of AMPA and NMDA subunits in animals that received rmCHI. The analysis was performed 30 days after the first injury. ∗∗P<0.01 compared with naive control, ^#P<0.05, ^##P<0.01 compared with TBI-vehicle (n=3).

Discussion

In the present study, we demonstrate that inhibition of 2-AG metabolism significantly reduced CTE-like neuropathologic changes, including increased expression of inflammatory cytokines, reactivation of astroglial cells, expression of APP and the enzymes that synthesize Aβ, production of Aβ, neurodegeneration, aggregation of TDP-43 protein, and phosphorylation of tau, impaired LTP and spatial memory, and altered expression of α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid and N-methyl-D-aspartate receptor subunits in a mouse model of rmCHI. Our results suggest that MAGL inhibition, which increases 2-AG and reduces 2-AG metabolites prostaglandins in the brain, may prevent development of CTE or halt disease progression.

Mild TBI accounts for ∼80% of all TBIs in athletes, military personnel, and the general population.^{7, 28} Recent evidence reveals that development of CTE most often occurs in persons who suffered mild TBI from multiple concussions or subconcussions.^{1, 4, 7, 29} The features of the clinical symptoms and neuropathologic changes in CTE are similar to those seen in AD.^{1, 2, 8, 10, 11} Clinically, the majority of mild TBI cases result from closed head injury that affects athletes, military personnel, and other individuals. Therefore, it is imperative to create the animal models of repetitive mild CHI that are clinically relevant and reflect the majority of rmTBI in humans to determine the molecular mechanisms underlying CTE and develop therapeutic interventions for this devastating disease. Several rodent models of rmCHI have been reported from two, three, five TBIs to 16 or 42 impacts at different intervals or frequencies.^{23, 24, 30, 31, 32, 33, 34, 35, 36, 37} In the present study, we chose a mouse model of rmCHI with three impacts at an interval of 24 hours, as reported previously.^{5, 23, 24} As shown in our results, our model displayed mild brain trauma with low NSS scores and the absence of intracranial hemorrhage and convulsions. Overall the model was noninvasive, simple, and reproducible. Importantly, our model resembled the most common features of neuropathology in CTE as presented in the results, including neuroinflammation, Aβ formation, TDP-43 aggregation, tau phosphorylation, neurodegeneration, and impairments in long-term synaptic plasticity, spatial learning and memory.

Contact sports player and military personnel are at high risk for developing CTE, a progressive neurodegenerative disease. Currently, no effective interventions are available for preventing or treating CTE. Neuroinflammation is a critical factor inducing neuropathogenesis following TBI.^{12, 13} This means that resolving neuroinflammation may prevent or reduce the incidence of CTE development after rmTBI. Endogenous cannabinoids display antioxidant, anti-inflammatory, and neuroprotective properties.³⁸ Earlier studies show that the levels of the endocannabinoid 2-AG are significantly increased in the brain following CHI, and administration of 2-AG attenuates TBI-induced neuropathologicchanges,^{16, 17, 18} indicating that 2-AG likely is an endogenous signaling mediator that maintains brain homeostasis against harmful insults.³⁸ The anti-inflammatory and neuroprotective behaviors of 2-AG suggest that manipulation of 2-AG signaling may provide novel therapeutic interventions to prevent TBI-induced neuropathologic events.²² MAGL is the primary enzyme degrading 2-AG in the brain.^{14, 15} Previous studies demonstrated that MAGL inactivation suppresses inflammatory cytokines in response to proinflammatory insults and neurodegeneration in an MPTP model of Parkinson's disease.²¹ Moreover, inhibition of 2-AG metabolism reduces Aβ production, gliosis, and neurodegeneration, and prevents deteriorations in synaptic plasticity and cognitive function in animal models of AD.^{19, 27} Current available information indicates that there are similarities and overlaps in neuropathology and neurobiology between CTE and AD, suggesting that inhibition of 2-AG metabolism is able to reduce neuroinflammation and neuropathology and prevents development of CTE after repetitive mild TBI. Inhibition of 2-AG degradation has been shown to reduce neuroinflammation and neurodegeneration and improve blood—brain barrier function following TBI.³⁹ In particular, the results from the present study showing that JZL184 significantly suppresses inflammatory cytokines, astroglial activation, Aβ, neurodegeneration, TDP-43 protein aggregation, and tau phosphorylation, and ameliorates impairments in synaptic and cognitive function in animals exposed to rmCHI suggest that inhibition of 2-AG metabolism may provide a pharmacotherapy for CTE.

