In this exploratory case-control study, a total of 30 untreated patients with relapsing-remitting multiple sclerosis (RRMS) and 20 RRMS patients treated with oral cladribine tablets were included. Blood samples were obtained from these patients, which were subsequently used for both ex vivo and in vitro analysis.

The diagnosis of MS was based on the McDonald 2017 criteria (36). The untreated patients with RRMS were newly diagnosed and had not received any disease-modifying therapy. They were included in the study at least one month after their last steroid treatment.

The administration of a complete cladribine treatment course comprised oral intake of 3.5 mg/kg in two annual courses, each consisting of two treatment weeks, with a four-week interval between. Patients who received cladribine tablets (Mavenclad^®, Merck KGaA, Darmstadt, Germany) were included in the study at week 52 (W52), i.e., one year after their first annual course. Additionally, they were included at W60, W72, and W96, i.e., 8, 20 and 44 weeks after the beginning of second annual course.

All participants provided informed, written consent to participate in the study, and the research protocol was approved by the regional scientific ethics committee (protocol number H-16047666).

Peripheral blood mononuclear cells (PBMCs) were obtained from freshly collected venous blood using a density gradient centrifugation technique (Lymphoprep; Axis-Shield, Oslo, Norway). After isolation, PBMCs underwent two consecutive washes in cold phosphate-buffered saline (PBS) supplemented with 2-mM-ethylenediaminetetraacetic-acid (EDTA). Freshly isolated PBMCs were used for flow cytometry phenotyping, and excess PBMCs were cryopreserved in fetal bovine serum (FBS) with 10% dimethyl sulfoxide (DMSO) for subsequent in vitro analysis.

A minimum of 350,000 freshly isolated PBMCs were incubated with an Fc receptor-blocking reagent (Miltenyi Biotec, Bergisch Gladbach, Germany) and the cells were stained with fluorochrome-conjugated antibodies specific for the surface molecules of interest. For B cells the following antibodies were used: CD19 (PerCP/Cy5.5; HIB19), CD27 (FITC; 323), CD38 (BV421, HIT2) and CD11c (PE/Cy7; Bu15), all from BioLegend (San Diego, CA, USA). For T cells, the following antibodies were used: CD3 (APC/Cy7; HIT3a), CD4 (PerCP/Cy5.5; RPA-T4), CD25 (PE; M-A251), CD127 (APC; A019D5), CXCR5 (AF488; J252D4), PD-1 (BV605; EH12.2H7), CXCR3 (PE/Cy7; G025H7) and CCR6 (BV421; G034E3) from BioLegend. Where applicable, corresponding isotype controls were used. Absolute counts of T and B cell subpopulations were measured using TruCount beads (BD Biosciences, Franklin Lakes, NJ, USA). Flow cytometry data acquisition was performed using a FACS Canto II flow cytometer (BD Biosciences), and data analysis was conducted using FlowJo software (Tree Star, Inc., Ashland, OR, USA).

Cryopreserved PBMCs from 20 of the untreated patients with RRMS, 20 cladribine-treated patients at W52, 20 cladribine-treated patients at W96, and an additional 20 healthy controls were thawed and cultured for 48 hours in RPMI 1640 medium supplemented with penicillin and streptomycin (50 U/ml Gibco, Waltham, MA, USA) and 10% human AB serum (Invitrogen, Carlsbad, CA, USA). Cells were either left unstimulated or stimulated with CpG-ODN 2006 TLR9 agonist (InvivoGen, San Diego, CA, USA) at a concentration of 2.5 μg/ml. To assess B cell cytokine production, the CpG-stimulated and unstimulated PBMCs were subsequently incubated with phorbol-myristate-acetate (PMA; Sigma-Aldrich, St. Louis, MO, USA) at a concentration of 10 ng/ml, 0.5 μg/ml ionomycin (Sigma-Aldrich), and 5 μg/ml brefeldin A (Sigma-Aldrich), at 37°C and 5% CO₂ for 4 hours. Following incubation, cells were surface stained with fluorochrome-conjugated anti-CD19 antibody (PE/Cy7; HIB19) and a live/dead stain. Subsequently, cells were fixed, permeabilized, and stained with anti-IL-10 (APC; JES3-19F1), anti-TGF-β (BV421; TW7-16B4) and anti-LTα (PE; 359-81-11) all from BioLegend. Flow cytometry data acquisition was performed using a FACS Canto II flow cytometer and analyzed using FlowJo software.

