Sphingosine 1-Phosphate Receptor 1 (S1P₁) Upregulation and Amelioration of Experimental Autoimmune Encephalomyelitis by an S1P₁ Antagonist

Abstract

Sphingosine 1-phosphate receptor 1 (S1P₁) is a G protein–coupled receptor that is critical for proper lymphocyte development and recirculation. Agonists to S1P₁ are currently in use clinically for the treatment of multiple sclerosis, and these drugs may act on both S1P₁ expressed on lymphocytes and S1P₁ expressed within the central nervous system. Agonists to S1P₁ and deficiency in S1P₁ both cause lymphocyte sequestration in the lymph nodes. In the present study, we show that S1P₁ antagonism induces lymphocyte sequestration in the lymph nodes similar to that observed with S1P₁ agonists while upregulating S1P₁ on lymphocytes and endothelial cells. Additionally, we show that S1P₁ antagonism reverses experimental autoimmune encephalomyelitis in mice without acting on S1P₁ expressed within the central nervous system, demonstrating that lymphocyte sequestration via S1P₁ antagonism is sufficient to alleviate autoimmune pathology.

Introduction

Sphingosine 1-phosphate receptor 1 (S1P₁) plays an important role in many physiologic systems, including vascular development, lymphocyte development, and lymphocyte recirculation (Liu et al., 2000; Allende et al., 2003, 2004; Matloubian et al., 2004; Cyster and Schwab, 2012). S1P₁ is required on developing lymphocytes to mature beyond a semimature CD69^hi, CD62L^lo state, rendering the blood and lymph of mice lacking S1P₁ on developing lymphocytes largely devoid of T cells. When S1P₁^−/− thymocytes are transferred into recipient mice, they are also retained from blood and lymphatic circulation. S1P₁ became a relevant drug target in the treatment of autoimmune disease following the discovery that 2-​amino-​2-​[2-​(4-​octylphenyl)ethyl]-​1,​3-​propanediol (FTY720; fingolimod, Gilenya), which was known to inhibit lymphocyte recirculation, is a sphingosine 1-phosphate (S1P) receptor prodrug that is phosphorylated in vivo to yield 2-amino-2[2-(4-octylphenyl)ethyl]-1,3-propanediol, mono dihydrogen phosphate ester (FTY720-P), a potent agonist of S1P₁, S1P₃, S1P₄, and S1P₅ (Mandala et al., 2002). S1P₁ selective agonists demonstrated that FTY720 acted via S1P₁ to induce lymphocyte sequestration (Sanna et al., 2004). The ability of FTY720-P and other S1P₁ agonists to induce sustained internalization and/or degradation of S1P₁ (Graler and Goetzl, 2004; Gonzalez-Cabrera et al., 2007, 2008), combined with the deficient egress of S1P₁-deficient lymphocytes, has led to the hypothesis that S1P₁ agonists act as functional antagonists (Graler and Goetzl, 2004). Several S1P₁-selective antagonists have also been generated, which inhibit agonist-dependent effects in vitro; stabilize the S1P₁ receptor, allowing for its structural determination; and induce pulmonary edema in vivo. In addition, initial antagonists could reverse agonist-induced lymphocyte sequestration while being unable to induce lymphocyte sequestration themselves (Foss et al., 2005; Wei et al., 2005; Sanna et al., 2006; Hanson et al., 2012). Recent work has shown that S1P₁ antagonists can indeed induce lymphocyte sequestration at high plasma concentrations (Tarrason et al., 2011) and S1P₁ antagonists can alleviate animal models of autoimmune arthritis (Fujii et al., 2012), cardiac allograft rejection (Angst et al., 2012), and multiple sclerosis (Quancard et al., 2012).

