CD4 T cell sphingosine 1-phosphate receptor (S1PR)1 and S1PR4 and endothelial S1PR2 regulate afferent lymphatic migration

See allHide authors and affiliations

Science Immunology  15 Mar 2019:
Vol. 4, Issue 33, eaav1263
DOI: 10.1126/sciimmunol.aav1263

Getting immune cells home

Immune cells are recruited to tissues from the lymphatic system via efferent lymphatics in response to insults ranging from skin allergies to flu infection. Immune cells recirculate back to draining lymph nodes via afferent lymphatics. Using a combination of in vitro systems and in vivo studies, Xiong et al. have examined the role of sphingosine 1-phosphate (S1P) and its receptors (S1PRs) in the egress of T cells from tissues. They found expression of S1PR1 and S1PR4 on T cells and engagement of S1PR2 on lymphatic endothelial cells to be important for this process. Whereas the role of S1PRs in recruitment of lymphocytes to tissues is well established, the current study extends their importance to recirculation of immune cells back to the lymphatic system.


Sphingosine 1-phosphate (S1P) and S1P receptors (S1PRs) regulate migration of lymphocytes out of thymus to blood and lymph nodes (LNs) to efferent lymph, whereas their role in other tissue sites is not known. Here, we investigated the question of how these molecules regulate leukocyte migration from tissues through afferent lymphatics to draining LNs (dLNs). S1P, but not other chemokines, selectively enhanced human and murine CD4 T cell migration across lymphatic endothelial cells (LECs). T cell S1PR1 and S1PR4, and LEC S1PR2, were required for migration across LECs and into lymphatic vessels and dLNs. S1PR1 and S1PR4 differentially regulated T cell motility and vascular cell adhesion molecule–1 (VCAM-1) binding. S1PR2 regulated LEC layer structure, permeability, and expression of the junction molecules VE-cadherin, occludin, and zonulin-1 through the ERK pathway. S1PR2 facilitated T cell transcellular migration through VCAM-1 expression and recruitment of T cells to LEC migration sites. These results demonstrated distinct roles for S1PRs in comodulating T cell and LEC functions in migration and suggest previously unknown levels of regulation of leukocytes and endothelial cells during homeostasis and immunity.


Sphingosine 1-phosphate (S1P) controls T cell migration from thymus to the blood across microvascular endothelium and egress from lymph node (LN) to lymphatics across lymphatic endothelium. These activities are primarily mediated by S1P receptor 1 (S1PR1) (1). The S1P/S1PR1 axis acts on T cells as a signal to leave the LN (2, 3) and on endothelial cells to alter barrier function (4, 5). Although less is known about the use of S1P for migration in peripheral tissues, we previously showed that S1P/S1PR1 acted as a stop signal for T cells, but the effects of S1P gradients were not evaluated (6).

S1P regulates endothelial cell homeostasis (710) and barrier function (11, 12), which are important for leukocyte trafficking. Therefore, unlike the responses to traditional chemokines where only leukocytes express the cognate receptors, both cell types express one or more receptors for S1P, making it possible for S1P-driven migration to be regulated differently. The variability in receptor expression and utilization suggests additional levels of complexity in the regulation of migration.

There are five G protein–coupled receptor (GPCR) S1PRs: S1PR1, S1PR3, S1PR4, and S1PR5 are linked to Gi, whereas S1PR2 preferentially signals via G12/13 (13). Pertussis toxin (PTX) inhibits S1PR1, S1PR3, S1PR4, and S1PR5 through Gi but does not inhibit G12/13 or S1PR2. The nonspecific S1PR antagonist FTY720 binds S1PR1, S1PR3, S1PR4, and S1PR5 with much higher affinity than S1PR2 and has no inhibitory effects on S1PR2 (4). S1PRs regulate diverse leukocyte activities. S1PR1 directs B cells to the splenic marginal zone (14) and controls immature B cell egress from bone marrow (15). S1PR1 promotes human B cell migration, which is in turn modulated by S1PR4 and S1PR2 (16). S1P regulates macrophage entry and egress from LN (17). Mature dendritic cells (DCs) migrate to S1P (18), and CD69 modulates S1P-induced migration from skin to draining LNs (dLNs) (19). S1PRs regulate type 2 innate lymphoid cells (ILC2) entrance into lymphatic vessels and egress from LN (20).

Here, we looked at the roles of S1PRs in T cells and in lymphatic endothelial cells (LECs) and showed that T cells responded to S1P gradients through S1PR1 and S1PR4 to migrate across afferent LECs. S1PR1 and S1PR4 had distinct roles in T cell motility and binding to vascular cell adhesion molecule–1 (VCAM-1). The T cell–LEC interaction engaged LEC S1PR2–dependent processes to promote T cell transcellular migration, which was distinct from chemokine-driven migration. S1PR2 signaled through extracellular signal–regulated kinase (ERK) to regulate lymphatic permeability and LEC surface and junction proteins. These results demonstrated that trans-lymphatic endothelial migration (TEM) relies on several receptors with integrated process of both T cell and LEC responses to a common ligand.


S1P selectively promotes trans-endothelial migration

We previously established a TEM assay in which primary murine LECs or the mouse endothelial cell line SVEC was seeded on the lower surface of a transwell insert (designated as iLEC or iSVEC) (fig. S1A), allowing establishment of junctions and polarity (21). Leukocytes loaded in the upper chamber migrated across the LEC layer from the basal (or abluminal) to the apical (or luminal) direction. Migration only proceeded in the basal-to-apical direction, recapitulating directional migration in vivo.

Only low numbers of T cells migrated across transwell plastic membranes to S1P (fig. S1B) (22). Using the TEM model, we found that far more CD4 T cells migrated across primary LECs and the SVEC line in a dose-dependent fashion (Fig. 1, A to C). However, CD4 T cell TEM to C-C chemokine ligand 19 (CCL19) or other chemokines and cytokines was not enhanced compared with plastic (Fig. 1, A and B, and fig. S1, C to H). LECs promoted migration toward S1P of various mouse CD4 T cell subsets, including memory (Tmem) and activated effector cells (Teff) (fig. S1, I to K). Human effector T cells and regulatory T cells (Treg) also migrated more across human iLECs than plastic in response to S1P, but not CCL19 (Fig. 1D and fig. S1, L and M). These results suggested that S1P-driven TEM involved actively stimulated interactions on both T cells and LECs, in contrast to chemokine-driven migration, where ligands engaged cognate receptors only on T cells.

Fig. 1 LECs specifically promote CD4 T cell TEM to S1P but not to other chemokines.

Murine CD4 migration: (A) Across plastic or skin or LN-derived iLECs or LECs to S1P or CCL19; (B) across plastic, iSVECs, or SVECs to medium, 100 nM S1P, or 53 nM CCL19; (C) across plastic or iSVECs to varying doses of S1P. (D) Migration of human Teff across plastic, human iLECs or human LECs to CCL19 or S1P. (E) Naïve murine CD4 migration toward S1P or CCL19 across iSVECs, with anti-CCL21 loaded in the upper chamber. (F) Naïve murine CD4 migration across iSVECs toward S1P with various doses of anti-S1P in the upper chamber. (G and H) Individual naïve CD4 migration toward S1P and CCL19, showing velocity (G) and migration time (H) of individual cells. At least three independent experiments, triplicate wells. Data presented as mean ± SEM. One-way ANOVA for multiple comparisons. *P < 0.05; ns, not significant.

