Histones of Neutrophil Extracellular Traps (NETs) Activate Brain Pericyte via Dectin-1 in Traumatic Brain Injury

Blood-brain barrier (BBB) disruption and leukocyte inltration are two pathological features post traumatic brain injury (TBI). However, the role of circulating leukocytes in BBB disruption and the crosstalk between them are not fully elucidated. Neutrophil is the most abundant circulating cell type that migrates into brain tissue when TBI occurs instantly, while brain pericyte occupies a strategic position between circulating cell and interstitial space in BBB. Understanding their interactions is essential to provide insight into the intrinsic relationship and identify biological targets for TBI treatments. and “histones-Dectin-1-CD11b” are possible cellular and molecular mechanisms for building connection between BBB damage and neutrophil inltration. Targeting at NETs formation and Dectin-1 are promising treatments for improving neurological outcomes of TBI patients. The c-Jun directly enhanced the promoter activity of CD11b gene. a Putative c-Jun binding sequence of mouse CD11b promoter gene. b Luciferace activity of MBVP co-transfected with indicated reporters under specic conditions. c Luciferace activity of MBVP co-transfected with mutated reporters under specic conditions. Control means mixed normal medium with neutrophil-cultured medium (without PMA stimulation), PMA means mixed normal medium with NETs-formed medium induced by PMA (100ng/ml), Histones means normal medium was added with mixed puried histone peptides (H1, H2 and H3, 2μg/ml). All transfected cells were treated under indicated conditions for 24 hours and lysed for dural-luciferace measurement. d CHIP assays of c-Jun binding sequence from CD11b promoter gene. After treating MBVP with indicated conditions for 24 hours, the total chromatin was collected and amplied as input (positive control). Antibody against c-Jun was used for pulling down the biding segments, of which IgG was introduced as negative control. Two pairs of specic primers that covered each binding site were used to amplify SF1 contained segment (-1250~-1244) and SF2 contained segment (-410~-403) respectively within 30 cycles. e&f Realtime-PCR of CHIP products. The quantitative Realtime-PCR was performed with GAPDH as internal reference gene and regarded input from control. Data are shown as mean±SEM, presented for 3 individual experiments and analyzed by two-tailed unpaired Student’s t test or ANOVA. **p<0.01, *p<0.05. Each 4 mice for group time c Survival rate of each group at the rst week after TBI impact in presence of different treatments. n=6 for Sham, n=10 for TBI. Sham represents the mouse went through the same surgical procedures without impact as control, TBI (saline) represents as placebo group compared with other drug treatment groups. d Footfault evaluation of TBI mice with administration of indicated drugs for 2 weeks. n 5 for each group. Heatmap of mice movement OFT 4 weeks after moderate TBI impact in the presence of different treatments. Quantication of OFT assessment (e) aspects of n g at n=4 each h Quantication of (g). i FACS pericyte (CD45-CD11b+CD13+) at in drugs (Laminarin, 50mg/kg; Cl-Amidine, 50mg/kg). n=4 for each group. j Quantication of (i). Data are presented as mean±SEM, and analyzed by One-way ANOVA. The animal numbers and p values of each group are shown in gures and intended legends.

neurological outcomes of TBI patients have been remarkably improved with utilization of fast and e cient ways toward optimizing physiology [1][2][3]. In contrast to risk factors and primary injuries that are unchangeable when impact occurs, secondary injuries are fundamental pathological processes that contribute to heterogeneous outcomes and medically tractable when patients are under supervised care at hospital. Two of the secondary injury processes activated by TBI are the unbalanced activation of immune cells and the dysfunction of the blood-brain barrier (BBB), which establish the connection between brain injury and subsequent neurodegenerative disorders in a proportion of TBI patients [4,5]. The widespread cellular debris and overwhelming cytokine release after brain mechanical damage activate innate immune system instantly after TBI [1]. The activation of cerebral immune cells (microglia, astrocyte) exerts profound effects on neuroin ammation, leaving the circulating peripheral immune cells underestimated. Neutrophil is the most abundant circulating leukocyte that migrates at the injury sites, and involved in the initiation and development of immunological response [6]. It was previously treated as short-lived cell that eliminated debris and pathogens at injury/infection sites, causing indiscriminative damage to the tissue within acute phase. While the time window that neutrophil plays in TBI might be signi cantly longer than previously thought since several groups reported that neutrophil was detectable in damaged brain 14 days or 1 year after injury/hemorrhage [7][8][9]. The authors speculated that neutrophils in parenchyma were anti-apoptotic after in ltration and migrated into damaged tissue persistently even though the BBB integrity had been restored. This prompts us to rethink the previous view that BBB remodeling after damage might be the cause rather than consequence that facilitates leukocytes in ltrating into brain tissue [10]. Neutrophil possibly reprogrammes BBB functions while transmigrating in unidenti ed manner, making it convenient for the following extravasation of peripheral immune cells (T cell, B cell) post TBI.
Back to 1996, a new form of neutrophil death was described when neutrophil was treated with phorbol 12-myritate 13-acetate (PMA), characterized as decondensed chromatin, perforated membrane and spilled nucleoplasm [11]. This phenomenon was ignored until Brinkmann V et al presented a novel defense way of neutrophil in 2004, named neutrophil extracellular traps (NETs), representing that this structure was disintegrated deoxyribonucleic acid (DNA) coated with histones and granules [12]. NETs are fundamental weapons for neutrophil catching and killing pathogens, as well as contributors to unintended damages [13,14]. In the past two decades, accumulating studies have con rmed the existence of NETs under different pathological conditions, including trauma, infection, cancer and so forth [15][16][17][18]. It is recently reported that NETs contributes to the development of several diseases from central nervous system (CNS) [8,19,20]. Bérézowski V et al showed that neutrophils were present in all brain tissues and 50% of them were NETs positive after autopsy of 14 spontaneous intracerebral hemorrhage (SAH) patients. Both neutrophils and NETs were detected within the haematoma but also in the surrounding tissue [7]. NETs formation markers, including citrullinated histone H3 (CitH3), neutrophilderived DNA segments as well as peptidylarginine deiminase 4 (PAD4) were signi cantly increased after TBI or SAH, and correlated with the severity [20,21]. Vaibhav K et al presented that NETs were key factors to exacerbate TBI severity through aggravating neurovascular injury. Degrading NETs formation by PAD4 inhibitor or deoxyribonuclease-1 signi cantly improved BBB integrity and neurological outcome [21]. Furthermore, Yipp BG et al proved that NETs generally formed during crawling and transmigrating, casted large interacted areas with vascular units/BBB [22]. However, the functional components and molecular mechanisms of NETs that in uence BBB and neurological recovery are still poorly understood due to complicated constituents of NETs. Therefore, nding out potential target cell and molecular mechanism of NETs is important to illustrate the effects on BBB mediated by neutrophil.