One of the important features in CTE is loss of memory and cognitive deficits, which are closely associated with the impairments in synaptic transmission and plasticity. The integrity of synaptic transmission and plasticity is largely dependent on expression and function of glutamate receptors at synapses. In the present study, we provided evidence that expression of glutamate receptor subunits in the hippocampus is significantly altered in TBI animals. Correspondingly, hippocampal basal synaptic transmission and LTP are impaired in TBI animals. Increased latency to find the platform, prolonged travelling distance, and reduced number of times crossing the target zone indicate the deficits in spatial learning and memory retention. Thus, alterations in expression and function of glutamate receptor subunits and impaired synaptic transmission and plasticity after TBI are important factors contributing to cognitive deficits seen in CTE. Importantly, recovery of the altered glutamate receptor subunit expressions and improvement of synaptic and cognitive function by inhibition of MAGL suggest that strengthening 2-AG signaling after TBI may prevent further brain damage, neuropathological changes, and development of CTE.

Although inhibition of 2-AG metabolism displays anti-neuroinflammatory and neuroprotective properties, the molecular mechanisms responsible for these beneficial effects are still not clear. Recent studies show that the anti-inflammatory and neuroprotective effects by inhibition of MAGL are not mediated by CB1 or by CB2 receptors.^{19, 21, 27} In addition, inhibition of CB1 receptors only partly blocks 2-AG-produced neuroprotective effects in TBI.²² This suggests that there may be other mechanisms mediating the anti-neuroinflammatory and neuroprotective effects produced by inhibition of 2-AG metabolism.³⁸ It has been proposed that these beneficial effects may result from the reduction in metabolites of 2-AG when MAGL is inhibited.²¹ A latest study shows that 2-AG is able to increase activity and expression of peroxisome proliferator-activated receptor γ (PPARγ), a nuclear receptor involved in regulation of lipid metabolism and inflammation, and that the beneficial effects of MAGL inactivation are attenuated by PPARγ antagonism.⁴⁰ Since the metabolites of 2-AG hydrolysis by MAGL are glycerol and arachidonic acid, a precursor of prostaglandins and leukotrienes, it is possible that both reduction of 2-AG metabolites prostaglandins and leukotrienes, which are proinflammatory and neurotoxic, and elevation of 2-AG, which increases PPARγ activity and expression, by MAGL inhibition contribute to the anti-inflammatory and neuroprotective effects,³⁸ which in turn prevent development of CTE.

The long-term consequences of repetitive mild TBI are neurodegeneration, impairments in synaptic and cognitive function, and dementia, which are similar to those seen in AD. The data presented in this paper provide the first evidence that inhibition of 2-AG is capable of reducing CTE-like neuropathologic changes in a mouse model of rmTBI, suggesting that MAGL is likely a therapeutic target for preventing occurrence of CTE or halting disease progression.

Footnotes

ACKNOWLEDGMENTS

The authors thank NIH Mental Health Institute Chemical Synthesis and Drug Supply Program for providing JZL184.

The authors declare no conflict of interest.

References

Blennow

Hardy

Zetterberg

The neuropathology and neurobiology of traumatic brain injury. Neuron 2012; 76: 886–899.

Goldstein

Fisher

Tagge

Zhang

Velisek

Sullivan

Chronic traumatic encephalopathy in blast-exposed military veterans and a blast neurotrauma mouse model. Sci Transl Med 2012; 4: 134ra60.

