Study subjects
Peripheral blood samples were collected from healthy controls in collaboration with the University Biobank Limburg (UBiLim, Hasselt, Belgium). In total, 5 donors were used (1 male and 4 female) with ages ranging between 23 and 44 years (mean ± SEM: 33.8 ± 4.3). This study was approved by the local ethical committee and informed consent was obtained from all donors.
Cell culture
Human cerebral microvascular endothelial cells (hCMEC/D3) were provided by Tebubio (le Perray-en-Yevelines, France) and were seeded at 1-1.2 ×106 cells/cm2 onto pre-coated plates or inserts with 75 µg/mL rat tail collagen type I solution (Merck) between passages 27 and 34. Cells were grown in EGM™-2MV Microvascular Endothelial Cell Growth Medium-2 BulletKit™ (CC-3202, Lonza) supplemented with 2.5% fetal bovine serum (FBS) (Gibco™, Thermo Fisher Scientific) up to 80–90% confluency to minimize apoptosis. Characteristics and stability of the used cell line is extensively described (29). For the production of EV, confluent cells were rinsed twice with PBS (Lonza) and a mixture of rhTNF-α and rhIFN-γ (Peprotech, Londen, UK) at a final concentration of 10 ng/mL were added in refreshed medium supplemented with 2.5% exosome-depleted fetal bovine serum (EXO-FBS-250A-1, System Bioscience) for 24h or left untreated in exosome-depleted medium. Exposure time and concentration of inflammatory cytokines were as previously described in Hermans et al (2022) (30). To prevent the induction of apoptosis in cells, and thus the presence of apoptotic bodies in EV preparation steps, cell culture supernatants were collected after 24h from approximately 6–8 × 106 cells/mL (from ~ 80–90% confluency) with 98% viability. For expression and functional studies, medium was changed to experimental medium: EBM™-2 Basal Medium (CC-3156, Lonza) supplemented with 5 ng/ml human fibroblast growth factor (hFGF), 1.4 µM hydrocortisone, 10 mg/ml gentamicin, 1 mg/ml amphotericin (A2942, all Merck) and 2.5% FCS after reaching 80% confluency at least 24 h before starting treatment. During treatment, cells were placed on serum-reduced (0.25% FCS) experimental medium (as described above) without hydrocortisone. All culturing took place in a humidified atmosphere condition of 37°C/5% CO2, on rat tail collagen type I coated plastics (75 µg/ml) and routine mycoplasma contamination tests were performed throughout the culturing.
Isolation of size-based EV populations
Isolation of sEV and lEV was performed on pooled cell culture supernatants of approximately 24 million cells using a differential (ultra)centrifugation method. Briefly, 40 ml of supernatant was pooled and centrifuged at 300 ×g for 10 min at 4°C to remove cell debris. To collect lEV (EV-2K and EV-10K), supernatants were first transferred to new tubes and centrifuged in a S-4-72 fixed angle rotor (Eppendorf- VWR) at 2000 ×g for 20 min at 4°C to obtain the EV-2K pellet. A second centrifugation step was done on 2K-free supernatant for 40 min at 10,000 ×g at 4°Cs to obtain the EV-10K pellet. To pellet the sEV (EV-100K), the 10K-free supernatant was ultra-centrifuged in a Ti-70 rotor (L-90 Beckman centrifuge, Fullerton, CA, USA) at 100,000 ×g for 3 h at 4°C. Depending on the downstream analysis, pellets were suspended in either 1 mL of exosome-depleted medium, Pierce RIPA (89900, Thermo Fisher Scientific, MA, USA) or extraction buffer (ab193970, Abcam Ltd., Cambridge, UK) and stored at − 20°C for short term storage or -80°C for long-term storage.
EV for functional assays were further purified using Sepharose CL-2B (17-0140-01, VWR) size exclusion chromatography (SEC) as described by (31). 1 mL pooled EV-enriched fractions (F4 and F5) were then upconcentrated using Amicon-Ultra 0.5 Centrifugal Filter Units with 10 kDa cutoff (UFC501096, Merck). In addition, to prove the role of EV in functional assays, EV were depleted from EV-2K, EV-10K and EV-100K fractions using Exosome Isolation Pan Kit: EV (130-110-912, Miltenyi Biotec) following the manufacturer's protocol.