The BBB is compact and dense capillary wall that allows few substances going through to the brain, which is made up by endothelial cells, pericytes and astrocytes [5]. Pericytes are surrounded by endothelial basement membrane and astrocytic pseudopodia, forming intensively tight junctions and ne-tuning normal functions of neurovascular unit [23]. Compared to endothelial cells, pericytes are more sensitive and versatile toward in ammatory stimulus, regulating cerebral blood ow dynamics and leukocyte recruitment [24,25]. Coordinated crosstalk between dysregulated pericytes and other inherent cell types (microglia, astrocyte) also pave the foundation for further pathogenesis, such as glial overactivation and neuronal hyperexcitability [25,26]. Interestingly, Proebstl D et al found that pericyte was the accomplice toward neutrophil under in ammatory conditions, losing its guard function and supporting neutrophil subendothelial crawling [27,28]. Based on these, it is tempting to speculate that pericyte could be the potential target cells of NETs, which is involved in neutrophil-mediated BBB dysfunction and neutrophil-related neuroin ammation.
To address this issue, we analyzed brain tissues from TBI patients and mouse TBI model to build the relationship between neutrophils, NETs and brain pericyte. Then, the components of NETs-related medium were deeply investigated by proteomics and metabolomics to decipher the factors directly regulating pericytes. The molecular mechanisms were deeply explored by WB/CHIP/RT-PCR in primary brain pericyte/pericyte cell line MBVP treated with NETs-formed medium or speci c NETs components. In mice TBI model, we explored the possible therapeutic approaches for TBI treatment that targeting at the axis of neutrophil-NETs-pericyte.

Materials And Methods:
TBI patients: All studies were approved by the Institutional Research Ethics Committee of Army Medical University, and written informed consent was obtained from each patient's relatives in Department of Neurosurgery, Daping Hospital (Army Special Medical Center). Brain specimens were collected from acute middle/severe TBI patients (GCS of 3-9, within 24 hours) while decompressive craniectomy and damaged brain tissue resection were necessary for patients to survive regardless for age, race, gender, or socioeconomic status (listed in Table S1). All samples were de-identi ed and coded by the attending surgeon before transport to the laboratory. The specimens were processed and stored according to the Principles of Human Samples Preservations from PRC. TBI mice model: All animal procedures were approved by The Institutional Animal Care and Use Committee at Army Medical University. Adult mixed-sex C57/BL6 mice (age from 6-8 weeks) were provided by Animal Center of Army Medical University and subjected to sham or controlled cortical impact as we did previously [29]. Brie y, mice were anesthetized using pentobarbital sodium (30mg/kg) and craniotomy was made in the left parietal bone midway (anterior-posterior 2 mm, medial-lateral 2 mm from bregma). The exposed cortex was impacted by automatic BSI impact machine (LinTech, Monrovia, CA, USA) with down stroke (velocity: 2.5 m/s, deformation depth: 3 mm, duration: 150 ms) to construct moderate/severe TBI mice model, which was roughly equal to the severity of TBI patients. Sham-operated mice underwent the same anesthetic and surgical procedures without impact. The skin incision was closed by sterile sutures and mice were put back to clean, warm cage to recover. For drug treatment studies, Laminarin (TLRL-LAM, InvivoGen) and Cl-amidine (S8141, Selleck Chem) dissolved by saline was administered via intraperitoneal injection.
Neutrophil isolation and treatment: Murine neutrophils were isolated from bone marrow as we did before [30]. The purity of harvested cells was more than 98% which was con rmed by FACS with combined markers (CD11b + Ly6G + ). Human neutrophils were sorted by Beckman Coulter with speci c markers (CD11b + CD14 -CD15 + ) to identify. These cells were washed with PBS for three times and resuspended in Dulbecco's modi ed Eagle medium (DMEM), containing high glucose and antibiotics without FBS. PMA (100ng/ml, P1585, Sigma-Aldrich) was added to culture medium for 30 min to induce NETs formation. Then the treated cells were washed with DMEM for 3 times and cultured with new DMEM medium (antibiotic and FBS free) for 6 hours. The NETs-formed medium was puri ed with high speed centrifuge (14000rpm, 4°C) for 20 min to remove neutrophil, and freshly prepared for the following experiments. The medium ratio of NETs-formed medium/ new DMEM was 1:3 according to preliminary test ( Figure S4). Under this condition, the medium could induce obvious CD11b expression without strong cytotoxic effects. Mice brain pericyte isolation and culture: Mouse brain capillary pericyte was freshly isolated and cultured as previously with some modi cations [31]. Brie y, the capillaries were seeded by gradual Percoll (P4937, Sigma-Aldrich) centrifugation and cultured with DMEM medium (antibiotic and 10% FBS contained) for 3-4 days until the ask was 70% con uent. Then primary pericyte was harvested by twice digestions: rst digestion was to remove contaminated endothelial cells since the endothelial cells were more easily detached from the ask, second digestion was to collect pericyte for the following steps. The collected pericytes were cultured with DEME medium supplemented with pericyte growth factors (1252, ScienCell Research Labaratories) for 3-5 days until the cells were ready for the next experiments.
Culture of murine brain pericyte cell line: The cell line of mouse brain vascular pericytes (MBVP) was purchased from ScienCell Research Labaratories (M1200, San Diego). MBVP cells were cultured with DMEME at subcon uent density according to the supplier's protocol. Histone peptides of Histone 1 (H1917), Histone 3 (12-357) and Histone 4 (12-347) were all purchased from Sigma-Aldrich and dissolved in ddH 2 O for the following treatments.
Brain tissue preparation and analytic uorescence activated cell sorting (FACS): Mice were sacri ced with CO 2 prior to brain dissection. FACS sorting of brain endothelial cells and brain pericytes was performed according to previous study [32]. In Brief, brain tissue was dissected from mice perfused with saline transcardially and digested with Liberase™ TL Research Grade (05401020001, Sigma Aldrich) and 2 µg/ ml DNase I (104159, Roche) for 30 min at 37°C. Cell suspensions subsequently homogenized with grinder and ltered with sieve (200 mesh) to remove undigested tissue blocks. Myelin and debris were removed by 22% Percoll (P4937, Sigma-Aldrich) centrifugation for 10 min, 560×g at 4°C. Pellets with single vascular cells were carefully collected and resuspended with PBS containing 2% FBS. Single cell suspensions from blood, spleen and brain tissue were stained for 30 min at 4°C with speci c antibodies (Table S2). The diagrammatic procedure of cell sorting was shown in Figure S5. Flow cytometry data were analyzed with FlowJo software version 11.
