Endothelial cells closely interact with circulating lymphocytes. Aggression or activation of the endothelium leads to an increased shedding of endothelial cell microparticles (MP). Endothelial MP (EMP) are found in high plasma levels in numerous immunoinflammatory diseases, such as atherosclerosis, sepsis, multiple sclerosis, and cerebral malaria, supporting their role as effectors and markers of vascular dysfunction. Given our recently described role for human brain microvascular endothelial cells (HBEC) in modulating immune responses, we investigated how HBEC-derived MP could interact with and support the proliferation of T cells. Like their mother cells, EMP expressed molecules important for Ag presentation and T cell costimulation, that is, β2-microglobulin, MHC II, CD40, and ICOSL. HBEC were able to take up fluorescently labeled Ags with EMP also containing fluorescent Ags, suggestive of Ag carryover from HBEC to EMP. In cocultures, fluorescently labeled EMP from resting or cytokine-stimulated HBEC formed conjugates with both CD4+ and CD8+ subsets, with higher proportions of T cells binding EMP from cytokine-stimulated cells. The increased binding of EMP from cytokinestimulated HBEC to T cells was VCAM-1 and ICAM-1 dependent. Finally, in CFSE T cell proliferation assays using anti-CD3 mAb or T cell mitogens, EMP promoted the proliferation of CD4+ T cells and that of CD8+ T cells in the absence of exogenous stimuli and in the T cell mitogenic stimulation. Our findings provide novel evidence that EMP can enhance T cell activation and potentially ensuing Ag presentation, thereby pointing toward a novel role for MP in neuroimmunological complications of infectious diseases.
The endothelial cells (EC) that line the microvasculature are in constant contact with blood cells such as T lymphocytes. CD4+ and CD8+ T lymphocytes play a critical role in cellular immunity, functioning synergistically to mount immune responses and eradicate infection. Nevertheless, the induction of adaptive cellular immunity is a function of professional APC such as dendritic cells (DC). APC provide signal 1 (peptide-MHC), signal 2 (costimulatory molecules), and signal 3 (instructive cytokines) to naive T cells upon Ag encounter (1).
A body of evidence supports the role of EC as APC (2–5) with the hypothesis based upon the intimate interactions between EC and T cells during their transendothelial migration to LNs or peripheral tissues. Moreover, EC may also qualify as APC, as they express MHC Ags and costimulatory molecules (3, 5), and secrete cytokines (6). T cell–EC interactions are central in diseases such as multiple sclerosis (MS), cerebral malaria (CM), and viral neuropathologies, although the precise mechanisms underlying these interactions remain unknown (7–9). We have previously demonstrated that human brain microvascular ECs (HBEC) take up Ags by macropinocytosis (5) and, in a CM model, can adopt Ags from Plasmodium-infected RBC, thereby becoming a target for the immune response (10).
EC express members of the Ig superfamily, including ICAM-1 and VCAM-1, that bind to leukocyte cell-surface Ags (11). ICAM-1 is a receptor for leukocyte cell-surface β2 integrins, such as LFA-1 and Mac-1, playing a key role in the adhesion and transmigration of blood leukocytes (12), whereas VCAM-1 is the endothelial receptor for VLA-4 (α4β1) and α4β7 (12, 13). HBEC are now known to express markers relevant for Ag presentation and T cell activation, such as β2-microglobulin (MHC I), MHC II, ICOSL, and CD40 (2, 5, 14–16). More recently, HBEC have been shown to display the potential for allo-antigen presentation (5).
Membrane vesiculation is a general physiological process that leads to the release of plasma membrane cell vesicles, called microparticles (MP). MP, a heterogeneous population of submicron elements, range in size from 100 to 1000 nm (17). MP are part of a family of extracellular vesicles, which may be characterized according to size range, phenotype, and function. Exosomes (30–100 nm) are derived from endocytic compartments within the cell and apoptotic bodies (≤4000 nm) are derived from endoplasmic membranes (18). MP can be generated by nearly every cell type during activation, injury, or apoptosis (19–22). In circulation, MP are derived from various vascular cell types, including platelets, erythrocytes, leukocytes, and, of particular interest, EC (20, 23). All MP, regardless of their cell of origin, have negatively charged phospholipids, such as phosphatidylserine, in their outer membrane leaflet, accounting for their procoagulant properties (24). MP also participate in homeostasis under physiological conditions. MP carry biologically active surface and cytoplasm nucleotides, allowing them to activate and alter the functionality of their target cells, thereby leading to the exacerbation of normal physiological processes, such as coagulation and inflammation (24).
