Interferon beta (IFNβ) (also known as Betaseron, Avonex, and Rebif) is a cytokine that is produced by components of the systemic immune system and also within the CNS by astrocytes (Javed and Reder 2006; Boutros et al. 1997). The finding that IFNβ is most clinically effective in the early, inflammation-driven stages of MS suggests primary effects via immunomodulation. IFNβ induces changes in expression of several hundred genes in immune cells that are relevant for cell activation, migration, proliferation, and effector functions (Javed and Reder 2006). As regards the latter, ligands that are members of the tumor necrosis factor (TNF) superfamily, namely, Fas and TNF-related apoptosis induced ligand (TRAIL), are upregulated by IFNβ (Greil et al. 2003). Both can induce caspase-dependent signaling in target cells that leads to programmed cell death. Fas- and TRAIL-deficient animals develop autoimmune disorders. The extent of Fas and TRAIL upregulation in lymphocytes has been correlated with a favorable disease course in MS (Wandinger et al. 2003). However, the same TRAIL-dependent mechanisms are also shown to underlie lymphocyte-mediated injury of neural cells in slice culture systems (Aktas et al. 2007). IFNβ would have the potential to worsen neural injury by upregulating the expression of TRAIL and Fas ligands on lymphocytes while the receptor counterparts are concurrently upregulated on oligodendroglia under proinflammatory conditions, as observed following exposure to IFNγ in vitro and in MS lesions in vivo (Wosik et al. 2003). Furthermore, upregulation of p53 in human adult CNS-derived OLs in vitro results in increased Fas and TRAIL receptor expression and enhanced susceptibility to ligand-mediated toxicity (Wosik et al. 2003). Increased levels of p53 expression might be expected to occur in response to an array of insults and has been observed in OLs in MS lesions that featured prominent OL loss (Wosik et al. 2003).

Whether IFNβ therapy could enhance repair by sparing neural cells from immune-mediated damage is unclear based on recent animal studies. IFNβ administration to EAE-afflicted animals can dampen or exacerbate clinical symptoms when disease is induced by Th1 versus Th17 lymphocytes, respectively (Axtell et al. 2010). Coadministration with estrogen receptor ligand β decreases EAE severity while preserving axonal densities in the spinal cord (Du et al. 2010), and IFNβ enhances regeneration of axons in a sciatic nerve crush model (Zanon et al. 2010). IFNβ can also induce neurotrophic factors in a number of cell types. Peripheral blood mononuclear cells (MNCs) derived from MS patients are reported to produce neurotrophic factors during IFNβ therapy (Caggiula et al. 2006; Yoshimura et al. 2010). Exposing human brain endothelial cells (HBECs) to MNCs pretreated in vitro with IFNβ, or to MNCs derived from IFNβ-treated patients, induces nerve growth factor (NGF) production by these cells (Biernacki et al. 2005). It has not been determined whether this represents the pro or mature form of the growth factor; the former may have deleterious effects for neural cells (Hempstead and Salzer 2002). IFNβ itself did not induce NGF production by HBECs (Biernacki et al. 2005). The induction of NGF production by HBECs was only observed when MNCs were derived from MS patients at the early onset phase of the disease and correlated inversely with EDSS and MRI lesion burden (Biernacki et al. 2005), suggesting a possible protective rather than injury-mediating function.

In vitro studies have demonstrated that IFNβ can directly influence neural cell properties. NGF is induced in astrocytes following exposure to IFNβ (Boutros et al. 1997). In addition, IFNβ decreases adult mouse neural progenitor apoptosis under death-inducing conditions in vitro (Hirsch et al. 2009) and decreases neurosphere proliferation (Lum et al. 2009). However the systemically administered recombinant cytokine is unlikely to reach high enough levels in the CNS to stimulate this process. IFNβ is also endogenously produced by astrocytes; its activity can be inhibited in vitro by anti-IFNβ antibodies that develop in some IFNβ-treated patients (Shapiro et al. 2006). The titers of the antibodies in serum are sufficiently high in some IFNβ-treated patients such that some access to the CNS would be predicted based on an expected 1:300 ratio between levels of serum and CSF antibodies. Corticosteroid administration at the onset of IFNβ therapy decreases the risk of development of these neutralizing antibodies (Zarkou et al. 2010). Together, this literature suggests that IFNβ therapy has the potential to indirectly influence neural cell properties by modulating peripheral immune components directly or by inducing anti-IFNβ antibodies which interfere with the regulatory function of endogenous IFNβ produced within the CNS. This can potentially have either advantageous or deleterious consequences for neural function or repair.

