Animal experiments were authorized by the Animal Care and Use Committee of Xi’an Peihua University, China. Male C57BL/6 mice were fed in the mouse-feeding facility with a 12-h/12-h light–dark cycle, with free foraging and activity.

The TBI model was established by a controlled cortical impact (CCI) device (Hatteras Instruments Inc., Cary, NC, USA). After anesthetized with 6% chloral hydrate (intraperitoneal injection), mice were placed on a brain stereotaxis instrument (Beijing Youchengjiaye Biotechnology Company, Limited, Beijing, China). The heads of the mice were wiped with 70% ethanol, the skin was cut at the midline, and a hole with 0.5 mm diameter in the right skulls, 2.0 mm posterior to the Bregma and 2.0 mm lateral to the sagittal suture, was drilled. The CCI device was used to induce the controllable TBI model with a general parameter (depth: 2 mm; velocity: 5.0 m/s; dwell time: 100 ms). After the injury, mice were immediately administered an intracerebroventricular injection of NEK7-shRNA virus or not. NEK7 (Locus ID 59125) Mouse shRNA (targeting sequence: CATTCTCGAAGAGTCATGCACAGAGATAT, Cat#TL515241, OriGene Technologies, Beijing, China) and a negative control (NC) pGFP-C-shLenti (Cat#TR30021V, OriGene Technologies) were prepared following the manufacturer’s instructions. The right lateral ventricles (depth: 3.0 mm) were injected with 2 μl (targeting NEK7 or NC) at 1 μl/min; the needle remained in place for 2 min and was slowly recovered once the injection was complete. Immediately after surgery, skull was sealed and the incision was sutured. The mice were placed on a heat pad until they regained consciousness and recovered gross locomotor function. After that, the mice were put back into normal feeding units and monitored. The sham group was subjected to normal surgical procedures but not CCI, and to virus via intracerebroventricular injection.

The experimenters were unaware of which mice received TBI or sham treatment. According to the previous reports (Liu et al., 2018), the modified neurological severity scoring (mNSS), Rotarod, and open-field tests were performed to evaluate neurological deficits.

As previously described (Kuriakose and Kanneganti, 2017; Chen et al., 2019), brain water content was assessed in 3-mm coronal sections of the ipsilateral cortex (or corresponding contralateral cortex), centered upon the impact site. Tissues were immediately weighed (wet weight) and then dehydrated at 100°C for 24 h to obtain the dry weight. Water content of the brain was calculated using the following formula: [(wet weight − dry weight)/wet weight] × 100.

Primary cortical neurons were prepared as previously described (Liu et al., 2018). Briefly, cerebral cortices from embryonic C57BL/6 mice (day 18) were used for culture. Cortical tissue was obtained and digested using papayotin. Cells were collected after termination of digestion and cultured on poly-L-lysine-coated dishes with the DMEM medium, which was then replaced with the neuronal base medium (supplemented with 2% B27, 0.2 mM L-glutamine) overnight. The medium was changed every 3 days, and cells were used for subsequent experiments at 7–14 days.

Neurons were cultured for 10 days and then treated with lipopolysaccharide (LPS) or LPS + ATP: neurons were incubated with 10 ng/ml LPS (Sigma-Aldrich, St. Louis, MO, USA) for 4 h and 5 mM ATP (Sigma-Aldrich) for an additional 30 min. Cells or supernatant was then harvested for the next assays.

NEK7 (Locus ID 59125) mouse shRNA (Cat#TL515241, OriGene Technologies) was used for NEK7 knock-down in primary cortical neurons, and pGFP-C-shLenti (Cat#TR30021V, OriGene Technologies) was used as an NC. NEK7 knock-down was performed following the manufacturer’s instructions. Knock-down efficiency was analyzed using a western blot (WB) analysis with anti-NEK7 antibody (Abcam, Cambridge, UK; ab133514).

As previously described (Rühl and Broz, 2015), cells were treated with LPS and washed three times with the leaner (4.2 mM Na₂CO₃, 0.8 mM Na₂HPO₄, 1.3 mM CaCl₂, 0.5 mM MgCl₂, and 10 mM D-glucose monohydrate, pH 7.4) supplemented with 137 mM NaCl and 5 mM choline chloride. The control was washed with the leaner supplemented with 137 mM NaCl and 5 mM KCl. Cells were then incubated for 3 h in the respective buffers and then inflammasome activation and K⁺ concentrations were conducted.

According to Rühl and Broz (2015), cells were primed with LPS or LPS + ATP and washed three times with the leaner. On the basis of the leaner, the solutions with different K⁺ concentrations (5, 10, 20, and 40 mM) stimulated each group of cells. After incubation for 3 h, inflammasome activation and K⁺ concentrations were conducted.

