Chemokines and chemokine receptors are required for T cell trafficking and migration. Recent evidence shows that sphingosine 1-phosphate (S1P) and S1PRs are also important for some aspects of T cell migration, but how these two important receptor-ligand systems are integrated and coregulated is not known. In this study, we have investigated CCL19-CCR7 and CXCL12-CXCR4-driven migration of both splenic and peripheral lymph node (PLN) nonactivated and naive T cells, and used both S1P and the S1PR ligand, FTY720, to probe these interactions. The results demonstrate that splenic T cell migration to CCL19 or CXCL12 is enhanced by, but does not require, S1PR stimulation. In contrast, PLN T cell migration to CXCL12, but not CCL19, requires both chemokine and S1PR stimulation, and the requirement for dual receptor stimulation is particularly important for steps involving transendothelial migration. The results also demonstrate that: 1) splenic and PLN nonactivated and naive T cells use different molecular migration mechanisms; 2) CCR7 and CXCR4 stimulation engage different migration mechanisms; and 3) S1P and FTY720 have distinct S1PR agonist and antagonist properties. The results have important implications for understanding naive T cell entry into and egress from peripheral lymphoid organs, and we present a model for how S1P and chemokine receptor signaling may be integrated within a T cell.
Migration of nonactivated and naive T lymphocytes to peripheral lymph nodes (PLNs)3 occurs mainly at the high endothelial venules (HEVs), through a process of tethering, rolling, activation, and arrest. Initial surface engagement of the αLβ2 and α4β7 integrins and L-selectin tethers the T lymphocyte, causing it to roll along the endothelium. The chemokines CCL19, CCL21, and CXCL12 presented on the endothelial cell surface then engage T lymphocyte CCR7 and CXCR4, activating LFA-1 and α4β7, causing firm adhesion or arrest on the HEVs, and eventual diapedesis across the endothelial cell monolayer. Within the PLNs, T lymphocytes are segregated to the paracortical T lymphocyte zone and are in constant motion seeking out their cognate Ags. Upon Ag recognition, T lymphocytes are activated and chemokine receptors and adhesion molecules are differentially up- and/or down-regulated. These changes in chemokine receptor expression putatively allow the now activated T lymphocyte to escape PLN chemokines, exit via the efferent lymphatic vessels, and migrate to the chemokine gradient of the inflammatory nidus (reviewed in Refs.1 and 2). Although the progression of a T cell from a naive to an activated phenotype, and mechanisms regulating its subsequent egress from the PLN, are well understood, the majority of T cells that enter the PLN do not encounter their cognate Ag, and therefore leave the PLN in a nonactivated or naive state. The mechanisms of migration of this subgroup of T cells are not well understood, and only recently have been shown to be dependent on the function of sphingosine 1-phosphate (S1P) receptors (S1PR) (3, 4, 5).
The role of S1P in T lymphocyte migration has been advanced through the discovery that the immunomodulator FTY720 is a sphingosine analog. FTY720 prolongs human and murine allograft survival by promoting lymphopenia, through sequestration of T lymphocytes within PLNs and Peyer’s patches, without affecting T cell activation and proliferation (6, 7, 8, 9, 10). FTY720 is phosphorylated by sphingosine kinases to the active compound phosphorylated FTY720 (P-FTY720), an S1P structural homologue which, along with S1P, is an agonist at four of the five known S1PRs, S1P1, S1P3, S1P4, and S1P5 but not S1P2 (7, 11, 12). S1P1 is constitutively expressed on all cell types, while S1P4 is found primarily in lymphoid cells (13). Although both S1P1 and S1P4 are the major S1PRs expressed on T lymphocytes, recent evidence suggests that only S1P1 is responsible for P-FTY720 and S1P-driven activities that stimulate T cell migration and PLN accumulation (3, 4, 5).
Recently, several reports have illustrated that T lymphocyte egress from PLNs is more complicated than described above, and may involve other events in addition to chemokine receptor or adhesion molecule expression, especially for the migration of nonactivated T cells. Matloubian et al. (3) demonstrated in an elegant chimeric model that T and B lymphocytes require S1P1 activity for egress from peripheral lymphoid organs in the presence of intact chemokine and chemokine receptor systems. However, the precise mechanism by which S1P1 releases or retains T cells within LNs is not known. Graler et al. (14) demonstrated not only that S1P acts as a direct chemotactic agonist, but also that S1P or FTY720 causes reversible S1P1 internalization from the cell surface, resulting in noncompetitive antagonism of S1PR. Thus, S1PR stimulation by S1P or FTY720 may initially amplify chemokine-driven migration into the PLN, and then cause S1P1 down-modulation with PLN retention. Although S1P1 activity, demonstrated either through the use of a knockout animal model or through desensitization of the receptor by FTY720 or S1P, appears to play a role in T lymphocyte recirculation and egress from secondary lymphoid organs, the interactions between chemokine and S1PRs are largely undefined. In particular, it is not known whether both chemokine and S1PR stimulation are obligate requirements for PLN entry or exit, how these two receptor systems are linked and regulated in the PLN or outside the PLN, and whether discrete steps in the migration process are regulated by one and/or the other receptor.
