Comparison of Immortalized bEnd5 and Primary Mouse Brain Microvascular Endothelial Cells as in vitro Blood–Brain Barrier Models for the Study of T Cell Extravasation

Abstract

Important insights into the molecular mechanism of T cell extravasation across the blood–brain barrier (BBB) have already been obtained using immortalized mouse brain endothelioma cell lines (bEnd). However, compared with bEnd, primary brain endothelial cells have been shown to establish better barrier characteristics, including complex tight junctions and low permeability. In this study, we asked whether bEnd5 and primary mouse brain microvascular endothelial cells (pMBMECs) were equally suited as in vitro models with which to study the cellular and molecular mechanisms of T cell extravasation across the BBB. We found that both in vitro BBB models equally supported both T cell adhesion under static and physiologic flow conditions, and T cell crawling on the endothelial surface against the direction of flow. In contrast, distances of T cell crawling on pMBMECs were strikingly longer than on bEnd5, whereas diapedesis of T cells across pMBMECs was dramatically reduced compared with bEnd5. Thus, both in vitro BBB models are suited to study T cell adhesion. However, because pMBMECs better reflect endothelial BBB specialization in vivo, we propose that more reliable information about the cellular and molecular mechanisms of T cell diapedesis across the BBB can be attained using pMBMECs.

Keywords

live-cell imaging physiologic shear flow T cell adhesion T cell crawling T cell diapedesis

Introduction

Homeostasis of the central nervous system (CNS) is critical for the proper function of neuronal cells. The endothelial blood–brain barrier (BBB) protects the CNS from the continuously changing milieu of the periphery by inhibiting the free paracellular diffusion of molecules into the CNS (through an elaborate network of tight junctions between endothelial cells), and their transcellular passage into the CNS (by extremely low pinocytotic activity) (Abbott et al, 2010). In addition, the BBB strictly controls immune cell entry into the CNS, which is low under physiologic conditions. Under various pathologic conditions in the CNS—such as multiple sclerosis or its animal model experimental autoimmune encephalomyelitis—leukocytes traverse the BBB and cause CNS inflammation. In fact, when injected into susceptible animals, in vitro activated myelin-specific T cells extravasate across the BBB into the CNS, where they induce experimental autoimmune encephalomyelitis (Engelhardt, 2008). Thus, interaction of circulating immunocompetent cells with the BBB endothelium is a critical step for CNS immunosurveillance, and also in the pathogenesis of CNS inflammation.

Extravasation of T cells has been characterized as a multistep process that involves T cell rolling along the vascular surface, T cell arrest and crawling on the endothelium, and diapedesis of T cells (Ley et al, 2007). Specific knowledge about the molecular mechanisms of immune cell extravasation across the BBB has been provided by a number of in vitro studies that used immortalized brain endothelial cell lines from not only rats and mice—as reviewed by Turowski et al (2005)—but also from humans (Afonso et al, 2007; Bahbouhi et al, 2009). In particular, we have previously used polyoma middle T-oncogen-immortalized mouse brain endothelioma cell lines (bEnd) to show that vascular endothelial cell adhesion molecule (VCAM)-1 is involved in T cell adhesion to, but not in T cell diapedesis across the BBB (Laschinger and Engelhardt, 2000), whereas endothelial intercellular cell adhesion molecule (ICAM)-1 and ICAM-2 are essential for T cell diapedesis across the BBB under static conditions in vitro (Lyck et al, 2003; Reiss et al, 1998; Reiss and Engelhardt, 1999).

With the aid of modern imaging tools, a distinct crawling behavior of T cells before their diapedesis was subsequently identified. In vitro, T cell crawling has been visualized on immortalized rat brain endothelial cells (Martinelli et al, 2009) or on primary human nonbrain endothelial cells (Carman et al, 2007; Shulman et al, 2009; Stanley et al, 2008). Very recently, Bartholomäus et al (2009) showed T cell crawling on the luminal face of leptomeningeal vessels in vivo. Subsequent to crawling, T cell diapedesis across the endothelium occurs either paracellular (i.e., through cell–cell junctions) or transcellular (i.e., through the endothelial cell body). For some time, it was generally held that the paracellular route would be favored. However, recent in vitro imaging studies have shown that both passageways exist. In fact, direct comparisons of endothelial cells isolated from different vascular beds led to the conclusion that the type of endothelium has a role in specifying the cellular route of T cell diapedesis (Carman, 2009). In particular, several electron microscopic studies have documented an exclusive transcellular migration across the BBB endothelium in vivo (Carman, 2009; Engelhardt and Wolburg, 2004).

As immortalized brain endothelial cell lines have lost the important BBB-specific property of being able to form a tight permeability barrier, BBB models consisting of primary brain endothelial cells have gained popularity for studying transport mechanisms across the BBB in vitro (Deli et al, 2005). These well-established in vitro BBB models use primary brain endothelial cells from various sources, such as bovine, porcine, rat, and even human brain tissues (Deli et al, 2005). However, to date, few mouse BBB in vitro models using primary brain endothelial cells have been established. Coisne et al (2005) recently described an in vitro model using primary mouse brain microvascular endothelial cells (pMBMECs) in coculture with primary mouse glial cells. Primary mouse brain microvascular endothelial cells form differentiated endothelial monolayers that retain numerous phenotypic properties of the CNS microvasculature in vitro, such as complex tight junctions and barrier formation (Coisne et al, 2005, 2006; Lyck et al, 2009).

Both pMBMECs and bEnd5—a polyoma middle T-oncogen-immortalized mouse brain endothelioma cell line—are well characterized for their expression of endothelial cell specific proteins, namely vascular endothelial-cadherin (VE-cadherin), von Willebrand factor, platelet endothelial cell adhesion molecule-1, endoglin, ICAM-2, and claudin-5. Moreover, both pMBMECs and bEnd5 have previously been characterized for cytokine-induced upregulation of P-selectin, VCAM-1, and ICAM-1, all three of which are known to be involved in T cell interaction with the brain endothelium upon proinflammatory stimuli (Coisne et al, 2005, 2006; Reiss et al, 1998; Rohnelt et al, 1997). However, although the expression of the tight junction protein occludin has been described for both in vitro BBB models (Coisne et al, 2005; Yang et al, 2007), we recently found a dramatically reduced occludin mRNA expression level in bEnd5 as compared with pMBMECs (Lyck et al, 2009).

For this study, we compared the suitability of bEnd5 and pMBMECs as model systems for the study of the cellular and molecular mechanisms involved in the multistep process of T cell extravasation across the BBB in vitro. We show that unlike pMBMECs, bEnd5 lacked localization of occludin to cellular junctions, and also failed to establish a tight permeability barrier. Thus as in vitro model pMBMECs better mimicked important BBB features than did bEnd5. Nevertheless, both in vitro BBB models equally promoted T cell adhesion under static and physiologic flow conditions, and comparably allowed T cell crawling with a preferential direction against the flow. In contrast, crawling distances of T cells on pMBMECs before diapedesis were significantly longer than on bEnd5, and T cell diapedesis rates across pMBMECs were reduced compared with bEnd5. Hence, our study shows that although both in vitro BBB models are appropriate to study T cell adhesion to the BBB, more reliable information about the cellular and molecular mechanisms involved in T cell diapedesis across the tight BBB endothelium seems to be obtained by using pMBMECs.

