Experimental autoimmune encephalopathy (EAE)-induced hippocampal neuroinflammation and memory deficits are prevented with the non-opioid TLR2/TLR4 antagonist (+)-naltrexone

Abstract

Multiple sclerosis (MS) is associated with burdensome memory impairments. Preclinical literature suggests that these impairments are linked to neuroinflammation. Previously, we have shown that toll-like receptor 4 (TLR4) antagonists, such as (+)-naltrexone [(+)-NTX], block neuropathic pain and associated spinal inflammation in rats. Here we extend these findings to first demonstrate that (+)-NTX blocks TLR2 in addition to TLR4. Additionally, we examined in two rat strains whether (+)-NTX could attenuate learning and memory disturbances and associated neuroinflammation using a low-dose experimental autoimmune encephalomyelitis (EAE) model of MS. EAE is the most commonly used experimental model for the human inflammatory demyelinating disease, MS. This low-dose model avoided motor impairments that would confound learning and memory measurements. Fourteen days later, daily subcutaneous (+)-NTX or saline injections began and continued throughout the study. Contextual and auditory-fear conditioning were conducted at day 21 to assess hippocampal and amygdalar function. With this low-dose model, EAE impaired long-term, but not short-term, contextual fear memory; both long-term and short-term auditory-cued fear memory were spared. This was associated with increased mRNA for hippocampal interleukin-1β (IL-1β), TLR2, TLR4, NLRP3, and IL-17 and elevated expression of the microglial marker Iba1 in CA1 and DG regions of the hippocampus, confirming the neuroinflammation observed in higher-dose EAE models. Importantly, (+)-NTX completely prevented the EAE-induced memory impairments and robustly attenuated the associated proinflammatory effects. These findings suggest that (+)-NTX may exert therapeutic effects on memory function by dampening the neuroinflammatory response in the hippocampus through blockade of TLR2/TLR4. This study suggests that TLR2 and TLR4 antagonists may be effective at treating MS-related memory deficits.

1. Introduction

Multiple sclerosis (MS) is a chronic debilitating disease affecting more than two million individuals worldwide [1]. It is pathologically characterized by inflammation, demyelination, and axonal degradation within the central nervous system (CNS) [2]. The clinical severity and symptoms can vary per individual and range from neuropathic pain and motor disturbances to sensory deficits and cognitive impairments [2-5]. The pathophysiology of MS is generally thought to involve components of immune dysregulation and autoimmunity [6] in which an initial unknown antigen triggers infiltration of antigen-reactive T lymphocytes across the blood brain barrier into the CNS [2]. The invasion of T lymphocytes into the CNS triggers chemoattractant cytokines to recruit other inflammatory cells such as monocytes and B-cells, and local microglial cells are also recruited within the spinal cord and brain [2,7]. All of these cells produce the principal proinflammatory cytokine interleukin-1 beta (IL-1β), and other cytokines, creating a highly inflammatory environment that is considered to be involved in demyelination and axonal degeneration [2,7-9].

Given this neuroinflammatory environment, and the large literature linking neuroinflammation with various cognitive impairments [10-12], it is not surprising that long-term and working memory deficits are exhibited in an overwhelming 40-70% of MS patients [13]. Indeed, the cognitive challenges experienced by MS patients are broad, including learning, memory, attention, executive function, processing speed, and visuospatial disabilities [4,14,15]. Importantly, many of these deficits have been recapitulated in various animal models, thus substantiating experimental autoimmune encephalomyelitis (EAE) as a useful model to study MS-related memory deficits in rodents [16-26].

In considering possible neural mechanisms, there is substantial supporting evidence linking these deficits to microglial activation and inflammation in the hippocampus and other connected brain regions [16,17,19,21-24,26-29]. The pattern recognition receptors toll-like receptor 2 and 4 (TLR2/TLR4), expressed primarily by microglia and astrocytes in CNS, but also T cells [30,31] bind damage-associated molecular patterns (DAMPs) to activate acute and adaptive immune responses, including activation of the NLRP3 inflammasome which causes processing and release of IL-1β [32-35].

Given that neuronal degeneration and myelin breakdown are key pathological processes of MS and EAE, and these processes release DAMPs [36], we hypothesize that TLR2 and TLR4 may play key roles in MS pathophysiology. This notion is supported by findings demonstrating elevated mRNA expression of TLR2 and TLR4 [30] and microgliosis in the brains of MS patients [37,38]. These findings encourage novel therapeutic development of TLR2/TLR4 antagonists. By specifically blocking TLR2/TLR4, MS/EAE-associated neuroinflammation might be dampened in a specific manner without interfering with relevant adaptive immunity or other TLR-mediated innate responses.

We have previously shown that (+)-naltrexone [(+)-NTX] functions as a highly selective TLR4 receptor antagonist with potent capacity to reverse neuropathic pain in various rodent models [39-42]. Notably, (+)-NTX contains only the positive enantiomer of NTX and thus does not disrupt signaling through classical opioid receptors, as these only bind the negative enantiomer. Moreover, (+)-NTX can cross the blood brain barrier [39,40]. These factors make (+)-NTX a clinically-relevant compound that may be effective for the treatment of various brain diseases/disorders associated with inflammation. In this study, we explored two objectives. First, we determined whether (+)-NTX might block other TLRs in addition to TLR4, and second, we determined, in two rat strains, the extent to which (+)-NTX would prevent EAE-induced memory deficits and the associated neuroinflammatory phenotype.

