FTY720 stimulates multidrug transporter– and cysteinyl leukotriene–dependent T cell chemotaxis to lymph nodes (original) (raw)
FTY720 stimulates multidrug transporter efflux activity. Given the importance of Abcc1 and Abcb1 in DC migration to the LNs and the fact that FTY720 bears structural features consistent with Abcb1 and Abcc1 substrates, we examined whether FTY720 would affect the function of these transporters in T cells in vitro. We assessed their functional activity by loading purified T cells with a fluorescent transport substrate in the presence of various doses of FTY720, and measuring the units of fluorescence retained in the cells over a period of time (23). Loss of fluorescence at 37°C indicates transport of the dye from the intracellular to the extracellular compartment, while cells held at 4°C do not export dye (7, 24). The substrate Fluo-3 requires Abcc1 for transport (13), while the substrate DiOC2 is transported by Abcb1 more efficiently than by Abcc1 (G.J. Randolph, unpublished data).
Figure 1a shows that T cells from wild-type mice treated with FTY720 had a dramatic increase in Abcb1-mediated DiOC2 efflux compared with controls. This dose of FTY720 had no adverse effects on cell viability. Abcc1 activity in resting T cells was very low but was significantly enhanced by FTY720 (Figure 1b). FTY720-mediated efflux was effectively blocked by the Abcb1 antagonist PSC833 (25) (Figure 1c) or the Abcc1 antagonist MK571 (26) (Figure 1d), indicating that FTY720 acts on the transporters to enhance efflux activity. Efflux assays performed on T cells purified from _Abcc1_–/– mice showed no efflux of Fluo-3 in cells either treated or not treated with FTY720 (Figure 1f). The effects of FTY720 on Abcb1 efflux remained intact in _Abcc1_–/– cells (Figure 1e). Flow cytometry using an antibody to Abcc1 showed that FTY720 had no effect on the level of Abcc1 expression (supplementary Figure 1, http://www.jci.org/cgi/content/full/111/5/627/DC1). Conversely, _Abcb1_–/– cells demonstrated little efflux activity for either substrate, and neither substrate’s efflux activity was significantly enhanced by FTY720 (Figure 1, g and h). These results suggest that FTY720 sequentially activates Abcb1 upstream of Abcc1.
FTY720 enhances efflux activity of the Abcb1 and Abcc1 transporters. (a–d) Abcb1 efflux activity (a and c) and Abcc1 efflux activity (b and d) in wild-type C57BL/6 T cells. (e and f) Abcb1 (e) and Abcc1 (f) efflux activity in _Abcc1_–/– T cells. (g and h) Abcb1 (g) and Abcc1 (h) efflux activity in _Abcb1_–/– T cells.
Efflux activity correlates with in vitro migration triggered by FTY720. Next, the relation of multidrug transporter efflux activity to cell migration was examined. In vitro chemotaxis to CCL19 and CCL21 was assessed, since DC migration to these chemokines is sensitive to multidrug transporter efflux (13). Dose-response experiments demonstrated effective migration of purified T cells to 1 μg/ml CCL19 or CCL21, with 20–40% of the total number of T cells migrating in individual experiments. Only about 10% of cells migrated to 0.5 μg/ml CCL19 or CCL21, and little migration above background occurred at lower doses. The migration response was chemotactic, since addition of chemokine to the upper chamber alone, or to both the upper and the lower chambers, resulted in no migration (not shown). Pretreating purified T cells from wild-type mice with doses of FTY720 that enhance efflux led to a shift of the migration curve, with enhanced chemotaxis to the lower doses of CCL19 and CCL21 (Figure 2, a and b) and 1–5 μg/ml FTY720 eliciting optimal responses. The results in Figure 2a also show evidence of desensitization with high doses of FTY720 plus high doses of chemokine, resulting in diminished chemotaxis. Other experiments demonstrated that viability was unaffected by this dose of FTY720 (see Figure 7) and that, as expected, desensitization was dependent on time of incubation with FTY720 (not shown). Furthermore, T cell chemotaxis occurred only in response to the biologically active R-enantiomer, not the L-enantiomer (not shown), proving specificity of the migration effect.
FTY720 and S1P cause Abcc1- and Abcb1-dependent migration to CCL19 and CCL21. In vitro chemotactic response of T cells from wild-type mice to CCL19 or CCL21 is shown. FTY720, S1P, PSC833, and MK571 were added as indicated.
FTY720 causes multidrug transporter–independent apoptosis. Percent apoptosis after the indicated treatments with FTY720 plus multidrug transporter blockers is shown.
