The future of multiple sclerosis treatments.

ABSTRACT Introduction: There are not many conditions in which the last few decades have brought such a major change in the landscape of treatments as is the case of multiple sclerosis (MS). A number of disease modifying treatments (DMTs) are presently available for the treatment of the inflammatory phase of this disabling disease; however, the need for treating neurodegeneration and halting the progression of disability is still unmet. Areas covered: In this paper we review the available information on existing and emerging DMTs and we discuss their place within the context of different treatment strategies in MS, Expert commentary: The future of MS treatments should include the development of new treatment strategies tackling disease progression, together with a better understanding of the side-effects and the best sequential strategy of implementation of available and emerging drugs.


Intoroduction
Multiple sclerosis (MS) is a chronic, immune-mediated condition causing inflammation and neurodegeneration in the brain and spinal cord and the most common nontraumatic cause of neurological disability in young people in the western world [1]. In most patients, MS initially has a relapsing-remitting course (RRMS) with bouts of inflammation (relapses) and periods of remission. In the majority of people with MS, the relapsing course will later be followed by a secondary progressive phase (SPMS) [2]. In around 15% of cases, the disease progresses from the beginning with a primary progressive course without superimposed relapses (PPMS) [2]. More recent classifications put emphasis on the inflammatory activity which can be present at all stages of the disease and can be targeted with disease-modifying treatments (DMTs) [3].
A number of DMTs are currently available for the treatment of RRMS, and their aim is to decrease the relapse rate (RR) and the inflammation within the central nervous system (CNS) [4]. The past 25 years have brought important changes in the treatment for MS. After several years in which first-line injectable DMTs interferon beta (IFNβ) and glatiramer acetate (GA) were the main treatment options, new medications with different regimens became available. The first oral DMT, fingolimod, was approved in 2010 in the US and soon after that in Europe. A number of other oral agents have been approved since or are currently in phase III trials or are due to be submitted to the regulatory agencies for approval [5]. Three monoclonal antibodies are now approved for MS treatment, and others are also in late-stage development. Nevertheless, the challenges raised by the protracted course of the disease, the presence of neurodegeneration in all MS stages, and the pathological burden (demyelination and axonal loss) inflicted over time by both inflammation and neurodegeneration raise challenges that are not yet completely tackled by the current therapies. In other words, despite a real breakthrough in treating MS, the available therapies are far from having sorted out the current unmet needs raised by the complexity of MS. Here, we provide an outline of the actual treatment landscape in MS and some prospects on its future development.

