MAGNIFY-MS was a 2-year, phase IV, open-label, single-arm study in which eligible participants with highly active RMS received CladT (3.5 mg/kg cumulative dose over 2 years). Participants received 2 weeks of active treatment per course (week 1 and 5 of each year), with the start of the first week of treatment in year 1 being considered baseline (BL, Supplementary Figure 1 ). The primary endpoint of this study, along with design and eligibility criteria have previously been reported (14, 15). Of particular note, participants were excluded if they had a lymphocyte count outside the normal laboratory limits, or if they had previous exposure to alemtuzumab, fingolimod, mitoxantrone, natalizumab, or ocrelizumab. Analyses for the 2-year period are described herein.

Peripheral blood mononuclear cells (PBMCs) were collected at indicated time points. Panels of B, T, and NK cells were analyzed using surface cell markers as described previously (15) at BL, month (M)3, M6, M12, M15, M18 and M24. Absolute cell counts and median percentage change from BL were assessed for cell subtypes and immunoglobulins. For the purposes of immunophenotyping, CD19⁺ B cells were analyzed as part of the TBNK cell panel, other B cell subtypes were analyzed as part of the full B-cell panel (Covance Central Laboratory Services, Inc. 8211 SciCore Drive, Indianapolis, IN 46214-2985, USA). Bulk RNA sequencing of CD19⁺ B cells and CD3⁺ T cells of PBMCs was carried out for 11 participants selected at random from the blood biomarker sub-study at BL, M3, M12, M15, and M24. Additional details are provided in Supplementary Appendix 1 .

Serum levels of immunoglobulin (Ig)G and IgM were analyzed by nephelometric assay using reference ranges provided by Covance at BL, M3, M6, M12, and M24 as previously described (15). sNfL levels were analyzed at BL, M3, M6, M12 and M24 using Quanterix NF-light chain assay method on Quanterix Simoa HD-1 Analyzer (Labcorp/Monogram, 345 Oyster Point Blvd, South San Francisco, CA 94080, USA). Obtained sNfL values were analyzed using Wilcoxon signed rank tests comparing data at the respective visit versus BL. sNfL values were normalized using Z-scores to adjust for age and body mass index (BMI) based on a control population, using the method previously described by Benkert et al. (16) Z-score interpretation as per Benkert et al: ≤-1, reduced relative to age- and BMI-matched reference population; ~0, similar relative to age- and BMI-matched reference population, >1, elevated relative to age- and BMI-matched reference population.

For participants of the blood biomarker sub-study, cytokine-producing T and B cells were analyzed at the previously described time points, with blood collected at three additional time points (M1, M2, and M14). For analysis of cytokines, 1×10⁶ of PBMC purified from whole blood was used. We studied stimulated and non-stimulated B and T-lymphocytes for individual intracellular cytokines comparing absolute values and median change from BL of each B and T cell intracellular cytokine by visit with 95% confidence intervals (CI) of the mean per stimulation method. B cell intracellular cytokines were stimulated with CpG and PMA/Iono. T cell intracellular cytokines were stimulated with Anti-CD3/CD28 or Anti-CD3/CD28 and PMA/Iono.