Intracellular K⁺ detection was conducted as previously described (Rühl and Broz, 2015). Briefly, cells were added to 96-well plates and then lysed with ultrapure 3% HNO₃ after stimulation. Intracellular K⁺ was then determined by inductively coupled plasma optical emission spectrometry using an Optima 2000 DV spectrometer (PerkinElmer Life Sciences, Waltham, MA, USA) using yttrium as an internal standard.

IL-1β, IL-18, and TNF-α were detected using the enzyme-linked immunosorbent assay (ELISA) kit (Anoric-Bio, Wuhan, China). Lactate dehydrogenase (LDH) was measured using a kit (Solarbio, Beijing, China). Syringe filters (0.2 μm) were used for filtering the supernatant to measure LDH secretion. The reaction mixture was added to the supernatant and incubated for 30 min. OD₄₅₀ was measured using a microplate reader (Thermo Scientific, Waltham, MA, USA).

TUNEL/NeuN stained as previously described (Chen et al., 2019), and brain tissues were obtained after transcardial perfusion, fixed with 4% paraformaldehyde, dehydrated, and embedded in paraffin. Paraffin coronal sections of thickness 4 μm (proximal to and located in the injury site) were stained using a DeadEnd™ fluorometric TUNEL system (Promega, Madison, WI, USA), followed by incubation overnight with primary anti-NeuN antibodies (Abcam, 1:400) and then incubated with secondary antibodies. Cell nuclei were stained with DAPI (Sigma-Aldrich). Images were obtained using a fluorescence microscope (Olympus, Osaka, Japan). NLRP3/caspase-1 staining was performed as the normal immunofluorescence operation, which was incubated with primary anti-NRLP3 antibodies (Abcam, 1:200) and caspase-1 antibodies (Abcam, 1:200).

After fixation and permeabilization, the different treatment samples were blocked with 5% goat serum. Cells were incubated with anti-ASC antibody (#67824; CST, Danvers, MA, USA; 1:800) and secondary antibody (Alexa Fluor 594-conjugated, Abcam, 1:1,000). Nuclei were stained with DAPI. Images were captured using an inverted fluorescence microscope (Leica, Oskar-Barnack, Germany; IX71).

The total protein of cortical tissue proximal to or located in the injury site was lysed, and a BCA protein kit (Thermo Scientific) was used to quantify protein concentration. Following SDS-PAGE electrophoresis and western transfer, a PVDF membrane (Millipore, Billerica, MA, USA) was blocked with 5% BSA (Sigma-Aldrich) and incubated at 4°C overnight with the primary antibodies: NEK7 (Abcam, 1:1,000), NLRP3 (Abcam, 1:1,000), caspase-1 (Abcam, 1:1,000), caspase-1 (p20; AdipoGen, 1:1,000), ASC (CST, 1:1,200), GSDMD (Abcam, 1:1,000), IL-1β (R&D Systems, Minneapolis, MN, USA; 1:1,000), and GAPDH (CST, 1:10,000). Samples were then incubated with secondary antibodies (Abgent, 1:30,000), followed by chemiluminescent substrate development (Bio-Rad, Hercules, CA, USA), and detection using an imaging system (Bio-Rad).

Syringe filters (0.2 μm) were used for filtering the supernatant for the measurement of protein secretion, and Centricon Plus-20 centrifugal filter devices (Millipore) were used to concentrate the samples. Concentrated samples were separated with 15% SDS-PAGE and evaluated by WB analysis, as above.

The ASC oligomerization assay was performed as previously described (Shi et al., 2016). Cells were lysed using lysis buffer [0.5% Triton X-100, 150 mM NaCl, and 20 mM Hepes (pH 7.5)]. After centrifugation at 6,000 g at 4°C for 15 min, supernatant and pellets were used as the Triton-insoluble and Triton-soluble fractions, respectively. For the detection of ASC oligomerization, crude pellet was re-suspended in 300 μl of TBS buffer, and chemically cross-linked with 4 mM non-cleavable disuccinimidyl suberate (DSS) cross-linker for 30 min at 37°C and then centrifuged for 15 min at 6,000 g. After dissolving in SDS sample buffer, the samples were used for WB analysis.

The total RNA of cortical tissue proximal to or located in the injury site was isolated using Trizol reagent (Invitrogen, Waltham, MA, USA). A HiFi-MMLV cDNA First Strand Synthesis Kit (CW Bio, Beijing, China) was used for reverse transcription. GoTaq qPCR Master Mix (Promega) was used for qRT-PCR analysis. Primer sequences were as follows: NEK7: F 5′-CCGTTACTCAGTTCCAGCCA-3′, R 5′-CTACCGGCACTCCATCCAAG-3′; NLRP3: F 5′-GCTAAGAAGGACCAGCCAGAGT-3′, R 5′-GAACCTGCTTCTCACATGTCGT-3′; ASC: F 5′-TGCTTAGAGACATGGGCTTAC-3′, R 5′-CTGTCCTTCAGTCAGCACACT-3′; caspase-1: F 5′-GACAAGGCACGGGACCTATGT-3′, R 5′-CAGTCAGTCCTGGAAATGTGC-3′; GAPDH: F 5′-AACTTTGGCATTGTGGAAGG-3′, R 5′-GGATGCAGGGATGATGTTCT-3′.