Injectable drugs
Three IFNβ preparations are in widespread use as first-line DMTs for relapsing MS (RRMS and SPMS with relapses) and in some countries for clinically isolated syndrome (CIS). Each of them was licensed following single multicenter, double-blind, placebo-controlled, phase III trials [6]. Two of these medications require subcutaneous administration and one is given intramuscularly. GA (Copaxone®, Teva) is a four amino acid synthetic copolymer based on the composition of myelin basic protein (MBP) [7]. GA was approved after a single, phase III, multicenter, randomized, placebo-controlled clinical trial [8]. GA is approved for RRMS and, in some countries, for patients with CIS. IFNβ and GA have different immunomodulatory effects but relatively comparable efficacy, reducing the RR by approximately 30% [9]. Data from a large observational cohort study recently showed that treatment with IFNβ and GA reduces disability progression measured by Expanded Disability Status Scale (EDSS) scores over 6 years of treatment [10]. Both IFNβ and GA are generally safe and well tolerated. Nevertheless, both IFNβ and GA require regular, longterm, self-injections. Side effects of IFNβ preparations include flu-like symptoms, an increase in liver enzymes, and injection-site reactions. Side effects of GA include local injection-site reactions and postinjection reactions which occur in about 15% of people [11]. Issues of adherence and tolerance may therefore reduce the likelihood of achieving durable treatment efficacy [12].
Approved therapies such as the humanized antibodies natalizumab, alemtuzumab [13], and daclizumab and the currently less used cytostatic agent mitoxantrone are more effective, but their use is associated with safety concerns. These drugs are administered parenterally and can have potentially severe side effects (e.g. progressive multifocal leukoencephalopathy [PML] for natalizumab; autoimmune-associated conditions for alemtuzumab; liver injury, colitis, and skin reactions for daclizumab; and cardiotoxicity and acute leukemia for mitoxantrone). They are, therefore, reserved for the treatment of highly active MS.
Natalizumab is a humanized recombinant monoclonal antibody directed against α4-integrin [13]. Natalizumab interferes with leukocyte migration from the peripheral blood into the CNS by preventing its binding via α4-integrin to the vascular cell adhesion molecule on endothelial cells [13]. This step has an impact on CNS inflammation as it blocks the adhesion and subsequent migration of lymphocytes across the blood-brain barrier (BBB). The drug was suspended in 2005 by the manufacturer [14] following two cases of PML in the SENTINEL (Safety and Efficacy of Natalizumab in Combination with Interferon Beta-1a in Patients with Relapsing Remitting Multiple Sclerosis) trial in which it was given in combination with intramuscular IFNβ-1a [15]. In the pivotal, placebo-controlled phase III trial which led to its approval, natalizumab administered in the dose of 300 mg intravenously (IV) monthly reduced RR by 68% and sustained progression of disability at 2 years by 42% [14] and magnetic resonance imaging (MRI) activity by 92% [16]. Natalizumab was reintroduced in 2006 with revised labeling and after the introduction of risk management programs [17]. The PML risk stratification for people with MS on natalizumab takes into account duration of treatment, prior immunosuppressant use, and the anti-JC virus (JCV) antibody status reflecting infection with JCV [18,19]. This is allowing further risk stratification during natalizumab treatment [20]. In 4-6% of cases, natalizumab treatment may induce the formation of persistent neutralizing antibodies (NABs), usually within the first 12 months. The NABs are associated with higher rates of infusion-related adverse events and can lower the efficacy of the treatment [21].
Alemtuzumab is a humanized monoclonal antibody targeting CD52 expressed on lymphocytes, natural killer (NK) cells, monocytes, and some granulocytes [22,23]. Alemtuzumab produces rapid and profound lymphopenia lasting for years via antibody-dependent cellular cytotoxicity [13]. Alemtuzumab was compared to IFNβ-1a administered subcutaneously three times a week in two phase III trials of RRMS [24,25]. Alemtuzumab reduced the annualized RR (ARR) by 49-55%, MRI gadolinium-enhancing lesions by 61-63%, and the rate of disability progression by 30-42% [25,26]. The major safety concern with alemtuzumab is the autoimmune conditions (cumulative risk between 22% and 47%) [27,28] involving mainly the thyroid gland and blood cells (thrombocytopenia, hemolytic anemia, and pancytopenia) and nephropathies (in 0.3% of patients [29]). Prophylaxis of herpetic infections with oral acyclovir is required during and for 28 days after alemtuzumab infusion [28]. In Europe, alemtuzumab is licensed as a first-line medication in active RRMS; however, some neurologists would use it as a second-line drug because of the risk of secondary autoimmunity [21].
Daclizumab is a humanized neutralizing monoclonal antibody against the interleukin-2 receptor subunit CD25 on T cells [30]. Its effect on reducing CD25+ T cells is minimal, but it expands CD56 bright NK cells, and this correlates with the clinical effect [30]. Daclizumab is given either IV once every 4 weeks or subcutaneously once every 4 weeks (daclizumab high-yield process [DAC HYP]). Daclizumab showed promising effects on MRI outcomes in randomized double-blind trials (two phase II and one phase III trial) [31][32][33] either as add-on therapy to IFNβ-1a or placebo, and there were no indications of rebound effects after treatment interruption. In a phase III trial, DAC HYP 150 mg was compared to intramuscular IFNβ-1a and led to a 45% reduction in AAR, but no statistically significant difference in the reduction of sustained disability. An extension phase is ongoing. There had been two deaths in the phase II trials (one from a psoas abscess and one from autoimmune hepatitis), and one death in the DAC HYP not deemed to be related to the medication. The safety profile includes an increased incidence of cutaneous adverse events, serious cutaneous events, serious infections, and elevations in liver function tests [33]. The cutaneous effects are distinctive adverse events of daclizumab. A review of the cutaneous adverse events in the DECIDE (Daclizumab HYP versus Interferon Beta-1a in Relapsing Multiple Sclerosis) trial has recently been published [34]. Cortese et al. followed 31 participants in the phase I study of DAC HYP (NCT01143441) over 42 months and observed cutaneous adverse events in 77% of patients treated with daclizumab [35]. The majority of skin events consisted of patches of eczema requiring no treatment [35], while moderate-to-severe rash (some with psoriasiform phenotype) developed in 19% and required treatment discontinuation in 13%. The skin biopsies from the lesions had nonspecific features of eczematous dermatitis, with important CD56+ lymphocytic infiltrates [35], which were not related to the clinical severity and with no histopathologic posttreatment changes [35].
Daclizumab (Zinbryta®, Biogen, Inc. of Cambridge, Massachusetts) has very recently been approved by the US FDA for RRMS [36]. Daclizumab should be given to patients who have had an inadequate response to two or more MS drugs. Because of the risk of side effects, daclizumab has a boxed warning and is available only through a restricted distribution program under a Risk Evaluation and Mitigation Strategy [36]. Monitoring of liver function is required before commencing on daclizumab and monthly before each dose and then for up to 6 months after the last dose [36]. The boxed warning highlights the risk of severe liver injury which could be potentially fatal, noninfectious colitis, skin reactions, and lymphadenopathy [36].
Mitoxantrone is an anthracenedione that inhibits type II topoisomerase and disrupts DNA synthesis. Mitoxantrone was approved by the FDA in 2000 for rapidly worsening RRMS or SPMS after several clinical trials [37,38]. Mitoxantrone can cross the disrupted BBB and may induce microglial death as shown by in vitro studies [39].
Mitoxantrone is given in infusions at doses of 12 mg/m 2 monthly; however, the cumulative dose is limited due to cardiologic and hematologic adverse effects. The use of mitoxantrone has rapidly decreased due to the risk of severe complications such as acute leukemia [40] and the increasing number of alternative highly effective and less toxic treatment options [21].