Materials and methods

Antibodies and Reagents

Rabbit antibodies against mouse claudin-5 and occludin were purchased from Zymed (Invitrogen, Basel, Switzerland). The hybridoma 11D4.1 (anti-mouse VE-cadherin) was a kind gift from D. Vestweber (Münster, Germany). Secondary antibodies AlexaFluor-488-conjugated goat anti-rat and goat anti-rabbit IgG and Cy-3-conjugated goat anti-rat IgG, as well as rhodamine-phalloidin were obtained from Molecular Probes (LuBioScience, Luzern, Switzerland). Fetal calf serum was obtained from Biowest (TECOmedical AG, Sissach, Switzerland), basic fibroblast growth factor was from Sigma (Buchs, Switzerland), and all other cell culture medium components were from Invitrogen.

T Cells

The proteolipid protein (PLP)-specific CD4⁺ T_H1 effector/memory T cell line SJL.PLP7 raised against the PLP peptide aa_{139 to 151} has previously been described in detail (Engelhardt et al, 1998). T cells were used 3 days after the third or fourth restimulation with the PLP peptide aa_{139 to 151} antigen.

Endothelioma Cell Line bEnd5

The bEnd5 mouse brain endothelioma cell line was described before (Reiss et al, 1998). The bEnd5 cells were grown in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal calf serum, 4 mmol/L L-glutamine, 1 mmol/L sodium pyruvate, 50 Units/mL penicillin, and 50 μg/mL streptomycin, 1% minimal essential medium nonessential amino acids exactly as described previously (Reiss et al, 1998). For all experiments, bEnd5 between passages 18 and 25 were used and cultured for at least 3 days on laminin (Roche, Basel, Switzerland) coated surfaces. Stimulated bEnd5 cells were cultured for 16 to 18 hours in the presence of recombinant human tumor necrosis factor (TNF)-α (25 ng/mL), kindly provided by Daniela Männel (Regensburg, Germany).

Primary Mouse Brain Microvascular Endothelial Cells

Primary mouse brain microvascular endothelial cells were isolated from gender-matched 4- to 6-week old C57BL/6 mice (Harlan Laboratories, Horst, The Netherlands), cultured in DMEM, 20% fetal calf serum, 1 mmol/L sodium pyruvate, 1% minimal essential medium nonessential amino acids, 50 μg/mL gentamycin, and 1 ng/mL basic fibroblast growth factor (Coisne et al, 2005; Lyck et al, 2009). For standardization of culture conditions between bEnd5 and pMBMECs, we used identical DMEM, and supplements for both types of brain endothelial cells were applicable. Primary mouse brain microvascular endothelial cells were used on day 5 or 6 after isolation. Previously we observed an apparent similarity between pMBMECs cultured with or without glial cells regarding the gene expression profile for BBB-specific proteins, the paracellular permeability and the localization of junctional proteins (Lyck et al, 2009). Thus, we focussed on pMBMECs kept in monoculture and omitted culture conditions with astrocyte-conditioned media throughout all experimental procedures in this study. Stimulated pMBMECs were cultured for 16 to 18 hours in the presence of recombinant human TNF-α (25 ng/mL). All animal procedures were performed in accordance with the Swiss legislation on the protection of animals and approved by the veterinary office of the Kanton of Bern.

Immunofluorescence Stainings

Confluent pMBMEC and bEnd5 monolayers grown on 8-well Lab-Tek chamber slides (Milian SA, Geneva, Switzerland) were fixed either with ice-cold methanol for 30 seconds or in 1% paraformaldehyde in phosphate-buffered saline for 10 minutes. Cells were permeabilized and blocked with 0.2% TX-100 and 10% normal goat serum in phosphate-buffered saline for 20 minutes and stained for 60 minutes with primary antibodies diluted in phosphate-buffered saline containing 10% normal goat serum. After washing, cells were incubated for 60 minutes with secondary antibodies or with rhodamine-phalloidin in phosphate-buffered saline with 10% normal goat serum and mounted with mowiol. Pictures were taken at a × 400 magnification using a fluorescence microscope (Eclipse E600 Nikon, Tokyo, Japan).

Evaluation of Density of Endothelial Cell–Cell Junctions

To assess the mean frequency of cell–cell junctions theoretically encountered by T cells crawling on the surface of bEnd5 or pMBMEC monolayers, we evaluated for each type of endothelial cells 10 independent immunofluorescent images obtained from cultures stained for junctional proteins (such as claudin-5, claudin-3, VE-cadherin). For counting the junctions, we placed a starlike construct with four lines of 100 μm in length each on the endothelial surface. The length of the lines was chosen on the basis of the average distance covered by T cells crawling on the brain endothelium over 30 minutes. The lines crossed in their center points and were shifted to angles of 45°, 90°, and 135° with regard to the first line. We overlaid each immunofluorescence image with three of these counting constructs arranged in a diagonally and nonoverlapping manner to achieve triplicate values for each image (Supplementary Figure S1). After manual counting of all junctions crossed by each line star, we calculated the mean number of junctions met by one such line star for bEnd5 (34.8 ± 5.0) and for pMBMECs (17.7 ± 2.0). We concluded that the relative density of cell–cell junctions is two-fold higher on bEnd5 than on pMBMECs.

Permeability Assay

Permeability assays were performed in triplicate as published previously (Coisne et al, 2005) with minor adaptations: pMBMEC and bEnd5 cells were grown on matrigel- or laminin-coated filter inserts (Transwell, 0.4 μm pore size, 6.5 mm diameter, VITARIS AG, Baar, Switzerland). After washing endothelial cells with wash buffer (HBSS, 10% calf serum, 25 mmol/L Hepes pH 7.2 to 7.5) permeability to AlexaFluor-680-dextran (3 kDa, 10 μg/mL, LuBioScience, Luzern, Switzerland) was measured in the presence of the migration assay medium (DMEM, 5% calf serum, 4 mmol/L L-glutamine, 25 mmol/L Hepes pH 7.2 to 7.5). Endothelial permeability was specifically measured at 16 to 18 hours of TNF-α stimulation of endothelial cells, because this was the optimal time point for TNF-α-induced upregulation of ICAM-1 and VCAM-1 on bEnd5 and pMBMECs and therefore the time point when T cell interaction assays were performed. Measuring the permeability at this time point allowed to correlate T cell diapedesis rates with endothelial monolayer permeability. Diffused AlexaFluor-680-dextran was quantified using the Odyssey Infrared Imaging System (LI-COR, Bad Homburg, Germany).

Static T cell Adhesion and Diapedesis Assays

T cell adhesion and diapedesis assays were performed as described before (Reiss et al, 1998; Rohnelt et al, 1997), except for an elongated time period to 6 hours of T cell diapedesis. For all experiments, endothelial cells were once washed with the wash buffer, and the assays were carried out under absolutely identical conditions for both types of endothelial cells in the presence of the migration assay medium (DMEM, 5% calf serum, 4 mmol/L L-glutamine, 25 mmol/L Hepes pH 7.2 to 7.5).