Our interests have focused on the mechanism of FTY720 and S1PR activity with respect to chemokine-chemokine receptor interactions. We demonstrated that CCR2 and CCR7 are important for PLN T lymphocyte accumulation in response to FTY720, and that S1PR engagement by FTY720 uncovered a role for CXCR4 in PLN T cell accumulation, although CXCL12-driven migration was not directly stimulated by FTY720 in vitro (Refs.9 and 10 , reviewed in Ref.15). These results led us to conclude that chemokine and S1PR signals are both important for driving T cells into PLNs. We have continued these studies, focusing on the requirements for CCR7, CXCR4, and S1P1 for migration by splenic and PLN T cells. In this study, we show that splenic and LN T cells have distinct requirements for, and responses to, both chemokine and S1PR-driven migration. In particular, CXCL12-driven, transendothelial migration of PLN T cells is highly regulated by S1PR, in contrast to other chemokine responses and other T cells which are not regulated in this fashion. These results suggest a model in which distinct chemokine receptors and alternative S1PR signaling pathways are integrated to determine PLN entry and egress of T lymphocytes.
Materials and Methods
C57BL/6 mice 8–10 wk of age were purchased from The Jackson Laboratory. All mice were housed in a specific pathogen-free facility in microisolator cages. All experiments were performed with age- and sex-matched mice in accordance with institutionally approved animal care criteria.
CyChrome-, allophycocyanin-, PE-, and FITC-conjugated rat anti-mouse CD4 mAb, CD8a mAb, CD44 mAb, CD45RB mAb, CD69 mAb, or CD62L mAb were purchased from BD Pharmingen. Murine CCL2, murine CCL19, and murine CXCL12 were purchased from R&D Systems. S1P was purchased from Biomol Research Laboratories. FTY720 and P-FTY720 were kind gifts from Dr. V. Brinkmann (Novartis Pharma, Basel, Switzerlan).
Mice (two to three per group) were sacrificed and spleen and LNs (cervical, periaortic, inguinal, and axillary), were removed and gently dissociated into single-cell suspensions. RBC were removed by Tris-NH4Cl lysis. If indicated, cell suspensions were passed through T cell enrichment columns (R&D Systems); these cells were routinely 85–95% T cells. Cells were placed in complete RPMI medium (RPMI 1640 supplemented with 10% FCS, 1 mM sodium pyruvate, 2 mM l-glutamine, 100 IU/ml penicillin, 100 mg/ml streptomycin, 1× nonessential amino acids, and 2 × 10−5 M 2-ME).
Cell washes and Ab dilutions were performed in PBS plus 1% BSA at 4°C. Flow cytometric analysis was performed on a FACScan flow cytometer (BD Biosciences). Forward- and side-scatter parameters were used to gate on live cells. Results are expressed as percentage of cells staining above background. Memory subsets were defined as CD4+CD44highCD45RBlow and naive subsets were defined as CD4+CD44lowCD45RBhigh.
Total cellular RNA was extracted from the indicated T cell populations using TRIzol reagent (Invitrogen Life Technologies), digested with RNase-free DNase I (Invitrogen Life Technologies), and reverse-transcribed into cDNA using the Sensiscript RT kit (Qiagen) and random primers (Invitrogen Life Technologies) according to the manufacturer’s protocol. CCR7, CXCR4, S1P1, and S1P4 mRNA levels were quantified by real-time PCR using the QuantiTect SYBR Green PCR kit (Qiagen) with the LightCycler (Roche). PCR consisted of a 15 min 95°C denaturation step followed by 45 cycles of 15 s at 94°C, 20 s at 56°C, and 15 s at 72°C. The following primers (Sigma-Genosys) were used: S1P1 forward primer 5′-GGAGGTTAAAGCTCTCCGC-3′, reverse primer CGCCCCGATCTTCAAC; S1P4 forward primer CCACAGCCTCCTCATTGTC, reverse primer TCAGCATCCCTAGCCCTC; CCR7 forward primer TGCTTCAAGAAGGATGTGCGG, reverse primer GAGGAAAAGGATGTCTGCCACG; CXCR4 forward primer AGTGGCTGACCTCCTCTTTGT, reverse primer GCCCACATAGACTGCCTTTTC; cyclophilin A forward primer AGGGTGGTGACTTTACACGC, reverse primer ATCCAGCCATTCAGTCTTGG. Normalized values for mRNA expression were calculated as the relative quantity of the tested receptor divided by the relative expression of cyclophilin A. All samples were run in duplicate.