Samples were lysed using a Pierce IP lysis buffer (Thermo Scientific) with proteinase inhibitor cocktail (Roche, Basel, Switzerland). After centrifugation with 4°C, 2,500 g for 5 min, suspensions were collected. Co-immunoprecipitation (CoIP) was performed using a Dynabeads™ CoIP Kit (Thermo Scientific), anti-NEK7 (Abcam, ab133514) or anti-NLRP3 (Abcam, ab214185) antibodies were coupled to dynabeads, and IgG served as the NC. Experimental operation was according to our previous study (Chen et al., 2019), and total proteins were mixed with antibody-coupled dynabeads and incubated overnight at 4°C. Adsorbed dynabeads were washed with wash buffer, and bound proteins were eluted with 20 μl of eluent, mixed with 20 μl of 2 × Laemmli buffer, and then boiled for 5 min. Finally, samples were loaded onto 4%–20% BeyoGel™ Plus PAGE (Beyotime, Shanghai, China) for electrophoresis.

Data are shown as means ± SEM. GraphPad Prism 7.0. software was used to perform the Student’s t-test or a one-way ANOVA followed by Tukey’s multiple comparison test. P < 0.05 was considered significant.

To evaluate whether NEK7 and NLRP3 inflammasomes were involved in TBI, the levels of NEK7 and NLRP3 inflammasomes in the peritraumatic cortex at different times were measured. As shown in Figures 1A,B, gene levels and protein expression were quantified at 2, 4, 8, 12, 24, 48, 72, 120, and 168 h post-TBI. NEK7 mRNA levels immediately increased within the first 2 h (P < 0.05) and reached the peak value at 24 h, and then remained at high levels until 168 h after TBI (P < 0.001). Levels of the proteins NEK7, NLRP3, and caspase-1 p20 were distinctly increased at 24 h; these high levels continued until 168 h post-TBI and following gradually decreased with time, and ASC and pro-caspase-1 levels appeared to increase slightly after TBI (Figure 1B). The IL-1β, IL-18, and TNF-α were analyzed (Figure 1C). IL-1β and IL-18 were immediately triggered at 2 h, TNF-α increased significantly by 4 h, and they all mildly decreased with time but still remained at high levels during 168 h post-TBI. These data indicate that NEK7 and NLRP3 inflammasomes show time dependence during the acute phase post-TBI.

Prior to surgery, there were no differences in mNSS, Rotarod, or open-field scores in each group. In the mNSS test (Figure 2A), the sham group had low scores and showed no differences at any time point, whereas TBI induced significantly higher mNSS at 1 day followed by a gradual decline over time in the TBI group relative to the sham group (P < 0.05). The scores remained with no difference between the TBI group and the TBI + NC group post-TBI. The TBI + NEK7-shRNA group exhibited less mNSS compared with the TBI + NC group (P < 0.05). In the Rotarod test, the sham group showed the best performance compared with the TBI group at 48 h post-TBI (P < 0.05; Figure 2B). The scores were almost the same between the TBI group and the TBI + NC group. However, the scores were higher in the TBI + NEK7-shRNA group compared with the TBI + NC group (P < 0.05). In the open-field test, TBI mice moved less in the perimeter zone and less total distance (P < 0.05; Figure 2C). There were modest yet statistically significant differences in total distance traveled and perimeter zones between the TBI + NEK7-shRNA group and the TBI + NC group (P < 0.05). There was indifference between the TBI and TBI + NC groups. The brains of the mice were assessed post mortem. TBI caused a high water content of about 79.97%, while NEK7 knock-down distinctly inhibited edema of the brain (P < 0.05, vs. the TBI + NC group). However, there was indifference in edema between the TBI + NC group compared with the TBI group (Figure 2D). TUNEL staining was used to demonstrate cortical damage to coronal sections (Figure 3A). Lots of cortical neurons were stained with green fluorescence, revealing serious nerve damage post-TBI. TUNEL-positive neurons were reduced in NEK7-shRNA mice, and the TBI + NC group showed the same damage as the TBI group (Figure 3). As shown in Figure 3B, TBI induced the interaction with NLRP3 and caspase-1 and increased NLRP3/caspase-1 tagged cell, while NEK7-shRNA suppressed the interaction with NLRP3 and caspase-1 after TBI.

