Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2013 Apr 1.
Published in final edited form as: J Neuroimmunol. 2012 Feb 29;245(1-2):48–55. doi: 10.1016/j.jneuroim.2012.02.004

Thrombin induces an inflammatory phenotype in a human brain endothelial cell line

Leah M Alabanza 1, Margaret S Bynoe 1,*
PMCID: PMC3608467  NIHMSID: NIHMS356998  PMID: 22381244

Abstract

In this study, we utilized the human brain endothelial cell line, hCMEC/D3, to determine the effects of the coagulation factor, thrombin, on the human blood–brain barrier (BBB). We show that thrombin increased the mRNA and cell surface levels of ICAM-1 and VCAM-1 in hCMEC/D3 cells. Thrombin similarly upregulated several chemokines implicated in human neurological conditions. Additionally, the paracellular permeability of the human BBB in vitro was also increased following thrombin treatment. Overall, this study demonstrates that thrombin can effectively induce an inflamed phenotype in an in vitro human BBB.

Keywords: Thrombin, Blood-brain barrier, hCMEC/D3, Endothelial cells

1. Introduction

The blood–brain barrier (BBB) is structured to limit the influx of potentially harmful blood-borne molecules into the central nervous system (CNS). The BBB endothelium, however, is responsive to various stimuli resulting in a BBB phenotype that can dynamically change based on the conditions in the microenvironment (de Boer and Gaillard, 2006). In neuroinflammatory settings, for instance, an inflamed BBB becomes more permeable to leukocytes in the circulation as a consequence of phenotypic changes that include the disassembly of tight junction molecules, increased expression of cell adhesion molecules (CAMs), and synthesis of chemotactic molecules (Webb and Muir, 2000; de Boer and Gaillard, 2006).

In recent years, there has been accumulating evidence of a complex, two-way interaction between the immune and coagulation systems (Levi and van der Poll, 2005; Esmon, 2008). Specific components of the coagulation system can directly influence the immune response. The coagulation factor, thrombin, in particular, is widely-regarded to have a major influence on inflammation (Naldini et al., 2002; Szaba and Smiley, 2002; Yanagita et al., 2007). Thrombin is a serine protease that is responsible for the formation of fibrin, the primary component of blood clot (Monroe and Hoffman, 2006). Beyond its central role in the coagulation process, thrombin can directly affect the cellular processes of various cell types, including leukocytes and endothelial cells, largely through the activation of G protein-coupled receptors known as protease-activated receptors (PARs) (Coughlin, 1999; Derian et al., 2002).

While the inflammatory influences of thrombin have been predominantly observed in the periphery, there are now growing reports that thrombin may also be involved in neuroinflammatory and neurodegenerative conditions (Chapman, 2006; Sokolova and Reiser, 2008). Thrombin in the circulation can enter the CNS in conditions where the BBB is compromised (Suo et al., 2004). Additionally several coagulation factors including prothrombin and its activator, factor X, are locally synthesized in the brain (Dihanich et al., 1991; Yamada and Nagai, 1996). Increased thrombin activity and/or expression in the CNS have been observed in a number of neurological diseases. In Parkinson’s disease and human immunodeficiency virus encephalitis, thrombin and prothrombin expression have been shown to be upregulated in astrocytes and neurons (Boven et al., 2003; Ishida et al., 2006). Ischemic brains are characterized by high thrombin levels as a result of thrombin extravasation from the circulation and local prothrombin synthesis (Riek-Burchardt et al., 2002; de Castro Ribeiro et al., 2006). Elevated thrombin levels have also been observed in brain microvessels in Alzheimer’s disease (Grammas et al., 2006). While the specific role of thrombin in the progression of these diseases is still not well-delineated, it is known that resident cells in the CNS can be responsive to thrombin activity. Thrombin can induce microglia and astrocytes to release proinflammatory cytokines including IL-1β, IL-6, and TNF-α (Choi et al., 2003; Ishida et al., 2006). High thrombin concentrations can also result in neurodegeneration (Rohatgi et al., 2004).

Similarly, in vivo studies in animal models have shown that the BBB can be responsive to high levels of thrombin in the CNS. For instance, intrathecal injection of thrombin into rat brains can result in brain edema as a result of BBB permeability (Chen et al., 2010; Liu et al., 2010). The collective effects of thrombin on the phenotype of the specialized endothelial cells that comprise the human BBB however have not been thoroughly investigated. While there are a number of studies that have examined thrombin’s effects on endothelial cells, these studies have exclusively utilized endothelial cells from the periphery. The specific responses of peripheral endothelial cells to thrombin, however, may not necessarily reflect the responses of the distinct endothelial cells that comprise the BBB. It is, therefore, warranted to particularly investigate the effects of thrombin activity on brain endothelial cells. Furthermore, given that thrombin’s presence is heightened in various human neuroinflammatory and neurological disorders it is pertinent to study thrombin’s effects specifically on human brain endothelial cells.

The purpose of this study is to utilize a recently developed in vitro model of the human BBB to examine the responses of human brain endothelial cells to thrombin activity. We utilized the hCMEC/D3 cell line, an immortalized primary human brain endothelial cell line that stably retain the morphological characteristics of the BBB without the need for supporting glial cells (Weksler et al., 2005), as an in vitro model for studying the responses of human brain endothelial cells to thrombin. In this study, we show that thrombin can induce specific responses in brain endothelial cells in vitro, effectively inducing an inflamed BBB phenotype that is conducive to the potential recruitment, capture, and diapedesis of leukocytes. The results of this study may provide insight into the possible contribution of thrombin in the pathogenesis of neuroinflammatory and neurodegenerative conditions.