Aggression or activation of the vascular endothelium leads to an increased shedding of endothelial MP (EMP). Although circulating EMP can be found in normal individuals, increased levels have been identified in a variety of pathological situations, including thrombosis, atherosclerosis, renal failure, diabetes, systemic lupus erythematosus, MS, and CM (21, 25–28). In these conditions, EMP express arrays of cell-surface molecules reflecting a state of endothelial dysfunction. These data highlight the link between endothelial damage, EMP release, and the modulation of inflammatory and/or immune responses.
Immune modulation by EMP has been described in very few settings. EMP induce plasmacytoid DC maturation and inflammatory cytokine production by DC (29) and can influence Th1 cell activation and secretion of cytokines in patients with acute coronary syndrome (30). Of note, MP isolated from Plasmodium-infected RBC contain Plasmodium Ags and activate macrophages and neutrophils (31, 32), demonstrating that MP from other cellular sources also can modulate the immune response. As endothelial alteration is associated with EMP release and T cell activation in numerous diseases, it raises the question of whether EMP carry molecules that enable them to bind to T cells that modulate their function. Moreover, given the potential for an HBEC line to modulate T cell responses (5), we aimed to investigate how EMP could interact with, and support the proliferation, of T cells.
Materials and Methods
Blood samples used in this study are from anonymous donors from the Australian Red Cross Blood Bank. Human ethics protocol was approved by the University of Sydney Human Ethics Committee (Approval #10218). The Animal Ethics Committee of the University of Sydney (K20/6-2011/3/5569 and K00/10-2010/3/5317) approved mouse experiments.
Cells and cell culture
Primary HBEC isolated from normal human brain tissue (HBEC, cAP-0002) were purchased from Angio-proteomie (Boston, MA; Supplemental Fig 4). The cAP-0002 HBEC are >95% positive (by immunofluorescence) for cytoplasmic von Willebrand factor/factor VIII, cytoplasmic uptake of Di-I-Ac-LDL, and cytoplasmic PECAM1. The HBEC are negative for HIV-1, hepatitis B virus, hepatitis C virus, and Mycoplasma. Mouse brain microvascular ECs (MBEC) from C57BL/6 mice (C1862) were purchased from Creative Bioarray (Shirley, NY). HBEC were cultured in supplemented EBM-2 medium (Lonza, Basel, Switzerland; Cat. no. CC-3156) and grown on plates precoated with rat-tail collagen type I (BD Biosciences, Franklin Lakes, NJ). MBEC were cultured in DMEM/F12 medium supplemented with 10% FCS (Life Technologies, Carlsbad, CA) and grown on plates precoated with 0.2% gelatin (Sigma-Aldrich, St. Louis, MO). THP-1 cells were obtained from American Type Culture Collection and grown in RPMI 1640 with 10% FCS. Cytokine-induced EMP production by HBEC and MBEC was performed by treating the cells with 100 ng/ml TNF and 50 ng/ml IFN-γ (PeproTech, London, U.K.) for 18 h.
Ag uptake analysis
The ability of EMP to carry Ags from mother cells in both resting and cytokine-stimulated conditions was assessed using the FITC-OVA (FITC-OVA) Ag uptake assay (33). Briefly, HBEC were incubated with 1 mg/ml FITC-OVA (Invitrogen, Carlsbad, CA) at 37°C for 45 min and washed three times with PBS. Cytokine-induced EMP release by HBEC was then performed by treating the cells with 100 ng/ml TNF and 50 ng/ml IFN-γ (PeproTech) for 18 h. FITC/CD105-positive EMP were enumerated on the Cytomics FC500 MPL flow cytometer (Beckman Coulter, Brea, CA), with the number of fluorescent-positive MP counted per 60 s. All cytometric analysis in this article was performed using FlowJo software (TreeStar, Ashland, OR).
