HEK-293T cells were obtained from the ATCC (CRL-11268) and human Jurkat leukemia CD4⁺ cells were kindly donated by Dr. J. Alcamí (Centro Nacional de Microbiología, Instituto de Salud Carlos III, Madrid, Spain). When needed, Jurkat cells lacking endogenous CXCR4 expression (Jurkat^-/-) (29) were transiently transfected with CXCR4-AcGFP (20 μg; JK^-/-X4) using a BioRad electroporator (20 × 10⁶ cells/400 μL RPMI 1640 with 10% fetal calf serum. 280V, 975 μF) and analyzed 24 hours later. Human peripheral blood mononuclear cells were isolated from buffy coats by centrifugation through Ficoll-Paque PLUS density gradients (GE Healthcare, Wakuesha, WI) at 760 × g for 30 minutes at room temperature (RT). They were then in vitro activated for 1 week with 20 U/mL of IL-2 (Teceleukin; Roche, Nutley, NJ) and 5 μg/mL phytohemagglutinin PHA (Roche) to generate T cell blasts (30).

The following antibodies were used: non-conjugated monoclonal anti-human CXCR4 (clone 44717), anti-CXCR4 (12G5) and anti-CD4 (OKT4) conjugated to PE and FITC accordingly, all from R&D Systems (Minneapolis, MN); anti-phospho-ERK1,2 (#9191), anti-ERK (#9102), anti-pAkt (#4060) and anti-Akt (#9272) from Cell Signaling Technology (Danvers, MA); phalloidin-TRITC (#P195) from Sigma-Merck (Darmstadt, Germany); and anti-GFP (JL-8) from Clontech Laboratories (Palo Alto, CA). Human CXCL12 and CCL19 were purchased from PeproTech (Rocky Hill, NJ). Human CXCR4 was cloned into pECFP-N1, pEYFP-N1 and pAcGFPm-N1 vectors (Clontech Laboratories), as described (8). Bacterial sphingomyelinase (bSMase, Staphylococcus aureus SMase, #S9396) was obtained from Sigma-Aldrich (Madrid, Spain) and the Di-4-ANEPPDHQ probe was from Invitrogen (Carlsbad, CA).

For lipid extraction from Jurkat and T cell blasts, cell pellets (1 x 10⁷ cells/pellet) were mixed with 200 μL of cold (-20°C) methanol:water (1:1, v/v) and sonicated with an ultrasonic homogenizer (UP200S, Hielscher Ultrasound Technology, HIELSCHER GmbH, Chamerau, Germany) for 16 bursts (0.5 second pulse) at 80% amplitude. Homogenates (100 μL) were mixed with 320 μL of cold (-20°C) methanol containing 1.6 ppm of sphinganine (d17:0) as the internal standard for positive-ion electrospray ionization (ESI) and 3.9 ppm of palmitic acid-d31 as the internal standard for negative-ion ESI. Samples were then vortex-mixed for 2 minutes, followed by the addition of 80 μL of methyl tert-butyl ether. Subsequently, samples were vortex-mixed (1 hour, RT). After centrifugation (16,000 × g, 15°C, 10 minutes), samples were used for ultra-high performance liquid chromatography (UHPLC; Agilent 1290 Infinity II, Agilent Technologies Inc., Santa Clara, CA) coupled with (ESI) quadrupole time-of-flight (QTOF) mass spectrometry (MS) (Agilent 6546): 100 μL of each sample was divided between two UHPLC-MS vials with inserts (50 μL/each) for direct injection into the system for LC-MS analyses in positive and negative ionization modes.

Quality control and blank samples were prepared for each cell population by pooling equal volumes of each sample. Samples were then randomized within each cell population, and quality controls were injected at the beginning, after every five experimental samples, and at the end of the batch. Blank samples were analyzed at the beginning and at the end of the analytical sequence. Finally, 10 iterative-MS/MS runs/group were performed for both ion modes at the end of the analytical run at 20 and 40 eV.