Glatiramer acetate (GA, also known as copaxone or copolymer-1) is a synthetic random polymer of four amino acids (glutamine, lysine, alanine, tyrosine) that are found in the same molar ratio as in MBP. The protective/reparative effects of GA have been linked with the demonstrated capacity of this agent to induce a shift in polarization of GA-reactive T cells from a Th1 to a Th2 phenotype in the periphery which is maintained in the CNS (Aharoni et al. 2000). Th2 polarized lymphocytes can alter the antigen presentation properties of microglia/macrophages, resulting in the subsequent polarization of naïve T cells into a Th2 rather than Th1 phenotype (Kim et al. 2004). CNS injuries have been suggested to demonstrate a shift toward a Th2 cytokine profile. This has been hypothesized to be a necessary component of repair in the CNS subsequent to injury by promoting neural regeneration and preventing autoimmune processes (Hendrix and Nitsch 2007), supporting the concept of “protective immunity” (Kipnis and Schwartz 2002). Th2-polarized lymphocytes can migrate across an endothelial barrier in vitro more effectively than Th1-biased lymphocytes (Biernacki et al. 2001) and may therefore be expected to more readily impact neural cell properties. To be established is whether “immune deviated” myelin antigen-reactive lymphocytes induced by antigen-specific therapies show similar protective/repair properties as GA-reactive T cells.

GA administration to MS patients results in fewer “black holes” on MRI scans, which can be interpreted as less axonal damage or enhanced repair (Bitsch et al. 2001). Systemic administration of GA has been shown to reduce the extent of neuronal injury in a number of experimental models including CNS trauma, neurodegenerative, and autoimmune diseases (Kipnis et al. 2000; Kipnis and Schwartz 2002; Schori et al. 2001; Aharoni et al. 2003; Benner et al. 2004). For instance, GA enhanced OPC numbers and promoted remyelination in a rodent spinal cord demyelination model (Skihar et al. 2009) and prevented demyelination and neuronal loss, stimulated neurogenesis and remyelination, and encouraged axonal sprouting/regeneration in relapse-remitting and chronic EAE (Aharoni et al. 2005, 2008, 2011). The protective/repair effects of GA-reactive T cells within the CNS, whether completely related to their Th2 properties or not, could be mediated directly on the target cell or indirectly via modulation of disease-relevant properties of glial cells. Both in vitro and in vivo studies have demonstrated that GA induces production of neurotrophins such as BDNF, NT3, NT-4 (Aharoni et al. 2003; Ziemssen et al. 2002; Skihar et al. 2009), growth factors such as IGF-1 (Skihar et al. 2009), and cytokines such as TGFβ (Aharoni et al. 2003, 2005). These can be predicted to protect neural cells against insults (Riley et al. 2004), rescue degenerating neurons (Riley et al. 2004), and affect regeneration of mature OLs and their progenitors (Althaus 2004). Administration of GA to EAE-afflicted animals at various clinical stages resulted in a sustained augmentation of levels of neurotrophins in infiltrating lymphocytes, neurons, astrocytes, and newly migrated neural progenitors within the lesion resulting in attenuation of axonal damage (Aharoni et al. 2003). GA also increases growth and neurotrophic factor production in remyelinating lesions of the rodent spinal cord, and by GA-reactive lymphocytes in vitro, resulting in enhanced repair and OPC specification/differentiation, respectively (Aharoni et al. 2008; Skihar et al. 2009; Ziemssen et al. 2002; Zhang et al. 2010).

Other glial cell functions that could be influenced by transmigrated immune cells in GA-treated patients would include the capacity to produce or remove potential injury mediators (such as glutamate and free radicals), production of trophic molecules, and expression of signaling molecules that modulate migration, differentiation, and myelin production by progenitor cells. Some have suggested that GA-carrying dendritic cells have the potential to cross the BBB (Liu et al. 2007), raising the potential of thus far uninvestigated direct effects of GA on neural cells.

