Acid Sphingomyelinase–Derived Ceramide Regulates ICAM-1 Function during T Cell Transmigration across Brain Endothelial Cells Free

Increasing evidence suggests that alterations in the sphingolipid pathway are involved in the pathogenesis of several neurologic disorders. Sphingolipids represent a class of lipids ubiquitously present in eukaryotic cells, particularly in cell membranes. Until recently, their main role was thought to organize specific domains in the plasma membrane. Nowadays, it becomes increasingly clear that sphingolipids are not just bystander molecules, but they provide an active role in modulating cellular events related to proliferation, differentiation, apoptosis, and inflammation (1, 2). Sphingolipid metabolism alterations were associated with Alzheimer’s, Huntington’s, Parkinson’s, and prion diseases, a process possibly associated with their pathogenesis (3–7).

Altered sphingolipid metabolism may also contribute to multiple sclerosis (MS). MS is a chronic inflammatory disorder of the CNS characterized by blood–brain barrier dysfunction, immune cell infiltration into the brain, astrogliosis, demyelination, and axonal damage (8, 9). Previously, we reported that a number of enzymes involved in sphingolipid metabolism, in particular the enzymes sphingomyelinases are deregulated in MS. Sphingomyelinases are rate-limiting enzymes involved in ceramide synthesis and have been classified in acid, neutral, or alkaline regarding their pH optimum for activity. Acid sphingomyelinase (ASM) is the acidic form and converts sphingomyelin into ceramide. Depending on its glycosylation pattern, ASM can be found in lysosomes or be secreted (10). It is thought that ASM resides in secretory lysosomes, which are directed to and fused with the plasma membrane upon certain stimuli. For instance, ASM has been shown to be activated by the inflammatory cytokines TNF-α and IL-1, and by oxidative stress (11–13). We previously demonstrated that the expression of ASM is upregulated in reactive astrocytes in MS lesions (14). Furthermore, ASM-derived ceramide decreased the barrier function in brain endothelial cells (BECs) (14). In fact, ceramide, an important second lipophilic messenger involved in receptor clustering, signal transduction, and apoptosis, is one of the main effectors of ASM signaling (15–17). Furthermore, ceramide is capable of self-aggregation leading to the formation of ceramide-enriched membrane rafts that fuse with each other creating bigger structures or platforms (18). These platforms are thought to trap and cluster plasma membrane receptors facilitating their signaling (16, 17, 19–22).

The intercellular adhesion molecule-1 (ICAM-1) regulates the adhesion to and transmigration of T lymphocytes across the brain endothelium (23, 24). Upon binding to its counterpart, LFA-1/macrophage-1 Ag present on T cells, ICAM-1 is clustered into lipid rafts (25). Moreover, filamin, an actin regulator, links ICAM-1 to the actin cytoskeleton upon clustering, required for the formation of so-called docking structures (26), which have been reported to be involved in T cell transmigration (27–30).

We have previously shown that exogenously added sphingomyelinase or ceramide increased monocyte transmigration through the brain endothelium (14). However, the mechanism by which sphingolipids regulate this process is not clearly understood. Because ICAM-1 mediates T lymphocyte diapedesis, a process detrimental during MS lesion formation, we hypothesized that ASM-derived ceramide regulates ICAM-1 function and thereby controls T cell transmigration in the brain endothelium.

Brain tissue from nonneurologic control cases and MS patients was obtained at rapid autopsy and immediately frozen in liquid nitrogen or fixed in formalin (in collaboration with the Netherlands Brain Bank, coordinator Dr. Huitinga) (Table I). The Netherlands Brain Bank received permission to perform autopsies for the use of tissue and for access to medical records for research purposes from the Ethical Committee of the VU University Medical Center (Amsterdam, the Netherlands). Tissue samples from control cases without neurologic disease were taken from the subcortical white matter. All control cases, or their next of kin, had given informed consent for autopsy and the use of their brain tissue for research purposes.