Low fluorescent PVDF FluoroSpot plates (Mabtech AB, Nacka Strand, Sweden) were activated with 35% ethanol for 1 minute and washed with sterile water. Plates were then coated with a dilution of 15 μg/ml IFN-γ, IL-13, IL-17 and IL-21 capture antibodies in sterile PBS supplied in Mabtech FluoroSpot kits (#3654-1-1) and incubated for 24 h at 4°C. After incubation, plates were washed in sterile PBS and blocked for 2 h at RT with RPMI-1640/5% human serum (Invitrogen). 250,000 PBMCs from each of 20 untreated patients with RRMS, 20 cladribine-treated patients with RRMS at W52, 20 cladribine-treated patients with RRMS at W96, and 20 healthy controls were added to each well in RPMI-1640/5% human serum with 30 µg/ml myelin basic protein (MBP; HyTest, Turku, Finland), 10 µg/ml myelin oligodendrocyte glycoprotein (MOG; AnaSpec, Fremont, CA, USA), 0.3 µg/ml RAS guanyl releasing protein 2 (RASGRP2; OriGene Technologies Inc., Rockville, MD, USA), 10 µg/ml alpha B crystallin (CRYAB; Abcam, Cambridge, United Kingdom), 5x10⁶ cells/ml heat-killed Candida albicans (HKCA; InvivoGen) or medium alone. A costimulatory monoclonal anti-CD28 antibody (Mabtech) was added to the wells in a concentration of 0.1 µg/ml and plates were incubated for 48 h at 37°C, 5% CO₂. Subsequently, cells were removed, and plates were washed with PBS and biotinylated-detection antibody in PBS/0.1% bovine serum albumin (BSA) was added for 2 h, at RT. Plates were washed and incubated for 1 h at RT with fluorophore-conjugated antibodies in PBS/0.1% BSA according to manufacturer (Mabtech AB). Subsequently plates were incubated with fluorescence enhancer for 15 minutes and afterwards emptied and left to dry protected from daylight for a minimum of 45 minutes. Analysis of spots was performed using the Mabtech IRIS reader system equipped with Apex™ software (Mabtech AB).

The included untreated patients with RRMS and HC were selected to match the group of cladribine treated patients with RRMS. There were no statistically significant differences in age or sex between any of the included groups.

Statistical analyses were performed using GraphPad Prism 7 software (GraphPad software Inc, La Jolla, CA, USA). Mann Whitney U test was applied for comparison of B and T cell subpopulations, between untreated patients with RRMS and patients treated with cladribine for 52 weeks. Friedman paired test was applied for comparison of B and T cell subpopulations between W52, W60, W72 and W96 of cladribine treatment. Kruskal-Wallis test was applied for comparison of B cell cytokine production and CNS antigen reactivity between untreated patients with RRMS and W52 and W96 of cladribine treatment where a Mann Whitney U test was applied for comparison between healthy controls and untreated patients with RRMS. A p value <0.005 was considered statistically significant and <0.05 as suggestive of significance (37).

For this manuscript the STROBE reporting guidelines for observational studies was used.

Data are available in anonymized form and can be shared by request form any qualified investigator. Sharing requires approval of a data transfer agreement by the Danish Data Protection Agency.

We included 30 untreated patients with RRMS, and 20 patients treated with oral cladribine tablets ( Table 1 ). The two groups showed no significant differences in age, sex, or EDSS scores. However, patients in the cladribine group had a longer disease duration (p<0.0001). Among the untreated patients, all had experienced either relapse or magnetic resonance imaging (MRI) activity in the previous year. In the cladribine-treated group, seven had neither relapse nor MRI activity in the year prior to initiating cladribine therapy. These patients had switched from another treatment due to side effects or the development of anti-JC virus antibodies. In the cladribine group, three patients had not received any previous treatment, while 12 had received oral therapies (teriflunomide, dimethyl fumarate, or fingolimod), and five had received monoclonal antibody treatment (natalizumab or rituximab). Before initiating cladribine treatment, all previously treated patients had normal lymphocyte counts, except for one patient previously treated with dimethyl fumarate who had grade 1 lymphopenia.