As shown in the results above, TBI induced NLRP3 inflammasome activation and increased inflammation at 48 h post-TBI. The expression of NLRP3 inflammasome-related molecules were also detected in TBI + NC and TBI + NEK7-shRNA mice. Following NEK7 knock-down, NLRP3 and caspase-1 mRNA was significantly reduced (P < 0.001), but no significant alteration in ASC was observed (Figure 4A). NEK7 knock-down reversed the secretion of IL-1β and IL-18 (Figure 4B, P < 0.01), but TNF-α was unaffected in NEK7-shRNA mice (Figure 4B). Figures 4C,D showed that NLRP3, ASC, GSDMD, pro-caspase-1, caspase-1 p20, pro-IL-1β, and IL-1β p17 were suppressed in NEK7-shRNA mice post-TBI, with the reversals of GSDMD, caspase-1 p20, and IL-1β p17 being the most significant, compared with the TBI group (P < 0.001). This suggests that NEK7 down-regulation blocks NLRP3 inflammasome activation and its downstream activity.

When reconstituted with NEK7-shRNA, primary cortical neurons lacked NEK7 expression, but showed normal expression of NLRP3, caspase-1, and ASC (Figure 5A). LPS induced NLRP3, pro-caspase-1, and pro-IL-1β expression, but the activation of caspase-1 and IL-1β release was just triggered in LPS + ATP groups. NEK7-shRNA abolished the secretion of caspase-1 p20 and IL-1β p17 in LPS + ATP-induced neurons (Figure 5A). To demonstrate that NEK7 was critical for the assembly of NLRP3 inflammasomes, CoIP analysis was performed (Figure 5B). Interestingly, LPS alone induced weak interactions between NLRP3 and caspase-1 and ASC, but not caspase-1 activation, and that was consistent with the study of Shi et al. (2016). LPS + ATP triggered NLRP3 complexes, while NEK7 down-regulation reduced the interaction between NLRP3 and NEK7, suppressing the assembly of NLRP3 inflammasomes, including NLRP3–ASC complexes and pro-caspase-1 recruitment (Figure 5B). As shown in Figure 6, following neuronal stimulation by LPS + ATP, large spherical intracellular ASC speck was rapidly formed in the cytosol. However, the formation of ASC speck was suppressed in NEK7-shRNA neurons (Figure 6A). Also, a failure of ASC oligomerization was detected in NEK7-shRNA neurons compared to WT cells (Figure 6B). ELISA analysis showed that NEK7 down-regulation reversed LPS + ATP-induced secretion of IL-1β (Figure 5C). In addition, intracellular K⁺ levels were analyzed. Although LPS + ATP induced K⁺ efflux, NEK7 knock-down did not inhibit K⁺ efflux in primary cortical neurons (Figure 5D). The data indicate that NEK7 acts as a crucial role in conducting the inflammatory response, downstream of NLRP3 inflammasomes.

To verify that NEK7-mediated NLRP3 activation is downstream of K⁺ efflux, primary cortical neurons were transduced with NC-shRNA or NEK7 shRNA, stimulated with LPS or LPS + ATP, and then treated with K⁺-free medium or high-K⁺ medium. In NEK7-shRNA neurons, K⁺-free medium induced K⁺ efflux (Figure 7A) and IL-1β release (Figure 7B), while K⁺ efflux was unchanged and IL-1β release was reduced. Importantly, K⁺ efflux directly increased up-regulation of NEK7 expression and caspase-1 activation, including the plentiful releasing of caspase-1 p20 into the supernatant (Figure 7C); it also triggered the NLRP3 complex containing NEK7 (Figure 7C). K⁺ efflux did not influence the caspase-1 activation and NEK7–NLRP3 complexes in NEK7 knock-down cells (Figure 7C). It is well known that NEK7 expression is an event downstream of K⁺ efflux. Stimulation of primary cortical neurons with LPS + ATP induced comparable K⁺ efflux and treated with different concentration of high-K⁺ medium to resist K⁺ efflux. Notably, the release of IL-1β (Figure 8A) and LDH (Figure 8B) was blocked by the presence of high-K⁺ medium, which inhibited K⁺ efflux. NEK knock-down caused low levels of IL-1β and LDH to be released under LPS + ATP and high-K⁺ stimulation. LPS + ATP-initiated caspase-1 activation was eliminated by treatment with 50 mM KCl, which inhibited K⁺ efflux (Figure 8C). Moreover, NEK7 expression triggered by LPS + ATP was also reversed in 50 mM K⁺ medium (Figure 8C), and NLRP3–NEK7 complexes were not seen in neurons with 50 mM K⁺ medium (Figure 8C). The data suggest that K⁺ efflux triggers NEK7 and that this is a requirement for the assembly of NLRP3 inflammasomes and activation of caspase-1.

The animal study was reviewed and approved by The Animal Care and Use Committee of Xi’an Peihua University.