MPO-DNA binding experiment: After harvesting brain tissues and peripheral blood from TBI patients, the lysate of brain tissue or neutrophil were incubated with peroxidase-labeled DNA primary antibody (1:25, MAB3868, Millipore) for 2 hours at 4°C. Then the incubated lysate was transferred into MPO (1:500, MA1-80878, Invitrogen) pre-coated ELISA 96-well plates for 1 hour. Finally the absorbance at 405nm were measured for quantify the MPO-DNA binding capacity to evaluate the level of NETs formation as detailed [21].
Transmission electron microscopy (TEM): Preparation of brain ultrathin sections were described previously [30]. Brie y, left brain tissues of sham and TBI mice were cut into small pieces (2mm×2mm), xed in 0.2% glutaraldehyde and dehydrated through gradient ethanol incubations. Post-xation was followed with 2% osmium tetroxide and propylene oxide. Then the samples were embedded and allowed to polymerize at 60°C overnight. Ultrathin sections (65-nm-thick) were cut on Leica UC6 ultramicrotome and placed on 200-mesh nickel grids. Finally, sections were washed with distilled water and stained with 2% uranyl acetate and lead citrate. Dried grids were observed and imaged using TECNAI-10 electron microscope.
H&E staining: Mice were anesthetized 24 hours after TBI and transcardially perfused by pre-warmed saline and 4% paraformaldehyde (PFA). Then brain tissues were immediately removed and xed in 4% PFA for 24 hours at 4°C. Para n-embedded sections (10 µm thick) were stained with hematoxylin and eosin (HE) and observed under a microscope (Olympus IX-81; Olympus, Tokyo, Japan).
LC-MS analysis of cell metabolites: A liquid-liquid extraction and liquid chromatography-electrospray ionization tandem mass spectrometry (LC-MS) method was utilized to determining metabolites from cell culture media. Primary isolated murine neutrophils were treated with or without PMA (100ng/ml) for 30 min, then incubated with DMEM (antibiotic and FBS free) for 6 hours. Subsequently, cultured medium were collected by high speed centrifuge and lyophilized for following steps [33]. Each group contains 7 individual samples and differential molecules with p value < 0.05 were screened out.
TMT proteomics analysis of cell cultured medium: Tandem Mass Tags (TMT) labeled quantitative proteomics method was performed to analyze total protein in neutrophil-cultured medium. Proteins with changes greater than 1.5-fold and P < 0.05 were considered differentially expressed, and bioinformatics analysis was performed subsequently as previously [34].
Immunohistochemitry: For tissue sections, mice were perfused with 0.9% NaCl transcardially after anesthetization. Then the brain tissues were immediately frozen and embedded with OCT (4583, Tissue-Tek). Brain sections (10µm) were incubated with primary antibodies (Table S2) diluted by 5% BSA (V900933, Sigma-Aldrich) containing 0.1% Triton X-100 (T8787, Sigma-Aldrich). For cell slices, cells were cultured on poly-L-lysine (P6282, Sigma-Aldrich) coated glass slides and treated with speci c conditions. Subsequently, cell slices were washed with PBS for 3 times and xed with cold 4% PFA for 30 min, following incubation of primary antibodies (Table S2) as tissue sections. Finally, samples were stained with DAPI and intended secondary antibodies and photographed by uorescence microscope (Olympus IX-81).
Quantitative Realtime-PCR: Total RNA extraction of intended cells was achieved by TRIzol Reagent protocol (15596018, Thermo Fisher). RNA samples were reverse transcribed to cDNA using a Reverse Kit (DRR047S, Takara). Quantitative reverse transcription polymerase chain reaction (qRT-PCR) was performed on a Bio-Rad iCycler (Version 3.0A). The sequences of primers were listed in Table S3. Data were presented as fold change versus control group.
Western blotting: Whole-cell lysates were collected and prepared as we did before [30], and western blot procedure was carried out by standard methods. The information of antibodies for speci c proteins were listed in Table S2. TEER measurement: Transwell inserts (polyester membranes, 3µm pore size, ø=6.5mm; Corning, NY, USA) were coated with pericyte cell line MBVP in 24-well plate (10 5 /cm 2 ). The Millicell-ERS (Electrical Resistance System, MERS00002, Millipore) was inserted between upper chamber and lower well while measurement took place. TEER values were obtained from continuous impedance measurements as described previously [35]. Once the TEER value of each insert reached above 200Ω/cm 2 (typically after 2-3 days), cells were treated with NETs-induced medium as described above or combined with other treatments (as in gure legends), and TEER measurements were obtained every two hours. TEER data were obtained as Ω/cm 2 and presented as percentage (%) by comparing to the value of control group at the beginning.
Evans Blue assay: 2% Evans Blue solution (4 mL/kg, E8010, Solarbio) was IV injected after TBI. The stain was allowed to circulate for 30 minutes, 6 or 24 hours. Afterwards, the mice were sacri ced and transcardially perfused with 50ml of ice-cold PBS. The left injured brain hemisphere was removed and weighed. Then brain tissues were homogenized in 1ml PBS, centrifuged for 30 min (14000 rpm at 4°C). The supernatant was collected in aliquots. Equal amount of 50% trichloroacetic acid was added to each 500µl supernatant and incubated over night at 4° C., Supernatant was subsequently centrifuged (30 min, 14000rpm at 4°C) and measured by In nite M200 plate reader (excitation at 620; emission at 680). Free PBS combined with 50% trichloroacetic acid was regarded as blank, and gradient doses of Evans Blue were measured to draw the standard curve. The results were quanti ed according to standard curve and presented as (µg of Evans Blue stain)/(g of brain tissue). Plasmids construction: The 5'-UTR sequence of mouse CD11b gene was achieved using Pubmed (http://www.ncbi.nlm.nih.gov/entrez/). The 5'-UTR sequence of mouse CD11b gene was further analyzed for potential c-Jun (AP-1) binding sides by using web-based algorithm (NUBIScan and hTFtarget) as we did before [36]. Mouse CD11b gene promoter region containing fragment (− 2900 to + 190) was chemically synthesized by Sangon (Shanghai, China) cloned into pGL3-basic vector. All the resulting plasmids were named as pGFs as Fig. 8A shows.