Human PBMC preparation
PBMC were separated from leukopacks by conventional Ficoll gradient and frozen in 10% DMSO in FCS and stored in liquid nitrogen. PBMC were then thawed and washed twice in cold medium before use in assays.
Mouse lymph node cell preparation
Mouse lymph node (LN) cells from the LNs and spleen of both CBA and C57BL/6 mice were isolated by mechanical disruption of the organs and incubated in RBC lysis buffer (0.156 M ammonium chloride, 0.01 M sodium bicarbonate, and 1 mM EDTA) on ice. Cells were washed twice in cold medium before use in assays.
T cell isolation
CD4+ and CD8+ T cells were isolated from freshly thawed PBMC, using an EasySep negative selection kit (Stemcell Technologies, Vancouver, BC, Canada) according to the manufacturer’s instructions. Purity was assessed by flow cytometric analysis of mAbs against CD4 and CD8 (eBioscience, San Diego, CA), with purity always >95%.
For labeling isolated T cells and whole PBMCs with CFSE (Invitrogen), cells (at a density of 107 cells per milliliter) were incubated for 10 min at 37°C in 5 mM CFSE in FCS-free RPMI 1640. Labeling was stopped with FCS, and cells were washed three times prior to use. For the quantification of cell proliferation, cells were analyzed by flow cytometry, with a reduction in CFSE mean fluorescence intensity indicative of cell division.
HBEC or MBEC culture supernatants were centrifuged for 5 min at 1800 g; large cellular debris was discarded and supernatants were spun for 45 min at 18,000 g. MP pellets were resuspended in cell culture–grade PBS, and samples were spun once more for 45 min at 18,000 g in PBS to ensure removal of any remaining soluble cytokine. Final MP pellets were resuspended in RPMI 1640/10% FCS for quantification and use in assays.
Enumeration of EMP by flow cytometry
Purified EMP were enumerated according to positivity for CD105. EMP were diluted in PBS and stained with either anti-human CD105-PE (Beckman Coulter) or anti-mouse CD105-PE (eBioscience) for 30 min at RT. Samples were further diluted with PBS prior to flow cytometric analysis and were analyzed on the FC500 flow cytometer, and the numbers of C105-positive MP were recorded per 60 s. All analysis parameters were set to log scale, as described previously (34). Light forward scatter was plotted against the side-angle scatter, and MP were gated in the forward scatter as a population of < 1 μm in size, using calibrated latex beads. Fluorescence was plotted against side-angle scatter, with the number of positive fluorescent events representing the number of EMP.
Fluorescent labeling of EMP
Purified EMP were labeled with either PKH26 (red) or PKH67 (green) fluorescent cell linker kit for general cell membrane labeling (Sigma-Aldrich), as previously described (35). The labeled EMP were then washed twice in RPMI 1640/10% FCS (18,000 g for 45 min) and leached overnight at 4°C. EMP were centrifuged (18,000 g for 45 min) and resuspended in RPMI 1640/10% FCS prior to EMP enumeration using CD105.
Surface phenotype by flow cytometry
For flow cytometric analysis of HBEC, cells were incubated in the presence of fluorochrome-conjugated mAbs against CD105 (SN6), VCAM-1 (STA), B7-1/CD80 (2D10.4), B7-2/CD86 (IT2.2), CD40 (5C3), HLA-DR/MHC II (LN3), and CD275 (MIH12) (all from eBioscience), ICAM-1 (5.6E; Beckman Coulter), and β2-microglobulin/MHC I (TÜ99; BD Biosciences), per the manufacturers’ instructions. Flow cytometric analysis was performed on the FC 500 flow cytometer. For flow multicolor cytometric analysis of the activation status of donor T cells, cells were incubated with the following Ab mixture: CD69-FITC (FN50), CD40L-PE (CD154, 24-31) (eBioscience), CD8–Alexa Fluor 700, CD4–Pacific Blue, CD45RA–Alexa Fluor 750, CD3–Krome Orange (Beckman Coulter), and CD62L-APC (Miltenyi Biotec). Cells were washed once prior to flow cytometry using a Gallios Flow Cytometer (Beckman Coulter). Naive and memory T cell populations were distinguished by CD45RA and CD62L expression, as described previously (36). For EMP phenotyping, HBEC supernatant following stimulation was incubated with mAb and diluted 1:1 in PBS prior to flow cytometric analysis on the FC 500 flow cytometer. EMP production per 100,000 HBEC was determined by triplicate supernatant analysis. The supernatant was collected from 400,000 cultured HBEC per well following 18-h stimulation.