When DCs cross afferent LECs, the interaction results in CCL21 secretion by the LECs to enhance entry into the lymphatics (2325). In our model, CD4 T cell TEM to CCL19 was significantly inhibited by anti-CCL21 (Fig. 1E), suggesting that LECs secreted CCL21 in response to CD4 T cells. Primary LECs and SVECs, along with afferent lymphatics, all constitutively expressed CCL21 (fig. S2, A and B). In contrast, S1P-driven migration was not inhibited by anti-CCL21, suggesting that migration mechanisms and their regulation by LECs were distinct from CCL21-driven migration (Fig. 1E).

S1P may stimulate LECs to express other chemotactic molecules (26). SVECs were treated with S1P and supernatants were collected separately from upper and lower chambers. These supernatants were applied to the upper and lower chambers, respectively, of a second TEM assay. Anti-S1P monoclonal antibody (mAb) was also added to supernatants to exclude an effect of the S1P added to the primary cultures. Neither upper-chamber nor lower-chamber supernatants enhanced migration (fig. S2C). To determine whether S1P induced chemotactic factors that remained bound to the endothelial cells, we pretreated SVECs with S1P and washed them, and T cells were added. S1P pretreatment of SVEC slightly, although not significantly, enhanced migration (fig. S2D). Last, anti-S1P added to the upper chamber inhibited TEM in a dose-dependent fashion (Fig. 1F). Together, these results excluded the possibility that enhanced T cell migration was due to additional chemotactic factors produced by LECs after S1P stimulation.

Fluorescence microscopy was used to monitor T cell real-time movements across LEC layers. The velocity and transit times of migration toward S1P and CCL19 were similar (Fig. 1, G and H), showing that once CD4 T cells engaged LEC in the presence of S1P, the T-LEC interactions resulted in kinetics similar to those of chemokine-driven T cell migration.

Trans-endothelial migration toward S1P is chemotactic, chemokinetic, and responsive to flow and inflammation

S1P loaded in the lower chamber or in both chambers, but not the upper chamber alone, resulted in enhanced TEM (Fig. 2A). Thus, S1P-driven TEM was chemokinetic and chemotactic, but not chemorepulsive (Fig. 2A). Real-time migration over LECs was measured under these conditions (Fig. 2B). T cells migrated with the same displacement for all groups, but migrated with higher velocities when S1P was in the lower chamber (chemotaxis) or in both chambers (chemokinesis) (Fig. 2, C and D). S1P induced a change in the velocity distribution; thus, more CD4 T cells had higher velocities (Fig. 2E). These results suggested that chemotactic and chemokinetic responses engaged similar migration mechanisms.

Fig. 2 S1P induces T cell chemotactic and chemokinetic TEM regulated by fluid flow and inflammation.

(A) Naïve CD4 migration across iSVECs; S1P loaded in different chambers as indicated. (B to E) Real-time imaging for naïve CD4 chemokinetic or chemotactic migration toward S1P loaded in the lower chamber or in both chambers, respectively. (B) Tracking for individual cells, (C) velocity, (D) migration times, and (E) velocity distribution. (F) Naïve CD4 migration across iSVECs to CCL19 or S1P under no-flow or flow conditions. (G) Migration of naive CD4 pretreated with or without anti–VLA-4 to S1P across iSVECs pretreated with anti–VCAM-1 and/or TNFα 3 hours before anti–VCAM-1. For (A) to (G), three independent experiments with triplicate wells, presented as mean ± SEM; one-way ANOVA for multiple comparisons. (H and I) Footpad migration with or without LPS treatment. One microgram of LPS was injected per footpad; 2 hours later, CD4 T cells without treatment (H) or mixed with control IgG or anti–VCAM-1 (I) were transferred into footpads, and migration to dLN was assessed after 16 hours (n = 6). For (H) and (I), statistical analyses, paired Student’s t test, *P < 0.05.

Trans-lymphatic fluid flow affects the phenotype and function of lymphatic monolayers (27, 28). We previously mimicked interstitial fluid flow by creating a pressure differential of 0.8 to 0.9 cm H2O that resulted in an average flow of 15.8 μl/hour (27, 28). The flow enhanced TEM toward S1P and CCL19 (Fig. 2F), suggesting that flow characteristics for both ligands were similar.

VLA-4–VCAM-1 interactions are important for chemokine-driven TEM, and migration could be partly inhibited with blocking mAbs (27). VCAM-1 expression on LECs was enhanced by tumor necrosis factor–α (TNFα), which increased TEM (6) (29). Similarly, TNFα enhanced S1P-driven TEM (Fig. 2G). Blocking anti–VLA-4 or anti–VCAM-1 mAbs partly inhibited migration (Fig. 2G), suggesting VLA-4–VCAM-1–dependent and –independent TEM, as noted for chemokine-driven migration. Overall, these results showed that S1P-driven migration shared several characteristics with chemokine-driven migration.

After inflammation induced by lipopolysaccharide (LPS) treatment of footpads, CD4 T cell migration into dLN was increased (Fig. 2H), and migration was blocked by anti–VCAM-1 mAbs (Fig. 2I). Thus, CD4 T cell homing into dLN under both homeostatic and inflammatory conditions was dependent on VLA-4–VCAM-1 interactions and S1P (Fig. 4, B and C).

Expression of S1PR1 and S1PR4 on CD4 T cells is required for migration into lymphatics

S1PR1 is linked to Gi, S1PR3 to S1PR5 are linked also to Gq, and S1PR2 is linked to G12/13, resulting in differential sensitivity to receptor and G protein antagonists (13). Reverse transcription polymerase chain reaction (RT-PCR) (fig. S2E) and Western blotting (fig. S2F) confirmed that primary LECs and SVECs expressed S1PR1 and S1PR2 (30). Naïve CD4 T cells Teff and Tmem expressed S1PR1 and S1PR4 (fig. S2, E and G to I) (13, 31).

Gi-coupled S1PR1, S1PR3, S1PR4, and S1PR5 are inhibited by PTX, and FTY720 induces functional antagonism of S1PR1 but not S1PR2. CD4 T cells pretreated with PTX or FTY720 had significantly decreased migration (Fig. 3A and fig. S3A), suggesting that T cells use S1PR1 and/or S1PR4 for TEM. Using specific pharmacologic inhibitors, treatment of T cells with a reversible S1PR1 antagonist (32) and an irreversible S1PR4 antagonist (33) inhibited migration (Fig. 3, B and C, and fig. S3, B and C). These inhibitors did not alter T cell viability (fig. S3D). We concluded that T cells used both S1PR1 and S1PR4 to respond to S1P-driven TEM.

Fig. 3 T cell S1PR1 and S1PR4 regulate TEM.