Raw LC-MS/MS data were reprocessed with Lipid Annotator software (Agilent Technologies). The MS/MS spectra of the 20 sphingomyelin and 20 ceramide candidates obtained were manually inspected using Agilent MassHunter Qualitative (v10.0), comparing the retention time and MS/MS fragmentation to the available spectral data across several databases included in the CEU Mass Mediator tool (CMM) (http://ceumass.eps.uspceu.es/mediator/) (31). MassHunter Profinder software (B.10.0.2, Agilent Technologies) was used to clean the raw LC-MS data from background noise and unrelated ions for feature building for the sphingomyelin and ceramide lipid species detected, and for feature alignment across the study samples (32). Data normalization was performed before statistical analysis. The raw data matrices were normalized according to the intensity of the corresponding internal standard to correct the unwanted variance related to sample preparation and the analytical run. Then, the coefficient of variation of the detected sphingomyelin and ceramide lipid species was evaluated in the quality controls, with all displaying a coefficient of variation <10%. The original and crude data from these experiments has been deposited at Metabolomics Workbench platform (https://www.metabolomicsworkbench.org) (access no. ST002150).

Cells (3 × 10⁶) were activated with CXCL12 (50 nM) at the indicated time points and then lysed in RIPA detergent buffer supplemented with 1 mM PMSF, 10 μg/mL aprotinin, 10 μg/mL leupeptin and 10 μM sodium orthovanadate for 30 minutes at 4°C. Extracts were analyzed by western blotting using specific antibodies. Densitometric evaluation of western blot experiments was performed using ImageJ software.

Cells were incubated with specific antibodies (30 minutes, 4°C) and mean fluorescence intensity was determined on a FC500 flow cytometer (Beckman Coulter Inc., Brea, CA). Receptor internalization was determined by flow cytometry after activation with CXCL12 (50 nM) at the indicated time points. Results are expressed as a percentage of the mean fluorescence intensity of treated cells relative to that of unstimulated cells.

Cultured cells were collected from microplates and washed twice in PBS buffer before staining with Annexin-V PE and 7AAD (Immunostep, Spain) following the manufacturer’s instructions or propidium iodide (3 μM, 30 minutes, RT) for analysis by flow cytometry (FC500, Beckman Coulter). Results are expressed as % of stained cells.

Cells (3 × 10⁵ in 0.1 mL of RPMI containing 10 mM HEPES and 0.1% BSA) were placed in the upper wells of uncoated 24-well transmigration chambers (3-μm pore or 5-μm pore; Transwell, Costar, Corning, NY). CXCL12 (12.5 nM) in 0.6 mL of the same medium was added to the lower well. Plates were incubated for 120 minutes (37°C, 5% CO₂) and cells that migrated to the lower chamber were counted by flow cytometry (FC500, Beckman Coulter), corrected for variations in input concentrations, and expressed as the mean (standard deviation, SD) percentage of cell migration. When required, cells were pretreated with bSMase (0.5 U/mL, 60 minutes, 37°C) and then maintained during cell migration.

Cells were seeded (30 minutes, 37°C) on Ibidi μ-well-chambers (Martinsried, Germany) coated with fibronectin (20 μg/mL, 30 minutes, 37°C, Sigma) or poly-L-lysine (20 μg/mL, 30 minutes, 37°C, Sigma). Adhered cells were untreated or treated with bSMase (0.5 U/mL, 60 minutes, 37°C) prior to imaging. The Di-4-ANEPPDHQ probe was added (5 μM) to wells just before imaging. Cells were imaged in phenol-free medium supplemented with 10 mM HEPES and 0.1% BSA, at 37°C and 5% CO₂.

Time-lapse images of cells were acquired on a Leica TCS SP5 inverted confocal microscope (Leica Microsystems, Wetzlar, Germany) fitted with an HCX PL APO 63×/1.2 NA water immersion objective. Di-4-ANEPPDHQ was excited using an argon white light laser at 488 nm. Emission signals were collected with two photomultiplier tubes (500–580, 620–750) and the pinhole was set to one Airy unit. Optimized acquisition was performed to retrieve membrane diffusion values as described (33, 34). Images of 256 × 256 pixels at 8-bit depth were collected using 80.4 nm pixel size and 4 μs dwell time, for 200 consecutive frames.

