Internalization and presentation of myelin antigens by the brain endothelium guides antigen-specific T cell migration

Melissa A Lopes Pinheiro; Alwin Kamermans; Juan J Garcia-Vallejo; Bert van het Hof; Laura Wierts; Tom O'Toole; Daniël Boeve; Marleen Verstege; Susanne MA van der Pol; Yvette van Kooyk; Helga E de Vries; Wendy WJ Unger

doi:10.7554/eLife.13149

. 2016 Jun 23;5:e13149. doi: 10.7554/eLife.13149

Internalization and presentation of myelin antigens by the brain endothelium guides antigen-specific T cell migration

Melissa A Lopes Pinheiro ¹, Alwin Kamermans ¹, Juan J Garcia-Vallejo ¹, Bert van het Hof ¹, Laura Wierts ¹, Tom O'Toole ¹, Daniël Boeve ¹, Marleen Verstege ¹, Susanne MA van der Pol ¹, Yvette van Kooyk ¹, Helga E de Vries ^1,^*,^†, Wendy WJ Unger ^1,^*,^†,^‡

Editor: Gary L Westbrook²

PMCID: PMC4959842 PMID: 27336724

Abstract

Trafficking of myelin-reactive CD4⁺ T-cells across the brain endothelium, an essential step in the pathogenesis of multiple sclerosis (MS), is suggested to be an antigen-specific process, yet which cells provide this signal is unknown. Here we provide direct evidence that under inflammatory conditions, brain endothelial cells (BECs) stimulate the migration of myelin-reactive CD4⁺ T-cells by acting as non-professional antigen presenting cells through the processing and presentation of myelin-derived antigens in MHC-II. Inflamed BECs internalized myelin, which was routed to endo-lysosomal compartment for processing in a time-dependent manner. Moreover, myelin/MHC-II complexes on inflamed BECs stimulated the trans-endothelial migration of myelin-reactive Th1 and Th17 2D2 cells, while control antigen loaded BECs did not stimulate T-cell migration. Furthermore, blocking the interaction between myelin/MHC-II complexes and myelin-reactive T-cells prevented T-cell transmigration. These results demonstrate that endothelial cells derived from the brain are capable of enhancing antigen-specific T cell recruitment.

DOI: http://dx.doi.org/10.7554/eLife.13149.001

Research Organism: Human, Mouse

eLife digest

The blood vessels in the brain help to control the entry of nutrients, cells and waste products into and out of the brain. In doing so, they create a protective barrier between the blood and the brain known as the blood-brain barrier. However, this barrier loses its protective function in individuals with multiple sclerosis or other disorders that affect the brain. Multiple sclerosis patients develop inflammation and their immune cells become able to enter the brain. These immune cells may then attack layers of insulation called myelin that surround nerve cells. Myelin helps nerve cells to work properly so the loss of this insulation can lead to tissue damage and cognitive problems. When immune cells called T cells enter the brain they can become primed to recognize myelin and attack it in the same way that they would attack viruses or bacteria. However, it is not clear precisely how these T cells develop the ability to cross the blood-brain barrier and attack myelin.

Now, Lopes Pinheiro et al. show that “endothelial” cells in the blood-brain barrier are able to present fragments of myelin to T cells, which enables the T cells to identify myelin and move into the brain. First, the blood-brain barrier cells absorb and break down proteins in the myelin, and then they present fragments of these proteins on their surfaces with the help of protein clusters called major histocompatibility complexes (MHCs). Other protein fragments that can also activate T cells in other parts of the body did not affect the blood-brain barrier when they were presented by MHCs, which suggests that the effect could be specific to myelin proteins.

The experiments also show that it is possible to stop T cells from crossing the blood-brain barrier by preventing them from interacting with myelin fragments presented by MHCs. This suggests that therapies that interfere with the ability of blood-brain barrier cells to break down myelin proteins and present them to T cells might help to protect the brains of patients with multiple sclerosis.

DOI: http://dx.doi.org/10.7554/eLife.13149.002

Introduction

In the neuro-inflammatory disorder multiple sclerosis (MS), the trafficking of immune cells into the brain is a crucial step not only in the onset but also the progression of the disease (Lublin et al., 2014; Sospedra and Martin, 2005). Once within the central nervous system (CNS), myelin-reactive T-cells induce severe neuronal and tissue damage and degeneration (Trapp et al., 1998). During the initiation of the disease, myelin-specific CD4⁺ T cells are differentiated into effector T helper (Th) 1 or Th17 cells (Bailey et al., 2007; Bielekova et al., 2000). There is compelling evidence that Th1 and Th17 cells, separately or in cooperation, mediate deleterious responses in MS (Carbajal et al., 2015).

To enter the CNS, immune cells have to cross the blood-brain barrier (BBB), which is composed of highly specialized brain endothelial cells (BECs) that are sealed by closely regulated tight junctions (Tietz and Engelhardt, 2015). Lymphocyte migration into the CNS parenchyma is a multi-step process that requires close contact between lymphocytes and BECs (Ransohoff et al., 2003). These cell-cell contacts are mediated by cell surface molecules on both the lymphocytes and BECs. During inflammation BECs upregulate the expression of the adhesion molecules Inter-Cellular Adhesion Molecule (ICAM)−1, Vascular cell adhesion molecule (VCAM)−1 and Activated leukocyte cell adhesion molecule (ALCAM), which are necessary for the firm adhesion of lymphocytes to the endothelium as well as for the trans-migration process (Larochelle et al., 2011). Although these adhesion molecules expressed by BECs have been shown important for the transendothelial migration of leukocytes, the complexity of this interaction and the molecules involved remain poorly understood.

Several studies provided evidence for an antigen-specific component in the transmigration process of encephalitogenic T-cells (Archambault et al., 2005; Galea et al., 2007; Ludowyk et al., 1992). Using an adoptive transfer model of murine MS (experimental autoimmune encephalomyelitis, EAE) it was demonstrated that only activated myelin-specific CD4⁺ T-cells accumulated in the CNS parenchyma, while non-CNS-specific T-cells failed to infiltrate (Archambault et al., 2005). Furthermore, expression of MHC-II on the recipient cells appeared to be required for CNS infiltration, as the myelin-specific T-cells did not transmigrate over CNS vascular endothelium when adoptively transferred in MHC-II deficient mice. However, whether the antigen-specific signal was provided by APCs or BECs was not elucidated. Also infiltration of CD8⁺ T-cells into the brain was shown to be an antigen-specific process: haemagglutinin-specific CD8⁺ T-cells were only detected in the CNS upon intra-cerebral injection of cognate, but not control, peptides in an haemagglutinin T-cell receptor transgenic mouse (Galea et al., 2007). A role for BECs in providing the antigen signal to T-cells was claimed by showing luminal expression of MHC-I on BECs and the fact that intra-venous injection of a blocking MHC-I antibody significantly reduced the CD8⁺ T-cell infiltration. However, since a soluble, nominal MHC-I epitope was used as antigen, exogenous binding of these peptides to MHC-I molecules on other cells in the CNS, or even in CNS-draining lymph nodes cannot be excluded (Weller et al., 1996). Thus, so far no direct evidence is provided for a role of BECs in processing and presenting CNS-derived antigens during inflammatory conditions.

BECs have been shown to express MHC-I while MHC-II is virtually absent. Inflammation induced activation of BECs causes increased expression of MHC-II as well as of the co-stimulatory molecule CD40 and enhanced their ability to stimulate the proliferation of allogeneic T-cells in-vitro (Wheway et al., 2013). Although BECs have been shown to take up soluble antigens by macro-pinocytosis and clathrin-coated pits (Wheway et al., 2013), not much is known about their capacity to process and present internalized antigens.

We therefore explored the potential of BECs as antigen-presenting cells and determined whether antigen-presentation by BECs contributes to transmigration of myelin-reactive T-cells. We here demonstrate that inflamed BECs take up and process myelin via the endo-lysosomal degradation pathway in a time-dependent manner. Importantly, these myelin-derived antigens are presented in de-novo expressed MHC-II molecules and facilitate the migration of antigen-specific Th1 and Th17 pathogenic T-cells through the brain endothelium. Better insight into the events that trigger T-cell migration into the brain is crucial for our understanding of MS pathogenesis and will aid the development of new treatments to prevent T-cell infiltrating the CNS.

Results and discussion

Brain endothelial cells internalize exogenous antigens irrespective of their activation status

To determine if BECs play a role in antigen-specific migration of CD4⁺ T cells by acting as APCs, we first assessed the expression of molecules necessary for antigen presentation and co-stimulation. Resting, non-inflamed, human BECs express MHC-I and PD-L1 while MHC-II, CD40 and VCAM−1 are expressed at low levels (Figure 1A). Upon inflammatory activation, BECs express high levels of VCAM−1, and significantly increased the expression levels of MHC-II (Figure 1A,B). Similarly, CD40 expression was increased upon activation. Both MHC-I and PD-L1 were highly expressed on resting as well as on activated BECs. Expression of the classical co-stimulatory molecules CD80 and CD86 were undetectable on resting and activated BECs (data not shown). Comparable changes in phenotype were observed when BECs were activated using IFN-γ instead of TNFα (Figure 1—figure supplement 1) Together, these results confirm and extend previous findings (Wheway et al., 2013) and indicate that BECs are equipped to present antigens under inflammatory conditions. Up-regulation of MHC class II molecules via inflammation induced CIITA activity has been associated with increased susceptibility of EAE, yet how increased MHC-II expression contributes to actual disease has so far not been described (Reith et al., 2005).

Figure 1. — Confluent monolayers of brain endothelial cells (BECs) were stimulated with 5 ng/ml TNFα for 24 hr. (A) Expression of MHC-I, MHC-II, CD40, PD-L1 and VCAM−1 was determined by flow cytometry. Histograms depict expression of indicated markers in resting (grey solid line) and activated (black solid line) BECs. Dashed lines indicate isotype controls. (B) The MFI of expression of the indicated markers is shown. Data are presented as the mean ± SD of duplicate values (n = 5 independent experiments). *p<0.05, **p<0.01, ***p<0.001 (Student t-test). (C–E) Fluorescent labeled human myelin was added to resting or activated BECs for 4 hr or 24 hr and uptake was analyzed by (C) flow cytometry or (D–E) imaging flow cytometry. (C) Representative facs plots of myelin uptake by BECs, numbers in plots indicate the MFI of myelin-positive cells. The percentage of myelin-positive resting and activated BECs at 4 and 24 hr after loading with antigen is shown in a graph. (D–E) Myelin-positive BECs internalized between 1–3 particles/cell. On average, BECs acquired 2–3 myelin particles/cell. Activation of BECs did not affect the number of internalized particles. The average number of internalized myelin particles per cell is shown in a bar graph. Data presented are the means of triplicate values ± SEM of at least three independent experiments. *p<0.05, **p<0.01, ***p<0.001 (Student t-test).