Dual-luciferase reporter assay: Murine brain pericyte cell line MBVP was transiently transfected with reporter plasmids (pGFs) by Lipofectamine 3000 (L3000008, Thermo Fisher). Two hours later the cultured medium was replaced with fresh complete DMEM medium. After 24 hours incubation, NETsformed or normal medium was added with speci c drugs for another 24 hours. Then the cell lysates were collected and luciferase activities were measured by Dual-Luciferase Reporter System (E1910, Promega). The transfection experiments were repeated three times in triplicate, and the transfection e ciency was normalized by dividing re y luciferase activity to renilla luciferase activity.
Chromatin immunoprecipitation assays: Chromatin immunoprecipitation (ChIP) assays were performed with Pierce™ ChIP assay kit (26157, Thermo Fisher) according to the manufacturer's instructions. The targeted protein-DNA mixture was gathered by PureProteome Magnet (LSKMAGS08, Millipore). Then the nal DNA extracts were qPCR or RT-PCR ampli ed by speci c primers (shown in Figure S7). The antibodies against c-Jun and the control IgG were obtained from CST and Beyotime respectively (listed in Table S2).
Neurobehavioral evaluations: Open-eld test: Mice were tested in a square box (40cm × 40cm × 40cm) for 10 min, and activity was digitally recorded as we did before [37]. Travel distance, mean velocity, and time spent in the center zone were analyzed with Ethovision XT video-tracking software (Noldus Information Technology, Asheville, NC). Foot-fault test: The experiments were carried out on an elevated beam (2.5 × 75 cm) raised 76 cm above the oor as previously described [38]. Mice were placed at the beginning of the beam and their movements were manual from the beginning to the end. It was considered an individual foot fault when a mouse paw slipped completely through the beam during its walking. The average foot fault score was calculated from the total steps within 75cm in three separate trials. Y-maze spontaneous alteration test: Mice were tested in a Y-shaped maze with three arms at 120° angle from each other, and allowed to explore the three arms freely for 8 min as previously [39]. The number of arm entries and triads were recorded to calculate the percentage of alteration.
Statistical analysis: All data were analyzed and presented by using GraphPad Prism version 5.01 (GraphPad Software). Unpaired two-tailed Student t test were evaluated for comparison of treated groups with vehicle controls. For analyzing parameters depending on two factors or more, two-way ANOVA/multivariate analysis of variance (MANOVA) was used with Bonferroni correction. A p value < 0.05 was considered statistically signi cant. Data in gures were all represented as mean ± SEM with *p < 0.05; **p < 0.01. The number of animals and experimental repeats were marked in the corresponding gure legends.

Results:
Neutrophil in ltration and NETs formation are associated with the CD11b induction in pericyte after TBI.
The leukocyte in ltration of injured brain tissue is an important pathological process during TBI [40,41]. In TBI patients and mouse model, we observed obvious in ltration of neutrophils in brain tissue within 24 hours (Fig. 1a, CD177; Fig. 1b, CD16). We also found that neutrophils were closely distributed around brain pericytes. This was in consistency with previous report claiming that neutrophil initial in ltration was in close proximity to pericyte [42]. It has been shown that integrin alpha M (CD11b) or intercellular cell adhesion molecule-1 (ICAM-1) is one hallmark of pericyte activation, which are fundamental for pericyte to increase permeability as well as recruit leukocytes [28,43,44]. To examine whether neutrophils/ NETs formation gave rise to pericyte activation, we analyzed the relationship between myeloperoxidase (MPO)-DNA binding capacity (a sensitive measurement of NETs formation) of in ltrated neutrophils and CD11b/ICAM-1 expression levels of brain pericytes from 10 TBI patients. Figure 1c&1d showed that there was signi cant positive correlation between NETs formation and CD11b + brain pericyte. The correlation ratio between NETs level and ICAM-1 expression showed no statistical difference. In mice, we found similar phenomenon that neutrophils were accumulated around morphologically altered pericyte following TBI, alongside with disrupted BBB structure ( Fig. 1e-1g). Moreover, marked with CitH3 and DAPI (combined marker for NETs), we saw decreased expression of pericyte marker PDGFRβ and enhanced level of CitH3 in injured brain tissue (Fig. 1h). Moreover, these website-like DNA structures were not only detectable in injured side, but also in the contralateral side from TBI mice (Fig. 1i&1j). This indicates that neutrophil in ltration and NETs formation are spatially and functionally related with brain pericyte after TBI, in which CD11b expression might be the reason to connect them.
CD11b + pericyte is endowed with pro-in ammatory pro les.
Since we showed that CD11b expression was a character of NETs-activated pericytes in human, we next con rmed the association between NETs and CD11b expression pericyte in mouse TBI model and investigated the characters of these CD11b + pericytes. FACS test presented that CD11b + brain pericyte population was signi cantly enhanced post TBI (Fig. 2a), which was not obvious in brain endothelial cell (Fig. 2b&2c), indicating that CD11b induction was speci c to brain pericyte alongside with NETs formation. After establishing a crucial link between NETs formation and CD11b induction of pericytes, we aimed to investigate the differences between CD11b + and CD11b − brain pericytes at a transcriptomic level. To begin with, we harvested brain pericytes (CD45 − CD13 + ) from brain tissue as previously did [45], and sorted CD11b + and CD11b − brain pericytes respectively for further study (Fig. 2d&S1). After culturing these two pericyte subpopulations with complete DMEM medium for 48 hours, CD11b + pericytes appeared to be round and huddled shape, whilst CD11b − pericytes were prolonged and stretched (Fig. 2e).
NETs directly induce CD11b expression on pericyte to facilitate increased adhesion and integrity damage in vitro.