2. Materials and methods

2.1. Cell culture

The hCMEC/D3 cell line was generated by Weksler et al. (2005) and kindly provided to us by Dr. Babette Weksler. The cells were cultured in EBM-2 (Lonza, Wokingham, UK) or MCDB 131 (Invitrogen, CA, U.S.A.) with 2.5% fetal bovine serum (FCS) and supplemented with hydrocortisone, ascorbate, basic fibroblast growth factor (bFGF), vascular endothelial growth factor (VEGF), insulin-like growth factor (IGF-1), epidermal growth factor (EGF), and gentamycin sulfate. Supplements were purchased from Lonza (Wokingham, UK), and concentrations used were according to the manufacturer’s protocol. The cells were maintained at 37 °C in a humidified atmosphere with 5% CO2. All culture wares were coated with 0.1 mg/ml rat-tail collagen type I (BD Pharmingen, CA, U.S.A.). In all the experiments, the cells were grown to confluency for 7–10 days in fully supplemented MCDB-131 or EBM-2 media, which was changed every 2–3 days. Eighteen to twenty hours before each assay, the media was switched to reduced FBS and growth-factor depleted medium (MCDB 131 or EBM-2 supplemented with 0.25% FBS, 1 ng/ml bFGF, 10 mM Hepes, and .55 μM hydrocortisone). This media was used throughout the entire duration of all experiments. For thrombin treatments, human thrombin (Sigma-Aldrich, MO, U.S.A.) was solubilized in molecular grade H2O and further diluted in media.

2.2. Quantitative real-time PCR

Cells were seeded on collagen-coated 12-well plates (BD Falcon, CA, U.S.A). After thrombin (0.5–60 U/ml) or vehicle treatments, total RNA was extracted using Trizol (Invitrogen, CA, U.S.A.) according to the manufacturer’s protocol. Extracted RNA was treated with Baseline-Zero DNAse (Epicentre, WI, U.S.A.) to remove contaminating genomic DNA. 1.5 μg of total RNA was reverse transcribed using High- Capacity cDNA Reverse Transcription kit (Applied Biosystems, CA, U.S.A.). Real Time PCR was performed using SYBR Green technology (SYBR FAST Master Mix KAPAbiosystems, MA, U.S.A.) on a CFX96 Real Time System Thermal Cycler (Bio-RAD, CA, U.S.A.). The cycling conditions were as follows: enzyme activation 95 °C for 3 min followed by 40 cycles of denaturation at 95 °C for 3 s, annealing at 60 °C for 30 s, and elongation at 72 °C for 5 s. BioRad CFX Manager software was used to determine cycle threshold (Ct) values, and gene expression levels were calculated using the 2−ΔΔCt method. The expression levels of all genes of interest were normalized to the internal control gene, GAPDH.

2.3. Flow cytometry analysis

Cells were seeded on 24-well plates (BD Falcon). After thrombin or vehicle treatments, cells were gently detached from culture plates using polyethylene cell scrapers (BD Falcon). The cells were washed and re-suspended in staining buffer (PBS with 0.5% BSA, and 0.09% sodium azide). Cell suspensions were incubated with antibody against the protein of interest for 30 min at 4 °C. Antibodies used are as follows: biotin anti-human ICAM-1, (BD Pharmingen, CA, U.S.A.), APC-conjugated anti-human VCAM-1 (Biolegend, CA, U.S.A.), and anti-human PAR-1 (Enzo Life Sciences, NY, U.S.A.). The cells were subsequently washed twice in staining buffer and incubated with the appropriate secondary antibody for 30 min at 4 °C. The cells were subjected to two final rounds of washing and resuspended in a final volume of 300 μl of staining buffer. The cells were acquired using a FACSCanto II flow cytometer (BD Biosciences, CA, U.S.A.). The raw data was evaluated using FlowJo Flow Cytometry Analysis Software (Treestar, OR, U.S.A.).

2.4. Western blot analysis

Cells were seeded on collagen-coated 12-well plates (BD Falcon). After thrombin or vehicle treatments, the supernatants and whole-cell lysates were collected. Cells were lysed on ice using lysis buffer (10 mM Tris–HCl, pH 7.4, 150 mM NaCl, 1 mM EDTA, 10mM Na3VO4, 10mM sodium fluoride, 10 mM sodium pyrophosphate, 1% NP-40, 1 mM PMSF, 1 μM Pepstatin A). Laemmli buffer at 5× (60 mM Tris–HCl pH 6.8, 2% SDS, 10% glycerol, 5% β-mercaptoethanol, 0.01% bromophenol blue)was added to the supernatant andwhole-cell lysate samples. Samples then were heated to 99 °C for 5 min. For detection of CCL2 or CXCL8, the supernatant samples were loaded in 18% SDS-PAGE gel (1.5M Tris–HCl pH 8.8, 20% SDS, acrylamide/bis-acrylamide, 10% ammonium persulfate, TEMED). For CX3CL1 detection, the whole-cell lysate samples were loaded in 7% SDS-PAGE gel. The gels were ran in running buffer (25 mM Tris, 192 mM glycine, 0.1% SDS pH 8.5) at 100 V. Proteins were transferred to nitrocellulose membrane (Whatman, NJ, U.S.A.) by standard wet transfer protocol. After transfer, membranes were washed with TBS-Tween (100 mM Tri-HCl pH 7.5, 0.9% NaCl, 0.1% Tween-20) for 5 min. Membranes were blocked with 5% BSA in TBS-Tween for 1 h at 4 °C. After blocking, membranes were incubated with primary antibody in 1% BSA in TBS-Tween overnight at 4 °C. Primary antibodies used are as follows: rabbit anti-human IL-8 (Peprotech, NJ, U.S.A.), rabbit anti-human MCP-1 (Hycult Biotech, U.S.A.), and rat anti-mouse CX3CL1 (R&D Systems). Membranes were washed 3 times with TBS-Tween and incubated with the appropriate HRP-conjugated secondary antibody for 1 h at room temperature. Membranes were washed 3 times with TBS-Tween and incubated with chemiluminescent substrate for HRP (Thermo Scientific Super Signal). X-ray films were exposed to membranes to visualize proteins.

2.5. Transendothelial electrical resistance measurement

Cells were seeded on collagen-coated transwell polyester membrane inserts (for 24-well plate)with 8 μm pore size (BD Falcon). Transendothelial Electrical Resistance (TEER) was measured using volt ohm meter, EVOMX, equipped with STX100 electrodes (World Precision Instruments, FL, U.S.A.). TEER readings were recorded 10min before thrombin (5, 20, 60 U/ml) or vehicle treatment and every 10min for 1.5 h after treatments. Cells were maintained on a 37 °C heat block for the duration of the TEER measurements. Changes in TEER at the given timepoints were expressed as fold change relative to TEER readings that were recorded 10 min before treatment.