A number of cytotoxic agents have been used typically for patients who fail conventional therapies. Examples of such therapies include cyclophosphamide, mitoxantrone, and cladribine. Subcutaneous administration of cladribine (2-chlorodeoxyadenosine) is associated with decreased levels of proinflammatory cytokines and chemokines in the CSF of treated MS patients in remission, possibly indicating a potential effect on CNS cells in addition to peripheral suppression (Bartosik-Psujek et al. 2004). A strategy being adopted from other autoimmune disorders is to administer such agents as initial induction therapies to be followed by immunomodulatory therapy. Such therapies can induce initial loss of brain volume that cannot be attributed to eliminating inflammation. This was observed in the Canadian immunoablative therapy protocol, in which patients underwent serial MRI studies before and after the immunoablation phase that included administration of CNS-accessible agents (busulfan), followed by infusion of autologous bone marrow-derived stem cells (Freedman 2007).

Minocycline is a semisynthetic tetracycline that has been in clinical use for nonneurologic indications (e.g., acne) that readily crosses the BBB. The agent has been shown to alleviate CNS pathology in a number of neurodegenerative and inflammatory animal models including stroke, spinal cord injury, Parkinson’s disease, and multiple sclerosis (Arvin et al. 2002; Brundula et al. 2002; Popovic et al. 2002; Wu et al. 2002; Lee et al. 2003; Fox et al. 2005; Nikodemova et al. 2007, 2010). Mechanisms held to account for this include direct effects on blockade of microglia/macrophage activation, inhibition of matrix metalloproteinases, and antiapoptotic effects (Yrjanheikki et al. 1998; Popovic et al. 2002; Elewa et al. 2006; Nikodemova et al. 2007), as well as indirect effects by limitation of CD4+/CD8+ T cell entry into the CNS and peripheral (but not central) immune deviation (Popovic et al. 2002; Nikodemova et al. 2010). Minocycline can also promote the survival and myelination of OPCs transplanted into rodents (Zhang et al. 2003). Clinical trials in MS patients have shown minocycline to beneficial in reducing the number of active scans and minimizing brain volume change (Zhang et al. 2008). However, the neuroprotective properties of minocycline are unclear at the moment given some reports of worsened disease or lack of neuroprotection in animal models (Saganova et al. 2008); improved outcome appears to depend on early initiation of treatment, high dosage, and direct delivery to the CNS (Xue et al. 2010). Of concern are recent clinical trial results showing that minocycline had a harmful effect on patients with amyotrophic lateral sclerosis, an effect not predicted by studies in the SOD1 animal model of this disease (Gordon et al. 2007).

Statins describe a class of drugs that competitively inhibit the rate-limiting enzyme in the mevalonate pathway, 3-hydroxy-3-methyl-glutaryl-coenzyme A (HMG CoA) reductase (HMGR), primarily reducing the biosynthesis of cholesterol and the posttranslational lipid moiety attachments, isoprenoids (Edwards and Ericsson 1999). Their beneficial effects in EAE were attributed to their systemic immune-regulatory actions resulting in changes in protein isoprenylation (Greenwood et al. 2006; Peng et al. 2006; Pahan et al. 1997; Youssef et al. 2002). Based on these studies, simvastatin was selected for a clinical trial in MS and was found to reduce the number of newly emerging lesions by MRI at a dose substantially higher than those typically prescribed for hypercholesterolemia (Vollmer et al. 2004). Statins dampen the expression of proinflammatory cytokines, chemokine receptors, adhesion molecules, and MMPs that would normally facilitate lymphocyte transendothelial migration (Lindberg et al. 2005; Pahan et al. 1997). Additionally, lipophilic statins (i.e., lovastatin, simvastatin) can passively diffuse through cell membranes and directly access the CNS (Saheki et al. 1994). The expression of HMGR in all cell types along with the penetration of this therapy into the brain highlights the potential direct effects of statins on neural cell properties. Accordingly, statins have been attributed neuroprotective properties in neurodegenerative models of stroke, vascular dementia, Alzheimer’s disease, and Parkinson’s disease, although the mechanisms through which this occurs are not yet fully understood (Wang et al. 2011).