**DOI:** http://dx.doi.org/10.7554/eLife.13149.003

Figure 1—figure supplement 1. — Confluent monolayers of brain endothelial cells (BECs) were stimulated with 5 ng/ml TNFα for 24 hr. (A) Expression of MHC-I, MHC-II, CD40, PD-L1 and VCAM−1 was determined by flow cytometry. Histograms depict expression of indicated markers in resting (grey solid line) and activated (black solid line) BECs. Dashed lines indicate isotype controls. (B) The MFI of expression of the indicated markers is shown. Data are presented as the mean ± SD of duplicate values (n = 5 independent experiments). *p<0.05, **p<0.01, ***p<0.001 (Student t-test). (C–E) Fluorescent labeled human myelin was added to resting or activated BECs for 4 hr or 24 hr and uptake was analyzed by (C) flow cytometry or (D–E) imaging flow cytometry. (C) Representative facs plots of myelin uptake by BECs, numbers in plots indicate the MFI of myelin-positive cells. The percentage of myelin-positive resting and activated BECs at 4 and 24 hr after loading with antigen is shown in a graph. (D–E) Myelin-positive BECs internalized between 1–3 particles/cell. On average, BECs acquired 2–3 myelin particles/cell. Activation of BECs did not affect the number of internalized particles. The average number of internalized myelin particles per cell is shown in a bar graph. Data presented are the means of triplicate values ± SEM of at least three independent experiments. *p<0.05, **p<0.01, ***p<0.001 (Student t-test).

**DOI:** http://dx.doi.org/10.7554/eLife.13149.003

Since myelin-derived antigens are the major target of auto-reactive T-cells in MS, we investigated if BECs can take up and process myelin. We therefore incubated BECs with fluorescent labeled myelin for different time-points under resting and inflammatory conditions and determined myelin uptake by flow cytometry. As depicted in Figure 1C a time-dependent increase in the proportion of myelin⁺ BECs was observed. Moreover, this process is not significantly affected by treatment with inflammatory stimuli as activated BECs showed a similar amount of internalized myelin as resting BECs. Using imaging flow cytometry, we assessed that BECs that were able to capture myelin increased the number of myelin particles over time to a maximum of three myelin particles/cell after a 24 hr incubation (Figure 1D). Moreover, the average amount of myelin particles per cell was the same in both resting and inflammatory conditions, again, demonstrating that this process is not significantly affected by treatment with inflammatory stimuli (Figure 1D,E). Of note, in order to measure whether the localization of the myelin signal was intracellular or membrane-bound, we designed a mask that excludes the cell membrane and calculated a ratio of the amount of fluorescence located in the mask vs the total amount of fluorescence, as previously reported (Garcia-Vallejo et al., 2015). The results indicate that the myelin fluorescence signal was intracellular, demonstrating that BECs are able to efficiently internalize myelin (Figure 1—figure supplement 2).

Myelin internalized by BECs is directed to the endo-lysosome compartments

The endo-lysosomes are the typical antigen-processing compartments of APCs (Blum et al., 2013; Roche and Furuta, 2015). This intracellular route allows optimal processing of exogenous protein antigens and transfer of antigen-derived peptides to the MHC-II compartment for loading and subsequent presentation to CD4⁺ T-cells. To determine whether internalized myelin is shuttled to these compartments in BECs, myelin-treated BECs were stained with antibodies against EEA1 (a marker of early endosomes) and LAMP1 (a marker of late endosomes and lysosomes) to measure co-staining with myelin using imaging flow cytometry. We observed that myelin co-localized with both EEA1 and LAMP1 as shown by a high co-localization score (Figure 2A,B). The co-localization with both markers was higher at 24 hr of exposure to myelin compared to 4 hr. Since the increase of the co-localization score for myelin-EEA1 was not as strong as shown for myelin-LAMP1 at 24 hr (Figure 2A,B), this suggests that at that time point the majority of myelin was present in lysosomes. However, non-internalized myelin fragments that are attached to the cell membrane, could potentially be 'internalized' as a consequence of trypsinization of adherent BECs. To demonstrate that myelin is actively taken up by BECs, we analyzed myelin uptake and intracellular routing in adherent BECs using confocal laser scanning microscopy. Similar to our experiments using imaging flow cytometry, we observe that 24 hr after loading of adherent BECs a proportion of cells show internalized myelin. Furthermore, it is clear that internalized myelin is present within LAMP1 positive vesicles and not with EEA1 positive organelles (Figure 2C–I).

Figure 2. — Resting or activated BECs were loaded with Atto-633 labeled myelin for 4 hr or 24 hr. Uptake of myelin particles and their co-localization with early endosomal (EEA1) or endosomal/lysosomal (LAMP1) compartments was analyzed by imaging flow cytometry and quantified using the brightfield similarity R3 feature (see methods for details). Myelin particles co-localized with (A) EEA1 and (B) LAMP1 in both resting (grey bars) and activated (black bars) BECs. Graphs represent the mean of triplicate values ± SEM of n = 3 independent experiments. (C–G) Adherent BECs were loaded with Atto−633-labeled myelin and 24 hr later, co-localization of myelin (in red) with EEA1 (in green, upper panels) or LAMP1 (in green, lower panels) was analyzed using CSLM. Nuclei were visualized with Hoechst (in blue) and the cytoskeletal F-actin bundles are shown in yellow. Representative images of adherent brain endothelial cells with subcellular localization of myelin with EEA1 (C,E) or LAMP1 (D,F). A magnification of indicated areas is shown in E–F. A cross-sectional study focusing in an myelin-rich area demonstrates the presence of the antigen surrounded by LAMP1 staining, indicating its presence within lysosomes. (G–H) Histograms were created for a selected area (indicated by a line) using ImageJ software (NIH, USA). Histograms were created from each fluorochrome and overlays were made by the program. (I) Quantification of myelin positive early-endosomal and lysosomal compartments. Percentage of myelin fragments associated with each marker was determined using ImageJ software (N = 6).

**DOI:** http://dx.doi.org/10.7554/eLife.13149.006

Together, these data suggest that myelin enters the endosomal/lysosomal pathway when internalized by BECs. Furthermore, it is clear that the efficiency of myelin uptake and internal routing to compartments associated with antigen processing by BECs is not affected by inflammation per se. This is in contrast to their professional counterparts: only in an immature state DCs possess high antigen internalization and processing capacities. Activation induced maturation of DCs strongly reduces these functions, and increases the presentation of antigens in MHC molecules (Inaba et al., 2000; Jin et al., 2004).

Migration of myelin-specific T-cells depends on presentation of myelin-antigens in MHC-II by BECs

Taking the lack of human myelin-specific T-cell clones as well as HLA-matching issues with BECs into account, we used murine BECs (mBECs) and MOG_35-55-specific CD4⁺ T-cells from 2D2 transgenic mice (Bettelli et al., 2003) as a model system to elucidate whether myelin antigens are processed and presented by BECs to facilitate T-cell transmigration. Notably, an increased expression of MHC-II was observed on cerebral blood vessels of mice in the active phase of EAE when compared to control adjuvant injected mice (CFA; not shown), demonstrating that mBECs, similar to the human counterparts, are properly equipped to present antigens to pathogenic CD4⁺ T cells. Since in the brain of MS patients and of EAE mice mainly Th1 and Th17 effector cells have been found (Carbajal et al., 2015), we generated MOG-specific Th1 and Th17 in-vitro (Figure 3A) and used them in a trans-well setting with myelin-loaded activated mBECs. To allow sufficient antigen processing, mBECs were loaded with myelin in the presence of TNFα 24 hr prior the co-culture with T-cells. To control for antigen-specificity, mBECs were loaded with the non-CNS antigen ovalbumin (OVA). Loading of mBECs with OVA did not significantly induce the migration of any of the MOG-specific T-cell subsets, similar to medium-control mBECs (Figure 3B,C). However, when mBECs were loaded with myelin, a significant increase in migrated Th1 and Th17 cells was observed, demonstrating that processing and presentation of myelin-derived peptides by mBECs specifically leads to migration of antigen-specific T-cells. Addition of an MHC-II-blocking antibody during the migration period significantly reduced the trans-migration of both Th1 and Th17 cells (Figure 3D,E), further providing evidence that presentation of myelin-derived antigens in MHC-II by mBECs facilitates T-cell migration. Using the nominal epitope for 2D2 T-cells (i.e. MOG_35–55) to load mBECs with, similar results were obtained as with myelin-loaded mBECs (Figure 3F). Moreover, our observation that OVA-specific Th1 and Th17 only trans-migrated when encountering OVA-loaded BECs and not when co-cultured with MOG_35–55-loaded or medium control BECs substantiates the finding that T-cell migration over the BEC monolayer occurs in an antigen-specific manner (Figure 3G,H). These data demonstrate that brain endothelial cells can internalize antigen and promote antigen-specific T cell transmigration in vitro.

Figure 3. — (A) Th1 and Th17 subsets were generated in-vitro from naive CD4⁺CD62L^high 2D2 T cells. Expression of IFN-γ, Il−17, IL−10, T-bet and RORγT was determined using qRT-PCR. Data are the means of triplicate values ± SEM of three independent experiments. (B) mBECs were seeded onto trans-wells, activated with TNFα and loaded with myelin for 24 hr. As a control, BECs loaded with the CNS-unrelated antigen OVA or unloaded BECs were used. (B) Th1 or (C) Th17 2D2 T-cells were added to the upper compartment and T-cell migration was quantified by flow cytometry 3 hr later using fluorescent labelled beads as reference. To block antigen recognition by T-cells, an MHC-II blocking antibody was added to mBECs one hour prior addition of the (D) Th1 or (E) Th17 cells. The MHC-II blocking antibody was present during the 3 hr incubation with the T-cells. (F) Transmigration of Th1 and Th17 cells over a monolayer of MOG_35-55 pulse-loaded activated mBECs was analyzed using Transwells. Migration was assessed in the presence of an MHC-II blocking antibody or control antibody. *p<0.05, **p<0.01, ***p<0.001 (ANOVA with Bonferroni correction). The average frequency of T cells that transmigrated in the control setting are 10.8% ± 1.2 for 2D2 Th1 and 11.6% ± 0.4 for 2D2 Th17. (G,H) Th1 and Th17 subsets were generated in-vitro from naive CD4⁺CD62L^high OT-II T cells. mBECs were seeded onto trans-wells, activated with TNF and loaded with OVA for 24 hr. As a control, BECs loaded with MOG_35-55 or unloaded BECs were used. Th1 (G) or Th17 (H) OT-II T-cells were added to the upper compartment and T-cell migration was quantified by flow cytometry 3 hr later using fluorescent labelled beads as reference. Average frequency of OT-II Th1 and OT-II Th17 that transmigrated in the control settings are 7.9% ± 1.9 and 12.5% ± 1.4, respectively.**p<0.01, ***p<0.001 (ANOVA with Bonferroni correction).

**DOI:** http://dx.doi.org/10.7554/eLife.13149.007

Mice lacking a functional class-II restricted antigen processing machinery are resistant to both active and adoptive transfer EAE (Tompkins et al., 2002), suggesting that proper processing of antigens is essential for disease initiation. Although these results could be due to the lack of activation of auto-reactive T-cells by peripheral APCs, the failure to induce disease by adoptive transfer of activated T-cells in this study could also be explained by the lack of a functional antigen processing machinery in BECs since trafficking of injected, ex-vivo activated, T-cells into the brain is impaired. Thus, together with our novel data, it seems likely that antigen-presentation by the brain endothelium facilitates the entry of antigen-specific CD4⁺ effector T cells into the brain. This phenomenon has also been proposed, but was never demonstrated, in other diseases suffering from the infiltration and destruction of tissue by auto-reactive T-cells. In Type 1 Diabetes, pancreatic islet antigen expression was shown to be a key factor in governing the ability of the autoantigen-specific T-cells to accumulate in the pancreatic islets (Hamilton-Williams et al., 2003; Van Halteren et al., 2005). Importantly, using human T-cell clones and humanized mice, it was demonstrated that only beta cell-specific T-cells reached the pancreatic islets where they destroyed the insulin-producing beta cells. By contrast, diabetes unrelated T-cells retained at the peri-vascular sites (Unger et al., 2012), demonstrating that beta-cell-specific T-cells, present in the circulation, need to cross the endothelium to access the pancreatic islets. Whether antigen-specific T-cell entry favors the entry of non-tissue specific T-cells is still a matter of debate. The entry of encephalitogenic T-cells into the brain has been shown to pave the way for non-CNS-specific T-cells (Lees et al., 2010; Ludowyk et al., 1992) yet the latter subset remained in an inactive state.