For immunohistochemical analysis, 5-μm cryosections were fixed in acetone for 10 min, incubated overnight at 4°C with primary Ab against ASM (rabbit; Santa Cruz Biotechnology, Santa Cruz, CA). Subsequently, sections were incubated with EnVision Dual Link (DAKO, Glostrup, Denmark) for 30 min at room temperature. Diaminobenzidine tetrachloride (DAKO, Glostrup, Denmark) was used as the chromogen. Between incubation steps, sections were thoroughly washed with PBS. After a short rinse in tap water, sections were incubated with hematoxylin for 1 min and extensively washed with tap water for 10 min. Finally, sections were dehydrated with ethanol followed by xylol and mounted with Entellan (Merck, Darmstadt, Germany). All Abs were diluted in PBS containing 0.1% BSA (Boehringer-Mannheim, Mannheimm, Germany). For colocalization studies, sections were incubated for 30 min with 10% normal goat serum (NGS), incubated overnight at 4°C with primary Ab against ASM (rabbit; Santa Cruz) after permeabilization with 0.05% Triton X-100, and Goat anti-rabbit Alexa 555 (Molecular Probes, Leiden, the Netherlands) was used as secondary Ab. To visualize endothelial cells, we incubated sections with an Ab directed against CD31 (mouse; DAKO, Glostrup, Denmark), for 1 h at room temperature. Then sections were incubated with secondary goat anti-mouse Alexa 488 (Molecular Probes, Leiden, the Netherlands). After washing, slides were covered with Vectashield (Vector Laboratories, Burlington, CA) supplemented with 0.4% DAPI to stain nuclei. Fluorescence analysis was performed with a Leica DM6000 (Leica Microsystems, Heidelberg, Germany).

The human brain BEC line hCMEC/D3 (31) was kindly provided by Dr. P.-O. Couraud (Institut Cochin, Université Paris Descartes, Paris, France). BECs were grown in endothelial growth basal medium 2 supplemented with human epidermal growth factor, hydrocortisone, GA-1000, vascular endothelial growth factor, human fibroblastic growth factor B, R3-IGF-1, ascorbic acid, and 2.5% FCS (Lonza, Basel, Switzerland). Human PBLs were recovered after monocyte isolation from buffy coats (remaining fraction) (32). These T cells were activated with 1 mg/ml PHA (Sigma Aldrich, Zwijndrecht, the Netherlands) and 10 ng/ml IL-2 (MACS; Miltenyi Biotec, Bergisch Gladbach, Germany) for 48 h.

For knockdown of ASM (National Center for Biotechnology Information Gene ID: SMPD1), we used a vector-based shRNA delivery system. shRNA expression plasmids from the TRC library (https://www.broadinstitute.org/rnai/trc/lib) targeting human SMPD1 were used to produce recombinant lentiviruses. To this end, subconfluent human embryonic kidney 293T cells were cotransfected with the shRNA lentivirus expression plasmid, packaging plasmids (pMDLg/pRRE and pRSV Rev), and the envelope vector pMD2.G using calcium phosphate as a transfection reagent. Human embryonic kidney 293T cells were cultured in DMEM supplemented with 10% FCS, 1% penicillin/streptomycin, in a 37°C incubator with 5% CO2. Infectious lentiviruses were collected 48 h after transfection, and the supernatant was centrifuged to remove cell debris. BECs were transduced with the lentivirus-containing shRNA. Forty-eight hours postinfection, stable cell lines were selected by puromycin treatment (2 μg/ml). The knockdown efficiency of all five SMPD1 library constructs was tested, and the most effective construct used in subsequent experiments was TRCN0000049014, encoding CCCAATCTGCAAAGGTCTATT targeting nt 458–478 of the NM_000543.4 SMPD1 RefSeq. The TRC SHC002 vector containing a nontargeting control (NTC) sequence of the human genome was used as a negative control.

Cell lysates of control and ASM-deficient cells were prepared with lysis buffer (Cell Signaling, Leiden, the Netherlands), and ASM activity of the supernatants was quantified using the Acid Sphingomyelinase Assay Kit (Echelon Biosciences, Salt Lake City, UT) according to manufacturer’s instructions and using the standard curve as reference. ASM activity was corrected for protein content in each sample after protein measurement (Pierce BCA Protein kit; Thermo Fisher).

mRNA was isolated from BECs 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: SMPD1 forward 5′-GCTGGAGCTGGAATTATTACCG-3′, SMPD1 reverse 5′-CGGCTCAGAGTCTCTTCATTCAT-3′, SMPD2 forward 5′-TGTCCGCATTGACTACGTG-3′, SMPD2 reverse 5′-ACAAACAGAGTAGCCATGAGG-3′, SMPD4 forward 5′-CTCATGGTGTTCCGAGTGG-3′, SMPD4 reverse 5′-AAGCTCCCAGTGAATGTGG-3′, GAPDH forward 5′-CCATGTTCGTCATGGGTGTG-3′, GAPDH reverse 5′-GGTGCTAAGCAGTTGGTGGTG-3′. Oligonucleotides were synthesized by Invitrogen (Bleiswijk, the Netherlands). Quantitative PCRs (qPCRs) were performed in an ABI7900HT sequence detection system using the SYBR Green method (Applied Biosystems, New York, NY). Expression levels were normalized to GAPDH expression levels.