2.6. FITC-dextran permeability assay

hCMEC/D3 cells were grown to confluence on collagen-coated transwell inserts (for 24-well plate) with 3 μm pore size (BD Falcon). Thrombin at 5, 20, 60 U/ml, or vehicle and 1 mg/ml of FITC-dextran in assay media were added to the apical chamber. The media from the basolateral lower chamber was sampled after 15, 30, 60, 90, and 180 min of treatment to assess the degree of FITC-dextran flux across the endothelial barrier. Fluorescence intensity was measured at excitation wavelength 488 nm and at emission wavelength 519 nm using Synergy 4 microplate reader (Biotek, VT, U.S.A.). The degree of diffused FITC-dextran at each given timepoint was expressed as fold change in fluorescence intensity relative to media from vehicle controls.

2.7. Statistical analysis

Results were represented as means±S.E. All experiments were done in triplicates. Two-tailed Student’s t-test was used to analyze the differences between two groups using GraphPad Software. A difference was considered statistically significant if the p value is <0.05.

3. Results

3.1. The hCMEC/D3 cell line expresses the necessary receptors to be responsive to thrombin

To determine whether hCMEC/D3 cells can be responsive to thrombin activity, we examined whether these cells constitutively express the cellular receptors for thrombin. We show by real-time quantitative PCR that hCMEC/D3 cells express all four PARs, three of which (PAR-1, PAR-3, and PAR-4) can be activated by thrombin (Fig. 1A). PAR-1, the predominant thrombin receptor on endothelial cells (Minami et al., 2004), is the most highly expressed among all three thrombin receptors in hCMEC/D3 cells. Additionally, we confirmed by flow cytometry analysis that hCMEC/D3 cells express PAR-1 at the cell surface (Fig. 1A). We also examined whether cell surface expression of PAR-1 is altered in response to thrombin activity. We observed a slight decrease in PAR-1 expression after 1.5 h of thrombin treatment (Fig. 1B). At 3 h, PAR-1 levels returned to baseline and remained close to these levels until 20–24 h of thrombin treatment. After 24 h of thrombin treatment, we observed a significant decrease in cell surface expression of PAR-1 (Fig. 1B). This data shows that thrombin can affect the expression of PAR-1 on hCMEC/D3 cells and is consistent with previous studies on other endothelial cell systems (Woolkalis et al., 1995). We also show that the hCMEC/D3 cell line express thrombomodulin and endothelial protein C receptor (EPCR) (Fig. 1C). Both receptors are involved in regulating thrombin activity; the former acts as a thrombin regulator through direct interactions with thrombin, while the latter is involved in the activation of protein C, a serine protease known to inhibit thrombin activation (Van de Wouwer et al., 2004). Overall, these data show that the hCMEC/D3 cells express the necessary receptors to respond to thrombin activity.

Fig. 1.

Fig. 1

Open in a new tab

hCMEC/D3 cells express the necessary receptors to respond and regulate thrombin activity. A) Real-time quantitative PCR was performed to assess the expression of PARs in hCMEC/D3 cells. The mRNA levels are expressed relative to the internal control gene, GAPDH. Data are represented as means±S.E. (n=3). Fluorescence histogram on the right represents hCMEC/D3 cells that were stained with either fluorescently-labeled antibody to PAR-1 or isotype control. Cell fluorescence was assessed by flow cytometry. B) hCMEC/D3 cells were stained with fluorescently-labeled antibody to PAR-1 after treatment with 20 U/ml of thrombin for the timepoints given. Cell fluorescence was examined by flow cytometry, and levels of cell surface PAR-1 at each timepoint was shown as fold change in mean fluorescence intensity relative to vehicle controls. Data represented as means±S.E. (n=3). * p<0.05 versus vehicle control (two-tailed Student’s t-test). Fluorescence histogram on the right represents cell surface expression of PAR-1 at 24 h after vehicle or thrombin treatment. C) The mRNA expression of thrombomodulin and EPCR was examined by real-time quantitative PCR. The mRNA levels are expressed relative to the internal control gene, GAPDH. Data are represented as means±S.E. (n=3).

3.2. Thrombin induces brain endothelial cells to express the mRNA and cell surface expression of ICAM-1 and VCAM-1

The BBB can increase the expression of cell adhesion molecules (CAMs) in response to stimulating factors in the microenvironment. The resting BBB typically expresses low levels of CAMs, however in certain conditions such as in neuroinflammatory settings, the BBB upregulates various CAMs to capture inflammatory immune cells and facilitate the entry of leukocytes into the CNS (Greenwood et al., 2011). ICAM-1 and VCAM-1 are believed to have a predominant involvement in leukocyte infiltration into the CNS (Greenwood et al., 2011). We examined whether thrombin is capable of altering the expression of both ICAM-1 and VCAM-1 in human brain endothelial cells. Our data show that thrombin can induce a dose-dependent increase in mRNA expression of both CAMs in hCMEC/D3 cells (Fig. 2A). The increase in mRNA levels reached statistical significance after treatment with thrombin at 0.5 U/ml, the lowest concentration used, and reached maximal upregulation after treatment with 60 U/ml of thrombin. The increase in adhesion molecule expression was observed after 1 h of thrombin treatment and further increase was observed at 3 h (Fig. 2B). The mRNA levels return close to basal levels after 6 h of thrombin treatment. We observed a second increase in mRNA levels after 18 h of treatment, subsequently declining to basal levels at 24 h (Fig. 2B).

Fig. 2.

Fig. 2

Open in a new tab

Thrombin upregulates the mRNA and cell surface expression of VCAM-1 and ICAM-1 in human brain endothelial cells. A) hCMEC/D3 cells were activated for 3 h with thrombin at the given concentrations, and the mRNA expression of VCAM-1 and ICAM-1 were assessed by real-time quantitative PCR. Expression levels are shown as fold change relative to vehicle-treated controls. Data represented as means±S.E. (n=3). * P<0.05, versus vehicle control (two-tailed Student’s t-test). B) hCMEC/D3 cells were treated with 10 U/ml of thrombin for the given timepoints. The mRNA expression levels of adhesion molecules at each timepoint were evaluated by real-time quantitative PCR. Mean expression levels are shown as fold change relative to vehicle treated controls for each given time point (n=3). C) and D) hCMEC/D3 cells were stained with fluorescently-labeled antibodies to VCAM-1 and ICAM-1 after treatment with the given concentrations of thrombin for 24 h. Cells were analyzed by flow cytometry, and changes in cell surface expression of VCAM-1 and ICAM-1 were assessed by comparing mean fluorescence intensity (MFI) of thrombin-treated samples to vehicle-treated controls. Fluorescence histograms for VCAM-1 and ICAM-1 are shown on the left. Bar graphs on the right show fold change in MFI relative to vehicle treated controls. Data represented as means±S.E. (VCAM-1 n=3, ICAM-1 n=3). * P<0.05 versus vehicle control (two-tailed Student’s t-test).