Together, the data presented in this study demonstrate for the first time that myelin enters the endosomal/lysosomal pathway when internalized by BECs, irrespective of their activation status. This observation is also different to findings on professional APCs such as dendritic cells, which mainly internalize antigens being in an immature state. The fact that BECs maintain to internalize and process exogenous antigens in an activated state is advantageous during infection-induced inflammation in the brain (e.g. meningitis) as it will facilitate the presence of antigen-specific effector T-cells to resolve the unwanted infection. However, this continuous facilitation of immune cell entry into the CNS is destructive in case of MS.

Overall, our results demonstrate that BECs can take up and process myelin particles in a time-dependent manner. Although the focus of the present study was to examine whether antigen-presentation by BECs contributes to transmigration of myelin-reactive T cells, it can be speculated that uptake of myelin, consisting of large particles, by brain endothelial cells predominantly occurs via phagocytosis. BECs have been shown to use different endocytosis mechanisms to internalize particles, which is dependent on the size and composition of the particle (Faille et al., 2012; Falcone et al., 2006; Georgieva et al., 2011). Furthermore, the upregulation of MHC-II expression under inflammatory conditions reinforces the idea of a non-professional antigen presenting cell role. Although we do not provide direct evidence for processing and presentation of internalized myelin, our data strongly suggest that myelin-derived antigens can be presented by brain endothelial cells in MHC-II to antigen-specific T cell subsets, aiding in the diapedesis of these cells in an MHC-II dependent fashion. These results demonstrate that the brain endothelium is an active contributor to disease pathogenesis. Furthermore, these findings have major implications in neuro-inflammatory disorders such as MS, since increased immune cell trafficking has a detrimental effect in disease progression. Therapies directed at antigen processing and presentation by BECs could be effective to dampen unwanted immune cell infiltration in MS.

Materials and methods

Cell culture

The human brain endothelial cell (BEC) line hCMEC/D3 (Weksler et al., 2005) was kindly provided by Dr PO Couraud (Institut Cochin, Universite Paris Descartes, Paris, France). BECs were grown in EBM−2 medium supplemented with hEGF, hydrocortisone, GA-1000, FBS, VEGF, hFGF-B, R3-IGF-1, ascorbic acid and 2.5% fetal calf serum (Lonza, Basel, Switzerland).

Flow cytometry

For antigen internalization experiments, resting or 24 hr rhTNFα activated (5 ng/ml, Peprotech, UK) BECs were seeded in collagen-coated plates and when confluent, incubated with 10 µg/ml labeled myelin (myelin−555) for 4 hr or 24 hr. Subsequently, cells were extensively washed with PBS to remove external myelin and fluorescence intensity was measured using a FACS Calibur flow cytometer (Becton and Dickinson, San Jose, CA).

The following antibodies were used to detect the presence of MHC and costimulatory molecules on resting or TNFα activated BECs: FITC-conjugated anti-HLA-ABC (clone DX-17) and -VCAM−1 (clone STA); PE-conjugated anti-HLA-DR (clone G46-6); -CD80 (clone L307.4); -CD86 (clone 2331). Binding of unconjugated anti-CD40 (clone TRAP-1) was detected using goat-anti-mouse IgG1-A488 (Life Technologies). All antibodies were obtained from BD Pharmingen, except anti-VCAM which was obtained from eBiosciences.

Imaging flow cytometry

Confluent BECs were seeded in 6-well plates (Corning, Amsterdam, The Netherlands) and stimulated with 5 ng/ml rhTNFα for 24 hr. 10 µg/ml of fluorescent-labeled human myelin was added to BECs for 4 hr or 24 hr. Cells were then extensively washed with ice-cold PBS, detached with trypsin and fixated with 4% formaldehyde. Cells were then permeabilized with 0.05% saponin for 30 min at RT and subsequently blocked with 10% goat serum in PBS/BSA. Cells were labeled with EEA1-FITC (BD Bioscience), LAMP1 (BD Pharmingen) and goat anti-mouse Alexa 488 (Molecular Probes, Eugene, OR). Cells were analyzed on the ImageStream X100 (Amnis-Merck Millipore) imaging flow cytometer as previously described (García-Vallejo et al., 2015). A minimum of 15,000 cells were acquired per sample. Internalization and co-localization scores were calculated as previously described (García-Vallejo et al., 2015). Briefly, cells were acquired on the basis of their area. Analysis was performed with single cells after compensation (with a minimum of 5000 cells). For standard acquisition, the 488 nm laser line (for EEA-1 and LAMP-1) was set at 10 mW and the 642 nm laser line (for myelin) was set at 5 mW.

Firstly, a mask was designed based on the surface of BECs in the brightfield image. This mask was then eroded to exclude the cell membrane. Finally, the resulting mask was applied to the fluorescence channel. The internalization score was then calculated on this mask using the Internalization feature provided in the Ideas v6.0 software (Amnis-Merck Millipore). Internalization can be interpreted as a log-scaled ratio of the intensity of the intracellular space versus the intensity of the entire cell. Cells that have internalized antigen typically have positive scores, while cells that show the antigen still on the membrane have negative scores. Cells with scores around 0 have similar amounts of antigen on the membrane and in intracellular compartments. Co-localization is calculated using the bright detail similarity R3 feature in the Ideas software. This feature corresponds to the logarithmic transformation of Pearson’s correlation coefficient of the localized bright spots with a radius of 3 pixels or less within the whole cell area in the two input images. Myelin particle counts were calculated using the peak mask in combination with the spot count feature as previously described (García-Vallejo et al., 2014).

Confocal microscopy

Confluent BECs were seeded in 8-well Ibidi slides (Ibidi, GmbH, Munchen, Germany) and incubated with 10 µg/ml Atto 633 labeled myelin for 24 hr. Subsequently, cells were extensively washed with PBS and fixated with 4% formaldehyde. Non-specific binding was blocked with 5% goat serum in PBS/BSA containing 0.3% Triton-X100. Cells were labeled with rabbit anti-EEA1 (Cell Signaling) or rabbit anti-LAMP1 (Cell Signaling). Antibodies were visualized after 1 hr incubation with goat anti-rabbit Alexa488 (Molecular Probes). Finally, sections were stained with Hoechst (molecular Probes, Invitrogen) to visualize cellular nuclei and with phalloidin rhodamine to visualize F-actin (Molecular Probes, Invitrogen). Sections were mounted with mounting medium. Co-localization was analyzed using a Confocal Laser Scanning Microscope (Leica DMI 6000, SP8, Leica, Mannheim, Germany); images were acquired using LCS software (version 2.61, Leica).

Isolation and culture of primary murine BECs (mBECs)

Primary mBECs were isolated from brains of C57BL/6 mice as described previously (Coisne et al., 2005). Brains were harvested and superficial blood vessel, meninges and cerebellum were removed. Brains were homogenized in isolation medium (HBSS supplemented with 10 mM HEPES and 0.1% BSA) in a potter and centrifuged. The pellet was resuspended in 15% dextran (70 kDa) and centrifuged at 3000 g for 25 min. Subsequently, the pellet was resuspended in 0,2% collagenase/dispase with 10 μg DNase in culture medium (DMEM supplemented with 20% FCS, 1% amino acids, 2% sodium pyruvate and 50 µg/ml gentamycin) and incubated for 30 min in a 37°C waterbath. After washing, the obtained fragments of blood vessels were seeded in collagen-coated dishes in culture medium containing puromycin to avoid contamination with pericytes. After 24 hr of culture, medium was supplemented with 1 ng/ml FGF. At the end of culture, endothelial purity was checked by qPCR for CD31 (endothelial), GFAP (astrocytes), and PDGF-receptor beta (pericytes) as described before (Reijerkerk et al., 2013) and cultures were found to be consisting of 95% endothelial cells.

Generation of MOG-specific T cell subsets

Single cells suspensions of spleens and lymph nodes from 2D2 Tg mice (generous gift from L. Berod, TWINCORE Institute, Hannover, Germany) were depleted of erythrocytes using ACK lysis buffer. Subsequently, CD4⁺ T-cells were enriched using the mouse CD4⁺ T-cell enrichment kit (eBiosciences) according to manufacturer’s instructions; stained with anti-CD4-PE and CD62L-APC antibodies and naive CD4⁺CD62L^high T cells were sorted using a MoFlow (DakoCytomation, Glostrup, Denmark). Naive T-cells (5 × 10⁴) were incubated with MOG_35–55/LPS loaded BMDCs (1 × 10⁴) to promote Th1 differentiation. Incubation of naive CD4⁺ T-cells with MOG-loaded BMDCs in the presence of PGN (10 µg/ml) promoted Th17 differentiation. Two days later, 10 U/ml rmIL−2 (Invitrogen, Bleijswijk, The Netherlands) was added to the Th1 promoting cultures and another three days later T-cells were harvested and used in functional assays.

Quantitative PCR

Messenger RNA was isolated from mBECs using the TRIzol method (Life Technologies, Bleiswijk, the Netherlands) and cDNA was synthesized with the Reverse Transcription System kit (Promega, Leiden, the Netherlands). The following primer sequences were used: IFN-γ FWD: TACTACCTTCTTCAGCAACAGC, IFN-γ REV: AATCAGCAGCGACTCCTTTTC, IL−10-FWD: GGCGCTGTCATCGATTTCTC; IL−10 REV: ATGGCCTTGTAGACACCTTGG, T-bet FWD: CAGGGAACCGCTTATATG, T-bet REV: CTGGCTCTCCATCATTCA, RORγT FWD: GGAGCAGAGCTTAAACCCCC; RORγT REV: TCCCAGATGACTTGTCCCCA, GAPDH FWD: GACAACTCATCAAGATTGTCAGCA; GAPDH REV: TTCATGAGCCCTTCCACAATG. Oligonucleotides were synthesized by Invitrogen (Bleiswijk, the Netherlands). Quantitative PCR (qPCR) reactions were performed in an ABI7900HT sequence detection system using the SYBR Green method (Applied Biosystems, New York, USA). Expression levels were normalized to GAPDH expression levels.

Transwell migration

Ex-vivo isolated mBECs were seeded on collagen-coated 5 µm pore size Costar transwells (Corning, Amsterdam, The Netherlands) for 5–7 days. mBECs were loaded with 72.5 µg/ml myelin, 10 µg/ml MOG_35–55 or 10 µg/ml OVA in the presence of 25 ng/ml TNFα for 24 hr. Cells were thoroughly washed and 1 × 10⁵ Th1 or Th17 were added per transwell. Anti-mouse MHC-II blocking antibody (#16-5321-81, eBioscience) was added at 5 µg/ml per transwell, 1 hr prior to addition of T cells. After 3 hr T-cells were recovered from the lower well and 20,000 beads (Beckman Coulter, USA) were added to each sample. Samples were analyzed by flow cytometry on a FACScalibur (BD, San Jose, USA) and by gating and counting 5000 beads, the number of migrated T-cells was determined.

Statistical analysis

Statistical analysis was performed using GraphPad Prism software (v5.01 GraphPad Software, La Jolla, CA) using either unpaired Student t test or one-way ANOVA followed by posthoc Bonferroni correction.