Clinical outcomes were assessed in all enrolled patients. Annualized relapse rate (ARR) of qualifying relapses was estimated from a Poisson regression model, adjusted for age and Expanded Disability Status Scale (EDSS) at BL (≤3, >3). EDSS, T25FW, and 9HPT scores were recorded at BL and M3, M6, M12, M15, M18, and M24. Definitions for 6mCDP, and progression on T25FW or 9HPT are provided in Supplementary Appendix 1 . All patients were included in the assessment of EDSS progression. Patients with relapse-associated worsening were not excluded from the analysis. MRI scans were performed at screening, BL, and at M1, M2, M3, M6, M12, M15, M18, and M24 following the initiation of CladT. Initial exploratory correlation analysis included participants with both B-cell and MRI assessments up to M6 (MAGNIFY-MS study primary endpoint). Reported changes in combined unique active (CUA), T1 gadolinium-enhancing (Gd+), and active T2 lesions were correlated with B regulatory (B_reg) and B_mem cell counts to investigate associations between fast CladT onset of action and lymphocyte dynamics. For measuring MRI activity, changes in CUA, T1 Gd+, and new or enlarging T2 lesion counts were compared between BL and post-BL visits using a mixed-effects linear model for repeated measures. The correlation between change in immune cell subsets and MRI from BL to M6 was assessed with Cramer’s V method ( Supplementary Appendix 1 ). We also performed correlation analysis using Spearman correlation coefficient between BL and each following timepoint from M3 to M24 to guide interpretation of the results. We compared fold-changes/percent changes of B-lymphocytes with changes of CUA between BL and follow-up. Association of clinical outcomes and changes in lymphocyte ratio dynamics were carried out as exploratory. For this purpose, presence of clinical progression was compared to immune cells ratios. The full list of disability status categories and immune cell subset ratios are presented in Supplementary Figure 2 .

Analyses were performed on the full analysis set (FAS), which included all eligible participants (people with highly active relapsing MS aged ≥18 years with an Expanded Disability Status Score ≤5.0) treated with at least one dose of CladT. All analyses were performed without adjustment for multiplicity. The study protocol and statistical analysis plan for MAGNIFY-MS have been published on ClinicalTrials.gov (NCT03364036) (17).

Ethical approval for the MAGNIFY-MS study (NCT03364036) was obtained from independent ethics committees for each trial site, and the study was performed in line with the principles of the Declaration of Helsinki. All participants provided written informed consent before participation in the study. Participants who elected to participate in the blood biomarker sub-study signed additional informed consent form.

Any requests for data by qualified scientific and medical researchers for legitimate research purposes will be subject to Merck's Data Sharing Policy. All requests should be submitted in writing to Merck's data sharing portal https://www.merckgroup.com/en/research/our-approach-to-research-and-development/healthcare/clinical-trials/commitment-responsible-data-sharing.html. When Merck has a co-research, co-development, or co-marketing or co-promotion agreement, or when the product has been out-licensed, the responsibility for disclosure might be dependent on the agreement between parties. Under these circumstances, Merck will endeavor to gain agreement to share data in response to requests.

Quantitative immune cell changes were observed following each yearly treatment course, during study visits outlined in Supplementary Figure 1 . All values presented are median percentage changes from BL.

B cell subtypes: In the FAS, changes observed in B cell counts after the first CladT treatment were similar to those previously reported for the sub-study (15) ( Figure 1A ; Table 2 ). At the first planned post-BL visit (M3) most B cell subsets reached nadir values: B_naïve (-76%), CD19⁺ (-80%), CD20⁺ (-81%) and activated CD69⁺ (-74%). B_mem, CD38⁺ plasmablasts and short-lived plasma cells reached nadir values only after the second treatment course in year 2 (M15). B_mem cell counts drastically reduced to -93% at M3 and to -97% at M15, this reduction was sustained throughout the observation period (M24, -89%).

At M3, we only saw repopulation for B_reg cell counts (+111%), which remained elevated at M6 (+93%) and M15 (+92%). At M24, B_reg cell counts repopulated to around BL level (+1.6%). B_trans cell counts at M3 were (-4.1%), differing from the previous report. However, these cells recovered to above BL by M6 (+15%) and had the highest increase after the second treatment course (M15, +29%) and remained above BL until M24 (+6.3%). Previous sub-study observations found B_naive cells reaching nadir at M2 (-90%), recovering to near baseline levels at M12 (-5.0%) (15). Here, for the total study population B_naïve cell counts recovered after completion of each yearly treatment course (+1.6%, M12; +11%, M24, highest value).