We next examined whether the cell surface levels of both CAMs are similarly increased after thrombin treatment. Flow cytometry analysis shows detectable increase in VCAM-1 cell surface expression after 24 h of thrombin treatment (Fig. 2C, left). The up-regulation in cell surface VCAM-1, as demonstrated by increase in mean fluorescence intensity (MFI), is observed after treatment with thrombin concentrations in the range of 0.5–10 U/ml (Fig. 2C, right). Higher concentrations of thrombin (20–60 U/ml), however, decreased cell surface VCAM-1 expression at 24 h (Fig. 2C, right). Conversely, we observed maximal increase of ICAM-1 cell surface levels at 24 h with higher concentrations of thrombin (20–40 U/ml) (Fig. 2D). We did not observe significant changes in cell surface ICAM-1 after treatment with lower concentrations of thrombin (data not shown). Overall these data demonstrate that thrombin can effectively induce brain endothelial cells to express VCAM-1 and ICAM-1, both at the mRNA and cell surface levels.

3.3. Thrombin induces brain endothelial cells to increase the expression of chemokines

Under pathological conditions, the endothelial cells that form the BBB can synthesize chemokines to recruit leukocytes to the site of CNS inflammation (Eugenin and Berman, 2003). In addition to their chemotactic functions, chemokines secreted by endothelial cells can enhance leukocyte firm adhesion to CAMs on the endothelium, thus directly facilitating the subsequent diapedesis of leukocytes to the CNS parenchyma (Greenwood et al., 2011). We examined whether thrombin can induce CNS endothelial cells to express chemokines that are known to be involved in various neuroinflammatory conditions. We examined the expression of CXC chemokines, a family of chemokines known to be primarily chemotactic for neutrophils. In particular, we looked at the mRNA expression of CXCL1 (GRO-alpha), CXCL2 (GRO-beta), CXCL8 (IL-8), and CXCL10 (IP-10); all have been reportedly involved in various neuropathologies (Cardona et al., 2008; Hamann et al., 2008). Thrombin activity significantly resulted in increased mRNA expression of all four CXC chemokines (Fig. 3A). CXCL8 a strong neutrophil chemoattractant, with likely involvement in neurological conditions like ischemic brain injury and traumatic brain injury (Semple et al., 2010), was the most highly upregulated with maximal mRNA upregulation reaching 40 fold. Western blot analysis confirms a time-dependent increase in soluble CXCL8 in the supernatant of thrombin treated hCMEC/D3 cells (Fig. 3B).

Fig. 3.

Fig. 3

Open in a new tab

Thrombin induces the expression of neuroinflammatory-implicated chemokines in human brain endothelial cells. hCMEC/D3 cells were activated with the given concentrations of thrombin for 3 h, and the mRNA levels of selected chemokines from A) CCL, C) CXC, and E) CX3C chemokine families were assessed by real time quantitative PCR. The mRNA levels are expressed as fold change relative to vehicle treated controls. Data represented as means±S.E. (n=3). *P<0.05 versus vehicle controls (two-tailed Student’s t-test). Arrows span the concentrations in which the chemokines indicated have been significantly upregulated. B) and D) The supernatant of thrombin-treated hCMEC/D3 cells at the given time-points were examined by Western blot for soluble CXCL8 or soluble CCL2. F) The protein expression of CX3CL1 in whole-cell lysates of thrombin-treated hCMEC/D3 cells was assessed by western blot.

We also examined the mRNA expression of selected members of the CC chemokine family which have been recognized to play roles in neuroinflammatory diseases; these include CCL2 (MCP-1), CCL3 (MIP-α), and CCL5 (RANTES) (Eugenin and Berman, 2003). Thrombin can induce a robust dose-dependent increase in mRNA levels of CCL2 (Fig. 3C), a potent chemoattractant for monocytes and activated T cells (Semple et al., 2010). CCL2 mRNA expression was increased by 15.2 fold after treatment with 60 U/ml of thrombin. Additionally, soluble CCL2 was also increased over time in the supernatant of thrombin treated hCMEC/D3 cells (Fig. 3D). Thrombin treatment, however, did not significantly affect the mRNA expression of either CCL3 or CCL5 (Fig. 3C).

We next examined the CX3C chemokine group, which is composed of CX3CL1 (fractalkine). CX3CL1 is a unique chemokine in that it is synthesized as a membrane-bound molecule allowing it to directly capture circulating leukocytes. It can also be cleaved from the cell membrane by metalloproteinases to generate a soluble form that can function as a traditional chemoattractant (White and Greaves, 2009). CX3CL1 has similarly been linked to the pathogenesis of several neuroinflammatory diseases, including Parkinson’s disease and experimental autoimmune encephalomyelitis (EAE) (Shan et al., 2009). Our data show that thrombin activity can dose-dependently induce brain endothelial cells to significantly upregulate themRNA expression of CX3CL1 (Fig. 3E). The highest upregulation was observed at 9.3 fold relative to vehicle control. We also show that CX3CL1 protein expression was increased by thrombin in whole-cell lysates of hCMEC/D3 cells (Fig. 3F).

3.4. Thrombin increases the paracellular permeability of brain endothelial cells

One of the more drastic changes in BBB functionality in response to factors in the microenvironment is the increased permeability to cells and compounds in the circulation (de Boer and Gaillard, 2006). Increase in BBB permeability is a consequence of various processes in endothelial cells which include cytoskeletal rearrangements and disassembly of junctional proteins, ultimately resulting in the formation of spaces in between endothelial cells to allow for paracellular diapedesis of leukocytes and paracellular flux of blood-borne molecules (Greenwood et al., 2011). We examined if thrombin activity can directly affect the permeability of the human BBB in vitro. We assessed permeability changes by measuring the transendothelial electrical resistance (TEER) of hCMEC/D3 cells after thrombin treatment. The cells were seeded in transwell inserts and treated with vehicle or thrombin at 5, 20, or 60 U/ml. TEER was measured over time before and after treatment; a decrease in TEER correlates with increase in barrier permeability. Thrombin at all three concentrations resulted in detectable decrease in TEER of hCMECD/3 cells within 30 min after treatment (Fig. 4A). Maximal and statistically significant TEER decrease was observed after 1 h of treatment at all three concentrations. We also observed that the decrease in TEER was not sustained; as TEER started to increase 70min after treatment (Fig. 4A).