Acknowledgements

We thank our biotechnicians for excellent care-taking of the animals. This work was supported by the European Research Council (ERCAdvanced339977; WWJU) and the MS research foundation (MS-09-358d, MLP; MS-14-358e, AK).

Funding Statement

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Funding Information

This paper was supported by the following grants:

European Research Council ERCAdvanced339977 to Wendy WJ Unger.
Stichting MS Research MS-09-358d to Melissa A Lopes Pinheiro.
Stichting MS Research MS-14-358e to Alwin Kamermans.

Additional information

Competing interests

The authors declare that no competing interests exist.

Author contributions

MALP, Designed and performed experiments, Analyzed and interpreted data, Wrote paper, Conception and design, Acquisition of data, Drafting or revising the article.

AK, Performed experiments, Analyzed and interpreted data, Acquisition of data.

JJG-V, Designed and performed imaging flow cytometric analysis, Analyzed and interpreted data, Conception and design, Acquisition of data.

BvhH, Performed experiments, Analyzed data, Acquisition of data.

LW, Performed experiments, Analyzed data, Acquisition of data.

TO, Performed flow cytometric cell sorting, Analyzed and interpreted data, Acquisition of data.

DB, Performed experiments, Analyzed and interpreted data, Acquisition of data.

MV, Performed experiments, Analyzed and interpreted data, Acquisition of data, Contributed unpublished essential data or reagents.

SMAvdP, Acquisition of data.

YvK, Designed experiments, Interpreted data, Analysis and interpretation of data.

HEdV, Designed experiments, Interpreted data, Wrote paper, Supervised the study, Conception and design, Drafting or revising the article.

WWJU, Designed experiments, Interpreted data, Wrote paper, Supervised the study, Conception and design, Acquisition of data, Analysis and interpretation of data, Drafting or revising the article.

References

Archambault AS, Sim J, Gimenez MA, Russell JH. Defining antigen-dependent stages of T cell migration from the blood to the central nervous system parenchyma. European Journal of Immunology. 2005;35:1076–1085. doi: 10.1002/eji.200425864. [DOI] [PubMed] [Google Scholar]
Bailey SL, Schreiner B, McMahon EJ, Miller SD. CNS myeloid DCs presenting endogenous myelin peptides 'preferentially' polarize CD4+ T(H)-17 cells in relapsing EAE. Nature Immunology. 2007;8:172–180. doi: 10.1038/ni1430. [DOI] [PubMed] [Google Scholar]
Bettelli E, Pagany M, Weiner HL, Linington C, Sobel RA, Kuchroo VK. Myelin oligodendrocyte glycoprotein-specific T cell receptor transgenic mice develop spontaneous autoimmune optic neuritis. Journal of Experimental Medicine. 2003;197:1073–1081. doi: 10.1084/jem.20021603. [DOI] [PMC free article] [PubMed] [Google Scholar]
Bielekova B, Goodwin B, Richert N, Cortese I, Kondo T, Afshar G, Gran B, Eaton J, Antel J, Frank JA, McFarland HF, Martin R. Encephalitogenic potential of the myelin basic protein peptide (amino acids 83-99) in multiple sclerosis: results of a phase II clinical trial with an altered peptide ligand. Nature Medicine. 2000;6:1167–1175. doi: 10.1038/80516. [DOI] [PubMed] [Google Scholar]
Blum JS, Wearsch PA, Cresswell P. Pathways of antigen processing. Annual Review of Immunology. 2013;31:443–473. doi: 10.1146/annurev-immunol-032712-095910. [DOI] [PMC free article] [PubMed] [Google Scholar]
Carbajal KS, Mironova Y, Ulrich-Lewis JT, Kulkarni D, Grifka-Walk HM, Huber AK, Shrager P, Giger RJ, Segal BM. Th cell diversity in experimental autoimmune encephalomyelitis and multiple sclerosis. Journal of Immunology. 2015;195:2552–2559. doi: 10.4049/jimmunol.1501097. [DOI] [PMC free article] [PubMed] [Google Scholar]
Coisne C, Dehouck L, Faveeuw C, Delplace Y, Miller F, Landry C, Morissette C, Fenart L, Cecchelli R, Tremblay P, Dehouck B. Mouse syngenic in vitro blood-brain barrier model: a new tool to examine inflammatory events in cerebral endothelium. Laboratory Investigation. 2005;85:734–746. doi: 10.1038/labinvest.3700281. [DOI] [PubMed] [Google Scholar]
Faille D, El-Assaad F, Mitchell AJ, Alessi MC, Chimini G, Fusai T, Grau GE, Combes V. Endocytosis and intracellular processing of platelet microparticles by brain endothelial cells. Journal of Cellular and Molecular Medicine. 2012;16:1731–1738. doi: 10.1111/j.1582-4934.2011.01434.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
Falcone S, Cocucci E, Podini P, Kirchhausen T, Clementi E, Meldolesi J. Macropinocytosis: regulated coordination of endocytic and exocytic membrane traffic events. Journal of Cell Science. 2006;119:4758–4769. doi: 10.1242/jcs.03238. [DOI] [PubMed] [Google Scholar]
Galea I, Bernardes-Silva M, Forse PA, van Rooijen N, Liblau RS, Perry VH. An antigen-specific pathway for CD8 T cells across the blood-brain barrier. Journal of Experimental Medicine. 2007;204:2023–2030. doi: 10.1084/jem.20070064. [DOI] [PMC free article] [PubMed] [Google Scholar]
García-Vallejo JJ, Bloem K, Knippels LM, Garssen J, van Vliet SJ, van Kooyk Y. The Consequences of Multiple Simultaneous C-Type Lectin-Ligand Interactions: DCIR Alters the Endo-Lysosomal Routing of DC-SIGN. Frontiers in Immunology. 2015;6:e13149. doi: 10.3389/fimmu.2015.00087. [DOI] [PMC free article] [PubMed] [Google Scholar]
García-Vallejo JJ, Ilarregui JM, Kalay H, Chamorro S, Koning N, Unger WW, Ambrosini M, Montserrat V, Fernandes RJ, Bruijns SC, van Weering JR, Paauw NJ, O'Toole T, van Horssen J, van der Valk P, Nazmi K, Bolscher JG, Bajramovic J, Dijkstra CD, 't Hart BA, van Kooyk Y. CNS myelin induces regulatory functions of DC-SIGN-expressing, antigen-presenting cells via cognate interaction with MOG. Journal of Experimental Medicine. 2014;211:1465–1483. doi: 10.1084/jem.20122192. [DOI] [PMC free article] [PubMed] [Google Scholar]
Georgieva JV, Kalicharan D, Couraud PO, Romero IA, Weksler B, Hoekstra D, Zuhorn IS. Surface characteristics of nanoparticles determine their intracellular fate in and processing by human blood-brain barrier endothelial cells in vitro. Molecular Therapy. 2011;19:318–325. doi: 10.1038/mt.2010.236. [DOI] [PMC free article] [PubMed] [Google Scholar]
Hamilton-Williams EE, Palmer SE, Charlton B, Slattery RM. Beta cell MHC class I is a late requirement for diabetes. PNAS. 2003;100:6688–6693. doi: 10.1073/pnas.1131954100. [DOI] [PMC free article] [PubMed] [Google Scholar]
Inaba K, Turley S, Iyoda T, Yamaide F, Shimoyama S, Reis e Sousa C, Germain RN, Mellman I, Steinman RM. The formation of immunogenic major histocompatibility complex class II-peptide ligands in lysosomal compartments of dendritic cells is regulated by inflammatory stimuli. Journal of Experimental Medicine. 2000;191:927–936. doi: 10.1084/jem.191.6.927. [DOI] [PMC free article] [PubMed] [Google Scholar]
Jin Y, Fuller L, Ciancio G, Burke GW, Tzakis AG, Ricordi C, Miller J, Esquenzai V. Antigen presentation and immune regulatory capacity of immature and mature-enriched antigen presenting (dendritic) cells derived from human bone marrow. Human Immunology. 2004;65:93–103. doi: 10.1016/j.humimm.2003.11.002. [DOI] [PubMed] [Google Scholar]
Khaw LT, Ball HJ, Golenser J, Combes V, Grau GE, Wheway J, Mitchell AJ, Hunt NH, Obeid S, Couraud PO. Endothelial cells potentiate interferon-γ production in a novel tripartite culture model of human cerebral malaria. PLoS One. 2013;8:e13149. doi: 10.1371/journal.pone.0069521. [DOI] [PMC free article] [PubMed] [Google Scholar]
Larochelle C, Alvarez JI, Prat A. How do immune cells overcome the blood-brain barrier in multiple sclerosis? FEBS Letters. 2011;585:3770–3780. doi: 10.1016/j.febslet.2011.04.066. [DOI] [PubMed] [Google Scholar]
Lees JR, Sim J, Russell JH. Encephalitogenic T-cells increase numbers of CNS T-cells regardless of antigen specificity by both increasing T-cell entry and preventing egress. Journal of Neuroimmunology. 2010;220:10–16. doi: 10.1016/j.jneuroim.2009.11.017. [DOI] [PubMed] [Google Scholar]
Lublin FD, Cutter G, Giovannoni G, Pace A, Campbell NR, Belachew S. Natalizumab reduces relapse clinical severity and improves relapse recovery in MS. Multiple Sclerosis and Related Disorders. 2014;3:705–711. doi: 10.1016/j.msard.2014.08.005. [DOI] [PubMed] [Google Scholar]
Ludowyk PA, Willenborg DO, Parish CR. Selective localisation of neuro-specific T lymphocytes in the central nervous system. Journal of Neuroimmunology. 1992;37:237–250. doi: 10.1016/0165-5728(92)90008-9. [DOI] [PubMed] [Google Scholar]
Ransohoff RM, Kivisäkk P, Kidd G. Three or more routes for leukocyte migration into the central nervous system. Nature Reviews Immunology. 2003;3:569–581. doi: 10.1038/nri1130. [DOI] [PubMed] [Google Scholar]
Reijerkerk A, Lopez-Ramirez MA, van Het Hof B, Drexhage JA, Kamphuis WW, Kooij G, Vos JB, van der Pouw Kraan TC, van Zonneveld AJ, Horrevoets AJ, Prat A, Romero IA, de Vries HE. MicroRNAs regulate human brain endothelial cell-barrier function in inflammation: implications for multiple sclerosis. Journal of Neuroscience. 2013;33:6857–6863. doi: 10.1523/JNEUROSCI.3965-12.2013. [DOI] [PMC free article] [PubMed] [Google Scholar]
Reith W, LeibundGut-Landmann S, Waldburger JM. Regulation of MHC class II gene expression by the class II transactivator. Nature Reviews Immunology. 2005;5:793–806. doi: 10.1038/nri1708. [DOI] [PubMed] [Google Scholar]
Roche PA, Furuta K. The ins and outs of MHC class II-mediated antigen processing and presentation. Nature Reviews Immunology. 2015;15:203–216. doi: 10.1038/nri3818. [DOI] [PMC free article] [PubMed] [Google Scholar]
Sospedra M, Martin R. Immunology of multiple sclerosis. Annual Review of Immunology. 2005;23:683–747. doi: 10.1146/annurev.immunol.23.021704.115707. [DOI] [PubMed] [Google Scholar]
Tietz S, Engelhardt B. Brain barriers: Crosstalk between complex tight junctions and adherens junctions. Journal of Cell Biology. 2015;209:493–506. doi: 10.1083/jcb.201412147. [DOI] [PMC free article] [PubMed] [Google Scholar]
Tompkins SM, Padilla J, Dal Canto MC, Ting JP, Van Kaer L, Miller SD. De novo central nervous system processing of myelin antigen is required for the initiation of experimental autoimmune encephalomyelitis. Journal of Immunology. 2002;168:4173–4183. doi: 10.4049/jimmunol.168.8.4173. [DOI] [PubMed] [Google Scholar]
Trapp BD, Peterson J, Ransohoff RM, Rudick R, Mörk S, Bö L. Axonal transection in the lesions of multiple sclerosis. New England Journal of Medicine. 1998;338:278–285. doi: 10.1056/NEJM199801293380502. [DOI] [PubMed] [Google Scholar]
Unger WW, Pearson T, Abreu JR, Laban S, van der Slik AR, der Kracht SM, Kester MG, Serreze DV, Shultz LD, Griffioen M, Drijfhout JW, Greiner DL, Roep BO, Mulder-van der Kracht MG, Serreze LD. Islet-specific CTL cloned from a type 1 diabetes patient cause beta-cell destruction after engraftment into HLA-A2 transgenic NOD/scid/IL2RG null mice. PLoS One. 2012;7:e13149. doi: 10.1371/journal.pone.0049213. [DOI] [PMC free article] [PubMed] [Google Scholar]
van Halteren AG, Kardol MJ, Mulder A, Roep BO. Homing of human autoreactive T cells into pancreatic tissue of NOD-scid mice. Diabetologia. 2005;48:75–82. doi: 10.1007/s00125-004-1613-2. [DOI] [PubMed] [Google Scholar]
Weksler BB, Subileau EA, Perrière N, Charneau P, Holloway K, Leveque M, Tricoire-Leignel H, Nicotra A, Bourdoulous S, Turowski P, Male DK, Roux F, Greenwood J, Romero IA, Couraud PO. Blood-brain barrier-specific properties of a human adult brain endothelial cell line. FASEB Journal. 2005;19:1872–1874. doi: 10.1096/fj.04-3458fje. [DOI] [PubMed] [Google Scholar]
Weller RO, Engelhardt B, Phillips MJ. Lymphocyte targeting of the central nervous system: a review of afferent and efferent CNS-immune pathways. Brain Pathology. 1996;6:275–288. doi: 10.1111/j.1750-3639.1996.tb00855.x. [DOI] [PubMed] [Google Scholar]