Oral drugs
Teriflunomide is the metabolite of leflunomide, which has been approved for mild-to-moderate rheumatoid arthritis. Teriflunomide is given as 14 mg tablets once daily, and the oral bioavailability is almost 100% [41]. It inhibits the ratelimiting mitochondrial enzyme dihydroorotate dehydrogenase (DHODH) in de novo pyrimidine synthesis [41]. This leads to reduced proliferation of cells that need de novo synthesis of pyrimidine to divide. A salvage pathway independent of DHODH is enough for resting lymphocytes; however fast-proliferating cells such as activated lymphocytes are dependent on de novo synthesis, thus being a selective target of teriflunomide [42]. Other immunological mechanisms of teriflunomide have been suggested [41].
In two phase III trials in RRMS [43,44], teriflunomide reduced the ARR compared to placebo by 31-36%, the MRI gadolinium-enhancing lesions by about 80%, and the rate of disability progression by 26-27%. Teriflunomide had similar effects on the ARR and on time to a new relapse or termination of treatment compared with subcutaneous IFNβ-1a [45]. Teriflunomide has been tested in a randomized, double-blind, placebo-controlled trial on CIS patients with silent MRI lesions [46], and it delayed the time to a second relapse and a reduction in new MRI lesions. Teriflunomide is generally well tolerated; however, common adverse events include alanine aminotransferase (ALT) increase, headache, diarrhea, hair thinning, and nausea [47]. These are usually mild-to-moderate in intensity and are self-limiting [47]. ALT elevation is the most common reason for treatment discontinuation and a relatively frequent (every second week) ALT screening during the first 6 months of treatment and thereafter every second month is recommended [47].
Delayed-release dimethyl fumarate (DMF) is the most recently approved oral DMT for RRMS. DMF is administered as a 240 mg capsule twice daily. Its mechanisms of action are not completely understood; however, data suggest that they are mediated via the activation of the nuclear factor (erythroid-derived 2)-like 2 (Nrf2) pathway [48]. DMF was compared to placebo in two phase III trials in RRMS [49,50], and it reduced the ARR by 44-53%, the MRI gadolinium-enhancing lesions by about 75-94%, and the rate of disability progression by 22-32% [51]. DMF reduced the ARR by 24% and the rate of disability progression by 17% compared to GA [50]; however, the differences were not significant, and the study was not powered to detect statistically significant differences in treatment effect. An interim analysis of an ongoing long-term extension of the phase III trials showed that treatment with DMF was followed by continuously low clinical and MRI disease activity [52]. Common adverse events include flushing, nausea, diarrhea, and abdominal pain [51], are usually transient and mild to moderate in severity, and can limit patient tolerance of DMT [53]. The treatment may induce leucopenia and increase liver transaminases, and regular blood tests are therefore recommended. Prolonged severe lymphopenia (<500 cells/mm 3 ) has been suggested as a risk factor for PML [54]. In 2014, a case of PML was reported in a patient treated with DMF [55]. Four other PML cases have been previously reported in patients with psoriasis who had received fumaderm [56]. In JCV-positive patients with persistent lymphopenia, DMF should be stopped. Up to now, PML has not been observed during DMF therapy without lymphopenia and in the presence of appropriate monitoring and drug-discontinuation rules [57].
Fingolimod was the first of the oral agents for relapsing forms of MS to be approved by the FDA in 2010. Fingolimod is administered as a once-daily 0.5-mg capsule. Fingolimod is a nonselective sphingosine-1-phosphate (S1P) receptor modulator [58,59]. It acts as a functional antagonist, by internalizing and degrading the S1P1 receptor on lymphocytes (a receptor responsible for T lymphocyte exit from the lymph nodes, circulation, and T-cell differentiation) [58]. This leads to sequestration of T lymphocytes in secondary lymphatic tissues and thus reduces inflammation in MS [60]. An additional direct effect on the CNS has been suggested by animal studies [60].
In two phase III trials in RRMS, fingolimod compared to placebo reduced the ARR by 48-55%, the MRI gadoliniumenhancing lesions by more than 80%, and the rate of disability progression by 25-30% [61]. Compared to IFNβ-1a 30 μg, intramuscularly once-weekly fingolimod reduced ARR by 52%, the MRI gadolinium-enhancing lesions by more than 50%, and the rate of disability progression by 25% [62]. In a phase III trial in PPMS, fingolimod failed to delay disability progression compared to placebo [63].
Common adverse events include upper respiratory tract infection, diarrhea, cough, headache, and back pain [64]. Because it can cause bradycardia and atrioventricular block at first administration, it is recommended to have electrocardiogram monitoring continuously for 6 h after the first dose [65]. After one death due to a fulminant primary varicella zoster in one of the phase III trials [66], screening for previous varicella zoster infection is advised in patients due to start fingolimod, and vaccination is recommended if negative. Rare adverse events include elevated liver enzymes and macular edema, and blood and ophthalmological follow-ups are recommended [67].

Monoclonal antibodies
Three anti-CD20 agents (rituximab, ocrelizumab, and ofatumumab) that deplete pre-B cells and mature B cells without affecting plasma cells or progenitor cells in the bone marrow have been studied in MS [23]. Rituximab is a human-mouse chimeric monoclonal antibody against CD20. Rituximab has been used off-label for MS as well as neuromyelitis optica. Rituximab has been approved to treat B-cell lymphomas, rheumatoid arthritis, Wegener's polyangiitis, and microscopic polyangiitis [23]. In a phase II double-blind, placebo-controlled trial, rituximab reduced new MRI gadolinium-enhancing lesions by 91% and the proportion of patients with relapses [68]. Seventy-eight percent of treated patients experienced an infusion-related adverse effect. Infection incidence was similar in both groups [68]. A phase II trial of rituximab in PPMS failed to meet its primary end point of delaying confirmed disability progression (CDP) [69] and showed a 3.5% increase in risk of serious infections in the treated arm. In patients using rituximab for other indications, PML cases have rarely been reported [70]. Although rituximab rapidly and consistently decreases the numbers of peripheral CD20+ and CD19+ cells [71], a small phase II trial of intrathecal rituximab was terminated early because of low efficacy on the CSF biomarkers [72]. Rituximab induced an incomplete and transient depletion of B cells in the CSF, while the effects on peripheral B cells were complete and lasting [72]. No phase III trials of rituximab for MS have yet been performed. Some suggest that rituximab can be an option in RRMS patients who have failed to respond to first-and second-line therapies, in those with concomitant autoimmune disorders [73], or in people with stable RRMS who switch from natalizumab to other DMT due to increased PML risk [74].
Ocrelizumab and ofatumumab are humanized anti-CD20 monoclonal antibodies. In a phase II, placebo-controlled RRMS trial [75], two doses of ocrelizumab met the primary outcome of reducing the number of MRI-enhancing lesions. Infection side effects were equivalent between the two groups, but more infusion-related side effects occurred in the ocrelizumab group than in the placebo arm. Two phase III clinical trials (A Randomized, Double-Blind, Double-Dummy, Parallel-Group Study To Evaluate the Efficacy and Safety of Ocrelizumab in Comparison to Interferon Beta-1a [Rebif®] in Patients With Relapsing Multiple Sclerosis, OPERA I and II) for RRMS were completed in 2015 [76]. The trials compared IV ocrelizumab 600 mg every 6 months to subcutaneous IFNβ-1a. The studies reported that ocrelizumab reduced the annual RR by 46% in OPERA I and by 47% in OPERA II [76]. Also, they showed a reduction of clinical disability by 40%, as measured by the EDSS. Ocrelizumab reduced the number of T1 gadolinium-enhancing lesions in the brain by 94% and 95%, respectively [76]. Additionally, it resulted in a 77% and 83% reduction of hyperintense T2 lesions compared with IFNβ-1a. At 96 weeks, 47.9% of patients treated with ocrelizumab in OPERA I and 47.5% in OPERA II achieved the composite measure of 'no evidence of disease activity' (NEDA) defined as no MS relapses, no CDP, and no new or enlarging T2 or gadolinium-enhancing T1 lesions compared to 29.2% and 25.1% of patients treated with IFNβ-1a [77].
A recent phase III clinical trial (A Phase III, Multicentre, Randomized, Parallel-group, Double-blind, Placebo Controlled Study to Evaluate the Efficacy and Safety of Ocrelizumab in Adults With Primary Progressive Multiple Sclerosis, ORATORIO) tested ocrelizumab for PPMS and compared IV ocrelizumab 600 mg every 6 months to placebo [78]. The primary end point was time to onset of 12-week CDP, while secondary end points included time to onset of 24-week CDP, change in timed 25-foot walk, T2 lesion volume, and change in whole brain volume and safety [78]. Ocrelizumab was the first trialed drug to meet primary and key secondary efficacy outcomes in a phase III PPMS study. Ocrelizumab significantly reduced the relative risk of 12-week CDP by 24% and 24-week CDP by 25% [78], decreased the volume of T2 hyperintense lesions, and reduced the whole brain volume loss compared with placebo [78]. The frequency of adverse events and serious adverse events was similar in both groups [78]. The results of the OPERA I, OPERA II, and ORATORIO trials will be submitted for the approval of ocrelizumab to the FDA.
Ofatumumab is a fully human monoclonal antibody currently used for lymphocytic leukemia, which interacts with the early activation of the B lymphocyte and has lower potential for antigenicity. It was tested in a small phase II clinical trial with promising results, showing a 99% reduction of MRI activity, with no serious adverse events [79]. A larger phase II trial, Ofatumumab Subcutaneous Administration in Subjects With Relapsing-Remitting Multiple Sclerosis (MIRROR), tested the safety and effectiveness of ofatumumab compared to placebo in 232 patients with RRMS [80]. This study also showed a 90% reduction of MRI lesions after 12 weeks of treatment [80]. Five serious adverse events were reported in the 60-mg dose regimen, but no cases of PML or opportunistic infections were reported [80]. The authors concluded that the results support the further study of ofatumumab in clinical RRMS trials [80].