In vitro migration assays were performed as previously described (10). A total of 5 × 105 splenic T cells was incubated with various doses of FTY720 or S1P for the indicated incubation time at 37° C. The cells were washed twice, resuspended in RPMI 1640 containing 0.5% fatty acid-free BSA (Sigma-Aldrich), and added in a volume of 100 ml to the upper wells of a 24-well Transwell plate with a 5-mm insert (Corning). Lower wells contained various doses of chemokines in 600 ml of RPMI 1640/0.5% fatty acid-free BSA. The number of T cells that migrated to the lower well following a 2-h incubation were counted in three high-power fields using a hemocytometer.
The MS-1 pancreatic vascular endothelial cell line (American Type Culture Collection) were cultured in DMEM containing 5% FCS. Confluent MS-1 monolayers grown in flasks were gently trypsinized and seeded on either the upper surface (regular transendothelial monolayer) or lower surface (reverse transendothelial monolayer) of a Transwell plate with a 5-mm insert previously coated with 0.1% gelatin at 7.5 × 104 cells per insert. After 3–5 days of culture the integrity of the confluent monolayer was assessed with a modified Wright-Giemsa stain. The MS-1 monolayers in these studies were used up to passage 5. T lymphocyte migration across the monolayer was performed as described previously. For some experiments, 1.0 × 107 T cells/ml were incubated in the dark at room temperature with 5.0 mM CFSE (Molecular Probes) for 1 min. The staining process was stopped by adding 20 ml of PBS with 1% BSA and washed twice with PBS. Confluent MS-1 monolayers grown on 0.1% gelatin-covered Transwell plates were stained with 4.0 mM PKH26 (Sigma-Aldrich) solution. The staining process was stopped by adding 20 ml of PBS with 1% BSA and washing twice with PBS. CFSE-labeled splenic or LN T cells were migrated through PKH26 labeled MS-1 monolayers for 4 h at 37°C. The migrated cells in the lower wells were discarded and the unmigrated cells in the top wells were washed with PBS with 1% BSA. The transwell inserts were fixed with 3.7% paraformaldehyde and analyzed with confocal microscopy to examine the unmigrated T lymphocytes that were adherent to and within the MS-1 monolayer. Quantification of the adherent cells was performed by counting the attached cells in three high power fields.
Adoptive transfer assays with PKH26-labeled cells
Single-cell suspensions of T cells from the spleens or PLNs of wild-type mice were made. PKH26 solution (4.0 mM; Sigma-Aldrich) was added to a single-cell suspension containing 1.0 × 107 cells/ml and incubated at room temperature for 5 min. The staining process was stopped by adding 20 ml of PBS with 1% BSA and washing twice with PBS. A total of 2 × 107-labeled T lymphocytes in 300 ml of PBS was injected into the tail veins of the recipient mice; 0.3 mg/kg FTY720 or water via p.o. gavage was given to the experimental groups of mice and PLNs were harvested 18 h later. Cell counts were performed with a hemocytometer, and cell subsets were determined by flow cytometry.
In vivo migration results represent samples from two to three mice per experiment. In vitro migration results represent mean values of triplicate samples. All experiments were performed two to five times. SDs and p values were calculated with the Student t test using Microsoft Excel software.
Chemokine-driven splenic T cell migration is sensitized by S1PR stimulation
To address the issue of what signals and receptors are required to drive nonactivated and naive T cells into PLN, we investigated the interactions of chemokines and S1PR agonists in splenic T cell migration. Previous reports demonstrated that peripheral blood or splenic T cell migration to chemokines is enhanced by incubation with S1P, and that S1P can also act as a chemotactic agent by itself (3, 9, 10, 13). However, whether S1P acts primarily as a T cell chemotactic agent, primarily as a sensitization agent for chemokine receptors, or has both activities is unknown. To assess this question we performed a series of in vitro migration assays. Over a range of chemokine and S1P doses, purified splenic T cells showed enhanced migration to CCL19 or CXCL12 if they had been pretreated with 100 nM S1P for 15 min at 37°C (Fig. 1,A, results of optimal doses shown, other data not shown or in Ref.10). Since previous reports demonstrated that S1P acts as a direct T cell chemoattractant, we assessed in vitro migration of splenic T cells to S1P and its analogues FTY720 and P-FTY720 (3, 13). Purified splenic T cells migrated weakly to S1P alone (average of 8.3% of input cells from three separate experiments, data from optimal concentration shown), and failed to migrate to the other S1P analogues above background control levels (Fig. 1 B). Addition of S1P or its analogues to the upper well or to both wells of the plate resulted in no migration, indicating that migration to S1P is chemotactic but not chemokinetic, and that FTY720 and P-FTY720 are neither chemotactic nor chemokinetic. Because the magnitude of the effect of S1P on splenic T cell migration is greater when used in conjunction with chemokines than when used alone, we conclude that S1P primarily sensitizes chemokine-driven migration of splenic T cells, and secondarily acts as a weak, direct chemotactic agent.