Cells were detached from 24-well culture plates with 1 mg/ml collagenase type I (Sigma Aldrich), washed, and blocked with 5% NGS in PBS/0.1% BSA for 30 min on ice. Subsequently, cells were stained with monoclonal mouse anti–ICAM-1 (Rek-1, 5 μg/ml, a kind gift from the Department of Tumor Immunology, University Medical Center St. Radboud, Nijmegen, the Netherlands) for 30 min on ice. After washing, staining was detected using a goat anti-mouse Alexa 488 (Molecular Probes, Eugene, OR). Omission of primary Abs served as negative control. Fluorescence intensity was measured using a FACSCalibur flow cytometer (Becton Dickinson, San Jose, CA). The mean fluorescence intensity was used as a measure for the expression of ICAM-1.

Cells were plated in collagen-coated μ-slide 8-well slides (Ibidi, Martinsried, Germany) and stimulated with recombinant human TNF-α (10 ng/ml; Peprotech) for 24 h. Ab-coated polystyrene beads (10 μm; Polysciences, Eppelheim, Germany) were pretreated with 8% (v/v) glutaraldehyde for 6 h, washed five times with PBS, and incubated with 200 μg/ml ICAM-1 (Santa Cruz) according to the manufacturer’s protocol. Free glutaraldehyde groups were blocked with 0.2 M ethanolamine and 10 mg/ml BSA. Anti–ICAM-1–coated beads were added for 30 min to the cells. After extensive wash, cells were fixated with 4% formaldehyde (Sigma, St. Louis, MO) in PBS (Life Technologies). Cells were washed and blocked with PBS/0.1% BSA (Sigma) and 5% NGS. Subsequently, cells were incubated with mouse anti-ceramide primary Ab (Alexis, Lausen, Switzerland) in PBS/0.1% BSA overnight and anti–ICAM-1. Cells were washed and incubated with goat anti-mouse Alexa 488 or goat anti-rabbit Alexa 555 secondary Abs (Molecular Probes, Eugene, OR). After washing the cells with PBS, Hoechst (Molecular Probes) was used for nuclear staining. Staining was analyzed using a confocal LSM microscope (Leica).

Control and ASM-deficient BECs were seeded in 96-well plates and stimulated with TNF-α (10 ng/ml) for 24 h after cells reached confluency. Human PBLs were fluorescently labeled with 0.5 μM calcein-AM (Molecular Probes, Eugene OR) and resuspended in RPMI complete medium + 0.5% BSA + 25 mM HEPES to a final concentration of 1 × 10⁶ T cells. A standard curve was made with 0, 12.5, 25, 50, and 100% of this T cell suspension. BECs were washed, and T cells were added to the endothelium for 30 min in a 37°C incubator with 5% CO₂. Nonadherent cells were washed and adherent cells were lysed with 0.1M NaOH. Fluorescence intensity was measured (FLUOstar Galaxy; BMG Labtechnologies, Offenburg, Germany; excitation 480 nm, emission 520 nm), and the number of adhered T cells was calculated using the standard curve.

BECs were grown to confluence in 96-well plates and subsequently exposed to TNF-α (10 ng/ml) for 24 h. After extensive washing, human T cells (7.5 × 10⁵ cells/ml) were added to BEC monolayers, and the number of migrated T cells was assessed after 4 h (33). To monitor T cell migration, we placed cocultures in an inverted phase-contrast microscope (Nikon Eclipse TE300; Lijnden, the Netherlands) housed in a temperature-controlled (37°C), 5% CO₂ gassed chamber. A field (220 × 220 μm) was randomly selected and recorded for 10 min for 50 times by using a digital video camera using Cell F imaging software (Olympus, Heidelberg, Germany). Diapedesis was assessed by enumerating the number of T cells within the field that had either adhered or migrated through the monolayer. Transmigrated cells (phase-dark) could be readily distinguished from those remaining on the cell surface by their highly refractive (phase-bright) morphology. To test the role of ASM inhibitor on T cell migration, 50 μM Imipramine hydrochloride (imipramine hydrochloride; Sigma Aldrich) was added simultaneously with the T cells to BEC monolayers.