Raster image correlation spectroscopy (RICS) analysis was performed with “SimFCS 4” software (Global Software, G-SOFT Inc., Champaign, IL), as described (35). RICS analysis was performed in regions of interest of 64 × 64 pixels at 4 random cytoplasmic areas per cell using a moving average (background subtraction) of 10 to discard possible artefacts due to cellular motion and slow-moving particles passing through. The autocorrelation 2D map was then fitted to obtain a surface map that was represented as a 3D projection with the residuals on top. As a general rule, we focused on those regions with intensity fluctuation events in which the intensity changes were following short increasing or decreasing steps, avoiding abrupt intensity decays or increases.

Jurkat^-/- cells were transiently transfected with CXCR4 receptor fused to the AcGFP monomeric protein (JK^-/-X4). Twenty-four hours after transfection, cells expressing low receptor-AcGFP⁺ levels were selected by cell sorting on a BD FACSAria Fusion (BD Biosciences) platform for detection and tracking analysis. CXCR4 expression levels were then evaluated using the Dako Cytomation Qifikit and flow cytometry as described (36). Transfected cells expressing ~8,500–22,000 receptors/cell, which equates to a density of <4.5 particles/μm², were selected for detection and tracking analysis.

Experiments were performed with cells untreated or treated with bSMase (0.5 U/mL, 60 minutes, 37°C) using a TIRF microscope (Leica AM TIRF inverted) equipped with a Hamamatsu Flash 4 digital sCMOS camera (Hamamatsu Photonics Europe GmbH, Herrsching, Germany), a 100× oil-immersion objective (HCX PL APO 100×/1.47 NA) and a 488-nm diode laser. The microscope was equipped with an incubator and temperature control units; experiments were performed at 37°C with 5% CO₂. Image sequences of individual particles (500 frames) were then acquired at 3.5% laser power with a frame rate of 10 Hz (90 ms/frame). Penetration depth of the evanescent field was 90 nm.

Particles were detected and tracked using described algorithms (U-Track2 (37);) implemented in MATLAB, as described (8). Mean spot intensity (MSI), number of mobile and immobile particles and diffusion coefficients (D_1:4) were calculated from the analysis of thousands of single trajectories over multiple cells (statistics provided in the respective figure captions), using described routines (38). Receptor number along individual trajectories was determined as described (8). We measured the average fluorescence intensity for the first 20 frames of each trajectory and used the intensity of the monomeric protein CD86-AcGFP as a reference in Jurkat^-/- cells transiently transfected with CD86-AcGFP. Distribution of monomeric particles intensities was analyzed by Gaussian fitting, rendering a mean value of 69.33 ± 3.26 a.u. A similar intensity value of the monomer was obtained for CXCR4-AcGFP particles showing a unique photobleaching step. Therefore, this value was used as the monomer reference to estimate the number of CXCR4-AcGFP molecules per particle ( Supplemental Figures 1A, B ).

HEK-293T cells transiently transfected at a fixed CXCR4-YFP : CXCR4-CFP ratio (15 μg and 9 μg, respectively) were treated with bSMase (0.5 U/mL, 60 minutes, 37°C) or its solvent (50 mM Tris-HCl; pH 7.5) as a control, and FRET efficiency was evaluated (n = 3, mean ± SEM, ∗∗∗p ≤ 0.001). Emission light was quantified using the Wallac Envision 2104 Multilabel Reader (Perkin Elmer, Foster City, CA) equipped with a high-energy xenon flash lamp (donor: receptor fused to C-CFP, 8-nm bandwidth excitation filter at 405 nm; acceptor: receptor fused to YFP, 10 nm bandwidth excitation filter at 510 nm).

Chambers (Ibidi μ−Slide Chemotaxis System; 80326) were first coated with fibronectin (20 μg/mL, 60 minutes, 37°C). Subsequently, untreated or bSMase-treated (0.5 U/mL, 60 minutes, 37°C) cells were diluted to 10 × 10⁶ cell/mL in RPMI medium containing 1% BSA and 20 mM HEPES (chemotaxis medium), seeded into the channel of the chemotaxis slide, and cultured (60 minutes, 37°C). The reservoirs were then filled with chemotaxis medium and 50 nM CXCL12 was added to the right reservoir. Phase-contrast images were recorded over 20 hours with a time lapse of 2 minutes using a Microfluor inverted microscope with a 10× objective (Leica) and equipped with an incubation system set to 5% CO₂ and 37°C. Single-cell tracking was evaluated by selecting the center of mass in each frame using the manual tracking plug-in tool in ImageJ (NIH, Bethesda, MD). Spider plots, representing the trajectories of the tracked cells, forward migration index (FMI), and straightness values were obtained using the chemotaxis and migration plug-in tool (https://ibidi.com/chemotaxis-analysis/171-chemotaxis-and-migration-tool.html, Ibidi).