Apart from induced NETs formation in brain parenchyma, the circulating neutrophils isolated from TBI mice exhibited higher frequency of NETs formation compared with neutrophil from sham group (Fig. 3a&3b), which was in consistency with Vaibhav K's work [21]. In brain tissues from TBI patient, we con rmed the existence of CD11b + brain pericyte in parenchyma by immuno uorescence (Fig. 3c). To detect whether NETs directly activated pericyte via regulating CD11b expression, then we stimulated primary brain pericyte or pericyte cell line MVBP with NETs-formed medium at established conditions (obvious effects without strong cytotoxicity according to Figure S3). After cultured with NETs-induced medium for 24 hours, we found that primary pericytes displayed an activated morphological characters of shrunken shape and round cell body (Fig. 3d) as CD11b + pericyte showed (Fig. 2e). Based on previous study, pericyte activation was associated with various dysfunctions, including decreased tight junction and increased adhesive capacity [24,50]. To further con rm the association between activation pro les and CD11b induction in pericyte, we utilized murine pericyte cell line (MBVP) to evaluate pericyte functions toward NETs-formed medium. Accordingly, NETs-formed medium signi cantly changed cell morphology with sparse cell contacts and elongated shapes (Fig. 4a), decreased tight junction (Fig. 4b) and increased neutrophil adhesion (Fig. 4c). Furthermore, NETs-induced medium strongly promoted formation of CD11b + population (Fig. 4d-4h) and jeopardized integrity (Fig. 4i) accompanied with obvious CD11b induction. All the effects mediated by NETs-induced medium could be reversed by Cl-Amidine (NETs formation inhibitor by decreasing PAD activity), suggesting that NETs formation was associated with pericyte activation alongside with CD11b induction.
Histones are the main components of NETs to drive CD11b induction on pericyte.
As we found that NETs-induced medium signi cantly changed morphology and functions of pericyte in indirect way, we aimed to investigate what components of NETs-formed medium drive this phenomenon. The effects of neutrophil functions in co-culture system are different depending on direct or indirect cell interactions [51]. The indirect interaction means cell communication occurs through releasing various substances without cell contact. Since we treated pericytes without neutrophil direct contact, these effects were mediated by substances released from neutrophil while forming NETs structure. In Fig. 5a, we collected the cultured medium after removing neutrophils with or without PMA stimulation for proteomics and metabolomics analysis. In proteomic analysis, overall 4098 proteins were identi ed in these two groups. Proteins with repeatable fold changes > 1.2 or < 0.8 were screened out as differentially expressed proteins (Fig. 5b&S2). In PMA-stimulated groups, there were 101 up-regulated and 3 downregulated proteins identi ed. Of note, histones accounted for 30% of the top up-regulated proteins and other neutrophil intracellular molecules (lactoferrin, complement 3, and metalloproteinase et al) were also detectable, which was consistent with reported NETs characters [14]. NETs formation induced by PMA caused dramatic alterations in amino acids too, such as N-acetyl-L-cycteine, L-isoleucine, L-glutamate (Fig. 5c&S3). In order to testify their effects on pericyte, we selected some of these proteins (histones and lactoferrin) or metabolites (N-acetyl-L-cycteine and L-isoleucine) to con rm if they could change CD11b expression in vitro. As shown in Fig. 5d-5f, only histones treatments could induce CD11b expression in dose-dependent manner. We did not see obvious alterations ( Figure S5) under intended concentrations of lactoferrin, N-acetyl-L-cycteine and L-isoleucine according to previous studies [52,53]. Recently, several researches have proved that histones released by necrotic cells or NETs were non-negligible factors that drove in ammatory responses [54][55][56]. The increased level of histones and segmented DNA was highly associated with BBB damage and brain edema [8,21], supporting our data that histones were the possible components of NETs to drive CD11b induction in brain pericytes.
Dectin-1 is fundamental for pericyte to respond toward histones of NETs in vitro.
After de ning histones as the main components of NETs to affect pericyte, we aimed to decipher the possible ways that pericyte might respond to. According to previous study, we converted our attention to C-type lectin receptors (CLRs), a large family of transmembrane receptors that recognized not only fungal moieties, but also histone-related molecules from dead cells [54,57]. Three types of CLRs are classi ed based on their molecular structures: type I, type II and soluble type. Of which, Type II CLRs are the main receptors for recognition of histone sequences or histone deacetylase complexes released from disintegrated DNA [54,58]. Type II CLRs carry a conserved carbohydrate-recognition domain and contain ve receptors: Dectin-1, Dectin-2, macrophage-inducible CLR (Mincle), dendritic cell-speci c ICAM3grabbing nonintegrin (DC-SIGN), as well as DC-NK lectin group receptor-1 (DNGR-1) [59]. Based on human protein atlas database, we found that the expressions of type II CLRs in CNS (yellow bars) were relatively low except Dectin-1 and DNGR-1 ( Figure S6a). The protein expression and glycosylation level of Dectin-1was obviously upregulated in pericyte cell line with PMA-induced medium or histones (Fig. 6a&S6b).
Meanwhile, we found that CD11b + brain pericyte was armed with higher expression of Dectin-1 compared to CD11b + population in human samples (Fig. 6b&6c), alongside with increased Dectin-1expression of brain protein after TBI ( Figure S6c). This proved that Dectin-1 expression was positively related to CD11b upregulation in brain pericyte. In order to con rm if Dectin-1 was indispensable for CD11b upregulation, the Dectin-1 siRNAs and speci c antagonist Laminarin (LAM) were introduced in the following experiments. WB showed that synthetic siRNAs was capable to interfere Dectin-1 protein expression in the presence of NETs-medium and histones (Fig. 6d). Furthermore, antagonizing Dectin-1 by LAM as well as blocking Dectin-1 expression by siRNAs decreased CD11b + pericyte population and CD11b expression compared to PMA-medium alone (Fig. 6e-6h). This phenomenon was consistent under histones-treated conditions (Fig. 6i-6l), indicating that histones of NETs were crucial factors to induce CD11 expression of brain pericyte in Dectin-1 dependent manner. TEER experiments also con rmed that targeting at Dectin-1 was bene cial for restoring pericyte integrity in vitro (Fig. 6m&6n). Here, we show that Dectin-1 antagonist LAM and siRNAs are competent to blunt CD11b expression mediated by NETs medium or histones stimulation, con rming that Dectin-1 is indispensable for pericyte to react toward histones of NETs.