Statins can prevent the release of proinflammatory mediators such as prostaglandin E2, nitric oxide, TNFα, IL-1β, and IL-6 from activated microglia and astrocytes (Pahan et al. 1997; Tringali et al. 2004). Treatment of astrocytes with lovastatin also impairs the function of glutamate synthetase which normally degrades excessive cytotoxic amounts of glutamate sequestered from the extracellular environment (Chou et al. 2003). In vitro studies have also indicated that statins can have variable effects on neuronal morphology and cell survival depending on dose, cell maturity, neuronal subtype, length of treatment, and environmental stress (Kumano et al. 2000; Michikawa and Yanagisawa 1999). Simvastatin has been shown to promote neuritic outgrowth on myelin substrates that would normally prevent regeneration and growth of axons (Holmberg et al. 2006). Statins also have beneficial effects on nerve conduction and generation of neuroblasts in neurodegenerative models (Tarhzaoui et al. 2009; Chen et al. 2009). In vitro studies have demonstrated the ability of lipophilic statins to influence cellular processes in human and rodent OPCs that have been implicated in the remyelination process (Miron et al. 2007; Paintlia et al. 2005). For instance, simvastatin treatment can induce an initial process extension, enhanced differentiation, and impaired spontaneous migration via inhibition of isoprenoid synthesis (Miron et al. 2007). Prolonged treatment induced process retraction via inhibition of cholesterol production and decreased viability and cell death caused by interference with both isoprenoid- and cholesterol-dependent functions in the cells (Miron et al. 2007). Mature oligodendrocytes treated with statins demonstrate similar responses to OPCs with regard to cytoskeletal dynamics and survival (Miron et al. 2007). Statin treatment may have altered the activity levels of the normally isoprenylated Rho GTPases to regulate effects on process extension, cell survival, differentiation, and migration. In addition, statin exposure may have altered the dynamics of cholesterol-rich lipid rafts important for OL process outgrowth (Decker and Ffrench-Constant 2004; Michikawa and Yanagisawa 1999).

Myelin contains a high content of lipids which consist of the majority of its dry weight, over a quarter of which is composed of cholesterol (Morell and Jurevics 1996). Importantly, brain-derived cholesterol cannot be replaced by dietary or peripheral sources (Jurevics and Morell 1995; Morell and Jurevics 1996; Saher et al. 2005). Short-term daily treatment of EAE-afflicted animals with an immunomodulatory dose of lovastatin was shown to increase myelin recovery in the spinal cord by enhancing survival as well as proliferation and differentiation of OPCs (Paintlia et al. 2004). This was attributed to the creation of an environment favorable to remyelination by altered inflammation and neurotrophic factor production. However, these studies could reflect either a direct impact of the statin on neural cells or a sparing of immune-mediated injury consequent to anti-inflammatory activity. This was addressed by using a relatively non-immune-mediated model of brain-based demyelination (oral cuprizone administration) to assess the direct impact of statins on myelin maintenance and remyelination (Klopfleisch et al. 2008; Miron et al. 2009). Both short-term and long-term administration of an immunomodulatory dose of the more potent simvastatin impaired remyelination when statin was administered during the period of concomitant de- and remyelination, the entire period of OPC responses to demyelination, or the remyelination period alone. Effects were associated with a decrease in OPC numbers when the drug was applied during the period of concomitant de- and remyelination, and an increase in OPCs in an immature state when statin was applied during the remyelination period suggesting a block in differentiation (Miron et al. 2009). Together, these findings suggest the potential of lipophilic statins to have both deleterious and advantageous effects on the function of various neural cell types and repair/regeneration in the CNS.