Live-Cell Imaging Under Shear Flow

For live-cell imaging, a parallel flow chamber (Stein et al, 2003) connected to an automated syringe pump (Harvard Apparatus, Holliston, MA, USA) was mounted on TNF-α-stimulated pMBMEC and bEnd5 cells, and placed on a heating stage of an inverted microscope (Axiovert 200, Carl Zeiss AG, Feldbach, Switzerland). Shear stress (dyn/cm²) was calculated according to τ = 3 μQ/2a²b (τ = wall shear stress, μ = coefficient of viscosity, Q = volumetric flow rate, a = half channel height, b = channel width) (Lawrence et al, 1990). The migration assay medium was used during the complete assay procedure and all conditions were kept identical for both types of endothelial cells throughout the complete experiment. T cells (5 × 10⁵ per mL) were allowed to accumulate for 4 minutes at low shear stress (0.25 dyn/cm²). Subsequently, dynamic T cell interactions with either pMBMECs or bEnd5 were recorded under physiologic shear stress (1.5 dyn/cm²) at × 100 magnification (objective: A-Plan 10 × /0.25), using a monochrome CCD camera (Cohu, San Diego, CA, USA) connected to a digital video recording system (Sony, DVCAM DSR-11, Kilchemann AG, Kchrsatz, Switzerland) and converted to time-lapse image sequences of one frame every 30 seconds (iMovie, Apple, Cupertino, CA, USA).

Every T cell that stopped on the brain endothelium in the field of view (FOV) (963 × 642 μm²) during the accumulation phase (0.25 dyn/cm²), and which resisted immediate detachment after increasing the shear flow rate to 1.5 dyn/cm², was counted as an arrested T cell (first frame after flow enhancement). Arrested T cells were then tracked manually, using ImageJ software (National Institute of Health, Bethesda, MD, USA) using the manual tracking and chemotaxis plugin. Arrested T cells that detached, entered, or left the FOV during the recording time were excluded from tracking.

Statistical Analyses

Statistical analysis was performed using Prism 5 software (Graphpad software, La Jolla, CA, USA). Differences between two groups were analyzed by the unpaired Student's t-test. P < 0.05 was considered as significant (*P < 0.05; **P < 0.01; ***P < 0.001). Mean values are expressed with s.e.m., except for representative experiments, in which mean values are expressed with s.d. of the mean.

Results

Primary Mouse Brain Microvascular Endothelial Cells and bEnd5 Differ in Both the Architecture of Their F-actin Cytoskeleton and Their Tight Junctions

When working with bEnd5 and pMBMECs, we observed a more spindle-shaped cell morphology for bEnd5 than for pMBMECs. When comparing F-actin stainings, we found that in pMBMECs, the F-actin cytoskeleton was organized in cortical rings, whereas in bEnd5 cells, F-actin presented as parallel stress fibers oriented longitudinally throughout the entire cell body (Figure 1). Thus, both in vitro BBB models clearly differ with regard to the architecture of their F-actin cytoskeleton.

Figure 1

Differences in junctional protein expression and F-actin cytoskeleton between pMBMECs and bEnd5. Images of pMBMECs and bEnd5 fluorescently stained for F-actin (rhodamine-phalloidin), VE-cadherin, claudin-5, and occludin are shown. Image acquisition: Plan Fluor 40 × /0.75 objective, Eclipse E600 Nikon. Bar, 20 μm. bEnd5, brain endothelioma cell line-5; pMBMEC, primary mouse brain microvascular endothelial cell; VE-cadherin, vascular endothelial-cadherin.

To better understand the different appearance of bEnd5 and pMBMECs, we next investigated the molecular composition of cellular junctions of both in vitro BBB models. In both bEnd5 and pMBMECs, the distribution of the transmembrane adherens junction protein VE-cadherin and the transmembrane tight junction protein claudin-5 proved similar, and was found to be restricted to cell–cell junctions (Figure 1). Similarly, the scaffolding proteins zonula occludens-1 and zonula occludens-2 showed junctional localization in both in vitro BBB models (data not shown). However, the tight junction protein occludin exclusively localized to cellular junctions in pMBMECs, but not in bEnd5, in which occludin staining was additionally observed within the cytoplasm (Figure 1). After combining these findings with our previous observation that gene expression levels of occludin mRNA were lower in bEnd5 than in pMBMECs (Lyck et al, 2009), we concluded that the junctional organization of bEnd5 differs clearly from that of pMBMECs.

Primary Mouse Brain Microvascular Endothelial Cells and bEnd5 Differ in Their Functional Barrier Characteristics

As the appearance of stress fibers has been shown to impair endothelial barrier formation (Bogatcheva and Verin, 2008), we compared the paracellular permeabilities of pMBMECs and bEnd5. Using fluorescently labeled 3 kDa dextran, we measured a permeability value of Pe_{3 kDa} = 0.13 ± 0.02 × 10⁻³ cm/min for pMBMECs. By contrast, bEnd5 failed to establish a comparable tight diffusion barrier for 3 kDa dextran: The permeability value of Pe_{3 kDa} 2.45 ± 0.33 × 10⁻³ cm/min was almost 19-fold higher than the value measured for pMBMECs (Figure 2). As various studies have reported that inflammatory mediators enhance the permeability of endothelial cell monolayers (Deli et al, 2005; Vandenbroucke et al, 2008), we next examined whether the proinflammatory cytokine TNF-α would alter the permeability characteristics of pMBMEC or bEnd5 cells. We found that incubation of pMBMECs or bEnd5 with TNF-α for 16 hours did not significantly influence the paracellular barrier characteristics of pMBMEC (Pe_{3 kDa} 0.13 ± 0.04 × 10⁻³ cm/min) or bEnd5 monolayers (Pe_{3 kDa} 1.59 ± 0.22 × 10⁻³ cm/min) (Figure 2). Taken together, these data show that unlike pMBMECs, bEnd5 cells failed to establish a similar restrictive permeability barrier in vitro, in accordance with their incomplete tight junction organization.

Figure 2

Monolayer permeabilities of pMBMECs and bEnd5 are different, but remain unchanged upon TNF-α stimulation. The permeability coefficients of 3 kDa (Pe_{3 kDa}) dextran for pMBMECs either unstimulated (w/o TNF-α) or TNF-α stimulated (+ TNF-α) are shown. The extreme differences in permeability between pMBMECs and bEnd5, indicated by a broken y axis must be noted. Pe_{3 kDa} values were calculated from diffused AlexaFluor-680-dextran at 4 time points of 10-minute intervals as published previously (Cecchelli et al, 1999). Bars represent mean ± s.e.m. of three or more independent experiments. ***P < 0.001, unpaired Student's t-test. bEnd5, brain endothelioma cell line-5; pMBMEC, primary mouse brain microvascular endothelial cell; TNF-α, tumor necrosis factor-α.

Primary Mouse Brain Microvascular Endothelial Cells and bEnd5 Equally Promote T cell Adhesion Under Static Conditions

Given the observed differences between bEnd5 and pMBMECs, we next examined whether both in vitro BBB models were equally suited to study T cell adhesion to the BBB. Adhesion of T cells to the brain endothelium depends on the presence of the cell adhesion molecules VCAM-1, ICAM-1, and ICAM-2 on the endothelium. Both bEnd5 and pMBMECs were shown to constitutively express ICAM-2 and to upregulate the cell surface expression of VCAM-1 and ICAM-1 upon TNF-α stimulation in a similar manner (Coisne et al, 2005, 2006; Reiss et al, 1998; Rohnelt et al, 1997). In this study, this was confirmed by immunofluorescence stainings (Figure 3C). To investigate T cell adhesion to both in vitro BBB models, we compared adhesion of PLP-specific CD4⁺ T_H1 cells to bEnd5 and pMBMECs under static conditions. Similar numbers of T cells were found to firmly adhere to unstimulated bEnd5 and pMBMECs, as shown in Figure 3A for a representative experiment with 114 ± 17 T cells per FOV for bEnd5 and 133 ± 3 T cells per FOV for pMBMECs. Tumor necrosis factor-α stimulation of endothelial cells increased adhesion of T cells to bEnd5 (222 ± 22 adherent T cells per FOV) and pMBMECs (226 ± 20 T cells per FOV) (Figure 3A). Taken together, in accordance with a comparable adhesion molecule expression profile, similar numbers of T cells were found to adhere to bEnd5 and pMBMECs under unstimulated or TNF-α-stimulated conditions. Hence, both models are equally suited for the study of the cellular and molecular mechanisms of T cell adhesion to the BBB in vitro under static conditions.

Figure 3

T cell adhesion to pMBMECs and bEnd5 is comparable, but diapedesis of T cells is enhanced across bEnd5 under static conditions. T cell adhesion (A) to and diapedesis (B) across unstimulated or TNF-α-stimulated pMBMECs and bEnd5 are shown. (Panel A) T cell adhesion under static conditions to pMBMECs and bEnd5 was determined by counting numbers of adherent T cells per field of view (FOV = 600 × 600 μm²). One representative experiment is depicted. Bars represent mean ± s.d. of triplicate values. (Panel B) T cell diapedesis across pMBMECs and bEnd5 was measured in a static two-chamber-based experimental setup. Migration time was 6 hours. One representative experiment is depicted. The number of T cells added to the upper chamber was set at 100% (1 × 10⁵ T cells) and values of diapedesed T cells are expressed as percentage of input. Bars represent mean ± s.d. of triplicate values. Asterisks above columns show significant differences between unstimulated and TNF-α-stimulated conditions, asterisk above crossbars show significant differences between pMBMECs and bEnd5 under identical conditions. (Panel C) Images of unstimulated and TNF-α-stimulated pMBMECs and bEnd5 fluorescently stained for ICAM-1 and VCAM-1, respectively. Overall, 16 to 18 hours TNF-α stimulation clearly upregulates the expression of both cell adhesion molecules. Image acquisition: Plan Fluor 40 ×/0.75 objective, Eclipse E600 Nikon. Bar, 20 μm, *P < 0.05, **P < 0.01, ***P < 0.001, unpaired Student's t-test. bEnd5, brain endothelioma cell line-5; ICAM-1, intercellular cell adhesion molecule-1; pMBMEC, primary mouse brain microvascular endothelial cell; TNF-α, tumor necrosis factor-α; VCAM-1, vascular endothelial cell adhesion molecule.

T cell Diapedesis Across Primary Mouse Brain Microvascular Endothelial Cells is Severely Reduced Under Static Conditions

Given the similar numbers of T cells adhering to both bEnd5 and pMBMECs, we next examined whether T cell diapedesis across both in vitro BBB models would also be comparable. To this end, we compared T cell migration across bEnd5 and pMBMECs under static conditions in a two-chamber assay system, as described previously (Rohnelt et al, 1997). In this case, we observed a striking difference between bEnd5 and pMBMECs. Whereas 32.3% ± 7.6% of T cells migrated across unstimulated bEnd5, only 4.3% ± 0.7% of T cells migrated across unstimulated pMBMECs (Figure 3B). Tumor necrosis factor-α stimulation increased T cell diapedesis across both in vitro BBB models to 58.1% ± 11.8% and 12.6% ± 0.3% of T cells that diapedesed across TNF-α-stimulated bEnd5 or TNF-α-stimulated pMBMECs, respectively (Figure 3B). Nevertheless, overall T cell diapedesis rates across both types of BBB endothelial cells still differed greatly by a factor of 4.6. Taken together, in correlation with their reduced paracellular permeability, pMBMECs were found to be less penetrable for T cells than bEnd5 under both inflammatory and noninflammatory conditions. Considering the smaller cell size and therefore higher junctional frequency in monolayers of bEnd5 compared with pMBMECs (Figure 1) enhanced T cell diapedesis across bEnd5 could be a consequence of higher paracellular T cell diapedesis rates across more frequent cell–cell junctions. Quantitative evaluation of cell–cell junctions in endothelial monolayers showed a 2.0-fold inreased frequency of cell–cell junctions on bEnd5 compared with pMBMECs (Supplementary Figure S1). Thus, the enhanced density of cell–cell junctions can only partly explain the differences in T cell diapedesis observed across both types of endothelial cells.

Primary Mouse Brain Microvascular Endothelial Cells and bEnd5 Equally Promote T Cell Arrest and Crawling—But Not T Cell Diapedesis Under Physiologic Shear Flow

The differences found between pMBMECs and bEnd5 in the static diapedesis experiment suggested that postarrest T cell–BBB interactions might not be equally modeled by both in vitro BBB setups. As physiologic shear flow has been shown to promote T cell diapedesis across endothelial cells in vitro (Cinamon et al, 2001), we next investigated T cell interaction with TNF-α-stimulated bEnd5 and TNF-α-stimulated pMBMECs under physiologic shear by live-cell imaging. T cells were first allowed to accumulate on the surfaces of bEnd5 or pMBMECs at low shear stress (0.25 dyn/cm²). Next, the dynamic interaction of T cells with bEnd5 or pMBMECs under physiologic shear stress (1.5 dyn/cm²) was recorded. Comparable numbers of T cells were found to arrest on bEnd5 (67 ± 17) and pMBMECs (87 ± 7) per FOV (Figure 4A), supporting the notion that both in vitro BBB models are equally suited to study T cell adhesion to the BBB.

Figure 4

Comparable T cell arrest on, but severe differences in, the dynamic T cell interaction with pMBMECs and bEnd5 under physiologic shear. Shear-resistant T cell arrest (A) on and (B) dynamic T cell interaction with TNF-α-stimulated pMBMECs or bEnd5, respectively, are shown. (Panel A) Numbers of arrested T cells per FOV (963 × 642 μm²) on pMBMECs and bEnd5 that formed initial contact with the endothelium during the accumulation phase (0.25 dyn/cm²) and resisted immediate detachment under enhanced shear stress (1.5 dyn/cm²) were counted by evaluation of the first image aquired after flow enhancement. Bars represent mean ± s.e.m. from three independent experiments of each group. (Panel B) Evaluation of dynamic T cell interactions with pMBMECs and bEnd5 under flow conditions during a period of 15 minutes is shown. The behavior of each arrested T cell was analyzed by eye and assigned to one category and expressed in percentage of initially arrested T cells. Arrested T cells that crawled into or out of the FOV during recording time were excluded from analysis. ‘Crawling and diapedesis’: T cells that polarized and crawled until they finally crossed the endothelial cell monolayer. ‘Continuous crawling’: T cells that polarized and crawled at least two T cell diameters but did not diapedese across the endothelium. ‘Detachment’: T cells that detached during the evaluation period. ‘Stationary’: T cells that remained stationary and did not polarize. Bars represent mean ± s.e.m. from three independent experiments for each group. **P < 0.01, unpaired Student's t-test. bEnd5, brain endothelioma cell line-5; FOV, field of view; pMBMEC, primary mouse brain microvascular endothelial cell; TNF-α, tumor necrosis factor-α.