T cell blasts, treated or not with bSMase (0.5 U/mL, 3 hours, 37°C, 5% CO₂), were plated on fibronectin (20 μg/mL, Sigma)-coated glass slides and were then stimulated or not with 100 nM CXCL12 or CCL19 (5 minutes at 37°C) and fixed with 4% paraformaldehyde (10 min, RT). Preparations were blocked with PBS containing 150 mM NaCl, 0.1% goat serum and 1% BSA (60 minutes, RT) before staining with anti-human ICAM-3 or anti-CXCR4 (44717) plus AlexaFluor 488 goat anti-mouse IgG (30 minutes, RT, Thermo Fisher Scientific). Cells were permeabilized with 0.25% saponin (10 minutes, RT) and stained with phalloidin-TRITC (Sigma-Merck; 30 minutes, RT). Preparations were analyzed using a ZEISS confocal multispectral microscope.

Data were analyzed using GraphPad PRISM (ns = not significant p>0.05; *p ≤ 0.05; ** p ≤ 0.01; *** p ≤ 0.001; **** p ≤ 0.0001). Data from experiments of cell migration, directional cell migration assays, RICS, and MSI, were analyzed using one-way ANOVA followed by Tukey’s multiple comparison test. A two-tailed Mann-Whitney non-parametric test was used to analyze the diffusion coefficient (D1-4) of single particles and FRET efficiency. We used contingency tables to compare two or more groups of categorical variables, such as the percentages of mobile or immobile particles, and these were compared using the Chi-square test with a two-tailed p-value. For lipidomic analysis, statistics were analyzed using Matlab (R2018a, MathWorks), by the Mann-Whitney U test (p ≤ 0.05) after normality testing with the Shapiro-Wilk test. The false discovery rate at the level α = 0.05 was inspected by the Benjamini–Hochberg correction test.

To investigate how sphingomyelin ablation affects CXCR4 dynamics at the cell membrane, we treated Jurkat cells and primary T cell blasts for prolonged periods with bSMase (0.5 U/ml, 3 hours, 37°C, 5% CO₂), and examined its effect on cell migration using Boyden chambers. Results showed a significant reduction of CXCL12-triggered migration in both cell types after bSMase treatment ( Figures 1A, B ). bSMase treatment was not toxic to cells, as evaluated by propidium iodide and annexin-V/7AAD staining and flow cytometry ( Supplemental Figure 2 ). Lipidomic analysis of cells using UHPLC-ESI-QTOF MS and MS/MS, in both positive and negative ion mode, demonstrated the complete breakdown of sphingomyelins under these experimental conditions, and the corresponding accumulation of derived ceramides: 18:0/15:0, 18:1/16:0, 18:1/17:0 and 18:2/24:1 ( Figures 2A, B ; Supplemental Figures 3A, B ). These results indicate that prolonged bSMase treatment promotes sphingomyelin depletion in Jurkat cells and T cell blasts with an accumulation of saturated and unsaturated long chain ceramides and suppressed CXCL12-mediated cell chemotaxis.