2. Materials and Methods

2.1. Cell-Culture Specificity Testing of (+)-NTX for TLRs

To determine whether (+)-NTX might have antagonistic activity on other TLRs, BV-2 murine microglia were grown in supplemented media (Dulbecco Modified Eagle Medium [DMEM], including 10% Fetal Bovine Serum [FBS], 50 unit/mL penicillin, and 50 μg/mL streptomycin), and detached from the flask by cell lifter when confluence was reached. Cells were seeded at a density of 4×l0⁴ cells per well in 96-well plates. After overnight incubation, media was aspirated and changed to DMEM media without FBS. Cells were then treated with indicated concentration of TLR agonist (TLR1/2: 200 ng/mL Pam₃CSK₄; TLR4: 200 ng/ml LPS; TLR2/6: 200 ng/mL Pam₂CSK₄; TLR7/8, 100 ng/mL R848) and (+)-NTX (0, 10, and 40 μM) (National Institute of Health, gifted by Dr. Kenner Rice). It should be noted that BV-2 cells are not responsive to other TLR agonists (TLR3: poly(I:C); TLR5, flagellin; TLR9, CpG DNA), thus these agonists were not tested. In order to explore (+)-NTX’s specificity for TLRs, the effect of (+)-NTX on TLR signaling downstream nitric oxide (NO) was measured. Briefly, 100 μL of supernatant media was removed after cells were treated for 24 h and added to flat black 96-well microfluor plates (Thermo Scientific, MA, USA). Subsequently, 10 μL of 2, 3-diaminonaphthalene (0.05 mg/mL⁻¹ in 0.62 M HCl) was added to each well and incubated for 15 min. The reaction was quenched by addition of 5 μL of 3 M NaOH and the plate was read on a SYNERGY H1 Micro-plate Reader (BioTek Instruments, Carlsbad, CA, USA) with excitation at 360 nm and emission at 430 nm. The NO levels of TLR agonist-treated group without (+)-NTX was set as 100%.

2.2. Experimental Animals

The following experiments were designed to determine whether EAE would induce hippocampal-dependent learning and memory deficits, and the extent to which (+)-NTX would prevent any such deficits and the associated neuroinflammatory phenotype.

Subjects were male Dark Agouti (DA) rats (225-250g; Envigo) and male Sprague-Dawley (SD) rats (285-325 g; Envigo) between 9-10 weeks old upon arrival. Subjects were housed two per cage on a 12 hour light/dark cycle (lights on at 7 am). Early pilot studies indicated that these two strains exhibited differential susceptibility to EAE induction, as measured by motor disturbances. That is, DA rats required 16 μg MOG and SD rats required 32 μg MOG to produce standard EAE-related motor disturbances. Unfortunately, these motor impairments confound learning and memory measurements, as physical exploration and freezing behavior are essential to the contextual fear conditioning paradigm. Further pilot studies indicated that 4 μg MOG in DA rats and 8 μg MOG in SD rats were optimal doses to produce minimal motor disturbances (allowing for unconfounded behavioral assaying), and memory impairments (see Supplemental Figure S1 for dose-response functions in DA & SD rats). Thus, we used both strains of rats with their corresponding low dose of MOG to explore the generalizability of the effects of EAE on cognitive function in two rat strains that exhibit such variable responses to MOG.

Four cohorts of rats were used in this study. One DA rat cohort was used for the long-term memory experiments and another DA rat cohort was used for the short-term memory experiments. Separate cohorts were necessary because exposure to the contexts during the short-term memory tests would initiate extinction of the fear memory and thus impact the long-term memory measurements. A third cohort of DA rats was used to measure serum anti-MOG antibody levels and mRNA expression of various inflammatory markers in the brain that may have been modulated by EAE and/or (+)-NTX, but not by the behavioral experience. These rats were euthanized to coincide with the time of long-term memory testing. Lastly, a fourth cohort of SD rats was used to measure long-term memory and assess additional mechanistic implications of EAE and (+)-NTX treatment via immunohistochemistry. A schematic timeline of the experimental design can be seen in Figure 1. All cohorts started with n=8 per treatment group. For the behavioral studies, a 2 × 2 (disease x treatment) experimental design was used. For the PCR experiments, we conserved animals by comparing EAE/Saline and EAE/+NTX groups to only a naïve control group. All rats received access to water and food ad libitum. Experiments were conducted between 0800 and 1600 hour. All procedures were conducted in accordance with protocols approved by the University of Colorado Boulder Institutional Animal Care and Use Committee.

Figure 1.

Schematic of timeline and experimental design. Four cohorts of rats were used for this study. Cohorts 1-3 were Dark Agouti (DA) rats, whereas cohort 4 was Sprague-Dawley (SD) rats. Motor scores were assessed daily from day −1 to day 26 in all cohorts. On day 0, all rats were injected with either MOG or vehicle. Fourteen days later, half of each group received 3 daily injections of either (+)-NTX or saline until the end of the study. On day 21, cohorts 1, 2, and 4 were conditioned in a contextual fear conditioning paradigm. Cohorts 1 and 4 were tested for long-term memory four days later (day 25), whereas cohort 2 was tested for short-term memory 1-2 hrs later (day 21). The third cohort of rats was used to measure serum anti-MOG antibody levels and mRNA expression of various inflammatory markers in the brain that may have been modulated by EAE and/or (+)-NTX, but not by the behavioral experience. This cohort was euthanized on day 26. Lastly, hippocampi from the fourth cohort (also euthanized on day 26) were assessed via immunohistochemistry.

2.3. EAE Induction

Given that learning and memory behavioral procedures require normal motor and sensory capacity, in choosing MOG concentrations, our goal was to avoid hind limb motor/sensory dysfunction that could interfere with these procedures and confound interpretations. Thus, as explained above, we chose doses 25% of those that we have found to induce more severe EAE motor symptoms, such as partial/full-hind limb paralysis/partial upper limb paralysis in these two strains of rats (i.e. 16 μg in DA rats and 32 ug in SD rats, see Supplementary Figure S1). Upon arrival, the rats were randomly assigned to either the MOG or saline group. At 10-12 weeks of age, the MOG-assigned DA rats received a 4 μg injection of recombinant rat MOG1-125 (VU University Medical Center, Netherlands, gifted by Dr. Anne-Marie Van Dam) in a vehicle consisting of sodium acetate and incomplete Freund’s adjuvant (IFA, Sigma; St. Louis, MO; [43]. Also at 10-12 weeks of age, the MOG-assigned SD rats received an 8 μg injection of recombinant rat MOG1-125 in a saline vehicle. The MOG or vehicle injection was given intradermally at the base of the tail and the syringe was left in place for 3 minutes to avoid leakage from the injection site. All animals were between 10-12 weeks of age at the time of injection.