The ability of EMP to form conjugates with T cells was assessed using an in vitro conjugation assay (5, 37). Briefly, EMP were coincubated with freshly thawed PBMC (human) or isolated LN cells (mouse) (ratio of 1 PBMC/3 EMP) for 45 min at 37°C. Cells were washed once prior to staining at 4°C with fluorochrome-conjugated mAbs against human CD4/mouse CD4 (eBioscience) and human CD8/mouse CD8 (BioLegend/eBioscience). Samples were analyzed by flow cytometry. Conjugates were deemed to be positive for PKH and either CD4 or CD8. To block binding of EMP to T cells through ICAM-1 and VCAM-1, EMP were incubated with specific blocking Abs prior to conjugation. αICAM-1 (clone BBIG-I1; 10 μg/ml; R&D Systems, Minneapolis, MN), αVCAM-1 (clone BBIG-V1; 10 μg/ml; R&D Systems), and isotype control (mouse IgG1) were added to purified EMP for 1 h at 37°C. EMP were cocultured with PBMC for conjugation, as outlined above. EMP binding to specific CD4+ and CD8+ T cell subtypes was examined by surface staining with the following Ab mixture after EMP conjugation to PBMC: CD8–Alexa Fluor 700, CD4–Pacific Blue, CD45RA–Alexa Fluor 750, CD3–Krome Orange (Beckman Coulter), and CD62L-APC (Miltenyi Biotec). Cells were washed once prior to flow cytometry using a Gallios Flow Cytometer. Naive and memory T cell populations were distinguished by CD45RA and CD62L expression, as described previously (36). Flow cytometric analysis of the naive/effector/memory subsets for both CD4+ and CD8+ donor T cells is shown in Supplemental Fig. 2.
Confocal microscopic analysis of EMP/T cell conjugates
CFSE-labeled isolated CD4+ and CD8+ T cells were cocultured with PKH26-labeled EMP at a ratio of 1:4 for 45 min at 37°C. Cell/EMP conjugates were then mounted in PBS and imaged with the Zeiss LSM 510 Meta Confocal microscope at ×63/1.40 oil with 1× and 1.6× optical zoom.
In vitro T cell proliferation assays
A total of 5 × 104 CFSE-labeled PBMCs were added per well with the following conditions: PBMC alone, 1 μg/ml Con A (Sigma) or 0.3 μg/ml for anti-CD3 mAb (eBioscience; clone HIT3a). The agonistic anti-CD3 mAb was added to the assay to mimic TCR stimulation (38), and Con A was used as a T cell mitogen (39). Resting or TNF + IFN-γ EMP were added to the proliferation assay at a ratio of 3 EMP:1 PBMC. THP-1, resting HBEC, or TNF + IFN-γ–stimulated HBEC were γ-irradiated (3000 rads) and added at a ratio of 1 cell:1 PBMC. The cocultures were incubated for 5 d at 37°C and then stained with PE-conjugated anti-human CD4 (clone OKT4; eBioscience) and PE-Cy5 anti-human CD8a (clone HIT8a; BioLegend).
All statistical analyses were performed using GraphPad Prism 6.0 (GraphPad Software). The Mann–Whitney U test was used to evaluate statistical significance; p < 0.05 was deemed significant.
