Report

Promotion of Lymphocyte Egress into Blood and Lymph by Distinct Sources of Sphingosine-1-Phosphate

See allHide authors and affiliations

Science  13 Apr 2007:
Vol. 316, Issue 5822, pp. 295-298
DOI: 10.1126/science.1139221

Abstract

Lymphocytes require sphingosine-1-phosphate (S1P) receptor-1 to exit lymphoid organs, but the source(s) of extracellular S1P and whether S1P directly promotes egress are unknown. By using mice in which the two kinases that generate S1P were conditionally ablated, we find that plasma S1P is mainly hematopoietic in origin, with erythrocytes a major contributor, whereas lymph S1P is from a distinct radiation-resistant source. Lymphocyte egress from thymus and secondary lymphoid organs was markedly reduced in kinase-deficient mice. Restoration of S1P to plasma rescued egress to blood but not lymph, and the rescue required lymphocyte expression of S1P-receptor-1. Thus, separate sources provide S1P to plasma and lymph to help lymphocytes exit the low-S1P environment of lymphoid organs. Disruption of compartmentalized S1P signaling is a plausible mechanism by which S1P-receptor-1 agonists function as immunosuppressives.

Sphingosine-1-phosphate (S1P) functions as an extracellular signaling molecule by engaging five G protein–coupled receptors, designated S1P1 to S1P5, that play a key role in vascular development and lymphocyte trafficking among other processes (1). Although S1P is believed to be generated intracellularly by all cell types during sphingolipid degradation (2), the sources of extracellular S1P in vivo have not been well defined. Lymphocytes exit from lymph nodes into lymph and from the spleen into blood and are thought to exit from the thymus into blood. Studies with S1P1-deficient mice revealed a necessary role for this receptor in lymphocyte egress from all three organ types (36). However, experiments with S1P-neutralizing antibodies and S1P1 antagonists have not revealed inhibitory effects on lymphocyte egress (710). By contrast, synthetic S1P1 agonists block lymphocyte egress and are in clinical development because of their immunosuppressive properties (11). It has been unclear how genetic S1P1 deficiency can have the same outcome as increased S1P1 agonism and a different outcome from S1P1 antagonism. Does S1P act directly on the lymphocyte to promote egress (46, 12), or does it act indirectly, as pharmacological manipulations of S1P1 function might suggest (711)?

To investigate the source of extracellular S1P and clarify its role in egress, we generated mice deficient in the two enzymes responsible for S1P production, sphingosine kinase-1 (Sphk1) and sphingosine kinase-2 (Sphk2) (2, 13). Because deletion of either kinase alone did not have a profound effect on circulatory S1P and combined deletion led to embryonic lethality at mid-gestation (1417), we used a conditional gene deletion approach (fig. S1, A to E). Sphk1f/–Sphk2–/– pups carrying the Mx1-Cre transgene were treated 3 to 5 days after birth with polyinosine polycytidylic acid (pI-pC) to activate the Mx1 promoter and Cre expression. Such induction of Mx1-Cre causes efficient excision of floxed alleles in hematopoietic cells, vascular endothelium, and liver, among other organs (fig. S1F) (18, 19). We further established that Mx1-Cre is active in lymphatic endothelium (fig. S1, G and H). pI-pC–treated Sphk1f/–Sphk2–/– Mx1-Cre mice, hereafter designated “Sphk-deficient” mice, survived to adulthood and were indistinguishable from littermates in appearance.

By liquid chromatography followed by tandem mass spectrometry (LC/MS/MS), S1P was undetectable in plasma and lymph of Sphk-deficient mice, whereas the concentrations of S1P in plasma and lymph of controls were in the micromolar and the 100-nanomolar range, respectively (Fig. 1A). Similar differences were obtained in a bioassay of S1P activity (fig. S1I). Furthermore, flow cytometric analysis of the lymphocytes that could be identified in blood and lymph of the Sphk-deficient mice revealed high surface S1P1 levels compared with controls, where receptor expression was undetectable (Fig. 1B). S1P1 is highly sensitive to ligand-induced internalization, and thymocytes down-modulate S1P1 almost completely when exposed to 1 nM S1P (12). The high surface S1P1 on circulating lymphocytes in the Sphk-deficient mice is consistent with a lack of circulatory S1P. S1P1 levels on cells from the thymus, spleen, and lymph nodes of Sphk-deficient mice overlapped with S1P1 levels on blood and lymph lymphocytes from these mice (Fig. 1C and fig. S1J), suggesting that extracellular S1P is equally low in all sites.

