Introduction
Fear memory plays a central role in the development and onset of trauma- and stress-related disorders, such as posttraumatic stress disorder (PTSD) [
1]. Fear memory has been one of the most intensively studied areas and has been investigated in multiple animal models. By fear conditioning, adverse unconditioned stimuli (US) can alter the impact of neutral conditioned stimuli (CS) in certain neuronal circuits, thereby eliciting specific fear behavior induced by US. Fear conditioning is accomplished relying on the process of associative learning [
2]. Meanwhile, mechanisms of fear inhibition exist. Instead of forgetting, extinction is believed to be a new learning process, during which a new association between CS and US is established. Extinction learning involves the modification of synaptic connections in neuronal circuits similar to fear conditioning [
2,
3]. In contextual fear memory, the direct association between the context and US is established under bidirectional communication between amygdala and hippocampus [
4‐
6]. The hippocampus is responsible for the contextual fear memory in the establishment and retrieval of detailed contextual representation in fear and extinction memories [
7‐
11]. Particularly, the projections from the hippocampus to the amygdala and the medial prefrontal cortex (mPFC) are involved in the context-dependent fear response after extinction through increasing neuronal activity [
12,
13].
Chronic low-grade neuroinflammation has been reported in PTSD, anxiety, and major depressive disorders (MDD) [
14‐
16]. Clinically, multiple studies have indicated that such disorders are related to elevated circulating concentrations of pro-inflammatory cytokines, such as tumor necrosis factor-α (TNF-α), interleukin 1β (IL-1β), IL-6, and interferon-γ (IFN-γ) [
16‐
18]. Meanwhile, altered immune responses and increased pro-inflammatory reactions are also observed in patients with PTSD [
16]. Despite the abundant discoveries of neuroinflammation in PTSD and related disorders, the underlying mechanisms still remain unclear.
As pattern recognition receptors (PRRs) in the innate immune system, NACHT, LRR, and PYD domains-containing protein 3 (NLRP3) can recognize a wide range of stimuli, including Nigericin, reactive oxygen species (ROS), extracellular adenosine triphosphate (ATP), and crystalline uric acid [
19]. NLRP3 inflammasome activation has been suggested a robust link to the onset and progression of a wide range of central nervous system (CNS) diseases, such as Alzheimer’s disease (AD) [
20,
21], Parkinson’s disease (PD) [
22‐
25], anxiety [
26,
27], and MDD [
27‐
30]. Importantly, previous studies showed that upon the stimulation of stress, danger-associated molecular patterns (DAMPs), such as ATP and heat shock proteins (HSPs), can induce the activation of NLRP3 inflammasome [
31,
32]. NLRP3 inflammasome is tightly regulated by a two-step signaling. A priming signal occurs leading to transcriptional upregulation of NLRP3 and pro-IL-1β. Upon activation, the NLRP3 inflammasome is assembled with apoptosis-associated speck-like protein containing a CARD (ASC), leading to the maturation of interleukin-1β-converting enzyme caspase-1, which subsequently cuts pro-IL-1β into its activated form (IL-1β) [
19]. In animal studies, stress exposure increases IL-1β concentrations in multiple brain regions [
33]. Both insufficient and excessive levels of IL-1β impair the formation of memory [
34], indicating that IL-1β is important for normal learning and memory formation. IL-1 receptor null mutant mice show enhanced fear memory and decreased anxiety behavior [
35]. Meanwhile, the central administration of IL-1β can induce anxiety-like behavior and enhance fear memory after stress encounters [
35,
36]. However, despite the increasing evidences linking IL-1β production, neuroinflammation, fear memory, and related disorders, it remains unclear whether NLRP3 inflammasome serves as a causal factor and how it acts on fear memory. In this study, using both genetic and pharmaceutical strategies, we characterized the important roles of NLRP3 inflammasome in the regulation of neuroinflammation in fear memory.
Material and methods
Mice
All mice used in this study were C57BL/6 mice, 10-week-old males, weighted around 20 g. Mice were kept under ambient photoperiod at 26 ± 1 °C, had free access to standard rodent chow and clean water. Nlrp3 knockout mice (C57BL/6 background) were a generous gift from Prof. Rongbin Zhou (University of Science and Technology of China, Hefei, China). All animal experiments were approved by the Institutional Animal Care and Use Committee at Beijing Institute of Basic Medical Sciences.