EMP express key molecules for Ag presentation and T cell activation
Given that EMP express markers of the parent EC, we sought to examine whether EMP released from our HBEC expressed such markers relevant to Ag presentation and T cell costimulation. Numbers of annexin V+- and CD105+ MP, both considered constitutive markers of MP, were significantly increased following stimulation with TNF and further increased with TNF + IFN-γ dual stimulation (Fig. 1). This increase in CD105+ EMP numbers following TNF or TNF + IFN-γ occurred without any change in CD105 expression on parent HBEC following the stimulations (Supplemental Fig 1). Resting HBEC released low levels of ICAM-1–positive EMP, compatible with their low basal expression, after TNF or TNF + IFN-γ stimulation; a significantly greater number of ICAM-1+ MP were detected following stimulation of HBEC (Fig. 1). Although VCAM-1 was not detected on EMP released from resting cells, nor on the resting cells themselves (Fig. 1, Supplemental Fig 1), significantly higher numbers of VCAM-1+ MP were detected following TNF or TNF + IFN-γ stimulation of HBEC (Fig. 1). Unlike the human CMEC/D3 line, in which the constitutive expression of MHC I (β2-microglobulin) is unchanged following cytokine stimulation (5), MHC I expression was upregulated by cytokine stimulation in primary cells (Supplemental Fig 1). In the case of the EMP, MHC I+ MP were released by resting HBEC, and their numbers increased following stimulation of the HBEC with either TNF or TNF + IFN-γ (Fig. 1). Most interestingly, both MHC II+ and CD40+ MP were detected following TNF +IFN-γ stimulation (Fig. 1), whereas ICOSL+ MP were detected in both TNF and TNF + IFN-γ stimulation conditions (Fig. 1). This EMP phenotype mirrors that seen on parent cells (Supplemental Fig 1). The primary cells used in this assay did not express other costimulatory molecules, such as B7-1 and B7-2 (Supplemental Fig 1).
EMP carry foreign Ags from the parent cell following vesiculation
As Ag presentation to T cells is a fundamental step in an adaptive immune response, in this study we examined whether EMP carried Ags from their mother cell following vesiculation, using the FITC-OVA Ag uptake assay. OVA is widely used as a model Ag for the characterization of Ag uptake and processing in APCs. Following Ag uptake with FITC-OVA, resting HBEC produced detectable levels of FITC-positive EMP (Fig. 2A), indicating transfer of the Ag to the MP following vesiculation. Moreover, when HBEC were stimulated with TNF + IFN-γ following Ag uptake, increased numbers of FITC-positive EMP were detected (Fig. 2A, 2B).
EMP bind to and form conjugates with CD4+ and CD8+ T cells
Given the expression of a number of surface molecules on EMP that have corresponding ligands/receptors on T cells, we next sought to examine these interactions in vitro. EMP isolated from both resting and cytokine (TNF + IFN-γ)–stimulated HBEC were fluorescently labeled with PKH26 (red) and cocultured with green fluorescently labeled (CFSE) CD4+ and CD8+ T cells prior to confocal microscopic analysis. Of note, flow cytometric analysis of the T cell activation markers CD69 and CD40L on both CD4+ and CD8+ donor T cells indicated that the donor T cells were not activated (Supplemental Fig. 2). EMP purified from both resting and cytokine-stimulated HBEC bound readily to both CD4+ and CD8+ T cells (Fig. 2C). Of interest, more than one EMP was found bound per T cell (Fig. 2C); however, not every T cell imaged had EMP bound (Fig. 2D).
As optimal T cell activation and differentiation in vivo requires long-lasting T cell–APC interaction, a classical in vitro conjugate-forming assay was adapted to assess the ability of EMP to form conjugate-like interactions with T cells. CD4+ or CD8+ T cells purified from PBMC were incubated in suspension with red fluorescently labeled (PKH26) EMP, and the adherence between EMP and T cells was examined using cell-specific staining and flow cytometry. Conjugates were defined as cells positive for both PKH26 and the subset-specific mAb for CD4 or CD8. EMP from both resting and TNF + IFN-γ–stimulated HBECs formed conjugates with CD4+ and CD8+ T cells (Fig. 3A). There were significantly higher percentages of T cell/EMP conjugates when EMP were purified from TNF + IFN-γ–stimulated than from resting cells (Fig. 3B). In the case of resting EMP, 18.5% of CD4+ and 17.4% of CD8+ T cells had EMP bound. When T cells were cocultured with EMP from TNF + IFN-γ–stimulated HBEC, the increase in conjugation led to 40.6% of CD4+ and 46.8% of CD8+ T cells with EMP bound.