Fig. 1.

Sphk-deficient mice have no detectable S1P in plasma or lymph. (A) LC/MS/MS quantification of plasma and lymph S1P. The limit of detection in this assay was about 50 nM (5 × 10–2 on ordinate). Δ indicates Sphk2–/– mice in which one Sphk1 allele was null and one Sphk1 allele was excised; F indicates Sphk2–/– littermate controls in which at least one functional Sphk1 allele was present. Both Δ and F mice were treated with pI-pC 3 to 5 days after birth. (B) S1P1 levels on CD4+CD62Lhi cells from blood or lymph of the indicated mice, measured by flow cytometry. Shaded histograms show staining of blood CD4+CD62Lhi cells with a control antibody. Each plot is representative of at least three experiments. (C) S1P1 levels on CD4+CD62Lhi cells from the indicated organs. Each plot is representative of at least three experiments.

The number of T cells in blood and lymph of Sphk-deficient mice was markedly reduced, whereas spleen and lymph node T cell numbers were similar compared to controls (Fig. 2, A to D). Furthermore, there was an accumulation of mature T cells in the thymus without apparent changes in other thymic subsets (Fig. 2E). These observations indicate that Sphk is required for T cell egress from peripheral lymphoid organs and from the thymus. The presence of close-to-normal T cell numbers in peripheral lymphoid organs despite the defect in thymic egress could be due to the persistence of some Sphk-expressing cells up to a week or more after birth. B cell counts in blood and lymph of Sphk-deficient mice were also significantly lower than those in controls (Fig. 2, A and C), although B cell numbers in the lymphoid organs of Sphk-deficient mice appeared normal (Fig. 2, B and D, and fig. S2A). B cell egress from the bone marrow, which occurs in the absence of S1P1 (4), presumably provided a source of peripheral B cells that entered but failed to exit lymphoid organs in Sphk-deficient mice. It has been suggested that S1P1, and thus possibly S1P, might be needed for normal lymphocyte development rather than for egress (11). However, short-term adoptive transfer experiments indicate that the defective egress reflects a proximal requirement for S1P (fig. S2B).

Fig. 2.

Lymphocytes fail to egress from thymus and secondary lymphoid organs in Sphk-deficient mice. (A to E) Cell numbers in the indicated organs or fluids from control or Sphk-deficient mice. Lymphnode numbers are for the mesenteric lymph nodes. All enumerated cells (except double-positive thymocytes) were CD62Lhi; additionally, splenic CD19+ cells were IgDhi and AA4.1lo, and mature CD4 and CD8 single-positive (SP) thymocytes were CD69lo. Asterisks indicate P ≤ 0.05.

To determine whether the source of S1P that promotes lymphocyte egress is hematopoietic in origin, we performed bone marrow reconstitution studies. In lethally irradiated wild-type mice transplanted with Sphk-deficient marrow, plasma S1P levels were reduced by 10-fold compared with controls, although lymph S1P was normal (Fig. 3A and fig. S3A). The remaining plasma S1P concentration was above the low nanomolar range sufficient to down-modulate lymphocyte S1P1 in vitro (12), and S1P1 on blood lymphocytes remained low (Fig. 3B). Lymphocyte counts in both blood and lymph were in the normal range, and mature T cells did not accumulate in the thymus (Fig. 3C and fig. S3, B and C). These data suggest that S1P in plasma is supplied mainly by hematopoietic cells and that S1P in lymph is supplied by an independent source. They also suggest that radiation-resistant cells generate sufficient S1P to support lymphocyte egress into blood and lymph.

Fig. 3.

A hematopoietic source supplies most plasma S1P, whereas a radiation-resistant source supplies lymph S1P. (A to C) Lethally irradiated wild-type (WT) mice were reconstituted with Sphk-deficient or littermate control bone marrow. (D to F) Lethally irradiated Sphk-deficient or control mice were reconstituted with wild-type bone marrow. [(A) and (D)] LC/MS/MS quantification of plasma and lymph S1P in the indicated mice. [(B) and(E)] S1P1 levels on CD4+CD62Lhi cells from the indicated organs and fluids, measured by flow cytometry. Each plot is representative of at least three experiments. Experiments shown in bottom graphs in (B) were done on the same day and used the same background control. [(C) and (F)] Cell numbers in the indicated organs and fluids. Lymph node numbers are for the mesenteric lymph nodes. All enumerated cells were CD62Lhi; additionally mature CD4 single-positive thymocytes were CD69lo. For lymph, relative concentration (rel. conc.) was determined as described in (13). Asterisks indicate P ≤ 0.05.