PKC and c-Jun activation mediated by Dectin-1 are responsible for CD11b induction of pericyte during NETs and histones treatments. Next, we explored the possible signaling pathways mediated by Dectin-1 that were involved in CD11b induction. After reading previous literature, we sorted out that protein kinase Cζ/λ (PKCζ/λ) and transcriptional factor c-Jun were possible molecules post Dectin-1 activation [59]. The levels of phosphorylated PKC (p-PKCζ/λ) and phosphorylated c-Jun (p-c-Jun), activation status of these two molecules, were signi cantly increased following NETs-formed medium treatment (Fig. 7a). Similar effects were observed in histones treatment group, reiterating the factor that histones were robust to mimic NETs-formed medium. In the presence of Dectin-1 inhibitor LAM and Dectin-1 siRNAs, the phosphorylation levels of PKC and c-Jun dropped dramatically (Fig. 7b&7c), showing that Dectin-1 was essential for PKC and c-Jun activation toward NETs-medium stimulation. PKC speci c antagonist GFX as well as c-Jun inhibitor T-5224 were further used to con rm their roles in Dectin-1 mediated pathway stimulated by NETs-formed medium and histones (Fig. 7d). The Realtime-PCR as well as ow experiments both demonstrated that PKC inhibitor or c-Jun inhibitor blunted CD11b expression mediated by Dectin-1 activation (Fig. 7e-7h). Therefore, Dectin-1 contributes to CD11b upregulation expression in brain pericyte toward NETs-formed medium/ histones through the Dectin-1/PKC/c-Jun signaling pathway.
c-Jun directly binds to the promoter sequence of CD11b to enhance its expression in pericyte.
Transcription factors (TF) bind to the binding site, upstream to the promoter of target gene, to activate or repress gene expression. The data above presented that c-Jun activation, an important TF, contributed to CD11b induction in pericyte, we'd like to explore if c-Jun could activate CD11b transcription by binding to its promoter gene. After searching the possible c-Jun binding sides in CD11b promoter gene region by NUBIScan, we screened out two possible sites with the "TGACTCA" motif (Fig. 8a&S7). Subsequently, three luciferase reporters containing different promoter regions of CD11b gene were constructed, as well as two mutation luciferase reporters that disrupted binding motif. They were named as pGF1 (-2900 to + 190), pGF2 (-1000 to + 190), pGF3 (-200 to + 190), pGFmut1(-1250 to -1244) and pGFmut2 (-410 to -403) respectively (Fig. 8a).Then luciferase reporter assays showed that the activity of pGF1 and pGF2 was signi cantly increased in the presence of NETs-formed medium or histones (Fig. 8b), revealing that these two potential binding sites of CD11b promoter region were capable to promote CD11b gene expression. Of note, the ability of histones to raise CD11b promoter activity was relatively low compared to NETsformed medium (Fig. 8b), indicating that there were other potential components in NETs that induced CD11b expression apart from histones. Moreover, mutating these two side separately both decreased the promoter activity (Fig. 8c), verifying that these two binding sites were required for CD11b induction toward NETs-formed medium or histones treatment in pericyte. Subsequently, ChIP assays further con rmed that c-Jun could bind to these two CD11b promoter regions (-1250 to -1244, -410 to -403) directly (Fig. 8d-8f). Taken together, these results indicate that c-Jun activates CD11b expression by binding to two sites (-1250 to -1244, -410 to -403) in CD11b gene promoter region.

Inhibition of NETs formation or Dectin-1 improves brain function recovery after TBI.
Previous study and our results above con rmed that NETs formation and brain pericyte dysfunction contributed to acute TBI pathogenesis [21]. Mechanistically, we proved that histones of NETs facilitated CD11b expression in Dectin-1/PKC/c-Jun pathways, leading to pro-in ammatory pro les and damaged integrity of brain pericyte. It had been shown that blocking Dectin-1 was bene cial for ischemia injury in heart and brain [60,61]. However, if blocking NETs formation and inhibiting Dectin-1 could decrease pericyte activation, restore BBB integrity and improve TBI long-term outcomes were undetermined. Thus, we explored the effects of Cl-Amidine and LAM at different time points after TBI (Fig. 9a). The mice were treated with or without Cl-Amidine (50mg/kg) and LAM (25 mg/kg and 50mg/kg) intraperitoneally every three days post TBI according to previous study [21,60]. Administration of Cl-Amidine and LAM signi cantly reduced BBB permeability and death rate compared with saline group in acute phase (Fig. 9b&9c). These two drugs also improved motor and psychiatric functions in TBI mice after two/four weeks (Fig. 9d-9f). However, we did not see statistically differences in these drug treatment groups at 8 weeks post TBI ( Figure S8), which might be the annulled pharmaceutical effect ( Figure S8). FACS results also showed that Cl-Amidine and LAM dramatically decreased the amounts of in ltrated neutrophils ( Fig. 9g&9h) and CD11b + brain pericytes in damaged parenchyma (Fig. 9i&9j), which contributed to neurobehavioral improvement. These data provide the evidences that targeting at NETs formation and Dectin-1 are effective to ameliorate pericyte dysfunction and neutrophil in ltration, facilitating neurological recovery after TBI.

Discussion:
In this study, we report that histones of NETs drive pericyte transformation, the important constituent cell of BBB, from a steady state to an active phenotype with high expression of CD11b. The CD11b + pericytes shrink terminal processes and lower their integrity, leading to decreased vascular coverage and increased permeability of BBB; secrete tremendous in ammatory mediators to attract more circulating leukocytes into brain parenchyma, deteriorating neuroin ammation post TBI. Mechanistically, Dectin-1 on pericytes is responsible for recognizing histones to induce CD11b expression in a PKC-c-Jun dependent manner.
Targeting at Dectin-1 and NETs formation signi cantly pericyte dysfunction and neutrophil in ltration, facilitating neurological recovery after TBI. These data indicate that peripheral blood neutrophil is an essential contributing factor to BBB breakdown by forming NETs, while brain pericyte of BBB is both a target cell and a bridge when neutrophil transmigrates through BBB to exert its pro-in ammatory effect. This builds up cellular and molecular connections between pericyte dysfunction and leukocyte in ltration, which may provide new pathological process after brain injury and novel treatment strategy against TBI.
