FTY720/Fingolimod, a Sphingosine Analogue, Reduces Amyloid-β Production in Neurons

Sphingosine-1-phosphate (S1P) is a pluripotent lipophilic mediator working as a ligand for G-protein coupled S1P receptors (S1PR), which is currently highlighted as a therapeutic target for autoimmune diseases including relapsing forms of multiple sclerosis. Sphingosine related compounds, FTY720 and KRP203 known as S1PR modulators, are phosphorylated by sphingosine kinase 2 (SphK2) to yield the active metabolites FTY720-P and KRP203-P, which work as functional antagonists for S1PRs. Here we report that FTY720 and KRP203 decreased production of Amyloid-β peptide (Aβ), a pathogenic proteins causative for Alzheimer disease (AD), in cultured neuronal cells. Pharmacological analyses suggested that the mechanism of FTY720-mediated Aβ decrease in cells was independent of known downstream signaling pathways of S1PRs. Unexpectedly, 6-days treatment of APP transgenic mice with FTY720 resulted in a decrease in Aβ40, but an increase in Aβ42 levels in brains. These results suggest that S1PR modulators are novel type of regulators for Aβ metabolisms that are active in vitro and in vivo.


Introduction
Bioactive lipids, such as sphingolipids, have effect on various neuronal activities, including signal transduction, inflammatory response, and neuronal survivals [1]. It has been reported that the sphingolipid metabolism in brain was altered under neurodegenerative conditions, e.g., Alzheimer disease (AD) [2,3,4]. However, the relationship between changes in brain sphingolipids with the pathological mechanisms of AD has remained largely unclear. Interestingly, the production of amyloid-b (Ab) peptide, the major component of senile plaques deposited in the brains of patients with AD, is known to be modulated by sphingolipids [5]. Ab is produced from amyloid-b precursor protein (APP) through a sequential cleavage by two aspartate proteases, band c-secretases [6,7]. BACE1 (b-site APP cleaving enzyme 1) [8] is a type-1 transmembrane protein responsible for the b-secretase activity, and c-secretase is comprised of four integral membrane proteins, Presenilin (PS) as the catalytic subunit associated with Nicastrin (Nct), Aph1, and Pen2 [9]. Both enzymes are located in lipid rafts [10], a membrane microdomain enriched in sphingolipids and cholesterol, and the activities of the secretases are affected by the lipid composition [11,12,13].
Sphingosine-1-phosphate (S1P) is produced from sphingosine by sphingosine kinase (SphK). S1P works as a ligand for a subset of Gprotein coupled receptor (S1PR) proteins, and functions on various cellular events including neurogenesis, angiogenesis, and immune response [14]. S1PR modulator FTY720 (Fingolimod/Gilenya) is a sphingosine-related molecule exhibiting an immunomodulatory function, which has recently been approved as an oral treatment for relapsing forms of multiple sclerosis [14,15]. FTY720 is phosphorylated by SphK2 to function as an agonist for S1P receptors, i.e., S1PR1, S1PR3, S1PR4, and S1PR5 ( Fig. 1) [16,17,18]. Despite its agonistic action, FTY720 promotes endocytosis and degradation of the S1P receptors, thereby resulting in functional antagonistic effects. FTY720 interferes with the neuroinflammatory responses of auto-active T-cells and glial cells [19] and ameliorates the symptoms of autoimmune encephalomyelitis in rodents, the latter being a model for multiple sclerosis [20,21].
Recently, we reported that SphK2, one of the rate limiting enzymes for the production of S1P, is upregulated in AD brains, and that S1P interacts with BACE1 to regulate its proteolytic activity [22]. To examine the effects of S1P receptor modulators on AD, we investigated the effects of FTY720, and another S1PR modulator KRP203 harboring a more specific agonist activity against S1PR1 [23], on Ab production in cultured neuronal cells and brain Ab levels in AD model mice.

Antibodies and immunological methods
A polyclonal antibody against presnilin1 (PS1) CTF (G1L3) was raised as described [25]. Following antibodies were purchased from indicated venders: 82E1 (Immuno-Biological Laboratories) against the N terminus of human Ab for detection of Ab and bCTF, anti-APP (C) (Immuno-Biological Laboratories) for detection of APP CTFs and AICD, anti-mouse/rat APP (597) for detection of sAPPa, anti-sAPPbwt (Immuno-Biological Laboratories), anti-a-Tubulin DM1A (Sigma Aldrich), anti-SphK2 P-19 (Santa Cruz Biotechnology), anti-Myc 9B11 and anti-phospho specific ERK antibody (Cell Signaling Technology). Protein samples were analyzed by immunoblotting or two-site ELISAs for the detection of Ab as previously described [26,27]. Specificities of the APP antibodies used have previously been shown [22]. To analyze Ab species with different C-terminal lengths, samples were separated by modified Tris/Tricine/8M urea gels as reported previously, followed by immunoblotting with 82E1 [28,29].