Fig. 4.

Fig. 4

Open in a new tab

Thrombin increases the paracellular permeability of human brain endothelial cells. A) hCMEC/D3 cells were seeded on transwell inserts and treated with vehicle or thrombin at 5, 20, or 60 U/ml. Permeability was assessed by measuring TEER changes every 10 min after treatment. TEER changes at each given timepoint are shown as fold change relative to TEER readings recorded 10 min prior to treatment. Data represented as means±S.E. (n=3). *P<0.05 versus vehicle control at each timepoint (two-tailed Student’s t test). B) hCMEC/D3 cells were seeded in transwell inserts; 10 KDa FITC-dextran and thrombin at the concentrations shown were added to the apical chamber. After treatment, media from the lower chamber were sampled at the timepoints given to assess the degree of FITC-dextran paracellular flux. Fluorescence intensity in the basolateral media is expressed as fold change relative to vehicle control for each respective timepoint. Data represented as means±S.E. (n=3) * P<0.05 versus vehicle control at each timepoint (two-tailed Student’ t test).

The paracellular permeability of thrombin-treated hCMEC/D3 cells was further investigated by assessing the paracellular flux of 10 kDa FITC-Dextran at various timepoints after treatment with thrombin. We observed statistically significant increases in dextran permeability 15 min after thrombin treatment at 5, 20, 60 U/ml (Fig. 4B). Maximal permeability was observed after 60–90 min of treatment. We did not observe continued increase in dextran permeability at 3 h (180 min) (Fig. 4B), suggesting that similar to what was observed with TEER measurement, the increased paracellular permeability to dextran molecules was not sustained for prolonged periods.

4. Discussion

There are growing reports that thrombin, the central enzyme in the coagulation process, is increased in the CNS during various neuroinflammatory and neurodegenerative disorders (Chapman, 2006; Sokolova and Reiser, 2008). In this study, we investigated whether increased thrombin presence in the CNS can potentially affect the phenotype and functionality of the BBB. We utilized the hCMEC/D3 cell line, a recently-developed model of the human BBB (Weksler et al., 2005), to assess the collective effects of thrombin on brain endothelial cells. We demonstrate that thrombin can induce the expression of chemokines and CAMs in hCMEC/D3 cells. These results corroborate previous studies that have been conducted on peripheral endothelial cells (Kaplanski et al., 1998; Okada et al., 2006). There is, however, an apparent difference in the degree of thrombin responsiveness in human brain endothelial cells compared to what has been reported on peripheral endothelial cells. A study conducted on human umbilical vein endothelial cells (HUVECS) reported seeing robust mRNA upregulation of ICAM-1 (16.0 fold), and VCAM-1 (12.2 fold) after treatment with a relatively low dose of thrombin (2 U/ml for 4 h) (Okada et al., 2006). In comparison, our study showed that hCMEC/D3 cells reached a comparable level of mRNA upregulation after treatment with 60 U/ml of thrombin. The upregulation in ICAM-1 and VCAM-1 cell surface levels were similarly reported to be more robust in thrombin-treated HUVECS (Kaplanski et al., 1998) compared to what we observed in hCMEC/D3 cells. It has been proposed that cells detect and respond to differences in thrombin concentrations based on the rates of receptor activation (Ishii et al., 1993). The observation that HUVECS are more responsive than hCMEC/D3 cells to thrombin at low concentrations suggests that thrombin receptors on HUVECS are likely being activated at a faster rate than on hCMEC/D3 cells. Any differences in the expression levels of thrombin receptors on the two cell systems may likely affect the rate of receptor activation. Moreover, thrombomodulin, a cell surface glycoprotein expressed on endothelial cells, including hCMEC/D3 cells (Fig. 1C) and HUVECS (Nan et al., 2005), is known to neutralize the pro-coagulant and proinflammatory activities of thrombin. Differences in thrombomodulin cell surface levels on hCMEC/D3 cells and HUVECS may also affect the activation rate of thrombin receptors on both cell types. It should be noted, however, that while hCMEC/D3 cells may be less responsive than HUVECS to thrombin activity, the thrombininduced cell surface levels of VCAM-1 (1.8 fold) and ICAM-1 (1.7 fold) on hCMEC/D3 are levels that have been shown to be biologically functional. Recent studies by Roh et al. (2011) and Rains and Jain (2011) show that a 1.25 to 2 fold increase in ICAM-1 expression on endothelial cells resulted in increased monocyte adhesion. Another group (Smedlund et al., 2010) showed that a 2 fold increase in VCAM-1 cell surface levels is likewise sufficient to increase monocyte adhesion.

While HUVECS and hCMEC/D3 may differ in the degree of thrombin responsiveness, our data showed that the two cell systems likely have similar mechanisms for clearance of activated thrombin receptors and re-expression of new intact receptors. In HUVECS, activation of thrombin receptors results in rapid internalization of receptors, resulting in 60% clearance of cleaved receptors within 10 min (Woolkalis et al., 1995). Moreover, HUVECS have intracellular pools of new thrombin receptors, allowing for immediate re-expression of new receptors on the cell surface (Woolkalis et al., 1995). HUVECS can restore nearly a full set of new receptors within a few hours after the initial thrombin receptor activation and clearance (Woolkalis et al., 1995). In hCMEC/D3 cells, PAR-1 surface expression decreased after treatment with thrombin for 1.5 h, suggesting that similar to HUVECS, hCMEC/D3 likely internalize thrombin receptors immediately after activation. We also observed that PAR-1 cell surface expression returned to basal levels within 3 h of thrombin activation, suggesting that hCMEC/D3 cells like HUVECS have an existing intracellular pool of new receptors that can immediately replace activated and internalized receptors. Treatment of hCMEC/D3 cells for various timepoints longer than 3 h did not significantly alter cell surface levels of PAR-1, likely due to the continued process of receptor internalization and receptor re-expression. We did eventually observe a significant decrease in PAR-1 expression after 24 h of continued thrombin treatment, whichmay suggest that, at this point, the intracellular reserve of intact receptors has been depleted.