(A) iSVECs or CD4 treated with PTX, FTY720. (B) CD4 treated with S1PR1 antagonist. (C) iSVECs or CD4 treated with S1PR4 antagonist. (D) S1PR1−/− or S1PR1+/+ [wild type (WT)]. (E) S1PR4−/− or S1PR4+/+, CD4 treated with S1PR1 or S1PR4 antagonists, migrated to S1P or CCL19. (F) S1PR1-Tg or (G) S1PR1 S5A CD4 treated with S1PR1 antagonist, migrated to S1P or CCL19. (H) CD4 treated with S1PR1, S1PR4, or combined antagonists; iSVEC; and S1P in the lower chamber or in both chambers. (I to M) Naïve CD4 footpad migration: (I) CD4 pretreated with S1PR4 antagonist (n = 12); (J) CD4 pretreated with S1PR1 antagonist (n = 12); (K) S1PR1+/+ or S1PR1−/− CD4 transferred to WT recipients (n = 6); (L) S1PR4+/+ or S1PR4−/− CD4 to WT recipients (n = 12); (M) S1PR1 WT or S5A CD4 to WT recipients (n = 10). (N to S) CD4 Teff or Tmem across iLEC: (N) Teff or (O) Tmem treated with S1PR1, S1PR4, or combined antagonists; S1PR1−/− or S1PR1+/+ (P) Teff or (Q) Tmem; S1PR4−/− or S1PR4+/+ (R) Teff or (S) Tmem. (T and U) Teff footpad migration: (T) Teff pretreated with S1PR4 antagonist (n = 7); or (U) S1PR1 antagonist (n = 6). In paired analysis footpad migration, T cells were transferred to both footpads, with one side treated with inhibitor and the other treated with control. (A) to (H) and (N) to (S), mean ± SEM, one-way ANOVA for multiple comparisons. (I) to (M) and (T) to (U), paired Student’s t test.

To confirm these findings, we assessed tamoxifen-induced S1PR1−/− conditional deletion and S1PR4−/− germline deletion T cells. Deletion of either receptor resulted in poor migration responses to S1P while migration to CCL19 remained intact (Fig. 3, D and E). The S1PR1 and S1PR4 antagonists were also inactive on the receptor-deficient cells. These results confirmed the findings with the pharmacologic inhibitors and ruled out off-target or nonspecific effects. T cells from S1PR1 transgenic (Tg) (Fig. 3F) or S5A (Fig. 3G) mice, in which there were higher levels of cell surface S1PR1 (7, 34), had enhanced CD4 T cell migration toward S1P, but not CCL19 (Fig. 3, F and G), showing the critical function of S1PR1 in TEM. T cells were pretreated with antagonists, and migration to S1P loaded into the lower chamber or into both chambers was assessed. The results showed that both S1PR1 and S1PR4 antagonists inhibited chemotaxis and chemokinesis (Fig. 3H). Thus, S1PR1 and S1PR4 regulated both types of movements and were additive for chemotaxis.

To assess pharmacologic blockade in vivo, we pretreated T cells with inhibitors and transferred them to hind footpads, and we assessed migration to popliteal dLN. The S1PR4 antagonist inhibited migration (Fig. 3I) while there was a trend for migration inhibition with the S1PR1 antagonist (Fig. 3J); lack of significance was likely due to the reversibility of this compound. The combination of S1PR1 and S1PR4 antagonists inhibited migration more completely (fig. S3E). S1PR1−/− and S1PR4−/− T cells migrated less to dLN compared with wild-type CD4 T cells (Fig. 3, K and L). In contrast, S1PR1 S5A CD4 T cells migrated more into dLN compared with wild type (Fig. 3M).

Tmem and Teff are the major subsets that migrate from tissues into dLN through afferent lymphatics (35, 36). Similar to naïve T cells, pretreatment of Teff (Fig. 3N) and Tmem (Fig. 3O) with S1PR1 and S1PR4 antagonists inhibited migration in vitro. S1PR1−/− Teff (Fig. 3P) and Tmem (Fig. 3Q) and S1PR4−/− Teff (Fig. 3R) and Tmem (Fig. 3S) migrated less. In vivo, S1PR1 (Fig. 3T) and S1PR4 (Fig. 3U) antagonists inhibited Teff migration to dLN. Overall, S1PR1 and S1PR4 were required for homing of various T cell subsets into dLN through afferent lymphatics.

S1P gradients and expression of S1PR2 on LECs are required for trans-endothelial migration

Whole-mount staining of ear pinnae showed that S1P was present around the afferent lymphatics (Fig. 4A), revealing proximity of S1P to the lymphatics, and that S1P distribution appears to establish a gradient. S1P-regulated chemotaxis was confirmed because anti-S1P mAb administered to the footpad inhibited migration into dLN under homeostatic and inflammatory conditions (Fig. 4, B and C). T cells migrated less into dLN of sphingosine kinase 1 (Sphk1)−/− and Sphk2−/− mice (Fig. 4D) deficient in S1P production (37). Ear whole mounts showed that S1P expression was diminished around afferent lymphatics of Sphk1−/− and Sphk2−/− strains (fig. S3F). Lyve-1 and CCL21 expression in Sphk1−/− and Sphk2−/− lymphatics did not differ from wild type (fig. S3F), and the morphology and density of lymphatic vessels appeared normal. Preincubation of anti-S1P mAb with S1P abolished staining, showing additional specificity (fig. S3G). Thus, morphologic alterations did not account for changes in migration.

Fig. 4 S1P gradient and LEC S1PR2 regulate T cell TEM.

(A) Whole-mount staining, S1P in lymphatics; 100×; scale bars, 8 μm (two ears per treatment per experiment, n = 8). LV, lymphatic vessel. (B) Naïve CD4, footpad migration, anti-S1P treatment (n = 11). (C) Naïve CD4, footpad (treated with LPS) migration, anti-S1P (n = 6). (D) Wild type (WT) into WT, S1PR2−/−, Sphk1−/−, or Sphk2−/− recipients (n = 8 WT, S1PR2−/−; n = 12 Sphk1−/−; n = 6 Sphk2−/−). (E to H) CD4 across iSVECs to S1P: (E) Naïve CD4 (F) Teff and (G) Tmem across iSVECs treated with S1PR2 antagonist. (H) iSVECs infected with shRNA lentivirus. (I to K) CD4 footpad migration: (I) Naive CD4 and (J) Teff footpads pretreated with S1PR2 antagonist (n = 15 naïve and n = 8 Teff); (K) endogenous T cell migration, anti-CD62L mAb intravenously, and S1PR2 antagonist in footpad; total CD4+CD69+ activated and CD4+CD44hi memory in dLN (n = 10). (L and M) CD4 across iLECs to S1P: (L) Naïve CD4 or (M) Teff pretreated with S1PR1, S1PR4, or combined antagonists; iLECs treated with S1PR2 antagonist. (N) Naïve CD4 pretreated with S1PR1 and S1PR4 antagonists into footpad pretreated with S1PR2 antagonist (n = 8). (O) Naïve CD4 treated as indicated; iSVECs treated with S1PR2 antagonist; S1P plus CCL19. (P) Anti-CCR7 or S1PR1 or S1PR4 antagonists added at indicated time points. For (E) to (H), (L) to (M), (O), and (P), mean ± SEM, three independent experiments, one-way ANOVA for multiple comparisons; for (B), (C), (I), (J), and (N), paired Student’s t test; for (K), unpaired Student’s t test; for (D), Mann-Whitney.

Treatment of LECs with PTX or FTY720 did not inhibit migration (Fig. 3A and fig. S3A), suggesting the use of S1PR2 instead of S1PR1. Pretreatment of LECs, but not naive T cells, with the S1PR2 antagonist specifically decreased migration (Fig. 4E and fig. S3H). Migration of Teff and Tmem was inhibited by LEC pretreatment with S1PR2 inhibitor (Fig. 4, F and G). LEC pretreatment with S1PR4 antagonist did not inhibit migration (Fig. 3C and fig. S3C), and S1PR2−/− CD4 T cell migration was not inhibited toward S1P or CCL19 (fig. S3I), showing cell specificity for receptor blockade. It was not possible to treat T cells or LEC alone with the S1PR1 inhibitor because it was reversible (32). However, S1PR2 but not S1PR1 knockdown of SVECs (fig. S3, J and K) resulted in decreased TEM toward S1P but not CCL19 (Fig. 4H and fig. S3L).