FTY720 treatment in AD model mice
All experiments using animals in this study were performed according to the guidelines provided by the Institutional Animal Care Committee of Graduate School of Pharmaceutical Sciences, The University of Tokyo. All animals were maintained on food and water with a 12 h light/dark cycle. A7 transgenic mice overexpressing human APP695 harboring K670N, M671L, and T714I FAD mutations in neurons under the control of Thy1.2 promoter were used as a mouse model of AD [37]. Female A7 mice at 6 months old were used for treatment of FTY720, in which FTY720 dissolved in saline was injected subcutaneously once a day for 6 days (0.5 mg/kg/day). Brain samples were solubilized with 10 mM Tris buffer containing 1% CHAPS, and subjected to the sandwich ELISA for Ab (Wako Chemical) [22].

S1P receptor modulators, FTY720 and KRP203, decreased Ab production in neuronal cells
To test if the S1PR modulators alter Ab production in neuronal cells, we treated mouse primary cortical neurons with FTY720 and KRP203 (structures shown in Figure 1A) and found that FTY720 and KRP203 decreased secretion of both Ab40 and Ab42 in a dose dependent manner (Fig. 1B). To further investigate the effect of these compounds, we tested a mouse neuronal N2a cell line stably expressing GFP, truncated Notch fused with Gal4/VP16 (NDE/gvp) and a luciferase reporter under UAS promoter [32,33,36] (N2aNH cell line), which enables us to simultaneously analyze the effects of compounds on Ab production, Notch signaling (as luciferase activity) and cell viability (as GFP fluorescence). Treatment with FTY720 decreased the production of Ab40 and Ab42 in a dose dependent fashion below the toxic concentration ( Fig. 1C and S1) without affecting the Notch signaling ( Fig. 1D and E).
To further identify the molecular target of S1PR modulators, we analyzed their effect on N2a cells transiently expressing SC100, corresponding to bCTF of human APP that is a direct substrate of c-secretase. SC100 is endoproteolyzed by c-secretase at e-site to release the intracellular domain (AICD), and resultant intramembrane stub is trimmed by carboxypeptidase-like activity of the csecretase at multiple c-sites to generate Ab. FTY720 and KRP203 decreased secretion of Ab40 and Ab42 from SC100 ( Fig. 2A), suggesting that S1PR modulators affected the c-secretase activity. Recently, small compounds that specifically lower Ab42, and Ab40 to a lesser extent, have been termed as c-secretase modulators [6,29]. However, FTY720 treatment decreased all Ab species with different Ab C-termini, including Ab38, Ab39, Ab40 and Ab42 (Fig. 2B). Concomitantly, a slight accumulation of aCTF (from endogenous mouse APP) and SC100 was observed (Fig. 2C-F), whereas the levels of sAPPa and sAPPb products of aor b-secretase-mediated APP cleavage of endogenous APP, respectively, were not altered by FTY720 treatment (Fig. 2G). Intriguingly, AICD production was not altered by FTY720 treatment (Fig. 2F) similarly to the processing of NDE, a direct Notch substrate for c-secretase (see Fig. 1D and E). Because both AICD and NICD are produced from e-cleavage by c-secretase, these data suggest that S1PR modulators specifically regulate the c-cleavage irrespective of the substrate.