We also investigated whether thrombin can affect the paracellular permeability of brain endothelial cells. For the endothelium to be permissive to paracellular diapedesis, endothelial cells must undergo various processes which include cytoskeletal rearrangements and disassembly of junctional proteins, ultimately resulting in the formation of spaces in between endothelial cells to allow for paracellular transmigration (Komarova et al., 2007). We assessed if thrombin can affect the paracellular permeability of brain endothelial cells by measuring TEER changes and permeability to dextran molecules after thrombin treatment. We observed that thrombin can rapidly but transiently increase the paracellular permeability of brain endothelial cells. These observations are consistent with previous studies on peripheral endothelial cells (Knezevic et al., 2009). Increases in endothelial permeability in response to thrombin activity have been attributed to the disassembly and disappearance of junctional proteins at cell to cell contacts (Komarova et al., 2007). The thrombin-induced retraction of junctional proteins, however, is not accompanied by protein degradation, and thus can be rapidly reversible (Knezevic et al., 2009). This is the likely mechanism behind the transient permeability observed in thrombin-treated hCMEC/D3 cells.

Thrombin’s effects on endothelial permeability and CAM expression are known to be mediated through different signaling pathways and therefore subject to different regulatory mechanisms (Minami et al., 2004; Komarova et al., 2007). This may explainwhy thrombin at certain concentrations can have opposing effects on CAM expression and paracellular permeability. For instance, we observed that thrombin at 60 U/ml rapidly resulted in increased permeability in hCMEC/D3 cells, but this same concentration decreased the expression of cell surface VCAM- 1 at 24 h (Fig. 2C right). Thrombin induction of VCAM-1 expression is mediated through the NF-κB signaling pathway, specifically the p65 subunit of NF-κB (Minami et al., 2004). However, it has also been observed that thrombin can induce the p50 subunit of NF-κB, which has been suggested to be a negative transcriptional regulator of the VCAM-1 gene (Minami et al., 2004). Thus, it has been proposed that there is a possible negative-feedback mechanism in the thrombin- VCAM-1 signaling pathway (Minami et al., 2004). The observed downregulation of VCAM-1 at higher concentrations of thrombin is likely due to this negative feedback mechanism.

The overall effect of thrombin on brain endothelial cells in vitro results in an endothelium phenotype that is conducive to the multiple stages of leukocyte extravasation. The question remains whether thrombin can have the same effects on the BBB in vivo. Previous studies have shown that intracerebroventricular injection of thrombin into rat brains resulted in increased brain edema (Liu et al.), suggesting that thrombin is capable of affecting the functionality of the BBB in vivo. It remains to be seen, however, if in vivo thrombin activity can result in increased leukocyte extravasation into the CNS. Furthermore, the source of endogenous thrombin that can affect the BBB during neurological diseases also needs to be considered. While thrombin is predominantly found in the circulation, thrombin can be locally synthesized in the CNS (Dihanich et al., 1991). Moreover, thrombin activity and/or expression have been shown to be increased in the CNS in a number of neurological diseases (Chapman, 2006). The fact that intracerebroventricularly-injected thrombin in rat brains can affect BBB permeability (Liu et al., 2010), suggests that thrombin from within the CNS is capable of affecting BBB function. Thus, it is possible for thrombin synthesized within the CNS to induce the phenotypic changes in brain endothelial cells that have been shown in this current study. Additionally, it also needs to be taken into account that thrombin can induce glial cells to produce proinflammatory cytokines such as IL-1β and TNF-α (Suo et al., 2004), which are cytokines that have been traditionally implicated in BBB dysfunction (Minagar and Alexander, 2003). While we have shown in this study that thrombin can independently affect the functionality of the BBB, it is very likely that thrombin can act in concert with various pro-inflammatory factors during neurological conditions to affect BBB functionality. In fact, it has been observed that thrombin and TNF-α can synergistically affect endothelial barrier function in vitro (Tiruppathi et al., 2001).

In summary, this study shows that thrombin can induce brain endothelial cells to express chemokines and CAMs, which can potentially augment the recruitment and capture of leukocytes from the circulation. In addition, thrombin can increase the paracellular permeability of brain endothelial cells, resulting in a BBB that is more permeable to leukocyte diapedesis. Therefore, thrombin can modify the functionality of brain endothelial cells in vitro, resulting in a BBB phenotype that can likely facilitate the successive stages of leukocyte extravasation. Overall, this current study lends better understanding to the implications of increased thrombin presence in the CNS on the pathology of neurological conditions.

Acknowledgments

The hCMEC/D3 cell linewas kindly provided by Dr. Babette Weksler. Funding for this work was supported by Grants from the National Institutes of Health, AI 072434-01A2 and R01 NS063011 (to M.S.B).