S1P receptor agonists have come of age with the Food and Drug Administration’s approval of FTY720 for the treatment of relapsing-remitting multiple sclerosis. The efficacy of FTY720 is not solely dependent on its ability to cause full lymphocyte sequestration via S1P₁, as it is effective at doses that maintain ∼50% lymphopenia. This efficacy probably involves both S1P₁ and other S1P receptors within the central nervous system (CNS) (Cohen and Chun, 2011; Hla and Brinkmann, 2011). S1P₁ agonists that can efficiently penetrate the CNS can induce receptor signaling and degradation of S1P₁ expressed on neurons and astrocytes (Gonzalez-Cabrera et al., 2012), and require lymphocyte sequestration for only one-third of a dosing interval to reverse experimental autoimmune encephalomyelitis (EAE) in mice. Additionally, mice lacking S1P₁ on astrocytes are refractory to developing EAE, and are suggested to be important targets of FTY720 (Choi et al., 2011). Several other S1P receptors are expressed within the CNS, and the activation and/or degradation of these receptors by FTY720 may also play important roles in reversing the immunopathology of multiple sclerosis (Miron et al., 2008, 2010).

In the present study, we demonstrate that S1P₁ antagonism sequesters lymphocytes in the peripheral lymph nodes but not the spleen, similar to that observed with S1P₁ agonists. S1P₁ antagonism also causes significant upregulation of S1P₁ expression on peripheral lymphocytes, mature thymocytes, and lung endothelial cells. Additionally, S1P₁ antagonism can alleviate EAE in mice despite the inability of the antagonist to penetrate the CNS. Thus, lymphocyte sequestration induced by S1P₁ antagonists is sufficient to ameliorate the autoimmune pathology observed in EAE, and does not require antagonism of S1P₁ expressed on neurons or astrocytes within the CNS.

Materials and Methods

Compounds and In Vitro Assays.

Example 26 [Ex26, 1-(5′-((1-(4-chloro-3-methylphenyl)ethyl)amino)-2'-fluoro-3,5-dimethyl-[1,1'-biphenyl]-4-ylcarboxamido)cyclopropanecarboxylic acid] was synthesized as a racemic mixture according to its published synthesis in the patent literature (Angst et al., 2009). RP-001 (3-(4-(5-(3-cyano-4-isopropoxyphenyl)-1,2,4-oxadiazol-3-yl)-2,3-dihydro-1H-inden-1-ylamino)propanoic acid) was synthesized as previously described (Cahalan et al., 2011). FTY720 was purchased from Cayman Chemicals (Ann Arbor, MI). Ex26 and RP-001 were solubilized in 50 mM Na₂CO₃, while FTY720 was solubilized in H₂O. In vitro assays for S1P receptor function were performed using the following cell lines: S1P₁, S1P₄, and S1P₅: Tango Human Osteosarcoma U2OS cells (Invitrogen, Carlsbad, CA) expressing the indicated receptor; S1P₂: Chinese hamster ovary cells expressing S1P₂ coupled to a cAMP response element coupled to beta-lactamase reporter; and S1P₃: Chinese hamster ovary cells expressing S1P₃ coupled to an Nuclear factor of activated T-cells promoter coupled to beta-lactamase reporter through G protein α 16. S1P₁ internalization and polyubiquitinylation were evaluated using human embryonic kidney cells expressing S1P₁ enhanced green fluorescent protein (eGFP) as previously described (Gonzalez-Cabrera et al., 2007), pretreating cells for 1 hour with Ex26.

Evaluation of Lymphocyte Sequestration, Pulmonary Edema.

Eight-week-old male C57Bl/6J mice were purchased from the The Scripps Research Institute mouse breeding facility (La Jolla, CA) for evaluation of lymphocyte sequestration and pulmonary edema. Mice were injected i.p. with Ex26 or 50 mM Na₂CO₃ vehicle, and blood was removed from the heart following euthanasia. Blood was lysed in 150 mM NH₄Cl, 10 mM KHCO₃, and 0.1 mM EDTA; washed with phosphate-buffered saline (PBS) containing 2% fetal bovine serum, 1 mM EDTA, and 0.1% NaN₃; counted using a ViCell-XR counter (Beckman Coulter, Brea, CA); stained with antibodies; and analyzed by flow cytometry. To evaluate pulmonary edema, mice were perfused with 15 ml of PBS through the right ventricle, then the lungs were removed, blotted dry to remove excess PBS, and weighed. All mouse experiments were performed using protocols approved by the Institutional Animal Care and Usage Committee.