Mode of action of inhibition of Ab production mediated by FTY720
Phosphorylation of FTY720 by SphK2 yields the active metabolite, FTY720-phosphate (FTY720-P), which is known as a potent agonist of the S1P receptors (Fig. 1A). To determine whether FTY720-P is involved in the regulation of Ab production, we treated N2a cells with FTY720 after RNAi knock-down of the endogenous expression of SphK2. We observed a ,60% of decrease in SphK2 expression after siRNA treatment ( Fig 3A). As reported previously, knockdown of SphK2 decreased Ab secretion [22]. However, additional decrease was not observed in FTY720treated SphK2 knockdown cells, suggesting that SphK2 is required for lowering Ab secretion by FTY720 (Fig. 3B). Next, we examined the effects of overexpression of SphK2 or its dominant negative mutant (G243D) in N2a cells. As reported previously, overexpression of SphK2 increased Ab production [22] (Fig. 3C). Intriguingly, FTY720 treatment significantly decreased Ab secretion from N2a cells that overexpress wild-type (WT) SphK2, but not dominant negative mutant, to levels lower than those of untransfected cells treated by FTY720. Quantitative comparison of the inhibitory effects of FTY720 revealed that an increase in SphK2 activity significantly sensitized N2a cells to the inhibitory effect of FTY720 on Ab production (Fig. 3D). These data strongly suggest SphK2 activity is involved in the mechanism of action of FTY720 to lower Ab production, raising the possibility that FTY720-P is the bona fide regulator of the c-secretase activities.
Phosphorylation of FTY720 by SphK takes place in the cytosol, and the resultant FTY720-P translocates to the extracellular side and acts as an agonist for S1PRs [38]. Next we examined whether known downstream signaling pathway of S1PRs was involved in the modulation of Ab production. S1PR1, a major target of FTY720-P, is a Gi coupled receptor [39]. FTY720-P caused a significant phosphorylation of ERK1/2, a known downstream event of S1PR1-Gi signaling cascade (Fig. S2) in a similar fashion to that by SEW2871 [40]. Phosphorylation of ERK1/2 induced by SEW2871 was decreased by an authentic S1PR1 antagonist, W123 [41]. However, neither SEW2871 nor W123 affected the Ab productions at indicated doses (Fig. S3). Then we tested cotreatment of W123 or suramin, the latter being known to work as a Gi protein inhibitor [42] together with FTY720. We found that both compounds failed to affect the decremental effect of FTY720 on Ab production ( Fig. 4A and B). In sharp contrast, extracellular addition of FTY720-P did not affect the Ab production from N2a cells (Fig. 4C). These results raise the possibility that the molecular mechanism whereby FTY720 lowers Ab production is independent of its antagonistic effects neither on S1PR1 nor Gi pathways and that intracellular FTY720-P lowers Ab by an as yet identified mechanism.
In vivo effects of FTY720 on Ab levels in brains of AD model mice FTY720 is known to cross the blood brain barrier and be accumulated in brains to suppress the inflammatory response in model mice of MS [19]. These data prompted us to test the in vivo effects of FTY720 in AD model mice. We subcutaneously injected FTY720 (0.5 mg/kg/day) for 6 days into 6-month-old female A7 transgenic mice overexpressing human APP carrying Swedish and Austrian double mutations [37]. At 6 months, A7 mice do not show amyloid plaques, whereas the levels of soluble Ab are moderately increased compared to those in younger animals. Unexpectedly, we found that the level of soluble Ab40 was decreased, whereas that of Ab42 was significantly increased by a 6days treatment with FTY720. This suggests that FTY720 treatment impacts on brain Ab levels in vivo, although with some difference to its cellular effects.