To assess the migration under physiological flow conditions, we cultured BECs to confluence in collagen-coated μ−Slide VI flow chambers (Ibidi). Cells were treated for 24 h with 10 ng/ml TNF-α, connected to a perfusion pump, subjected to the physiological flow speed of 0.8 dyn/cm², and 5 × 10⁵ T cells were perfused per condition. During this time leukocyte adhesion and transmigration under flow were recorded for 30 min with a Zeiss Axiovert 200 microscope (10× objective) equipped with a motorized stage. Images were recorded with Zeiss Zen 2008 software. Live Imaging was performed at 37°C and 5% CO₂. Transmigrated T cells were distinguished from adhered T cells by their transition from bright to dark morphology. Quantification was performed by using the cell-counter plugin in the ImageJ software (v1.38r).

For immunofluorescence, Control (NTC shRNA) and ASM-deficient BECs (ASM shRNA) were grown on collagen-coated Lab-Tek (Thermo Fisher Scientific, Waltham, MA), treated for 24 h with 10 ng/ml TNF-α, fixed, and stained for F-actin and endogenously expressed ICAM-1. mAb against ICAM-1 was purchased from R&D Systems (Minneapolis, MN), F-actin filaments were visualized by using Acti-stain-555 phalloidin, Cytoskeleton (Cytoskeleton, Denver, CO), and secondary Alexa-coupled Abs were purchased from Invitrogen. Imaging was performed using an LSM510 Meta confocal laser-scanning microscope (Carl Zeiss MicroImaging, Jena, Germany) using Zeiss Zen 2008 software. Detailed Z-stacks of BECs were created and combined into a maximum intensity projection image (ImageJ). Microvilli were quantified using ImageJ.

BECs were seeded in collagen-coated eight-well μ-slide (Ibidi) and transfected with 0.5 μg ICAM-1–GFP construct using Dharmafect4 (Thermo Fisher Scientific) as a transfecting reagent. Cells were then stimulated with 10 ng/ml TNF-α 24 h before imaging. Slides were mounted onto a heating block connected to a confocal microscope (Zeiss LSM510). Fluorescent recovery after photobleaching (FRAP) experiments were performed using 30 iterations with 488-nm laser illumination, at maximum power (25 mW). Fluorescence recovery was measured by time-lapse imaging. Image analysis was performed with LSM510 software (Carl Zeiss MicroImaging). The first points before photobleaching were set at 100%, and the first measurement after bleaching was set to 0% to normalize the fluorescence recovery from each window.

Magnetic goat anti-mouse IgG-coated Dynabeads (Life Technologies, Bleiswijk, the Netherlands) were coated with anti–ICAM-1 mAb (R&D Systems, Abingdon, U.K.), according to the manufacturer’s protocol. In short, beads were washed with ice-cold PBS (with 0.1% BSA, 2 mM EDTA) and incubated with 1.6 μg Ab for 45 min at 4°C under constant head-over-head rotation. The beads were subsequently washed twice with PBS (with 0.1% BSA, 2 mM EDTA) to remove any unbound Ab. Endothelial cells were seeded in six-well plates, stimulated with TNF-α, and Ab-coated magnetic beads were incubated for 30 min at 37°C. For subsequent immunoprecipitation, cells were gently washed in cold PBS (supplemented with 1 mM CaCl₂, 0.5 mM MgCl₂) to remove unbound beads and lysed in cold RIPA buffer (1% Nonidet P-40, 10% glycerol, 100 mM NaCl, 10 mM MgCl₂, 50 mM Tris, pH 7.4, 1% deoxycholate, and 0.1% SDS). Beads were extracted from cell lysates using a magnetic holder (PickPen 1M; BioNobile) and subsequently washed five times in a Nonidet P-40–based lysis buffer (1% Nonidet P-40, 10% glycerol, 100 mM NaCl2, 10 mM MgCl2, 50 mM Tris, pH 7.4). Finally, beads were resuspended in 30 μl SDS sample buffer for SDS-PAGE followed by Western blot analysis for ICAM-1 (rabbit polyclonal; Santa Cruz), filaminA (mouse monoclonal; Serotec, Puchheim, Germany), p-ezrin/radixin/moesin (p-ERM), and ezrin (Cell Signaling).