Compound Concentrations in Plasma and Tissues.

Ex26 plasma concentrations were determined using methanol extraction as previously described (Cahalan et al., 2011), detecting an m/z value of 495.2 for Ex26 using an Agilent 6410 triple quadrupole mass spectrometer coupled to an Agilent 1100LC system (Agilent Technologies, Santa Clara, CA). Ex26 concentrations in the brain were determined by disruption of brain tissue in water by probe sonication, followed by extraction with acetonitrile and filtration through MultiScreen hydrophilic polytetrafluoroethylene 0.45 μm filters (EMD Millipore, Billerica, MA). Filtrates were analyzed using a API 4000 liquid chromatography-tandem mass spectrometer (AbSciex, Framingham, MA) and quantified using a positive-ion multiple reaction monitoring method (495.1/242.1, m/z).

Continuous Administration of S1P₁ Antagonist.

Six-week-old S1P₁-eGFP mice were anesthetized with isoflurane, and their backs were shaved, cleaned with 70% ethanol to remove any excess hair, then wiped with povidone iodine. An incision was made on the lower back of the mice, and micro-osmotic pumps (Alzet model 1003D; Alzet, Cupertino, CA) containing either 50 mM Na₂CO₃ vehicle or 2 mg/ml Ex26 were implanted, yielding a dose of ∼0.1 mg/kg per hour. Mice were given an i.p. dose of 3 mg/kg Ex26 or vehicle immediately following surgery.

Flow Cytometry, S1P₁ Expression, and Statistical Analysis.

Fluorescently labeled antibodies specific to CD4 and CD8 were obtained from Biolegend (San Diego, CA). Fluorescently labeled antibodies specific to CD19, CD31, CD45.2, CD62L, and CD69 were obtained from Beckton-Dickinson (San Diego, CA). Data were collected using an LSRII flow cytometer (Beckton-Dickinson) and analyzed using FlowJo (Treestar, Ashland, OR). S1P₁ expression by flow cytometry was measured using S1P₁-eGFP knockin mice (Cahalan et al., 2011). S1P₁ expression in the CNS in EAE experiments was evaluated using a C-terminal–specific S1P₁ antibody (H-60, Santa Cruz Biotechnology, Santa Cruz, CA; used at 1:500 dilution). All statistical analyses were performed using GraphPad Prism Software (GraphPad, La Jolla, CA).

EAE Induction and Scoring.

EAE was induced in female 10-week-old C57Bl/6J mice purchased from Jackson Laboratories (Bar Harbor, ME). EAE was induced using a Hooke Laboratories EAE induction kit (Lawrence, MA; EK-0114 for EAE, CK-0114 for control) according to the manufacturer’s instructions. Mice were scored by the following criteria: 0.5 (weak tail), 1 (limp tail), 1.5 (weak tail + weak hind limbs), 2 (limp tail + weak hind limbs), 2.5 (limp tail + unilateral hind limb paralysis), 3 (limp tail + bilateral hind limb paralysis), 4 (limp tail + bilateral hind limb paralysis + partial front limb paralysis), and 5 (moribund or dead). Mice scoring 4 for two consecutive days were euthanized and recorded as 5 for the remaining days of the experiment. Mice were injected i.p. daily with 50 mM Na₂CO₃ vehicle, 30 mg/kg Ex26, or 10 mg/kg FTY720 in a volume of 10 µl per gram weight of mouse beginning the first day on which clinical signs were observed in that mouse.

Results and Discussion

Ex26 is an S1P₁ Antagonist that Inhibits Lymphocyte Egress.