Transmission electron microscopy (TEM)
The protocol for TEM-imaging of EV has been previously described (31). Briefly, droplets of the sample were placed on clean Parafilm and a Nickel TEM grid was placed on top of the droplets for 60 minutes. The grids with adherent EVs were fixed with 2% glutaraldehyde for 10 minutes, washed 5 times and transferred to 2% uranyl acetate for 15 minutes. The grids were then incubated in 0.13% methyl cellulose and 0.4% uranyl acetate for 10 minutes and dried at room temperature before examination with Tecnai G2 Spirit BioTWIN (FEI, Eindhoven, The Netherlands). Images were taken at 120kV.
Nanoparticle tracking analysis (NTA)
The mean size, concentration and size distribution of EV were quantified using the NanoSight NS300 system (Malvern Ltd, Sysmex Belgium N.V.) equipped with a 532 nm laser. EV suspensions were diluted with PBS over a range of concentrations to obtain between 10 and 100 particles per frame. The camera level was set to 14. Samples were injected into the sample chamber and measured five times for 60 s with a syringe pump speed of 75 u. Acquisitions were captured and analyzed using NTA software 3.2 (NanoSight. Malvern Ltd) and the threshold was set to 9. Five measurements were taken for each sample (at least nine biological replicates), mean and standard deviation were calculated and plotted using GraphPad Prism 9 software (GraphPad Software, San Diego, CA, USA). EV concentration was reported as particles/mL and EV size (nm) in mean values.
Fluorescence labeling of EV and their uptake
The membrane and RNAs loaded in EV were fluorescently labelled. Free dyes were removed from labeled EV using Sepharose CL-2B (17-0140-01, VWR) SEC as described above. Pooled EV fractions (F4 and F5) were concentrated using Amicon-Ultra 0.5 Centrifugal Filter Units with 10 kDa cutoff (UFC501096, Merck) at 4°C.
To compare the uptake of different sized based fractions of EV, hCMEC/D3 cells were grown in pre-coated eight-well culture plates, for 24 h before EV incubation. 1xE9 labeled EV were added to cells and incubated for 24 h. After incubation, cells were washed with PBS and fixed with PFA 2%. Nuclear staining was performed with 4’,6-diamidino-2-phenylindole (DAPI) at a final concentration of 10 µg/ml for 30 min at 37°C. Following the staining, images were captured using a Leica DM2000 LED (Leica Microsystems, Heidelberg, Germany) attached to a digital camera and Leica Application Suite X (LAS X) software (Leica Microsystems).
Protein extraction and quantification
Protein content of cells and EV lysates in either RIPA (89900, Thermo Fisher Scientific, MA, USA) supplemented with Protease Inhibitor Cocktail (P8340, Sigma-Aldrich, MO, USA) or extraction buffers (ab193970, Abcam) was determined using Pierce BCATM protein assay reagent Kit (Thermo Fisher Scientific, MA, USA) and micro Pierce BCATM protein assay reagent kit (Thermo Fisher Scientific, MA, USA), respectively. The procedure was done following the manufacturer’s specifications. Optical densities of standards and samples were read at OD595 nm using a Multiskan™ FC microplate absorbance reader (Thermo Fisher Scientific, MA, USA).