FTY720 (fingolimod/Gilenya) is a recently clinically approved oral agent for the treatment of relapsing MS that induces profound reduction in numbers of lymphocytes circulating in the peripheral blood due to its capacity to enhance ingress and reduce egress of CCR7-expressing naïve and central memory lymphocytes from regional lymph nodes (Compston and Coles 2002). This agent was initially evaluated (and concluded to be ineffective) as a therapy to prevent organ rejection in transplant patients. Based on these immunomodulatory properties, this agent has been successfully used in clinical trials in relapsing–remitting MS and is being evaluated in primary progressive MS clinical trials (Kappos et al. 2006a, b). FTY720 decreases the rate of neurodegeneration in patients compared to an IFNβ-1a-treated cohort, as measured by brain volume loss (Khatri et al. 2011). The clinically administered native form of FTY720 can readily cross the BBB by virtue of its lipophilicity (Sanchez and Hla 2004). Once it has entered the brain parenchyma, it is rapidly phosphorylated by sphingosine kinase (SphK) and consequently biotransformed into an analogue of the endogenous bioactive lysophospholipid, sphingosine-1-phosphate (S1P) (Billich et al. 2003; Zemann et al. 2006). The active phosphorylated form of the drug can bind to four of five known G protein-coupled receptors which belong to the endothelial differentiation gene-related (EDG) family, termed S1P1, 3, 4, and 5 or EDG 1, 3, 6, and 8, respectively (Davis et al. 2005; Mandala et al. 2002). Generally, S1P3/4/5 activation leads to G_12/13 signaling and process retraction via the cytoskeletal modulator RhoA GTPase (Jaillard et al. 2005; Toman et al. 2004). S1P1 signaling to G_i/o is associated with Rac1, Ras, and extracellular signal-related kinase (ERK) 1/2 activation and subsequent process extension, survival, migration, and proliferation (Goetzl and Rosen 2004; Toman et al. 2004). S1P receptors are ubiquitously expressed throughout the immune and central nervous systems. The biological basis behind its anti-inflammatory effects reflects the ligation of the bound S1P receptor on circulating naïve and central memory lymphocytes, with eventual endocytosis of the bound receptor, downregulation of the receptor at the mRNA level, and subsequent blockade of S1P1-dependent efflux from secondary lymph nodes to target organs such as the brain (Goetzl and Rosen 2004; Liu et al. 1999; Sawicka et al. 2005). Fingolimod has previously been shown to have neuroprotective effects in experimental models of spinal cord injury, stroke, and multiple sclerosis (Zhang et al. 2009; Hasegawa et al. 2010; Foster et al. 2007).

Audoradiographical analysis of rodents that were orally administered C¹⁴-labelled FTY720 for 1 week indicate that both the native drug and the phosphorylated metabolite entered the brain parenchyma, steadily increased in concentration over time within this compartment, and was concentrated in myelin sheaths (Foster et al. 2007). Levels of the bioactive form of FTY720 are within the micromolar range in the brain, whereas subnanomolar concentrations are measured in the CSF (Foster et al. 2007).

FTY720 treatment of relapsing EAE animals, either prophylactically or during the chronic phases of the disease, can prevent or reverse clinical disability, respectively (Fujino et al. 2003; Webb et al. 2004; Kataoka et al. 2005; Al-Izki et al. 2011), whereas a recent study has demonstrated that clinical deterioration is not prevented when FTY720 is administered during progressive EAE (Al-Izki et al. 2011). Microarray profiling has revealed that FTY720 treatment of EAE animals also results in a downregulation of proinflammatory cytokine and chemokine-related genes and an increase in transcript levels of myelin-related genes (Foster et al. 2007). Given that the clinically administered native form of FTY720 can readily cross the BBB, such responses in the EAE model could reflect effects in the periphery, at the BBB, and/or on neural cells. The efficacy of FTY720 in relieving clinical disability in the later stages of EAE where inflammation has subsided suggests a neuroprotective effect (Fujino et al. 2003; Webb et al. 2004; Foster et al. 2007).

The combination of FTY720’s crossing of the BBB and the expression of S1P1 and S1P3 on human-derived endothelial cells indicates the potential of this agent to impact on endothelial barrier properties in treated patients (Lin et al. 2007). FTY720 has been shown in vivo to promote adherens junction assembly between endothelial cells and decrease BBB permeability via S1P1 signaling (Lin et al. 2007; Foster et al. 2009). Astrocytes and microglia also have the potential to respond to S1P receptor signaling, with the expression of S1P1, 3, and 5 receptor isoforms. FTY720 treatment of astrocytes in vitro promotes their migration and survival via S1P1 signaling (Mullershausen et al. 2007; Osinde et al. 2007). Selective loss of S1P1 expression in astrocytes abrogates the protective effect of FTY720 in EAE (Choi et al. 2011). These findings suggest that S1P receptor signaling on astrocytes can influence cellular events that contribute to the inflammatory response and/or to a nonsupportive microenvironment, i.e., a parallel with the glial scar found in MS lesions. Studies using rat-derived microglia have indicated that their relative S1P receptor levels are modulated based on their activation state and that S1P receptor engagement on these cells alters their cytokine profile (Tham et al. 2003). This agent also increases microglia and astrocyte cell numbers ex vivo and in vivo during remyelination (Miron et al. 2010; Kim et al. 2011). Furthermore, FTY720 inhibits macrophage migration into inflammatory lesions (Zhang et al. 2007). Together these data indicate that FTY720 may influence their roles in phagocytosing debris within lesions and regulating the immune responses within the CNS. FTY720 also normalizes electrophysiological responses in EAE-afflicted animals (Balatoni et al. 2007). Although FTY720 has been shown to induce apoptosis in rat neurons, this only occurs at a concentration which would be supraphysiological in the CNS (Oyama et al. 1998). Although FTY720 does enter the brain parenchyma, it is not primarily localized to neurons suggesting limited potential to impact neuronal properties (Foster et al. 2007).