Cladribine
Cladribine is a cytotoxic drug, an adenosine deaminase-resistant purine nucleoside, used as a first-line chemotherapeutic agent in the treatment of hairy cell leukemia and other neoplasms, in its parenteral formulation [81]. Cladribine enters cells via purine nucleoside transporters [82]. Cladribine works preferentially on lymphocytes and monocytes by disrupting cellular metabolism, resulting in cell death [83], being incorporated into the DNA of the dividing cells. It selectively depletes the number of circulating T cells and B lymphocytes, having only a minor effect on NK cells [84]. The study on cladribine in MS started more than 20 years ago. Intravenous cladribine was initially evaluated for progressive MS (PPMS and SPMS) in two randomized, double-blind clinical trials [85,86] with some promising results. Later, cladribine was tested as an oral agent for RRMS; tablets being taken in two cycles of few days per year [84]. Cladribine for RRMS was tested in a placebocontrolled (A Safety and Efficacy Study of Oral Cladribine in Subjects With Relapsing-remitting Multiple Sclerosis, CLARITY) trial and reduced both the frequency and the severity of relapses and suppressed the MRI-enhancing lesions at 6 months [87]. The recently reported results of a 120-week extension demonstrated that in a majority of patients, the clinical benefits on relapses and disability as well as on MRI outcome measures of 3.5 mg/kg cladribine given in the first 2 years of the trial can be maintained for at least 4 years [88,89]. ONWARD (A Phase 2 Study of Cladribine Add-on to Interferon-beta (IFN-beta) Therapy in Multiple Sclerosis (MS) Subjects With Active Disease) was a 2-year, randomized, double-blind phase IIb study of cladribine tablets (3.5 mg/kg) as an add-on to IFNβ therapy in relapsing MS patients (including SPMS with ongoing relapses). The mean number of relapses was lower (23%) in patients treated with cladribine 3.5 mg/kg and IFNβ than in patients on placebo plus IFNβ (56%) [90]. The mean numbers of new enhancing T1 lesions and active T2 were reduced in the cladribine + IFNβ treatment arm vs. placebo plus IFNβ [90]. In the phase III trial Oral Cladribine in Early Multiple Sclerosis (ORACLE), 616 CIS patients received 1:1:1 cladribine 5.25 mg/kg, cladribine 3.5 mg/kg, or placebo [91]. Both doses of cladribine considerably delayed MS diagnosis compared with placebo [91]. A more recent analysis applied the 2010 McDonald criteria to the patient cohort at baseline and showed that the risk of further relapse and disability worsening was significantly reduced with cladribine 3.5 mg/kg compared to placebo [92]. Cladribine failed to get regulatory approval by the European Medicines Agency (EMA) because of concerns over the risks of cancers in the CLARITY active arm [93]. Cladribine was used for RRMS treatment in 2010 in Russia and Australia, but it was withdrawn afterwards. However, a recent meta-analysis of phase III trials of licensed DMTs for RRMS and the CLARITY trial did not support an increased cancer risk from cladribine in the doses used in CLARITY and ORACLE MS, which previously contributed to refusal of market authorization of cladribine by the EMA [94]. The authors concluded that longer-term follow-up is required to assess the safety profile of both cladribine and the currently approved DMDs to definitively assess cancer risk [94].

Laquinimod
Laquinimod is an orally available carboxamide derivative, derived from linomide, a drug that was proved to reduce the activity in RRMS, but with the cost of severe adverse events [95]. Laquinimod was or currently is tested for neurodegenerative disease such as Huntington's disease and also for relapsing remitting and progressive MS [96]. Studies on the animal model of MS (experimental autoimmune encephalomyelitis [EAE]) showed that laquinimod decreased inflammation, demyelination, and axonal injury [97]. It appears to reduce infiltration of CD4+ T cells and macrophages into the CNS; it also seems to increase the serum level of brainderived neurotrophic factor which may protect against neuronal injury [98]. Laquinimod efficacy in RRMS patients has been assessed in two phase III trials, Assessment of Oral Laquinimod in Preventing Progression in Multiple Sclerosis (ALLEGRO) [99] and Benefit-Risk Assessment of Avonex and Laquinimod (BRAVO) [100]. In ALLEGRO, laquinimod was compared to placebo and showed a modest but significant reduction of the annual RR and also a reduced rate of progression [99]. Initially, the primary outcome measure of reducing annual RR in BRAVO was not reached [100], but following an adjustment for an imbalance between the groups in the number of patients with enhancing lesions and of mean T2 lesion volume, both predictors of relapses, a reduction in RR in the laquinimod arm (21% reduction vs. placebo, p = .026) was obtained. Overall, laquinimod had more pronounced effects on disability progression and brain atrophy than on relapses and new MRI lesion formation [101]. Two ongoing studies are evaluating the safety and efficacy of laquinimod: CONCERTO (phase III, RRMS; NCT01707992) and ARPEGGIO (phase II, PPMS; NCT02284568). Both aimed to compare two doses of laquinimod, 0.6 and 1.5 mg/day, to placebo; however, in January 2016, TEVA announced the discontinuation of higher doses of laquinimod after the occurrence of cardiovascular events (none fatal) in eight subjects [102]. The study of lower-dose laquinimod will continue in both trials.

