Antagonism of Protease-Activated Receptor 4 Protects Against Traumatic Brain Injury by Suppressing Neuroinflammation via Inhibition of Tab2/NF-κB Signaling

Abstract

Traumatic brain injury (TBI) triggers the activation of the endogenous coagulation mechanism, and a large amount of thrombin is released to curb uncontrollable bleeding through thrombin receptors, also known as protease-activated receptors (PARs). However, thrombin is one of the most critical factors in secondary brain injury. Thus, the PARs may be effective targets against hemorrhagic brain injury. Since the PAR1 antagonist has an increased bleeding risk in clinical practice, PAR4 blockade has been suggested as a more promising treatment. Here, we explored the expression pattern of PAR4 in the brain of mice after TBI, and explored the effect and possible mechanism of BMS-986120 (BMS), a novel selective and reversible PAR4 antagonist on secondary brain injury. Treatment with BMS protected against TBI in mice. mRNA-seq analysis, Western blot, and qRT-PCR verification in vitro showed that BMS significantly inhibited thrombin-induced inflammation in astrocytes, and suggested that the Tab2/ERK/NF-κB signaling pathway plays a key role in this process. Our findings provide reliable evidence that blocking PAR4 is a safe and effective intervention for TBI, and suggest that BMS has a potential clinical application in the management of TBI.

Electronic supplementary material

The online version of this article (10.1007/s12264-020-00601-8) contains supplementary material, which is available to authorized users.

Introduction

Traumatic brain injury (TBI) is one of the leading causes of death in people under 45 years of age worldwide, with high morbidity and mortality [1]. Survivors often suffer multiple severe neurological dysfunctions [2]. And the incidence of TBI seems to be on the rise globally [3].

The damage caused by TBI is mainly composed of two parts, primary mechanical damage and secondary damage [4]. These two parts are very different but partially overlap. The primary injury occurs immediately after the trauma, that is, the devastating damage of the structural integrity of the brain caused by mechanical forces [5]. Primary mechanical injury results in the loss of vascular integrity, particularly the destruction of the blood-brain barrier (BBB), resulting in the entry of many toxic factors into the brain [6], which then results in the rapid development of secondary injury such as oxidative stress, inflammation, necrosis, and apoptosis [7–10]. Nothing can be done about the primary mechanical damage, but the gradual progression of secondary damage provides a time window for intervention and treatment. Unfortunately, the mechanisms of secondary injury are not fully understood, and its most appropriate treatment is still controversial [11]. It is important to explore the detailed mechanisms of secondary injury in TBI, in order to find new therapeutic targets.

Destruction of the BBB after TBI causes the extravasation of serum proteases such as thrombin into the parenchyma [12, 13], which is one of the main factors leading to secondary injury. Thrombin works by binding to its receptors, which are also called protease-activated receptors (PARs) and are members of the superfamily of seven transmembrane-domain g-protein-coupled receptors [14]. Four PARs (PAR1-4) are activated by cutting off the extracellular n-terminal receptor. These n-terminal receptors are ligands and initiate intracellular signal transduction by binding to PARs. PAR1, PAR3, and PAR4 are activated by thrombin and PAR2 by trypsin [15]. Previous studies have shown that PAR1 and PAR4 exist in the platelets of humans, non-human primates, and guinea pigs [14, 16], while PAR3 and PAR4 occur in rodents and rabbits [17], but signal only through PAR4 [18]. The activation of PARs in platelets has been studied in numerous reports [19–21]. The PAR1 antagonist Vorapaxar (brand name Zontivity) has been approved by the FDA for anti-thrombotic therapy. It is currently the only drug that antagonizes the activity of thrombin receptors. However, antithrombotic drugs that target the PAR1 receptor often cause bleeding [22, 23], so Vorapaxar cannot be used in patients with hemorrhagic injury. The evidence suggests that PAR4 antagonists could be potential, safe, and effective antithrombotic drugs. In addition to platelets, PAR4 has also been found in the brains of mammals, and a high correlation between thrombin-induced secondary injury and PAR4 has been demonstrated. Also, cerebral ischemia/reperfusion injury is attenuated in PAR4-deficient mice [24]. However, whether PAR4 plays an important role in TBI and its underlying mechanism remain unclear.

In this study, we explored the changes of PAR4 expression after TBI and tested weather blockade of PAR4 (using BMS, a first-in-class and highly effective small molecule antagonist of PAR4) has a neuroprotective effect after TBI. We also explored its potential mechanism in vivo and in vitro, so as to provide new targets and new ideas for future clinical treatment of brain trauma.