To determine whether chemokine- and FTY720-induced migration was dependent on multidrug transporter function, wild-type cells were pretreated with MK571 or PSC833, with or without FTY720, and chemotaxis was then measured. Cells pretreated with the multidrug transporter blockers failed to migrate in response to CCL19, with or without FTY720 (Figure 2c). In vitro migration assays with T cells purified from _Abcc1_–/– and _Abcb1a/1b_–/– showed some migratory response to CCL19 alone, but FTY720 did not enhance migration (supplementary Figure 2, http://www.jci.org/cgi/content/full/111/5/627/DC1). Thus, multidrug transporter activity plays a pivotal role in both chemokine-mediated and FTY720-mediated migration of T cells.
Since phosphorylated FTY720 and S1P initiate S1PR activation and T cell homing (6), in vitro migration assays with wild-type or knockout T cells, in response to S1P or phosphorylated FTY720, were performed. Dose-response studies showed that, like FTY720, S1P enhanced the migration response to chemokines and shifted the migration curve (Figure 2d). S1P also enhanced the migration to CCL19 in wild-type and _Abcb1_–/– cells and cells treated with PSC833 (an Abcb1 blocker), but not _Abcc1_–/– cells or cells treated with MK571 (an Abcc1 blocker) (Figure 2e and supplementary Figure 2, http://www.jci.org/cgi/content/full/111/5/627/DC1). Similar experiments with phosphorylated FTY720 likewise demonstrated its dependence on Abcc1 but not Abcb1 to enhance migration (Table 1). Since Abcb1 transports ceramides and sphingosines (8), these results suggest a molecular sequence in which sphingosines are directly or indirectly dependent on Abcb1 for transport to the extracellular space so that phosphorylated sphingosines can then act in an autocrine fashion to activate S1PRs. The results suggest that Abcb1 acts upstream of Abcc1. These results are consistent with the transporter studies shown in Figure 1, in which FTY720 sequentially activates Abcb1 proximal to Abcc1.
Migration to phosphorylated FTY720 depends on Abcc1 but not Abcb1
Multidrug transporters are required for FTY720-driven in vivo migration. The role of the multidrug transporters in in vivo migration to FTY720 was examined in _Abcc1_–/–, _Abcb1a/1b_–/–, and wild-type control mice. Animals were treated orally with either 60 μg/d FTY720 or vehicle alone. After 6 days, LNs, peripheral blood, and spleens were harvested, and total lymphoid cells were counted, stained for CD4 and CD8, and analyzed by flow cytometry. As expected, wild-type mice treated with FTY720 showed an increase in total cell number and total T cell number in the LNs, with concurrent decreases in the spleen and peripheral blood. On the other hand, both _Abcc1_–/– and _Abcb1_–/– mice treated with FTY720 showed a decrease in peripheral blood T cell counts but did not demonstrate an increase in LN T cell counts (T cell results, Figure 3; total cell counts, supplementary Figure 3, http://www.jci.org/cgi/content/full/111/5/627/DC1). Interestingly, the depletion of spleen cells that is normally observed in FTY720-treated wild-type mice was not observed in _Abcb1_–/– mice and was less substantial in _Abcc1a/1b_–/– mice. These data suggest that FTY720 may redirect T cell trafficking to the spleen in the absence of a capacity to promote LN homing of T cells. Also, while T cell counts changed in peripheral blood or spleen, the total number of cells in the three lymphoid compartments remained insensitive to FTY720. Overall, the results demonstrate that multidrug transporter activity is critical for FTY720-induced LN homing or sequestration of T cells in vivo.
FTY720 causes Abcc1- and Abcb1-dependent LN homing in vivo. In vivo migration from peripheral blood and spleen to LNs in response to FTY720 in fvb wild type (a), Abcc1–/– (b), and Abcb1–/– (c). Total T cells are indicated.
FTY720-induced in vitro migration is dependent on 5-LO. Abcc1-mediated migration of DCs is dependent on the efflux of cysLTs, specifically LTC4, through the multidrug transporters (13). Since 5-LO is normally expressed in both resting and activated T cells (27) and mice that lack 5-LO are incapable of producing cysLTs, we examined whether the cysLTs played a similar role in FTY720- and S1P-mediated T cell migration. The results demonstrate that purified T cells from _5-LO_–/– mice, or wild-type T cells treated with the 5-LO blocker AA-861, did not migrate to CCL19 or CCL21 plus FTY720 or S1P in vitro. In contrast, T cells pretreated with the 12-LO blocker CDC demonstrated normal FTY720-mediated migration (Figure 4, a–d; and supplementary Figure 4, http://www.jci.org/cgi/content/full/111/5/627/DC1). The addition of the exogenous cysLT LTD4, following pretreatment with AA-861, restored migration (Figure 4e). In contrast to DCs, T cells do migrate to the higher doses of CCL19 alone when 5-LO is absent (Figure 4b) or blocked (Figure 4a). Additional controls showed that neither 5-LO deficiency nor the 5-LO blockers affect the activity of the multidrug transporters in response to FTY720 (not shown). In conjunction with migration results showing that neither FTY720 nor S1P enhances migration in cysLT transporter–deficient _Abcc1_–/– T cells (Figures 2 and 3), these findings support a molecular sequence in which S1PR activation directly or indirectly enhances 5-LO activity, LTC4 synthesis, and cysLT transport by Abcc1 from the cytosol to the extracellular space with autocrine activation of the cysLT receptor (cysLTR). In vivo migration studies using _5-LO_–/– mice were performed as well. The results showed that FTY720 was not able to cause LN homing in these mice (Figure 4j). In sum, 5-LO is required for in vitro and in vivo migration to FTY720.