After the initial shear-resistant arrest on bEnd5 or pMBMECs, most T cells acquired a polarized cell shape and started to crawl on endothelial surfaces. To compare the respective T cell behavior on bEnd5 or pMBMECs in a quantitative manner, we performed a visual frame-by-frame offline analysis of the time-lapse videos. The number of T cells arrested on bEnd5 or pMBMECs per FOV was set at 100%, and each T cell was categorized into a behavioral group as follows: Most T cells either crawled for a certain distance and then underwent diapedesis (group 1), or crawled on the surfaces of bEnd5 or pMBMECs continuously during the complete observation period (group 2). A few T cells were observed to detach owing to low shear resistance (group 3), or to remain stationary without polarized cell morphology (group 4) (Figure 4B).

Interestingly, the majority of T cells (70.9% ± 5.2%) crawled and successfully diapedesed across the bEnd5 monolayer within an observation period of 15 minutes. Only 21.6% ± 3.9% of T cells failed to diapedese, and remained continuously crawling on the monolayer of bEnd5. This contrasted strongly with the results for pMBMECs, in which the majority of T cells (64.7% ± 4.6%) continuously crawled on the endothelial surface. Only 29.1% ± 5.7% of T cells diapedesed after a distinct crawling phase across the more tightened pMBMEC monolayer (Figure 4B). Thus, 2.4-fold more T cells diapedesed across bEnd5 than across pMBMECs. Even if T cells would exclusively take the paracellular pathway across both types of the brain endothelium, the 2.0-fold increased junctional density of monolayers formed by bEnd5 compared with monolayers formed by pMBMECs cannot be the only explanation for the obvious failure of T cells to find sites permissive for diapedesis across pMBMECs.

T Cell Crawling Velocity and Distance are Increased on Primary Mouse Brain Microvascular Endothelial Cells

Reduced T cell diapedesis across pMBMECs compared with bEnd5 could be caused by either a lower frequency of sites permissive for T cell diapedesis or by a reduced T cell crawling velocity. To address this, we manually tracked the crawling paths of individual T cells on bEnd5 and pMBMECs (Figure 5A). From these tracks, we then calculated the respective crawling velocity and crawling distance of each T cell from its initial site of shear-resistant arrest to the site of its diapedesis, or until the end of the observation period (Figures 5B and 5C). On bEnd5, T cells crawled a mean distance of 25.3 ± 2.1 μm before diapedesis, whereas on pMBMECs, T cells crawled significantly longer distances of 55.2 ± 3.0 μm before diapedesis (Figure 5B). The specific behavior of continuous crawling of T cells that did not diapedese within the observation period was documented by the increased distances covered by these T cells with 105.0 ± 7.3 μm on bEnd5 and 136.9 ± 5.2 μm on pMBMECs, compared with the respective crawling distances of T cells that finally diapedesed (Figure 5B).

Figure 5

Detailed analysis of T cell crawling on pMBMECs or bEnd5: Equal crawling against the direction of flow but different crawling distances and velocities. (A) T cell crawling tracks, (B) distances of T cell crawling, (C) T cell crawling velocity, and (D) T cell crawling directionality against shear forces on TNF-α-stimulated pMBMECs or bEnd5 within an observation period of 30 minutes are shown. For evaluations, T cells were divided into two groups: group 1 represents T cells that crawled and finally underwent diapedesis, and group 2 describes T cells that continuously crawled during the total observation period. Crawling paths of T cells on pMBMECs and bEnd5 were manually tracked. (Panel A) Diagrams of T cell crawling tracks (group 1, upper row; group 2, lower row) are depicted for one representative experiment each. The initial site of T cell arrest was set to the center point of the respective diagram. End points of tracks are indicated by dots and correspond to the relative site of diapedesis (upper row) or to the T cell position after 30 minutes (lower row). Open arrow shows the direction of shear flow. (Panel B) Mean T cell crawling distances (μm) from the site of initial arrest to the site of diapedesis (group 1, left) or to the end point after 30 minutes (group 2, right) were calculated from three independent movies of pMBMECs or bEnd5. Each data point represents the crawling distance of one T cell. T cells that remained stationary were excluded from this analysis. **P < 0.01; ***P < 0.001, unpaired Student's t-test. (Panel C) Mean T cell crawling velocities (μm/min) for T cells which crawled and finally diapedesed (group 1, left) or for T cells that continuously crawled (group 2, right) were calculated from three independent movies of pMBMECs and bEnd5, respectively. T cells that remained stationary were excluded from this analysis. Each data point represents the velocity of one T cell. **P < 0.01; ***P < 0.001, unpaired Student's t-test. (Panel D) Directionality with regard to shear forces of T cell crawling expressed as x-forward migration index (xFMI = D_x/D_acc; D_x: straight x axis distance covered by the T cell, D_acc: accumulated total distance of T cell movement). With relation to the FOV, the direction of flow was along the x axis from plus to minus (open arrow). Therefore, a positive xFMI represents a directed crawling against the orientation of shear flow. The xFMI was calculated from T cell crawling tracks (group 1) of three independent movies per type of the endothelium. Data are expressed as mean ± s.e.m. *P < 0.05, one sample t-test against xFMI = 0. bEnd5, brain endothelioma cell line-5; FOV, field of view; pMBMEC, primary mouse brain microvascular endothelial cell; TNF-α, tumor necrosis factor-α.

At 2.9 ± 0.1 μm/min, the mean crawling velocity of T cells that finally underwent diapedesis on bEnd5 was significantly lower than the respective T cell crawling velocity on pMBMECs (3.7 ± 0.1 μm/min, Figure 5C). This finding clearly showed that the reduced T cell diapedesis rate across pMBMECs was rather caused by a lower frequency of permissive sites than by impaired T cell crawling. Interestingly, T cells that failed to diapedese (and therefore continuously crawled on endothelial surfaces) also crawled at a significantly lower velocity on bEnd5 than on pMBMECs (Figure 5C).

Finally, to exclude any influence of the crawling directionality against shear forces on T cell diapedesis rates, we analyzed the direction of T cell crawling with regard to the orientation of the shear flow. To this end, we calculated the FMI (forward migration index) of T cell crawling tracks toward the x axis, which was orientated along the direction of flow in the FOV. On both types of brain endothelial cells, T cells crawled preferentially against the direction of flow, as documented by an xFMI of 0.2 ± 0.07 for bEnd5 and 0.15 ± 0.03 for pMBMECs (Figure 5D). Taken together, our data show that both types of brain endothelial cells allowed for equal T cell adhesion behavior and the same directionality of T cell crawling against shear forces. By contrast, pMBMECs supported an enhanced T cell crawling velocity combined with a reduced T cell diapedesis rate compared with bEnd5. Therefore, a lower number of permissive sites for T cell diapedesis must be concluded for pMBMECs.

Discussion

Studies of the molecular and cellular mechanisms of T cell extravasation across the BBB are facilitated by in vitro BBB models that allow multiple conditions to be studied in parallel, and are easy accessible for microscopic documentation of the ongoing process. Of the broad panel of established in vitro BBB models, mouse models are advantageous because of the availability of transgenic and gene-targeted animals, and the wide range of antibodies. Previously, we have successfully used mouse brain endothelioma cell lines to analyze the important roles of endothelial VCAM-1, ICAM-1, and ICAM-2 for T cell adhesion and diapedesis across the BBB in vitro under static conditions (Laschinger and Engelhardt, 2000; Lyck et al, 2003; Reiss et al, 1998; Reiss and Engelhardt, 1999). Whereas these immortalized brain endothelial cells are available in convenient quantities without time-consuming isolation procedures, BBB models based on primary brain endothelial cells qualify by virtue of their enhanced tight junction complexity, increased monolayer tightness, and more appropriate BBB-like phenotype (Coisne et al, 2005; Deli et al, 2005; Lyck et al, 2009). Given the differences between primary brain endothelial cells and brain endothelioma cell lines, we analyzed the suitability of both systems for studying T cell extravasation across the BBB in vitro in this study. As culture conditions critically influence endothelial cell differentiation and endothelial barrier formation and thus can likewise influence T cell extravasation, we carefully applied optimal growth and differentiation protocols for each type of brain endothelial cells (Coisne et al, 2005; Rohnelt et al, 1997). However, to achieve most similar growth conditions, we used identical media components where applicable. All experiments were carried out under identical conditions.

To analyze the junctional architecture of bEnd5 and pMBMECs, we first verified the junctional localization of the adherens junction protein VE-cadherin in both BBB models. Although VE-cadherin was properly localized to the cell junctions, we cannot exclude a differential phosphorylation pattern of the cytoplasmic domain of VE-cadherin; this would influence tight junction protein expression, endothelial cell monolayer permeability, and lymphocyte extravasation (Dejana et al, 2009; Taddei et al, 2008; Turowski et al, 2008; Vestweber, 2008). Therefore, we investigated the localization of the tight junction proteins, claudin-5 and occludin. In both types of brain endothelial cells, claudin-5 was exclusively localized to the junctions. However, occludin failed to properly localize in bEnd5. Combined with our previous observation that bEnd5 have strongly decreased occludin mRNA levels compared with pMBMECs (Lyck et al, 2009), it can therefore be assumed that the tight junction complexity is severely altered in bEnd5. Consistently, an impaired junctional localization of occludin has already been described for human umbilical vein endothelial cells (Man et al, 2008) and for the immortalized rat brain endothelial cell line GPNT (Turowski et al, 2008). The presence of actin stress fibers in bEnd5 cells, as compared with cortical actin rings in pMBMECs, prompted us to compare paracellular permeabilities of both types of endothelial cells. Our results clearly showed that for 3 kDa dextran, the paracellular diffusion rate across bEnd5 was significantly higher than across pMBMECs. Although occludin-deficient cells develop complex tight junctions, an influence of occludin for the proper functionality of tight junctions has been described earlier (Schulzke et al, 2005). Thus, the appearance of actin stress fibers and the failure to properly localize junctional proteins (such as occludin) into appropriate junctional complexes at cell borders coincided with impaired barrier formation in bEnd5.

The cellular and molecular differences observed between bEnd5 and pMBMECs drove us to provide a detailed comparison of the individual steps of T cell extravasation between both types of brain endothelial cells. To this end, we performed classic T cell adhesion and diapedesis experiments under static conditions (Rohnelt et al, 1997), as well as modern live-cell imaging to follow dynamic T cell behavior under physiologic flow conditions (Stein et al, 2003). For all experiments, we used a previously well-characterized antigen-specific CD4⁺ effector/memory T_H1 cell line. These T cells have a high predisposition for interaction with the BBB, because these T cells migrate across the BBB in vivo and induce experimental autoimmune encephalomyelitis when injected into susceptible mouse strains (Engelhardt et al, 1998; Laschinger and Engelhardt, 2000; Vajkoczy et al, 2001). We found that similar numbers of T cells adhered to pMBMECs and bEnd5 under static conditions. Similarly, equal numbers of T cells arrested on pMBMECs and bEnd5 under shear conditions. These results indicate a comparable capacity of pMBMECs and bEnd5 to promote shear-resistant T cell arrest and firm adhesion. Our results stand in apparent contrast to another study that described a severe impairment of SV40-immortalized human endothelial cell lines, but not of primary human endothelial cells to support rolling and adhesion of peripheral blood mononuclear cells (Oostingh et al, 2007). However, in that study, failure to support shear-resistant initial contact formation of peripheral blood mononuclear cells could be assigned to a defect of endothelial cell lines in the upregulation of important adhesion molecules, which is not the case for bEnd5. The fact that both types of brain endothelial cells tested in our assay inducibly express ICAM-1 and VCAM-1 might explain the equal capacities of both in vitro BBB models to support T cell adhesion (Coisne et al, 2006; Reiss et al, 1998).

On both pMBMECs and bEnd5, arrested T cells polarized immediately and started to crawl on the endothelial surface until they underwent diapedesis. In our study, the velocity of T cell crawling on the endothelial surfaces was between 2.9 and 4.4 μm/min. A recent two-photon imaging study visualized the intraluminal crawling of encephalitogenic T cells on meningeal microvessels at a higher velocity of 12.5 μm/min in vivo (Bartholomaus et al, 2009). However, in vitro studies of T cell crawling on immobilized ICAM-1 under static conditions (Jacobelli et al, 2009; Smith et al, 2005) documented highly variable T cell crawling velocities that ranged from 4 to 9 μm/min. In particular, the density of ICAM-1 on the crawling substrate (Smith et al, 2005) or the interplay between T cell adhesion and its actomyosin contractility during crawling (Jacobelli et al, 2009) were definitely shown to influence the speed of movement. Thus, different velocities of T cell crawling on brain endothelial cells could be caused by variations in the spatial organization of endothelial cell adhesion molecules. Consequently, the reduced crawling velocity on bEnd5 compared with pMBMECs could be caused by an increased avidity of cellular interaction between the T cell and the bEnd5 surface.

On the surfaces of both of our in vitro BBB models, T cell crawling was observed preferentially against the direction of shear flow. This observation is in line with the study of Bartholomaus et al (2009) showing that T cells preferentially crawl against the direction of blood flow within meningeal microvessels during the onset of experimental autoimmune encephalomyelitis in vivo. Therefore, the directionality of T cell crawling with regard to shear forces can be modeled by bEnd5 and pMBMECs in vitro. The question of whether the orientation of T cells against the direction of flow stands in direct correlation with the promoting effects of flow to T cell extravasation as observed in vitro remains to be elucidated (Cinamon et al, 2001).

In this study, T cell diapedesis rates were found to be strikingly lower across pMBMECs than across bEnd5 under static conditions. Live-cell imaging under shear flow showed that T cell crawling distances were elongated on pMBMECs, and that therefore, a higher percentage of T cells remained continuously crawling on the pMBMEC surface. Although T cell crawling velocity was slightly increased on pMBMECs, this did not compensate for the elongated crawling paths before diapedesis. In conclusion, the reduced diapedesis rate across pMBMECs can be attributed to a lower frequency of sites permissive for diapedesis on pMBMECs than on bEnd5. The number of permissive sites could be determined by the concentration of chemokines presented on the endothelial surface. An elegant in vitro study using human umbilical vein endothelial cells showed that the availability of basal CCL5 and apical CXCL12 significantly influences the dynamic interaction of T cells during extravasation (Schreiber et al, 2007). Therefore, the longer crawling distance and low diapedesis rate observed on pMBMECs compared with bEnd5 could be caused by a differential availability of chemokines on the respective endothelial surfaces, which would be the subject of further investigations. Importantly, for direct comparison differential external influences of in vitro cultured endothelial cells such as astrocytic-conditioned medium must always be considered. For standardization, we made every attempt to keep culture conditions of both types of brain endothelial cells most similar by using identical media components and ommitting conditioned medium for both cell types. However, a differential influence on gene expression exerted by the differences in fetal calf serum and basic fibroblast growth factor media content cannot be excluded.

One alternative explanation for the reduced capability of T cells to diapedese across pMBMECs could be a different cellular pathway of diapedesis supported by bEnd5 and pMBMECs. Both the paracellular and the transcellular pathways of T cell diapedesis have been shown to exist, and to depend on the origin of the endothelium (Carman, 2009). In particular, it has been shown that across TNF-α-activated macrovascular endothelium (human umbilical vein endothelial cells), 90% of T cells migrated through the paracellular pathway and only 10% of T cells took the transcellular route for diapedesis (Carman et al, 2007; Millan et al, 2006). By contrast, the percentage of transcellular T cell diapedesis across microvascular endothelial cells from the human skin and lung has been observed to be significantly increased to 30% (Carman et al, 2007; Millan et al, 2006). Although in our study, the high paracellular permeability of bEnd5 and the increased density of cell–cell junctions present on bEnd5 monolayers suggest that endothelial junctions are readily accessible for T cell diapedesis through the paracellular pathway, pMBMECs might restrict diapedesis to the transcellular pathway because of their complex tight junctions and low paracellular permeability (Coisne et al, 2005). To account for the obvious differences in cell shape and in consequence the different frequencies of cellular junctions between both types of the brain endothelium, we evaluated the relative densities of cell–cell junctions. However, as T cell diapedesis across TNF-α-stimulated bEnd5 was 4.6-fold above T cell diapedesis across TNF-α-stimulated pMBMECs in the static experimental setup and 2.4-fold above T cell diapedesis across TNF-α-stimulated pMBMECs in the flow chamber experimental setup, the 2.0-fold increased junctional density of monolayers formed by bEnd5 compared with monolayers formed by pMBMECs can—even if T cell diapedesis would solely occur through the paracellular pathway—only partially account for the enhanced T cell diapedesis observed. Further studies will therefore be required to analyze the precise passageway of T cells across bEnd5 and pMBMECs.

Taken together, our study shows that both in vitro BBB models examined offer valuable model systems for the study of the molecular mechanism of shear-resistant adhesion of T cells to the BBB. However, bEnd5 and pMBMECs seem to harbor different molecular cues for T cell crawling and diapedesis. This in turn leads to lower diapedesis rates of T cells across pMBMECs. As pMBMECs better reflect endothelial BBB specialization in vivo, we propose that reliable information about the cellular and molecular mechanisms of T cell diapedesis across the BBB can only be obtained by using pMBMECs.

Footnotes

Acknowledgements

We thank Dr Urban Deutsch for providing mice as a source of pMBMECs. We are grateful for critical manuscript reading by Dr Charaf Benarafa, and we thank Mark Liebi for his excellent technical assistance.

The authors declare no conflict of interest.

Contribution: OS and RL performed experiments, made the figures, and wrote the paper. CC added essential knowledge for BBB endothelial cell isolation and culturing. RL and BE designed and supervised the research and wrote the paper.

Notes

Supplementary Information accompanies the paper on the Journal of Cerebral Blood Flow & Metabolism website ()

References

Abbott

Patabendige

Dolman

Yusof

Begley

(2010) Structure and function of the blood-brain barrier. Neurobiol Dis 37:13–25

Afonso

Ozden

Prevost

Schmitt

Seilhean

Weksler

Couraud

Gessain

Romero

Ceccaldi

(2007) Human blood-brain barrier disruption by retroviral-infected lymphocytes: role of myosin light chain kinase in endothelial tight-junction disorganization. J Immunol 179:2576–83

Bahbouhi

Berthelot

Pettre

Michel

Wiertlewski

Weksler

Romero

Miller

Couraud

Brouard

Laplaud

Soulillou

(2009) Peripheral blood CD4+ T lymphocytes from multiple sclerosis patients are characterized by higher PSGL-1 expression and transmigration capacity across a human blood-brain barrier-derived endothelial cell line. J Leukoc Biol 86:1049–63

Bartholomaus

Kawakami

Odoardi

Schlager

Miljkovic

Ellwart

Klinkert

Flugel-Koch

Issekutz

Wekerle

Flugel

(2009) Effector T cell interactions with meningeal vascular structures in nascent autoimmune CNS lesions. Nature 462:94–8

Bogatcheva

Verin

(2008) The role of cytoskeleton in the regulation of vascular endothelial barrier function. Microvasc Res 76:202–7

Carman

(2009) Mechanisms for transcellular diapedesis: probing and pathfinding by ‘invadosome-like protrusions’. J Cell Sci 122:3025–35

Carman

Sage

Sciuto

de la Fuente

Geha

Ochs

Dvorak

Springer

(2007) Transcellular diapedesis is initiated by invasive podosomes. Immunity 26:784–97

Cecchelli

Dehouck

Descamps

Fenart

Buee-Scherrer

Duhem

Lundquist

Rentfel

Torpier

Dehouck

(1999) In vitro model for evaluating drug transport across the blood-brain barrier. Adv Drug Deliv Rev 36:165–78

Cinamon

Shinder

Alon

(2001) Shear forces promote lymphocyte migration across vascular endothelium bearing apical chemokines. Nat Immunol 2:515–22

10.

Coisne

Dehouck

Faveeuw

Delplace

Miller

Landry

Morissette

Fenart

Cecchelli

Tremblay

Dehouck

(2005) Mouse syngenic in vitro blood-brain barrier model: a new tool to examine inflammatory events in cerebral endothelium. Lab Invest 85:734–46

11.

Coisne

Faveeuw

Delplace

Dehouck

Miller

Cecchelli

Dehouck

(2006) Differential expression of selectins by mouse brain capillary endothelial cells in vitro in response to distinct inflammatory stimuli. Neurosci Lett 392:216–20

12.

Dejana

Orsenigo

Molendini

Baluk

McDonald

(2009) Organization and signaling of endothelial cell-to-cell junctions in various regions of the blood and lymphatic vascular trees. Cell Tissue Res 335:17–25

13.