Although S1P can enhance migratory responses to chemokines, we next asked whether S1P-S1PR is necessary for chemokine responses (9, 10). It has been shown that incubation of T cells with FTY720 or with S1P at varying doses or times produces reversible down-modulation and internalization of S1P1 from the cell surface, and S1P1 modulation then prevents in vitro migration responses to S1P. However, whether S1P1 modulation then desensitizes or in other ways affects other G protein-coupled receptors (GPCR), such as chemokine receptors, the function is unknown. To assess this issue, T cells were preincubated with FTY720 or S1P, washed, and then migrated to chemokines. The results demonstrate that murine splenic T cells show enhanced migration to CCL19 or CXCL12 after stimulation with various concentrations and incubation times of FTY720 or S1P, with the optimal concentration of 100 nM for S1P and 0.05 mg/ml FTY720, and the optimal incubation between 15 and 60 min at 37°C for S1P and FTY720 (Fig. 1, C and E). These results confirm the conclusions from the data in Fig. 1, A and B, that S1P primarily sensitizes chemokine-driven migration of splenic T cells. At FTY720 or S1P concentrations and incubation times greater than optimal conditions, which should cause S1PR desensitization and down-modulation (14), splenic T cell migration fails to exhibit enhanced migration. However, migration does not fall below the basal level obtained with chemokines alone, suggesting that desensitization or down-modulation of S1P1 does not affect chemokine receptor function or chemokine-driven migration of splenic T cells.
Analysis of the starting and migrated populations for T cell subsets by flow cytometry and for receptor expression by quantitative RT-PCR for CCR7, CXCR4, S1P1, and S1P4 was performed to determine whether there were subset preferences in the migration responses. The results revealed a preference for memory T cell migration to CXCL12 plus S1P compared with controls (Fig. 2,A), but this represents only a small subset of the whole T cell population (5–8%). Otherwise, there were no other significant differences among any of the other subsets or receptors for the various migration conditions, indicating that nonactivated and naive T cell subsets are responsive to chemokine plus S1PR stimulation (Fig. 2). Together, these results support the conclusions that S1PR stimulation sensitizes chemokine-driven migration of splenic T cells, sensitization is not required for chemokine-driven migration, desensitization does not impede chemokine-driven migration, and this is true for a variety of T cell subsets migrating to either CCL19 or CXCL12.
Peripheral T cell entry into PLNs involves migration through an endothelial cell monolayer from a luminal to abluminal direction (reviewed in Ref.2). To determine whether our in vitro migration results are relevant for this process, we next performed a series of in vitro transendothelial migration assays, using the murine MS-1 vascular endothelial cell line. Endothelial cell monolayers were grown to confluence on the upper surface of the transwell (designated as a regular monolayer) to simulate luminal to abluminal migration. Purified splenic T cells migrated across the endothelial monolayers to both CCL19 and CXCL12 in a fashion similar to wells that lacked the monolayer (Fig. 3). S1P enhanced T cell migration to CCL19 or CXCL12 across the endothelial monolayer. Because S1P enhances both transendothelial and nontransendothelial splenic T cell migration, it is likely that S1P enhanced, chemokine-driven, splenic T cell migration observed in vitro is relevant for the mechanisms of T cell entry into PLNs in vivo.