2.4. Motor Scoring

Motor behavior was scored on weekdays prior to (+)-NTX dosing and daily after the initiation of (+)-NTX dosing in all rats to assess the severity of their EAE symptoms as previously described [43-46]. The motor score quantified physical paralysis and scoring was based on the following designations: 0 = no signs of paralysis, 1 = partial tail paralysis, 2 = full tail paralysis, 3 = hind limb weakness, 4 = partial hind limb paralysis, 5 = full hind limb paralysis, 6 = partial upper limb paralysis. Rats that reached a score of 6 were euthanized if paralysis exceeded one day. These rats received a score of 7. MOG-treated rats with a motor score of 3 or higher on day 21 or later were excluded from the study.

2.5. (+)-Naltrexone Administration

Based on the presence of motor scores at two weeks post-MOG, rats were equally divided into (+)-NTX and saline treatment groups to achieve n=8 per treatment group. (+)-NTX (gifted generously by Dr. Kenner Rice, National Institute of Health) administration began 2 weeks post MOG administration. The animals received subcutaneous (+)-NTX (6 mg/kg) or saline injections three times daily (0900, 1200, and 1500 hour) for 2 weeks throughout behavioral testing for a total of 18 mg/kg/day. The volume of each injection administered was 1 mL/kg. This dosing regimen was selected due to the relatively short half-life of (+)-NTX, and has been used successfully in a previous study to treat a rat model of central neuropathic pain [39]. Indeed, in a separate pilot study, the effects of 18 mg/kg (+)-NTX once a day were less effective in decreasing IL-1β mRNA than when administered 3 times per day (unpublished data).

2.6. Contextual Fear Conditioning and Testing

Contextual fear conditioning was used to assess hippocampal and amygdalar memory function. The hippocampus-specific mechanisms that mediate contextual fear conditioning (conjunctive representations and pattern completion) are the same ones that enable the hippocampus to support episodic declarative memory in humans [47]. Since MS patients are known to exhibit declarative memory impairments [48], using this behavioral paradigm is appropriate with this disease model. Behavioral conditioning and testing were conducted as previously described [49]. Contextual fear conditioning of subjects occurred 3 weeks post MOG administration and one week into (+)-NTX or vehicle treatment. Two rats were taken from their home cages and placed in conditioning apparatuses. Each apparatus consisted of a clear plastic chamber (26 cm [L] X 21 cm [W] X 24 cm [H]) inside a larger sound-attenuating enclosure. The top of the plastic chamber consisted of wire mesh and the floor was constructed from stainless steel wire rods that connected to a shock generator and scrambler (Coulbourn Instruments). Mounted inside each enclosure were a speaker, a fan, one red light bulb, and one white light bulb. After the rats had spent two minutes exploring the chamber, a 15 second tone (76 dB) was sounded followed immediately by a 2 second footshock (1.5 mA). To assess obvious signs of lethargy or sickness, locomotion was scored during conditioning. After termination of shock, rats were returned to their home cage.

All rats were later tested for fear of the conditioning context (a hippocampal-dependent task) and fear of the tone (an amygdalar-dependent task) either 1 hour (for short-term memory) or 96 hours (for long-term memory) following conditioning. For the contextual fear test, rats were placed in the same context in which they were conditioned and observed for freezing behavior. For the auditory-cued fear test (conducted 1-2 hours after the contextual fear test), rats were placed in an altered context (e.g., different shaped and sized chamber, no grid floor, different lighting in the room) and observed for freezing behavior in response to the tone that had been paired with the footshock during conditioning. Long-term memory tests began at the same time of day (i.e. 9 am) as the conditioning trials 96 h prior, and rats were tested in same pairs/order as they were conditioned in order to maintain time-of-day consistency through both phases of the behavioral paradigm. Freezing is the rat’s dominant defensive fear response, and it is a common measure of conditioned fear [50]. Freezing was defined as the absence of all visible movement, except for respiration. Freezing scores were recorded for each rat every 10 second period for 6 minutes. Inter-rater reliability exceeded 97% for all experiments. On behavioral days, rats received their morning injections of saline and (+)-NTX immediately after conditioning/testing.

2.7. Tissue Collection

Two cohorts of rats were used for tissue studies in these experiments. The first cohort was a non-behavioral group of DA rats used for serum anti-MOG antibody and PCR assays, and the second cohort consisted of SD rats used for both behavioral and immunohistochemistry assays. Approximately four weeks (26 days) post-MOG administration while under sodium pentobarbital anesthesia, DA rats were transcardially-perfused with saline (pH 7.4), while SD rats were transcardially-perfused with saline (pH 7.4) followed by 5 min transcardial perfusion with 4% paraformaldehyde (pH 7.4) (Sigma-Aldrich, St. Louis, MO). Cardiac blood was also collected immediately prior to perfusion for serum anti-MOG antibody analysis in the DA rat cohort (described below). In an effort to conserve animals, we chose to omit the saline-saline and saline-(+)-NTX groups for the DA rat tissue studies, and instead used a naïve group of rats as controls. All rats continued to receive their (+)-NTX and saline injections on the day of the dissection until the time of tissue collection for a total of 12 days (+)-NTX. In the DA rat cohort, the hippocampus and amygdala were dissected from whole brain and stored at −80°C until time of assay. In the SD rat cohort, whole brain was extracted and then post-fixed as described in the immunohistochemistry section below.