MCC950 (Selleck, 1 mg/kg) administration was carried out via intraperitoneal (i.p.) injection for 3 consecutive days before the start of the contextual fear memory paradigm and extinction training. And then 1 h prior to each extinction train at the test days 1-3. Sterile saline i.p. injection was set as a control group.
Contextual fear memory paradigm
Experimental procedures were performed according to Zhang et al. with modifications [
37]. Generally, in the contextual fear memory paradigm, mice were introduced to the fear conditioning chambers (35 cm × 20 cm × 20 cm, Jiliang Tech) followed by a 5-min adaptation period. A total of 15 intermittent inescapable electric foot shocks (0.8 mA, 10 s with 10 s interval) were delivered. Mice in control groups were exposed to the fear conditioning chambers for an equivalent amount of time without the electric foot shocks. Fear conditioning chambers were wiped clean with 75% ethanol solution between tests. Forty-eight hours after the electric foot shocks, fear-conditioned mice were reintroduced to the same fear conditioning chambers once a day for 4 consecutive days for extinction training (test days 1-4). Mice in the control group were placed to the same fear conditioning chambers for an equivalent amount of time. Spontaneous activity (5 min) was recorded during extinction training. Freezing behavior of mice associated with contextual fear memory induced by adverse experience [
37]. Percentages of cumulative freezing time during spontaneous activity were used to reflect the fear responses of mice. All mice were tested throughout the procedure. Spontaneous activities were recorded and analyzed by Jiliang Tech analysis system.
Elevated plus maze
Elevated plus maze (EPM) contains 2 open arms (35 × 5 cm) and 2 enclosed arms (35 × 5 cm) connected by a center area (5 × 5 cm). The apparatus was lifted up 50 cm above the floor. Tests were carried out under a quiet and dimly lit environment. The apparatus was wiped clean with 75% ethanol between tests. Mice were introduced to EPM 24 h after the end of the extinction procedure. Mice were placed in the center area gently, facing to one of the open arms. Spontaneous activities were monitored for a 5-min period. Number of entries, time spent, and distance traveled in the open arms by the mice were analyzed by the ANY-maze software (Global Biotech).
Open-field test
Open-field (OF) apparatus (50 × 50 × 20 cm) was placed in a quiet and dimly lit environment, wiped clean with a 75% ethanol solution between tests. Twenty-four hours after EPM, mice were introduced to OF. Spontaneous activities were monitored for a 5-min period. Number of entries, time spent, and distance traveled in the center area (25 × 25 cm) by mice were analyzed by the ANY-maze software.
Stereotaxic surgery
Mice were anesthetized by pentobarbital sodium (70 mg/kg, dissolved in saline) via i.p. injection and immobilized on the stereotaxic apparatus (RWD). A guide cannula (OD 0.41 mm, C = 2.2 mm, RWD) was stereotaxically positioned into the lateral ventricle at the following coordinates from bregma: AP, −0.4 mm; ML, 1 mm; DV, −2.2 mm. The guide cannula was secured to the skull with dental cement and steel screws (M1.0 × L2.0 mm, RWD). Cap (OD 0.2, G = 0.5 mm, RWD) was screwed into guide cannula. Mice were maintained individually in cage and allowed to recover for 7 days before the start of infusion.
Intracerebroventricular infusion
The delivery system contains an injector cannula (OD 0.21 mm, C = 4 mm, G = 0.5 mm, RWD) fixed on a polyethylene (PE50) tube (OD 0.85 mm, RWD, filled with mineral oil) and connected to a syringe microinjector (5 μl, Hamilton). The infusion procedure was conducted 30 min before the start of the contextual fear memory paradigm and each extinction training. Mice were kept at the state of being conscious and able to move freely during infusion. Cap was removed before infusion. Injector cannula was inserted into guide cannula, secured on by fixing screws (OD 5.5 mm). The infusion procedure was programmed at the rate of 0.1 μl/min (total volume 1 μl) and delivered by a micro flow rate syringe pump (Longer Pump). After each infusion, the injector cannula was remained in the guide cannula for 10 min. Recombinant mouse IL-1 receptor antagonist (IL-1ra, R&D Systems, dissolved in sterile saline) was infused at a dose of 90 μg/kg each time at test days 0-3. This dose of IL-1ra was chosen according to previous researches proving it was able to block depressive-like behavior in mice [
38‐
40]. Recombinant mouse IL-1β (R&D Systems, dissolved in sterile saline) was infused at a dose of 1 ng/kg each time (test days 0-3). Same volume of sterile saline was infused as control.