Materials and Methods

Animals and Ethics

All experimental procedures were approved by the Ethics Committee of the Fourth Military Medical University. And all experiments were performed according to the guidelines of the National Institutes of Health Guide for the Care and Use of Laboratory Animals. Experimental animals were healthy adult male wild-type C57BL/6J mice (6–8 weeks old) weighing 20–25 g, provided by the Animal Center of the Fourth Military Medical University. For consistency, the mice were raised in a specific pathogen-free center with constant conditions (temperature 25 ± 2°C; humidity 60% ± 5%), a light: dark cycle of 12:12 h, and food and water ad libitum for two weeks before experiments began.

TBI Procedure and Drug Administration

All animals were randomly divided into four groups (n = 6/group): Sham, TBI, TBI+BMS 1 mg/kg, and TBI+BMS 2 mg/kg [25]. A TBI model was established as described previously [26]. The controlled cortical impact (CCI) model is a widely used pre-clinical rodent model in which to study the complex secondary and severity-specific injury response to TBI [27–30]. The CCI method has consistent intra-laboratory outcomes due to precise control of the depth of cortical penetration, dwell time, and velocity of impact. The correlation between the parameters of injury induction and severity of injury has been discussed [31]. Above all, the CCI model is the most commonly used and is well-recognized for inducing both moderate and severe TBI [32]. Briefly, mice suffered TBI injury by a CCI device (68099 Precision Strike, RWD, Shenzhen, Guangzhou, China) after anesthesia with 2% pentobarbital sodium. Then the mice were fixed on a mouse stereotaxic apparatus. A 3-mm square on the right of the skull was abraded (2 mm anterior to 1 mm caudal to bregma and from the midline to 2.50 mm lateral). The exposed cortical surface was impacted perpendicularly by a rounded metal needle tip (diameter, 2.0 mm; velocity, 3 m/s; duration, 0.2 s; depth, 1.5 mm). After impact, the injured region was covered by tissue adhesive (3M). Mice in the Sham group were also anesthetized, and a craniotomy was implemented. However, these mice did not suffer CCI injury. After surgery, the mice were monitored in an incubator (33°C with 35% humidity) for 2 h, and then allowed to recover in their respective cages. Mice were given BMS orally 3 h after TBI with vehicle, a mixture of 0.6% Methocel and 0.9% saline. BMS was administered intragastrically daily until the mice were sacrificed.

Calculation of Lesion Volume

The calculation was as previously described [32]. Twenty-four hours after TBI, the mice were sacrificed and perfused transcardially with saline and 4% PFA. Ten 1-mm-thick slices were cut from olfactory bulb to cerebellum, and the area of the lesion on each slice was quantified using ImageJ. The lesion volume was calculated as: lesion volume = sum of lesion area × distance between sections. All measurements were performed blindly.

Brain Water Content

Brain water content was measured to assess brain edema 24 h after TBI as described previously [33]. Briefly, after cardiac perfusion the mice were decapitated the brain carefully removed and divided into four parts: brain stem, cerebellum, injured hemisphere, and contralateral hemisphere. To obtain the wet weight, each part was weighed immediately after separation. The dry weight was measured after drying in an oven (95°C, 72 h). Tissue water content = [(wet weight − dry weight)/wet weight] × 100%.

Modified Neurological Severity Score (mNSS)

We used the mNSS to evaluate the functional impairment of the mice as described previously [34]. The mNSS score consists of three parts: sensory function, motor function, and reflex scores. According to the corresponding clinical manifestations and severity, neurological function was scored between 0 and 18 (the worse the performance, the higher the score). Twenty-four hours after TBI, the tests were carried out by two investigators blinded to the experimental groups.

Tissue Preparation

Mice were sacrificed 24 h after TBI. Briefly, after anesthesia with an overdose of 2% pentobarbital, the mice were perfused through the heart and fixed with 4% PFA (4°C, overnight). Then, the brain was dehydrated in 10%, 20%, and 30% sucrose and sectioned at 20–25 μm (from −2.0 to +1.0 mm anterior to bregma). Tissue within 0.5 mm of the edge of the impact was identified as lesion area tissue. For all staining, the images were captured from at least three independent experiments on 6 mice.