2.8. Real-Time Polymerase Chain Reaction (PCR)

To examine whether our low dose MOG protocol evoked a proinflammatory phenotype in key brain regions, and to explore potential mechanisms through which (+)-NTX may work, we examined mRNA expression of various inflammatory markers in hippocampus and/or amygdala. We measured IL-1β, TLR2, TLR4, and NLRP3 due to their key roles in acute immune response [32,34,35]. We also explored IL-17 as an indicator of Th17 T cell and/or γδ T cell activity, which have been shown to be critical for EAE disease expression [31,51-53] and also have been demonstrated to interact with TLR2 and TLR4 signaling [31,53-55].

2.8.1. RNA Extraction

Hippocampal and amygdala tissues were homogenized on ice with 800 μL of trizol (Thermo Fischer Scientific; Waltham, MA) and allowed to sit at room temperature for 5 minutes. To separate the organic and aqueous layers, 160 μL of chloroform (Sigma; St. Louis, MO) was added to the samples before vortexing for 15 seconds and then incubating at room temperature for 3 minutes. The samples were spun at 12,000 g at 4°C for 15 minutes. The aqueous phase was separated from the organic phase and added in a 1:1 ratio to 100% 2-propanol where the samples incubated at room temperature for 10 minutes. The samples were then centrifuged again at the same temperature and speed for 10 minutes. The supernatant was decanted from each sample, 1 mL of 75% ethanol was added and the samples were spun in the centrifuge at 7,500 g for 5 minutes. The samples were decanted for a second time and the prior step was repeated. After the two cycles, the supernatant was decanted once more and the pellet was allowed to air dry. The pellet was then resuspended in 40 μL of nuclease free water (Bio-Rad; Hercules, CA). The concentration and purity of each RNA sample was measured using a NanoDrop One (Thermo Fischer Scientific; Waltham, MA).

2.8.2. cDNA Synthesis

The volume of a sample containing 3 μg of RNA was added to polymerase chain reaction (PCR) grade water for a total of 10 μL as calculated from the previously determined RNA concentrations (260/280: 1.8-2.1; 260/230: > 1.8). The samples were then incubated in an iCycler (Bio-Rad; Hercules, CA) at 65°C for 5 minutes in Mastermix 1 consisting of 50% random primers (Thermo Fischer Scientific; Waltham, MA) and 50% dNTP mix (Thermo Fischer Scientific; Waltham, MA). The samples received a quick chill on ice before a second incubation at 25°C for 2 minutes in Mastermix 2. Mastermix 2 consisted of 66% 5X first strand buffer (Thermo Fischer Scientific; Waltham, MA) and 33% 0.1M DTT (Thermo Fischer Scientific; Waltham, MA). Immediately after the incubation, 1 μL of Superscript II Reverse Transcriptase (Thermo Fischer Scientific; Waltham, MA) was added to each sample. The samples were then incubated at 25°C for 10 minutes, 42°C for 50 minutes, 70°C for 15 minutes, and then maintained at 4°C.

2.8.3. RT-PCR

Samples were analyzed in duplicate on 96-well PCR plates (Bio-Rad; Hercules, CA) in a Real Time System Thermocycler C100Touch (Bio-Rad; Hercules, CA). Each well contained 13 μL SYBR Green mix (Qiagen; Hilden, Germany), 1 μL forward primer, 1 μL reverse primer, 10 μL nuclease free water, and 1 μL sample cDNA. Primer sequences were designed using GenBank, National Center for Biotechnology Information (www.ncbi.nlm.nih.gov) to span exon/exon boundaries and thus exclude amplification of genomic DNA, and they were obtained from Invitrogen. Primers included GAPDH (forward: GGAGAAACCTGCCAAGTATG; reverse: GTCATTGAGAGCAATGCCAG); IL-1β (forward: CCTTGTGCAAGTGTCTGAAG; reverse: GGGCTTGGAAGCAATCCTTA); TLR2 (forward: TGGAGGTCTCCAGGTCAAATC; reverse: ACAGAGATGCCTGGGCAGAAT); TLR4 (forward: TCCCTGCATAGAGGTACTTC; reverse: CACACCTGGATAAATCCAGC); NLRP3 (forward: AGAAGCTGGGGTTGGTGAATT; reverse: GTTGTCTAACTCCAGCATCTG); IL-17 (forward: TCCATCCATGTGCCTGATGC; reverse: ACTCTGAGCCGCAATGAGGA). PCR cycling conditions consisted of a hot start activation of Hot Start Taq DNA polymerase (94°C , 15 minutes) and 40 cycles of denaturation (95°C , 15 seconds), annealing (57°C , 30 seconds), and extension (72°C , 30 seconds). PCR product was denatured (95°C , 1 minute) and annealed (55°C , 1 minute) before melt curve analysis was conducted. mRNA levels of IL-1β, TLR2, TLR4, NLRP3, and IL-17 were semi-quantified using the ΔC_T method relative to the housekeeping gene GAPDH. There were no group differences in GAPDH expression levels.

2.9. Immunohistochemistry.

To further explore mechanisms through which (+)-NTX dampens the neuroinflammatory response in the hippocampus, we assessed microglial and astrocyte immunoreactivity in the CA1 and dentate gyrus (DG) regions of the hippocampus using standard immunohistochemistry procedures. Brains were extracted, post-fixed in 4% paraformaldehyde/0.1 M phosphate buffer (PB, pH 7.4) for 24 h, and then cryo-protected in increasing concentrations (15, 20, 30%) of sucrose in PB (pH 7.4) (Sigma-Aldrich, St. Louis, MO). Brains were then blocked in OCT (Fisher Scientific, Waltham, MA), frozen at −80°C, and sectioned at 30 μm on a cryostat (Leika CM1850). Sections were then thaw-mounted and stored at −20°C until staining.