Microglia isolation
Mice were anesthetized by pentobarbital sodium via i.p. injection and cardiac perfused with saline. Microglia were isolated according to the procedure described previously [
41,
42]. Whole-brain tissue was freshly harvested, cut into small pieces, suspended in Dounce buffer (1.5 mM HEPES, 0.5% glucose in HBSS buffer), and homogenized gently by Dounce homogenizer. Homogenates of brain tissue were suspended in phosphate-buffered saline (PBS; 8 g/l NaCl, 0.2 g/l KCl, 1.44 g/l Na2HPO4, 0.24 g/l KH2PO4), filtered with cell strainers (70 μm), and centrifuged at 600×
g for 6 min (4 °C) to collect the cell pellets. One hundred percent of percoll solution was prepared with an absolute percoll (GE Healthcare) dissolved in 10× PBS (9:1), and further diluted (v/v) to 70%, 37%, and 30% with PBS. Cell pellets were suspended in 37% percoll solution. Microglia were isolated by density gradient centrifugation. Density gradient was added into 15 ml centrifuge tubes, by layers of percoll solution from bottom to top containing: 70%, 37% (with cell suspension), 30% percoll solution, and PBS. Centrifuge was carried out in a horizontal centrifuge at 2000×
g for 30 min (4 °C). Microglia were converged on the interphase between 37% and 70% percoll solution. Isolated microglia were washed with 10× volumes of PBS, and centrifuged at 600×
g for 6 min (4 °C). Cell pellets were ready for mRNA extraction.
Tissue harvesting
Mice were anesthetized by pentobarbital sodium via i.p. injection and cardiac perfused with saline. Hippocampal tissue was harvested freshly from the brain and stored at −80 °C until use.
Quantitative real-time PCR
Total RNA of hippocampal tissue was extracted by Trizol regent (Invitrogen). Total RNA of isolated microglia was extracted by NucleoSpin RNA Plus XS kit (Macherey-Nagel). A 0.5 μg aliquot of total RNA of each sample was reversely transcribed using a one-step first strain cDNA synthesis kit (Transgen). Primer sequences used for quantitative real-time PCR (qPCR) were listed in Table S1. qPCR reactions were performed on QuantStudio 3 (Applied Biosystems) using 2× SYBR Green PCR master mix (Genestar). Data were quantified with comparative Ct method (2−ΔΔCt) based on Ct values normalized to β-actin. All tests were performed in triplicates.
Immunoblot analysis
Samples were prepared in SDS loading buffer (50 mM Tris HCl, 2% SDS, 0.1% bromophenol blue, 10% glycerol, 1% β-mercaptoethanol, pH = 7.0), fractionated by SDS-PAGE and transferred to nitrocellulose film (GE Healthcare). The film was then blocked in 5% skimmed milk. Following, primary and secondary antibodies were used for immunoblotting: NLRP3 (AdipoGen, AG-20B-0014, 1:1000), Caspase-1 (AdipoGen, Ag-20B-0042, 1:1000), PSD95 (Cell Signaling, 2507, 1:1000), Shank2 (Absin, abs134803, 1:1000), Shank3 (Cell Signaling Technology, 64555, 1:1000), β-actin (Invitrogen, MA5-11869, 1:3000), Goat anti-rabbit IgG (Jackson ImmunoResearch Laboratories, 1:5000), Goat anti-mouse IgG (Jackson ImmunoResearch Laboratories, 1:5000). Immunoblot was visualized by ECL (Thermo Scientific) and analyzed using software Image J.
Immunohistochemistry and immunofluorescence
All procedures were performed as previously described [
43,
44]. Briefly, mice were anesthetized by pentobarbital sodium and perfused with saline. Brains were fixed with 4% paraformaldehyde (w/v) for 7 days. Fixed brains were transferred to 30% sucrose solution for 3 days. Coronal sections were cut throughout the whole brain. In immunohistochemistry (IHC), brain slices were blocked with 10% goat serum (Abcom, ab7481) in PBS containing 0.2% Triton X-100 (Sigma, V900502), and incubated with Iba1 primary antibody (WAKO, 019-19741, 1:400), biotinylated goat anti-rabbit IgG, and streptavidin-conjugated horseradish peroxidase using VECTASTAIN® ABC-HRP kit (Vector Laboratories, PK-4000). Iba1 stains were visualized with 3,3′-diaminobenzidine (Sigma-Aldrich), scanned and analyzed by stereo investigator (MicroBrightfield). In immunofluorescence, brain slices were blocked, and incubated with primary and secondary antibodies as follow: Iba1 (Novus Biologicals, NB100-1028, 1:400), ASC (Cell Signaling Technology, D2W8U, 1:500), Alexa Fluor 546-conjugated (Invitrogen, A-11056, 1:400), and FITC-conjugated (Abcam, ab6798, 1:400). Hoechst 33258 (Sigma) were incubated for the visualization of nuclear morphology.