The samples were analyzed by 7.5% SDS-PAGE. Proteins were transferred to 0.45 μm nitrocellulose membrane (Schleicher and Schuell, Munich, Germany) and the blots were first blocked with 5% (w/v) BSA in TBST for 1 h, subsequently incubated at 4°C overnight with the appropriate Abs, followed by 30-min incubation with HRP-conjugated secondary Abs (DAKO, Glostrup, Denmark) at room temperature. Between the various incubation steps, the blots were washed five times with TBST and finally developed with an ECL detection system (GE Healthcare, Amersham, U.K.).

Results are shown as mean values with SEM. 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 post hoc Bonferroni correction. All statistical tests are described in the figure legends.

To investigate the regulation of ASM under inflammatory conditions, we exposed BECs to the inflammatory cytokine TNF-α and measured mRNA levels of ASM. The results showed that TNF-α significantly increased mRNA expression of ASM in BEC (p < 0.001; Fig. 1A). The expression of ASM by BEC was also studied in postmortem brain material from MS patients in active and chronic active lesions (Fig. 1B). Immunohistochemical analysis indicated a vascular staining pattern, although no significant difference in ASM expression comparing control subjects with MS patients with active or chronic active lesions was observed (Table I). Using ex vivo brain sections, we underscored the vascular localization of ASM by double-labeling studies of ASM with the vascular marker CD31 (Supplemental Fig. 1).

ASM has been shown to convert sphingomyelin into ceramide at the plasma membrane. To investigate whether ceramide generation at the plasma membrane occurs at sites where ICAM-1 is present, we used beads that were coated with anti–ICAM-1 Abs to mimic T cell–induced clustering of ICAM-1, a prerequisite for T cell transmigration (27). Upon attachment, we found ceramide to accumulate at the plasma membrane around anti–ICAM-1 Ab–coated beads, suggesting recruitment of ceramide to sites of bead-induced ICAM-1 clustering (Fig. 2A, Supplemental Fig. 2).

To further understand the importance of endothelial ceramide during T cell transmigration, we determined the effect of impaired ASM activity on T cell trafficking across the brain vasculature. Upon addition of the specific ASM inhibitor imipramine, a significant reduction of transendothelial migration of T cells was detected (p < 0.001; Fig. 2B). However, upon ASM blockade in the migration experiments, we could observe a decreased motility of the T cells (Supplemental Video 1). Therefore, to specifically assess the function of endothelial ASM in T cell trafficking, we reduced ASM expression in BECs using a lentiviral approach. The knockdown of ASM resulted in a decreased RNA (p < 0.001; Fig. 3A) and protein expression (p < 0.01; Fig. 3B) when compared with NTC. Furthermore, this downregulation was confirmed by a significant decrease in specific ASM activity in ASM-deficient cells (p < 0.01; Fig. 3C). Importantly, silencing of ASM did not affect the expression of neutral sphingomyelinase (Fig. 3D).

Subsequent transmigration experiments showed a significant decrease in transmigration of T cells across ASM-deficient endothelial cells when compared with control cells (p < 0.05; Fig. 4A). To address whether the decreased transmigration observed in ASM-deficient cells was due to a defect in cell adhesion, we performed adhesion experiments in both control and ASM-deficient cells. We measured a minimal but significant decrease in adhesion of immune cells to ASM-deficient endothelium (p < 0.05; Fig. 4B). Interestingly, TNF-α–induced expression of ICAM-1 and VCAM-1 (data not shown) was slightly but significantly increased in ASM-deficient endothelium (p < 0.01; Fig. 4C). To mimic T cell migration under physiological conditions, we performed transendothelial migration assays under physiological flow. Strikingly, the total number of T cells that adhered and transmigrated was significantly reduced in ASM-deficient cells (p < 0.05; Fig. 4D, Supplemental Video 2). In line with this observation, fewer transmigrated T cells also were observed (p < 0.01; Fig. 4E). However, from the fraction of T cells that is able to adhere under flow in the absence of ASM, the migration capacity is not affected (Fig. 4F). These data indicate that loss of ASM mainly affects the adhesion of T cells to the endothelium.