To assess blockade in vivo, pretreatment of footpads, but not T cells, with S1PR2 antagonist inhibited naïve and effector T cell migration (Fig. 4, I and J, and fig. S3M). To validate these findings for endogenous T cells, we administered anti-CD62L intravenously to prevent leukocyte entrance into LN through blood, and the S1PR2 antagonist was injected into the footpad (27, 29). Analysis of the dLN showed that S1PR2 antagonist reduced endogenous CD4+CD44hi memory and CD4+CD69+ activated T cell homing (Fig. 4K). Wild-type CD4 T cells migrated less in S1PR2−/− (Fig. 4D), whereas S1PR2−/− CD4 T cells migrated the same as wild type (fig. S3N). Ear whole mounts showed that S1P, Lyve-1, CCL21, morphology, and lymphatic density were normal for S1PR2−/− afferent lymphatics (fig. S3F). Overall, the results demonstrated that LEC S1PR2 regulated in vitro and in vivo migration of multiple CD4 T cell subsets.

S1PRs and CCR7 have distinct roles in trans-endothelial migration

Combined treatments of LECs with S1PR2 antagonist and T cells with S1PR1 and S1PR4 antagonists almost completely blocked naïve T cell (Fig. 4L) and Teff (Fig. 4M) migration in vitro. T cell migration into dLN was completely blocked with the combination of antagonists in vivo (Fig. 4N). CCR7 is a potent regulator for T cell afferent lymphatic migration (6). To test for an interaction or preference for S1PRs and CCR7, we assessed CD4 T cell migration to S1P plus CCL19. When both were combined, there was an additive effect (Fig. 4O). S1PR1, S1PR2, and/or S1PR4 antagonists blocked migration, and to a greater extent than the component contributed by S1P alone (Fig. 4O). Anti-CCR7 blocked CD4 T cell migration toward S1P plus CCL19 (Fig. 4O) and CD4 T cell homing into dLN (fig. S4A). These results suggested that S1PRs and CCR7 were independently required by CD4 T cells for TEM, although CCR7 appeared more potent compared with the various S1PRs. None of the S1PR antagonists inhibited TEM driven by CCL19 (excluding off-target events and cross-talk among different GPCR; fig. S4B).

To distinguish why different receptors were required for TEM, we added S1P or CCL19 alone or combined to the migration assays, and we added antagonists at various times. The S1PR antagonists only blocked migration when added early, whereas anti-CCR7 blocked migration at both early and late times, and this was true for single or combined ligands (Fig. 4P and fig. S4, C and D). These findings suggested that S1P-S1PRs were required for early migration events and interactions between T cells and LECs, whereas CCL19-CCR7 were required for later interactions.

S1P-S1PRs regulate afferent lymphatic entry

Immunohistochemistry of the dLN showed that carboxyfluorescein diacetate succinimidyl ester (CFSE)–labeled, transferred CD4 T cells had the same distribution patterns among groups treated with vehicle and S1PR1, S1PR4, and S1PR2 antagonists (fig. S5A), and S1PR2 antagonist did not change LN architecture (fig. S5B). This suggested that inhibitors prevented migration across lymphatic endothelium and into lymphatic vessels, rather than block downstream events, such as entry into the subcapsular space or dLN parenchyma.

To verify that S1PRs regulated CD4 T cell migration across lymphatic endothelium in vivo, we transferred CD4 T cells treated with antagonists into pinnae. Whole-mount staining showed that the S1PR4 antagonist inhibited CD4 T cell migration nearby and into the afferent lymphatic vessel lumens (Fig. 5A and fig. S5C). S1PR4−/− CD4 T cells migrated less efficiently nearby or into lymphatics (Fig. 5B and fig. S5D). S1PR1 antagonist prevented migration nearby or into the afferent lymphatics (Fig. 5C and fig. S5E). In contrast, S1PR1-Tg and S1PR1 S5A CD4 T cells migrated into afferent lymphatics more rapidly and completely; thus, only 2 hours after transfer, there was already significant migration into vessels (Fig. 5, D and E). CD4 T cells failed to migrate into vessels in ears pretreated with anti-S1P mAb or S1PR2 antagonist (Fig. 5, F and G). Wild-type CD4 T cells failed to enter lymphatics of S1PR2−/− recipients (Fig. 5H). The role of CCR7 in afferent lymphatic distribution was also assessed: Anti-CCR7 treatment of the pinnae prevented T cell migration nearby or into the afferent lymphatics (fig. S5F). These results demonstrated the locus of S1P-S1PR as TEM and entry into the vessel.

Fig. 5 S1PRs regulate CD4 T cell afferent lymphatic distribution.

Naïve CD4, ear pinnae migration. (A) CD4 treated with S1PR4 antagonist, into pinnae. (B) Wild type (WT) and S1PR4−/− CD4, into WT pinnae. After 16 hours, pinnae were stained for Lyve-1. (C) CD4 treated with S1PR1 antagonist, into pinnae. (D) S1PR1-Tg or (E) S1PR1 S5A or littermate CD4, WT pinnae. After 2 hours, pinnae were stained for Lyve-1. (F) CD4, into pinnae treated with anti-S1P. (G) CD4, into pinnae treated with S1PR2 antagonist; and (H) WT CD4, into WT or S1PR2−/− pinnae. After 16 hours, pinnae were stained for Lyve-1. Samples were analyzed by fluorescence microscopy; magnification, ×40; scale bars, 8 μm. Percentage of CD4 in lymphatics out of total CD4, or number of lymphatic vessel lumens with clustered groups of CD4, calculated per high-powered field. Four independent experiments, two ears per treatment per experiment, 6 to 8 mice per group, 10 high-powered fields from each pinna. Statistical analyses, Student’s t test.

S1PR2 signals through ERK to regulate LEC permeability and junction molecule expression

S1PR2 activates the mitogen-activated protein kinases (MAPKs) p38 and c-Jun N-terminal kinase (JNK) in blood vascular endothelial cells (30, 38). Similarly, S1P induced phosphorylation of p38, JNK, and ERK1/2 in LECs (Fig. 6A). S1PR2, but not S1PR1, antagonism specifically inhibited phosphorylation (Fig. 6A). Endogenous S1P synthesis also regulates LECs (39), and the Sphk inhibitor (Sphki) decreased ERK phosphorylation in untreated cells, whereas p-ERK was restored by exogenous S1P (Fig. 6B). Together, these results showed that exogenous or endogenous S1P stimulated LEC S1PR2 and MAPKs.

Fig. 6 S1PR2 regulates LEC layer integrity and junction molecules.

(A and B) Western blot: SVECs pretreated with S1PR2 or S1PR1 antagonists, stimulated with S1P (A); SVECs pretreated with Sphki, stimulated with S1P (B). (C to F) Permeability assays: Evans Blue, upper chamber; S1P, lower chamber; SVECs treated as indicated; three independent experiments. (G) In vivo permeability, footpad treated with S1PR2 antagonist, Evans Blue to dLN measured at indicated times. (H to J) Junction molecule immunohistochemistry: ears treated with S1PR2 antagonist; Lyve-1, VE-cadherin (H); VE-cadherin, occludin (I); ZO-1 (J). (K) VE-cadherin by WT and S1PR2−/− lymphatics (two ears per treatment per experiment, n = 6); magnification, ×40; scale bars, 21 μm. (L to N) Junction molecule expression, SVECs treated with inhibitors. Mean ± SEM: (C) to (F) and (K), one-way ANOVA for multiple comparisons; (G) to (J), Student’s t test.