Discussion
In this study, we examined the effect of S1PR modulators, FTY720 and KRP203, on brain Ab and showed these compounds decreased Ab production in neuronal cells. S1PR modulators decreased Ab production also from bCTF, an APP-derived direct substrate of c-secretase, without affecting AICD production nor the cleavage of Notch, suggesting that these reagents inhibited the carboxypeptidase-like c-secretase activity. Sphk2 activity was required for the FTY720-mediated decrease of Ab production, whereas the signaling cascade downstream of S1PR was dispensable for the effects of the FTY720. FTY720 was active also in vivo and decreased the levels of Ab40 but increased those of Ab42 in brains of APP transgenic mice. These results shed light on the complex regulatory function of S1PR modulators on brain Ab.
Phosphorylation of FTY720 or KRP203 by SphK2 yields the active metabolite FTY720-P or KRP203-P, respectively, which acts as the ligand for S1PR1 [14,15,38]. It has been shown that phosphorylated S1PR modulators induce endocytosis of S1PR1 from the cell surface after binding to S1PR1, thereby causing the antagonistic effects [14,15]. Our RNAi suppression and overexpression studies showed that the SphK2 level was correlated to the capacity of FTY720 to decrease Ab production (Fig. 3), supporting the view that phosphorylated forms of S1PR modulators are the bona fide active species that lower Ab production. However, neither an authentic S1PR1 receptor an agonist SEW2871 nor antagonist W123 altered the Ab production (Fig. S3). Moreover, cotreatment with W123 [41] or suramin [42] failed to cancel the inhibitory effect of FTY720 (Fig. 4), raising the possibility that the unconventional GPCR cascade is involved in the modulation of Ab production. Interestingly, it has been shown that b2 adrenergic receptor [43] and an orphan G-protein receptor GPR3, the latter exhibiting a significant homology to S1PRs [44], affect the csecretase activity through the b-arrestin pathway [45,46,47]. b-Arrestins redistributes the c-secretase complex toward detergentresistant membranes, thereby increasing the catalytic activity of the complex [45,47]. This raises the possibility that S1PR modulators also alter the distribution of c-secretase to regulate the APP-specific processing to generate Ab. Another possibility would be that FTY720-P directly binds to c-secretase or APP. FTY720 is known to accumulate within endosomes [48], in which mature c-secretase and APP reside [10]. Interestingly, it has been shown that phosphatidylinositol 4,5-diphosphate (PI(4,5)P2), a representative phospholipid with a second messenger activity, directly binds to c-secretase and inhibits its activity [49]. However, we observed extracellularly treated FTY720-P did not significantly alter the Ab production. It would be important to note that controlled phosphorylation of S1PR modulators at appropriate intracellular locations is critical to their function [50]. Moreover, we are unable to rule out the possibility that other S1PRs or unknown receptor is involved in the FTY720-mediated c-secretase inhibition. Further study is required to determine the precise intracellular site of phosphorylation, and distribution, of FTY720-P in specific membrane microdomains or organelles. The binding of FTY720-P to c-secretase or APP should also be examined.
In contrast to the in vitro results in cells, a 6-days treatment of FTY720 decreased Ab40, but increased Ab42 in the brains of APP transgenic mice in vivo (Fig. 4D). We used a similar dosage of FTY720 that had been adopted in autoimmune model mice, which is expected to yield submicromolar levels of FTY720 in brains [19]. Thus, the unexpected rise in Ab42 levels caused by FTY720 in vivo in brains might have been due to other mechanisms that are distinct from the modulation of the csecretase activity in cultured cells. Another possibility is that FTY720 affected the Ab42 levels by negatively regulating the inflammatory responses in the central nervous system. FTY720 has been reported to inhibit the egress of T cell into the spinal cord in autoinflammatory response [19], as well as into the ischemic lesions in brain ischemia [51]. Interestingly, FTY720 also inhibited the migration of human monocytes induced by Ab42 [52]. It is tempting to speculate that the FTY720 might have impacted on Figure 3. SphK2 activity is required for FTY720 mediated decrease of Ab secretion. (A) N2a cells were transfected with siRNA against murine SphK2. After 48 hrs transfection of siRNA, levels of SphK2 was detected by immunoblotting (upper panel) and quantified (lower graph n = 3, mean 6 SEM). (B) After 48 hrs transfection of siRNA, cells were treated with FTY720 for 24 hrs. Levels of secreted Ab were quantified by ELISA (n = 3, mean 6 SEM; *P,0.05, **P,0.01, ***P,0.001 compared with DMSO treatment or siRNA against SphK2 (indicated by line)). One-way ANOVA with Tukey's post hoc test for individual treatment differences was used for statistical analysis. (C, D) N2a cells were transiently transfected with LacZ, wildtype (WT) or dominant negative mutant (G243D) SphK2. After 24 hrs transfection, cells were treated with FTY720 for 24 hrs. (C) Levels of secreted Ab (n = 4, mean 6 SEM; *P,0.05, **P,0.01, ***P,0.001 compared with DMSO treatment or SphK2 (indicated by line)). (D) The inhibitory efficiency of FTY720 on Ab secretion compared with DMSO treatment in each transfection of (C). Secreted Ab levels of FTY720 were standardized by vehicle control in each group (mean 6 SEM; **P,0.01). doi:10.1371/journal.pone.0064050.g003 the Ab42 metabolism in vivo through altering the clearance of Ab42 by microglial cells, overriding the inhibitory effects on neurons. As FTY720 has been approved for MS therapy in clinics, the regulatory mechanisms whereby S1PR modulators impact on Ab42 metabolism, as well as on the inflammatory responses in AD brains, should further be characterized, in the light of therapeutics as well as adverse effects in humans. In sum, we identified a novel role of S1PR modulators on Ab, which may open up a novel aspect in Ab metabolism and lead to a novel therapeutic strategy for AD. Figure S1 AlamarBlue assay of N2aNH cells treated with S1PR modulators. After 24 hr treatment with S1PR modulators, N2aNH cells were incubated with cultured medium containing almarBlue (Invitrogen). Medium was collected to monitor fluorescence at 530-560/590 nm excitation/emission wavelengthes. (TIF) Figure S2 Effect of S1PR1 agonist and antagonist on ERK1/2 phosphorylation. Immunoblotting analysis of N2a cell lysates for ERK1/2 phosphorylation. Prior to stimulation with the indicated compounds, N2a cells were starved in serum free medium for 6 hr. Results of densitometric analysis of phosphorylated ERK1/2 (compared with control) are shown below the columns. (A) N2a cells were incubated for 10 min with the FTY720 (1 mM), FTY720-P (1 mM), SEW2871 (1 mM), and PKC activator PMA (Phorbol 12-Myristate 13-acetate; 1 mM), which in known as an activator of ERK1/2 phosphorylation. (B) N2a cells were preincubated with or without 1 mM W123 for 10 min. After addition of 1 mM SEW2871, cells were further incubated for 30 min and harvested for immunoblotting. (TIF) Figure S3 Dose dependent responses of S1PR1 agonist and antagonist on N2aNH cells. N2aNH cells were treated with SEW2871 or W123 for 24 hr at indicated doses. These reagents have no toxicity at 0.3, 1, 3, and 10 mM (A; alamarBlue assay), and failed to affect the Ab40 production at indicated doses (B). (TIF)