Slides were rinsed 3x with PBS, permeabilized with 0.3% hydrogen peroxide, rinsed 3x in PBS, blocked for 1 h with 10% NGS, 0.3% Triton-X in PBS, and then incubated overnight at 4°C for 24 h in 2% normal goat serum together with primary antibodies at the following dilution ratios: rabbit Iba1; 1:1000 (Wako, Richmond, VA) and rabbit GFAP; 1:500 (Dako, Carpinteria, CA). Slides were then rinsed 3x with PBS and incubated for 2 h in the secondary antibody at the following dilution ratio: biotinylated goat anti-rabbit IgGs; 1:200 (for Iba1) and 1:500 (for GFAP), (Jackson Immuno Research, West Grove, PA). Slides were then rinsed 3x in PBS, incubated in avidin-biotin complex (ABC) solution at 1:250 (Vector Laboratories, Burlingame, CA) for 2 h, rinsed 3x in PBS, and incubated in inactive 3,3′-diaminobenzidine (DAB) (Sigma-Aldrich, St. Louis, MO) for 10 min. DAB was then activated with B-D glucose (10 mg/ml) and slides were incubated for 8 min, rinsed 3x with PBS, and dried overnight. Slides were dehydrated in increasing concentrations of ethanol (50, 70, 95 and 100%), cleared in Citrisolv (Fisher Scientific, Waltham, MA), and then dried and covered with DPX mountant (Sigma-Aldrich, St. Louis, MO).

Black and white Images were acquired using an Olympus BX61 microscope (Olympus America, Center Valley, PA) with Olympus Suite CellSens Dimension software. All images of comparison were taken using the same exposure and other acquisition settings. Images were captured at 20x magnification. To conduct densitometry, images were converted to 32-bit, background subtracted, and then adjusted for threshold while blinded to the treatment conditions in NIH Image J software. Data were expressed as total area positive for staining within the selected area. Four images per animal were taken with one selection within each image analyzed, resulting in 4 areas per region of interest (CA1 and DG) of analysis per animal.

2.10. Anti-MOG Antibody ELISA

To rule out the possibility that (+)-NTX may interfere with the ability of rats to produce anti-MOG antibodies, an anti-MOG antibody assay was performed. Blood was collected via cardiac puncture at the time of tissue dissection from DA rat cohort 3 (i.e. day 26 post MOG). Samples were then centrifuged at 14,000 rpm at 4°C for 10 min. Supernatants containing isolated serum were then removed and stored at −80°C until ELISA was performed. SensoLyte Anti-Rat MOG(1-125) IgG quantitative ELISA kit (Anaspec, San Jose, CA) was used to assess anti-MOG antibody levels. The assay was performed according to manufacturer’s instructions. The assay sensitivity was 7.8125-500 ng/ml. An initial dilution curve was run to determine the proper dilution for optimal antibody detection at dilution ranges of 1:1000-1:125,000 according to the manufacturer’s instructions (data not shown). As a result of this initial dilution curve, samples were diluted 1:25,000 and a total of 100 μl of diluted sample were used for the assay. The analyte concentrations are presented as mg per ml of serum.

2.11. Statistical Analysis

Statistical analyses were conducted using GraphPad Prism v.8.30 software. Behavioral, cell culture, and immunohistochemistry data were analyzed with two-way ANOVA and serum antibody and PCR data were analyzed with one-way ANOVA. Mann-Whitney U non-parametric test was also used to analyze motor scores on days 23-25 in cohort 1. A significant ANOVA was followed by Tukey’s post hoc test to assess differences between specific experimental groups. For all tests, statistical significance was set to p < 0.05.

3. Results

3.1. (+)-NTX reduced TLR2- and TLR4-mediated nitric oxide (NO) release from BV-2 microglial cells in vitro.

We previously demonstrated that (+)-NTX is a TLR4 antagonist [40-42]. Here, the goal was to determine if (+)-NTX might also block the inflammatory responses induced by other TLRs. A two-way ANOVA [TLR agonist x dose (+)-NTX] revealed a significant interaction (F_{3, 32} = 8.79, p < 0.0001; Figure 2). Post-hoc analysis revealed that (+)-NTX (40 μM, but not 10 μM) significantly reduced nitrite in supernatant (a stable indicator of NO release) from vehicle control-treated BV-2 cells treated with Pam₃CSK₄ (p < 0.05), Pam₂CSK₄ (p < 0.05), and LPS (p < 0.0001) (control values represented by the dotted line as 100%), indicating that (+)-NTX blocked both TLR1/2 and TLR2/6 (heterodimers required for TLR2 signaling) as well as TLR4.

Figure 2.

Reduction of TLR nitric oxide release by (+)-NTX. Percent nitrite in supernatant (a stable indicator of nitric oxide release) from BV-2 cells treated with agonists of TLR4, TLR1/2; TLR2/6, or TLR7/8, and the indicated concentrations of (+)-NTX. Data are presented as mean ± SEM and analyzed using a two-way ANOVA, n=4 wells per group. Experiments were replicated 5 times. * p < 0.05; **** p < 0.0001.

3.2. (+)-NTX prevented hippocampal-dependent long-term memory impairments in two EAE rat models.

In the DA cohort (Figure 3A-C), a two-way ANOVA (EAE x treatment) revealed a significant interaction between EAE and treatment group (F_{1, 24} = 8.991, p < 0.01). Tukey’s multiple comparisons post-hoc test confirmed that saline-treated rats with EAE showed a significant decrease in freezing compared to rats without EAE (p < 0.01), indicating a significant deficit in long-term contextual memory. In addition, (+)-NTX-treated rats with EAE exhibited greater amounts of freezing compared to their saline-treated counterparts (p < 0.05), indicating that long-term contextual memory deficits in rats with EAE were rescued with (+)-NTX (Figure 3A). There were no significant differences between any of the other groups. Auditory cued-fear memory revealed no differences between the groups (Figure 3B). Lastly, all groups showed robust exploration during the conditioning session. No significant differences were observed between any of the groups, indicating normal locomotion and exploration (Figure 3C).

Figure 3.