To determine whether the signaling regulated LEC structure and function, we treated cell layers with S1P and/or inhibitors, and we assessed layer permeability by dye transport. The S1PR2 antagonist decreased permeability, whereas exogenous S1P, anti-S1P, or FTY720 did not (Fig. 6, C to E). The ERK inhibitor also decreased permeability (Fig. 6F). The antagonists did not alter LEC viability (fig. S5G). The S1PR2 antagonist inhibited lymphatic dye transport from tissues into dLN (Fig. 6G). Thus, S1PR2 and ERK regulate LEC permeability.

The permeability changes suggested that S1P alters cell junctions (39). Overlapping flaps at endothelial borders of initial lymphatics are connected by discontinuous button-like junctions, composed of VE-cadherin and other junction-associated proteins (40); DCs penetrate into afferent lymphatics through the portals between button junctions (41). To assess these junctions, we treated pinnae or LECs with inhibitors, and we measured the effects on junctional proteins and structures. S1PR2 inhibition resulted in increased numbers of VE-cadherin–positive buttons (Fig. 6H), along with increased expression of the junction molecules occludin and zonulin-1 (ZO-1) (Fig. 6, I and J). In S1PR2−/− pinnae, the afferent lymphatic terminals expressed more VE-cadherin and buttons (Fig. 6K). The S1PR2 antagonist and ERK inhibitor, but not the other S1PR antagonists, increased VE-cadherin, ZO-1, and occludin expression (Fig. 7, A to C). Together, these results suggested that S1P-S1PR2 triggered ERK signaling, thereby altering LEC layer permeability through junction molecule expression.

Fig. 7 S1PR2 regulates junction molecules and S1PRs mediates T cell paracellular and transcellular migrations.

(A) VE-cadherin; (B) ZO-1; (C) occludin. SVECs pretreated with S1PR2 or S1PR1 antagonists; ERK inhibitor; stimulated with S1P. Magnification, ×100; scale bars, 8 μm; three independent experiments. (D) CD4 migration, iSVECs treated with ERK inhibitor, (E) ROCK or Rho inhibitors; three independent experiments. (F) CFSE-labeled CD4 toward S1P and CCL19 (i); T cells pretreated with S1PR1 or S1PR4 antagonist; eFluor 670–labeled SVECs pretreated with S1PR2 antagonist toward S1P (ii); eFluor 670–labeled SVECs pretreated with anti–VCAM-1 (iii); magnification, ×63; scale bar, 14 μm; arrowheads indicate paracellular (white) and transcellular (yellow) migration. Percentage of T cells in paracellular and transcellular positions (right); >159 cells counted per group; three independent experiments. Mean ± SEM: (L) to (P), one-way ANOVA for multiple comparisons.

S1PR2 signals through ERK to regulate transcellular TEM

Because S1P-S1PR2-ERK regulated permeability and junction structure, we next determined how this affected TEM. The ERK inhibitor blocked TEM (Fig. 7D). We previously showed that S1PR2 signaling in human umbilical cord endothelial cells (HUVECs) involved Rho or Rho-associated protein kinase (ROCK) and phosphatase and tensin homolog (PTEN) (38) and regulated endothelial permeability and junction integrity. This additional pathway was active here as Rho and ROCK inhibitors decreased TEM (Fig. 7E). Together, these results suggested that S1P-S1PR2 triggered ERK signaling, thereby altering LEC layer permeability through junction molecule expression, which regulated TEM.

Leukocytes cross endothelium via paracellular and transcellular routes, which depend on endothelial cell junction structures (42, 43). We previously demonstrated that CCL19-driven TEM used both routes equally (6). Here, we found that the transcellular position was preferentially engaged for S1P-driven T cell migration (Fig. 7F, i). Not only did the S1PR1, S1PR2, and S1PR4 antagonists inhibit overall migration, but each also reduced the percentage of transcellular migration (Fig. 7F, ii). Anti–VCAM-1 and Rho and ROCK inhibitor treatments also reduced the percentage of transcellular migration (Fig. 7F, iii, and fig. S5H) while the ERK inhibitor did not change the preference for migration route (fig. S5H). Thus, changes in permeability and junction molecule expression are reflected in the physical pathways for TEM. Overall, S1P-S1PR2 signaling recruited all these molecules, with preferences for use of several in transcellular migration.

S1PR2 regulates endothelial expression of, and S1PR1 and S1PR4 regulate T cell binding to, VCAM-1

As noted above, S1P-mediated TEM was dependent on VCAM-1–VLA-4 (Fig. 2G). S1PR2 up-regulates VCAM-1 expression on blood vascular endothelial cells (44). To investigate S1PR regulation of VCAM-1 expression, we noted that VCAM-1 expression on SVECs (Fig. 8, A and B and fig. S6, A and B) and primary LECs (fig. S6, C and D) was down-regulated by S1PR2 but not by S1PR1 inhibitor, without affecting other receptors such as Lyve-1. ERK inhibitor U0126 decreased VCAM-1 expression (Fig. 8, A and B). VCAM-1 expression was decreased by Sphki inhibition of endogenous S1P synthesis (Fig. 8C) and partially restored by exogenous S1P. VCAM-1 on the LECs surrounded and facilitated T cell transmigration (45), and T cells did not express VCAM-1 (fig. S6E). The S1PR2 antagonist and the ERK inhibitor both inhibited VCAM-1 expression and recruitment around the T cells (Fig. 8, D and E). The S1PR2 antagonist also inhibited VCAM-1 expression in vivo (Fig. 8F). Like the S1PR2 inhibitor, the ERK inhibitor decreased CD4 T cell migration (Fig. 6O). Together, these results confirmed S1P-S1PR2-ERK regulation of VCAM-1 expression by LECs, a tight association between migrating T cells and LEC VCAM-1, and the importance of VLA-4–VCAM-1 in TEM.

Fig. 8 LEC S1PR2 regulates VCAM-1 expression.

(A) VCAM-1 expression, SVECs treated with S1P, S1PR1 antagonist, S1PR2 antagonist, or ERK inhibitor for 3 hours. Flow cytometry analysis, three independent experiments. (B) VCAM-1 expression, SVEC layers pretreated with S1PR2 antagonist, S1PR1 antagonist, or 5 μM ERK inhibitor, and then treated with S1P. Fluorescence microscopy analysis; magnification, ×40; scale bars, 21 μm. (C) VCAM-1 expression, SVECs pretreated with Sphki and then treated with S1P. Fluorescence microscopy analysis; magnification, ×100; scale bars, 8 μm. (D and E) CD4 position, LEC VCAM-1, and actin distribution. CD4 migration to S1P for 1 hour over iSVECs pretreated for 1 hour with (D) S1PR2 antagonist or (E) ERK inhibitor, at least three independent experiments. (F) VCAM-1 expression, afferent lymphatics treated with 0.5 nM S1PR2 antagonist overnight; magnification, ×40; scale bars, 21 μm. (G) Naïve CD4, treated with S1PR1 or S1PR4 antagonists, migrated to S1P, across transwell inserts coated with VCAM-1–Ig, three independent experiments. (H) Naïve CD4, treated with S1PR1 or S1PR4 antagonists, migrated to S1P, across transwell inserts coated with VCAM-1–Ig. Real-time imaging, tracking path (left), velocity (middle), and distances (right) of individual cells. One-way ANOVA for multiple comparisons (B, C, F, and G).