In lethally irradiated Sphk-deficient mice reconstituted with wild-type bone marrow, the S1P level in plasma was restored to the normal range, but S1P in lymph remained undetectable (Fig. 3, D and E, and fig. S3D). Egress from the thymus was largely restored: Mature CD4 single-positive thymocytes accumulated only ∼50% above control (Fig. 3F) compared with a greater than fourfold accumulation in Sphk-deficient mice (Fig. 2E). Although some studies have suggested thymic egress can occur via lymphatics (2022), this finding supports the view that exit can occur into the blood. T and B cell numbers were dramatically reduced in the lymph (Fig. 3F and fig. S3, E and F), suggesting a persistent block in lymphocyte egress from lymph nodes. Furthermore, there was a substantial decrease in the number of T and B cells in the spleen and blood and an increase in the lymph nodes (Fig. 3F and fig. S3, E and F). This suggested that splenic lymphocytes were exiting into the blood, traveling to lymph nodes, and then failing to exit, resulting in a general block in recirculation. Because the number of CD4+ T cells in the entire lymph node compartment is substantially larger than the number in the spleen, their fractional decrease in spleen was easier to detect than their fractional increase in lymph nodes. These data are again consistent with a model in which plasma S1P is supplied predominantly by hematopoietic cells and lymph S1P is derived from a separate source(s). Moreover, although the hematopoietic source is sufficient to support lymphocyte egress from thymus and spleen, only the separate Mx1-Cre–sensitive, radiation-resistant source can support lymphocyte egress from lymph nodes.

Previous studies have suggested that platelets might be an important source of plasma S1P (2325). However, NF-E2–deficient mice had normal plasma S1P concentrations (Fig. 4A) despite having virtually no circulating platelets (26). Analysis of RAG1-deficient mice that lack T and B cells also did not reveal any reduction in plasma S1P (Fig. 4A). These observations and our finding that plasma and lymph S1P are separately maintained led us to investigate whether red blood cells (RBCs) are a source of plasma S1P. Wild-type RBCs (fig. S4, A to F) were transferred to Sphk-deficient mice in amounts sufficient to represent at least 20% of circulating RBCs in the recipients. Within 36 hours of wild-type RBC transfer, plasma S1P in Sphk-deficient mice was restored (Fig. 4B and fig. S4G). RBCs have abundant sphingosine kinase activity (fig. S4H), as expected (27). Moreover, although all cell types except platelets have been thought to express the S1P-degrading enzyme S1P lyase (2, 24), we observed that RBCs contain minimal lyase activity (Fig. 4C). This low lyase activity, together with their abundance, suggests a role for RBCs as the main source of plasma S1P and accords with the inability of lyase inhibitors to alter blood S1P levels (12).

Fig. 4.

Plasma S1P promotes lymphocyte egress into the blood in an S1P1-dependent manner. (A) LC/MS/MS quantification of plasma S1P from the indicated mice. (B) Sphk-deficient or control mice received 0.35 ml of packed RBCs (about 2 × 109 cells) or vehicle 36 hours before analysis. Plasma S1P, as well as S1P in the RBC preparation, was quantified by LC/MS/MS. (C) S1P lyase activity. Tritiated dihydroS1P (dhS1P) was incubated with lysates of RBCs, splenocytes (splen.), or homogenization buffer alone (substr.). The hexadecanal (Hex) product was separated by thin-layer chromatography and visualized by autoradiography. Some dephosphorylation of dhS1P to dihydrosphingosine (Sph) also occurred. Results are representative of three experiments. (D and E) Sphk-deficient mice were infused intrajugularly with S1P or vehicle for 4 hours. (D) S1P1 levels on blood CD4+CD62Lhi cells, measured by flow cytometry. Plot is representative of two experiments. (E) Blood counts of the indicated cell types were measured before and after infusion, and the fold change is plotted. All enumerated lymphocytes were CD62Lhi. (F) Experimental design for (G) to (I). (G and H) Cell numbers in the indicated organs of RBC-transfused or control mice. Enumerated cells were CD4+CD62Lhi; in addition, mature CD4SP thymocytes were CD69lo. In (H), the left plot shows the number of mature CD4SP thymocytes, and the right plot shows the ratio of the number of mature CD4SP thymocytes in the thymus of a Sphk-deficient mouse to the number in its identically treated Sphk-sufficient littermate control; each point represents an individual pair. (I) Mature CD4SP S1P –/–1 and S1P +/1 thymocytes were cotransferred into Sphk-deficient mice that also received RBC or vehicle. Transferred CD4+CD62Lhi cells were enumerated in spleen and lymphnodes after 36 hours. Four experiments are shown; lines connect pairs of mice in the same experiment. Circles and squares in (G) to (I) represent individual mice as indicated in (F). Asterisks indicate P ≤ 0.05.