T cell subtypes: Like previously published data (15), a substantial decrease in absolute values for all T cell counts was observed immediately after CladT treatment in the FAS. T cell counts decreased further after the second treatment course with nadir values at M15 for most cell subsets. CD4⁺ effector memory (EM), CD4⁺ regulatory (reg), CD4⁺ T helper (Th)17 and CD8⁺ EM reached nadir values at M18 (-49%, -49%, -45%, and -49% respectively). No T cell subtypes had recovered to BL levels at M24 ( Figure 1B ; Table 2 ). The most substantial change from BL levels was observed for CD4⁺ naïve (-84%) and CD8⁺ naïve (-85%) cells at M15. CD4⁺ Th1 and Th2 cell count reductions were like other T cell subtypes across the post-BL period, with the largest median percentage change at M15 for CD4⁺ Th1 and Th2 (-64% and -76%).

NK cell subtypes: Most NK cell subsets reached nadir values at M3 ( Figure 1C ; Table 2 ); however, these counts generally remained above the lower limits of normal (>100×10⁶/L) (18). The most substantial reductions were observed at M3 for CD16^brightCD56^dim (-36%), CD16⁺ (-33%), CD16^lowCD56^bright (-8.9%), and NK (-31%) cell counts, while NKp46 cell counts were most reduced at M6 (-22%). Repopulation to above BL values was observed for CD16^lowCD56^bright and NKp46 cells starting at M6 and M12 respectively. At M18, CD16^lowCD56^bright cells increased to 30% above BL levels, differing from the previous report where CD16^lowCD56^bright cells remained at -9.0% at M12 (15). NKp46 cells reached BL levels at M12 and continued to increase in number after the second of two-yearly treatment courses (M24, 78%).

Serum proteins: Throughout the 2-year study, serum IgG and IgM levels for most participants remained within the normal ranges of 5.65–17.65 g/L and 0.40–2.30 g/L, respectively ( Table 2 ). For IgG, at BL, M3, M24: >50% – <75% were within the normal range. M6, M12: all participants were within the normal range. M15, M18: >75% of participants were within the normal range. For IgM, at all visits approximately >50% – <75% of participants were within the normal range. Mean sNfL levels at BL were 13.11, and 7.53 and 7.87 ng/L at M12 and M24, respectively. During the study, sNfL levels decreased by 25% at M12 and 23% at M24 (p<0.0001). This reduction was also evident in the normalized Z-scores, at M12 the median sNfL Z-score was 0.1 (a z-score of zero is similar relative to the age- and BMI-matched reference population). sNfL Z-score reduction was sustained at M24 ( Figure 2 ).

B cells: In the blood biomarker sub-study population, treatment with CladT reduced the level of pro-inflammatory interleukin (IL)-6 producing B_mem cells in PBMCs after the first and second yearly treatment courses. IL-6⁺ B_mem expression did not reach BL levels at M24 ( Figure 3A ; Supplementary Table 1 ). The reductions for IL-6⁺ B cells and IL-6⁺ B_mem cells were most pronounced at the end of the first year of treatment with CladT, with the most pronounced decrease at M12 (median percentage change from BL: -16% and -12%, respectively). There was an increase at M6 in anti-inflammatory cytokine-producing B cell counts (IL-10⁺) following the first course of CladT; however, after the second treatment course, there was a reduction in IL-10⁺-producing B cells. Overall, a slight decrease in IL-10⁺ B cell counts was seen after M12, with a nadir at M15 (-1.6%). IL-10⁺ B_mem cells were similarly reduced compared to BL, with a nadir at M15 (-2.5%).