Siponimod and ozanimod
Siponimod (BAF312) is an oral modulator of sphingosine more selective than fingolimodacts selectively on S1P-1 and S1P-5 [103]. In a phase II trial in RRMS, siponimod was shown to reduce brain MRI lesions and relapses by up to 80% compared to placebo [104]. The results of a phase III trial in SPMS (NCT01665144) were reported at the 32nd ECTRIMS (European committee for the treatment and research in multiple sclerosis) Congress in London in September 2016 by Kappos et al. Patients with SPMS on siponimod had the risk of 3-month CDP reduced by 21% compared with patients in the placebo arm (p = 0.013). Another oral selective S1P receptor modulator, ozanimod, was recently successfully tested in a phase II trial with positive MRI outcomes [24].

Agents in trial for progressive MS
Finding effective treatments for progressive MS is a major priority and a challenge. The current treatment candidates and approaches for progressive MS have been recently reviewed in detail by Shirani et al. [105]. An important point in testing drugs for progressive forms of MS is to adapt trial methodology and clinically meaningful outcomes to this particular phenotype of MS based on lessons learned from prior clinical trials [106].
It has been hypothesized that high-dose biotin can act on demyelinating axons by increasing energy production promoting synthesis of myelin [107]. Treatment with high-dose biotin was tested in an open-label pilot study [108] and a randomized, double-blind, placebo-controlled trial in progressive MS (23 and 154 patients, respectively) [109]. Both studies reported positive results on disability progression and a good safety profile [107]. Biotin can interfere with some laboratory tests (e.g. false-positive or false-negative thyroid tests [110,111]), and teratogenicity was reported in rabbits. Further studies are needed to clarify the duration of response to biotin over time and the responder profile in people with progressive MS.

Autologous bone marrow transplantation
Hematopoietic stem cell transplantation (HSCT) has been reviewed in detail elsewhere [112][113][114][115][116]. Studies on animal models showed that strong immunosuppression followed by syngeneic bone marrow transplantation can induce long-term antigen-specific tolerance [117]. In the last 15 years, due to the high rate of complications related to the procedure, autologous bone marrow transplantation was reserved only to MS patients that failed all other therapies and had a poor prognosis [116]. There is evidence that high-dose immune ablation and autologous HSCT could renew the immune system repertoire and reinforce immune tolerance mechanisms [118], thus having a clinical impact. Phase I clinical trials have shown that autologous HSCT may improve the quality of life of MS patients [112,116]. A recent phase II trial of HSCT vs. mitoxantrone in RRMS and SPMS showed that HSCT reduced the number of new T2 and enhancing lesions and the AAR compared to mitoxantrone [119]. In a very recent phase II trial, 70% of the patients who received an aggressive immuneablative treatment followed by an HSCT graft depleted of autoreactive lymphocytes did not have any signs of disease activity (relapses, new MRI lesions, or EDSS progression) after a median follow-up of 6 to 7 years [120]. Although these results are promising, and progress has been made over the last decade in mitigating risks, there are many unknowns regarding the use of HSCT as a possible second-line therapy for refractory MS [121,122]. There is lack of consensus on the optimal conditioning regimen, patient selection (including the stage of the disease or what other prior treatments should they have failed to), and the HSCT graft manipulation. Different conditioning regimens following the harvesting of the stem cells were used in the studies of autologous HSCT in people with MS [123]. The choice of the conditioning regimens (myeloablative or non-myeloablative) can impact the outcomes in terms of both efficacy and toxicity [113]. The myeloablative conditioning regimens can completely eliminate the activated immune cells before the HSCT treatment but would expose the patients to aplasia-related complications and death [113]. The use of lymphoablative but non-myeloablative conditioning regimens would be in keeping with 'the rationale of auto-HSCT [. . .] to revive an antigen-naive immune system from the patient's HSCs' [113] while mitigating adverse effects such as neurotoxicity [124]. Freedman and Atkins recently noted that, although systematic comparison of regimens is lacking, the available data would suggest that the more intense conditioning regimens are followed by more durable responses but with more toxicity as well [121]. In their recently published study, Atkins et al. removed the mature lymphocytes from the graft prior to transplantation by an ex vivo cell selection technology, avoiding graft-mediated immune effects, and thus apparently impacting disease activity [120]. However, the benefits of this procedure would only be seen with intense conditioning regimens which achieve a near-complete immune ablation [121]. Intense immune ablation with a regimen including a cytotoxic agent crossing the BBB would translate in an accelerated whole-brain atrophy rate although the rate of atrophy would further slow to that expected from normal aging [125]. In a recent editorial, Sorensen suggests that, taking into account the HSCT benefit-risk profile and the availability of highly effective treatments with monoclonal antibodies which can achieve disease control in patients with active disease, intense immunosuppression with HSCT should remain a third-line therapy [122]. However, as Mowry notes in an editorial in the same journal [123], the comparison in terms of efficacy between the phase III studies with monoclonal antibodies and the observational or single-arm HSCT studies does not provide evidence for or against the use of HSCT in MS, and further studies designed to comparing these treatments are warranted [123]. It is likely that the place of HSCT in MS will be reevaluated over the next years, in light of the continuously growing spectrum of available therapies and of new pragmatic, prospective, controlled multicenter trials.