S1P but not FTY720 sensitizes CXCL12-driven splenic T cell migration
We concluded above that S1PR stimulation primarily sensitizes CCL19 and CXCL12-driven splenic T cell migration. We also previously observed that FTY720 did not enhance CXCL12-driven in vitro migration, but did reveal a role for CXCR4 in PLN T cell accumulation in vivo (10). Because S1P and FTY720 function by binding the same S1PRs, we readdressed this apparent contradiction by examining splenic T cell migration to CCL19 and CXCL12 while stimulating S1PR with S1P or FTY720. As shown in Fig. 4,A, both FTY720 and S1P enhance splenic T cell migration to CCL19, but only S1P and not FTY720 enhances migration to CXCL12. The inability of FTY720 to enhance CXCL12-driven migration was true over a range of FTY720 concentrations and incubation times (not shown). To determine whether this response was a reflection of specific T cell subsets migrating preferentially to CCL19 or CXCL12 following S1P or FTY720 administration, the starting and migrated populations were analyzed by flow cytometry and quantitative RT-PCR. No significant differences were found in comparing the groups, indicating that movement of all the different nonactivated and naive T cell subsets to chemokines occurs regardless of the chemokine used (CCL19 or CXCL12), the S1PR agonist used (S1P or FTY720), or chemokine (CCR7, CXCL4) or S1P (S1P1, S1P4) receptor transcriptional expression (Figs. 2, A–E, and 4, B and C). We conclude that S1P and FTY720 stimulation of S1PR are functionally and mechanistically different, resulting in different migration responses depending on the chemokine receptor costimulated at the same time as S1PR. This distinction between S1PR ligands is supported by our previous results and those of others that show differences between S1P and FTY720 in their interactions with multidrug transporters and in causing S1PR internalization (3, 4, 5, 14). The results also demonstrate that CXCR4-driven migration is mechanistically different from CCR7-driven migration, and that this difference is not simply due to different responding T cell subsets. Lastly, these findings resolve the contradiction between in vitro and in vivo findings with regard to the role of CXCR4 and S1PR in T cell migration.
S1P sensitizes only CXCL12-driven but not CCL19-driven PLN T cell migration
To address the issue of what signals and receptors are required to drive nonactivated and naive T cells out of PLNs, we investigated the interactions of chemokines and S1PR agonists in PLN T cell migration. As shown by our results above and other published reports, S1P enhances chemokine-driven peripheral blood and splenic T cell migration in vitro and affects S1PR expression (9, 10, 14). However, it is not known whether other sources of T lymphocytes behave in a fashion similar to peripheral blood and splenic T cells, or whether the passage of nonactivated T cells through an endothelial monolayer, such as PLN-derived T cells, alters chemokine and/or S1P-driven migration. To assess these questions, we conducted a series of in vitro migration assays using PLN, rather than splenic, T cells as the source of migrating cells. The results demonstrate that PLN T cells migrate to both CCL19 and CXCL12, but S1P treatment enhances T cell migration only to CXCL12 but not CCL19 (Fig. 5,A, results of optimal doses shown). Further, S1P alone is not chemotactic for PLN T cells. S1P-enhanced migration is dose- and incubation time-dependent, with the optimal dose of 100 nM and incubation time of 15 min at 37°C, similar to splenic T cells. T cell migration after supraoptimal doses or incubation times with S1P does not fall below basal T cell migration to CXCL12 (Fig. 5 B), indicating that desensitization or down-modulation of S1P1 receptors does not affect CXCR4-CXCL12 signaling and migration, and that S1PR stimulation is not required for PLN T cell chemokine-driven migration. Thus, the results seem to suggest that PLN T cells, like splenic T cells, require chemokines for migration; and S1P enhances but is not required for migration. There is also the striking finding that splenic but not PLN T cell migration to CCL19 is enhanced by S1P, while both cell populations migrate similarly to CXCL12 plus S1P, supporting the conclusion above that CXCR4- and CCR7-driven migration are mechanistically distinct, and the additional conclusion that naive and nonactivated splenic and PLN T cells have distinct migration mechanisms.
Flow cytometric analysis of the starting and CXCL12-migrated populations following S1P treatment shows enhanced migration of all T cell subsets without a subset preference as a result of S1P stimulation (Fig. 5,C). Flow cytometric analysis of S1P-treated PLN T cells migrating to CCL19 shows that all T cell subsets also migrate in response to this chemokine, and that there is enhanced migration of memory T cells to CCL19 plus S1P (Fig. 5,D). Because the memory subset is a small minority of the entire population, there is no enhancement of migration of the whole PLN T cell population to CCL19 plus S1P (Fig. 5, A and D). Quantitative RT-PCR for chemokine and S1PRs of the starting and migrated PLN T cells to CXCL12 or CCL19 following S1P treatment demonstrates no significant subset differences, indicating that chemokine or S1PR transcriptional expression are not the primary determinants of S1P-enhanced T cell migration (Fig. 5, E–G). We conclude that the differential response of PLN T cells to CXCL12 or CCL19 with S1P is not due to different responding T cell subsets.
In vitro transendothelial migration of PLN T cells through either regular murine MS-1 monolayers, representing luminal to abluminal migration, or reverse monolayers, representing abluminal to luminal migration, confirms that S1P treatment enhances migration to CXCL12 but not to CCL19 (Fig. 6). Because S1P enhances PLN T cell transendothelial and nontransendothelial migration, it is likely that S1P enhanced, chemokine-driven PLN T cell migration observed in vitro is relevant for the mechanisms of egress of peripheral T cells from PLNs in vivo.