eLife. 2016 Jun 23;5:e13149. doi: 10.7554/eLife.13149.016

Decision letter

Editor: Gary L Westbrook¹

In the interests of transparency, eLife includes the editorial decision letter and accompanying author responses. A lightly edited version of the letter sent to the authors after peer review is shown, indicating the most substantive concerns; minor comments are not usually included.

Thank you for submitting your work entitled "Internalization and presentation of myelin antigens by the brain endothelium guides antigen-specific T cell migration" for consideration by eLife. Your article has been reviewed by two peer reviewers, and the evaluation has been overseen by Gary Westbrook as the Senior Editor. The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission. The following individual involved in review of your submission has agreed to reveal his identity: Christopher Carman (peer reviewer).

Summary of essential revisions:

Both reviewers thought this was an important and clinically-relevant topic that has been discussed for some time, but without convincing direct evidence. Both reviewers thought that additional information is required to make the conclusions convincing, but that we are willing to consider a revised manuscript if the critiques can be addressed. The key point in Reviewer 1's review concerns Figure 4. Please also consider the points raised re Figure 1,2 and given that the results are in vitro, the conclusions in the Discussion should be tempered. The two points in Review 2 may require additional experiments that seem reasonable in scope and potentially possible a 2-month timeframe. In consultation between the editor and the reviewers, we further discussed some of these points that we urge you to address in your revisions:

Author response image 1. — The human brain endothelial cells (cell line hCMEC/D3) were cultured on Ibidi-slides. After reaching confluence, cells were washed, fixated with 4% formaldehyde and stained with FITC-labeled antibodies specific for A. the tight junction protein Claudin-5 (green). Nuclei were visualized using Hoechst (blue) or B. the adherens junction protein VE-Cadherin (green) and subsequently cells were analyzed by confocal laser scanning microscopy. From the images it is clear that both Claudin-5 and VE-Cadherin are highly present at cell-cell contacts, indicative of tight barrier formation by the brain endothelial cells.

**DOI:** http://dx.doi.org/10.7554/eLife.13149.008

1) Point 1 of reviewer 2: The issue that the internalization of myelin observed by fluorescence could be due to internalization of membrane as a consequence of cell rounding when they come off their substratum. This issue should really be ruled out by either rigorous co-localization experiments and/or by a control using a fluorescent molecule that sticks to plasma membrane but is not processed by MHCII.

2) Point 2 of reviewer 2: The reviewers were surprised that MHCII is upregulated in ECs by TNFa. Many studies have shown that TNFa increases expression of ICAM-1 and VCAM-1, but not MHCII. In contrast, IFNγ is universally accepted as the major cytokine elevating MHCII in EC. A side-by-side comparison of the two would be instructive and thus the data may be more robust.

3) There is a need to examine internalization on adherent cells and specifically under confluent settings.

Reviewer #1:

Lopes Pinheiro, et al. postulate that endothelial cells (EC) of the blood brain barrier (BBB) present antigen to T cells, which promotes their migration into the brain. They study myelin antigens and EC derived from the CNS in a series of mostly in vitro studies. The authors present data to show that EC from the CNS pulsed with myelin, but not ovalbumin, stimulate the transmigration of myelin-specific 2D2 T cells in Transwell assays. The stimulated transmigration is blocked by antibodies against MHCII. The interpretation is that EC internalize antigen and present it on MHCII to stimulate the antigen-specific T cells. This is an interesting idea, although the concept of antigen presentation by MHCII-expressing endothelial cells has been studied in the setting of transplantation for decades by Jordan Pober's group, and the idea of antigen presentation by BBB EC has been studied, as well, although most of the authors concluded that confluent BBB EC could not successfully present antigen by themselves. (See below). As the data in this manuscript stand, they are consistent with the premise that BBB EC take up myelin, process it in lysosomes, present it to cognate T cells, and stimulate antigen-specific T cell migration. However, the data fall short of direct proof. I do not ask for more experiments. However, I believe better versions of the experiments they have performed are required to prove their point.

1) Figure 1. Could the authors show that the monolayers are indeed confluent? The BBB endothelium is supposed to show very low levels of endocytic activity. How are the antigens taken up and processed? Indeed, several reports show that stimulation of T cells by MHCII expressing BBB EC is far more efficient when the EC are subconfluent than when they are confluent (e.g. J Neuroimmunol. 1995 Sep;61(2):231-9) and that expression of high levels of MHCII by IFNgamma are required for optimum stimulation (J Neuroimmunol. 1999 Jan 1;93(1-2):81-91 and J Neuropathol Exp Neurol. 2000 Feb;59(2):129-36). In contrast, uptake of myelin (Figure 1) and delivery to lysosomes (Figure 2) is said to be independent of cytokine activation in this study. The authors should discuss this.

2) Figure 2. The studies demonstrating delivery to endosomes and lysosomes are difficult to see. Moreover, either the authors have not explained their FRET assay correctly or are not using FRET correctly. In any case, what would be required to show that the EC are processing antigen correctly would be to show that the proper peptides are expressed on MHCII on the surface of these EC in a functional manner. This would ideally require demonstrating that the peptides are expressed on the MHC. This is difficult. I would settle for showing that they can stimulate the antigen-specific proliferation of antigen-specific T cells, not just stimulate migration in a transwell.

3) Figure 3 shows that mouse brain EC in EAE are Class II positive. This has been shown many times before, as far back as 1984 (J Immunol. 1984 May;132(5):2402-7). This figure is nice, but unnecessary, since all the definitive experiments are in vitro.

4) Figure 4 shows that Th1 and Th17 cells generated from MOG-specific 2D2 TCR transgenic mice migrate better across murine EC isolated from the CNS and incubated with myelin than they do across the same EC incubated with ovalbumin (OVA). This does show some biological relevance and is blocked by anti-MHCII antibody. However, the assumption is that the increased migration is due to presentation of cognate antigen by the EC, and this is never demonstrated (see Figure 2). Moreover, the proper control would be to show that either OVA is processed and presented to the same extent by these EC as is myelin and/or to show that OVA-specific T cells would transmigrate the EC pulsed with OVA, but not those pulsed with myelin. The T cells added to the assay have been activated by dendritic cells and expanded with IL-2. They are highly activated. My understanding is that the antigen-specific T cells that would come into the CNS in response to myelin antigens presented by the BBB EC in multiple sclerosis or EAE would be memory cells but not activated at the time but become activated upon presentation of their cognate antigen by the endothelial cells. What happens if the T cells are allowed to rest before being added to the Transwells?

Reviewer #2:

This study investigates an important problem with significant clinical implications. Specifically, the authors examine the underlying basis for auto-reactive myelin-specific inflammatory T cell migration into the brain, a critical event in multiple sclerosis (MS) pathogenesis. Vascular endothelial cells have long been recognized as a type of non-professional antigen presenting cell that may be able to present peptide antigens via MHCII to influence antigen-specific responses in effector/memory CD4⁺ T cells. Studies in MS, and murine models of MS, have suggested that the brain endothelial cells (BEC), which T cells must migrate across in order to enter the brain, may act to selectively recruit myelin-reactive T cells through their ability to present processed myelin peptide via MHCII. However, existing studies have yet to provide direct evidence for such functions. In the current studies the authors demonstrate in vitro, through flow cytometric methods, that BEC are able to constitutively uptake myelin protein where they are targeted to classical MHCII peptide processing compartments (i.e., endosomes and lysosomes). They further show that BEC upregulated MHCII and the co-stimulatory molecule CD40 in response to the inflammatory cytokine TNF-α. The authors go on to demonstrate for the first time that such intact myelin is processes and functionally presented to Th1 and Th17 cells in order to promote their migration across the endothelium in an MHCII-dependent manner. These studies provide new unambiguous evidence that endothelial cells play active roles as non-professional APC in MS pathogenesis and point toward new therapeutic strategies for this disease. This study is generally well performed and presented. However, a couple of important issues/questions remain to be addressed.

1) While the studies to investigate myelin internalization are reasonably well preformed including use of a FRET-based assay to co-localized myelin with endosomal and lysosomal markers. These all use flow cytometric methods (including imaging-based Amnis imaging cytometry). There is some concern that the transition from highly flattened adherent cells to trypsinized cells in suspension (i.e., in order to perform flow cytometry) will almost certainly drive some degree of surface membrane internalization to accommodate the massive reduction in the cell-surface area to volume ratio, that may influence the degree of observed myelin internalization. Thus, this analysis would be strengthened if it also included complementary imaging studies of adherent endothelial cells (e.g., including myelin, EEA1 and/or LAMP1 with Pearson's co-localization analysis).

2) Additionally, the authors show that TNF-α promotes MHCII upregulation on mouse BEC. However, a multitude of studies have shown in human, murine and other endothelial cell types the interferon-γ is the principle driver of endothelial MHCII expression and that TNF-α generally promotes adhesion molecule expression without induction of MHCII. The authors should comment on this as well, ideally as include analysis using interferon-γ.

[Editors' note: further revisions were requested prior to acceptance, as described below.]

Thank you for resubmitting your work entitled "Internalization and presentation of myelin antigens by the brain endothelium guides antigen-specific T cell migration" for further consideration at eLife. Your revised article has been favorably evaluated by Gary Westbrook (Senior editor) and two reviewers. The manuscript has been improved but there are some remaining issues that we would like you to address as outlined below before we can make a final determination of acceptance.