Cholesterol interacts with the saturated acyl chains of sphingolipids and phospholipids, conferring specific biophysical properties that increase cohesion and packing of neighboring lipids and proteins. We determined whether the accumulation of ceramides affected the conformation of CXCR4 at the cell membrane. Untreated and bSMase-treated JK cells and T cell blasts were stained with two distinct anti-CXCR4 conformational mAbs (12G5 and 44717 mAbs) (39) and their binding was analyzed by flow cytometry. Results indicated that bSMase treatment reduced 12G5 mAb binding, whereas the binding of 44717 mAb remained unaltered ( Figures 3A, B ). These data suggest that bSMase treatment did not alter CXCR4 levels at the cell membrane but promoted a conformational change on the receptor affecting the epitope recognized by 12G5 mAb. In addition, we performed FRET experiments on untreated or bSMase-treated (0.5 U/ml, 3 hours, 37°C, 5% CO₂), HEK-293T cells transiently cotransfected with CXCR4-CFP and CXCR4-YFP. Results showed that FRET efficiency was higher in cells treated with bSMase than in untreated cells ( Figure 3C ), confirming the notion that aberrant ceramide accumulation at the cell membrane modifies the conformation of CXCR4 dimers that might be due to an increase of the number of CXCR4 complexes at the cell surface, thus corroborating a relationship between CXCR4 and ceramide enriched-platforms.

To further establish that sphingomyelin depletion per se does not affect CXCR4 levels at the cell membrane, we examined the subcellular distribution of transfected CXCR4-AcGFP in Jurkat-/- cells untreated or bSMase-treated (0.5 U/ml, 60 minutes, 37°C, 5% CO₂) prior to lysis. Western blotting of subcellular fractions indicated that the bulk of CXCR4 was present in the membrane fraction in both untreated and bSMase-treated cells ( Figure 3D , left). Analysis of the distribution of IKB (cytosolic protein) and the transferrin receptor (membrane protein) confirmed no cross-contamination of fractions ( Figure 3D , right). Overall, the data suggest that ceramide accumulation does not modify the levels of CXCR4 at the cell membrane, but alters its conformation.

Ceramides influence membrane fluidity and regulate cell migration (40, 41). The effects of ceramide on membrane fluidity are, nonetheless, complex and depend on several factors such as the acyl chain length (42), saturation (43), and the ratio of long chain and very-long chain species present in the cell membrane (44). We thus used RICS with the Di-4-ANEPPDHQ lipid probe to evaluate membrane fluidity by means of membrane diffusion. This lipophilic dye is also a reporter for lipid lateral packing (33, 34). Jurkat cells ( Figures 4A, C ) and T cell blasts ( Figures 4B, D ) untreated or treated with bSMase (0.5 U/ml, 3 hours, 37°C, 5% CO₂) were labeled with Di-4-ANEPPDHQ and RICS was performed after confocal microscopy. Results indicated that membrane diffusion was higher in cells treated with bSMase than in untreated controls ( Figure 4 ), suggesting a more fluid environment. The increase in unsaturated ceramides (C18:1 and C18:2) detected in the cells as a consequence of bSMase treatment ( Figure 2A ) might explain the evident higher membrane diffusion (43). To analyze the influence of the aberrant ceramide accumulation on CXCR4 clustering and lateral diffusion, we next generated SPT trajectories of CXCR4-AcGFP using TIRF microscopy in transiently-transfected Jurkat^-/- cells (JK^-/-X4) (29), untreated or treated with bSMase. We first determined appropriate expression conditions for detecting and tracking individual CXCR4 spots, as described (8). Analysis of receptor trajectories (37), indicated that bSMase treatment had no effect on CXCR4 dynamics under steady-state conditions. A high proportion of CXCR4 particles were mobile (~82% in untreated cells vs ~85% in bSMase-treated cells) in both bSMase-treated and untreated JK^-/-X4 cells ( Figure 5A ). The median value of the short time-lag diffusion coefficient (D_1-4) for CXCR4 trajectories was 0.017 μm²/s in untreated cells and was slightly slower (0.016 μm²/s) in bSMase-treated cells ( Figure 5B ). However, whereas CXCL12 promoted a significant reduction in overall receptor diffusivity in control cells (CXCL12, median D_1-4 = 0.007 μm² s^–1) and increased the percentage of immobile particles from ~17% (basal) to ~30% (CXCL12), it had no effect on bSMase-treated cells (CXCL12, median D_1-4 = 0.016 μm² s^-1), with ~15% of immobile particles (basal) and ~18% after CXCL12 stimulation ( Figure 5A ). To determine the receptor number in individual trajectories, we measured the average fluorescence intensity for the first 20 frames of each trajectory and used the intensity of the monomeric protein CD86-AcGFP as reference (45) ( Supplemental Figures 1A,B ). In steady-state, we found predominantly monomers and dimers in both type of cells (~65% for control vs ~62% for bSMase-treated cells), with a similar percentage of complexes of ≥3 receptors (~35% vs ~37%, respectively) ( Figures 5C, D ). Accordingly, basal MSI for CXCR4 was comparable in both cases (135.18 a.u. in control vs 156.99 a.u. in bSMase-treated cells). As previously shown (8), CXCL12 activation promotes CXCR4 membrane nanoclustering in control cells (~60% of nanoclusters of ≥3 receptors), but it failed to do so in bSMase-treated cells (~37%) ( Figures 5C, D ). These findings indicate that the lipid equilibrium at the cell membrane is essential for the spatiotemporal regulation of CXCR4 nanoclustering and dynamics.