ELISA
Mouse IL-1β was measured in hippocampal homogenates according to the manufacturer’s instructions (R&D Systems). Hippocampal homogenates were prepared in cell lysis buffer (50 mM Hepes pH 7.4, 150 mM NaCl, 1% Nonidet P-40, 0.1% deoxycholate, 0.05% SDS, 0.1 M NaF, 1 mM EGTA, and protease inhibitor cocktail), and aligned according to protein concentrations measured by Pierce™ BCA Protein Assay kit (Thermo Scientific). All tests were performed in triplicates.
RNA sequencing and analysis
mRNA of hippocampal tissue was extracted and purified by the Dynabeads mRNA purification kit (Invitrogen, 61006). Aliquots of mRNA extracted from five mice were mixed in each group and fragmented for cDNA library constructing. cDNA library was sequenced by Illumina HiSeq 4000 sequencing platform. Low-quality reads in the raw data were trimmed using Trimmomatic v0.33 [
45]. Clean data were mapped against Mus musculus reference genome (GRCm38) by STAR, then read counts were used for the expression level quantification of each gene using RSEM [
46]. The differentially expressed genes (fold change > 1 and the adjust
p value < 0.05) between two groups were analyzed by DESeq v1.10.1 (adjust
p < 0.05) [
47]. Pathway and process enrichment analysis was conducted by Metascape (
http://metascape.org/gp/index.html) [
48]. Gene Set Enrichment Analysis (GSEA) from the target passway was performed using the GSEA v2.0.14 software (
http://www.broadinstitute.org/gsea/index.jsp). Heat-map representation of gene expression was generated by the “pheatmap” package of R (
https://CRAN.R-project.org/package = pheatmap).
Statistical analysis
All results were expressed as mean ± SEM. Statistical analysis was performed by GraphPad Prism 6. Data were subjected to t tests (unpaired t test with Welch’s correction) or ANOVA for significance of difference, to where appropriate. A significant level was set to p < 0.05 for all statistical analysis.
Discussions
In this study, we demonstrated that stress exposure by electric foot shocks can significantly increase the hippocampal concentration of cleaved caspase-1 and IL-1β in 3 h, suggesting an immediate activation of the NLRP3 inflammasome shortly after the stress encounter. Previous studies have found that the NLRP3 inflammasome is activated in stress-induced depressive animal models and MDD patients [
28], and that antidepressant treatments inhibited it through autophagy [
51]. In accordance with these findings, our results revealed the activation of the NLRP3 inflammasome in an electric food shocks-induced animal model of PTSD. Most importantly, the activation of the NLRP3 inflammasome was observed earlier than the appearance of systematic inflammation was. In the CNS, the expression of NLRP3 can be detected in microglia, astrocytes, and neurons during severe pathological conditions such as spinal cord injury [
52], but can only be found in the microglia/macrophages under physiological conditions [
53,
54]. A previous study has reported that the expression of NLRP3 is restricted to microglia in a study of animal models of depressive disorder [
55]. In isolated microglia, we detected the transcriptional upregulation of
il-1β and
tnf-α 3 h after the electric foot shocks, indicating that the process of initiating neuroinflammation after stress stimulation may predominately occur in microglia. However, we cannot exclude the possible effects of peripheral immunity during this process, as the animals we used in our study were
Nlrp3 knockout and MCC950 i.p. injected mice. The activation of the NLRP3 inflammasome in mononuclear blood cells has also been reported in patients with MDD [
56]. Additionally, a recent study has revealed the effect of peripheral CD4
+ T cells in stress-induced anxiety-like behavior in mice [
57]. Despite the early reaction of the NLRP3 inflammasome, microglia were activated 72 h after the electric foot shocks, as revealed by the significant increase in Iba1
+ cell numbers in the hippocampus. This observation was also supported by GSEA, where the significant upregulation of both toll-like receptor and RIG-I-like receptor signaling occurred at 72 h after the electric foot shocks, suggesting that neuroinflammation continually developed after the stress encounter. Abnormalities in microglia can directly cause neuropsychiatric disorders in mice. Mutation of Homeobox B8 (Hoxb8, only expresses in the microglia in mice) in mice causes pathological grooming, hyperanxiety, and social impairment deficits, which are similar to the obsessive-compulsive disorder (OCD) and autism spectrum disorders (ASDs) observed in human [
58,
59]. Concurrently, the loss of PSD proteins, including PSD95 and Shank proteins, suggested impaired neuronal function, which was associated with the development of neuroinflammation. Interestingly, defection in Shank proteins has been intensively studied as one of the most important causes of ASDs.