Rodent-derived mature OLs and OPCs express S1P receptors both in vitro and in vivo, in the relative abundance of S1P5  >  S1P1  =  S1P2  >  S1P3 for OLs and S1P1  =  S1P2  =  S1P5  >  S1P3 for OPCs (Yu et al. 2004; Jaillard et al. 2005; Novgorodov et al. 2007; Miron et al. 2007, 2008b). FTY720 has been shown to induce transient S1P3/5-dependent process retraction with subsequent S1P1-dependent process extension in human OPCs and adult mature OLs but only in rodent OLs matured in vitro when coapplied with neurotrophic factors (Miron et al. 2007, 2008a, b; Coelho et al. 2007). FTY720 induced cyclic down- and upregulation of S1P1 and S1P5 mRNA transcripts over time in cultured oligodendroglial cells, in contrast to the sustained downregulation of S1P1 observed in lymphocytes exposed to the drug in vivo (Miron et al. 2007, 2008a, b). FTY720 inhibits OPC differentiation and spontaneous migration yet does not influence directed migration to the chemoattractant PDGF (Jung et al. 2007). FTY720 was able to enhance OPC and OL survival under death-inducing conditions via S1P1 and S1P5 signaling, respectively (Miron et al. 2007, 2008a, b; Coelho et al. 2007; Jung et al. 2007). Contrasting reports exist on the ability of oral administration of FTY720 to promote OL survival and myelin integrity in experimental models of de- and remyelination (Kim et al. 2011; Hu et al. 2011). Kim et al. reported that 1 mg/kg/day FTY720 orally administered during 6 weeks of cuprizone-induced demyelination attenuated injury to OLs, neurons, and myelin, causing an increase in the percentage of myelinated fibers compared to control and associated with a reduced astro- and microgliotic response. However, when FTY720 was administered during the entire period of OPC responses (6 weeks during cuprizone administration and return to normal diet), an increase in astrogliosis was observed, thereby indicating that FTY720 can induce differential neural responses depending on the time it is administered following injury. Accordingly, Mi et al. observed that FTY720 (1 mg/kg/day) orally administered for 2 weeks at 4 weeks after the onset of cuprizone diet did not have any protective effect on the proportion of myelinated axons or lesion volume (Hu et al. 2011). Neither group observed an increase in remyelination in vivo, perhaps reflecting the difficulty in assessing enhanced remyelination in young animals where this process is rapid and effective. In contrast, application of physiological doses of fingolimod to organotypic mouse cerebellar slice cultures, which maintain the physiological cell–cell interactions and have mature compact myelin, enhanced remyelination following toxin-induced demyelination via S1P3/5 signaling (Miron et al. 2010). This was associated with process extension in both OPCs and mature OLs, as well as an increase in microglial and astrocytic components, which may have contributed to the repair process by production of factors beneficial to OPC and OL function. Differences with in vivo observations may indicate limited bioavailability of the drug in an animal compared to direct application in vitro. Mi et al. (2011) showed that FTY720 caused a decrease in developmental myelination in slices of P17 rat corpus callosum, which may reflect either differential susceptibility of myelin and OLs to FTY720 under myelinating and remyelinating conditions, or distinct responses of specific white matter tracts to the drug. Overall, the net effect of FTY720 on neural cells will reflect the relative levels of the S1P receptor subtypes on cells of interest, the modulation of these receptor levels in response to the environment, and counteracting signals emanating from the inflamed brain. Efforts are underway to assess more selective S1PR agonists, such as BAF312, which specifically binds S1P1 and S1P5, shows encouraging results in decreasing disease severity in EAE (Nuesslein-Hildesheim et al. 2010).