S1P sensitizes while FTY720 inhibits CXCL12-driven PLN T cell migration
As noted above, the results in Fig. 5,B seem to suggest that PLN like splenic T cells require chemokines for migration, while S1P-S1PR enhances but is not required for migration. This conclusion seems to contradict reports which demonstrate that absence of S1P1 prevents PLN egress, and that FTY720 administration in vivo prevents PLN T cell egress due to S1P1 down modulation or desensitization (3, 4, 5). Because FTY720 more effectively and persistently down-modulates S1P1 compared with S1P (14), to resolve this issue we evaluated whether FTY720 could influence chemokine-driven PLN T cell migration. As shown in Figs. 7, A and B, S1P enhances, whereas FTY720 partially inhibits, PLN T cell migration to CXCL12. In eight experiments, the inhibition varied from 18 to 25%. Neither S1P nor FTY720 influence CCL19-driven PLN T cell migration (Figs. 5,A, 6, and 7D), and FTY720 does not influence CXCL12-driven splenic T cell migration (Fig. 4,A). Flow cytometric analysis and quantitative RT-PCR of the starting and migrated PLN T cell populations following CXCL12 and FTY720 treatment show no significant subset differences, indicating that the diminished T cell movement is not associated with particular T cell subsets, or chemokine or S1PR transcriptional expression (Figs. 5, E and F, and 7C). These results show that S1P sensitizes both splenic and PLN T cell migration to CXCL12, while FTY720 is not able to sensitize the response of either T cell population to this chemokine. In fact, FTY720 partially inhibits PLN T cell migration to CXCL12. Conversely, splenic T cell responses to CCL19 are sensitized by both S1P and FTY720, while PLN T cell migration to CCL19 is not sensitized by either S1PR ligand. We conclude as above that splenic and PLN T cell migration responses are functionally and mechanistically distinct, that CXCR4- and CCR7-driven migration mechanisms are distinct, and that S1P and FTY720 are functionally distinct S1PR ligands. Although PLN T cell migration to CXCL12 is only modestly inhibited by FTY720, these results partly resolve the apparent contradiction and suggest that by using the more potent S1P1 modulator there may be a requirement for S1PR function in PLN T cell migration.
FTY720 preferentially inhibits CXCL12-driven transendothelial migration of PLN T cells
Because migration also involves traversing endothelial cell layers, we next determined whether there was an additional effect of FTY720 on this step. The results in Fig. 8,A show that the inhibition of migration is even more pronounced if PLN T cells are treated with FTY720 and then migrated to CXCL12 across an endothelial cell layer (35–50% inhibition in six experiments). The decreased migration of FTY720-treated PLN T cells to CXCL12, both in the presence and absence of endothelial cells, compared with FTY720-untreated T cells suggests that modulation of S1P1 by FTY720 does impair migration, and that both T cell autonomous and T cell-endothelial cell interactions are targets of the inhibition. This further resolves the apparent contradiction noted above, in that S1PR is required for efficient transendothelial migration but is relatively less important for nontransendothelial migration, and S1P is a less effective S1PR modulator than FTY720. Importantly, the same FTY720 treatment of splenic T cells does not inhibit transendothelial or nontransendothelial migration to CXCL12 (Figs. 8 B and 4A), demonstrating that the S1P1 function required for CXCL12-driven transendothelial migration is specific to PLN T cells.
To assess more specific effects of FTY720 on transendothelial migration to CXCL12, CSFE-labeled splenic or PLN T cells were treated or untreated with FTY720 or S1P, migrated across PKH26 labeled MS-1 endothelial cells, washed, and confocal microscopy was used to analyze the plates for adherent T cells. Fig. 8, C and D, demonstrate that FTY720-treated PLN T cells are retained within the endothelial monolayer compared with S1P-treated or control PLN T cells. This finding suggests that T cells initiate migration after chemokine stimulation, are able to enter the endothelial layer, but are unable to complete transmigration, perhaps due to changes in T cell adhesion or T cell-stimulated disruption of endothelial adherens junctions. In contrast, splenic T cells treated with S1P but not FTY720 were also found in increased numbers within the endothelial cells. Because S1P but not FTY720 enhances CXCL12-driven splenic T cell migration (Fig. 4 A), this could represent more T cells in transit through the endothelial layer.