Because ceramide-rich platforms are implicated in a variety of signaling cascades in immune cells, including those related to chemokine receptors (46), we next evaluated whether the effect of bSMase treatment on CXCR4 nanoclustering and dynamics influences CXCR4-mediated functions. Possible differences in CXCR4 internalization due to bSMase treatment were discarded by flow cytometry analysis of Jurkat cells stained with an anti-CXCR4 monoclonal antibody ( Figure 6A ), which showed that the receptor could still bind CXCL12 and activate some cell functions independently of the effects of ceramide accumulation on CXCR4 dynamics. We also observed that CXCL12-triggered Ca²⁺ flux was similar in both untreated and bSMase-treated Jurkat cells ( Figure 6B ) as was CXCL12-mediated Akt phosphorylation ( Figures 6C, D ). We also detected that ERK1/2 phosphorylation, although slightly reduced was not completely impaired by bSMase treatment ( Figures 6C, D ). Altogether these data indicate that sphingomyelin depletion blocked some signaling pathways whereas other were still active.

Because the absence of CXCR4 nanoclustering has been related to defective directed cell migration (29), we analyzed ligand-mediated directed cell migration on fibronectin-coated μ-chemotaxis chambers. Results showed that whereas untreated Jurkat cells sensed the CXCL12 gradient, bSMase-treated Jurkat cells did not ( Figure 7A ; Supplemental Videos 1 , 2 ). Quantification of the results indicated that, compared with untreated Jurkat cells, CXCL12 failed to increase the forward migration index and track straightness in bSMase-treated Jurkat cells ( Figure 7B ). These data further show that while some chemokine-mediated signaling events remain operative in cells with aberrant accumulation of ceramides, directed cell migration does not.

To evaluate whether these changes also occur in primary T cells, we reproduced the prolonged bSMase treatment (0.5 U/ml, 3 hours, 37°C, 5% CO₂) on T cell blasts. We confirmed the complete breakdown of their sphingomyelins and the accumulation of the corresponding ceramides ( Figure 2B ; Supplemental Figure 3B ). We also observed a defect in their migratory behavior ( Figure 1B ) and RICS analysis of Di-4-ANEPPDHQ indicated higher membrane diffusion in bSMase-treated T cell blasts than in controls ( Figure 4D ). As we expected, bSMase treatment abrogated the ability of CXCL12 to promote cell polarization in T cell blasts, as shown by ICAM-3 and Phalloidin staining ( Supplemental Figures 4A–C ). The same effect was observed in cells treated with CCL19, indicating that bSMase treatment also diminishes CCR7-mediated cell polarization, as previously shown (26) ( Supplemental Figure 5 ). In similar experiments we observed that CXCR4 redistribution to the leading edge in response to CXCL12 or CCL19 was also altered upon pre-treatment with bSMase ( Supplemental Figure 5 ). We then determined CXCL12-mediated directed cell migration on fibronectin-coated chemotaxis chambers and observed that whereas untreated T cell blasts sensed a CXCL12 gradient, bSMase-treated cells did not ( Figure 7C ; Supplemental Video 3 , 4 ). Under these conditions, CXCL12 failed to increase the forward migration index and track straightness ( Figure 7D ). Overall, these data indicate that the aberrant accumulation of ceramides also abrogate the ability of primary T cell blasts to sense CXCL12 gradients.