Western blotting
EV and cell pellets were lysed in RIPA buffer (89900, Thermo Fisher, MA, USA) containing a protease inhibitor cocktail (P8340, Sigma-Aldrich, MO, USA) and denatured at 95°C for five minutes. Five µg of protein was first separated by SDS-PAGE with either 12% or 4–7,5% polyacrylamide gels at 80 V for 30 minutes, followed by 140 V for 45–60 minutes. The proteins were then transferred onto a polyvinylidene fluoride membrane (Immobilon R, Merck Millipore Ltd) for minimum 1 hour at 350 mA. The membranes were blocked in 5% Powdered Milk/PBS or 5% Powdered Milk /TBS 0.1% Tween (TBS-T) for two hours and incubated overnight at 4°C with primary antibodies: mouse monoclonal anti-human ICAM-1 (1:1000; clone 15.2, sc-107, Santa Cruz Biotechnology), Annexin II (1:1000; clone C-10, sc-28385 Santa Cruz Biotechnology), CD63 (1:1000; clone Ts63, Thermo Fisher Scientific), CD9 (1:1000; clone Ts9, Life Technologies), and PECAM-1 (1:1000; sc-365804, Santa Cruz); rabbit anti-BAX antibody (1:500; clone E63, ab32503, Abcam), and Claudin-5 (1:1000; 34-1600, Life Technologies) in 5% Powdered Milk /PBS or 5% Powdered Milk /TBS-T. Next, the secondary antibodies (either rabbit anti-mouse HRP-conjugated or goat anti-rabbit HRP-conjugated; both from DAKO) at 1:1000 dilution were added to the membrane for 1 h at room temperature after three times washing with PBS-T. The blots were developed using Western Bright TM Sirius (K-12043-C20, Advansta, CA, USA) for the EV lysate or Western Bright TM Quantum (K-12042- C20, Advansta, CA, USA) for the cell lysate samples. The corresponding bands were detected by the ImagerQuant™TL (Amersham Imager 680, GE Healthcare) detection system. Band intensities were determined by quantifying the mean pixel gray values using ImageJ software.
RNA isolation, cDNA synthesis, and quantitative real-time PCR
For quantitative polymerase chain reaction (qPCR) of BBB-EC treated with different fractions of size-based EV, hCMEC/D3 were grown as described in pre-coated 6 well plates at 500.000 cells/2 ml. On the day of treatment, cells were washed twice with PBS (Lonza) and treated by adding an equal amount of EV (1x109 EV/mL) or 10 ng/ mL TNF-α/IFN-γ for 24 h in the refreshed serum-reduced experimental medium. Cells were collected from the plate by scraping in RLT buffer containing 1% β-mercapto-ethanol. Total RNA was isolated from cell pellets using the RNeasy Mini Kit from (Qiagen, Germany) according to the manufacturer's protocol. RNA concentrations were checked by Nanodrop™ 2000/2000c Spectrophotometer (Thermo Fisher Scientific). Afterwards, complementary DNA (cDNA) was generated from 1 µg of total RNA using qScript cDNA SuperMix (Quanta bioscience, CA, USA) following the manufacturer's protocol. PCR reactions contained 10 ng cDNA, Fast SYBR Green (Applied Biosystems, 4385612), and 10 µM forward and reverse primer mixture (Integrated DNA Technologies, Leuven, Belgium). Quantitative PCR was performed utilizing a StepOnePlus Real-Time PCR detection system (Life technologies) and universal cycle conditions (20s at 95°C, 40 cycles of 3 s at 95°C and 30 s at 60°C). Data as ΔΔCt was normalized to the two most stable housekeeping genes. Primers used in this study are listed in Table 1.