Summary:

There were still some concerns about possible limitations of the conclusions. The most important experiment to justify the conclusions has now been included, i.e. the specificity of their transwell assay for the antigen (Figure 3). The authors have shown antigen-specific stimulation of T cell transmigration in vitro across endothelial cells derived from brain in a manner that can be blocked by anti-MHCII antibody. However, the evidence is less convincing that the antigen has been internalized and presented by the endothelial cells, although some of the data are consistent with that hypothesis. A balanced discussion of this issue is important. Please also address these issues:

1) The authors seem to have misunderstood the point of reviewer #2 about their internalization assay. Masking out the surface fluorescence has nothing to do with it. The concern is that when cells go from being flat to round (in this case upon trypsinization) there could be a decrease in cell surface area if membrane is rapidly internalized as vesicles. If these portions of plasma membrane had myelin fragments attached, they would appear as "internalized" even though they were only internalized as part of the procedure. The experiment with flat cells is more convincing. But then, these are not confluent endothelial cells. The was the point 1 of reviewer 2 in the original submission. This concern should be addressed with discussion or with additional data if necessary.

2) One of the reviewers regarded the description of FRET throughout the manuscript and figure as incorrect. The issue was what is described as "FRET" is really illumination of a red fluor by 488nm light, probably as a result of band pass filters. True FRET would be a decrease in green signal due to energy transfer to the red fluor. The authors should either perform this technique correctly or omit this part.

3) The images of "confluent" endothelial cells do show some areas of true confluence with cell borders marked by claudin-5 or VE-cadherin. However, at least half of the monolayers do not show these markers, so they are not truly confluent. Please discuss this issue.

[Editors' note: further revisions were requested prior to acceptance, as described below.]

Thank you for resubmitting your work entitled "Internalization and presentation of myelin antigens by the brain endothelium guides antigen-specific T cell migration" for further consideration at eLife. Your revised article has been favorably evaluated by Gary Westbrook (Senior editor) and one of the original reviewers. The manuscript has been improved but there are some remaining issues that need to be addressed as outlined below:

1) Figure 2 now shows internalization and co-localization of myelin with endosomal and lysosomal markers in adherent cells. However, the co-localization shown with LAMP-1 could also be interpreted as a cluster of LAMP-1 positive lysosomes surrounding a myelin-containing endosome. Overlap of the fluorescence should appear yellow. Their images seem to be green blobs surrounding a red one. Because lysosomal diameters are much closer to 1 μm than the 3 μm shown in the scale, the authors should really show an image that shows true overlap of signals rather than what appear in Figure 2H as three separate peaks. Moreover, no quantification of these data (2C-H) are presented, just one set of micrographs for each. Readers need to know what percentage of total myelin fragments were internalized and associated with each marker. The authors should be able to obtain these numbers from their existing data.

2) In Figure 3 the results for transmigration are presented as% of control. The authors should state at least in the figure legend what% of added T cells transmigrated in the control. That will show how robust the assay is.

3) In the subsection “Migration of myelin-specific T-cells depends on presentation of myelin-antigens in MHC-II by BECs”, end of first paragraph: The authors really don't demonstrate that the phenomena of antigen presentation and transmigration are linked, as they admit later in the Discussion. Please reword this sentence to say that the data e.g. "demonstrate that brain endothelial cells can internalize antigen and promote antigen-specific T cell transmigration in vitro".

eLife. 2016 Jun 23;5:e13149. doi: 10.7554/eLife.13149.017

Author response

Reviewer #1:

A) The reviewer asks whether the monolayers are indeed confluent (concerning data presented in Figure 1 and 2).

All experiments in this study were performed upon confluency of the brain endothelial monolayers. Within the research team, we have ample knowledge on blood-brain-barrier function studied in confluent monolayers, which mimic blood vessels’ physiological characteristics in vitro. We have extensively tested their characteristics in functional assays as well as performed confocal microscopy analysis on confluent monolayers, as published in van Doorn et al., J Neuroinflam 2012 (doi:10.1186/1742-2094-9-133) and Mizee et al., J Neuroscience 2013 (doi: 10.1523/JNEUROSCI.1338-12.2013). Please see an example of confluent brain endothelial cells as analyzed by confocal microscopy in Author response image 1. The brain endothelial cell line hCMEC/D3 we used in our assays is described in Weksler et al., FASEB 2005 (doi:10.1096/fj.04-3458fje).

B) How are the antigens taken up and processed?

As shown by many studies, brain endothelial cells use different endocytosis mechanisms to internalize particles: phagocytosis; micropinocytosis and receptor-mediated endocytosis (Faille et al., J Cell Mol Med, 2012; Georgieva et al., Mol therapy 2011; Falcone et al., J. Cell Science 2006). It can be speculated that uptake of myelin, consisting of large particles, by brain endothelial cells predominantly occurs via phagocytosis. However, the aim of our study was to determine whether antigen-presentation by brain endothelial cells contributes to transmigration of myelin-reactive T cells, and not to study the mechanisms used to internalize antigens in detail. We have added this part to the Results & Discussion section (subsection “Migration of myelin-specific T-cells depends on presentation of myelin-antigens in MHC-II by BECs”, last paragraph).

C) The reviewer asks to discuss an apparent discrepancy between the cytokine-dependent increase of MHC-II expression and optimum stimulation of T cells and the cytokine-independent uptake of myelin (Figure 1) and delivery to lysosomes (Figure 2).

Internalization and intracellular routing of antigen are different processes than T cell stimulation and differently influenced by cytokines. Uptake, delivery to lysosomes and subsequent presentation of antigens in MHC molecules also occurs on brain endothelial cells under homeostatic conditions and independent from the presence of inflammatory cytokines (see for example in “Chapter 6: Barriers of the CNS” in Cerebral Circulation, Cipolla MJ., Morgan & Claypool Lifesciences, 2009). However, during inflammation, the processing as well as the presentation of myelin-derived epitopes is increased, as cytokine-mediated activation of the brain endothelial cells results in enhanced expression of MHC-II molecules. Together with enhanced expression of costimulatory molecules (e.g. CD40), this will lead to enhanced T cell activation.

We apologize to the reviewer for being unclear on how the FRET assay was executed. We have adapted this part of the Materials and methods section with the following:

“Fluorescence resonance energy transfer (FRET) was used as a measure of transfer of energy between the EEA1- or LAMP1-conjugated FITC fluorochromes with the myelin-associated Atto 633 fluorochrome. […] This phenomenon could only be possible if the emitted photons from FITC had excited the atto633-labeled myelin, demonstrating that the proximity of the EEA1/LAMP1 and myelin fluorochromes was close enough to allow FRET.”

Furthermore, the specific images in Figure 2 have been enlarged and the controls are shown as supplementary figure (Figure 2—figure supplement 1) in the manuscript.

In any case, what would be required to show that the EC are processing antigen correctly would be to show that the proper peptides are expressed on MHCII on the surface of these EC in a functional manner. This would ideally require demonstrating that the peptides are expressed on the MHC. This is difficult. I would settle for showing that they can stimulate the antigen-specific proliferation of antigen-specific T cells, not just stimulate migration in a transwell.

We have tested this possibility in co-cultures of in-vitro polarized MOG-specific Th1 or Th17 cells and BECs that were loaded with MOG or myelin in the presence of TNFα. TNFα-activated unloaded BECs were used as controls. When measuring proliferation of the T cells 3 days later, it was clear that the antigen-loaded BECs did not induce proliferation of Th1 nor of Th17 cells. This was found irrespective of using 2D2 or OT-II tg T cells (see Author response image 2A showing proliferation of 2D2 Th1 and Th17 cells). However, we hypothesize that this function is likely provided by other cells in the perivascular space, such as the perivascular APCs (Greter et al., Nat Med 2005; Kivisäkk et al., Annals of Neurol 2008; Hickey and Kimura, Science 1988. However, analysis of the activation status of the T cells in these co-cultures revealed that the activation marker CD25 was only expressed when T cells encountered antigen-loaded BECs (see Author response image 2B). Since induction of CD25 expression is a consequence of TCR signaling (Szamel et al., JI 1998), these data suggest that the epitopes are expressed in MHCII on the surface of BECs in a functional manner.

Author response image 2. — (A) Brain endothelial cells were loaded with myelin or MOG35-55 in the presence of TNF and subsequently co-cultured with in vitro generated myelin-specific Th1 or Th17 cells. As a control, none antigen loaded TNF activated brain endothelial cells were used. Proliferation of T cells was determined by incorporation of 3H-Thymidine, which was present during the last 18h of a three day culture period. (B) Activation of T cells was determined by analyzing expression of CD25 after 24 hr co-culture with brain endothelial cells that were pulsed with antigen in the presence or absence of TNF prior. Controls included brain endothelial cells that were treated with TNF or medium. Expression of CD25 on T cells was determined using flow cytometry.

**DOI:** http://dx.doi.org/10.7554/eLife.13149.009

We agree with the reviewer and have removed the figure from the manuscript.

We agree with the reviewer and performed these experiments. We observed the same phenomenon when using antigen-loaded BECs and OT-II T cells as for the 2D2 T cells: OVA-specific T cells only transmigrated when encountering OVA-loaded BECs and did not when cultured with MOG-loaded or none antigen-loaded BECs (see Author response image 3). These data have been added to the manuscript (Results & Discussion section, subsection “Migration of myelin-specific T-cells depends on presentation of myelin-antigens in MHC-II by BECs”, first paragraph and as Figure 3—figure supplement 1).

Author response image 3. — Th1 and Th17 subsets were generated in-vitro from naive CD4⁺CD62L^high OT-II T cells. mBECs were seeded onto trans-wells, activated with TNF and loaded with OVA for 24h. As a control, BECs loaded with MOG_35-55 or unloaded BECs were used. Th1 (left graph) or Th17 (right graph) OT-II T-cells were added to the upper compartment and T cell migration was quantified by flow cytometry 3h later, using fluorescent labelled beads as reference. **p<0.01, ***p<0.001 (ANOVA with Bonferroni correction).

**DOI:** http://dx.doi.org/10.7554/eLife.13149.010

The T cells added to the assay have been activated by dendritic cells and expanded with IL-2. They are highly activated. My understanding is that the antigen-specific T cells that would come into the CNS in response to myelin antigens presented by the BBB EC in multiple sclerosis or EAE would be memory cells but not activated at the time but become activated upon presentation of their cognate antigen by the endothelial cells. What happens if the T cells are allowed to rest before being added to the Transwells?

To differentiate the naive T cells into Th1 or Th17 cells as well as to expand them, the T cells were activated in vitro by antigen-loaded DCs. However, at time of addition to the antigen-loaded BECs, which is 7 days after the initial stimulus, the T cells were rounded up again and receptive for re-stimulation. Moreover, upon activation these cells secreted significant amounts of effector cytokines (as measured by intracellular staining and flow cytometry), which would be hampered if they would still be activated.

Reviewer #2:

This study is generally well performed and presented. However, a couple of important issues/questions remain to be addressed.