Pericyte is generally treated as indispensable participant during angiogenesis and microcirculation regulation. Recent years, it has raised increasing interests due to its heterogeneity and plasticity in terms of multiple morphology and protein expression pattern. Brain pericyte is more unique and speci c since it has the highest vascular coverage [62] and forms a compact barrier between circulating cell and interstitial space in neurovascular unit [23]. Here, we con rm that CD11b + brain pericyte is highly induced after TBI, which is characterized with increased ability of leukocyte attraction and damaged integrity. This is also consistent with previous studies that increased expression of G-protein signaling 5 (RGS5), NG2 and CD11b endowed pericytes microglia-like phenotypes under stroke or sterile in ammation conditions [46,63]. CD11b and ICAM-1 are two important molecules for pericyte to actively instruct immune cells toward in ammatory stimulation [28,63]. The unapparent correlation between ICAM-1 expression on pericytes and NETs formation does not exclude the possibility that CD11b and ICAM-1could facilitate pericyte activation synergistically in TBI. What's more, activated pericyte mediates neuroin ammation through diverse ways apart from leukocyte recruitment and BBB disruption [25]. The pericyte-endothelium communications [41], ability of brotic scarring [64], as well as in uence on endocytosis [65] that CD11b + brain pericyte might exhibit within TBI pathogenesis are desperately needed in the future study. Xie X et al demonstrated that targeting at pericyte might be bene cial for patients with TBI, and potentially reduce the risk of developing AD [66]. Here we did not see statistical differences in Y maze between treatment groups (Cl-Amidine, LAM) and saline groups 12 weeks post TBI ( Figure S8), even though they decreased the amount of CD11b + pericyte in acute phase (Fig. 9i). We speculate that these drug's effects may not last that long since we only treated these mice for 2 weeks. Thus, continuous drug administration for longer time post TBI as well other behavioral evaluations about spatial learning and memory function are substantial in the future. In brain diseases, it has been shown that NETs are associated with cerebral edema, hypoperfusion, BBB damage, neurotoxicity, amyloid β plaque deposition and so forth [19,21]. However, the detailed mechanisms about how cytotoxic proteins and cytokines carried by NETs perform in brain are still elusive. Here, we have analyzed the component of NETs-formed medium by proteomics and metabolomics methods, con rming histones from NETs are important mediators for brain pericyte activation (Fig. 5). Over time, our knowledge of histones has changed dramatically: from DNA packaging proteins to antimicrobial agents, from inherent factors to self-detrimental agents [56]. Circulating histones from necrotic or dead cells are highly associated with the severity and development of acute in ammatory diseases and ischemia-reperfusion injury [55,67]. In pre-clinical studies, neutralizing or degrading extracellular histones with speci c antibody or heparan sulfates have shown promising effects even though our understandings about histone-induced pathologies are still at the beginning [68]. We build the connection between NETs histones and CD11b induction in brain pericyte, illuminating the possible histone-related molecular mechanism in neuroin ammation. While, it is better to distinguish the differences of histone subtypes upon this phenomenon and determine the functional amino acid sequences that stimulate these responses, providing potential therapeutic and diagnostic targets for TBI and other neuroin ammation-related diseases.
The histones and free nucleic acids from NETs are generally recognized by innate immune system through binding to damage-associated molecular pattern recognition receptors, such as Toll-like receptors (TLRs), cytosolic DNA sensors (CDSs), and Nod-like receptors (NLRs) [69]. Interacting with TLRs and NLRs, Dectin-1 also plays fundamental roles of regulating phagocytosis, ROS and cytokine production in myeloid cells, such as macrophage and neutrophil [61,70]. Moreover, the recognition for histones [54] and alarmin for neuroin ammation [71,72] broaden our understanding about Dectin-1 in innate immune system as well as CNS. Latest studies and our data ( Figure S6) have showed that Dectin-1 expression is highly enhanced in brain tissue following ischemic injury or TBI, causing overwhelming activation of microglia [60,73]. This provides the hint that Dectin-1might affect neuroin ammation in multiple ways.
Apart from the previous report that Dectin-1 facilitates CD11b activation in neutrophil by forming Vav-PLCγ complex during fungal clearance [74], we for the rst time show that Dectin-1 directly induces CD11b expression by promoting c-Jun binding to promoter gene sequence in pericyte (Fig. 8), elucidating the possible molecular mechanism for pericyte activation. However, our data did exclude the possibility that Dectin-1 inhibition could decrease neutrophil in ltration directly apart from pericyte modulation since Dectin-1 contributes neutrophil accumulation based on previous research [75]. Thus, more speci c treatment or conditional brain pericyte Dectin-1 knockout will strengthen the conclusion in the future.
Targeting at Dectin-1 and its family receptors were previously focused on infection, and were at the initiation stage of several clinical trials (HBV and fungal infection, oral squamous cell carcinoma treatment et al) [76]. Our result that Dectin-1 is involved in NETs-mediated pericyte dysfunction provides the innovative approaches for TBI treatment, broadening the biological effects and therapeutic targets of Dectin-1 in CNS. Also, the price of Dectin-1 inhibitor (LAM) is lower compared to PAD inhibitor (Cl-Amidine) and rhDNase, which might be low-cost and high-reward candidate for TBI patients. There are still some aspects that might limit Dectin-1 applications: inappropriate Dectin-1antagonist administration could induce severe immune de ciency (vulnerable for fungal infection); there is no FDA-approved drugs in clinic targeting at Dectin-1. Further investigations about effective and securable Dectin-1-related drugs are desperately needed.

Conclusions
In conclusion, our data support that activated phenotype (CD11b + ) of brain pericyte is highly induced by NETs-associated histones after TBI. Recognition of histones by Dectin-1 in pericyte contributes to CD11b induction in a PKC-c-Jun dependent manner. Targeting at NETs formation and Dectin-1 are e cient for restoring BBB integrity and attenuating neuroin ammation to improve neurological outcomes of TBI.  Neutrophil in ltration, NETs formation and pericyte dysfunction post TBI. a&b Immunostaining of pericyte (PDGFRβ, α-SMA) and neutrophil (CD177, CD16) in brain tissues from TBI patients. White scale bar 100μm. c&d Correlation analysis between NETs formation (MPO-DNA binding capacity) and CD11b/ICAM-1 expression level of pericyte isolated from injured brain tissues (n=10). Pearson's correlation coe cient (r), effective R square (R2) and p value are shown as insets. e Immuno uorescence of pericyte (PDGFRβ) and in ltrated neutrophil (LY6G) in mouse brain at 24 hours after TBI insults. Sham represents the mouse went through the same surgical procedures without impact as control. White scale bar, 100μm. f&g HE stain (f) and TEM (g) of mouse brain vascular unit at 24 hours after TBI. The asterisk (*) represents lumen, the arrow (>) represents endothelial cell, the triangle (Δ) represents astrocyte and the cross (+) represents pericyte. White bar in HE (f) 50μm. White bar in TEM (g) 1μm. h Detection of NETs formation (Cit H3 combined DAPI) and pericyte (PDGFRβ). White arrows point the NETs characters of web-like nucleic acids (patulous and elongated shape of DAPI) and enhanced citrullinated histone 3 (Cit H3) in brain tissue at 24 hours after TBI. White scale bar, 100μm. i Morphology of NETs formation in different brain regions at 24 hours after TBI. White scale bar, 100μm. j FACS analysis of NETs formation (LY6G+MPO+) rates in different brain regions of (i). Each bar represents 5 individual experiments. Cell nuclei were all stained with DAPI.  NETs formation and its impact on primary isolated brain pericyte after TBI. a Analysis of NETs formation from peripheral blood. The neutrophils were collected from peripheral blood of Sham and TBI mice at 24 hours post injury. Then the cells were treated with gradient concentrations of PMA (10 ng/ml, 100 ng/ml, 500 ng/ml) for 6 hours and stained with Cit H3 and DAPI. White scale bar 200μm. b Quali cation of NETs formation rate in (a). The NETs structures were counted from 5-10 random eld in each group. **p<0.01. c Immuno uorescence of pericyte (red) that expressing CD11b (green) in injured brain from TBI patient. White scale bar 10μm. d Impact on primary-isolated brain pericyte mediated by NETs-formed medium (detailed process is described in method). The freshly-isolated and cultured mouse brain pericyte was cultured with or without NETs-formed medium. None means pericyte was cultured under normal condition. The same amounts of neutrophils were treated with or without PMA (100ng/ml) for 6 hours, then the medium was collected (detailed in method neutrophil isolation and treatment) as PMA and control group. Then the collected medium was mixed with normal culture medium by the ratio of 1:3 and incubated with primary pericyte cells for 24 hours. Subsequently, the cells were xed and stained with PDGFRβ (red) and CD11b (green). Cell nuclei were all stained with DAPI. White scale bar 100μm. Effects of NETs-formed medium on pericyte cell line MBVP. Murine brain pericyte cell line MBVP was incubated with intended mixed-culture medium as described above for 48 hours. None means normal medium, control means neutrophil without PMA stimulation, PMA means NETs formation stimulated by PMA (100ng/ml), PMA+Cl-Amidine means PMA stimulation (100ng/ml) combined with NETs inhibitor Cl-Amidine (10μM). a Morphological features of MBVP in different groups observed by light microscope.