Remyelination strategies
Remyelination occurs initially in MS lesions but is inadequate, and the mechanism of repair in the CNS fails with time, especially in chronic disease stages [126]. The differentiation of oligodendrocyte precursor cells (OPCs) into mature cells is essential [126]. Remyelination develops in two steps: the colonization of the lesions by the OPCs and the OPC differentiation into mature oligodendrocytes able to generate functional myelin sheath [127]. The factors that could interfere with the OPC abilities to remyelinating in MS have been reviewed [128][129][130]. The immunological-OPC cross talk has specific features in MS [131,132]. Age and disease duration matter. In older animals, remyelination is significantly slowed, possibly because of the decreased response of monocytoid immune cells which are necessary for the clearance of the myelin debris that inhibit remyelination [133]. Neurodegeneration itself occurs early in the course of the disease, and repeated episodes of demyelination could conceivably lead to local wearing out of myelin-forming OPCs [134]. Most information on these processes comes from studies on animal models (EAE and the cuprizone model of MS) [135]. Remyelination can be promoted either by intrinsic (altering intrinsic signaling pathways) or by extrinsic (acting on lesion environment) repair mechanisms [136,137].

Intrinsic targets for remyelination
A way of promoting remyelination is modulation of specific signaling pathways such as Notch, the Wnt/β-catenin pathway, and the retinoid X-receptor (RXR) signaling pathway [136] within oligodendrocytes to outweigh the inhibition of remyelination. Tocopherol derivative TFA-12, a synthetic long-chain fatty alcohol and a member of the vitamin E family, with anti-inflammatory properties [138], stimulates OPC differentiation and myelin repair in experimental models of MS through the inhibition of the Notch/Jagged1 intrinsic signaling pathway [138]. Lithium chloride stimulates myelin gene expression in oligodendrocytes via Wnt/β-catenin and Akt/CREB pathways [139]. Indomethacin, a nonsteroidal anti-inflammatory drug that penetrates the BBB, promotes the differentiation of OPC into mature cells, hence stimulating remyelination in animal models via modulation of Wnt/β-catenin pathway [140]. In vitro studies showed that the action of indomethacin relies on the Glycogen synthase kinase 3 beta (GSK3β) activity [140]. The nuclear RXR-γ regulates positively the endogenous remyelination, by stimulating OPC differentiation [141]. RXR-γ binds to receptors inside the OPC, including the vitamin D receptor, and the complex RXR-vitamin D receptor enhances OPC differentiation [142]. Hence, vitamin D might have a role in remyelination. Finally, miconazole and clobetasol have been recently shown to enhance the generation of human oligodendrocytes from human OPC in vitro, possibly through mitogen-activated protein kinase and glucocorticoid receptor signaling, respectively [143].

Extrinsic targets for remyelination
Leucine-rich repeat and immunoglobulin domain-containing 1 (LINGO-1) is a nervous-system-specific transmembrane protein that may be a therapeutic target for remyelination. LINGO-1 is expressed by oligodendrocytes and was shown to inhibit their ability to differentiate and myelinate; it is also expressed on axons where it limits axonal regeneration [144,145]. Antibodies blocking LINGO-1 promote OPC differentiation in demyelinated lesions, reduce axonal damage, and restore function in EAE, cuprizone, and lysolecithin animal models [146,147]. BIIB033 (Li 81; opicinumab) is a fully human monoclonal antibody that binds LINGO-1. BIIB033 enhances remyelination and restores function in animal models although brain concentrations were of <0.5% of those in blood [148]. Anti-LINGO-1 antibodies have been the first remyelinating therapy evaluated in humans. Two randomized placebo-controlled phase I trials tested the safety, tolerability, and pharmacokinetics of BIIB033 administered via IV infusion or subcutaneous injection in healthy volunteers and people with MS [149]. In these studies, one or two doses up to 100 mg/kg were tolerated, with no serious adverse events and low immunogenicity [149]. RENEW (A Randomized, Double-Blind, Parallel-Group, Placebo Controlled Study to Assess the Efficacy, Safety, Tolerability, and Pharmacokinetics of BIIB033 in Subjects With First Episode of Acute Optic Neuritis) was a randomized, double-blind, placebo-controlled phase II trial evaluating the ability of anti-LINGO-1 "to promote the repair of an acute optic nerve lesion after a first episode of acute optic neuritis and examining the effects on remyelination over 24 weeks through the measurement of the latency of nerve conduction between the retina and the visual cortex" [150]. The treatment with anti-LINGO-1 in acute optic neuritis prevented the amplitude loss of the multifocal visual-evoked potential observed in the fellow eye visual pathway of placebo-treated subjects over 32 weeks and had some possible positive effects on amplitude preservation on the affected eye [150].
Another phase II trial SYNERGY (Study to Assess the Efficacy, Safety, Tolerability, and Pharmacokinetics of BIIB033 in Participants With Relapsing Forms of Multiple Sclerosis When Used Concurrently With Avonex) tested the safety and effectiveness of opicinumab in association with IFNβ-1a injections once weekly in people with RRMS or SPMS (NCT01864148). The trial has recently completed, and the results were presented at the 32nd ECTRIMS Congress in London in September 2016 by Cadavid et al. The trial involved 416 participants receiving IFNβ-1a weekly and one of five different doses of anti-LINGO-1 per kilogram body weight or a placebo, once every 4 weeks for 72 weeks. The trial missed its primary end point (the percentage of subjects with confirmed improvement of neuro-physical and/or cognitive function and/or disability).
Wang et al. differentiated OPCs from human-induced pluripotent stem cells and engrafted them in a myelin-deficient mouse model [151]. The transplanted OPCs differentiated into astrocytes and oligodendrocytes, myelinated the brains of the animals, and increased their survival [151]. However, since MS is a multifocal disease, it would probably require repeated transplantation of the OPC in all the demyelinated regions [134]. These techniques are still under study, and their safety and efficacy are yet unknown. OPC recruitment is reduced in MS lesions, and chemotactic molecules such as Sema3A receptor neuropilin-1 could be a new group of drug targets to improve remyelination [152]. It is not yet clear whether remyelination completely prevents neurodegeneration, but it does appear to restore neuronal function and at least limit neuronal degeneration [134]; therefore, remyelination strategies are likely to be part of the MS treatments in the future (Table 1). Clinical studies suggest neuroprotection as measured by brain atrophy during treatment compared with pretreatment.