Role of MHC mismatch in EMP/T cell binding
In the human system, syngeneic EMP/T cells are not available; therefore, as the donor of HBEC is different from that of PBMC, potential allogeneic differences may result in the EMP/T cell binding and conjugation observed. To ensure this was not the case, EMP purified from MBEC were put into a conjugation assay with donor T cells from two different mouse strains. The C1862 MBEC line used was isolated from C57BL/6 (H-2b). EMP from C1862 cells were purified and conjugated to LN cells from two different donor strains: C57BL/6 (H-2b; syngeneic) and CBA/J (H-2k; allogeneic). Mouse EMP conjugated at equivalent levels to both syngeneic and allogeneic T cells (Fig. 3C), suggesting the binding of human EMP to donor T cells was not due to any MHC mismatch between the EMP and T cells.
Identification of receptors involved in EMP/T cell binding
As EMP purified from HBEC (Fig. 1B) and other sources (23, 40) express ICAM-1 and VCAM-1, we sought to determine whether these molecules play a role in EMP binding to T cells. Blocking VCAM-1, ICAM-1, or both on EMP isolated from resting HBEC prior to conjugation to PBMC had no effect on their binding to CD4+ and CD8+ T cells (Fig. 4A). Similarly, blocking VCAM-1 on EMP isolated from TNF + IFN-γ–stimulated cells did not change the level of EMP binding to either T cell subset (Fig. 4A). However, blocking ICAM-1 on TNF + IFN-γ EMP resulted in significantly decreased conjugation to both CD4+ and CD8+ T cells (Fig. 4A). Moreover, blocking both VCAM-1 and ICAM-1 on TNF + IFN-γ EMP resulted in an even greater reduction in conjugation to both T cell subsets, with the level of binding almost equivalent to that seen with resting EMP (Fig. 4A).
EMP bind preferentially to memory T cells
To further examine which specific populations of T cells EMP bind to, six-color flow cytometric analysis was conducted using the following gates on both CD4 and CD8 subsets: naive (N; CD45RA+, CD62L+), central memory (TCM; CD45RA−, CD62L+), effector memory (EM;. CD45RA−, CD62L−), and terminal effector memory (TEM; CD45RA+, CD62L−) cells. For both CD4+ and CD8+ subsets, there was significantly more EMP binding to effector memory subsets than to the naive and terminally differentiated effector memory subsets (Fig. 4B). In addition, significantly more EMP were bound to central memory than to naive T cells.
EMP promote the proliferation of CD4+ and CD8+ T cells
Because HBEC support and promote the proliferation of T cells (5) and because EMP were capable of binding to and forming conjugates with both CD4+ and CD8+ T cells, the ability of HBEC to support T cell proliferation was assessed by coculturing CFSE-labeled donor PBMCs with EMP isolated from either resting or TNF + IFN-γ–stimulated HBEC. Five days following coculture, the percentage of proliferating CD4+ and CD8+ T cells was determined by measuring the reduction in CFSE mean fluorescence intensity. In the case of CD4+ T cells, EMP had no effect in resting conditions (Fig. 5A, 5B). However, in the presence of TCR or T cell mitogenic stimulation, both EMP from resting cells and TNF + IFN-γ MP increased the percentage of proliferating cells (Fig. 5A, 5B) when compared with stimuli alone. Of note, no significant difference was found between EMP from resting and cytokine-stimulated HBEC in the ability to promote CD4+ T cell proliferation. Of interest, both resting EMP and TNF + IFN-γ EMP significantly increased the percentage of proliferating CD8+ cells in the absence of any mitogen or TCR activation (Fig. 5A, 5B). Similarly, when CD8+ T cells were stimulated with Con A and cocultured with EMP from both resting and TNF + IFN-γ–stimulated HBEC, a significant increase in the percentage of proliferating cells was observed (Fig. 5A, 5B). As for CD4+, no significant difference was found between EMP from both resting and TNF + IFN-γ–stimulated HBEC in the ability to promote resting CD8+ T cell proliferation. In addition, EMP had no effect on CD8+ proliferation when T cells were stimulated with anti-CD3 mAb (Fig. 5A, 5B). Interestingly, both resting and TNF + IFN-γ EMP did not increase the proliferation of CD4+ and CD8+ T cells to the same level as that seen by the parent HBEC or by professional APC such as THP-1 cells (Supplemental Fig. 3).