Omalu

Chronic traumatic encephalopathy. Progr Neurol Surg 2014; 28: 38–49.

Smith

Johnson

Stewart

Chronic neuropathologies of single and repetitive TBI: substrates of dementia?

Nat Rev Neurol 2013; 9: 211–221.

Creeley

Wozniak

Bayly

Olney

Lewis

. Multiple episodes of mild traumatic brain injury result in impaired cognitive performance in mice. Acad Emerg Med 2004; 11: 809–819.

DeKosky

Blennow

Ikonomovic

Gandy

Acute and chronic traumatic encephalopathies: pathogenesis and biomarkers. Nature Rev Neurol 2013; 9: 192–200.

McKee

Robinson

. Military-related traumatic brain injury and neurodegeneration. Alzheimer's Dementia 2014; 10: S242–S253.

McKee

Stern

Nowinski

Stein

Alvarez

Daneshvar

The spectrum of disease in chronic traumatic encephalopathy. Brain 2013; 136: 43–64.

Jordan

. The clinical spectrum of sport-related traumatic brain injury. Nat Rev Neurol 2013; 9: 222–230.

10.

Lakis

Corona

Toshkezi

Chin

. Chronic traumatic encephalopathy - neuropathology in athletes and war veterans. Neurol Res 2013; 35: 290–299.

11.

Stern

Riley

Daneshvar

Nowinski

Cantu

McKee

. Long-term consequences of repetitive brain trauma: chronic traumatic encephalopathy. PM & R 2011; 3: S460–S467.

12.

Zetterberg

Smith

Blennow

Biomarkers of mild traumatic brain injury in cerebrospinal fluid and blood. Nat Rev Neurol 2013; 9: 201–210.

13.

Woodcock

Morganti-Kossmann

. The role of markers of inflammation in traumatic brain injury. Front Neurol 2013; 4: 18.

14.

Blankman

Simon

Cravatt

. A comprehensive profile of brain enzymes that hydrolyze the endocannabinoid 2-arachidonoylglycerol. Chem Biol 2007; 14: 1347–1356.

15.

Long

Nomura

Cravatt

. Characterization of monoacylglycerol lipase inhibition reveals differences in central and peripheral endocannabinoid metabolism. Chem Biol 2009; 16: 744–753.

16.

Panikashvili

Mechoulam

Beni

Alexandrovich

Shohami

CB1 cannabinoid receptors are involved in neuroprotection via NF-kappa B inhibition. J Cerebr Blood Flow Metab 2005; 25: 477–484.

17.

Panikashvili

Shein

Mechoulam

Trembovler

Kohen

Alexandrovich

The endocannabinoid 2-AG protects the blood-brain barrier after closed head injury and inhibits mRNA expression of proinflammatory cytokines. Neurobiol Dis 2006; 22: 257–264.

18.

Panikashvili

Simeonidou

Ben-Shabat

Hanus

Breuer

Mechoulam

An endogenous cannabinoid (2-AG) is neuroprotective after brain injury. Nature 2001; 413: 527–531.

19.

Chen

Zhang

Wang

Feng

Tang

Monoacylglycerol lipase is a therapeutic target for Alzheimer's disease. Cell Rep 2012; 2: 1329–1339.

20.

Zhang

Chen

Endocannabinoid 2-arachidonoylglycerol protects neurons by limiting COX-2 elevation. J Biol Chem 2008; 283: 22601–22611.

21.

Nomura

Morrison

Blankman

Long

Kinsey

Marcondes

Endocannabinoid hydrolysis generates brain prostaglandins that promote neuroinflammation. Science 2011; 334: 809–813.

22.

Shohami

Cohen-Yeshurun

Magid

Algali

Mechoulam

Endocannabinoids and traumatic brain injury. Br J Pharmacol 2011; 163: 1402–1410.

23.

Luo

Nguyen

Villeda

Zhang

Ding

Lindsey

Long-term cognitive impairments and pathological alterations in a mouse model of repetitive mild traumatic brain injury. Front Neurol 2014; 5: 12.