Mesenchymal stem cells
Mesenchymal stem cells (MSCs) can be harvested from adult bone marrow and can be transplanted securely without the need for immunosuppression and with a low risk of aberrant proliferation. The IV route is preferred as it is less invasive and has fewer adverse effects compared to intrathecal administration. It is likely that the potential therapeutic efficacy of MSC could be based on systemic effects as recently shown [164]. Mechanisms of action would include immunoregulation and anti-inflammatory changes of the cellular environment. The phase II trials of MSC in MS did not report major side effects related to treatment [165]. A phase I/II, open-safety, clinical trial in patients with MS and with amyotrophic lateral sclerosis showed that intrathecal and IV administration of autologous MSCs is a clinically feasible and a relatively safe procedure which produced immediate immunomodulatory effects [166].

T-cell-directed strategies
It was suggested that the immune response in MS is directed at least in part against myelin proteins including MBP, myelin oligodendrocyte protein (MOG), and proteolipid protein (PLP) [167][168][169][170]. Although there are differences in the activation state or precursor frequencies of T cells from patients with MS and healthy subjects, people without MS also have immune responses against such antigens. However, it was suggested based on indirect evidence that molecular mimicry, epitope spreading, and bystander activation are possible mechanisms to initiate and maintain disease activity. Consequently, an alternative treatment approach in MS could aim to selectively restore self-tolerance to autoantigens via immunization [171] and epitopespecific induction of T-cell tolerization [172] or specifically by targeting regulatory T cell signaling [173]. Antigen-specific therapies in MS have been reviewed [174,175]. A first phase I trial in humans published in 2013 showed that the antigen-coupled cell tolerization in MS is feasible and safe [176]. A number of animal studies and human clinical trials of T-cell vaccination (TCV) in MS have been conducted and reviewed [177,178]. TCV includes an attenuation step to preventing the encephalitogenicity of myelin-reactive T cells and is safe. Two phase II, placebo-controlled clinical trials of TCV in MS were reported [178,179]. Fox et al. showed that treatment with myelin-reactive T cells against up to six peptides of MBP, MOG, and PLP was safe in 100 RRMS and CIS patients, of which 44% were previously treated with DMTs [179]. Although no statistically significant clinical or radiological benefits were obtained when compared with the placebo arm, some post hoc evidence of clinical efficacy in the active arm in the subgroup of patients DMT-naive was seen [179]. Karussis et al. performed the first placebo-controlled, double-blind TCV trial in progressive MS [180]. Twenty-six patients with relapsingprogressive MS were included in the trial, of which 19 were treated with four injections with a mixture of attenuated autologous T-cell lines reactive to three to nine different peptides from the MBP, PLP, and MOG sequences [180]. The TCV was safe, without serious adverse events. There was a strong clinical effect on relapses and disability measured by EDSS and the timed 10-m walking test. The authors suggested that the use of multiple vaccinations, peptides, and several anti-myelin cell lines accounted for the positive clinical effects obtained in this trial [180]. The ongoing or completed clinical trials using T-cell-directed approaches are listed in Table 2.

Drug repurposing in MS
The failure to deliver successful neuroprotective therapies in MS has led to alternative strategies such as drug repurposing [181]. Vesterinen et al. list some of the factors contributing to this failure: 'the incomplete understanding of disease biology, pathogenic complexity and heterogeneity, limited predictive value of animal models, [. . .] a lack of established trial methodologies compounded by a wider context of chronically declining productivity in drug development based on target-based approaches, declining resources for drug development in the neurosciences, and the growing costs of clinical trials' [181]. In a phase II trial, high-dose simvastatin compared with placebo reduced the annualized rate of whole-brain atrophy [182]. Encouraging recent data suggest that phenytoin in acute optic neuritis [183] and amiloride in PPMS [153] could have a neuroprotective effect. We and others showed that repurposing molecules such as angiotensin-converting enzyme inhibitors or retinoids can be effective in the animal model of MS or promote remyelination [141,[184][185][186]. It is probable that repurposing drug development will play a bigger role in the future despite the inherent pitfalls which impact its feasibility [187].