Several reports have indicated that ICAM-1–rich microvilli are involved in leukocyte adhesion and migration (29, 34, 35). Therefore, we analyzed the formation of microvilli in ASM-deficient BECs and control cells. Whereas under control conditions (NTC shRNA), TNF-α stimulation induced the formation of ICAM-1–rich F-actin⁺ microvilli, we observed that in ASM-deficient BECs (ASM shRNA), the formation of microvilli was reduced with a rather dispersed ICAM-1 localization (Fig. 5A, Supplemental Fig. 3). A significant reduction in microvilli-positive cells (p < 0.001; Fig. 5B) is observed as well in ASM-deficient BECs. In addition, if ASM-deficient cells express microvilli, the number of microvilli per cell is diminished as well (p < 0.01; Fig. 5C). These data demonstrated the importance of ASM for proper microvilli formation. Importantly, Oh and coworkers (34) showed that ezrin is required for the induction of these ICAM-1–rich microvilli structures. Depletion of ezrin perturbed the formation of ICAM-1–rich microvilli and reduced T cell adhesion. In addition, Kanters and coworkers (26) showed that ICAM-1 clustering resulted in the recruitment of the actin adapter filamin. Interestingly, the presence of endothelial filamin was shown to be required for ICAM-1–mediated T cell adhesion and transmigration (36). Therefore, we tested whether ASM may interfere with the association of ICAM-1 with filamin and ezrin-mediated ICAM-1–rich microvilli.

Clustering of ICAM-1 using the Ab-coated beads in ASM-deficient endothelial cells showed impaired recruitment of filamin to ICAM-1 compared with control endothelial cells (Fig. 6A). In addition, ICAM-1 clustering induced the phosphorylation of ezrin, in line with data reported by Oh and colleagues (34). Our data showed that ezrin that is precipitated in the ICAM-1 fraction is specifically phosphorylated. We did not observe increased phosphorylation of ezrin in the total cell lysate samples, indicating that only a specific ezrin pool is phosphorylated upon ICAM-1 clustering. However, in ASM-deficient endothelial cells, phosphorylation of this specific ezrin pool downstream from ICAM-1 clustering was reduced compared with the control cells (Fig. 6A).

We additionally tested whether the dynamics of ICAM-1 have changed in the absence of ASM activity. Therefore, we transfected ASM-deficient endothelial cells with ICAM-1–GFP and studied ICAM-1 mobility using FRAP. The results showed that there was a slight decrease in ICAM-1–GFP recovery in the absence of ASM activity, although this was not significant (Fig. 6B). These data indicate that ASM activity is required for proper ICAM-1 function to remodel the actin cytoskeleton to facilitate T cell adhesion and consequently transmigration.

Sphingolipids are becoming increasingly recognized for their involvement in neurodegeneration and neuroinflammation. However, the exact role of sphingolipids and, in particular, ASM in MS pathogenesis is still not fully understood. The development of MS lesions is largely dependent on immune cell infiltration into the brain. Upon adhesion of lymphocytes, the endothelial cells that line the lumen of brain vessels become activated and, through endothelial cell adhesion molecules such as ICAM-1 and VCAM-1, transmigrate across the blood–brain barrier. Although much is known about ICAM-1 function, not all processes concerning its regulation are fully understood. In this study, we demonstrate how proper ICAM-1 clustering in BECs is dependent on a functional ASM and its product ceramide.

We found that ASM was expressed by the inflamed brain vasculature. Next to our earlier reports on the presence of ASM in reactive astrocytes in MS lesions (14), ASM was also detected in the vasculature, although we did not detect a significant change in ASM expression compared with nonneurologic controls. This may be because of its secretion from the brain endothelium (37). We could observe that under inflammatory conditions, ASM activity in BEC supernatants is increased (M.A. Lopes Pinheiro, unpublished data), which may be due to its increased secretion. This result is in line with other studies that have demonstrated that endothelial cells are high secretors of ASM, both in naive and in inflamed conditions (37). However, it remains to be established whether the secreted form of ASM plays a role in membrane dynamics through the production of ceramide.