Immunofluorescence

The selected sections were treated with 0.1% Triton X-100 for 30 min. After permabilizing the cell membranes, 5% goat serum (Gibco, Grand Island, NK, USA) dissolved in PBS was used for blocking for 30 min. Then, the sections from all groups were incubated with first antibodies overnight at 4°C. The antibodies were: goat anti-chicken GFAP (1:200; Invitrogen, Carlsbad, CA, USA), anti-NeuN purified guinea pig polyclonal protein (1:100, Millipore, Temecula, CA, USA), Iba1 monoclonal antibody (GT10312) (1:200; Invitrogen), and PAR4 polyclonal antibody (1:200; Invitrogen). The next day, after rinsing in PBS for 3 × 5 min, the sections were incubated with the corresponding secondary antibodies (25°C, 1 h): Alexa Fluor 594 donkey anti-chicken IgG (Invitrogen), Alexa Fluor 594 donkey anti-guinea pig IgG (Millipore), Alexa Fluor 594 donkey anti-mouse IgG (Invitrogen), and Alexa Fluor 488 donkey anti-rabbit IgG (Invitrogen). After washing, sections were stained with DAPI (Invitrogen). All sections were observed under a microscope (A1 Si confocal, Nikon, Japan) in a blinded manner. The relative position of the field of view is shown in Fig. S1B.

Fluoro-Jade C (F-JC) Staining

F-JC was used to evaluate neurodegeneration in the lesion area 24 h after TBI following the manufacturer’s instructions. The sections were immersed in 80% and 70% ethanol (with 1% NaOH) for 5 and 2 min sequentially, then oxidized with 0.06% KMnO₄ for 10 min. After careful rinsing in ddH₂O for 3 min, the sections were incubated with 0.0001%F-JC (Millipore) for 15 min. After sealing, their images were captured under the microscope. The field of view we chose was all around the injury of the same level in different groups of mice. Finally, we counted the F-JC-positive neurons.

TUNEL

TUNEL staining was used to assess the cell death around the lesion. An In Situ Cell Death Detection Kit (Basel, Switzerland) was used for TUNEL inspection according to the manufacturer’s protocol. Briefly, after incubation with 0.3% H₂O₂ for 30 min the sections were immersed in a solution of 0.25% proteinase K for 30 min at 37°C. After rinsing with ddH₂O, the sections were reacted in the dark (37°C, 1 h). Finally, the nuclei were stained with DAPI (37°C, 10 min). The degree of apoptosis = (TUNEL-positive cells/DAPI-stained cells) × 100%.

Quantitative Real-Time PCR (qRT-PCR)

qRT-PCR was used to measure the relative mRNA expression levels following the procedure of previous studies [35]. We used TRIzol (Invitrogen) to extract total RNA from cell samples. The HiScript II Q RT SuperMix was used for the reverse transcription of RNA. And the qRT-PCR was performed on an iQTM 5 Optical Module Real-Time PCR Detection System (Bio-Rad, Hercules, CA, USA). Then, the level of mRNA was evaluated by Cham QTM SYBR qPCR master mix. The primers used are listed in Table S1. Finally, the 2^-ΔΔCt method was used to obtain relative mRNA levels.

Primary Astrocyte Preparation and Treatment

Primary astrocytes were isolated from the newborn mouse (1–2 days old) mouse brain and cultured as previously described [36]. Briefly, the whole brain except the cerebellum was carefully isolated and cleaned of meninges in re ice-cold Hank’s Mg²⁺ solution (Ca²⁺ and Mg²⁺-free) and digested in 0.25% trypsin for 20 min. Then, the homogenate was re-suspended in DMEM with 10% FBS and cultured in a humidified incubator (5% CO₂/95% air). Penicillin (100 mg/mL) and streptomycin (100 U/mL) were added to the cultivation system. Six days later, to obtain purified astrocytes, the  culture flask was shaken at 220 rpm for 24 h. Astrocytes were cultivated with 50 μmol/L thrombin for 6 h to establish an in vitro model. Next, the thrombin-containing medium was replaced with fresh DMEM. BMS was dissolved in a mixture of 5% DMSO and 95% DMEM, and stored at −20°C. Before use, DMEM was used to dilute the BMS to a final working concentration of 100 nmol/L [37].

Construction of Plasmid and Cell Transfection

The Tab2 silencing plasmid (short hairpin Tab2, shTab2), silencing control plasmid (shCon), PAR4-overexpression plasmid (OE-PAR4), and overexpression control plasmid (Con) were all synthesized by GENE (Shanghai, China). The shRNA sequence of Tab2 is shown in Supplementary Materials 2. Mouse primary astrocytes were transfected with shTab2 to knock down Tab2, and cells were simultaneously transfected with shCtrl as controls. OE-PAR4 and Ctrl were used to overexpress PAR4 in primary astrocytes. After transfection for 48 h using Lipofectamine 2000 (1.0 μL/well, Invitrogen), the cells were collected and used for subsequent experiments.