The ability of hematopoietic cell–derived S1P to support lymphocyte egress from the thymus and spleen and the evidence that RBCs are the major hematopoietic S1P source suggested that thymocytes and spleen cells egress in direct response to plasma S1P. Thus, we asked whether providing S1P in the circulation might be sufficient to restore egress of cells from lymphoid tissues into blood. S1P infusion into Sphk-deficient mice down-modulated S1P1 surface expression on blood T cells and indeed led to an increase in the number of B and T cells in the blood within 4 hours (Fig. 4, D and E). To perform longer-term experiments, we used RBC transfusion. Three RBC transfusions over 9 days led to an increase in plasma but not lymph S1P (fig. S4I). RBC transfusion by this protocol or transfusion once with analysis at 36 hours produced a decrease in T and B cell numbers in the spleen and an accumulation in the lymph nodes of Sphk-deficient mice (Fig. 4, F and G). There was also a reduction in the number of mature T cells in the thymus that was most obvious when the data were analyzed as the ratio of counts from identically treated littermate pairs (Fig. 4H). To test whether the RBC transfusions act by providing S1P to stimulate S1P1 on the lymphocyte or by some indirect mechanism, we cotransferred fluorescently labeled S1P –/–1 and +/ (+/+ or +/–) lymphocytes into Sphk-deficient mice, either 1 day after the third RBC transfusion in the 9-day protocol or 12 hours before the single RBC transfusion in the 36-hour protocol (Fig. 4F). RBC transfusion reduced the number of S1P1+/ cells in the spleen and increased their number in lymph nodes, but the numbers of S1P1–/– cells in the spleen and lymph nodes were unaffected (Fig. 4I). These findings suggest that RBC-derived plasma S1P is able to promote lymphocyte egress from the spleen in a manner that depends on lymphocyte-intrinsic expression of S1P1.

The above findings identify a previously unknown, endocrine-like function for RBCs as a major producer of plasma S1P. We also make the unexpected finding that plasma and lymph S1P are maintained as separate compartments. Albumin and high-density lipoproteins, major S1P carriers, are thought to be supplied to lymph from transudated plasma and are present in lymph at concentrations within a fewfold of those in plasma (28). Our data suggest that S1P is removed from transudated plasma and “resupplied” to lymph by a distinct source, perhaps lymphatic endothelial cells, which we have shown are a radiation-resistant Mx1-Cre target (fig. S1, and H). This may be part of a general mechanism that keeps S1P abundance in tissues low, ensuring S1P-receptor–expressing cells are sensitive to local increases in S1P, as may occur at a site of tissue injury or inflammation (2, 24).

Taken together, our data are consistent with a model in which egress from thymus, spleen, and lymph nodes is promoted by S1P produced by radiation-resistant cells at egress sites, as well as, in the case of thymus and spleen, by plasma S1P. S1P at exit sites enables egress by acting on S1P1 expressed by lymphocytes. In accord with studies of S1P1-deficient mice (46), this model predicts that antagonism of S1P1 should inhibit egress; the inability of the antagonists reported thus far to do so (710) is likely due to incomplete receptor blockade or ligand neutralization, because our studies of wild-type mice reconstituted with Sphk-deficient marrow revealed that even when plasma S1P levels were reduced by 95%, lymphocyte egress occurred normally. This model also predicts that activation and down-regulation of lymphocyte S1P1 within lymphoid tissue, either by elevating tissue S1P by inhibition of lyase activity (12) or by direct activation of S1P1 with the lyase-insensitive agonist FTY720 (11), would disrupt compartmentalization of S1P signaling, which is normally low in lymphoid organ interstitium and high at exit sites, and inhibit egress. Thus, this model accommodates the paradox that SIP1 deficiency and SIP1 agonism can both yield defective lymphocyte egress.

Supporting Online Material

www.sciencemag.org/cgi/content/full/1139221/DC1

Materials and Methods

Figs. S1 to S4

References

References and Notes

View Abstract

Stay Connected to Science

Navigate This Article