Most existing S1P₁ antagonists are S1P analogs with IC₅₀ values in the double-digit nanomolar range that possess relatively short half-lives. Recently, new S1P₁ antagonists have been described, including a series of biaryl benzylamines by Novartis (Angst et al. 2009). We synthesized and characterized one of these compounds, Ex26, and confirmed it to be a potent and selective antagonist of S1P₁ (Fig. 1A; Table 1), similar to a recently published antagonist (Quancard et al., 2012). Ex26 could inhibit RP-001–induced S1P₁ internalization and polyubiquitinylation in vitro (Supplemental Fig. 1). Similar to other previously described S1P₁ antagonists, Ex26 induced dose-dependent and time-dependent pulmonary edema in vivo (Fig. 1, B and C), and had a relatively short in vivo half-life of approximately 73.5 minutes (Supplemental Fig. 1C).

Fig. 1.

Ex26 is a potent, selective S1P₁ antagonist. (A) Dose response in vitro of Ex26 on S1P₁-expressing cells in the presence of 5 nM S1P. The structure of Ex26 is depicted on the right. (B) Ex26 induces dose-dependent pulmonary edema 2 hours following i.p. treatment. (C) Pulmonary edema induced by 3 mg/kg Ex26 i.p. resolves by 16–24 hours following treatment. All data are representative of at least two experiments, with (B) and (C) having four mice per group per experiment. Graphs are plotted as the mean ± S.E.M.

TABLE 1.

Selectivity of Ex26 on S1P receptors

Ex26 displays excellent selectivity for S1P₁ over other S1P receptors. It also does not exhibit any detectable agonist activity on any S1P receptor.

Receptor	Antagonist IC50	Agonist EC50
	nM	nM
S1P1	0.93	>10,000
S1P2	>10,000	>10,000
S1P3	>10,000	>10,000
S1P4	4900	>10,000
S1P5	3100	>10,000

Ex26, 1-(5′-((1-(4-chloro-3-methylphenyl)ethyl)amino)-2'-fluoro-3,5-dimethyl-[1,1'-biphenyl]-4-ylcarboxamido)cyclopropanecarboxylic acid; S1P, sphingosine 1-phosphate.

Earlier work showed that the S1P-like S1P₁ antagonists W146 and VPC44116 reversed agonist-induced lymphocyte sequestration while not causing lymphocyte sequestration (Sanna et al., 2006; Foss et al., 2007). Recent work has found that W146 induces transient lymphocyte sequestration at high doses (Tarrason et al., 2011), which we replicated (unpublished data). Ex26 induced lymphocyte sequestration at low doses, possessing an ED₅₀ of ∼0.06 mg/kg when examined 2 hours following treatment (Fig. 2A). Lymphocyte sequestration by Ex26 resolved with similar kinetics as did Ex26-evoked pulmonary edema (Fig. 2B). To examine the effects of extended antagonist treatment, we implanted mice with micro-osmotic pumps to continuously deliver Ex26 at a dose of 0.1 mg/kg per hour for 3 days following a loading dose of 3 mg/kg. Extended treatment with Ex26 led to significant retention of T and B cells within the lymph nodes and significant decreases in T and B cells within the spleen, similar to S1P₁ agonists (Fig. 2, C and D). Continuous administration of Ex26 also led to thymic retention of mature CD62L^Hi single-positive thymocytes, also similar to the effects induced by S1P₁ agonists (Fig. 2E). These data demonstrate that disruption of S1P₁ signaling by S1P₁ antagonism leads to the inhibition of lymphocyte and thymocyte egress.

Fig. 2.

S1P₁ antagonism by Ex26 induces lymphocyte sequestration in the lymph nodes and thymus. (A) Ex26 induces dose-dependent lymphocyte sequestration 2 hours following i.p. treatment. (B) Lymphopenia induced by 3 mg/kg Ex26 i.p. resolves by 24 hours following treatment. (C and D) Continuous administration of Ex26 in 6-week-old mice by micro-osmotic pumps sequesters T and B cells in the peripheral lymph nodes (pLN) (C), leaving the spleen depleted of lymphocytes (D). pLN cell numbers derive from combined inguinal, axillary, and brachial lymph nodes. (E) Ex26 leads to accumulation of mature CD62L^Hi, but not immature CD62L^Lo, SP thymocytes. All graphs are representative of three experiments, with 3–4 mice per group per experiment. Graphs are plotted as the mean ± S.E.M. *P < 0.05, **P < 0.01 as determined by an unpaired, two-tailed t test.