FTY720-mediated migration and S1P-mediated migration are dependent on cysLTs. (a and b) In vitro chemotactic response of T cells from wild-type or 5-LO–/– mice to CCL19 and FTY720 plus the 5-LO blocker AA-861 or the 12-LO blocker CDC. (c and d) Chemotactic response of T cells from 5-LO–/– and wild-type mice to CCL19 and S1P plus AA-861 or CDC. (e) Chemotactic response of T cells treated with CCL19 and with FTY720, FTY720 plus AA-861, or FTY720 plus AA-861 plus LTD4. (f) In vitro chemotactic response of T cells treated with FTY720 or FTY720 supernatant. (g) Chemotactic response of T cells treated with supernatants from FTY720, FTY720 plus AA-861, or FTY720 plus CDC-treated T cells. (h and i) Chemotactic response of fvb wild-type or fvb Abcc1–/– T cells treated with CCL19 and FTY720 or supernatant from FTY720-treated T cells. (j) In vivo migration from peripheral blood and spleen to LNs in response to FTY720 in 5-LO–/– mice. Total T cells are indicated.
5-LO activity causes cysLT secretion. Since this model predicts that cysLT secretion is necessary for migration, supernatants from FTY720-treated cells should contain cysLTs and enhance in vitro migration. To test this, purified T cells from wild-type mice were treated with FTY720, and the supernatant was collected and diluted 1:5 to make the final concentration of FTY720 in the supernatant too low to enhance migration. The supernatant was then added, along with CCL19, to purified T cells from wild-type mice. The results show that FTY720 supernatant but not the direct addition of FTY720 caused enhanced migration to the suboptimal dose of chemokine (Figure 4f). The activity of the supernatant was abrogated by addition of the 5-LO blocker AA-861 to the cells, but not by addition of the 12-LO blocker CDC (Figure 4g). The undiluted supernatant also enhanced migration in both fvb and _Abcc1_–/– cells, while FTY720 was effective only in both fvb wild-type and _fvb Abcc1_–/– cells (Figure 4, h and i; and Figure 2), giving further evidence against direct effects of FTY720. Since LTC4 is the only 5-LO metabolite transported by Abcc1, these data are most consistent with cysLT secretion via Abcc1 as the mechanism for FTY720-mediated increases in T cell migration.
T cell migration in response to FTY720 is dependent on CCL19 and CCL21. As shown previously and above, CCL19- and CCL21-driven migration of DCs and T cells is dependent on the multidrug transporters (13). The results above also demonstrate that FTY720 enhances T cell migration to suboptimal doses of CCL19 and CCL21 in a manner also dependent on the multidrug transporters. To determine whether other chemokine responses are sensitive to this pathway, we examined in vitro migration to the chemokines CCL5 (RANTES) and CXCL13 (BCA-1), both of which are involved in T cell migration. The results show that both CCL5 and CXCL13 caused T cell migration at a dose of 1 μg/ml, but not at a dose of 0.5 μg/ml. Pretreating the cells with 1 or 5 μg/ml FTY720 for 1 hour did not enhance migration to either chemokine, indicating that these chemokines and their receptors CCR1, CCR3, CCR5, and CXCR5 are unlikely to be involved in LN homing mediated by FTY720, multidrug transporters, or cysLTs (supplementary Figure 5, http://www.jci.org/cgi/content/full/111/5/627/DC1).
FTY720 enhances migration to CCL19 and CCL21 but not to other chemokines. FTY720 causes T cell LN homing in vivo in C57BL/6 wild-type (a) and CX3CR1–/– (c) but not plt (b) mice.