In order to measure whether the localization of the myelin signal was intracellular or membrane-bound, we designed a mask that excludes the cell membrane and calculated a ratio of the amount of fluorescence located in the mask vs the total amount of fluorescence, as previously reported (Garcia-Vallejo et al., Front Immunol 2015). The results indicate that the myelin fluorescence signal was intracellular, demonstrating that BECs are able to efficiently internalize myelin (Author response image 4). Using Imaging flow cytometry we could clearly see that myelin was internalized and not located on the membrane. However, the reviewer is correct that if there was myelin attached to the membrane but not internalized at the moment of trypsinization, this signal would be gone and would be undetectable as not-internalized material. We therefore performed, as suggested, experiments on adherent BECs. Using confocal laser-scanning microscopy (CSLM) we ascertained that myelin is indeed internalized by the BECs and does not remain on the membrane. Furthermore, it is clear that 24 hours after loading, internalized myelin is contained within LAMP1 positive vesicles and not with EEA1 positive organelles (Author response image B-G), which corroborate our observations made with Imaging flow cytometry.

Author response image 4. — (A) BECs were loaded with Atto 633-labeled human myelin for 24h and internalization of myelin particles was assessed by imaging flow cytometry. To determine internalization scores, a mask was designed based on the surface of BECs in the brightfield image. This mask was then eroded to exclude the cell membrane. The resulting mask was applied to the fluorescence channel. The internalization score, interpreted as a ratio of the intensity of the intracellular space versus the intensity of the whole cell, was calculated on this mask using the internalization feature of the Ideas v6.0 software (AMNIS Merck Millipore). Cells that have internalized antigens have positive scores, as depicted here for BECs. (**B-G**) Adherent brain endothelial cells were incubated with Atto-633-labeled myelin and 24h later, co-localization of myelin (Red) with early endosomal (EEA1, Green, upper panels) or endosomal/lysosomal (LAMP1, Green, lower panels) compartments was analyzed using CSLM. Nuclei were visualized with Hoechst (blue). Representative images of adherent brain endothelial cells with subcellular localization of Myelin with EEA1 (or LAMP1 (C). A magnification of indicated areas is shown in D B)+E. F+G. Histograms were created for a selected area (indicated by a line) using ImageJ software (NIH, USA). Histograms were created from each fluorochrome and overlays were made by the program. Histograms clearly show the lack of association of myelin with early endosomes but the enclosure of myelin within LAMP1 positive vesicles.

**DOI:** http://dx.doi.org/10.7554/eLife.13149.011

The Imaging flow cytometry data showing the internalization score of myelin have been added to the manuscript (Results, subsection “Brain endothelial cells internalize exogenous antigens irrespective of their activation status”, last paragraph and as Figure 1—figure supplement 1).

The new CLSM data obtained on the internalization and intracellular routing of myelin in adherent brain endothelial cells are added to the Results & Discussion section (subsection “Myelin internalized by BECs is directed to the endo-lysosome compartments”, first paragraph) and as Figure 2—figure supplement 2.

We have additionally activated brain endothelial cells with IFNγ and analyzed the expression of MHC-II as well as MHC-I, VCAM-1 and CD40 using flow cytometry. As shown in Author response image 3, IFN-γ increases, similar to TNF, expression of all markers analyzed. Although the expression of MHC-I, VCAM-1 and CD40 seems to be more enhanced by IFN-γ than by TNFα, in our hands both cytokines increase expression of these costimulatory and MHC molecules.

Author response image 5. — Confluent monolayers of brain endothelial cells (BECs) were stimulated with 10 ng/ml IFN-γ for 24 hr. Expression of MHC-I, MHC-II, CD40 and VCAM-1 was determined by flow cytometry. Histograms depict expression of indicated markers in resting (orange line) and activated (green line) BECs. Red and blue lines indicate isotype controls.

**DOI:** http://dx.doi.org/10.7554/eLife.13149.012

[Editors' note: further revisions were requested prior to acceptance, as described below.]

Summary:

We understand the concerns from the reviewers regarding the confluency of the endothelial cells.

Both claudin-5 and VE-cadherin are highly present at cell-cell contacts in the confluent human brain endothelial cells, indicative of tight barrier formation by the brain endothelial cells. We agree that there appear to be differences in the expression level of claudin-5 and VE-cadherin at the cell-cell border throughout the monolayers, which may be the result of the presence of serum upon the culturing. To demonstrate that the brain endothelial cells form confluent monolayers we performed additional staining for the cytoskeletal F-actin, which reveals its localization of the cell-cell contacts, indicative of confluency (Author response image 6).

Author response image 6. — **DOI:** http://dx.doi.org/10.7554/eLife.13149.013

We performed additional FRET experiments using a more conventional setup to study FRET, i.e. FilterFRET (Figure 2A) and acceptor photo-bleaching (Figure 2B), by confocal laser scanning microscopy on adhered brain endothelial cells. Using this approach, we could not show the occurrence of FRET in our studies. We therefore agree with and sincerely apologize to the reviewers that the FRET as described previously (measured using imaging flow cytometry), is no “true” FRET, revealing the limitations of this techniques (possibly due to the lower resolution in comparison to the CLSM). However, this technique remains suitable to demonstrate the colocalization between myelin particles and EEA1 or LAMP1. We have adapted Figure 2 and the respective Results section in the manuscript accordingly.

See response to point #1.

[Editors' note: further revisions were requested prior to acceptance, as described below.]

The manuscript has been improved but there are some remaining issues that need to be addressed as outlined below:

We understand the concerns about the co-localization between LAMP1 and Myelin. For visualization of the lysosomes, we used LAMP1, which is a transmembrane protein of the lysosome. We thus stain the membrane of lysosomes. The myelin that has been taken up by the lysosomes is inside the lumen of the lysosome. This results in the observed images where we see myelin surrounded by LAMP1 staining. To show that this is also the case in the Z-direction, we included orthogonal views of the lysosomes in Figure 2F of the manuscript.

Due to the optical sectioning of the confocal microscope the LAMP1 positive lysosomes are observed as ring shaped lysosomes, at which the intensity of the Lamp1 staining is very bright at the lysosomal membrane, while in the lumen the intensity drops significantly. The intensity of myelin is very high in the lumen of the lysosome. Due to these difference in intensity, the overlap of the fluorescence does not appear yellow. Yellow overlap only occurs when the fluorescence intensities of both channels is equal. To illustrate this, we generated green and red images of different intensities and their corresponding merge images and histograms (see Author response image 7).

Author response image 7. — (A) co-localization of high intensity green and high intensity red image shows a yellow merge image. Histogram shows overlap of intensities. (B) Co-localization of low intensity green dot with bright green ring with a high intensity red dot, Merge image does not show yellow colour. Histogram shows high intensity of green channel at the ring and high intensity of red channel at the inside. (C) Green LAMP1 positive lysosome shows high intensity border and lower intensity lumen together with high intensity myelin in red. The merge image does not appear yellow.

**DOI:** http://dx.doi.org/10.7554/eLife.13149.014

Because lysosomal diameters are much closer to 1 μm than the 3 μm shown in the scale, the authors should really show an image that shows true overlap of signals rather than what appear in Figure 2H as three separate peaks.

The shown lysosome is indeed relative large. However, it has been reported that the size of a lysosome differs depending on cell type, and in most cells lysosomes can be up to 1.5 μm in diameter (Lüllmann-Rauch R. History and morphology of the lysosome. In: Saftig P., editor.Lysosomes. New York: Springer; 2005). It has also been reported that lysosomes increase in size as a result of accumulation of undigested material (Appelqvist et al., J Mol Cell Biol. 2013). To support this statement, we measured the size of myelin containing (LAMP1-positive) lysosomes as well as none myelin containing lysosomes. The average size of myelin containing lysosomes is 1.4 ± 0.08 µm while the size of none myelin containing lysosomes is around 1.1 ± 0.1 µm (P=0.004; N=10; please see Author response image 7).

The image in the previous Figure 2F of the manuscript showed a lysosome with a size of 1.8 µm, we therefore exchanged the image for another image that more closely represents the average measured lysosome size.

Author response image 8. — The diameter of lysosomes was determined using ImageJ software and shown as histograms. The average size of 5 lysosomes per cell of a total of 10 cells that contain myelin is compared with that of none myelin containing lysosomes and represented in the bar graph.

**DOI:** http://dx.doi.org/10.7554/eLife.13149.015

Moreover, no quantification of these data (2C-H) are presented, just one set of micrographs for each. Readers need to know what percentage of total myelin fragments were internalized and associated with each marker. The authors should be able to obtain these numbers from their existing data.

We apologize for not providing the quantification of the CSLM data. Using both CSLM and imaging flow cytometry (ISX) we find that approximately 50% of brain endothelial cells contain myelin. Moreover, the majority of the myelin is present in LAMP1 positive lysosomes.We have added this quantification as panel I to Figure 2 in the manuscript. The data obtained by ISX were already included in Figure 1C of the manuscript.

The average frequency of T cells that transmigrated in the control setting are 10.8% ± 1.2 for 2D2 Th1; 11.6% ± 0.4 for 2D2 Th17; 7.9% ± 1.9 for OT-II Th1 and 12.5% ± 1.4 for OT-II Th17. The indicated percentages of migrated control cells have been added to the legend of Figure 3.

Although these percentages may seem low in comparison to other studies on endothelial transmigration by T cells, we want to emphasize that these studies used cell lines (of both brain endothelial cells as well as of T cells), in contrast to our study using primary brain endothelial cells and primary T cells. Of note, in a study by Reiss and Engelhardt (Int Immunol. 1999) similar frequencies of transmigrated T cells in the controls were found when using primary T cells.

We have adapted this sentence as requested.

[bib1] Archambault AS, Sim J, Gimenez MA, Russell JH. Defining antigen-dependent stages of T cell migration from the blood to the central nervous system parenchyma. European Journal of Immunology. 2005;35:1076–1085. doi: 10.1002/eji.200425864. [DOI] [PubMed] [Google Scholar]

[bib2] Bailey SL, Schreiner B, McMahon EJ, Miller SD. CNS myeloid DCs presenting endogenous myelin peptides 'preferentially' polarize CD4+ T(H)-17 cells in relapsing EAE. Nature Immunology. 2007;8:172–180. doi: 10.1038/ni1430. [DOI] [PubMed] [Google Scholar]

[bib3] Bettelli E, Pagany M, Weiner HL, Linington C, Sobel RA, Kuchroo VK. Myelin oligodendrocyte glycoprotein-specific T cell receptor transgenic mice develop spontaneous autoimmune optic neuritis. Journal of Experimental Medicine. 2003;197:1073–1081. doi: 10.1084/jem.20021603. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib4] Bielekova B, Goodwin B, Richert N, Cortese I, Kondo T, Afshar G, Gran B, Eaton J, Antel J, Frank JA, McFarland HF, Martin R. Encephalitogenic potential of the myelin basic protein peptide (amino acids 83-99) in multiple sclerosis: results of a phase II clinical trial with an altered peptide ligand. Nature Medicine. 2000;6:1167–1175. doi: 10.1038/80516. [DOI] [PubMed] [Google Scholar]

[bib5] Blum JS, Wearsch PA, Cresswell P. Pathways of antigen processing. Annual Review of Immunology. 2013;31:443–473. doi: 10.1146/annurev-immunol-032712-095910. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib6] Carbajal KS, Mironova Y, Ulrich-Lewis JT, Kulkarni D, Grifka-Walk HM, Huber AK, Shrager P, Giger RJ, Segal BM. Th cell diversity in experimental autoimmune encephalomyelitis and multiple sclerosis. Journal of Immunology. 2015;195:2552–2559. doi: 10.4049/jimmunol.1501097. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib7] Coisne C, Dehouck L, Faveeuw C, Delplace Y, Miller F, Landry C, Morissette C, Fenart L, Cecchelli R, Tremblay P, Dehouck B. Mouse syngenic in vitro blood-brain barrier model: a new tool to examine inflammatory events in cerebral endothelium. Laboratory Investigation. 2005;85:734–746. doi: 10.1038/labinvest.3700281. [DOI] [PubMed] [Google Scholar]