S1P₁ Antagonism Upregulates S1P₁ Expression.

Since S1P₁ agonists downregulate S1P₁, we wanted to determine whether S1P₁ antagonism could conversely upregulate S1P₁. Continuous S1P₁ antagonism in mice expressing S1P₁-eGFP from the S1P₁ locus (Cahalan et al., 2011) for 3 days by micro-osmotic pumps caused significant upregulation of S1P₁-eGFP on lymphocytes within the lymph node (Fig. 3A). This suggests that the low concentration of S1P within the lymph node (Schwab et al., 2005) under normal physiologic conditions is sufficient to suppress the expression of S1P₁. We observed similar upregulation within the spleen (unpublished data) and a modest upregulation of S1P₁-eGFP on fully mature CD62L^hi SP thymocytes (Fig. 3B). S1P₁ agonists cause a loss of surface expression of CD69 on mature thymocytes (Alfonso et al., 2006). In contrast to the effects seen with agonists, continuous Ex26 treatment led to significant upregulation of CD69 (Fig. 3B), indicating that S1P₁ signaling, not only expression of S1P₁ (Bankovich et al., 2010), is critical for suppressing the surface expression of CD69; thus, downregulation of CD69 by S1P₁ agonists is a measure of agonism, not functional antagonism. Upregulation of S1P₁-eGFP was not limited to lymphocytes, as blood endothelial cells within the lung significantly upregulated S1P₁-eGFP expression (Fig. 3C). Unlike many S1P₁ agonists, including FTY720-P, Ex26 did not cause any changes in the expression of S1P₁-eGFP within the brain (Fig. 3D), due to the fact that Ex26 was almost undetectable within the CNS (plasma: 6.8 ± 0.3 µM, brain: 0.01 ± 0.005 µM; 2/3 animals below the level of detection; mean ± S.E.M.).

Fig. 3.

S1P₁ antagonism by Ex26 upregulates S1P₁ and CD69, but not in the central nervous system. (A) S1P₁-eGFP expression on lymphocytes from lymph nodes from S1P₁-eGFP knockin mice continuously administered Ex26 by micro-osmotic pump for 3 days. Gray shaded histograms represent background fluorescence in wild-type mice. The graph on the right represents the mean fluorescence intensity of S1P₁-eGFP on the indicated cell type. (B) Mean fluorescence intensity of S1P₁-eGFP (left) and CD69 (right) on CD4 SP thymocytes following continuous treatment with Ex26. (C) Mean fluorescence intensity of S1P₁-eGFP on lung endothelial cells following continuous treatment with Ex26. (D) Fluorescent scan of SDS-PAGE gel from the brains of mice following 3 days of treatment with Ex26. The graph on the right is obtained by densitometric analysis of the gel on the left. All histograms and graphs are representative of three experiments, with 3–4 mice per group per experiment. Graphs are plotted as the mean ± S.E.M. **P < 0.01, ***P < 0.001 as determined by an unpaired, two-tailed t test. a.u., arbitrary units; MFI, mean fluorescence intensity; n.s., not significant.

S1P₁ Antagonism Ameliorates EAE.

