The human brain endothelial cell line hCMEC/D3 is described in 23 . hCMEC/D3 cells were seeded on Transwell® filters (polycarbonate 12 well, pore size 3.0 μm, Corning, Lowell, MA) coated with type I collagen (BD Biosciences, Paris, France), at a density of 350,000 cells/cm² in commercially available complete medium EGM®-2 (Lonza, Walkersville, MD), supplemented with vascular endothelial growth factor, insulin-like growth factor 1, epidermal growth factor, basic fibroblast growth factor (FGF), hydrocortisone, ascorbate, penicillin-streptomycin, and 2.5% FCS, (all from Lonza) in an incubator at 37°C with 5% CO₂. For differentiation and expression of junction-related proteins, the hCMEC/D3 cells were grown at confluence in a growth-factor-depleted medium.

Primary HCMECs (Cell Systems, Kirkland, WA) were grown on 0.2% gelatin-coated (Fisher Scientific, New York, NY) tissue-culture plates in M199 medium supplemented with 20% fetal bovine serum (Gibco/BRL, Grand Island, NY), 5% heat-inactivated human serum (Invitrogen, Carlsbad, CA), 1% penicillin-streptomycin (Gibco/BRL), and 12 ng/ml endothelial cell growth factor (Sigma Aldrich, St. Louis, MO).

Human umbilical vein endothelial cells (HUVECs) and a human acute monocytic leukemia cell line (THP-1) were obtained from ATCC (Molsheim, France) and were cultivated, respectively, in EBM-2 medium supplemented with EBM-2 bullet kit (Lonza) and RPMI 1640 supplemented with 10% FCS and 1% penicillin-streptomycin (Invitrogen).

For stimulation assays, cells were incubated for 3 h, 12 h, or 24 h with recombinant human TWEAK (100 ng/ml, Peprotech, Neuilly-Sur-Seine, France), Fc-TWEAK (1 μg/ml, BiogenIdec, Cambridge, MA), its isotype control P1.17 (BiogenIdec), or recombinant human TNF (10 ng/ml, Peprotech). In some experiments, cells were incubated with recombinant human MMP-9 (rhMMP-9) from Calbiochem. All reagents were endotoxin-free.

After trypsination, differentiated unstimulated or TWEAK-stimulated hCMEC/D3 cells were pre-incubated on ice for 20 min with a solution containing PBS, 1% FCS, 0.02% sodium azide, and 25% purified human serum Immunoglobulin G(Sigma Aldrich) to inhibit binding to Fc receptors. After washes with a solution containing PBS, 1% FCS, and 0.02% sodium azide, cells were incubated on ice for 20 min with fluorescein-conjugated anti-human ICAM-1 antibody, fluorescein-conjugated anti-human E-selectin antibody (both from R&D Systems, Minneapolis, USA), or anti-human Fn14 phycoerythrin-conjugated antibody (eBioscience, Paris, France). After three more washes, cells were centrifuged and resuspended in PBS with 2% paraformaldehyde. Fluorescence-activated cell sorting (FACS) analysis was performed on a FACSCanto II (Becton Dickinson, Le Pont-De-Claix, France) using BD FACSDiva software.

Total RNA was prepared from cultures of hCMEC/D3, HUVECs, and THP-1 using the RNeasy Lipid Tissue Mini kit (Qiagen, Courtaboeuf, France). Single-strand cDNA was synthesized from 1 μg of total RNA using oligo(dT)_12–18 primers (Invitrogen) and Moloney murine leukemia virus reverse transcriptase (Invitrogen) under the conditions indicated by the manufacturer. The sequences of the specific forward (F) and reverse (R) primers were as follows: TWEAK-F (5′-ATATATAGATCTATGGCCGCCCGTCGGAGC-3′), TWEAK-R (5′-AGCCTTCCCCTCATCAAAGT-3′), Fn14-F (5′-CCAAGCTCCTCCAACCACAA-3′), Fn14-R (5′-TGGGGCCTAGTGTCAAGTCT-3′), GAPDH-F (5′-GTCAGTGGTGGACCTGACCT-3′), and GAPDH-R (5′-TGCTGTAGCCAAATTCGTT-3′). Each cDNA was amplified by Taq recombinant DNA polymerase (Invitrogen). The cDNAs were first denatured 3 min at 94°C, then 40 PCR cycles were carried out with the following profile: 45 s denaturation at 94°C, 30 s annealing at 59°C for Fn14 or 52°C for GAPDH, and 1 min elongation at 72°C. Cycles were followed by a 10 min final elongation at 72°C (Verity 96-well thermal cycler, Applied Biosystems Foster City, CA). Controls were performed with template-free PCR reactions. PCR products were analyzed by electrophoresis on a 2% agarose gel containing ethidium bromide. The expected sizes of the TWEAK, Fn14, and GAPDH PCR products were 522 base pairs (bp), 242 bp, and 226 bp, respectively.

Real-time PCR (qPCR) experiments were carried out with the 7500 Fast Real-Time PCR System (Applied Biosystems). All reactions were performed using TaqMan Fast Universal PCR Master Mix and two probes from the TaqMan® Gene Expression Assays (MMP-9, Hs00234579_m1 and Abelson (ABL): ABL-F (5′-TGGAGATAACACTCTAAGCATAACTAAAGGT-3′), ABL- TaqMan reverse probe (5′-Fam6CCATTTTTGGTTTGGGCTTCACACCATT-Tamra-3′), ABL-R (5′-GATGTAGTTGCTTGGGACCCA-3′, used as reference) according to the manufacturer’s instructions (Applied Biosystems). Each experiment used 25 ng of previously prepared hCMEC/D3 cDNA. Samples were run in duplicates on the same 96-well plates and analyzed with the 7500 Software v2.0 (Applied Biosystems). The thermal cycling conditions started with initial denaturation at 95°C for 20 s, followed by 40 cycles of denaturation at 95°C for 3 s and annealing and extension at 60°C for 30 s. Relative expression levels are determined according to the ΔΔCt method where the expression level of the mRNA of interest is given by 2^-ΔΔCT where ΔΔCT = ΔCT target mRNA − ΔCT reference mRNA (Abelson) in the same sample.

hCMEC/D3 cell proliferation was determined by measurement of bromodeoxyuridine (BrdU) incorporation during DNA synthesis by chemiluminescence detection using the Cell Proliferation ELISA BrdU kit (Roche Applied Science, Meylan, France) according to the manufacturer’s instructions.

hCMEC/D3 differentiated cells were stimulated with TWEAK or TNF for 24 h. Supernatants were collected, centrifuged, and stored at −80°C until analysis. CCL-2, IL-8, Il-6, and IL-10 levels were evaluated using commercially available ELISA kits (Peprotech, and BD Biosciences, San Jose, CA) according to the manufacturers’ instructions. All samples were analyzed in triplicate. The detection threshold was 16 pg/ml of cytokine.

For transport experiments, we tested the passage of two distinct molecules, Lucifer yellow (LY) (Sigma) and BSA-FITC (Accurate Chemical and Scientific, Westbury, NY). hCMEC/D3 cells or HCMECs were seeded and differentiated on coated Transwell® filters. Both the upper and lower chambers were washed with pre-warmed Ringer-HEPES (RH) at 37°C. At time t = 0, LY or BSA-FITC was applied in the apical compartment. After 60 min, detection of the fluorescent molecules was carried out with a Beckman DTX800 luminometer with excitation at 430/485 nm, and emission at 535 nm. Permeability coefficients (Pe) take into account the relation between the permeability of the monolayer and the permeability of empty filters (pre-coated, without cells). Each condition was tested in triplicate in each experiment.

The assay setup of HCMECs for transepithelial electric resistance (TEER) was the same as for the BBB permeability assay. The HCMEC TEER was measured using the Electrical Resistance System (Millicell-ERS-2; Millipore, Bedford, MA), following the manufacturer’s instructions. In brief, coated inserts without cells were used as a blank (minimum resistance). The electrical resistance of each insert following treatment with TWEAK, P1.17, or TNF was calculated by subtracting the blank from each reading. Each condition was run in duplicate, and the resistance measured twice for each well.

The following antibodies were used: goat anti-TWEAK (R&D system, Abingdon, UK), rabbit anti-ZO-1 (1/200, Invitrogen), and mouse anti-β-actin (1/3000, Sigma Aldrich). Protein concentrations were determined using the Lowry method (Bio-Rad, Hercules, CA). After boiling, aliquots containing equal amounts of protein were loaded in Laemmli buffer and separated by 8.5% sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS PAGE, Bio-Rad) using a MiniBlot system (Bio-Rad). Proteins were transferred onto nitrocellulose membranes (Amersham Biosciences, Buckinghamshire, UK) in transfer buffer (25 mM Tris, 192 mM glycine, and 20% ethanol). Membranes were incubated overnight in blocking buffer at 4°C and then probed with the primary antibody against ZO-1, β-actin, or TWEAK diluted in blocking buffer (Roche Diagnostics, Mannheim, Germany). After washing, membranes were incubated with a peroxidase-conjugated secondary antibody (Jackson Immunoresearch, West Grove, PA). Finally, proteins were detected using a chemiluminescence kit (Roche Diagnostics). Films were digitized using GeneTools software (Syngen, San Carlos, CA), and optical densities of the bands were assessed using Scion Image software (Scion Corporation, MA).

To assess gelatinolytic activity of MMPs, hCMEC/D3 cells were grown on glass coverslips and the medium was supplemented to a final concentration of 5 mM CaCl₂ and 10 μg/ml of intramolecularly quenched FITC-labeled DQTM-gelatin (EnzCheck Collagenase kit from Molecular Probes), as previously described 24 25 . After 2 h at 37°C in a humidified atmosphere containing 5% CO₂, cells were rinsed in PBS, fixed with 4% paraformaldehyde (PFA) for 5 min, and incubated for 5 min with DNA intercalant Hoechst #33258 (Molecular Probes, Eugene, OR). Cells were observed with a Nikon E800 upright epifluorescence microscope and digital images were acquired at 1,024 × 1,024 pixels and saved in TIFF format. Fluorescence levels were measured at the level of individual cells using image J software; image editing was performed using Adobe Photoshop (Adobe Systems, Paris, France).

Standard methodology for gelatin zymography was used to detect MMP-2 and MMP-9 expression levels via their activity in cell supernatant or cell lysate samples, as described previously 25 . Serum-free culture supernatants and lysates were collected and protein concentration was normalized as mentioned above. Equal amounts of protein were subjected to 8% SDS PAGE containing 1 mg/ml gelatin (Sigma Aldrich) in nondenaturing, nonreducing conditions. After electrophoresis, gels were washed twice for 30 min in 2.5% Triton X-100 to remove SDS and incubated for 48h in MMP-activating buffer, 50 mM Tris–HCl, pH 7.5, with 10 mM CaCl₂ at 37°C. Gels were then stained with 0.1% Coomassie Brilliant Blue R-250 (Bio-Rad) for 3h and destained with a solution containing 5% acetic acid until clear bands of gelatinolysis appeared on a dark background. Gels were digitized using GeneTools software.

hCMEC/D3 cells grown on glass coverslips were fixed in 4% PFA and were incubated for 1 h at room temperature with rabbit anti-ZO-1 (Invitrogen) or goat anti-MMP-9 (Abcam, Cambridge Science Park, UK) primary polyclonal antibodies. Subsequently, cells were incubated with Alexa Fluor 488 or 594 anti-mouse or anti-goat secondary antibodies (Invitrogen) followed by Hoechst and mounted in fluorescent mounting medium (DAKO, Glostrup, Denmark). The mounted slides were observed with a Leica TCS SP2 confocal microscope (Leica Microsystems, Heidelberg, Germany). High magnification images were acquired using a 633 HCX PL APO oil immersion objective by sequential scanning to minimize the crosstalk of fluorophores. For each channel, photomultiplier gains and offsets were adjusted to use full image dynamic range. Images acquired at 1,024 × 1,024 pixels and saved in TIFF format were processed for colocalization analysis using ImageJ plug-in processing software. Image editing was performed using Adobe Photoshop.

To investigate the involvement of mitogen-activated protein kinases (MAPKs) in MMP-9 expression following TWEAK stimulation, we blocked c-RAF1 and MEK signaling pathways using ERK2 and MEK1/2 specific inhibitors (Sigma Aldrich) at working concentrations of 5 μM GW5074 and 0.5 μM U0126v. hCMEC/D3 cells were cultured for 24h in the presence or absence of TWEAK (100 ng/ml) and in the presence or absence of the MAPK inhibitors; cell lysates and supernatants were then collected and tested by gel zymography.

In a first step, we used RT-PCR to study, in the hCMEC/D3 cells, the expression of the mRNAs encoding TWEAK and its receptor Fn14. We found that hCMEC/D3 cells express both TWEAK and Fn14 mRNAs (Figure 1A). We next used flow cytometry to assess TWEAK and Fn14 expression at the membrane in the same cells. We show that hCMEC/D3 cells do not constitutively express membrane TWEAK but constitutively express Fn14 on their surface (Figure 1C). TWEAK exposure did not up-regulate TWEAK or Fn14 expression at the cell surface (data not shown). Similarly, using ELISA, we were not able to detect soluble TWEAK in the culture supernatants of hCMEC/D3 (data not shown). In agreement with data issued from Western blot analysis of TWEAK expression by hCMEC/D3 cells (Figure 1B), these results lead us to conclude that this cell line expresses neither membrane nor soluble TWEAK.

Next, we studied the effects of TWEAK on hCMEC/D3 proliferation using a BrdU incorporation test. As indicated in Figure 2, a 24h TWEAK exposure of the cells induced proliferation. It is worth noting that the proliferative effects of soluble TWEAK are significantly higher than those of TNF.

To assess the potential inflammatory effects of TWEAK on hCMEC/D3 cells and BBB inflammation, we used ELISA to measure cytokine secretion in the culture supernatants after TWEAK or TNF exposure during 24h, compared with nonstimulated cells (Figure 3). We found that TWEAK significantly up-regulated proinflammatory cytokine (CCL-2, Il-8, and Il-6) production in the hCMEC/D3 cells (Figure 3A). We obtained the same results with primary HCMECs (Figure 3B). We also showed that nonstimulated cells can produce low levels of the anti-inflammatory cytokine Il-10 and that this secretion is not significantly modulated by TWEAK or TNF.

We also studied the effects of TWEAK stimulation on the expression of hCMEC/D3 adhesion molecules. We chose to explore E-selectin, which is involved in leukocyte rolling, and ICAM-1, which is also up-regulated during inflammation, especially during firm leukocyte adhesion to the endothelium. Using flow cytometry, we showed in hCMEC/D3 that E-selectin membrane expression is not significantly up-regulated by TWEAK, while ICAM-1 membrane levels are clearly increased following a 24h TWEAK exposure (Figure 3C).

Lucifer yellow is a small hydrophobic molecule that presents low cerebral penetration and that is classically used as a molecular marker of paracellular passage. To determine the effects of TWEAK on the permeability of a monolayer of brain endothelial cells, hCMEC/D3 cells were grown on Transwell filters and were exposed for 24h to TWEAK or TNF. We observed that TWEAK induced a significant increase in Pe compared with the nonstimulated condition Figure 4A. Similar results were obtained with human primary HCMEC (Figure 4B). In addition, we evaluated the TEER of the in vitro BBB model upon TWEAK or TNF stimulation; we observed decreased TEER upon TWEAK stimulation (Figure 4C). These data, associated with results from transport experiments, lead us to conclude that soluble TWEAK increases permeability of the human monolayers of brain endothelial cells in the in vitro BBB model.

Published reports indicate that TWEAK induces proteolytic activity, notably metalloproteinase activity, in different cell types 18 26 . We have also observed, in a transcriptomic analysis of hMEC/D3 cells treated for 24 h by TWEAK, that the mRNAs encoding several MMPs (MMP-12, MMP-17, MMP-28) were up-regulated, including MMP-9 (data not shown). We thus evaluated the effects of TWEAK on the gelatinolytic activity of hCMEC/D3 cells. We used in situ zymography following exposure to TWEAK (100 ng/ml) for 24 h. We found significantly increased gelatinolytic activity in TWEAK- or TNF-treated hMEC/D3 cells, compared with nontreated cells. In some cells, gelatinolytic activity appeared to delineate cells, suggesting localization at the plasma membrane (Figure 5A). Among proteinases that have gelatinolytic activity are the gelatinases MMP-2 and MMP-9. hCMEC/D3 cells were exposed to TWEAK and TNF for 24 h and gelatin-based gel zymography was performed on cell lysates and serum-free media conditioned by the treated and nontreated cells (Figure 5B). Nontreated cells showed constitutive expression of a 68 kDa gelatinase corresponding to the pro-form of MMP-2 and weaker expression of the 105 kDa gelatinase corresponding to the molecular weight of pro-MMP-9. Densitometric scanning of the zymograms indicated that MMP-2 levels (cellular and secreted forms) remained unchanged following TWEAK and TNF treatment. Secreted levels of MMP-9 also remained identical to control. In contrast, cellular MMP-9 pro-form levels, as well as active MMP-9, increased significantly following TWEAK and TNF treatment. To assess whether increased MMP-9 levels correlated with increased steady-state MMP-9 mRNA levels as suggested by the transcriptomics analysis, we used qPCR; an increase in MMP-9 mRNA expression was indeed observed in hCMEC/D3 cells 24 hours after treatment with TWEAK and TNF (Figure 5C).

In fibroblast and skeletal muscle cells, TWEAK is reported to activate MAPKs and regulate the expression of MMP-9 27 28 . To assess whether this is also the case in brain endothelial cells, cultures were incubated for 1 h before adding TWEAK in the presence GW5074 and U0126, which are respectively inhibitors for ERK1/2 and MEK2 involved in the MAPK signaling pathways. Densitometric scanning of zymograms indicates that MAPK inhibitors efficiently suppressed the up-regulation of MMP-9, whose expression remained at basal levels (Figure 5D) and had no effect on the expression or activity levels of MMP-2 (data not shown). These results suggest that in human brain endothelial cells, TWEAK induces MMP-9 up-regulation via the MAPK signaling pathways.

We next evaluated the effects of exogenous recombinant human MMP-9 on the permeability of the hCMEC/D3 monolayer. The addition of 250 ng/ml rhMMP-9 for 24h enhanced the permeability of the endothelial cell monolayer by 20% (Figure 5E). Considering that TWEAK increases both Pe of the hCMEC/D3 monolayer and MMP-9 levels and that exogenously applied rhMMP-9 also increases Pe, we assessed whether TWEAK-increased Pe could be modulated by MMP inhibitors. TWEAK and the broadband MMP inhibitor RXPO3 were applied to the hCMEC/D3 monolayer for 24h and Pe was assessed. We find that inhibition of MMP activity neither prevents barrier impairment nor leads to a significant barrier recovery (Figure 6F).

To investigate the cellular distribution of MMP-9 in nontreated and TWEAK-treated endothelial cells, we used an antibody for MMP-9 whose specificity had been previously validated on neuroblastoma N2a cells transfected with MMP-9-GFP-constructs 25 . In nontreated and TWEAK-treated hCMEC/D3 cells, MMP-9 showed a punctuate, vesicular-like pattern distributed throughout the cytoplasm (Figure 6A). MMP-9 immunolabeling was clearly increased in the TWEAK- and TNF-treated cells. Detailed analysis of the labeled endothelial cells also indicates perinuclear accumulation of MMP-9, presumably in the Golgi and trans-Golgi network. We show that MMP-9 is localized in part at the membrane of brain endothelial cells. In agreement with our previous findings in other cell types of the CNS 25 29 , MMP-9 was also localized in the nucleus of brain endothelial cells (Figure 6B).

Because ZO-1 is exclusively located in tight junctions but also constitutes a substrate for MMP-9 30 , we assessed expression of this protein in the hCMEC/D3 cells under TWEAK exposure. Using double labeling experiments combining antibodies against MMP-9 and against ZO-1, we show that MMP-9 and ZO-1 can be colocalized at discrete areas of the plasma membrane (Figure 6A, 6C). Using Western blot analysis, we show that soluble TWEAK induced a down-regulation of the two isoforms, α+ and α−, of ZO-1 (Figure 6D).