Male NLRP3 knock-out (KO) mice (NLRP3^Tm1Bhk/g) (n = 20) and their normal littermates (C57Bl6/J n = 27) (Charles River, Calco, Italy) were used in this study. Animals were housed in plastic cages, maintained under controlled environmental conditions (12 h light–dark cycle, 24°C), and provided with standard food and water ad libitum in the Animal Facility of the Department of Clinical and Experimental Medicine of the University of Messina.

All experiments were carried out according to the standards for care and use of animals as stated in the Directive 2010/63/EU, and the ARRIVE guidelines (Kilkenny et al., 2010). The procedures were evaluated and approved by the Ethic Committee of the University of Messina (#08/15).

Animals were randomly assigned to the following groups: Sham WT (n = 6), Sham NLRP3^-/- (n = 6), WT-TBI (n = 14), WT-TBI + BAY 11-7082 (20 mg/kg/i.p.; n = 7), NLRP3^-/- -TBI (n = 14).

Both NLRP3^-/- and WT animals were anesthetized by intraperitoneal injection of ketamine/xylazine (80 and 10 mg/kg, respectively) and subjected to an impact-acceleration model of diffuse TBI, based on the model described by Marmarou et al. (Marmarou et al., 1994), with some modifications. Briefly, an incision was made along the sagittal midline to expose the skull and a 3 mm thick steel disk was placed on between λ and bregma sutures. Animals were arranged on a 10-cm foam bed to soften the impact. Briefly, a 8 g weight with a 5 mm diameter and 5.5 cm length was dropped from a distance of 1.27 m, to produce TBI. After TBI, the steel disk was removed and the skin immediately sutured; a 2% lidocaine jelly was spread on to the impact site to reduce pain. Animals in WT-TBI and in NLRP3^-/--TBI group were randomly killed at 24 h (n = 7 from each group) or 7 days (n = 7 from each group) following procedures. The WT-TBI + BAY 11-7082 group received the NLRP3 inflammasome inhibitor BAY 11-7082 (Sigma–Aldrich S.r.l., Milan, Italy) 10 min after TBI procedures and were killed at 24 h or 7 days for histological and biochemical assessments.

At 24 h following TBI, both NLRP3^-/- and WT mice were subjected to an object recognition task to evaluate recognition memory (Antunes and Biala, 2012). Before TBI, in the “acquisition phase”, animals were individually put into an open field box (59 × 59 × 20) cm containing two identical objects (A and B) located in a symmetric position for 5 minutes to get acquainted with the new space. The two objects were adequately heavy and high so that mice could not move or climb over them. Twenty-four hours after TBI, one object, A or B, was substituted for a novel one (C) and exploratory behavior was again evaluated for 5 min. For each evaluation, both box and all objects were carefully cleaned with 70% ethanol to remove every odor recognition. Exploration of an object was evaluated whenever the mice rearing on or sniffing an object at a distance of less than 2 cm and/or touching it with the nose. It was considered a successful recognition if the animal preferred the exploration of the novel object. Discrimination of visual novelty was assessed by a preference index (Dix and Aggleton, 1999), calculated as follows: (time near the new – time near the old object)/(time near the new + time near the old object).

Brains were collected 24 h following TBI and 100 mg tissue was rinsed with 1X PBS, homogenized in 1 mL of 1X PBS and stored overnight at -20°C. After two freeze-thaw cycles were performed to break the cell membranes, the homogenates were centrifuged for 5 min at 5000 × g at 4°C. The supernatant was removed and assayed immediately to determine IL-1b and Caspase-1, using mouse-specific commercially available ELISA kits (Cusabio, College Park, MD, United States). Samples were run in duplicate, and the absorbance was read at 450 nm; the results were compared to the standard curves and expressed in pg/mL.

Brains were collected 24 h or 7 days following TBI. Hippocampus was obtained from brains and samples were homogenized in 400 μl of RIPA buffer with proteases inhibitors [25 mM Tris/HCl (pH 7.4), 1.0 mM EGTA, 1.0 mM EDTA, 0.5 mM phenyl methylsulfonyl fluoride, aprotinin, leupeptin, pepstatin A (10 μg/mL each) and NP-40 (10 μl/mL)]. Homogenate samples were centrifuged and the supernatants were collected and used for protein determination using the Bio-Rad protein assay kit (Bio-Rad, Richmond, CA, United States).

Protein samples were denatured in reducing buffer [62 mm Tris (pH 6.8), 10% glycerol, 2% sodium dodecyl sulfate, 5% β-mercaptoethanol, and 0.003% bromophenol blue] and separated by electrophoresis on an sodium dodecyl sulfate polyacrylamide gel (12%). The separated proteins were transferred onto a PVDF membrane using a transfer buffer [39 mm glycine, 48 mm Tris (pH 8.3), and 20% methanol] at 200 mA for 1 hour. The membranes were stained with Ponceau’s (0.005% in 1% acetic acid) to evaluate equal amounts of protein and were blocked with 5% non-fat dry milk in Tris-buffered saline (TBS)-0.1%, washed and incubated with a primary antibody for EGR-1 (1:500; Abcam, Cambridge, United Kingdom) as marker of memory function; BAX, BCL-2, β-actin (1:500; Cell Signaling, Beverly, MA, United States), BCL-XL (1:500, Milpitas, CA, United States). The membranes were then incubated with a secondary antibody peroxidase-conjugated for 1 h at RT and were analyzed by chemiluminescence (KPL, Gaithersburg, United States). Protein signals were quantified by scanning densitometry using a bio-image analysis system (C-DiGit Blot Scanner, with Image Studio software; Lincoln, NE, United States) and the results were expressed as relative integrated intensity compared to controls. β-actin was used to confirm equal protein loading and blotting.