24.

Meehan

3rd Zhang

Mannix

Whalen

. Increasing recovery time between injuries improves cognitive outcome after repetitive mild concussive brain injuries in mice. Neurosurgery 2012; 71: 885–891.

25.

Chen

Zhang

Fan

Teng

Yang

Delta9-THC-caused synaptic and memory impairments are mediated through COX-2 signaling. Cell 2013; 155: 1154–1165.

26.

Flierl

Stahel

Beauchamp

Morgan

Smith

Shohami

Mouse closed head injury model induced by a weight-drop device. Nature Protoc 2009; 4: 1328–1337.

27.

Piro

Benjamin

Duerr

Gonzales

Wood

A dysregulated endocannabinoid-eicosanoid network supports pathogenesis in a mouse model of Alzheimer's disease. Cell Rep 2012; 1: 617–623.

28.

Levin

Robertson

. Mild traumatic brain injury in translation. J Neurotrauma 2013; 30: 610–617.

29.

Stein

Alvarez

McKee

. Chronic traumatic encephalopathy: a spectrum of neuropathological changes following repetitive brain trauma in athletes and military personnel. Alzheimer's Res Ther 2014; 6: 4.

30.

Klemenhagen

O'Brien

Brody

. Repetitive concussive traumatic brain injury interacts with post-injury foot shock stress to worsen social and depressionlike behavior in mice. PloS One 2013; 8: e74510.

31.

Mouzon

Chaytow

Crynen

Bachmeier

Stewart

Mullan

Repetitive mild traumatic brain injury in a mouse model produces learning and memory deficits accompanied by histological changes. J Neurotrauma 2012; 29: 2761–2773.

32.

Mouzon

Bachmeier

Ferro

Ojo

Crynen

Acker

Chronic neuropathological and neurobehavioral changes in a repetitive mild traumatic brain injury model. Ann Neurol 2014; 75: 241–254.

33.

Petraglia

Plog

Dayawansa

Chen

Dashnaw

Czerniecka

The spectrum of neurobehavioral sequelae after repetitive mild traumatic brain injury: a novel mouse model of chronic traumatic encephalopathy. J Neurotrauma 2014; 31: 1211–1224.

34.

Shitaka

Tran

Bennett

Sanchez

Levy

Dikranian

Repetitive closed-skull traumatic brain injury in mice causes persistent multifocal axonal injury and microglial reactivity. J Neuropathol Exp Neurol 2011; 70: 551–567.

35.

Yoshiyama

Uryu

Higuchi

Longhi

Hoover

Fujimoto

Enhanced neurofibrillary tangle formation, cerebral atrophy, and cognitive deficits induced by repetitive mild brain injury in a transgenic tauopathy mouse model. J Neurotrauma 2005; 22: 1134–1141.

36.

Kane

Angoa-Perez

Briggs

Viano

Kreipke

Kuhn

. A mouse model of human repetitive mild traumatic brain injury. J Neurosci Methods 2012; 203: 41–49.

37.

Mannix

Meehan

Mandeville

Grant

Gray

Berglass

Clinical correlates in an experimental model of repetitive mild brain injury. Ann Neurol 2013; 74: 65–75.

38.

Chen

Endocannabinoids in synaptic plasticity and neuroprotection. Neuroscientist advance online publication, 25 February 2014; doi: 10.1177/1073858414524632 (e-pub ahead of print).

39.

Tchantchou

Zhang

Selective inhibition of alpha/beta-hydrolase domain 6 attenuates neurodegeneration, alleviates blood brain barrier breakdown, and improves functional recovery in a mouse model of traumatic brain injury. J Neurotrauma 2013; 30: 565–579.

40.

Zhang

Teng

Tang

Chen

Synaptic and cognitive improvements by inhibition of 2-AG metabolism are through upregulation of microRNA-188–3p in a mouse model of Alzheimer's disease. J Neurosci 2014; 34: 14919–14933.