[bib8] Faille D, El-Assaad F, Mitchell AJ, Alessi MC, Chimini G, Fusai T, Grau GE, Combes V. Endocytosis and intracellular processing of platelet microparticles by brain endothelial cells. Journal of Cellular and Molecular Medicine. 2012;16:1731–1738. doi: 10.1111/j.1582-4934.2011.01434.x. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib9] Falcone S, Cocucci E, Podini P, Kirchhausen T, Clementi E, Meldolesi J. Macropinocytosis: regulated coordination of endocytic and exocytic membrane traffic events. Journal of Cell Science. 2006;119:4758–4769. doi: 10.1242/jcs.03238. [DOI] [PubMed] [Google Scholar]

[bib10] Galea I, Bernardes-Silva M, Forse PA, van Rooijen N, Liblau RS, Perry VH. An antigen-specific pathway for CD8 T cells across the blood-brain barrier. Journal of Experimental Medicine. 2007;204:2023–2030. doi: 10.1084/jem.20070064. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib11] García-Vallejo JJ, Bloem K, Knippels LM, Garssen J, van Vliet SJ, van Kooyk Y. The Consequences of Multiple Simultaneous C-Type Lectin-Ligand Interactions: DCIR Alters the Endo-Lysosomal Routing of DC-SIGN. Frontiers in Immunology. 2015;6:e13149. doi: 10.3389/fimmu.2015.00087. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib12] García-Vallejo JJ, Ilarregui JM, Kalay H, Chamorro S, Koning N, Unger WW, Ambrosini M, Montserrat V, Fernandes RJ, Bruijns SC, van Weering JR, Paauw NJ, O'Toole T, van Horssen J, van der Valk P, Nazmi K, Bolscher JG, Bajramovic J, Dijkstra CD, 't Hart BA, van Kooyk Y. CNS myelin induces regulatory functions of DC-SIGN-expressing, antigen-presenting cells via cognate interaction with MOG. Journal of Experimental Medicine. 2014;211:1465–1483. doi: 10.1084/jem.20122192. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib13] Georgieva JV, Kalicharan D, Couraud PO, Romero IA, Weksler B, Hoekstra D, Zuhorn IS. Surface characteristics of nanoparticles determine their intracellular fate in and processing by human blood-brain barrier endothelial cells in vitro. Molecular Therapy. 2011;19:318–325. doi: 10.1038/mt.2010.236. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib14] Hamilton-Williams EE, Palmer SE, Charlton B, Slattery RM. Beta cell MHC class I is a late requirement for diabetes. PNAS. 2003;100:6688–6693. doi: 10.1073/pnas.1131954100. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib15] Inaba K, Turley S, Iyoda T, Yamaide F, Shimoyama S, Reis e Sousa C, Germain RN, Mellman I, Steinman RM. The formation of immunogenic major histocompatibility complex class II-peptide ligands in lysosomal compartments of dendritic cells is regulated by inflammatory stimuli. Journal of Experimental Medicine. 2000;191:927–936. doi: 10.1084/jem.191.6.927. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib16] Jin Y, Fuller L, Ciancio G, Burke GW, Tzakis AG, Ricordi C, Miller J, Esquenzai V. Antigen presentation and immune regulatory capacity of immature and mature-enriched antigen presenting (dendritic) cells derived from human bone marrow. Human Immunology. 2004;65:93–103. doi: 10.1016/j.humimm.2003.11.002. [DOI] [PubMed] [Google Scholar]

[bib17] Khaw LT, Ball HJ, Golenser J, Combes V, Grau GE, Wheway J, Mitchell AJ, Hunt NH, Obeid S, Couraud PO. Endothelial cells potentiate interferon-γ production in a novel tripartite culture model of human cerebral malaria. PLoS One. 2013;8:e13149. doi: 10.1371/journal.pone.0069521. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib18] Larochelle C, Alvarez JI, Prat A. How do immune cells overcome the blood-brain barrier in multiple sclerosis? FEBS Letters. 2011;585:3770–3780. doi: 10.1016/j.febslet.2011.04.066. [DOI] [PubMed] [Google Scholar]

[bib19] Lees JR, Sim J, Russell JH. Encephalitogenic T-cells increase numbers of CNS T-cells regardless of antigen specificity by both increasing T-cell entry and preventing egress. Journal of Neuroimmunology. 2010;220:10–16. doi: 10.1016/j.jneuroim.2009.11.017. [DOI] [PubMed] [Google Scholar]

[bib20] Lublin FD, Cutter G, Giovannoni G, Pace A, Campbell NR, Belachew S. Natalizumab reduces relapse clinical severity and improves relapse recovery in MS. Multiple Sclerosis and Related Disorders. 2014;3:705–711. doi: 10.1016/j.msard.2014.08.005. [DOI] [PubMed] [Google Scholar]

[bib21] Ludowyk PA, Willenborg DO, Parish CR. Selective localisation of neuro-specific T lymphocytes in the central nervous system. Journal of Neuroimmunology. 1992;37:237–250. doi: 10.1016/0165-5728(92)90008-9. [DOI] [PubMed] [Google Scholar]

[bib22] Ransohoff RM, Kivisäkk P, Kidd G. Three or more routes for leukocyte migration into the central nervous system. Nature Reviews Immunology. 2003;3:569–581. doi: 10.1038/nri1130. [DOI] [PubMed] [Google Scholar]

[bib23] Reijerkerk A, Lopez-Ramirez MA, van Het Hof B, Drexhage JA, Kamphuis WW, Kooij G, Vos JB, van der Pouw Kraan TC, van Zonneveld AJ, Horrevoets AJ, Prat A, Romero IA, de Vries HE. MicroRNAs regulate human brain endothelial cell-barrier function in inflammation: implications for multiple sclerosis. Journal of Neuroscience. 2013;33:6857–6863. doi: 10.1523/JNEUROSCI.3965-12.2013. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib24] Reith W, LeibundGut-Landmann S, Waldburger JM. Regulation of MHC class II gene expression by the class II transactivator. Nature Reviews Immunology. 2005;5:793–806. doi: 10.1038/nri1708. [DOI] [PubMed] [Google Scholar]

[bib25] Roche PA, Furuta K. The ins and outs of MHC class II-mediated antigen processing and presentation. Nature Reviews Immunology. 2015;15:203–216. doi: 10.1038/nri3818. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib26] Sospedra M, Martin R. Immunology of multiple sclerosis. Annual Review of Immunology. 2005;23:683–747. doi: 10.1146/annurev.immunol.23.021704.115707. [DOI] [PubMed] [Google Scholar]

[bib27] Tietz S, Engelhardt B. Brain barriers: Crosstalk between complex tight junctions and adherens junctions. Journal of Cell Biology. 2015;209:493–506. doi: 10.1083/jcb.201412147. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib28] Tompkins SM, Padilla J, Dal Canto MC, Ting JP, Van Kaer L, Miller SD. De novo central nervous system processing of myelin antigen is required for the initiation of experimental autoimmune encephalomyelitis. Journal of Immunology. 2002;168:4173–4183. doi: 10.4049/jimmunol.168.8.4173. [DOI] [PubMed] [Google Scholar]

[bib29] Trapp BD, Peterson J, Ransohoff RM, Rudick R, Mörk S, Bö L. Axonal transection in the lesions of multiple sclerosis. New England Journal of Medicine. 1998;338:278–285. doi: 10.1056/NEJM199801293380502. [DOI] [PubMed] [Google Scholar]

[bib30] Unger WW, Pearson T, Abreu JR, Laban S, van der Slik AR, der Kracht SM, Kester MG, Serreze DV, Shultz LD, Griffioen M, Drijfhout JW, Greiner DL, Roep BO, Mulder-van der Kracht MG, Serreze LD. Islet-specific CTL cloned from a type 1 diabetes patient cause beta-cell destruction after engraftment into HLA-A2 transgenic NOD/scid/IL2RG null mice. PLoS One. 2012;7:e13149. doi: 10.1371/journal.pone.0049213. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib31] van Halteren AG, Kardol MJ, Mulder A, Roep BO. Homing of human autoreactive T cells into pancreatic tissue of NOD-scid mice. Diabetologia. 2005;48:75–82. doi: 10.1007/s00125-004-1613-2. [DOI] [PubMed] [Google Scholar]

[bib32] Weksler BB, Subileau EA, Perrière N, Charneau P, Holloway K, Leveque M, Tricoire-Leignel H, Nicotra A, Bourdoulous S, Turowski P, Male DK, Roux F, Greenwood J, Romero IA, Couraud PO. Blood-brain barrier-specific properties of a human adult brain endothelial cell line. FASEB Journal. 2005;19:1872–1874. doi: 10.1096/fj.04-3458fje. [DOI] [PubMed] [Google Scholar]

[bib33] Weller RO, Engelhardt B, Phillips MJ. Lymphocyte targeting of the central nervous system: a review of afferent and efferent CNS-immune pathways. Brain Pathology. 1996;6:275–288. doi: 10.1111/j.1750-3639.1996.tb00855.x. [DOI] [PubMed] [Google Scholar]

PERMALINK

Internalization and presentation of myelin antigens by the brain endothelium guides antigen-specific T cell migration

Melissa A Lopes Pinheiro

Alwin Kamermans

Juan J Garcia-Vallejo

Bert van het Hof

Laura Wierts

Tom O'Toole

Daniël Boeve

Marleen Verstege

Susanne MA van der Pol

Yvette van Kooyk

Helga E de Vries

Wendy WJ Unger

Roles

Abstract

eLife digest

Introduction

Results and discussion

Brain endothelial cells internalize exogenous antigens irrespective of their activation status

Figure 1. Human brain endothelial cells internalize myelin particles.

Figure 1—figure supplement 1. Human brain endothelial cells increase MHC and costimulatory molecule expression upon activation by IFN-γ.

Figure 1—figure supplement 2. Brain endothelial cells internalize myelin particles.

Myelin internalized by BECs is directed to the endo-lysosome compartments

Figure 2. Myelin particles are preferably routed to the endo-lysosomes.

Migration of myelin-specific T-cells depends on presentation of myelin-antigens in MHC-II by BECs

Figure 3. Migration of myelin-specific T-cells depends on presentation of myelin-antigens in MHC-II by BECs.

Materials and methods

Cell culture

Flow cytometry

Imaging flow cytometry

Confocal microscopy

Isolation and culture of primary murine BECs (mBECs)

Generation of MOG-specific T cell subsets

Quantitative PCR

Transwell migration

Statistical analysis

Acknowledgements

Funding Statement

Funding Information

Additional information

Competing interests

Author contributions

References

Decision letter

Roles

Author response image 1. Confluent human brain endothelial cells form an impermeable barrier.

Author response

Author response image 2. Induction of T cell activation but not proliferation by brain endothelial cells.

Author response image 3. OVA-specific T-cells only transmigrate upon encounter of BECs presenting OVA-antigens in MHC-II.

Author response image 4. Myelin is internalized and routed to lysosomal compartments by brain endothelial cells.

Author response image 5. Human brain endothelial cells increase MHC and costimulatory molecule expression upon activation by IFN-γ.

Author response image 6. Confluency of cultured brain endothelial cells shown using F-actin, Claudin-5 and VE-Cadherin staining on samples shown in Figure 2 of the manuscript.

Author response image 7. Co-localization of green and red channels.

Author response image 8. Average size of lysosomes in myelin-loaded brain endothelial cells.

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases