The sequence of disease-modifying therapies in relapsing multiple sclerosis: safety and immunologic considerations

The mechanism(s) of action of the beta interferons are not yet fully established, although these therapies have been available since the 1990s. All beta interferons are known to exert autocrine and paracrine actions via activation of the interferon receptor on leukocytes (Table 2). Production of proinflammatory cytokines is reduced, and production of anti-inflammatory cytokines is induced [18]. Attachment of a polyethylene glycol side chain to the parent interferon beta-1a molecule yields peginterferon beta-1a, which, when administered subcutaneously, has a longer half-life, higher systemic exposure, and lower immunogenicity potential than intramuscular interferon beta-1a [57]. Glatiramer acetate, a synthetic polymer of four amino acids (l-glutamate, l-lysine, l-alanine, and l-tyrosine), is a mimetic for the MS autoantigen myelin basic protein (MBP), and thus competes with MBP antigens for binding to class II major histocompatibility complex (MHC II) [58]. With formation of the MBP antigen/MHC II complex impeded by glatiramer, helper T cells have less opportunity for activation and potential to destroy myelin [58]. In addition, glatiramer binding to MHC II inhibits the interaction of MHC II with CD4⁺ molecules located on the surface of helper T cells (Th1 and Th2). Consequently, there is reduced production of proinflammatory cytokines (interferon gamma) by Th1 cells, increased production of anti-inflammatory cytokines (interleukin 10) by Th2 cells (promoting a less inflammatory state), and induction of antigen-specific expansion of FOXP3⁺ regulatory T cells [58, 59].

The effects of beta interferons and glatiramer acetate on the immune system likely only endure for as long as a patient is exposed to therapeutic drug concentrations (i.e., less than five times the elimination half-life; Table 2). The safety profiles of the beta interferons and glatiramer acetate are well established (Table 5).

Following oral administration, dimethyl fumarate is rapidly metabolized to its active metabolite monomethyl fumarate, which is primarily responsible for its efficacy in MS [60, 61]. Similar to other fumarate derivatives, administration of dimethyl fumarate activates nuclear factor-erythroid 2-related factor 2, resulting in differential effects involving antioxidant responses (Table 2) [60]. It is likely that dimethyl fumarate has several additional immunomodulatory effects underpinning its efficacy in MS, including directing the immune response away from Th1 [62].

Integrated analyses of patient-level data (N = 2470) from Phase IIb, Phase III, and long-term extension studies of dimethyl fumarate showed that mean absolute lymphocyte count decreased by 30%, but generally remained above the lower limit of normal during the first year of treatment, before stabilizing [63]. Separate observational data indicated that the dynamics of the absolute lymphocyte count generally correlate with CD4⁺ and CD8⁺ counts [64], with the reduction of CD8⁺ T cells greater than that of CD4⁺ T cells (−55 vs. −39%) reflected in a 36% increase in the CD4/CD8 ratio [65]. In patients without severe lymphopenia (i.e., <0.5 × 10⁹ cells/L), there is evidence of improvement in lymphocyte counts following discontinuation of dimethyl fumarate, but full restoration takes >4 weeks [63, 66, 67]. For patients who become severely lymphopenic on dimethyl fumarate, lymphocyte counts may take a long time to recover [66]; this delay in lymphocyte recovery may complicate the switch to a subsequent DMT with myelosuppressive effects. Six percent of patients experienced lymphocyte counts <0.5 × 10⁹/L (grade ≥3 lymphopenia) in placebo-controlled trials [67]. The risk of developing moderate to severe lymphopenia while on dimethyl fumarate may be increased by the following: increasing age, lower baseline absolute lymphocyte count, and recent natalizumab exposure (there is a greater percentage reduction in absolute lymphocyte count due to the lymphocytosis induced by prior natalizumab) [64, 66]. Although the risk may be slightly different in patients of older age or with lower baseline absolute lymphocyte count, all patients remain at a small risk of lymphopenia. Based on 7250 cumulative patient-years of exposure, the overall incidence of serious infections was low, and there was no apparent correlation between the incidence of infection and grade of lymphopenia [63].

Of >230,000 patients treated with dimethyl fumarate globally in the 3 years following commercial availability (representing >330,000 patient-years), there have been five cases of PML in the setting of moderate to severe prolonged lymphopenia (absolute lymphocyte count <0.8 × 10⁹/L), a rare opportunistic brain infection caused by John Cunningham (JC) virus that is normally harmless in immunocompetent hosts [25, 68]. There are established risk stratification and mitigation strategies for patients on dimethyl fumarate in light of its safety profile. In the US, complete blood counts at baseline and every 6 months thereafter are mandatory to identify patients who may have developed severe prolonged lymphopenia [67]. In Europe, a baseline MRI also should be performed within 3 months of initiating therapy [69].

Teriflunomide is the active metabolite of leflunomide, which has anti-inflammatory, antiproliferative, and immunosuppressive properties [70]. Teriflunomide selectively and reversibly inhibits dihydroorotate dehydrogenase, a key mitochondrial enzyme in the de novo pyrimidine synthesis pathway, resulting in lymphocyte cell cycle impairment, without causing cell death [58]. Teriflunomide reduces neutrophils (during the first 6 weeks) and lymphocytes (during the first 3 months) by ~15%, although mean counts remain in the normal range [71]. Teriflunomide also inhibits protein tyrosine kinases, leading to decreased T cell proliferation, and a shift in the cytokine profile to a more anti-inflammatory cytokine milieu [58].

The long elimination half-life (18–19 days) of teriflunomide [72], due in part to its extensive enterohepatic recirculation, means that it can take approximately 8 months for the body to eliminate the drug, although plasma drug concentrations can still be detected up to 2 years after administration of the last dose [72]. If a rapid switch to another DMT is required, then an accelerated elimination procedure with either administration of cholestyramine or activated charcoal for 11 days should be considered [72], including in patients with cytopenia. Both elimination regimens result in a 98% decrease in plasma teriflunomide concentrations (i.e., <0.02 mg/L) [72, 73].

No new safety signals beyond those detected in individual trials (and summarized in Table 5) were identified with treatment duration exceeding 12 years and a cumulative exposure to teriflunomide exceeding 6800 patient-years [71]. Nonfebrile neutropenia and lymphopenia were reported in 5.9% and 0.5% of patients receiving teriflunomide over 1500 patient-years of cumulative treatment exposure, respectively [71]; no association between neutrophil count decrease and infection occurrence was detected [71]. Cases of thrombocytopenia, including rare cases with platelet counts <50,000/mm³, have been reported in the postmarketing setting [74]. Serious skin reactions, including severe generalized major skin AEs [Stevens–Johnson syndrome or toxic epidermal necrolysis (Lyell’s syndrome)] have been reported [74]. Teriflunomide is causally linked to one fatal case of toxic epidermal necrolysis [75], but has not been linked with PML. Complete blood count, tuberculin skin tests (to identify latent tuberculosis infection), liver function tests, and blood pressure measurements are required at baseline; liver function and blood pressure also should be monitored monthly for the first 6 months and then regularly thereafter with continued treatment [72].

Fingolimod is phosphorylated following oral administration to fingolimod phosphate, a mimetic for the naturally occurring extracellular lipid sphingosine-1-phosphate (S1P) [76]. S1P is an extracellular signaling molecule that regulates trafficking of many types of T and B cells from the lymph nodes to the blood (Table 2) [58]. Blood T cell levels decrease when S1P receptors are activated, as naive T cells are sequestered within secondary lymphoid organs after their egress from peripheral blood [58]. Lymphopenia is an expected pharmacodynamic effect of fingolimod due to the increased movement of CCR7⁺ lymphocytes into secondary lymphoid organs [77]. Fingolimod induces a rapid and reversible dose-dependent reduction in peripheral lymphocyte count to 20–30% of baseline values [13]. Peripheral lymphocyte reconstitution following fingolimod discontinuation occurs over 1–2 months [13], but this period may be extended with fingolimod use exceeding 1 year [78]. Rarely, an abrupt rise in lymphocyte count occurs during the fingolimod postwithdrawal period (immune reconstitution inflammatory syndrome), putting patients at risk for MS disease reactivation [79].

Cases of PML have occurred in patients with MS receiving fingolimod in the postmarketing setting who had not been previously treated with natalizumab nor who were taking immunosuppressive or immunomodulatory medications concomitantly [13]. In addition, affected patients had no other ongoing identified systemic medical conditions resulting in compromised immune system function [13]. For this reason, fingolimod should be withheld at the first sign or symptom suggestive of PML [13]. In addition, patients with signs and symptoms consistent with other opportunistic pathogens, including herpes simplex viruses 1 and 2, varicella zoster virus, cryptococci, and atypical mycobacteria, should undergo prompt diagnostic evaluation and appropriate treatment [13, 80].

Fingolimod phosphate acts on four of the five known S1P receptor subtypes expressed on a variety of cell types, including endothelial cells, lymphocytes, smooth muscle and cardiac myocytes, and neural cells [76]. Consequently, fingolimod administration elicits significant off-target pharmacology, resulting in rare but serious AEs (Table 5) [16]. For these reasons, in Europe, fingolimod is considered a second-line DMT following failure of interferon beta or glatiramer acetate, or a first-line agent for patients with highly active disease [12]. No such restrictions are placed on its use in the US, although numerous changes have been made to the warnings and precautions listed in the fingolimod product label to guide proper use (Table 5) [80]. Hence, patient selection and monitoring is of paramount importance to increase the likelihood that the benefits of fingolimod outweigh its risks. Baseline laboratory tests include complete blood count, liver enzymes, pregnancy test, and varicella zoster virus status [12, 13]. Tests that are required before dosing, during, and/or posttreatment with fingolimod include cardiac and blood pressure monitoring, complete blood counts, and examination of the fundus for macular edema [13]. Monitoring for signs of infection and suspicious skin lesions in case of basal cell carcinoma also should be conducted, as per the fingolimod product label [13]. Case reports of severe disease reactivation following fingolimod withdrawal [46‐48] also must be considered when treating patients with highly active disease and in women who wish to become pregnant (in whom it is advised to continue therapy only if the potential benefit justifies the potential risk to the fetus) [12].

Daclizumab beta is a humanized monoclonal antibody approved for the treatment of RMS as a monthly subcutaneous self-injectable [81‐83]. Daclizumab beta works in the periphery by binding CD25 (alpha subunit of the high-affinity interleukin 2 receptor mostly expressed on activated T cells) to modulate interleukin 2 signaling (Table 2) [84‐87]. Blockade of CD25 by daclizumab beta limits interleukin 2 consumption by activated T cells and facilitates cells that express the intermediate-affinity interleukin 2 receptor [i.e., natural killer (NK) cells and precursors of innate lymphoid cells] to receive more interleukin 2 signal, as this receptor does not feature CD25 [88, 89]. Consequently, there is a substantial expansion of immunoregulatory CD56^bright NK cells that penetrate the blood–brain barrier and eliminate important mediators of MS immunopathology, activated T cells, leaving resting T cells intact (Table 2) [84, 86, 90, 91].

The pharmacodynamic effects of daclizumab beta are sustained in patients with relapsing–remitting MS, as evidenced by a fivefold expansion of CD56^bright NK cell levels that plateau by the end of the first year of treatment [92, 93]. Modest 10% reductions in circulating total lymphocytes, CD4⁺ and CD8⁺ T cells, and B cells were observed in patients with MS after 1 year of daclizumab beta 150 mg treatment, and regulatory T cell levels were reduced by approximately 50% after 8 weeks [92‐94]. The remaining regulatory T cells are functionally active, as evidenced by stable cytokine production, maintained active cell cycling, and retention of a regulatory T cell-specific demethylated region in the FOXP3 promoter, albeit with a significant decrease in CD25 expression [94]. The effects of daclizumab beta are reversible; after treatment cessation, total lymphocyte counts return to baseline levels within 12 weeks, and CD56^bright NK cell and regulatory T cell counts return to baseline levels within 24 weeks [93, 95]. There was no apparent impact of antidrug antibodies or neutralizing antibodies on the pharmacodynamics, efficacy, or safety of daclizumab beta [82, 96].

The safety profile of daclizumab beta was determined over a 5-year period and was consistent with that from the pivotal Phase III trial (Table 5) [97]. Given the recent approvals for daclizumab beta, there are only limited data from real-world use, although the product label provides specific warnings regarding hepatic injury—elevations of serum transaminases and serious events, including fatal cases of autoimmune hepatitis and liver failure—and other immune-mediated disorders (including skin reactions, lymphadenopathy, and noninfectious colitis) as well as depression, infections, autoimmune hemolytic anemia, gastrointestinal AEs, and lymphopenia [81, 82]. These warnings are based on safety and tolerability information from an integrated analysis of six clinical studies (primarily randomized controlled trials and their extensions) encompassing 2236 patients with 5214 patient-years of daclizumab beta exposure [98]. Although this database was not large enough to detect rare events, the analysis did show that daclizumab beta had an acceptable safety profile without evidence of cumulative toxicity [98]. Daclizumab beta should be discontinued in cases of significant transaminase elevation [i.e., alanine transaminase (ALT) or aspartate transaminase (AST) >5 × the upper limit of normal (ULN) only; or total bilirubin greater >2 × ULN; or ALT or AST ≥3 but <5 × ULN and total bilirubin >1.5 and <2 × ULN] [82]. Discontinuing daclizumab beta should be considered if severe depression or suicidal ideation occurs [81, 82]. If serious infection develops, daclizumab beta should be withheld until the episode resolves [81, 82].

The humanized monoclonal antibody alemtuzumab targets the extracellular glycoprotein CD52, resulting in antibody-dependent cytolysis and complement-mediated lysis of T and B lymphocytes, monocytes, NK cells, macrophages, and dendritic cells (Table 2) [99, 100]. Alemtuzumab elicits rapid, profound, and prolonged B and T cell lymphopenia followed by a reconstituted immune system different in composition from that before treatment, which may rationalize its long-term efficacy, given patients only receive two medication cycles that are 1 year apart (Table 3) [99, 100]. It can take ~8 months for B cells and up to 3 years for T cell subsets to recover to the lower limits of the normal range after a single course of alemtuzumab, and T cells may not recover fully to baseline values [101]. It is worth noting that B cell recovery was rapid in one study; levels of ‘mature naive’ B cells (CD19 and CD23 positive but CD27 negative) returned to baseline by 3 months and rose to 165% of baseline values by 12 months after the first course of alemtuzumab treatment [102]. Conversely, CD27-positive memory B cell recovery was slow, reaching only 25% of baseline levels by month 12 [102]. The immunosuppressive effects of alemtuzumab on CD4⁺ T cell subsets lasted for up to 4 years in 29 patients who participated in CARE-MS I and CARE-MS II [103]. Differential lymphocyte reconstitution after alemtuzumab treatment may be a biomarker for MS relapse, as patients with active disease showed an accelerated recovery of CD4⁺ cells (p = 0.001), with a difference in absolute CD4⁺ counts at 24 months (p = 0.009), while CD4⁺ counts <388 × 10⁶ cells/mL predicted MRI stability [104]. Anti-alemtuzumab antibodies reduce plasma alemtuzumab concentrations during course 2 but not course 1, although they do not appear to affect clinical outcomes, total lymphocyte count, or AEs [100]. Alemtuzumab induces long-term immunodepleting effects, which must be considered when planning subsequent therapies for maintenance treatment or if a patient does not respond adequately. There are no data on sequencing therapies after alemtuzumab use to guide the clinician, who must, therefore, rely on intensive patient monitoring to individualize care.

Although the advantages of long-lasting efficacy and extremely high patient adherence are positive attributes of alemtuzumab, Table 5 shows that this DMT is associated with several serious AEs that may arise years after starting treatment and are, therefore, not reflective of alemtuzumab’s pharmacokinetic profile (elimination half-life, 2 weeks) [100]. Alemtuzumab is usually reserved for patients with unfavorable prognostic indicators because it is difficult to reconcile its superior efficacy over interferon beta with exposure to serious AEs in patients with less severe disease. Even in patients with highly active disease, diligent patient selection and strict adherence to risk monitoring programs is required. The alemtuzumab product label recommends regular laboratory monitoring up to 4 years after the last alemtuzumab dose (and beyond if warranted) for the detection of secondary autoimmune conditions (e.g., immune thrombocytopenia, antiglomerular basement membrane disease, and thyroid disorders) [100]. Laboratory testing includes differential blood count, serum creatinine, and urine analysis before administration and monthly thereafter [100]. Pretreatment thyroid-stimulating hormone level is mandatory and requires rechecking every 3 months until 4 years after the last infusion [100]. Patients with active or uncontrolled infections are not candidates for therapy [100]. Prophylactic oral acyclovir should be taken until CD4⁺ count is >200 cell/mm³ to reduce the risk for herpes infections [100, 105].

Natalizumab is a humanized monoclonal antibody that binds to the integrin molecule very late activation antigen 4 (α4β1), a glycoprotein surface molecule found on all leukocytes except neutrophils (Table 2) [58]. Blockade of α4β1 prevents adhesion of leukocytes to vascular cell adhesion molecule 1, a protein expressed on the surface of vascular endothelial cells in the brain and spinal cord, and thus blocks entry of leukocytes into the central nervous system across the blood–brain barrier [58]. Natalizumab increases the number of circulating leukocytes (due to inhibition of transmigration out of the vascular space), but does not affect the absolute count of circulating neutrophils [20].

Natalizumab has an elimination half-life of 11 days [20], although plasma natalizumab concentrations can be reduced by 92% within 1 week of plasma exchange sessions to treat PML if required [106]. The reversibility of natalizumab’s pharmacologic effects on peripheral immune cells is evident starting at weeks 8–12, with levels returning to those observed or expected in non-natalizumab patients by 16–20 weeks after the last natalizumab dose [107]. This is consistent with the reduction in plasma natalizumab concentrations to below the limit of detection by 16 weeks postdose [107]. Lymphocyte counts remain within the normal range at all times both for patients receiving natalizumab and for those who have stopped natalizumab treatment [107]. Patients who develop anti-natalizumab antibodies are more likely to have hypersensitivity reactions during drug administration [20].

An intensive risk stratification program is in place to help prescribers weigh the clear efficacy benefits of natalizumab against the development of PML [19, 20]. Three main factors drive the risk of developing PML in patients undergoing natalizumab therapy: (1) therapy ≥24 months; (2) previous use of immunosuppressant treatment; and (3) JC virus antibody positivity [108]. The anti-JC virus antibody index value and duration of natalizumab treatment are two key factors that enable clinicians and JC virus-positive immunosuppressant-naive patients with MS to make informed treatment and monitoring decisions [109‐111].

Natalizumab withdrawal often leads to an MS relapse and return of inflammatory disease activity on MRI [49‐53]. Younger patients (<40 years of age) were 3.8-fold more likely to have increased MRI activity during 24 weeks of natalizumab treatment interruption, as were those with one to five Gd⁺ lesions (2.7-fold increase) and >five Gd⁺ lesions (6.2-fold) before natalizumab initiation (vs. no lesions) [52]. Initiating interferon beta within 30 days postnatalizumab dosing in patients who had been free of disease activity [112], or initiating fingolimod ~4 months postnatalizumab dosing in patients with stable Expanded Disability Status Scale scores [113], was associated with clinical and radiologic disease recurrence. A therapeutic gap of no more than 3 months between discontinuing natalizumab and initiating fingolimod appears to minimize the risk of relapse [113]. Switching from natalizumab to alemtuzumab (n = 16) [114] or off-label rituximab (n = 114 [115] and n = 118 [116]) may be a feasible option to maintain disease control, including in those at high risk of PML. It is currently unknown if the switch-to therapy selection impacts the risk of PML in this context. Data on treatment selection after a PML event on natalizumab are limited, but there are reports of successful use of both dimethyl fumarate and fingolimod in this situation [117].

Ocrelizumab is a B cell-directed cytolytic monoclonal antibody with a humanized immunoglobulin G1 tail indicated for the treatment of patients with relapsing or primary progressive forms of MS as a twice-yearly intravenous infusion [118]. This recombinant monoclonal antibody binds to a different but overlapping CD20 epitope expressed on B cells to rituximab [119]. By binding to the cell surface antigen CD20 present on pre-B and mature B lymphocytes, it is believed that ocrelizumab induces antibody-dependent cellular cytolysis, complement-mediated lysis, and/or apoptosis via crosslinking membrane CD20 on the target cell surface [118, 119].

Assays for CD19⁺ B cells are used as a surrogate for B cell counts because ocrelizumab interferes with the CD20 assay. As such, ocrelizumab reduces circulating CD19⁺ B cell counts 14 days postinfusion to negligible levels [118]. In clinical studies, B cell counts rose to above the lower limit of normal or above baseline counts between infusions of ocrelizumab at least once in 0.3–4.1% of patients [118]. Median (range) time for B cell counts to return to either baseline or the lower limit of normal was 72 (27–175) weeks after the last ocrelizumab infusion (Table 2) [118]. Within 2.5 years after the last infusion, B cell counts rose to either baseline or the lower limit of normal in 90% of patients [118].

Treatment-emergent ocrelizumab AEs observed in a pooled analysis of the two identical 96-week Phase III OPERA trials included infections, infusion-related reactions, and an incidence rate of first neoplasm of 0.40 per 100 patient-years of exposure, based on data from 6467 patient-years of exposure (Table 5) [120]. The ocrelizumab prescribing information states that breast cancer occurred in 6 of 781 females treated with ocrelizumab versus none of 668 females treated with subcutaneous interferon beta-1a or placebo [118]. Hence, patients receiving ocrelizumab should be encouraged to follow standard breast cancer screening guidelines. The long-term effects and risks of B cell depletion on malignancy risk will remain uncertain until long-term real-world follow-up data are available, and may currently be underrecognized. As with alemtuzumab, long-term B cell depletion with ocrelizumab may limit subsequent treatment options. For instance, initiation of a later-line therapy while B cell levels remain depleted may result in cumulative, and presently undocumented, effects on immune system function. The appropriate timing for initiating other DMTs after a patient has received ocrelizumab should be considered by the physician. The immunogenicity of ocrelizumab appears low, based on the incidence of formation of treatment-emergent antidrug antibodies (~0.4%) [120].