In this article, we provide evidence that MP isolated from human brain MVEC possess the minimal machinery required to present Ags and to interact with and support the activation of T cells.
Our analysis of MHC and costimulatory molecule expression on EMP shows for the first time, to our knowledge, that EMP are endowed with a costimulatory ligand of the B7 family and TNFR superfamily, ICOSL and CD40, respectively. This feature, in conjunction with their expression of MHC II, supports the idea of brain EMP being able to present Ags to and costimulate T cells promoting effector CD4+ and T cell responses. In addition, with high numbers of MHC I+ EMP produced by HBEC, EMP possess the minimal requirement for Ag presentation to CD8+ T cells. Combined, these findings demonstrate expression of molecules relevant to Ag presentation on EMP and indicate that EMP have the capacity to interact with and modulate target T cells.
Previously, expression of MHC molecules on MP has been restricted to professional APCs. MP released from mature DCs may modulate local or distant adaptive immunologic responses by transferring MHC molecules to resting DCs, thereby allowing them to present alloantigens to T cells (41). Similarly, MP from bovine B cells confer Ag-presentation capabilities to bovine polymorphonuclear leukocytes by shuttling MHC class II molecules (42). Finally, MPs from both Mycobacterium tuberculosis–infected and uninfected macrophages express MHC II and CD40 (43). The data presented in this article provide evidence for endothelial alteration being able to modulate distal Ag recognition and costimulation in times of inflammation or infection through expression of MHC and costimulatory molecules.
Ag uptake is the first step in Ag-presenting pathways, leading to the initiation of an adaptive immune response. EC have been previously shown to take up Ag using macropinocytosis, clathrin-coated pits (5), and the mannose receptor (44). In this article, we show that EMP are able to transfer the soluble fluorescent Ag FITC-OVA following vesiculation, providing further evidence to support the immunogenic capacity of EMP, particularly to distal target cells. These data indicate that EMP, following Ag uptake by their mother cell HBEC, have the capacity to carry Ag, thereby participating in a fundamental initial step in the primary immune response (45).
Despite evidence for EMP modulation of Th1 cell function (30), there are no published data showing EMP interacting directly with T cells. In this study, using confocal microscopy and flow cytometry, we are able to demonstrate binding between EMP and both CD4+ and CD8+ T cells. Of interest, EMP produced by HBEC following TNF + IFN-γ stimulation bound more readily to T cells, with twice as many T cells binding these EMP compared with those from resting cells. Using mouse EMP in coculture with syngeneic and allogeneic donor T cells, we were able to show equivalent conjugation to T cells, demonstrating that the binding, in the human EMP system, is not due to any allogeneic differences between HBEC and PBMC donors.
Using blocking Abs against the endothelial adhesion molecules VCAM-1 and ICAM-1 enabled us to define the specific interactions involved in the increased binding of EMP from cytokine-stimulated cells to T cells. Both CD4+ and CD8+ T cells displayed increased binding of TNF + IFN-γ EMP that was mediated predominantly through ICAM-1, thus indicating that EMP isolated from TNF + IFN-γ HBEC bind to LFA-1 and Mac-1 expressed on target T cells (12). The addition of both blocking Abs brought the binding of TNF + IFN-γ EMP back to levels equivalent to those of resting EMP, suggesting that the increased binding is only partially mediated though the VCAM-1 ligands VLA-4 and α4β7 expressed on target T cells (12, 13). Given the high number of TNF + IFN-γ EMP that express ICAM-1, it is not unsurprising that this is the adhesion molecule–mediated binding in these conditions. Blocking ICAM-1 and VCAM-1 on the EMP had no effect on the binding of resting EMP to T cells, indicating that this binding occurs through other receptor/ligand pairs or phosphatidylserine at the surface of MP (46). Interestingly, binding experiments with EMP and plasmacytoid DC have been shown to be dependent on temperature, sodium–proton exchanges, and an intact actin cytoskeleton (29). Moreover, in this study EMP bound to CD4+ T cells at both 37°C and 4°C (29); however, the ratio of EMP/T cell used was 100:1, substantially higher than the 3:1 ratio used in our study. Finally, further flow cytometric analysis showed that EMP from cytokine-stimulated cells have a greater propensity for binding to effector memory T cells over any other subset, with the least binding observed to naive CD4+ and CD8+ T cells. These differences may in part be due to increased expression of LFA-1 and VLA-4 on memory T cells versus naive cells (47, 48). Physiologically, preferential EMP binding to memory and effector T cells, particularly during inflammatory situations, such as that mimicked in this study with TNF + IFN-γ, would lead to an exacerbated effector response to infection.