Deli

Abraham

Kataoka

Niwa

(2005) Permeability studies on in vitro blood-brain barrier models: physiology, pathology, and pharmacology. Cell Mol Neurobiol 25:59–127

14.

Engelhardt

(2008) Immune cell entry into the central nervous system: involvement of adhesion molecules and chemokines. J Neurol Sci 274:23–6

15.

Engelhardt

Laschinger

Schulz

Samulowitz

Vestweber

Hoch

(1998) The development of experimental autoimmune encephalomyelitis in the mouse requires alpha4-integrin but not alpha4beta7-integrin. J Clin Invest 102:2096–105

16.

Engelhardt

Wolburg

(2004) Mini-review: transendothelial migration of leukocytes: through the front door or around the side of the house? Eur J Immunol 34:2955–63

17.

Jacobelli

Bennett

Pandurangi

Tooley

Krummel

(2009) Myosin-IIA and ICAM-1 regulate the interchange between two distinct modes of T cell migration. J Immunol 182:2041–50

18.

Laschinger

Engelhardt

(2000) Interaction of alpha4-integrin with VCAM-1 is involved in adhesion of encephalitogenic T cell blasts to brain endothelium but not in their transendothelial migration in vitro. J Neuroimmunol 102:32–43

19.

Lawrence

Smith

Eskin

McIntire

(1990) Effect of venous shear stress on CD18-mediated neutrophil adhesion to cultured endothelium. Blood 75:227–37

20.

Ley

Laudanna

Cybulsky

Nourshargh

(2007) Getting to the site of inflammation: the leukocyte adhesion cascade updated. Nat Rev Immunol 7:678–89

21.

Lyck

Reiss

Gerwin

Greenwood

Adamson

Engelhardt

(2003) T-cell interaction with ICAM-1/ICAM-2 double-deficient brain endothelium in vitro: the cytoplasmic tail of endothelial ICAM-1 is necessary for transendothelial migration of T cells. Blood 102:3675–83

22.

Lyck

Ruderisch

Moll

Steiner

Cohen

Engelhardt

Makrides

Verrey

(2009) Culture-induced changes in blood-brain barrier transcriptome: implications for amino-acid transporters in vivo. J Cereb Blood Flow Metab 29:1491–502

23.

Man

Ubogu

Williams

Tucky

Callahan

Ransohoff

(2008) Human brain microvascular endothelial cells and umbilical vein endothelial cells differentially facilitate leukocyte recruitment and utilize chemokines for T cell migration. Clin Develop Immunol 2008:384982

24.

Martinelli

Gegg

Longbottom

Adamson

Turowski

Greenwood

(2009) ICAM-1-mediated endothelial nitric oxide synthase activation via calcium and AMP-activated protein kinase is required for transendothelial lymphocyte migration. Mol Biol Cell 20:995–1005

25.

Millan

Hewlett

Glyn

Toomre

Clark

Ridley

(2006) Lymphocyte transcellular migration occurs through recruitment of endothelial ICAM-1 to caveola- and F-actin-rich domains. Nat Cell Biol 8:113–23

26.

Oostingh

Schlickum

Friedl

Schon

(2007) Impaired induction of adhesion molecule expression in immortalized endothelial cells leads to functional defects in dynamic interactions with lymphocytes. J Invest Dermatol 127:2253–8

27.

Reiss

Engelhardt

(1999) T cell interaction with ICAM-1-deficient endothelium in vitro: transendothelial migration of different T cell populations is mediated by endothelial ICAM-1 and ICAM-2. Int Immunol 11:1527–39

28.

Reiss

Hoch

Deutsch

Engelhardt

(1998) T cell interaction with ICAM-1-deficient endothelium in vitro: essential role for ICAM-1 and ICAM-2 in transendothelial migration of T cells. Eur J Immunol 28:3086–99

29.

Rohnelt

Hoch

Reiss

Engelhardt

(1997) Immunosurveillance modelled in vitro: naive and memory T cells spontaneously migrate across unstimulated microvascular endothelium. Int Immunol 9:435–50

30.

Schreiber

Shinder

Cain

Alon

Sackstein

(2007) Shear flow-dependent integration of apical and subendothelial chemokines in T-cell transmigration: implications for locomotion and the multistep paradigm. Blood 109:1381–6

31.

Schulzke

Gitter

Mankertz

Spiegel

Seidler

Amasheh

Saitou

Tsukita

Fromm

(2005) Epithelial transport and barrier function in occludin-deficient mice. Biochim Biophys Acta 1669:34–42

32.

Shulman

Shinder

Klein

Grabovsky

Yeger

Geron

Montresor

Bolomini-Vittori

Feigelson

Kirchhausen

Laudanna

Shakhar

Alon

(2009) Lymphocyte crawling and transendothelial migration require chemokine triggering of high-affinity LFA-1 integrin. Immunity 30:384–96

33.

Smith

Carrasco

Stanley

Kieffer

Batista

Hogg

(2005) A talin-dependent LFA-1 focal zone is formed by rapidly migrating T lymphocytes. J Cell Biol 170:141–51

34.

Stanley

Smith

McDowall

Nicol

Zicha

Hogg

(2008) Intermediate-affinity LFA-1 binds alpha-actinin-1 to control migration at the leading edge of the T cell. EMBO J 27:62–75

35.

Stein

Soriano

M'Rini

Nombela-Arrieta

de Buitrago

Rodriguez-Frade

Mellado

Girard

Martinez

(2003) CCR7-mediated physiological lymphocyte homing involves activation of a tyrosine kinase pathway. Blood 101:38–44

36.

Taddei

Giampietro

Conti

Orsenigo

Breviario

Pirazzoli

Potente

Daly

Dimmeler

Dejana

(2008) Endothelial adherens junctions control tight junctions by VE-cadherin-mediated upregulation of claudin-5. Nat Cell Biol 10:923–34

37.

Turowski

Adamson

Greenwood

(2005) Pharmacological targeting of ICAM-1 signaling in brain endothelial cells: potential for treating neuroinflammation. Cell Mol Neurobiol 25:153–70

38.

Turowski

Martinelli

Crawford

Wateridge

Papageorgiou

Lampugnani

Gamp

Vestweber

Adamson

Dejana

Greenwood

(2008) Phosphorylation of vascular endothelial cadherin controls lymphocyte emigration. J Cell Sci 121:29–37

39.

Vajkoczy

Laschinger

Engelhardt

(2001) Alpha4-integrin-VCAM-1 binding mediates G protein-independent capture of encephalitogenic T cell blasts to CNS white matter microvessels. J Clin Invest 108:557–65

40.

Vandenbroucke

Mehta

Minshall

Malik

(2008) Regulation of endothelial junctional permeability. Ann NY Acad Sci 1123:134–45

41.

Vestweber

(2008) VE-cadherin: the major endothelial adhesion molecule controlling cellular junctions and blood vessel formation. Arterioscler Thromb Vasc Biol 28:223–32

42.

Yang

Roder

Abbruscato

(2007) Evaluation of bEnd5 cell line as an in vitro model for the blood-brain barrier under normal and hypoxic/aglycemic conditions. J Pharm Sci 96:3196–213

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.77 MB