Long-term memory in Dark Agouti (A-C) and Sprague Dawley (D-F) control or rats with EAE that were treated with (+)-NTX or saline. (A,D) Percentage of time spent freezing in the conditioning context (hippocampal-dependent memory), or (B,E) in a novel context in the presence of the tone (amygdala-dependent memory) 4 days after conditioning. (C,F) Percentage of time spent exploring the context during conditioning prior to shock. Data are presented as mean ± SEM and analyzed using a two-way ANOVA. DA rat: n=8 vehicle-saline and vehicle-(+)-NTX; n=6 MOG-saline and MOG-(+)-NTX. SD rat: n=8 per group. * p < 0.05; ** p < 0.01.

Similar results were found in the SD cohort (Figure 3D-F). A two-way ANOVA (EAE x treatment) revealed a significant interaction between EAE and treatment group (F_{1, 28} = 4.70, p < 0.05). Tukey's multiple comparisons post-hoc test confirmed that saline-treated rats with EAE showed a significant decrease in freezing compared to rats without EAE (p < 0.05) and that (+)-NTX treatment prevented this reduction (p < 0.01), thus rescuing long-term contextual memory deficits (Figure 3D). There were no significant differences between any of the other groups. Also similar to DA rats, SD rats did not have impaired auditory cued-fear memory (Figure 3E), suggesting impairment of hippocampal-dependent but not amygdalar-dependent memory. Lastly, similar to DA groups, SD rats showed good locomotion and did not have any differences in exploratory behavior during conditioning (Figure 3F).

3.3. EAE did not impair short-term contextual memory

To investigate whether the memory deficits observed in the previous experiment were specific to long-term memory, the effects of EAE on short-term memory were examined. To minimize the number of animals used, this was only conducted in a separate cohort of DA rats. A two-way ANOVA (EAE x treatment) showed no significant effects of condition, treatment, and or interaction, suggesting no impairments in short-term contextual fear memory (Figure 4A) or short-term auditory-cued fear memory (Figure 4B). As expected, no significant differences in exploratory behavior during conditioning were observed between any of the groups (Figure 4C). Notably, the data in Figure 4C, together with Figures 3C and 3F, support that no freezing is occurring during the conditioning trials in either the short- or long-term memory experiments, as the rats are at approximately 100% locomotion (i.e., absence of freezing).

Figure 4.

Short-term memory in control or rats with EAE that were treated with (+)-NTX or saline. (A) Percentage of time spent freezing in the conditioning context (hippocampal-dependent memory), or (B) in a novel context in the presence of the tone (amygdala-dependent memory) 1-2 hours after conditioning. (C) Percentage of time spent exploring the context during conditioning prior to shock. Data are presented as mean ± SEM and analyzed using a two-way ANOVA, n=8 vehicle-saline, vehicle-(+)-NTX, and MOG-saline; n=6 MOG-(+)-NTX group.

3.4. (+)-NTX reduced the increased IL-1β mRNA expression in the hippocampus of EAE rats

Here we assessed IL-1β mRNA expression in hippocampus and amygdala of MOG-inoculated rats treated with either saline or (+)-NTX. A significant one-way ANOVA (F_{2, 15} = 12.84, p < 0.0001; Figure 5A) and Tukey’s post hoc test revealed that saline-treated rats with EAE had significantly increased IL-1β levels in the hippocampus compared to naïve controls (p < 0.001). Furthermore, (+)-NTX treatment prevented this elevation (p < 0.01). IL-1β expression in rats treated with (+)-NTX did not differ from that of naïve rats (p > 0.05). In the amygdala, IL-1β expression was not significantly elevated in rats with EAE, and (+)-NTX did not alter those levels (F_{2, 20} = 2.25, p > 0.05; Figure 5B).

Figure 5.

Assessment of IL-1β mRNA in (A) hippocampus and (B) amygdala of naïve rats or rats with EAE that were treated with (+)-NTX or saline. Data are presented as mean ± SEM and analyzed using a one-way ANOVA. Hippocampus: n=8 naïve; n=4 MOG-saline; n=7 MOG-(+)-NTX. Amygdala: n=8 naïve, n=7 MOG-saline; n=8 MOG-(+)-NTX. * p < 0.05; ** p <0.01; ***p< 0.001.

3.5. (+)-NTX decreased EAE-induced Iba1 expression in the hippocampus.

Given that IL-1β was elevated in response to EAE in the hippocampus, but not in amygdala, and because the amygdala-mediated memory task was unimpaired, microglial and astrocytic expression and the additional inflammatory markers were only examined in the hippocampus. Immunohistochemistry was used to determine the effects of (+)-NTX on EAE-induced Iba1 expression in the CA1 and DG regions of the hippocampus, as these regions are known to be important for contextual memory [56]. Two-way ANOVA analyses (EAE x treatment) revealed significant interactions between EAE and treatment group for Iba1 in both CA1 (F_{1, 124} = 7.34, p < 0.05; Figure 6A) and DG (F_{1, 120} = 8.88, p < 0.001; Figure 6C). Tukey's multiple comparisons post-hoc tests confirmed that saline-treated rats with EAE showed a significant increase in Iba1 immunoreactivity compared to rats without EAE in both CA1 (p < 0.05) and DG (p < 0.0001) and that (+)-NTX treatment prevented these increases in both CA1 (p < 0.001) and DG (p < 0.001). Neither EAE nor (+)-NTX treatment affected GFAP in either CA1 (Figure 6B) or DG (Figure 6D). These results demonstrated that this dose of MOG increased microglia immunoreactivity, but not astrocytes in these regions of the hippocampus at this time point, and that (+)-NTX robustly decreased this microglial immunoreactivity.

Figure 6.

Microglial (Iba1) and astrocyte (GFAP) cell immunoreactivity of hippocampal CA1 and DG of control rats or rats with EAE that were treated with (+)-NTX or saline. Data are presented as mean ± SEM and analyzed using a two-way ANOVA, n=8 for all groups except n=7 for vehicle-(+)-NTX group in DG tissue. * p < 0.05; *** p < 0.001; **** p < 0.0001.