Western Blot Analysis

The tissue or cell samples were lysed in a mixture of lysis buffer, protease inhibitor, and phosphatase inhibitors for 20 min, and then further homogenized with an ultrasonic homogenizer. After 15 min precipitation, the homogenate was centrifuged at 15,000 rpm for 20 min and the supernatant was preserved. The supernatants were heated at 96°C for 10 min after chelating with loading buffer. After centrifugation, the resulting supernatant was preserved and stored at −80°C. The protein concentrations were measured with a BCA Protein Assay kit (Thermo Scientific, Waltham, MA, USA; UA276918) following the manufacturer’s protocol. SDS-PAGE gels were used to separate the proteins (25 μg total proteins per lane), and then the proteins were transferred from SDS-PAGE gels to PVDF membranes (Millipore, Billerica, MA, USA). After blocking with 2% non-fat milk diluted with TBST, the membranes were soaked in a mixture of TBST with primary antibodies at 4°C overnight. Then the corresponding secondary antibodies (1:10000, as014, Abclonal, Wuhan, China) were used for chemiluminescence detection with a BioRad imaging system (Bio-Rad, Hercules, CA, USA). Each incubation was followed by three 5-min TBST washes. The following primary antibodies were used: NF-κB p65 (D14E12) rabbit mAb (1:1000, #8242, CST); phospho-p44/42 MAPK (Erk1/2) (D13.14.4E) rabbit mAb (1:1000, #4370, CST); p44/42 MAPK (Erk1/2) (137F5) rabbit mAb (1:1000, #4695, CST); SAPK/JNK antibody (1:1000, #9252, CST); phospho-JNK (G9) mouse mAb (1:1000, #9255, CST); p38 MAPK (D13E1) rabbit mAb (1:1000, #8690, CST); phospho-p38 MAPK (D3F9) rabbit mAb (1:1000, #8632, CST); phospho-IKKα/β(16A6) rabbit mAb (1:1000, #2697, CST); TNF-α (D1G2) rabbit mAb (1:1000, #11948, CST); IL-1β (D3U3E) rabbit mAb (1:1000, #12703, CST); TAB2 rabbit pAb (1:1000, A9867, Abclonal); F2RL3(PAR4) rabbit pAb (1:1000, A8471, Abclonal); F2RL3(PAR4) rabbit pAb (1:1000, 25306-1-AP, Proteintech); Bax rabbit pAb (1:1000, gtx32465, Gene Tex); and Bcl2 rabbit pAb (1:1000, gtx100064, Gene Tex).

Statistical Analysis

All data are presented as mean ± SD. The mNSS was analyzed by Kruskal–Wallis one-way analysis and the Student–Newman–Keuls test. Student’s t test, Tukey’s post hoc test and ANOVA analysis were used for paired comparisons and multiple comparisons. P < 0.05 were considered to be statistically significant. All analyses were performed using GraphPad Prism 7.0 software (GraphPad, USA).

Results

PAR4 Expression is Up-Regulated After TBI at the Transcriptional and Translational Levels in Mice

First, we used Western blot (WB) to semi-quantitatively detect changes in PAR4 expression in mouse brain homogenate of the peripheral penumbra area at different time points after TBI. The results showed that PAR4 expression increased gradually in the peripheral penumbra after injury (Fig. 1A, B). More specifically, PAR4 expression doubled at 24 h after trauma and peaked at 72 h (Fig. 1B). In the sham group, PAR4 immunostaining was weak and evenly distributed throughout the cerebral cortex. However, after TBI, PAR4 staining was significantly enhanced in the peripheral penumbra (Fig. 1C, D). And the fluorescence intensity also increased significantly, indicating an increased level of PAR4, consistent with the results of WB. Then, the PAR4 mRNA level was verified by qRT-PCR (Fig. 1E). Compared with the Sham group, the mRNA level of PAR4 increased significantly to 1.5 times the normal level 24 h after injury, and remained at this high level until 72 h. These results suggest that PAR4 is up-regulated after TBI at both the transcriptional and translational levels.

Fig. 1.

PAR4 is up-regulated after TBI from 24 to 72 h. A, B Western blots and analysis of PAR4 expression 24, 48, and 72 h after TBI. C, D Representative immunofluorescence images and density of PAR4 in the mouse brain around the lesion area at 24, 48, and 72 h in the TBI model. Scale bar, 100 μm. E Results of qRT-PCR show that the PAR4 mRNA level is also up-regulated after TBI from 24 to 72 h. Values are represented as the mean ± SD, n = 6 per group. *P < 0.05 and **P < 0.01 vs sham; ^#P < 0.05 and ^##P < 0.01 vs TBI.

Treatment with BMS Alleviates Cerebral Edema, Reduces Lesion Volume, and Improves Neurological Function After TBI in Mice

The molecular structure of BMS is shown in Fig. S1A. Compared with BMS treatment (1 mg/kg) given immediately after injury, BMS treatment given by gavage at 3 h after injury was beneficial to coagulation, as evidenced by clotting time (Fig. S1B). And then we found that the lesion volume in the treatment group was significantly reduced by calculating the hematoma and penumbra areas (Fig. 2A, B). We also assessed the brain edema and neuronal function score in each group 24 h after TBI. Notably, compared with the TBI group, mice in the BMS treatment (1 and 2 mg/kg) groups showed alleviated brain edema (Fig. 2C). Furthermore, reduced mNSS scores suggested that BMS treatment also reduced the neurological damage (Fig. 2D). The results of these experiments suggest that BMS treatment has some neuroprotective effects against TBI in mice.

Fig. 2.

Effects of BMS on lesion volume, brain water content, and neurological function scores in mice 24 h after TBI. A Typical images of brain slices from mice in each group. The dotted line indicates the lesion area. B Statistical analysis of lesion volume in brain slices using ImageJ. C Brain water content in each group. D Modified neurological severity scores in each group. Values are represented as the mean ± SD, n = 6 per group. *P < 0.05 and **P < 0.01 vs sham; ^#P < 0.05 and ^##P < 0.01 vs TBI.

BMS-986120 Ameliorates Neuronal Degeneration and Apoptosis

TUNEL and F-JC staining were used to evaluate weather BMS treatment mitigates neural damage 24 h after TBI. The results showed that TUNEL staining was visibly enhanced after TBI, and this was reversed by BMS treatment, showing that BMS significantly alleviates the apoptosis induced by TBI (Fig. 3A, B). F-JC staining was used to detect degenerated neurons: F-JC-positive neurons stained green. As shown in Fig. 3C, D, the number of F-JC-positive cells increased significantly after TBI, but with additional BMS treatment the number decreased - fewer degenerated neurons. These results suggested that BMS significantly ameliorates the TBI-induced neuronal damage in mice.

Fig. 3.

BMS attenuates neural degeneration and apoptosis in mice 24 h after TBI. A, B Representative F-JC staining images (A) and analysis of F-JC-positive neurons in each group (B). Scale bar, 200 μm. C, D Representative TUNEL staining images (C) and analysis of TUNEL-positive cells in each group (D). Scale bar, 100 μm. Values are represented as the mean ± SD. n = 6 for each group. *P < 0.05 and **P < 0.01 vs. the sham group, ^#P < 0.05 and ^##P < 0.01 vs. the TBI +vehicle group.

The Effect of BMS-986120 Treatment is Associated with the NF-κB Signaling Pathway in Primary Astrocytes

PAR4 is widely and evenly expressed in the normal cerebral cortex, but its expression has not been well defined after TBI. We found that the expression of PAR4 after TBI was differentially up-regulated in cortical neurons, microglia, and astrocyte, as shown by immunofluorescence staining (Fig. 4A). The number of astrocytes with co-localized PAR4 and GFAP+ was significantly greater than NeuN+ neurons and Iba1+ microglia, which implied that PAR4 is more closely related to astrocyte function after TBI. As BMS acts by binding to PAR4, we conclude the effect of this drug on astrocytes is more marked than on neurons and microglia.

Fig. 4.

Analysis of mRNA-Seq result shows that the effect of BMS treatment is related to the NF-κB signaling pathway in primary astrocytes. A PAR4 is differentially expressed in the mouse brain after TBI, and its co-localization with GFAP is more evident than with NeuN and Iba1. B Volcano plots showing differentially-expressed genes (red) in Con vs TBI and TBI vs BMS. C Venn diagram showing the overlap in differential gene expression between groups. Ninety-seven differentially expressed genes were identified among the groups. D STRING networks of differentially-expressed genes that overlap between groups (disconnected nodes are hidden.) indicated two clear clusters: inflammation-related nodes (red and pink) based on correlation, and mitochondrial respiration-related nodes (green). E KEGG analysis showing that the genes located at the core node are highly correlated with the NF-κB signaling pathway. F Heat map showing the changing trend in the expression of inflammation- and mitochondrial respiration-related genes in each group. G Validation of the changing trend in the expression of inflammation-related genes by qPCR. Values are represented as the mean ± SD, n = 6 per group. *P < 0.05 and **P < 0.01 vs sham; ^#P < 0.05 and ^##P < 0.01 vs TBI.

To elucidate the effect of BMS on astrocytes, we isolated and cultured primary astrocytes from mice. Based on previous studies, we selected BMS concentrations (100 and 150 nmol/L) that did not show appreciable cytotoxicity [25], We chose 100 nmol/L BMS as the working concentration in the following experiment. Then thrombin was used to simulate astrocytes after TBI.

In order to further investigate how BMS inhibits the thrombin-mediated activation of astrocytes, we constructed profiles of astrocyte transcriptomics changes of Con, thrombin-treated, and thrombin with BMS groups using RNA-seq. In brief, total RNA extracted from each group of primary astrocytes was used to establish a cDNA library. Then an Illumina Hiseq2500 high-throughput sequencing platform was used for RNA-seq analysis. Afterward we used DEseq2 analysis to find differentially-expressed genes, many of which were detected (Fig. 4B). We found that, compared with the normal group, the expression of 3,258 genes changed significantly after thrombin treatment, while 387 differed between the thrombin and thrombin-BMS treatment groups. We identified 186 differentially expressed genes that overlapped in both comparisons (Fig. 4C). Proteins are the ultimate executors of biological functions, so we used STRING interaction analysis and GO/KEGG enrichment analysis to investigate which among these 186 genes were critical, and found that inflammation-related genes such as Tab2, Plau, and CD40 were located at crucial nodes in the network (Fig. 4D; disconnected nodes are hidden), and enriched in the NF-κB signaling pathway (Fig. 4E). Then, a heat map showing the expression of these pivotal genes was generated to analyze the variation of each gene among different treatments (Fig. 4F). The results suggest that the inflammatory response is involved in the protection by BMS against TBI injury. BMS significantly attenuated the activated inflammatory response at the transcriptional level. Then, qRT-PCR of each gene was used to verify the sequencing results (Fig. 5E). Also, we found that the mRNA levels of genes associated with mitochondrial function (Fig. 4D, green) changed in the thrombin treatment group, and greatly improved after BMS treatment (Fig. SlD). ATP production also recovered after treatment (Fig. S1E), suggesting that BMS treatment protects mitochondrial function. These results indicated that the protective effect of BMS treatment against TBI is associated with the inhibition of inflammation derived from astrocytes, especially through the NF-κB signaling pathway.

Fig. 5.

Additional treatment with BMS attenuates inflammation after TBI. A, B Representative western blots of the inflammatory factors IL-1β and TNF-α in vivo (A) and in vitro (B). Values are represented as the mean ± SD, n = 6 per group. *P < 0.05 and **P < 0.01 vs sham; ^#P < 0.05 and ^##P < 0.01 vs TBI +vehicle.

BMS-986120 Inhibits TBI-Induced Inflammation In Vivo, and Inhibits Thrombin-Induced Inflammation in Primary Astrocytes

To verify that BMS treatment can reduce the inflammatory response after trauma, we assessed the relative level of the classical inflammatory mediators TNF-α and IL-1β in vivo and in vitro. As shown in Fig. 5A, in the BMS treatment group, the WB results showed decreased production of TNF-α and IL-1β in thrombin-induced primary astrocytes, and Fig. 5B shows the results in vivo. These experiments indicated that BMS has broad inhibitory effects on the production of inflammatory cytokines both in vivo and in vitro.

BMS-986120 Inhibits Activation of the NF-κB Signaling Pathway in Astrocytes Induced by Thrombin Through Tab2/ERK/ NF-κB Signaling

Sequencing results suggested that the NF-κB signaling pathway, especially Tab2-related pathways, were highly correlated with BMS treatment. NF-κB is known to be a major regulator of several inflammatory mediators [38, 39]. In order to determine whether BMS treatment affects the NF-κB signaling pathway in vitro after thrombin treatment, we investigated their expression after treatment. Compared with controls, the expression of nuc-NF-κB in the thrombin treatment group was indeed significantly increased. And this increase was reversed by BMS treatment (Fig. 6A). We also confirmed that the expression of Tab2 was significantly increased after thrombin treatment, and this increase was reversed by additional BMS treatment, as evidenced by WB (Fig. 6A). By using shRNA, we efficiently and specifically degraded intracellular Tab2, thereby blocking its expression in primary astrocytes. Compared with normal primary astrocytes, nuc-NF-κB was significantly reduced after thrombin treatment in cells with knockdown of Tab2 (Fig 6B). These results confirmed that BMS treatment is related to NF-κB signaling pathway, and was likely to act through Tab2. In order to demonstrate that the activation of PAR4 is a factor directly affecting the high expression of Tab2 and activation of the NF-κB signaling pathway, we used PAR4-AP and knocked down Tab2 in vitro. The results showed that the expression of Tab2 and nuc-NF-κB were significantly increased in the case of PAR4-AP treatment (Fig. 6C). But in cells with PAR4-AP and Tab2 knockdown, the activation of NF-κB was inhibited (Fig. 6C). To verify the effect of elevated PAR4 expression on Tab2 and NF-κB, we transfected primary astrocytes with PAR4 plasmid. The results of WB indicated that elevated PAR4 expression had no effect on Tab2 and NF- κB (Fig. S1G).

Fig. 6.

Antagonism of PAR4 inhibits thrombin-mediated inflammation through Tab2/ERK/NF-κB signaling in vitro. A Representative western blots and statistics of Tab2 and nuc-NF-κB after thrombin or/and BMS treatment in primary astrocytes. B Representative western blots and statistics of Tab2 and nuc-NF-κB after knocking down Tab2, with additional thrombin treatment. C Representative western blots and statistics of PAR4, Tab2, and nuc-NF-κB after knocking down Tab2 or/and activation of PAR4 by PAR4-AP. D Representative western blots and statistics of P-ERK, ERK, P-JNK, JNK, P-P38, and P38 after thrombin or/and BMS treatment in primary astrocytes. E Representative western blots and statistics of P-IKK, P-ERK, and ERK after knocking down Tab2, with additional thrombin treatment. F Representative western blots and statistics of P-IKK, P-ERK, and ERK after knocking down Tab2 or/and treatment of PAR4-AP. Values are represented as the mean ± SD, n = 6 per group. *P < 0.05 and **P < 0.01 vs sham (in vivo)/vehicle (in vitro); ^#P < 0.05 and ^##P < 0.01 vs TBI (in vivo)/Thrombin (in vitro).

Previous studies have shown that the MAPK signaling pathway is activated after TBI and activates the phosphorylation of NF-κB [40]. We tested the expression pattern of JNK, ERK, and P38, three main subgroups of the MAPK signaling pathway after thrombin treatment [41, 42]. The results showed that the total amount of JNK, ERK, and P38 did not changed significantly, but their phosphorylation levels increased significantly (Fig. 6D), consistent with the results of previous studies. But BMS treatment only reversed the phosphorylation of ERK, and had little effect on the phosphorylation levels of JNK and p38 (Fig. 6D) as suggested by RNA-Seq. We also found that the phosphorylation of IKK, the major upstream modulator of the phosphorylation of NF-κB, was up-regulated after thrombin (Fig. 6E) [43]. And after knocking down Tab2, the thrombin-mediated phosphorylation of ERK and IKK were mitigated. PAR4-AP treatment was similar to thrombin treatment (Fig. 6F). We speculate that the series of changes in the Tab2/ERK/NF-κB signaling pathway after thrombin treatment are caused, at least in part, by high expression of PAR4.

Discussion

Our study demonstrated that: (1) After TBI, PAR4 expression in the penumbra of the cerebral cortex increased gradually at both the transcriptional and protein levels within 1 to 3 days in mice, and this increase was more pronounced in astrocytes. (2) We provided the first evidence that BMS, a selective inhibitor of PAR4, is a promising therapeutic drug for TBI, as evidenced by better coagulation function of blood, reduced cerebral lesion volume, decreased brain edema, improved neurological function, and reduced neural apoptosis. (3) mRNA-Seq indicated that the neuroprotective action of BMS is mainly achieved by regulating the inflammatory response of astrocytes, which was then further confirmed in vivo and in vitro. (4) We clarified that the anti-inflammatory properties of BMS rely, at least in part, on inhibition of the Tab2/ERK/NF-κB signaling pathway after TBI.

Considerable evidence indicates that PAR4 is an efficacious antithrombotic target. Previous study has shown that the expression PAR4 increases significantly after cerebral ischemic injury, and knocking out PAR4 attenuates cerebral ischemia/reperfusion injury in mice. Besides the recognized biological properties, PAR4-/- mice show less neuronal apoptosis, oxidative stress, and cerebral microvascular inflammation [24]. But the expression pattern of PAR4, and its role in the secondary injury mechanisms after TBI remain unclear. In the current study, we found that the expression of PAR4 increased significantly after TBI in a time-dependent manner from 24 to 72 h, and its expression on perivascular astrocytes was more evident than on neurons or microglial cells. Inhibiting the high expression of PAR4 may reduce the secondary injury of TBI, and may be a intervention target for TBI.