Table 1
Table 1 – Primer sequences Gene | Forward primer (5’ – 3’) | Reverse primer (5’ – 3’) | mTBP* | ATGGTGTGCACAGGAGCCAAG | TCATAGCTACTGAACTGCTG | mYWHAZ* | GCAACGATGTACTGTCTCTTTTGG | GTCCACAATTCCTTTCTTGTCATC | mZO-1 | GGCATTCCTGCTGGTTACA | AGGACACCAAAGCATGTGAS | mClaudin-5 | GCCCTGCTCAGAACAGACTA | GGCAGTTTGGTGCCTACTTC | mIL-1β | GCTGAAAGCTCTCCACCTCA | AGGCCACAGGTATTTTGTCG | mIL-10 | CTTTAAGGGTTACCTGGGTTGC | ATTAAAGGCATTCTTCACCTGC | mIL-17 | CAGCGATCATCCCTCAAAGC | GCGCCAAGGGAGTTAAAGAC | mCCL2 | CAGCAGGTGTCCCAAAGAAG | CATTTGGTTCCGATCCAGGTT | mCXCL10 | CTGCCCACGTGTTGAGATCA | TGGTCTTAGATTCCGGATTCAGA | mCD3 | AACACTTTCTGGGGCATCCT | CTGTCTAGAGGGCACGTCAA | mCD4 | GTTCAGGACAGCGACTTCTG | GAAGGAGAACTCCGCTGACT | mCD8 | AGTGAAGGGGACCGGATTG | GGACATTTGCAAACACGCTT | mFoxp3 | CAGAGAGGTATTGAGGGTGGG | GCAGAGTCAGGAGAAGTTGC | hTBP* | TATAATCCCAAGCGGTTTGC | GCTGGAAAACCCAACTTCTG | hCYCA* | AGACTGAGTGGTTGGATGGC | TCGAGTTGTCCACAGTCAGC | hIL-1β | GATGAAGTGCTCCTTCCAGG | GCATCTTCCTCAGCTTGTCC | hIL-18 | GACTCTTGCGTCAACTTCAAGG | CAGGCTGTCTTTTGTCAACGA | hIL-6 | GAGGAGACTTGCCTGGTGAA | GCTCTGGCTTGTTCCTCACT | hICAM-1 | AGCTTCGTGTCCTGTATGGC | ACAGTCACTGATTCCCCGAT | hVCAM-1 | AATGTTGCCCCCAGAGATACAACCG | GAGCTGCCTGCTCCACAGGA | hE-selectin | GCACTGTGTGCAAGTTCGC | GGCTTTTGGTAGCTTCCGTC | hClaudin-1 | TTGACTCCTTGCTGAATCTGAG | TTCTGCACCTCATCGTCTTC | hClaudin-3 | CATCACGTCGCAGAACATCT | GAGTCGTACACCTTGCACTG | hClaudin-5 | ACATTGTCGTCCGCGAGTTT | ACTTCTGCGACACGGGCA | hVE-Cadherin | AAACACCTCACTTCCCCATC | ACCTTGCCCACATATTCTCC | hPECAM | ATTGCAGTGGTTATCATCGGAGTG | CTCGTTGTTGGAGTTCAGAAGTGG | hZO-1 | CCCGAAGGAGTTGAGCAGGAAATC | CCACAGGCTTCAGGAACTTGAGG | * Housekeeping genes |
Human inflammation antibody arrays
To simultaneously detect 40 inflammation-associated proteins in EV and cell lysates, a membrane based human inflammation antibody C3 array (C-Series – AAH-INF-3 – RayBiotech Norcross-GA) was purchased from RayBiotech (Boechout, Belgium). Assays were performed according to the manufacturers’ instructions in two biological replicates. Briefly, 25 µg of total EV proteins in extraction buffer (Abcam) was subjected to a pre-blocked membrane and incubated overnight at 4°C with gentle shaking. Afterwards, the membrane was incubated with the primary biotin-conjugated antibody for 2 h, followed by incubation with HRP conjugated streptavidin antibodies for 1h at room temperature. Finally, the signal intensity of each array was imaged using the ImageQuant™TL detection system. Intensity of each dot was then quantified using ImageJ open source software (National Institutes of Health, USA). The spot densitometry data generated were imported into Microsoft Excel software to calculate normalized Z score, log2 changes. Then, heatmaps of inflammation-related protein expression were generated using GENE-E open source software. STRING analysis was conducted using high confidence (score 0.7). Cluster analysis was conducted using k-means with a value of k = 3.
Transendothelial electrical resistance
Transendothelial Electrical Resistance (TEER) of hCMEC/D3 cell monolayers was measured in real time to quantify BBB integrity. In short, 8250 cells in 250 µl growth medium were grown to confluency on collagen-coated 16-well RTCA E-Plates (Agilent, Santa Clara, CA, USA), containing interdigitated gold microelectrodes. The monolayers of hCMEC/D3 cells at 100% confluency were treated for a period of 48 h with TNF-α/IFN-γ solution (10 ng/mL, positive control), and increasing dose (from 5x107 to 6x108 particles/ml) of uEV, and tEV size-based EV subpopulations. Resistance values (in Ω) were collected at multiple frequencies, ranging from 1Hz to 1000 kHz at 5 frequencies per decade, by a palmSens 4 impedance analyzer, controlled by PSTrace software (palmSens BV, Houten, The Netherlands). TEER values were analyzed at a frequency of 6309.57 Hz, reflecting intercellular junctions, and data are depicted as Ω x cm², based on the well’s surface area (0.196 cm2). Data were normalized to the time point 0 hours of each condition individually and resistance curves were generated using GraphPad Prism software.