To determine whether CCL19 and CCL21 are required for in vivo LN sequestration, FTY720 was given orally at 60 μg/d for 5 days to plt mice, natural mutants that lack the genes for CCL19 and stromal cell CCL21 but not for lymphatic CCL21. It was expected that FTY720 would not cause LN sequestration in these animals. Indeed, the numbers of LNs and of T cells in plt mice receiving FTY720 were similar to those in plt mice receiving vehicle alone (Figure 5a and supplementary Figure 5, http://www.jci.org/cgi/content/full/111/5/627/DC1). FTY720 did, however, cause a decrease in the numbers of splenic and peripheral blood T cells and total cells, similar to the effects of FTY720 in _Abcc1_–/– (Figure 3). It is possible that lymphatic CCL21 contributes to peripheral blood or splenic egress; it is also possible that other uncharacterized deletions in the plt mutant contribute to the observed defect. FTY720 was also given to mice that lacked the chemokine receptor CX3CR1, which binds fractalkine. FTY720 caused an increase in the number of LN T cells in these mice compared with mice that received vehicle alone, and there was a decrease in the number of peripheral blood and splenic lymphocytes (Figures 5c). Thus, FTY720 requires CCL19 and/or stromal cell CCL21, but not a variety of other chemokines and their receptors, in order to promote LN sequestration or homing in vivo.
T cell migration in response to FTY720 requires specific adhesion receptors. While FTY720 acts on the multidrug transporters to cause cysLT efflux and subsequent migration to CCL19 and CCL21, the cells still require adhesion receptor molecules in order to attach to specific anatomic sites such as the LNs. We sought to determine which adhesion receptors are involved in FTY720-mediated LN sequestration, using in vivo migration assays. Wild-type mice were treated orally with 60 μg/d FTY720, and intravenously on days 0 and 1 with 100 μg/d of anti-CD62L (L-selectin), anti-CD44, or anti–VLA-4α mAb. Control groups either received the mAb’s alone or were untreated. Spleen, peripheral blood, and LNs were harvested and analyzed. The mice that received anti-CD62L mAb plus FTY720 showed a decrease in spleen and peripheral blood T cells, with a concurrent increase in LN T cells, indicating that FTY720-mediated LN sequestration is not dependent on the CD62L adhesion receptor (Figure 6 and supplementary Figure 6, http://www.jci.org/cgi/content/full/111/5/627/DC1). This finding is particularly notable since this mAb is able to drive naive T cells out of the LNs and into the spleen and peripheral blood (28, 29). However, mice that received anti-CD44 or anti–VLA-4α mAb plus FTY720 showed a decrease in spleen and peripheral blood T cells, without an increase in LN cell number, indicating the importance of VLA-4α and CD44 in FTY720-induced LN sequestration of T cells. We further examined the importance of adhesion receptors using knockout mice. Mice lacking L-selectin (_CD62L_–/–) or an L-selectin ligand (_FucTVII_–/–) and treated with FTY720 likewise showed an increase in LN cell numbers, with a decrease in cell numbers in the spleen and peripheral blood. These findings are similar to those of Chiba et al. (3), who showed in a rat model that the actions of FTY720 were independent of CD62L but not VLA-4α. Flow cytometric analysis was performed to determine whether FTY720 affected the level of expression of the adhesion receptors.
FTY720 causes CD44- and VLA-4α–dependent T cell LN homing in vivo. C57BL/6 wild-type (a), L-selectin–/– (b), and FucTVII–/– (c) mice and mice treated with anti-CD62L (d), anti-CD44 (e), or anti–VLA-4α (f) mAb received either vehicle or FTY720.
Purified T cells were treated with FTY720 or medium and stained for CD44, VLA-4α, or CD62L. The results showed similar levels of expression of VLA-4α, CD44, and CD62L in cells treated or not treated with FTY720 (not shown). Since naive cells are generally CCR7+, while many memory and/or activated T cells are CCR7–, CD44+, and VLA-4α+, the mAb’s may be interfering with LN adhesion after entry. Further, CCL21 can stimulate integrin α4–dependent adhesion (30). Overall, the results from these in vivo migration assays demonstrate that FTY720-mediated LN homing is L-selectin– and L-selectin ligand–independent, and VLA-4α– and CD44-dependent.
The effect of FTY720 on efflux is not related to apoptosis. Previous studies described an apoptotic effect of FTY720 on T cells in vitro. In order to test whether apoptosis was occurring at the doses of FTY720 used in our in vitro assays, cells were treated for 4 hours with 1, 5, 10, and 50 μg/ml FTY720, 4 μM camptothecin as a positive control, or complete medium alone. Apoptosis was measured using annexin V and 7-AAD staining. Cells that received 1 or 5 μg/ml FTY720 showed minimal apoptosis, while cells that received 10 μg/ml showed increased apoptosis (Figure 7). Dramatic apoptosis was seen with a dose of 50 μg/ml FTY720. Importantly, adding the multidrug transporter inhibitors MK571 and/or PSC833 had little effect on the levels of apoptosis seen at the various doses of FTY720. Additional experiments demonstrated that the 5-LO blocker and the CCL19 ligand did not inhibit the apoptotic response (not shown). These results demonstrate that FTY720 causes multidrug transporter–independent apoptosis, but not at the doses used in our study.