It has been reported that VCAM-1 forms a transmigratory cup to direct leukocyte TEM (46). To further investigate the role of T cell–VCAM-1 interactions for migration, we coated transwell inserts with VCAM-1–immunolglobulin (Ig) without LECs, and we assessed T cell migration to S1P and VCAM-1. Both S1P and VCAM-1 promoted T cell migration in an additive manner (Fig. 8G). The S1PR4 but not the S1PR1 antagonist prevented S1P plus VCAM-1–driven migration (Fig. 8G). To verify the specificity, anti–VLA-4 treatment also blocked T cell migration (fig. S6F). This regulation was highly specific because ICAM-1–Ig promoted CCL19 but not S1P-driven migration (fig. S6G). Real-time imaging across LECs showed that S1PR1 antagonism reduced T cell migration distances, but not migration velocities, although the highest velocities were inhibited, whereas S1PR4 antagonism decreased both distances and velocities (Fig. 7H). The results suggested that the specific role of S1PR4 was to regulate the interaction of T cells with VCAM-1. Because the S1PR1 antagonist actually increased migration to a small extent, the results suggested that S1PR1 may regulate adhesion and/or release of the T cell from adhesion to permit continued migration.


T cell S1PR1 and S1PR4 plus LEC S1PR2 were necessary for TEM. S1P gradients not only acted as chemotactic signals for T cells but also regulated LECs to facilitate T cell TEM. This suggested active coengagement of T cells and LECs due to S1P signaling. In contrast to chemokines, S1P-driven migration required LECs, was chemotactic and chemokinetic, preferentially used transcellular routes, but was not facilitated by LEC-derived CCL21. Because LEC-derived CCL21 is necessary for DC tissue egress (23, 47), it was possible that S1PRs mobilized an intracellular CCL21 pool that anti-CCL21 failed to target. The preferential use of transcellular migration may support such a hypothesis. S1PR2 regulated downstream signal p38, ERK and JNK activation, endothelial permeability, junction proteins and VCAM-1 expression and distribution, LEC buttons, and transcellular TEM. S1P-driven migration was similar to chemokines for use of VCAM-1–VLA-4, responsiveness to fluid flow, and modulation by inflammation.

There are many investigations of S1PR1 in thymus, blood, and LN, demonstrating that it controls T cell trafficking at multiple stages of T cell development and responses (1, 2, 13). The present findings add to the roles played by S1P and S1PRs in T cell migration. S1PR1 and S1PR4 had similar magnitudes of effect on migration or migration inhibition, were required early in the migration response, regulated transcellular migration, regulated chemotaxis and chemokinesis, and altered T cell positions around lymphatic vessels and lumens. Their major functional differences were in T cell motility and interaction with VCAM-1. Enforced surface expression of S1PR1 showed that S1PR1 promoted CD4 T cell afferent lymphatic TEM, whereas down-modulation prevented this. A recent study of human T cell migration across plastic showed that different T cell subsets used both S1PR1 and S1PR2 (48), suggesting different S1PRs expression by human and murine T cells. In our previous report (6), high local concentrations of S1P in inflammation acted as a stop signal for TEM, by down-modulation of S1PRs and possibly altering the S1P gradient. The ability of S1P to stimulate chemokinesis suggests that tissue egress may still occur despite increased local S1P concentrations.

S1P regulates important aspects of endothelial cells, including blood vascular endothelial cell spreading, maturation and stabilization, and barrier integrity (49), independent of VE-cadherin (49) and through S1PR1 and S1PR3 (49). Less is known about S1P and S1PRs in regulating LECs. We demonstrated that S1PR2 regulated LECs and lymphatic barrier permeability and junction molecule expression. S1PR1 did not play the same role, although S1PR1 can regulate blood endothelial cell permeability and mobility (49); hence, S1PR1 might have complementary roles to S1PR2 in other circumstances. Exogenous S1P did not change basal barrier function, suggesting that endogenous S1P maintained tonic barrier integrity, which was inhibited by S1PR2 blockade. S1P-S1PR2 induced ERK activation and regulated VCAM-1 and VE-cadherin, ZO-1, and occludin expression. Inhibition of S1PR2 altered VCAM-1 and VE-cadherin expression, along with changes in the density and distribution of the button and zipper endothelial junctions and changes in the interaction of CD4 T cells and LECs during TEM. These findings suggested that S1P-S1PR2-ERK regulated junctional and adhesion molecule expression, distribution, and function, which controlled LEC permeability and transcellular TEM. The results also suggested a simultaneous role for Rho and ROCK, which are regulated by S1PR2 in blood endothelial cells (38).

Paracellular and transcellular TEM involve dissimilar mechanisms, although both rely on recruitment of membrane from the lateral border recycling compartment (42, 43). During paracellular migration, VE-cadherin moves away from the migration site (42). S1PR2 decreases while its antagonist increases VE-cadherin on HUVECs (38). We found that S1PR2 and ERK antagonists increased VE-cadherin expression and reduced migration, suggesting that VE-cadherin was involved in both migration routes. We previously demonstrated that CCL19-driven migration across LECs used both routes (6), and FTY720 caused preferentially transcellular migration, suggesting that S1PRs differentially regulate the two pathways. We found that transcellular migration was preferentially engaged by S1P-S1PR2, which may explain why S1P but not CCL19 promoted T cell–LEC interactions for enhanced migration. Anti–VCAM-1 mAbs and ROCK and Rho inhibitors reduced the proportion of T cells migrating through the transcellular route toward S1P, whereas ERK inhibitor did not significantly change this proportion. Thus, S1P-S1PR2 signaling recruited all these molecules, with preferences for use of some in transcellular migration.

Several receptors are needed for TEM, and distinct roles and functions were defined. The S1PR1 inhibitor increased migration through VCAM-1, suggesting that S1PR1 regulated binding and/or release of T cells from adhesion. In contrast, the S1PR4 inhibitor blocked migration; thus, S1PR4 is responsible for migration through VCAM-1. S1PR1 regulated migration distances but not velocities, whereas S1PR4 regulated both distances and velocities. There was sequential use of S1PRs and CCR7: S1PRs were engaged early while CCR7 was engaged later. LECs promoted S1P- but not CCL19-driven migration, and T cells migrated toward CCL19 equally through paracellular and transcellular routes, whereas S1P-driven migration preferentially used the transcellular route. These observations suggest that S1PRs mediated T cell–LEC interactions and initiated T cell transcellular migration at early time points that could not be reversed by S1PR antagonists. In contrast, CCR7-driven migration depended on the CCL19 gradient and could be blocked at later times. Our findings suggest new areas requiring more investigation. The precise T cell–LEC interactions will require more definition of the receptors and ligands involved at the cell surfaces and events that determine cytoskeletal and membrane movements. Such investigations will likely shed more light on how and why multiple receptors must be used simultaneously for migration. The pathways engaged by S1PR1, S1PR4, and CCR7 for T cell movements to and across LEC will require investigation at higher magnification to visualize subcellular and molecular movements. It will be even more challenging to confirm in vitro observations with in vivo models and to determine whether efferent and afferent lymphatics function similarly or not or whether other lymphatic beds have unique responses (50). The intercellular interactions seem dependent on each other; hence, analyses will have to account for the complexity of complementary T cell and endothelial cell responses, particularly in teasing out the decision of both cells to use paracellular versus transcellular migration pathways.