White scale bar, 100μm. b Immunostaining of tight junction protein (ZO-1, green) and skeleton protein (βactin, red). White scale bar, 10μm. c Neutrophil adhesion test of MBVP. After 48 hours incubation with intended medium, MBVP was cocultured with 104 freshly-isolated neutrophil in each 6-well plate for 2 hours. Then the coculture plates were washed with PBS for 5 times and observed under light microscope.
The big outstretched shapes with low cell diopter were MBVPs, the small round shape with high cell  Analysis of functional components from NETs-formed medium affecting pericyte phenotype. a Work ow for proteomics and metabolomics comparisons between NETs-formed medium (PMA) and control medium (neutrophils without PMA stimulation). Proteomics n=3 for each group. Metabolomis n=7 for each group. b Pie chart of differential proteins and top 20 up-regulated proteins in NETs-formed medium.
Changes greater than 1.5-fold and p < 0.05 were considered signi cantly differential. c Hierarchical Dectin-1 in pericyte recognized histones from NETs. Pretreated medium: Control means neutrophil without PMA stimulation, PMA means NETs formation stimulated by PMA (100ng/ml), Histones means mixed puri ed histone peptides (H1, H2 and H3) were dissolved at 2μg/ml. a WB analysis of Detin-1expression in pericyte cell line MBVP incubated with mixed-culture medium for 48 hours. The pre-treated medium was mixed with normal culture medium at the ratio of 1:3 as described above. b Relative phosphorylated-PKCζ/λ and phosphorylated-c-Jun in GFX (PKC inhibitor, 5μM) pre-treated MBVP combined with NETs-formed medium (PMA) and Histones (2μg/ml) treatment for 24 hours. e Relative expression of CD11b mRNA in MBVP treated with NETs-formed medium (PMA) and Histones (2μg/ml) for 24 hours in the presence of GFX (PKC inhibitor, 5μM). f FACS analysis of CD11+ population in MBVP treated with NETs-formed medium (PMA) and Histones (2μg/ml) for 48 hours in the presence of GFX (PKC inhibitor, 5μM). g Relative expression of CD11b mRNA in MBVP treated with NETs-formed medium (PMA) and Histones (2μg/ml) for 24 hours in the presence of T-5224 (c-Jun inhibitor, 10μM). h FACS analysis of CD11+ population in MBVP treated with NETs-formed medium (PMA) and Histones (2μg/ml) for 48 hours in the presence of T-5224 (c-Jun inhibitor, 10μM). Data are shown as mean±SEM, presented for 3 individual experiments and analyzed by two-tailed unpaired Student's t test. **p<0.01, *p<0.05.  Evaluations of BBB integrity, pericyte activation, neutrophil in ltration and neurological recovery by targeting at NETs-Dectin-1 axis post TBI. a Schematic work ow of animal experiments. After constructing moderate brain impact model, mice were immediately treated with NETs inhibitor Cl-Amidine (50mg/kg) and Dectin-1 antagonist Laminarin (25mg/kg and 50mg/kg) by intraperitoneal injection every three days.
Two weeks later, the treatments were terminated and mice were fed without any interferences. b Quanti cation of Evan Blue in left brain hemisphere. Evans Blue (2% diluted in saline) was injected into tail vein (4 ml/kg) and allowed to circulate for intended time (30min, 6h and 24 h). The data were presented as (µg of Evans Blue stain)/ (g of injured brain hemisphere). Each group contains 4 mice for each group at intended time point. c Survival rate of each group at the rst week after TBI impact in the presence of different treatments. n=6 for Sham, n=10 for TBI. Sham represents the mouse went through the same surgical procedures without impact as control, TBI (saline) represents as placebo group compared with other drug treatment groups. d Footfault evaluation of TBI mice with administration of indicated drugs for 2 weeks. n≥5 for each group. e Heatmap of mice movement in OFT assessment at 4 weeks after moderate TBI impact in the presence of different treatments. f Quanti cation of OFT assessment (e) in the aspects of frequency to cross central zone and time that mice spent in center zone. n≥5 for each group. g FACS analysis of in ltrated neutrophils (CD45+CD11b+LY6G+) from injured brain tissue at 24 hours in TBI mice treated with indicated drugs (Laminarin, 50mg/kg; Cl-Amidine, 50mg/kg). n=4 for each group. h Quanti cation of (g). i FACS analysis of activated brain pericyte (CD45-CD11b+CD13+) from injured brain tissue at 24 hours in TBI mice treated with indicated drugs (Laminarin, 50mg/kg; Cl-Amidine, 50mg/kg). n=4 for each group. j Quanti cation of (i). Data are presented as