Because Ex26 did not enter the CNS or cause any change in S1P₁ expression within the CNS, it allowed us to determine whether lymphocyte sequestration alone was able to reverse EAE. Whereas 3 mg/kg Ex26 induced relatively short-duration lymphocyte sequestration, we found that a single dose of 30 mg/kg caused lymphocyte sequestration and pulmonary edema that lasted 24 hours in naïve mice (Supplemental Fig. 2, A and B). To examine whether S1P₁ antagonism could ameliorate EAE similar to S1P₁ agonism, we induced disease using the myelin oligodendrocyte glycoprotein residues 33–55 peptide model, and, upon development of clinical signs of disease, treated mice i.p. once daily with 30 mg/kg Ex26, 10 mg/kg FTY720, or 50 mM Na₂CO₃ vehicle, which we found to be indistinguishable from water, the usual vehicle for FTY720 (unpublished data). We found that treatment of mice with 30 mg/kg Ex26 daily significantly reduced the severity of EAE as assessed by examining clinical signs (Fig. 4A). We observed significant lymphocyte sequestration 3 hours following the last treatment with both 30 mg/kg Ex26 and 10 mg/kg FTY720; however, unlike its effect in naïve mice, 30 mg/kg Ex26 did not cause lymphocyte sequestration that lasted a full 24 hours in mice with EAE, whereas 10 mg/kg FTY720 did (unpublished data), suggesting that treatment with pertussis toxin used in the induction of EAE, or repeated dosing of Ex26, reduced the efficacy of Ex26, potentially by upregulating S1P₁ expression on lymphocytes. The reduction in the severity of EAE was seen in the spinal cord, as 30 mg/kg Ex26 inhibited both lymphocyte infiltration and destruction of the white matter in the spinal cord of mice euthanized at the end of the experiment (Fig. 4B). Consistent with the lack of CNS penetration of Ex26, we did not observe any changes in S1P₁ expression within the brains of mice euthanized at the end of the experiment that were treated daily with 30 mg/kg Ex26 compared with those treated daily with vehicle, whereas mice treated daily with 10 mg/kg FTY720 exhibited a complete loss in S1P₁ within the brain (Fig. 4C). This indicates that antagonism of S1P₁ expressed on neurons or astrocytes within the CNS is not required for the amelioration of EAE by S1P₁ antagonists, implying that lymphocyte sequestration by S1P₁ antagonists is sufficient to reverse the pathology of EAE, in keeping with the efficacy of lymphocyte migration inhibitory agents, such as natalizumab, that successfully treat multiple sclerosis.

Fig. 4.

S1P₁ antagonism by Ex26 alleviates EAE. (A) Average EAE scores from myelin oligodendrocyte glycoprotein residues 33-55–induced mice injected daily i.p. with vehicle, 30 mg/kg Ex26, or 10 mg/kg FTY720 following the onset of symptoms. ****P < 0.0001 compared with vehicle as calculated by one-way repeated measures analysis of variance with Bonferroni’s multiple comparison post test. The graph represents two separate experiments as the mean ± S.E.M with 9–10 mice per group. (B) Representative spinal cord sections stained with H&E from control mice without EAE (top left) or mice with EAE that had been treated daily as indicated following the onset of clinical signs. (C) Western blot for S1P₁ on the brains of mice with EAE treated daily with vehicle (50 mM Na₂CO₃), 30 mg/kg Ex26, or 10 mg/kg FTY720 following the onset of symptoms. The graph represents S1P₁ expression as determined by densitometry. a.u., arbitrary units; n.d., not detectable.

Abbreviations

CNS

central nervous system

EAE

experimental autoimmune encephalomyelitis

Ex26

1-(5′-((1-(4-chloro-3-methylphenyl)ethyl)amino)-2'-fluoro-3,5-dimethyl-[1,1'-biphenyl]-4-ylcarboxamido)cyclopropanecarboxylic acid

FTY720

2-​amino-​2-​[2-​(4-​octylphenyl)ethyl]-​1,​3-​propanediol

PBS

phosphate-buffered saline

S1P

sphingosine 1-phosphate

S1P_1–5

sphingosine 1-phosphate receptors 1–5

S1P₁-eGRP

S1P₁ enhanced green fluorescent protein

RP-001

3-(4-(5-(3-cyano-4-isopropoxyphenyl)-1,2,4-oxadiazol-3-yl)-2,3-dihydro-1H-inden-1-ylamino)propanoic acid

Authorship Contributions

Participated in research design: Cahalan, Gonzalez-Cabrera, Rosen.

Conducted experiments: Cahalan, Gonzalez-Cabrera, Nguyen, Cisar, Leaf, Brown.

Contributed new reagents or analytic tools: Guerrero, Roberts.

Performed data analysis: Cahalan, Gonzalez-Cabrera, Cisar, Brown, Rosen.

Wrote or contributed to the writing of the manuscript: Cahalan, Gonzalez-Cabrera, Rosen.