Flow cytometry
After 24 h treatment of hCMEC/D3 with equal amounts of EV subpopulations or 10ng/ml TNF-α/IFN-γ (positive control), cells were collected by scraping and stained with titrated amounts of mouse monoclonal anti-human antibodies against ICAM-1 (PE/Dazzle™ 594 anti-human CD54 Antibody, cat 353117, Bio Legend; 1/500), VCAM-1 (APC anti-human CD106 Antibody, cat 305810, Bio Legend; 1/250), and VE-Cadherin (FITC Mouse anti-human CD144, Clone 55-7H1, BD Bioscience; 1/10) for 15 min at room temperature. Acquisition was performed on a BD LSRFortessa™ cell analyzer (BD Bioscience, New Jersey, US) using the BD FACSDiva software (BD Bioscience). Data were analyzed with FlowJo software (BD Bioscience) version 10.7.1. Our gating strategy excluded doublets (using classical gating strategy), and defined positive gates using fluorescence minus one controls.
Transwell memory CD4+ T cell migration assays
CD4 + memory selection
All experiments were done using fresh peripheral blood mononuclear cell (PBMCs) isolated from whole blood by density gradient centrifugation (Cedarlane lympholyte, Sheffield, UK). Positive selection of CD4+ memory T cells from PBMCs using magnetic beads was performed according to the manufacturer’s protocol (130-091-893, Miltenyi Biotec, Leiden, The Netherlands). CD4 + memory T cells were used for Boyden chamber migration assays directly following isolation.
Migration assay with treated hCMEC/D3 cells
For migration assays, hCMEC/D3 were cultured in collagen coated Thincerts (24 well, translucent, 3 µm, Greiner Bio-One, Vilvoorde, Belgium) at a density of 25x103 cells/ cm². On days 3 and 5, hCMEC/D3 were replenished with experimental medium. Growth of the monolayer was followed using measurement of TEER with the EVOM2 Epithelial Voltohmmeter (World precision instruments, Hertfordshire, UK) and reached a plateau phase on day 6. Then, hCMEC/D3 were replenished with serum-reduced experimental medium, and either treated with tEV (sEV or lEV), inflammatory cytokines as described previously or left untreated for 24 hours. Prior to adding T cells, hCMEC/D3 on the inserts were washed with serum-reduced experimental medium and transferred to a new plate with fresh serum-reduced experimental medium. Memory CD4 + T cells (2.97x105-5x105 per insert, 3 inserts per condition) were allowed to migrate for 24 hours. After migration, T cells from the well (migrated) and from the insert (non-migrated) were collected, counted using an automatic cell counter (Moxi – Orflo technologies, cat.: MXZ000, Ketchum, US), and phenotyped with flow cytometric analyses. Cells were stained with titrated amounts of live dead (eF506, fixable viability dye (FVD), cat. 65-0866-14, Invitrogen), CD4 (APC eFluor 780 anti-human CD4 Antibody, cat. 47-0048-42, Invitrogen), CD45RO (PerCP/Cy5.5 anti-human CD45RO Antibody cat 304222, Bio Legend), CXCR3 (Brilliant Violet 711 anti-human CD183 Antibody, cat. 353732, Bio Legend), CCR4 (PE anti-human CD194 Antibody, cat. 359411, Bio Legend), CCR6 (PE/Cy7 anti-human CD196 Antibody, cat. 353417, Bio Legend) after which they were incubated for 30 min at room temperature. Next, cells were analyzed by flow cytometry as described previously. Flowcytometric gating is based on the presence and or absence of markers CCR6, CXCR3 and CCR4 the gating strategy can be found in Figure S1 .