3.6. (+)-NTX decreased EAE-induced TLR2, TLR4, NLRP3, and IL-17 mRNA expression in the hippocampus

A one-way ANOVA indicated significant differences between the groups for all genes analyzed (TLR2 (F_{2, 16} = 6.19, p < 0.05), TLR4 (F_{2, 16} = 4.16, p < 0.05), NLRP3 (F_{2, 16} = 15.13, p < 0.0001), and IL-17 (F_{2, 16} = 10.38, p < 0.01)). Tukey’s post hoc tests revealed that saline-treated rats with EAE expressed significantly elevated levels of TLR2 (p < 0.05), TLR4 (p < 0.05), NLRP3 (p < 0.001), and IL-17 (p < 0.01) compared to naïve controls and (+)-NTX -treated EAE rats (TLR2 (p < 0.05); NLRP3 (p < 0.01); and IL-17 (p < 0.01)) (Figures 7A-D). Expression levels of all genes was comparable between (+)-NTX -treated and naïve control groups (p > 0.05).

Figure 7.

Gene expression of the pattern recognition receptors (A) TLR2 and (B) TLR4, (C) the inflammasome NLRP3, and (D) the proinflammatory cytokine IL-17 in the hippocampus of naïve rats or rats with EAE that were treated with (+)-NTX or saline. Data are presented as mean ± SEM and analyzed using a one-way ANOVA, n=8 naïve; n=4 MOG-saline; n=7 MOG-(+)-NTX. * p < 0.05; ** p < 0.01; *** p < 0.001.

3.7. Low-dose MOG increased serum anti-MOG antibodies that were not reversed by (+)-NTX.

To rule out the possibility that (+)-NTX treatment might interfere with the ability of rats with EAE to produce anti-MOG antibodies, we measured serum anti-MOG IgG levels. A one-way ANOVA indicated that low-dose MOG-treated rats produced significantly elevated levels of anti-MOG IgG antibodies compared to naïve rats (p < 0.01; Figure 8). These levels were not significantly different from those produced by (+)-NTX-treated rats (p > 0.05), indicating that (+)-NTX does not interfere with the ability of MOG to produce a normal autoimmune response.

Figure 8.

Increased serum anti-MOG IgG antibody concentration in rats with EAE that were treated with (+)-NTX or saline (mg/ml). Data are presented as mean ± SEM and analyzed using a one-way ANOVA, n=6 naive; n=5 MOG-saline; n=4 MOG-(+)-NTX. * p < 0.05; *** p <0.001.

3.8. (+)-NTX does not alleviate EAE-induced motor deficits.

In addition to assessing the therapeutic effects of (+)-NTX on memory deficits associated with EAE, we also measured the extent to which (+)-NTX might also ameliorate EAE-induced clinically relevant motor disturbances. Figure 9 shows the motor scores from the four cohorts of rats used in this study. As expected, and by design, low-dose MOG administration in DA rats resulted in less severe motor disturbances than those observed in standard-dose MOG models (Figure S1), but nonetheless resulted in similar relapsing-remitting disease progression (Figures 9A-9C). In the SD rat model, however, motor impairments were practically absent (Figure 9D). A two-way ANOVA indicated that (+)-NTX treatment had no effect on motor scores in any of these four cohorts (p > 0.05), suggesting that (+)-NTX is likely not a promising treatment for motor impairments associated with MS. In addition to the two-way ANOVA, a Mann-Whitney non-parametric test was also used to analyze the results in Fig 9A (between days 23-25) to assess a possible effect of (+)-NTX in that cohort. This analysis also resulted in a non-significant difference between saline and (+)-NTX-treated groups. Collectively, these data indicate that (+)-NTX did not alleviate EAE-induced motor deficits.

Figure 9.

Assessment of (+)-NTX administration on motor disturbances in Dark Agouti (DA) and Sprague-Dawley (SD) rats with EAE. Motor scores throughout (A) Cohort 1 (DA rat), (B) Cohort 2 (DA rat), (C) Cohort 3 (DA rat), and (D) Cohort 4 (SD rat). Data are presented as mean ± SEM and analyzed using a repeated measures two-way ANOVA, n=8 for all groups except MOG-(+)-saline (n=7) and MOG-(+)-NTX (n=6) groups in Cohort 1 and MOG-saline (n=7) group in Cohort 3.

4. Discussion

This study revealed three main findings. First, that (+)-NTX likely acts as an antagonist to TLR2 in addition to TLR4. Second, we determined, in two distinct rat models of EAE, that (+)-NTX was effective at preventing EAE-induced long term memory deficits, and third, that these ameliorative effects coincide with reductions in the EAE-induced inflammatory response in the CNS.

Using BV-2 cells treated with either TLR2, TLR4, or TLR7/8 agonists, we demonstrated that (+)-NTX treatment selectively reduced NO release, their primary output, only in cells treated with TLR2 and TLR4 agonists, confirming previous studies demonstrating that (+)-NTX acts as a selective antagonist of TLR4 [40-42], and extending those findings to also include TLR2.