To further investigate the importance of ASM in ICAM-1 function, we first determined whether ceramide and ICAM-1 colocalize upon triggering. It is believed that ASM generates ceramide in the outer leaflet of the plasma membrane. Indeed, we observed that ICAM-1 colocalized with ceramide upon clustering. These results are in line with a previous study demonstrating the importance of ceramide for the endocytosis of ICAM-1–coated carriers (38). However, no functional relation has been made between ASM activity/ceramide and ICAM-1 function concerning cell migration. By blocking ASM via a selective inhibitor (imipramine) or through a lentiviral knockdown approach, we demonstrated that T cell migration was deficient in these conditions. Importantly, this was mainly due to reduced adhesion of T cells in ASM-deficient BECs. T cell diapedesis is highly dependent on adhesion molecules in the endothelium and it is well established that ICAM-1 plays a major role in this process (39). For proper intracellular signaling, ICAM-1 requires clustering and association to the actin cytoskeleton through actin adaptor proteins such as filamin (25). In this study, we demonstrate that the decreased adhesion and migration of T lymphocytes across ASM-deficient BECs is due to dysfunctional ICAM-1. Indeed, our data show that ASM-deficient BECs have reduced number of microvilli. Importantly, filamin recruitment to ICAM-1 is decreased in cells that lack ASM activity. Furthermore, ERM proteins, which are essential for microvilli and docking structure formation, were also less phosphorylated upon ICAM-1 clustering in ASM-deficient cells. Interestingly, it has been shown that phosphorylation of ERM proteins can be regulated by ASM and ceramide (40). In that study, increased ASM or ceramide was associated with increased ERM dephosphorylation, rendering these proteins to an inactive state. However, we observe that a decrease of ASM and ultimately ceramide does not change the total p-ERM pool in the cell, but are, in fact, less associated with ICAM-1. This discrepancy may be because of the type of cells used, because the effect of sphingolipids was analyzed in cancer cells, which are highly proliferative in comparison with confluent monolayers of BECs. Furthermore, we studied specific protein–protein interactions as opposed to membrane fractions, because these interactions are more informative and are related to their particular function.

Thus, based on these data, we put forward the idea that ASM activity controls proper ICAM-1 clustering to provide a perfect actin network across which T cells can adhere and transmigrate. If this delicate basis is affected, for example, by a deficient recruitment of filamin or ERM, this results in an improper association of proteins necessary for docking structure formation, ultimately slowing down or even inhibiting T cell transmigration.

Proper membrane structure is essential for the function of adhesion molecules such as VCAM-1 and ICAM-1. In fact, it has been proposed that ICAM-1 is already organized in the plasma membrane in endothelial adhesive platforms (41). In contrast, ceramide has been shown to displace cholesterol from rafts with a consequent alteration of its structure (42). Therefore, we believe that ASM-derived ceramide rearrange raft composition, possibly bringing together endothelial adhesive platforms for optimal ICAM-1 function. ASM would then be a modulator of raft dynamics and membrane structure via ceramide. Likewise ICAM-1, CD40, CD95, death receptor 5, FcγRII, and platelet-activating factor receptor are some examples of receptors for which activation is dependent on ceramide-enriched membrane platforms (16, 17, 19–22).

Our data implicate that ASM plays a role in the transmigration process of T cells into the CNS under inflammatory conditions, like seen during MS. Previous data revealed that inhibitors of ASM, such as fluoxetine (43), attenuate the development of a severe form of the mouse model of MS, experimental autoimmune encephalomyelitis (44). Similar phenotype was observed in ASM-deficient mice that develop a milder clinical course of experimental autoimmune encephalomyelitis together with decreased immune cell infiltrates in the spinal cord (45). This phenotype may be caused by the function of ASM in the release of cytotoxic granules by CD8⁺ T cells (46) or decreased inflammatory status in glial cells (14, 45). In line with our present findings, we postulate that ASM activity in the brain endothelium also contributes to the process of deficient transmigration capacity of immune cells into the brain.

In summary, we provide evidence for the involvement of ASM and ceramide in the process of T cell migration via ICAM-1 function. ICAM-1 triggering in the plasma membrane, possibly because of ligation to VLA-4 or LFA-1, activates ASM, where ceramide rearranges the lipid structure, providing the conditions for proper ICAM-1 clustering, association with actin-binding proteins, and intracellular signaling. The understanding of processes that mediate cell migration into the brain will open new avenues for the development of new therapeutic approaches to block this process.

We thank Dr. Jos van Rijssel for some technical support in ICAM-1 immunoprecipitation assays.

The authors have no financial conflicts of interest.