Mice
Animal experiments were performed using 10- to 12-week old female wild-type (WT) C57BL/6 mice obtained from Envigo. Mice were housed in an accredited conventional animal facility under a 12 h light/dark cycle with free access to food and water. All procedures were conducted in accordance with the EU directive 2010/63/EU and with prior approval from the Hasselt University Ethics Committee for Animal Experiments.
Experimental Autoimmune Encephalomyelitis (EAE) and peripheral EV administration
EAE was induced in wild-type C57BL/6 mice using Hooke Labs EAE induction kit (EK-2110) according to the manufacturer’s instructions (Hooke Laboratories, Lawrence, MA, USA). Briefly, mice between 10 to 12 weeks of age were subcutaneously injected with myelin oligodendrocyte glycoprotein (MOG)35−55 peptide emulsified in complete Freund’s adjuvant (CFA) containing Mycobacterium tuberculosis. Directly after immunization, mice were injected intraperitoneally (i.p.) with 40ng/100µl pertussis toxin (PTX). Weights and neurological deficits were recorded daily by a blinded investigator. Neurological deficits were scored using a standard 5-point scale (0: no symptoms; 1: limp tail; 2: weakness of hind legs; 3: complete paralysis of hind legs; 4: complete hind and partial front leg paralysis; 5: death).
For systemic EV administration, cohorts were randomized according to mean disease severity and weight. Mice were intravenously (i.v.) injected in the tail vein with 100 µL of 1 x109 EV/mL at the onset of clinical EAE. Control mice received injections of 100 µl PBS on the same day as the EV injection (n = 12 per group). For post-mortem analysis after transcardial perfusion with PBS-heparin at the chronic phase of the disease (28 dpi), transversal halves of the spinal cords were snap-frozen using liquid nitrogen for immunohistochemistry analysis. Additionally, lumbar lymph nodes, spleen and remaining halves of spinal cord and brain were isolated and snap-frozen for gene expression analysis. Mice were selected for further analysis based on the mean clinical score of the experimental group (n = 4 per group).
Immunohistochemistry
Murine spinal cord tissue was cut into 10 µm serial sections using the Leica CM3050S cryostat (Leica Microsystems). The sections were post-fixed with frozen acetone for 10 min, followed by three washes with PBS. Sections were then incubated with the Dako protein blocking buffer (DAKO) for 1 h at RT. To determine leakage of blood vessels, sections were incubated with rabbit anti-laminin (1:2000, Abcam) and donkey anti-IgG Alexa 488 (1:800, Thermo Fisher Scientific). Binding of the primary antibody against laminin was visualized using goat-anti-rabbit Alexa 555-conjugated secondary antibody (Life Technologies) and nuclear staining was performed with DAPI. Finally, sections were incubated with 0.3% Sudan Black (Merck) in 70% ethanol to limit autofluorescence. All slides were mounted with Fluoromount-G™ Mounting Medium (Invitrogen™, Thermo Fisher Scientific). Microscopic analysis was performed using Leica DM2000 LED and Leica Application Suite X (LAS X) software (Leica Microsystems, Heidelberg, Germany). All analyses were performed in ImageJ (FIJI). In all experiments, a control staining was performed by omitting the primary antibody.
RNA isolation
RNA was isolated from snap-frozen CNS and spleen tissue using the RNeasy® Mini Kit (Qiagen, Venlo, The Netherlands) according to the manufacturer’s protocol. cDNA syntheses and qPCR were performed as described previously in this paper.
Statistical analysis
Data were presented as mean ± Standard error of means (SEM) of at least three independent samples. One-way analysis of variance (ANOVA) with a Dunn’s multiple comparisons test (Kruskal-Wallis test) and Student’s test using the statistical packages GraphPad Prism (version 10) was applied to evaluate the statistical significance between different treatments. Two-tailed tests at value of * p < 0.05 and were considered as statistically significant. NS represented as not significant, p > 0.05.
Data availability
We have submitted all relevant data of our experiments to the EV-TRACK knowledgebase (EV-TRACK ID: EV220166) (Van Deun J, et al. EV-TRACK: transparent reporting and centralizing knowledge in extracellular vesicle research. Nature methods. 2017;14(3):228 − 32).