T cells: Decreases in cell counts for pro-inflammatory cytokine producing T cells such as granulocyte-macrophage colony-stimulating factor (GMCSF⁺), interferon (IFN)y⁺, and tumor necrosis factor alpha (TNFα⁺) T cells were observed. The reductions in GMCSF⁺ and IFNy⁺ producing CD4⁺ T cell counts were observed from M3 (median percentage change from BL: -3.3% and -3.6%) and continued until M15. For GMCSF⁺ and IFNy⁺ producing CD8⁺ T cells, counts were reduced following the first treatment course, reached a nadir at M15 (-12% and -19%, respectively, Figure 3B , Supplementary Table 1 ). These cells did not recover to BL values. Increases or no substantial reduction in cell counts were observed for anti-inflammatory cytokine-producing T cells (IL-4⁺ and IL-10⁺ T cells, respectively). IL-4⁺ CD4⁺ T cell counts increased slightly following the first treatment course (M3, +0.8%) and remained above BL levels to M24 (+1.7%). IL-4⁺ CD8⁺ T cell counts were also slightly increased over BL up to M24 (+0.6%).

Differentially expressed genes and the associated pathways were analyzed at each of the reported time points. The most prolific changes were observed after the completion of each yearly CladT treatment course ( Figure 4 ). At the end of the 1^st treatment course (BL to M12) both B and T cells showed upregulation in humoral immune response genes and pathways ( Figures 4 , 5 ). While B cell expression was upregulated (n=229), most T cell transcripts (n=260) were downregulated at this time point ( Supplementary Table 2 ). B cells showed an increase in CD19, CD74, CD22, and NF-κB inhibitor (NFκBI) genes, while an increase in granzyme A (GZMA), IL-32 and CD27 genes were seen in T cells, and a decrease in nuclear factor (NF)-κB2. Adaptive and innate immune responses were detected in B cell genes. At the time of completion of the second treatment course (comparing M15 to BL), an increase in B cell immunity and humoral response was seen for both B and T cells. Upregulated genes for CD3⁺ cells included: CDK6, CD27, IL-32, GZM(A, B, H and K), NKG 7, and DNA-binding factor TOX. At this timepoint, transcripts associated with adaptive immune response and regulation of IFNy+ production were upregulated in T cells only. Transcripts upregulated in B cells included SOX4, TNFRSF21, CCl5, and SOD 2. CD19⁺ cells had downregulated chemokine encoding genes at this time point (CXCR3). At the end of the study (BL to M24), upregulation was detected in two pathways only: oxidative phosphorylation and humoral immune response for B cells, otherwise gene expression was like BL. The full list of genes at M12, M15 and M25 compared with BL is presented in Supplementary Appendix 2 .

In the FAS, the ARR was 0.11 (95% CI: 0.09, 0.15), with 83% of participants free of qualifying relapses and 91.8% free of 6mCDP during the 24M study period. 92.9% and 96.9% of participants did not have confirmed progression on T25FWT and 9HPT, respectively ( Supplementary Table 3 ). Among participants with MRI data (n=265), significant reductions in the annualized CUA count were seen from M2 ( Supplementary Figure 3A ). The maximum mean change was reached at M6 and was maintained until M24. In the total population, 61.7% of participants had no T1 Gd+ lesions and 57.7% had no active T2 lesions (without T1 Gd+) at BL. By M24, this proportion increased to 92.8% and 86.9%, respectively ( Supplementary Figure 3B ). The safety profile observed in this study was generally consistent with that previously reported, with no new safety findings reported during the study. Incidences of study treatment related infections were reported in 55 (20.4%) participants and were either mild or moderate. No severe study treatment related infections were observed ( Supplementary Table 4 ).

In participants with both MRI and B cell data, the correlation was found between B_mem or B_reg cell counts and CUA, T1 Gd+ or active T2 lesions at M3 or M6 was not statistically significant ( Supplementary Figure 4 ). Most participants were grouped into the ≥90% category for both correlation analyses. A significant but weak negative correlation was observed between changes in CUA and changes in B_reg cells from BL to each M12 and M15 timepoints. No further significant correlations were observed for B_mem or other timepoints for B_reg cells ( Supplementary Figure 5 ). No significant correlations were found between disability status categories and selected immune cell subset ratios, including CD16^brightCD56^dim/NK cells and CD16^lowCD56^bright/NK cells ( Supplementary Figure 2 ).