PSD95 (coded by
Dlg4) belongs to the membrane-associated guanylate kinase (MAGUK) family and is able to interact with N-methyl-D-aspartate (NMDA) receptors, α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptors and potassium channels at the postsynaptic site [
60‐
62], and plays important roles in synaptic plasticity and synaptic stabilization in long-term potentiation (LTP) [
63].
Dlg4−/− mice display abnormalities in multiple behavior tests, including increased responses related to stress and anxiety [
64]. Shank proteins are scaffold proteins predominantly localized within the PSD of glutamatergic synapses in the CNS. Direct association of Shank2 and PSD95 has been reported [
65]. Shank proteins also interact with AMPA receptors, NMDA receptors, and mGLu receptors and have been associated with ASDs and other neurological diseases in the CNS [
66]. Moreover, Shank3 insufficiency in the anterior cingulate cortex (ACC) can generate dysfunctions in excitatory synaptic and social interaction deficits in mice [
67]. On the molecular level of fear memory, NMDA receptor-dependent neuron excitability is necessary for this process. Antagonizing or enhancing the NMDA receptor can modulate the processes for both fear conditioning and extinction [
68‐
70]. Activation of the NDMA receptor in the hippocampus is essential for the formation and trace of contextually regulated fear memory [
71]. In the hippocampus, the NMDA receptor is upregulated during extinction training in a rat model of PTSD [
72]. NMDA receptor agonist D-cycloserine administration eliminates the upregulation of NMDA receptor subunits and alleviates impaired fear extinction memory [
72]. Simultaneously, fear conditioning induces the trafficking of AMPA receptors in the lateral amygdala (LA), thereby affecting associative learning [
73]. Importantly, it has been reported that the pro-inflammatory cytokine IL-1β can regulate the trafficking of AMPA receptors, leading to depression-like behaviors in mice after chronic social defeat stress (CSDS) [
38]. In
Nlrp3 knockout mice, stress exposure induced a level of fear response (without extinction) similar to that of WT mice, suggesting an unscathed associative learning. Most remarkably, both alleviated neuroinflammation and PSD proteins lost were observed in the
Nlrp3 knockout mice 72 h after electric foot shocks, suggesting a direct impact of neuroinflammation on the integrity of neuronal functions. Improved extinction learning in
Nlrp3 knockout mice has been further proven in the extinction training procedure, as revealed by the significant lower fear response and less anxiety-like behaviors. Both MCC950 treatment and IL-1ra administration further confirmed this phenomenon by displaying similar improvements during the extinction training.
As an important form of learning and memory, extinction is essential for the regulation of fear memory. In the clinic, PTSD is often treated by exposure therapy, which is supported by extinction learning [
74]. However, extinction learning deficits have been implicated in PTSD and other related neuropsychiatric disorders [
75]. In addition to psychological interventions, most PTSD patients also receive pharmacologic treatments. Drugs that alleviate specific symptoms (e.g., insomnia, nightmares, and alcohol abuse) have also been recommended in the treatment of PTSD. However, these agents are usually symptoms-based and rarely induce remission. The relapse of the symptoms often occurs upon discontinuation of the agents [
1]. MCC950 acts as a selective inhibitor of the NLRP3 inflammasome by targeting its ATP-hydrolysis motif [
49,
50] and has been investigated to treat a wide range of CNS diseases related to the NLRP3 inflammasome [
20]. The protective effects of genetic knockout and pharmaceutical inhibition of the NLRP3 inflammasome and IL1-ra suggest a novel strategy in the treatment of PTSD and related neuropsychiatric disorders. Our study provides evidence for the development of new treatments for PTSD by targeting the NLRP3 inflammasome. Further studies are required for the clinical treatment of PTSD and related disorders.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.