C57BL/6 mice aged 8 to 10 weeks were purchased from the Jackson Laboratory (Bar Harbor, ME). S1PR1 flox × Rosa26-Cre-ERT2 (backcrossed with C57BL/6 for eight generations) (7), S1PR1-Tg mice (backcrossed with C57BL/6 for eight generations) (34), S1PR1S5A/S5A mice (backcrossed with C57BL/6 for five generations) (7), S1PR1 flox × UBC-Cre-ERT2 (C57BL/6 for both S1PR1 and UBC-CreERT2) (51, 52), S1PR2−/− (53), S1PR4−/− (54), and SphK1−/− and SphK2−/− (37) were used. All experiments were performed with age- and sex-matched mice and approved by the Institutional Animal Care and Utilization Committee at the University of Maryland at Baltimore.

Cell lines and primary cells

SVEC4-10 (designated as SVEC) (CRL-2181) were from the American Type Culture Collection (Manassas, VA). C57BL/6 mouse skin primary LECs and human skin primary LECs were purchased from Cell Biologics (Chicago, IL) and were cultured according to the manufacturer’s instructions. Human Teff, iTreg, and nTreg were generated as previously published (55). Mouse CD4 from spleen and LNs were prepared according to the manufacturer’s protocols (CD4 enrichment kit, STEMCELL Technologies Inc., Vancouver, Canada). Naive CD4 (CD4+CD44low) cells were activated with plate-bound anti-CD3 [precoated plate with anti-CD3 Ab (5 μg/ml) overnight], anti-CD28 (1 μg/ml), and IL-2 (10 ng/ml) for 3 days. Naïve, nTreg (CD4+CD25hi), activated CD4 T cells (CD4+CD69+), and memory CD4 T cells (CD4+CD44hi) were isolated with flow cytometry sorting (27, 29).


S1P was from Avanti Polar Lipids (Alabaster, AL). Recombinant murine CCL2, CCL5, CCL19, CCL21, CCL22, CXCL10, CXCL12, IL-6, and TNFα were from R&D Systems (Minneapolis, MN). S1PR antagonists W146 (S1PR1), JTE013 (S1PR2), CYM 50358 hydrochloride (S1PR4), ERK inhibitor U0126, Sphki Ski II, ROCK-selective inhibitor Y-27632, and Rho-selective inhibitor Rhosin were purchased from Tocris (Bristol, UK). CFSE and eFluor 670 were purchased from Molecular Probes (Eugene, OR). VCAM-1–Ig and ICAM-1–Ig were purchased from Sino Biological (Beijing, China). Anti-S1P mAb (clone LT1002) was supplied by R. A. Sabbadini (Lpath Inc., San Diego, CA). All of the antibody resources are indicated in table S1.

In vitro migration

A total of 7.5 × 104 SVECs or 15 × 104 primary mLECs and hLECs were seeded on the upper surface of a 5-μm pore size transwell insert coated with 0.2% (w/v) gelatin and cultured as described previously (21). For VCAM-1–Ig coating experiments, transwell inserts were coated with VCAM-1–Ig (5 μg/ml) for 2 hours at 37°C. A total of 2 × 105 freshly isolated CD4 T cells were added in a volume of 100 μl (for flow conditions, 340 μl) to the upper chambers of a 24-well transwell plate with a 5-μm insert (Corning International, Corning, NY). Lower wells contained various concentrations of chemokines: 53 nM CCL19, 42 nM CCL21, 12.5 nM CXCL12, 50 nM IL6, 25.3 nM CCL5, 23.5 nM CCL2, 100 nM S1P, or 18.2 nM CCL22 in 600 μl (for flow conditions, 360 μl) of migration medium [RPMI 1640/0.5% fatty acid–free bovine serum albumin (Sigma-Aldrich)]. The number of T cells that migrated to the lower well after 4 hours was counted with a hemocytometer. The percentage of migrated cells was calculated as the total number of transmigrated cells divided by the total cell input.

Real-time imaging

A total of 20 × 103 CFSE-labeled CD4 T cells in 100 μl of migration medium were loaded into the upper chamber of transwell inserts coated with or without SVEC layers and then migrated toward 53 nM CCL19 or 100 nM S1P. CD4 T cells migrating across LECs were visualized by the EVOS FL Auto Imaging System (Life Technologies, Carlsbad, CA) with a 20× objective. One image was captured every 5 min for 2 hours. Distance from origin, velocity, displacement, and individual cell tracking times were analyzed with Volocity version 6.1.1 (PerkinElmer).

Footpad migration assay

As previously described (6), 2 × 106 CFSE-labeled CD4 (treated with or without 10 μM S1PR antagonist or 10 μM S1PR4 antagonist) in 30 μl of phosphate-buffered saline (PBS) were injected into the footpads that were treated with or without 0.5 nM S1PR2 antagonist. For endogenous T cell migration, mice were pretreated with 100 μg of anti-CD62L (MEL-14, BioXCell, West Lebanon, NH) intravenously; 24 hours later, footpads were treated with 0.5 nM S1PR2 antagonist, and dLNs were assessed 16 hours later (29). For LPS (Sigma-Aldrich, St. Louis, MO)–induced inflammation, 1 μg of LPS in 20 μl of PBS was injected per footpad 2 hours before T cell transfer. CFSE-labeled CD4 T cells in 30 μl of PBS were mixed with 1 μg of mouse control IgG or anti-S1P, or 1 μg of rat control IgG or anti–VCAM-1, and injected into the footpads. In the paired analysis for footpad migration, T cells were transferred to both hind footpads of a recipient mouse. T cells on one side were treated with active reagent (e.g., receptor blocker), whereas T cells on the other side were treated with vehicle or control compound. Each mouse thus acts as its own control. Two or 16 hours after injection, mice were euthanized, the draining popliteal LNs were harvested, and single-cell suspensions or tissue sections were prepared for flow cytometric or immunohistochemical analysis, respectively.

Whole-mount staining of mouse ear pinnae

A total of 1 × 106 CFSE-labeled naïve CD4 T cells treated with or without S1PRs antagonists (10 μM S1PR antagonist or 10 μM S1PR4 antagonist) in 10 μl of PBS were injected into mouse ear pinnae treated with or without 0.5 nM S1PR2 antagonist or 10 μg/ml anti-CCR7 (in 10 μl of PBS) intradermally. Twelve hours later, ears were collected and peeled into two halves, fixed with 3% paraformaldehyde (PFA) in PBS for 5 min at 4°C, permeabilized with 1% Triton X-100 in PBS for 30 min at 4°C, incubated with 5% donkey serum (Jackson ImmunoResearch, West Grove, PA) for 30 min at 4°C, and incubated with the primary antibody at 4°C overnight. The ears were then washed with PBS, incubated with secondary antibody for 4 hours at 4°C, washed with PBS, and fixed with 3% PFA at 4°C for 10 min. For S1P staining, mouse ears were fixed with 10% buffered formalin for 2 min, permeabilized with 0.1% Triton X-100 for 10 min, blocked with 2% casein (Sigma-Aldrich, St. Louis, MO) solution in PBS for 10 min at room temperature, incubated with anti-S1P (Lpath Inc., San Diego, CA) and anti–Lyve-1 for 1 hour at room temperature, and then washed and incubated with secondary antibody donkey anti-rabbit Cy3 at 5 μg/ml (Jackson ImmunoResearch, catalog. no. 711-165-152) and donkey anti-mouse Alexa Fluor 488 at 5 μg/ml (Jackson ImmunoResearch, catalog. no. 705-546-147) for 1 hour at room temperature. The stained ears were analyzed by fluorescence microscopy (Nikon Eclipse E800, Nikon Co., Tokyo, Japan). For three-dimensional (3D) confocal microscopy, ear pinna samples were visualized with a Zeiss LSM5 Duo with a 63× or 40× objective, respectively. Z-stack images were acquired every 1 μm with 10- to 16-μm slice thicknesses.