To determine whether in vitro and in vivo findings are consistent, we adoptively transferred PKH26-labeled splenic and PLN T cells to recipients treated or untreated with FTY720. Fig. 8,E demonstrates that adoptively transferred splenic and PLN T cells can migrate to PLNs of control recipients. Following FTY720 administration, splenic, but not PLN, T cells are further sequestered within PLNs. This finding is consistent with the results in Fig. 7, A and B, that PLN T cell migration is not enhanced by FTY720. These findings further confirm the relevance of the in vitro results to in vivo trafficking.
The findings presented here demonstrate that the role of S1PR-driven sensitization or enhancement of chemokine-driven migration of nonactivated and naive T cells is dependent on the anatomic source of the T cells, the specific chemokine receptor stimulated, the specific S1PR agonist, and the type of migration measured. With regard to the anatomic source of the nonactivated migrating T cells, splenic T cells can be sensitized by either S1P or FTY720 depending on the chemokine receptor engaged, while PLN T cells can be sensitized only by S1P, and sensitization is also chemokine receptor specific. With regard to the specific chemokine receptor responses, CCR7 can be sensitized by S1P or FTY720 but only in splenic T cells, while CXCR4 can be sensitized only by S1P but in either splenic or PLN T cells. We have observed virtually identical results for CCL2-CCR2 as for CCL19-CCR7 migration (our unpublished results). With regard to the specific S1PR agonist, S1P can sensitize both splenic and PLN T cells and both CCR7 and CXCR4 responses, while FTY720 can sensitize only splenic T cells and only CCR7. The distinction between FTY720 and S1P confirm a previous report that FTY720 has more potent antagonist effects than S1P on S1PRs (14), and as shown here the antagonist activity reveals a specific requirement for S1PR function for PLN T cell transendothelial migration. With regard to the type of migration measured, S1PR stimulation is most important for PLN T cell transendothelial migration to CXCL12, and less important for PLN T cell nontransendothelial migration. S1PR stimulation does not appear to be required for transendothelial or nontransendothelial migration of splenic T cells to CCL19 or CXCL12. These results have important implications for our understanding of T cell entry into and egress from PLNs, the mechanisms of chemokine receptor and S1PR signaling integration, and transendothelial T cell migration.
T cell entry into PLNs
Naive T cell entry into PLNs occurs following the rolling, tethering, arrest, and diapedesis of the T cell on HEVs. CXCL12, CCL19, and CCL21 are presented by the endothelial cells and are responsible for the arrest of the cell and eventual migration across the monolayer and into the PLN (reviewed in Refs.1 and 2). Previous reports demonstrated that S1P and FTY720 sensitize or enhance peripheral T cell migration to chemokines suggesting that sequestration of T cells within PLNs is dependent on both chemokine and S1PR engagement on the T cell surface (9, 10, 16). Recently, Matloubian et al. (3) demonstrated that adoptively transferred S1P1−/− thymocytes obtained from fetal liver chimeras readily enter secondary lymphoid organs and migrate to appropriate T cell compartments. Conversely, in a variety of studies, a lack of chemokines or chemokine receptors prevents T cell trafficking and accumulation in the PLN, even if S1PR are stimulated (9, 10, 17). Our current results indicate that peripheral splenic T cells show enhanced migration to chemokines in the presence of S1PR stimulation, but can still migrate if S1P agonists are absent or if S1P1 is desensitized or down-modulated. Taken together, these results demonstrate that although S1P and chemokine receptors both play a role in promoting migration to the PLN, chemokines do, and S1PR do not, have an obligate role in PLN entry because in the absence of S1P1 naive peripheral T cells are still able to enter PLNs.
T cell egress from PLNs
Following activation, naive T cells down-regulate the CCR7 and/or CXCR4 chemokine receptors needed for PLN T cell entry, and up-regulate the chemokine receptors needed for PLN egress via efferent lymphatics and subsequent migration to the inflammatory nidus (reviewed in Ref.1). Whether there are other signals such as fugetaxis, adhesion molecules, or other entities that further regulate activated T cell PLN egress is not certain (18, 19, 20, 21, 22, 23). Further, the mechanisms of PLN egress of nonactivated or naive T cells are still mostly unknown. This is an issue of great importance because only a small percentage of T cells that enter the PLN are activated by cognate Ag, so that most cells leaving the PLN are nonactivated.