In two distinct rat strains in which we induced EAE with low doses of MOG we demonstrated a robust deficit in hippocampal-mediated long-term contextual memory, but unimpaired amygdala-mediated auditory-cued memory, in both models. Importantly, short-term memory for both memory tests was spared, indicating EAE interfered with memory consolidation, and not learning. Normal locomotion observed during conditioning, and modest motor scores throughout the study in all animals suggested no physical deficits or overt sickness, and unimpaired short-term memory function served to further confirm that EAE did not interfere with the rats’ ability to explore, sample, and encode the context, or to express fear for what was learned. All of these findings are consistent with what others have reported previously [16,18-22,24,26,27,29,57-59]. Here we add the discovery that systemic administration of the non-opioid TLR4 antagonist, (+)-NTX, completely prevented this long-term memory impairment. It is important to note that it is unlikely (+)-NTX interfered with the ability of the rats to sense the footshock, as freezing was not altered by (+)-NTX treatment alone in any of the groups. Importantly, we confirmed that systemic (+)-NTX administration did not alter rats’ ability to produce anti-MOG antibodies, as these were comparable with those produced in saline-treated rats. Thus, these protective effects were not achieved by reducing reactivity to the antigen. Instead, our results suggest that (+)-NTX may have mitigated the memory impairments by dampening the neuroinflammatory response in the hippocampus. EAE induced a potent increase in IL-1β mRNA in the hippocampus, as many others have also reported [19,21,22,24,26], and this increase was robustly attenuated in (+)-NTX-treated rats. EAE did not evoke elevations of IL-1β mRNA in the amygdala. These findings are congruent with the observed EAE-induced impairments to hippocampal-mediated memory, but spared amygdala-mediated memory, and with the large literature showing that long-term memory processes are degraded by heightened neuroinflammation and preserved when neuroinflammation is tempered or absent [10-12].

To further characterize the neuroinflammatory phenotype in the hippocampus in this low-dose EAE model, we measured microglial- and astrocytic-immunoreactivity markers (Iba-1 and GFAP, respectively) in the CA1 and DG regions of the hippocampus. We found, consistent with what others have reported previously [19,21,22,24,26,28,60,60], robust EAE-induced elevations in microglial immunoreactivity in both regions of the hippocampus. EAE did not evoke changes in astrocyte immunoreactivity. These results were somewhat surprising, given the known role of astrocytes in EAE-induced disease expression [61-64], including in memory deficits [17,21]. It is possible that the threshold for increased astrocyte immunoreactivity requires higher MOG doses than for increased microglial immunoreactivity, and that higher doses would have resulted in increased astrocyte immunoreactivity as has previously been reported [17,21,63,64]. Importantly, (+)-NTX treatment effectively blunted the microglial immunoreactivity in both regions. Because microglia are one of the main cell types that produce IL-1β, this (+)-NTX-induced reduction in microglial immunoreactivity is consistent with the dampened IL-1β response observed in the hippocampus. Furthermore, given the large literature demonstrating a neuroinflammatory role in the cognitive deficits associated with EAE and MS [16,17,19,21-24,26-29], the current findings support the notion that (+)-NTX may have prevented the EAE-induced hippocampal-dependent memory deficits through a dampening of this neuroinflammatory phenotype in the hippocampus.

In addition to elevations in hippocampal IL-1β and microglial expression, EAE triggered substantial elevations of TLR2, TLR4, NLRP3, and IL-17 mRNA in the hippocampus, and these were all also blunted by (+)-NTX treatment, providing additional evidence that (+)-NTX may have exerted protection to hippocampal memory function via a dampening of the neuroinflammatory response in the hippocampus. TLR2 and TLR4, which are predominately expressed on microglia and astrocytes in the CNS, are typically activated by DAMPs, such as high mobility group box 1 (HMGB1) and heat shock protein 70 (HSP70), derived from various cellular compartments upon local damage [36]. Activation of these receptors on microglia and astrocytes then signals to form the NLRP3 inflammasome, which cleaves the inactive form pro-IL-1β into its active form IL-1β for release, leading to elevated neuroinflammation [32,44]. Thus, the blockade of these receptors, and subsequent downregulation of these receptors, with (+)-NTX leading to diminished inflammatory signaling was not unexpected. IL-17 is a Th17 T cell and gamma delta (γδ) T cell cytokine that has been shown to be critical for EAE disease expression [31,51-53]. It is unclear whether (+)-NTX would prevent IL-17 expression indirectly, as a result of normalized inflammatory signaling from microglia for example, and/or directly, as a result of binding TLR2 or TLR4 on Th17 and/or γδ T cell cells themselves. Indeed, TLR2/TLR4/TLR9-dependent crosstalk between microglia and neurotoxic IL-17+ γδ T cells has recently been described [65], and there are also TLR4-related signaling cascades in Th17 cells that might play a role in (+)-NTX’s therapeutic effects [53] [55].

The mechanisms by which (+)-NTX downregulates TLR2 and TLR4 expression remain to be investigated. Previous studies examining other therapeutic agents have pointed to downregulation of MyD88 and/or NF-kB/ERK/JNK and p-p38 pathways as mechanisms associated with decreased TLR2 and TLR4 expression [66-69]. Understanding the mechanism of action of (+)-NTX is the aim of future studies.

Finally, it is worth noting that the low-dose EAE models used here evoked pronounced memory deficits irrespective of the presence of motor disturbances. This dissociation is not unprecedented. Previous studies have demonstrated significant memory deficits in early pre-symptomatic EAE, late-phase EAE after remission, or in low-dose EAE where significant motor symptoms were also absent [16,18-22,24,26,27,29,57-59]. Indeed, SD rats which exhibited almost no motor disturbances nonetheless exhibited robust memory deficits. These findings collectively highlight the idea that EAE mechanisms underlying motor disturbances may be independent of those underlying memory function. Therefore, it is not too surprising that (+)-NTX treatment did not ameliorate even modest motor deficits.

This study provides the first evidence of a TLR2/TLR4 antagonist functioning as an effective inhibitor of memory deficits in EAE and furthermore demonstrates the effectiveness of a systemic TLR2/TLR4 antagonist at reducing inflammation in the hippocampus. As long-term and working memory deficits are displayed in 40-70% of MS patients at some point during their disease [13], these findings could open novel pathways to improvement in memory and related disabilities in MS patients.

Highlights.

• (+)-NTX (naltrexone) functionally blocks TLR2 and TLR4

• Low-dose EAE impairs long-term memory, but not short-term memory

• Low-dose EAE causes greater inflammation and microglial expression in hippocampus

• (+)-NTX treatment prevents EAE-induced long-term memory deficits

• (+)-NTX treatment prevents EAE-induced hippocampal neuroinflammatory phenotype