Fluorescent flow cytometry

SVECs and mLECs were treated under various conditions and stained with indicated Abs. Samples were analyzed with an LSR Fortessa Cell Analyzer (BD Biosciences, San Diego, CA). For apoptosis assays, cells were stained with the 7-AAD and annexin V apoptosis kit (BD Pharmingen, San Diego, CA) according to the manufacturer’s protocols and analyzed by flow cytometry. Data were analyzed with FlowJo software v 8.8.7 (Tree Star, Ashland, OR).

Immunofluorescent staining

SVEC monolayers were stained with anti–VCAM-1, anti–ZO-1, or anti–VE-cadherin for 1 hour at 4°C and then incubated with fluorescein isothiocyanate or Cy3-labeled donkey anti-rat secondary Ab (Jackson ImmunoResearch, West Grove, PA) for 30 min at 4°C. The inserts were washed with PBS and fixed with 3% PFA at 4°C for 10 min. The insert membranes were removed from transwell and mounted on slides. For paracellular and transcellular migration, SVECs were stained with eFluor 670 (Molecular Probes, Eugene, OR), and then CFSE-labeled CD4 T cells migrated toward CCL19 or S1P across the iSVEC layers. One hour later, the inserts were washed with PBS, fixed with 3% PFA at 4°C for 10 min, permeabilized with 0.1% Triton X-100 in PBS for 5 min at 4°C, and stained with Alexa Flour 555–labeled phalloidin for 30 min at 4°C. Fluorescence microscopy was performed with a Nikon Eclipse E800. 3D confocal microscopy was performed on cell layers as noted above for ear pinna microscopy with a 63× objective. Quantitative analysis was performed with Volocity 3D Image Analysis Software to demonstrate the distribution of different molecules on cells and measure the density of different stains (PerkinElmer, Waltham, MA).

Permeability assays

For in vitro permeability, iSVEC inserts were pretreated under various conditions noted in the text and figure legends, and 100 μl of Evans Blue (0.67 mg/ml; Sigma-Aldrich) in migration medium was added to the top chamber and 600 μl of migration medium (21) without phenol red was added to the lower chamber. After 0.5, 1, 2, and 3 hours, lower-chamber medium was collected, and the optical density (OD) at 620 nm was measured in a microplate reader (TECAN, San Jose, CA). For in vivo permeability, vehicle or the S1PR2 antagonist was injected into hind footpads; 1 hour later, 30 μl of Evans Blue (0.67 mg/ml) in PBS was administered in the footpad, and 16 hours later, the popliteal dLNs were collected and dissociated in 100 μl of PBS, and OD at 620 nm was measured in a microplate reader.

Lentivirus and infections

Viruses were packaged target cells infected according to established protocols (56). S1PR1 and S1PR2 short hairpin RNAs (shRNAs) were purchased from Sigma-Aldrich (St. Louis, MO). Lentiviral particles were produced by cotransfecting human embryonic kidney (HEK) 293 T cells with shRNA plasmid, pCMV-ΔR8.2 packaging plasmid (Addgene, Cambridge, MA), and pCMV-VSVg envelope plasmid (Addgene). For infection, lentivirus preparations were mixed with polybrene (6 μg/ml) (American Bioanalytical, Natick, MA) and applied to SVECs at a multiplicity of infection of 2. Knockdown of S1PR expression was analyzed by RT-PCR, and the cells were used for in vitro migration assays.

Western blotting

A total of 1 × 106 cells were lysed with buffer containing 20 mM Hepes (pH 7.4), 1% Triton X-100, 150 mM NaCl, 12.5 mM β-glycerophosphate, 50 mM NaF, 1 mM DTT, 1 mM sodium orthovanadate, 2 mM EDTA, 1 mM PMSF, and protease inhibitor mixture (Roche Applied Science). Protein concentrations were measured with the Quick Start Bradford Protein Assay (Bio-Rad, Philadelphia, PA), and equal amounts of proteins were loaded and separated on 4 to 20% mini-gels (Invitrogen), electrotransferred to Immobilon-P membranes (Millipore, Darmstadt, Germany), blocked, and probed with the indicated Abs.

Statistical analysis

In vitro transwell migration assays were performed at least three times for individual experiments, and results represent mean values of triplicate samples. In vivo migration experiments were performed at least two times to collect enough data points to perform statistical analysis, usually 6 to 15 mice per group. Immunohistochemistry and hematoxylin and eosin staining were performed at least three times for individual experiments (two inserts per experiment), and 10 fields per insert were acquired. All flow cytometry experiments were performed at least three times. Fluorescence images were analyzed with Volocity software for quantification of fluorescence density. Other results were analyzed by GraphPad Prism Software (version 5, GraphPad Software Inc., La Jolla, CA) and presented as the mean ± SEM. Statistical analyses were performed using Student’s t test. One-way analysis of variance (ANOVA) was used for multiple comparisons. P < 0.05 was considered statistically significant. No statistical tests were used to predetermine the size of experiments. Sample and experiment sizes were determined empirically for sufficient statistical power. No samples were excluded specifically from analysis, and no randomization or blinding protocol was used.


Materials and Methods

Fig. S1. LECs do not promote CD4 T cell migration toward other chemokines and cytokines.

Fig. S2. LECs and lymphatics express CCL21 but not other chemotactic factors.

Fig. S3. S1PR antagonists inhibit migration to S1P but not CCL19, and do not induce CD4 T cell apoptosis; Spkh but not S1PR2 deletion decreases S1P expression by afferent lymphatics.

Fig. S4. S1PRs, CCR7, and receptor inhibitors regulate migration to their cognate ligands.

Fig. S5. CD4 T cell distribution in LNs after treatment with S1PR antagonists.

Fig. S6. S1PR2 regulates VCAM-1 expression.

Fig. S7. Model of T cell lymphatic TEM.

Table S1. Antibodies for flow cytometry, Western blotting, and immunohistochemistry.


Acknowledgments: We thank performed at the University of Maryland Marlene and Stewart Greenebaum Cancer Center Flow Cytometry Shared Service. Funding: This study was supported by R01AI062765 and AI114496 (to J.S.B.); Division of Intramural Research Programs, NIAID (to A.O.); R35 HL135821 (to T.H.); RO1 HL11879 (to B.R.B.); and R01AI085166 and R01AI123308 (to S.R.S.). Author contributions: Y.X. and J.S.B. designed the research, planned and analyzed the experiments, and wrote the manuscript; Y.X., W.P., C.C.B., and L.L. performed the experiments; J.M.K. and A.O. provided S1PR2−/−, S1PR4−/−, Sphk1−/−, and Sphk2−/− mice; A.C. and T.H. provided S1PR1−/−, S1PR1-Tg, and S1PR1S5A mice; K.L.H. and B.R.B. provided human T cells; S.R.S. provided S1PR1−/− mice; all collaborators made important suggestions on experimental design and manuscript review. Competing interests: The authors declare that they have no competing interests. Data and materials availability: All data needed to evaluate the conclusions in the paper are present in the paper or the Supplementary Materials.
View Abstract

Navigate This Article