References

  1. Boven LA, Vergnolle N, Henry SD, Silva C, Imai Y, Holden J, Warren K, Hollenberg MD, Power C. Up-regulation of proteinase-activated receptor 1 expression in astrocytes during HIV encephalitis. J Immunol. 2003;170:2638–2646. doi: 10.4049/jimmunol.170.5.2638. [DOI] [PubMed] [Google Scholar]
  2. Cardona AE, Li M, Liu L, Savarin C, Ransohoff RM. Chemokines in and out of the central nervous system: much more than chemotaxis and inflammation. J Leukoc Biol. 2008;84:587–594. doi: 10.1189/jlb.1107763. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Chapman J. Thrombin in inflammatory brain diseases. Autoimmun Rev. 2006;5:528–531. doi: 10.1016/j.autrev.2006.02.011. [DOI] [PubMed] [Google Scholar]
  4. Chen B, Cheng Q, Yang K, Lyden PD. Thrombin mediates severe neurovascular injury during ischemia. Stroke. 2010;41:2348–2352. doi: 10.1161/STROKEAHA.110.584920. [DOI] [PubMed] [Google Scholar]
  5. Choi SH, Joe EH, Kim SU, Jin BK. Thrombin-induced microglial activation produces degeneration of nigral dopaminergic neurons in vivo. J Neurosci. 2003;23:5877–5886. doi: 10.1523/JNEUROSCI.23-13-05877.2003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Coughlin SR. How the protease thrombin talks to cells. Proc Natl Acad Sci U S A. 1999;96:11023–11027. doi: 10.1073/pnas.96.20.11023. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. de Boer AG, Gaillard PJ. Blood–brain barrier dysfunction and recovery. J Neural Transm. 2006;113:455–462. doi: 10.1007/s00702-005-0375-4. [DOI] [PubMed] [Google Scholar]
  8. de Castro Ribeiro M, Badaut J, Price M, Meins M, Bogousslavsky J, Monard D, Hirt L. Thrombin in ischemic neuronal death. Exp Neurol. 2006;198:199–203. doi: 10.1016/j.expneurol.2005.11.017. [DOI] [PubMed] [Google Scholar]
  9. Derian CK, Damiano BP, D’Andrea MR, Andrade-Gordon P. Thrombin regulation of cell function through protease-activated receptors: implications for therapeutic intervention. Biochemistry (Mosc) 2002;67:56–64. doi: 10.1023/a:1013900130415. [DOI] [PubMed] [Google Scholar]
  10. Dihanich M, Kaser M, Reinhard E, Cunningham D, Monard D. Prothrombin mRNA is expressed by cells of the nervous system. Neuron. 1991;6:575–581. doi: 10.1016/0896-6273(91)90060-d. [DOI] [PubMed] [Google Scholar]
  11. Esmon CT. Crosstalk between inflammation and thrombosis. Maturitas. 2008;61:122–131. doi: 10.1016/j.maturitas.2008.11.008. [DOI] [PubMed] [Google Scholar]
  12. Eugenin EA, Berman JW. Chemokine-dependent mechanisms of leukocyte trafficking across a model of the blood–brain barrier. Methods. 2003;29:351–361. doi: 10.1016/s1046-2023(02)00359-6. [DOI] [PubMed] [Google Scholar]
  13. Grammas P, Samany PG, Thirumangalakudi L. Thrombin and inflammatory proteins are elevated in Alzheimer’s disease microvessels: implications for disease pathogenesis. J Alzheimers Dis. 2006;9:51–58. doi: 10.3233/jad-2006-9105. [DOI] [PubMed] [Google Scholar]
  14. Greenwood J, Heasman SJ, Alvarez JI, Prat A, Lyck R, Engelhardt B. Review: leucocyte-endothelial cell crosstalk at the blood–brain barrier: a prerequisite for successful immune cell entry to the brain. Neuropathol Appl Neurobiol. 2011;37:24–39. doi: 10.1111/j.1365-2990.2010.01140.x. [DOI] [PubMed] [Google Scholar]
  15. Hamann I, Zipp F, Infante-Duarte C. Therapeutic targeting of chemokine signaling in multiple sclerosis. J Neurol Sci. 2008;274:31–38. doi: 10.1016/j.jns.2008.07.005. [DOI] [PubMed] [Google Scholar]
  16. Ishida Y, Nagai A, Kobayashi S, Kim SU. Upregulation of protease-activated receptor-1 in astrocytes in Parkinson disease: astrocyte-mediated neuroprotection through increased levels of glutathione peroxidase. J Neuropathol Exp Neurol. 2006;65:66–77. doi: 10.1097/01.jnen.0000195941.48033.eb. [DOI] [PubMed] [Google Scholar]
  17. Ishii K, Hein L, Kobilka B, Coughlin SR. Kinetics of thrombin receptor cleavage on intact cells Relation to signaling. J Biol Chem. 1993;268:9780–9786. [PubMed] [Google Scholar]
  18. Kaplanski G, Marin V, Fabrigoule M, Boulay V, Benoliel AM, Bongrand P, Kaplanski S, Farnarier C. Thrombin-activated human endothelial cells support monocyte adhesion in vitro following expression of intercellular adhesion molecule-1 (ICAM-1; CD54) and vascular cell adhesion molecule-1 (VCAM-1; CD106) Blood. 1998;92:1259–1267. [PubMed] [Google Scholar]
  19. Knezevic N, Tauseef M, Thennes T, Mehta D. The G protein betagamma subunit mediates reannealing of adherens junctions to reverse endothelial permeability increase by thrombin. J Exp Med. 2009;206:2761–2777. doi: 10.1084/jem.20090652. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Komarova YA, Mehta D, Malik AB. Dual regulation of endothelial junctional permeability. Sci STKE. 2007:re8. doi: 10.1126/stke.4122007re8. [DOI] [PubMed] [Google Scholar]
  21. Levi M, van der Poll T. Two-way interactions between inflammation and coagulation. Trends Cardiovasc Med. 2005;15:254–259. doi: 10.1016/j.tcm.2005.07.004. [DOI] [PubMed] [Google Scholar]
  22. Liu DZ, Ander BP, Xu H, Shen Y, Kaur P, Deng W, Sharp FR. Blood–brain barrier breakdown and repair by Src after thrombin-induced injury. Ann Neurol. 2010;67:526–533. doi: 10.1002/ana.21924. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Minagar A, Alexander JS. Blood–brain barrier disruption in multiple sclerosis. Mult Scler. 2003;9:540–549. doi: 10.1191/1352458503ms965oa. [DOI] [PubMed] [Google Scholar]
  24. Minami T, Sugiyama A, Wu SQ, Abid R, Kodama T, Aird WC. Thrombin and phenotypic modulation of the endothelium. Arterioscler Thromb Vasc Biol. 2004;24:41–53. doi: 10.1161/01.ATV.0000099880.09014.7D. [DOI] [PubMed] [Google Scholar]
  25. Monroe DM, Hoffman M. What does it take to make the perfect clot? Arterioscler Thromb Vasc Biol. 2006;26:41–48. doi: 10.1161/01.ATV.0000193624.28251.83. [DOI] [PubMed] [Google Scholar]