In the coculture assays presented in this article, EMP were able to support and promote the proliferation of CD4+ and CD8+ T cells. Nonstimulated CD4+ cells were unresponsive to both resting and TNF + IFN-γ EMP. However, when the CD4+ T cells were stimulated via their TCR (anti-CD3 mAb) or with Con A, a significant increase in proliferation was observed in the presence of both resting and TNF + IFN-γ EMP, with, interestingly, no difference in the proliferation induced by the two types of EMP. This finding may be due to the threshold of activation and proliferation being at a maximal level, with no increase thus able to be observed in the presence of TNF + IFN-γ EMP. Conversely, when nonstimulated CD8+ T cells were cocultured with EMP, a significant increase in their proliferation was observed over no MP control, with this effect possibly owing to differences in MHC between PBMC donor and HBEC donor. Similar increases in proliferation were observed for Con A stimulation, but not for anti-CD3 mAb, and this may result from the fact that a large majority of the CD8+ T cells alone were proliferating 5 d after coculture in response to anti-CD3. In combination, the data in these experiments indicate that EMP, regardless of source (resting or cytokine activated), are able to support and promote the proliferation of both CD4+ and CD8+ T cells. Given that elevated numbers of circulating EMP are often associated with disease severity in a number of T cell–dependent immunoinflammatory diseases such as MS and CM (26, 49), these findings provide a novel mechanism for pathogenesis and pathology in such diseases.
The demonstration of Ag-specific activation of human T cells by EMP is hampered by the requirement for MHC-matched HBEC and T cells. Some studies using MHC-matched donors support the model that cultured human EC are able to present Ag to activated CD4+ T cells (50–52). However data indicate that mouse T cell clones or T cells from TCR-transgenic mice can be stimulated to proliferate in a peptide-antigen–specific manner by coculture with MHC-matched EC and the relevant protein Ag (53, 54), providing evidence for Ag-specific activation by EC. Finally, the in vivo relevance of our observations remains to be explored further.
In summary, we have shown that MP isolated from human brain microvascular endothelium express molecules important for T cell stimulation and activation, including CD40, ICOSL, and MHC II. Moreover, they transfer soluble Ag following vesiculation and bind readily to T cells using both ICAM-1–dependent and –independent mechanisms. Moreover, EMP promote the proliferation of T cells in vitro. The data presented in this article support the hypothesis that EMP are immunomodulatory and show direct interactions between EMP and target T cells. These data add to the evidence for EMP as effectors, as they have previously been shown to be involved in inflammatory processes, blood coagulation, and regulation of vascular function (55). The findings presented in this article offer a novel mechanism for the role of EMP in neuroimmunological complications and are pertinent to our understanding of vascular dysfunction in immune and inflammatory responses.
This article is dedicated to the memory of Dr. Charlotte Behr (University of Bordeaux, Bordeaux, France).
This work was supported by National Health and Medical Research Council Project Grants APP1028241 and APP571014, National Health and Medical Research Council Training Fellowship APP571397 (to J.W.), National Institutes of Health Grant RO1NS0789873-1, and the Rebecca L. Cooper Medical Research Foundation.
The online version of this article contains supplemental material.
Abbreviations used in this article:
human brain microvascular EC
mouse brain microvascular EC
The authors have no financial conflicts of interest.