In order to block PAR4 we needed an efficient and specific inhibitor that can penetrate the BBB. Several PAR4 inhibitors, including P4pal-10 pepducin, indazole small-molecule YD-3, ML354, and BMS have been reported [44–47]. BMS is the most specific reversible PAR4 inhibitor with good efficiency [48]. Previous studies have demonstrated that it can significantly inhibit the activity of PAR4 and has an anti-microthrombotic effect. In this research, we found that BMS also had protective effects against TBI as evidenced by decreased brain edema, reduced neural apoptosis, and improved neurological function. Moreover, administration of BMS did not affect bleeding time in mice, indicating that its use does not increase the risk of additional re-bleeding after TBI.

Previous research has confirmed that PAR4 is widely and evenly expressed in the mammalian brain, normally at a low level. Notably, in our study, immunofluorescence indicted that the increased PAR4 mainly co-localized with GFAP. PAR4 on astrocytes increased more than on neurons or microglia after injury. mRNA-Seq analysis showed that inflammation-related genes were negatively regulated by BMS treatment. And follow-up in vitro and in vivo experiments verified the potential activity of BMS to protect against TBI by inhibiting astrocyte-derived inflammation to a large extent. Sequencing results also indicated that the NF-κB signaling pathway was highly involved in BMS treatment. After thrombin treatment, experiments showed clearly increased levels of P-IKK, which directly induced the phosphorylation of NF-κB. Moreover, we found that the amount of NF-κB in the nucleus was significantly increased. These increases were reversed by BMS treatment. Previous studies have shown that the MAPK signaling pathway is activated after TBI, and it is also directly upstream of IKK. We found that the most marked change in the MAPK signaling pathway after BMS treatment was in the ERK pathway, suggesting that BMS might play a protective role through the ERK/NF-κB signaling pathway. The sequencing results also suggested that Tab2 was at the core node in a series of inflammation-related genes, and the protein encoded by this gene is an activator of MAP3K7/TAK1, which is required for the activation of MAPK and NF-κB [49]. After inhibiting Tab2 with shRNA, the thrombin-mediated activation of the ERK/NF-κB pathway was inhibited, further suggesting the important role of Tab2/ERK/NF-κB in inflammatory activation in vitro.

To demonstrate that PAR4 activates NF-κB through the Tab2/ERK pathway, we transfected a plasmid that carries the PAR4 gene, to generate primary astrocytes that strongly express the PAR4 gene. The results showed that overexpression of PAR4 was similar to thrombin treatment, and activated the Tab2/ERK/NF-κB signaling pathway. But this activation was blocked in primary astrocytes treated with Tab2 siRNA. From the above results, it is concluded that the increased expression of PAR4 directly activates the Tab2/ERK/NF-κB signaling pathway.

There are also several limitations in the current study. First, we studied the protective effects of BMS treatment during the early period after TBI, but the long-term effects of this drug are not yet clear. Second, we focused mainly on changes in astrocytes but PAR4 is also expressed in other brain cells, although not as strongly. We have not studied the effect of BMS on microglia and neurons, and it is not clear whether this drug affects intercellular communication between astrocytes and other cells. Third, in the mRNA sequencing section, we found that BMS also had a protective effect on mitochondrial function, but the specific mechanism has not been investigated.

Although BMS is effective in experimental TBI in mice, there is still long way to go before this drug can be transferred to clinical use. Some of the challenges to the clinical transformation of this treatment are as follow: (1) In the clinic, the occurrence of brain trauma is sudden, and the condition is very heterogeneous, such as different trauma sites, different degrees of severity, and different combinations with fractures and injuries of other organs. In our experiments, the TBI mouse model is relatively stable and cannot fully simulate the complexity of actual clinical situations. Whether this drug is clinically effective is a critical challenge, and further studies of its usage and dosage are needed. (2) Oral administration is an advantage of BMS, but on the other hand, it is also a disadvantage for patients with severe trauma and coma. Further studies of injection or fluid administration are needed. (3) For patients with diabetes, coronary heart disease, anemia, and other diseases, it is not clear whether the drug has additional side-effects. The above problems urgently need to be solved before actual transformation. We need to design and implement targeted clinical trials to solve the problems and promote the clinical application of BMS.

In summary, we demonstrated for the first time that blockade of PAR4 with BMS provides neuroprotective effects against TBI in mice by alleviating the activation of inflammation derived from astrocytes. Importantly, we revealed that BMS might be a vital regulator of the Tab2/NF-κB signaling pathway. All in all, our study provides reliable experimental evidence for BMS as a safe and efficient therapeutic agent against TBI, with a high possibility of clinical transformation, although many challenges remain to be conquered.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Conflict of interest

The authors have no conflicts of interest to declare.