Brains were fixed in 10% buffered formalin, embedded in paraffin, cut in coronal sections (5 μm) and stained with hematoxylin and eosin to observe morphologic changes.

Injured neurons were counted and a six-point scale score was applied to evaluate the histopathologic changes of the gray matter (Kawai and Akira, 2007): 0, no lesion observed; 1, gray matter contained one to five eosinophilic neurons; 2, gray matter contained five to 10 eosinophilic neurons; 3, gray matter contained more than 10 eosinophilic neurons; 4, small infarction (less than one third of the gray matter area); 5, moderate infarction (one third to one half of the gray matter area); 6, large infarction (more than half of the gray matter area). An average of the scores for each group was recorded to give a cumulative score. At least three slides for each animal were examined by a blinded pathologist.

All data are expressed as means ± SD. Comparisons between different treatments were analyzed by one-way or two-way ANOVA for non-parametric variables with Tukey’s post-test for intergroup comparisons. Mann–Withney U test was used to analyze the histological scores and preference index. The possibility of error was set at p < 0.05 and it was considered statistically significant. All analyses were performed using Stata/IC 12.0 (StataCorp LP, College Station, TX, United States). Graphs were drawn using GraphPad Prism (version 5.0 for Windows).

The Novel Object Recognition was performed to study the differences in cognitive performance between WT and NLRP3^-/- mice. WT-TBI mice had reduced memory ability at 24 h (Figure 1A, p < 0.0001), while WT-TBI+BAY 11-7082 and NLRP3^-/- mice showed better performance results (Figure 1A, p < 0.005).

To confirm if this data could be related to a memory preservation, a Western blot analysis was carried out to evaluate EGR1 expression in hippocampus samples at 24 hours. TBI-WT mice had significantly reduced EGR-1 expression as compared to Sham WT and Sham KO mice (p < 0.0001). BAY 11-7082 administration significantly improved EGR-1 in hippocampal tissue as compared to either WT-TBI and Sham WT animals (p < 0.0001). NLRP3^-/- mice had increased EGR-1 levels following TBI as compared to Sham KO (p < 0.0005) and TBI-WT (p < 0.01) animals, thus suggesting an improvement in the initiation of mechanisms underlying consolidation of memory (Figure 1B, F = 297.2).

IL-1b (Figure 2A) and caspase-1 (Figure 2B) levels were increased 24 h following TBI in brain lysates of WT mice, compared to Sham-WT (p < 0.0001). BAY 11-7082 administration in WT-TBI mice caused a marked decrease of both IL-1b and caspase-1 (p < 0.001), thus confirming its role as NLRP3 inhibitor. NLRP3^-/- mice showed almost undetectable levels of both IL-1b and caspase-1 and a not significant increase was observed following TBI (Figure 2A, F = 137.5; and Figure 2B, F = 116.2). These data indicate that NLRP3 is activated during TBI in WT mice, whereas the lack of NLRP3 in TBI-NLRP3^-/- mice and NLRP3 blockade is protective, reducing in turn inflammation and ameliorating the prognosis of the injury.

At 24 h post-TBI the damage was evident in the cortex of both WT (p < 0.0001) and NLRP3^-/- (p < 0.0005) mice, compared to sham animals (Figures 3, 4). TBI-WT mice showed histological alterations with scattered axonal injury, perilesional infarcted areas, disorganized tissue a strong inflammatory infiltrate and oedema (Figures 3, 4). At hippocampal level neuronal loss was evident in the CA1 and CA3 with neuronal shrinking and eosinophilia (Figures 3, 4). The use of BAY 11-7082 as NLRP3 blocking agent resulted in efficient reduction of inflammatory infiltrate and damage at both cortex and hippocampus level (Figures 3, 4K), as compared to TBI-WT animals (p < 0.01).

TBI-NLRP3^-/- animals had decreased (p < 0.01) inflammatory infiltrate and oedema compared to TBI-WT mice (Figures 3, 4) with more preserved neurons, especially in the hippocampus (Figures 3, 4), suggesting that mice lacking inflammasome might be protected during TBI, at least in the first 24 h.

At day 7 post-injury the impact point was still evident in both strain of mice; WT animals showed an altered structure of nerve fibers and diffused oedema (Figure 5). In the hippocampal areas shrunken neurons and vacuolar degeneration were suggestive of extensive damage (Figure 5). In TBI-NLRP3^-/- mice the cortex architecture was more conserved with a strong decrease (p < 0.01) in both oedema and inflammation in the perilesional area. In the hippocampus the reduced eosinophilia and the almost complete absence of shrunken neurons encourages the hypothesis that NLRP3 inflammasome absence might be protective for the relatively long-term sequelae of TBI.

The pro-apoptotic BAX protein was significantly expressed in hippocampal lysate from WT-TBI mice compared to Sham WT animals (p < 0.0001). NLRP3^-/- mice subjected to TBI demonstrated a reduced activation of BAX compared to WT-TBI (p < 0.01), suggesting that the lack of the inflammasome lead to a blunted apoptotic activation (Figure 6A, F = 187.32). On the other hand, the anti-apoptotic proteins Bcl-2 and BCL-XL were markedly reduced in both WT (p < 0.0001) and NLRP3^-/- (p < 0.005) animals subjected to TBI as compared to the relative Sham mice (Figure 6B, F = 30.94; Figure 6C, F = 22.35).