  26. Naldini A, Pucci A, Carney DH, Fanetti G, Carraro F. Thrombin enhancement of interleukin-1 expression in mononuclear cells: involvement of proteinase-activated receptor-1. Cytokine. 2002;20:191–199. doi: 10.1006/cyto.2002.2001. [DOI] [PubMed] [Google Scholar]
  27. Nan B, Lin P, Lumsden AB, Yao Q, Chen C. Effects of TNF-alpha and curcumin on the expression of thrombomodulin and endothelial protein C receptor in human endothelial cells. Thromb Res. 2005;115:417–426. doi: 10.1016/j.thromres.2004.10.010. [DOI] [PubMed] [Google Scholar]
  28. Okada M, Suzuki K, Takada K, Nakashima M, Nakanishi T, Shinohara T. Detection of up-regulated genes in thrombin-stimulated human umbilical vein endothelial cells. Thromb Res. 2006;118:715–721. doi: 10.1016/j.thromres.2005.11.008. [DOI] [PubMed] [Google Scholar]
  29. Rains JL, Jain SK. Hyperketonemia increases monocyte adhesion to endothelial cells and is mediated by LFA-1 expression in monocytes and ICAM-1 expression in endothelial cells. Am J Physiol Endocrinol Metab. 2011;301:E298–E306. doi: 10.1152/ajpendo.00038.2011. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Riek-Burchardt M, Striggow F, Henrich-Noack P, Reiser G, Reymann KG. Increase of prothrombin-mRNA after global cerebral ischemia in rats, with constant expression of protease nexin-1 and protease-activated receptors. Neurosci Lett. 2002;329:181–184. doi: 10.1016/s0304-3940(02)00645-6. [DOI] [PubMed] [Google Scholar]
  31. Roh HC, Yoo do Y, Ko SH, Kim YJ, Kim JM. Bacteroides fragilis enterotoxin upregulates intercellular adhesion molecule-1 in endothelial cells via an aldose reductase-, MAPK-, and NF-kappaB-dependent pathway, leading to monocyte adhesion to endothelial cells. J Immunol. 2011;187:1931–1941. doi: 10.4049/jimmunol.1101226. [DOI] [PubMed] [Google Scholar]
  32. Rohatgi T, Sedehizade F, Reymann KG, Reiser G. Protease-activated receptors in neuronal development, neurodegeneration, and neuroprotection: thrombin as signaling molecule in the brain. Neuroscientist. 2004;10:501–512. doi: 10.1177/1073858404269955. [DOI] [PubMed] [Google Scholar]
  33. Semple BD, Kossmann T, Morganti-Kossmann MC. Role of chemokines in CNS health and pathology: a focus on the CCL2/CCR2 and CXCL8/CXCR2 networks. J Cereb Blood Flow Metab. 2010;30:459–473. doi: 10.1038/jcbfm.2009.240. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Shan S, Hong-Min T, Yi F, Jun-Peng G, Yue F, Yan-Hong T, Yun-Ke Y, Wen-Wei L, Xiang-Yu W, Jun M, Guo-Hua W, Ya-Ling H, Hua-Wei L, Ding-Fang C. New evidences for fractalkine/CX3CL1 involved in substantia nigral microglial activation and behavioral changes in a rat model of Parkinson’s disease. Neurobiol Aging. 2009;32:443–458. doi: 10.1016/j.neurobiolaging.2009.03.004. [DOI] [PubMed] [Google Scholar]
  35. Smedlund K, Tano JY, Vazquez G. The constitutive function of native TRPC3 channels modulates vascular cell adhesion molecule-1 expression in coronary endothelial cells through nuclear factor kappaB signaling. Circ Res. 2010;106:1479–1488. doi: 10.1161/CIRCRESAHA.109.213314. [DOI] [PubMed] [Google Scholar]
  36. Sokolova E, Reiser G. Prothrombin/thrombin and the thrombin receptors PAR-1 and PAR-4 in the brain: localization, expression and participation in neurodegenerative diseases. Thromb Haemost. 2008;100:576–581. [PubMed] [Google Scholar]
  37. Suo Z, Citron BA, Festoff BW. Thrombin: a potential proinflammatory mediator in neurotrauma and neurodegenerative disorders. Curr Drug Targets Inflamm Allergy. 2004;3:105–114. doi: 10.2174/1568010043483953. [DOI] [PubMed] [Google Scholar]
  38. Szaba FM, Smiley ST. Roles for thrombin and fibrin(ogen) in cytokine/chemokine production and macrophage adhesion in vivo. Blood. 2002;99:1053–1059. doi: 10.1182/blood.v99.3.1053. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Tiruppathi C, Naqvi T, Sandoval R, Mehta D, Malik AB. Synergistic effects of tumor necrosis factor-alpha and thrombin in increasing endothelial permeability. Am J Physiol Lung Cell Mol Physiol. 2001;281:L958–L968. doi: 10.1152/ajplung.2001.281.4.L958. [DOI] [PubMed] [Google Scholar]
  40. Van de Wouwer M, Collen D, Conway EM. Thrombomodulin-protein C-EPCR system: integrated to regulate coagulation and inflammation. Arterioscler Thromb Vasc Biol. 2004;24:1374–1383. doi: 10.1161/01.ATV.0000134298.25489.92. [DOI] [PubMed] [Google Scholar]
  41. Webb AA, Muir GD. The blood–brain barrier and its role in inflammation. J Vet Intern Med. 2000;14:399–411. [PubMed] [Google Scholar]
  42. Weksler BB, Subileau EA, Perriere N, Charneau P, Holloway K, Leveque M, Tricoire-Leignel H, Nicotra A, Bourdoulous S, Turowski P, Male DK, Roux F, Greenwood J, Romero IA, Couraud PO. Blood–brain barrier-specific properties of a human adult brain endothelial cell line. FASEB J. 2005;19:1872–1874. doi: 10.1096/fj.04-3458fje. [DOI] [PubMed] [Google Scholar]
  43. White GE, Greaves DR. Fractalkine: one chemokine, many functions. Blood. 2009;113:767–768. doi: 10.1182/blood-2008-11-189860. [DOI] [PubMed] [Google Scholar]
  44. Woolkalis MJ, DeMelfi TM, Jr, Blanchard N, Hoxie JA, Brass LF. Regulation of thrombin receptors on human umbilical vein endothelial cells. J Biol Chem. 1995;270:9868–9875. doi: 10.1074/jbc.270.17.9868. [DOI] [PubMed] [Google Scholar]
  45. Yamada T, Nagai Y. Immunohistochemical studies of human tissues with antibody to factor Xa. Histochem J. 1996;28:73–77. doi: 10.1007/BF02331429. [DOI] [PubMed] [Google Scholar]
  46. Yanagita M, Kobayashi R, Kashiwagi Y, Shimabukuro Y, Murakami S. Thrombin regulates the function of human blood dendritic cells. Biochem Biophys Res Commun. 2007;364:318–324. doi: 10.1016/j.bbrc.2007.10.002. [DOI] [PubMed] [Google Scholar]

RESOURCES