Expert commentary
The best treatment strategy in MS is still debated. Escalation strategy implies starting with less effective but safer treatments and escalating to more efficacious but riskier DMTs if the inflammation is not controlled [188,189]. This approach would select only patients with a more active MS to be commenced on riskier drugs. The downside of this approach is that often in practice some patients have already accumulated some degree of disability at the time of escalation, which may be irreversible [190]. This disadvantage is taken into account by those advocating starting treatment with a more potent drug from the outset (induction strategy). This approach applied indiscriminately would, in turn, expose people with less active MS to unwanted drug side effects. The treatment selection should take into account how aggressive and active the inflammation is; medications such as natalizumab and alemtuzumab would be preferred in people with active MS. However, alemtuzumab was registered by the EMA for patients 'with active disease defined by clinical or imaging features' [29], and this allows treatment in people with early disease and moderate RR if active disease is demonstrated (defined as at least two relapses within the prior 2 years).
A new strategy aims to attain NEDA. NEDA emerged as therapeutic expectations and targets evolved to enclose possible remission from evolving disease. NEDA is defined as absence of relapses, progression of disability, and MRI activity such as new or newly enlarging T2 lesions and new enhancing lesions [191]. Achieving the NEDA status at 2 years has a good prognostic value over a 7-year period [192]. NEDA is criticized of not reflecting patients' needs in clinical practice [193]. The absence of brain atrophy measured by MRI as a marker of neuroaxonal loss was proposed as an additional criterion for NEDA (NEDA-4) [194]. The effects of MS treatments on disability progression relate with the effects on atrophy over 2 years, independently from the effect on active MRI lesions [195]. This would support the use of brain atrophy as a surrogate marker of disability progression [196]. However, routinely assessing atrophy is difficult due to logistical and technical restraints (e.g. variability between scans and scanners). As Ziemssen et al. note, 'the increasing focus on NEDA as an aim of MS therapy implies that regular, systematic, monitoring should be a central aspect of the management of the condition' [189]. This raises the issue of disease monitoring: how often and what measures to use. How each of the detected MRI changes would entail a specific and timely change in the treatment strategy is still a matter of debate. Moreover, it is not clear yet if an escalation approach aiming of achieving NEDA would certainly improve disease outcome over the long term [190].
Possibly, additional outcome measures (i.e. cognitive impairment) and MS patient-reported measures may prove useful to identify non-responders to DMT and therefore could be candidate measures to be incorporated in NEDA. Nevertheless, data from clinical trials back the idea of a benefit of early initiation of treatment in MS [197][198][199][200][201]. Early treatment would have an impact on long-term prognosis in terms of mortality [202]. Although it is intuitive that MRI measures can reflect breakthrough disease, there is a lack of clinical consensus on the quantitative aspects of these measures. Two new enhancing lesions were predictive of disability worsening in a 15-year follow-up study of patients enrolled in a pivotal IFNβ trial [203]. A recent systematic review of studies examining differential response to IFNβ showed that patients with two or more new hyperintense T2 lesions or new gadolinium-enhancing lesions had significantly increased risk of both future relapses and progression [204]. A study by Prosperini et al. suggested that MRI alone could be a 4-year good predictor of outcome after the first year of treatment with IFNβ-beta [205]. The Rio score was developed by the Barcelona group and included the number of clinical relapses, disability progression, and new T2 or enhancing MRI lesions [206]. Patients with RRMS who met at least two of these criteria (but not relapses or MRI criteria alone) had a higher risk of having relapses or disability progression in the next 2 years. A new version of the Rio score was validated in the original data set and classified patients in low-, medium-, and high-risk of progression based on relapses or MRI activity only [207,208]. While changes in clinical and MRI measures of disease activity can lead to a treatment change, the diseaserelated outcomes reported by the patient could too. It is probable that over the next years, more information from real-life experience regarding specific drug efficacy, longterm safety concerns, cost-benefit considerations, and patient-perceived therapy burden will be available.

Five-year view
In a recent review, Coles [190] describes the current era of treatments of MS as 'the era of complexity.' This followed the 'era of nihilism' when no DMTs were available and the 'era of modest efficacy' which started after 1993 with the IFNβ and GA trials (which were proven modestly effective on RR but safer than immunosuppressants such as mitoxantrone and cyclophosphamide) [190]. The current 'era of treatment' started with the introduction of natalizumab and the first reports of PML as the main drug side effect which mitigated the enthusiasm for this medication [190]. Over the next decade, the oral drugs fingolimod, DMF, and teriflunomide and the injectable alemtuzumab have become available in many countries. As mentioned above, it is likely that daclizumab and ocrelizumab will soon follow. The development of the newer therapies for MS raises new challenges to the treating neurologist. The questions the doctors are facing regard the correct sequence of drugs to be offered to a given patient and the appropriate and safe sequence of switching between drugs. Since all the available DMTs are directed against the inflammatory phase of MS, the major unmet need at this time is for treatments for progressive disease. It is likely that in the next and fourth era of DMTs in MS, the efforts of tackling progressive disease will become successful [190]. It is also likely that more evidence would become available to support the different strategies for the sequencing and timing of treatments. Clarifying the monitoring strategies for adverse effects and breakthrough MS would be crucial. The development of radiological and biological biomarkers could allow in the future personalizing treatment and precision medicine in MS [209]. Features such as detection of central veins in the MS lesions could improve diagnostic quality [210]. Biomarkers such as neurofilament light chains [211] and chitinase 3-like proteins in the CSF [212] could be used as potential prognostic markers after CIS, while markers such as CSF lipid-specific immunoglobulin M bands [213] and L-selectin [214] could be incorporated in the PML risk stratification in patients under natalizumab [213]. Over the past 50 years, the largest impact on increasing life expectancy in the general population came from primary prevention. MS is the result of the interaction between the genetic background and environmental factors. Among the latter, vitamin D deficiency, infection with the Epstein-Barr virus (EBV), cigarette smoking, obesity in youth, and lack of exposure to intestinal parasites are environmental predictors of MS risk [215]. Considering that research into preventing MS currently has a high priority, developments in prevention strategies are anticipated. Studies of anti-EBV vaccination in people at risk of developing MS, clinical trials of smoking cessation in MS, and more clinical trials with vitamin D are expected. The prevalence of MS is inversely correlated with helminth infections, and studies of helminth treatments in MS could help in understanding the mechanisms of parasiteinduced immunomodulation [216].
Treating the right patient with the right drug early in the disease course, before disability has been acquired, could yield long-term benefits. It is realistic to expect over the next decade more data from approaches aiming to prevent MS and on methodologies to protect neurons or promote remyelination. Finally, the future of MS treatments would rely on a better understanding of the immunopathogenesis of MS.

Key issues
• MS is an immune condition of the CNS characterized by focal damage (inflammation) which manifests clinically mainly as relapses and diffuse damage (neurodegeneration) and brain volume loss) which are both responsible for disability. • There is no cure for MS. A number of disease modifying treatments (DMTs) have been approved which are primarily directed against inflammation. There are a number of additional DMTs in trial or being submitted for approval. • There are different treatment strategies in MS. In patients with active disease despite treatment it has been recommended to switch early to a therapy of higher efficacy. • Despite important advances in the treatment of MS, the burden of progressive disability and premature mortality remains considerable. There is no currently approved treatment for progressive MS, but recently a monoclonal antibody (ocrelizumab) was reported to be effective in a primary progressive MS trial. • New treatment strategies involving remyelination or neuroprotection are under study. It is likely that the next decade brings substantial changes in the understanding and ways of approaching the treatment in MS.

Funding
This paper was not funded.

Declaration of interest
C. Coclitu has received educational support from Biogen. R. Tanasescu has received research support, support for travel to scientific meetings, or consultancy honoraria from Biogen, Genzyme, Novartis, Abvie and Teva. C. S. Constantinescu has received research support, support for travel to scientific meetings, or consultancy honoraria from Bayer, Biogen, Genzyme, GW, Merck, Morphosys, Novartis, Roche, Sanofi-Pasteur MSD, and Teva. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.