Although there are five known S1PRs, only S1P1 has been implicated as playing a role in T cell migration (3, 4, 5). Graler et al. (13, 14) demonstrated that naive T cells obtained from the spleen express S1P1 and after treatment with either FTY720 or S1P reversibly internalize or down-regulate the receptor, inhibiting the in vitro chemotactic response to S1P. Presence or absence of S1P1 is a dominant regulator of egress of nonactivated T cells from the thymus and secondary lymphoid organs (3, 4, 5). Therefore, homeostatic migration of nonactivated cells from PLNs seems dependent on S1P-S1PR. Our results here show that PLN T cells migrate in response to both chemokine and S1PR activation. Both signals are required to varying degrees for transendothelial and nontransendothelial migration, and interference with S1PR function inhibits chemokine-driven migration. Inhibition of PLN T cell migration by FTY720 even in the absence of endothelial cells suggests heterologous desensitization of the chemokine GCPR by the S1P GCPR (24). Further, this is specific for CXCL12-CXCR4 but not CCL19-CCR7-driven migration (or CCL2-CCR2-driven migration, our unpublished data). These findings demonstrate that different chemokine receptors are not coupled to S1PR signaling pathways in identical manners, so that the molecular pathways of migration are not the same for all chemokines. These findings also demonstrate that, unlike splenic T cells, PLN T cells do require S1PR stimulation or function for transendothelial migration in response to CXCL12. The differences between splenic and PLN T cell chemokine receptors and migration apparati that result in such different migration responses are not currently known, but the results here suggest that the mechanisms are distinct from simple changes in chemokine receptor and S1PR expression. The results also imply that CXCL12 may be one of the chemokines responsible for PLN egress of recirculating nonactivated T cells. Because CXCL12 also plays an important role in naive T cell PLN entry (reviewed in Ref.2), this particular issue is not easily addressed by simple blockade or deletion of CXCL12 or CXCR4.
Endothelial cell and PLN T cell interactions
T cell egress from PLN is dependent on migration across an endothelial cell monolayer into the efferent lymphatics (reviewed in Ref.2). FTY720 and S1P promote sequestration of T cells within PLNs through mechanisms related to down-modulation of S1P1 on the T cell membrane following FTY720 or S1P binding to the receptor (3, 4, 5). Our data suggest that the primary locus of S1PR activity is not inhibition of chemokine-driven migration by heterologous desensitization, but rather the interaction of T cells with lymphatic endothelial cells. Although only a little is known about the molecular mechanisms of T cell transendothelial migration, far more is known about monocyte and neutrophil transendothelial migration, although this is mostly limited to HEV or other vascular endothelia and not the lymphatic endothelium about which less is known (18, 25, 26). It is important to acknowledge that a limitation to our results is the use of vascular endothelial cells, because murine lymphatic endothelial cells and lines are not readily and reproducibly available, and with uncertainty as to whether those cells reliably reproduce the in vivo environment. Specific homophilic and heterophilic receptor interactions between leukocytes and endothelial cells result in both outside-in and inside-out signaling, initiating a complex process that results in transient dissolution and separation of endothelial adherens junctions to permit leukocyte transmural migration (17, 18, 27). The data here suggest that S1PR function is required for this step in T cell PLN egress. Shimonaka et al. (28) demonstrated that activation of the regulatory molecule Rap1 by CXCR4 induces redistribution of CXCR4 and CD44 and promotes T cell migration across an endothelial cell monolayer. Whether sensitization of S1PRs modulates this activity or other molecules, such as PI3K or protein kinase C, is unknown and is an avenue for further investigation (23, 29). It is also important to note that S1P1 is also present on endothelial cells, and S1P and FTY720 may therefore alter T cell migration indirectly through endothelial cells. Activation of endothelial S1P1 induces VE-cadherin and B-catenin and promotes adherens junction formation, thereby preventing T cell transendothelial migration (30, 31). It is unlikely that this mechanism explains PLN T cell sequestration because lymphatic endothelial cells do not express VE-cadherin and B-catenin (26). Lymphatic endothelium does express other components of tight junctions, but it is not known whether activation of S1PR induces tight junction formation (26).
Our data demonstrate that the variables of source of T cells (splenic vs PLN), chemokine ligand (CCL19 vs CXCL12), and S1PR agonist (S1P vs FTY720) determine the presence or absence of migration. In Fig. 9, we propose a model where T cell migration is dependent on the activation of different signaling pathways that vary according to the anatomic source of T cells and the expression or function of chemokine and S1PRs. In this model S1P and FTY720 activate different signals in different cell types, and different chemokine receptors are susceptible to different signals in different cell types. It is not currently known what secondary messenger systems are responsible for the differences observed in the patterns of S1P1 sensitized chemokine-driven migration across an endothelial cell monolayer, and this will clearly be an area of active investigation in the future.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by National Institutes of Health Grant R01 AI41428 (to J.S.B.).
Abbreviations used in this paper: PLN, peripheral lymph node; HEV, high endothelial venule; S1P, sphingosine 1-phosphate; GPCR, G-protein coupled receptor.