We found no difference in absolute count of CD19⁺ B cells between 30 untreated patients with RRMS and 20 patients treated with cladribine at W52 ( Figures 1A, B ). However, at W60, i.e., eight weeks after the initiation of the second treatment cycle we found a significantly lower number of CD19⁺ B cells, that was partly maintained to W72, but at W96 CD19⁺ B cell levels were reconstituted to levels comparable to W52 ( Figure 1B ). When analyzing the specific B cell subsets according to their expression of CD27 and CD38 ( Figure 1C ) the number of transitional ( Figure 1D ) and naïve ( Figure 1E ) B cells were higher in patients treated with cladribine at W52 compared to untreated, and although there was a significant decrease in numbers of transitional B cells at W72 and naïve B cells at W60 and W72, W96 numbers were comparable to W52 numbers.

Conversely, looking at the memory B cell compartment we found a significantly lower number of CD27^-CD38^- ( Figure 1F ), CD38^- ( Figure 1G ) and CD38⁺ memory ( Figure 1H ) B cells and plasmablasts ( Figure 1I ) at W52 of cladribine treatment compared to untreated patients. We also found that the number of all memory B cell subsets were significantly lower at W60 and W72 compared to W52, except for CD27^-CD38^- memory B cells that only showed a decrease in numbers at W60 and the CD38⁺ memory B cells which remained significantly lower at W96 than at W52.

Investigating atypical, CD11c⁺ memory B cells ( Figure 1J ) showed that at W52 of cladribine treatment there were significantly lower numbers of CD11c⁺CD38^- and CD38⁺ memory B cells whereas no difference was observed for CD38^-CD27^- B cells compared to untreated patients with RRMS ( Figure 1K ). In all three CD11c⁺ B cell memory subpopulations there was a significant decrease in numbers at W60 and for the CD38⁺ memory B cells the decrease persisted at W70, but at W96 of cladribine treatment CD11c⁺ memory B cells were reconstituted to levels comparable to W52 ( Figure 1K ).

Cladribine treatment induced long-term depletion of CD4⁺ T cells persisting from W52 to W96 with even lower numbers of circulating cells at W60 and W72 than W52 ( Figures 2A, B ). We measured the absolute count of Tfh and Tfr cells and non-follicular regulatory T cells (Treg) in order to investigate the B cell activation potential. We defined Tfh cells as CD4⁺CD127⁺CXCR5⁺, Tfr cells as CD4⁺CD127^-CD25^hiCXCR5⁺, activated Tfh and Tfr cells as PD-1⁺ and non-follicular Tregs as CD4⁺CD127^-CD25^hiCXCR5^- T cells ( Figures 2C, D, F ). The number of circulating CD25^- and CD25^int Tfh cells was significantly lower at W52 compared to untreated patients with RRMS ( Figure 2E ). While the number of CD25^- Tfh cells was unchanged from W52 to W96, the number of CD25^int Tfh cells decreased significantly at W60 and W72 (Figure E). There were no changes in numbers of PD-1⁺ Tfh cells between untreated patients with RRMS and W52 of cladribine treatment, but after the second treatment cycle PD-1⁺ CD25^int Tfh cells were significantly reduced at W60, W72 and W96 compared to W52 ( Figure 2E ).

Looking at the regulatory potential we found a suggestively significant lower number of Tfr cells at W52 in cladribine-treated patients compared to untreated, and furthermore we found a suggestively significant decrease in the number of Tfr cells at W72 and a significantly lower number after W96 compared to W52 in the cladribine-treated patients ( Figure 2G ). There were no changes in numbers of PD-1⁺ Tfr cells between untreated and W52 of cladribine-treated patients with RRMS. There was a significantly lower number of PD-1⁺ Tfr cells at W72 compared to W52 of cladribine treatment, but at W96 the number of PD-1⁺ Tfr cells were reconstituted to levels comparable to W52 ( Figure 2G ). Total numbers of non-follicular Tregs were lower at W52 in cladribine-treated patients compared to untreated and significantly lower at W72 and W96 of cladribine treatment than at W52 ( Figure 2G ).

Analyzing the Tfh: Tfr ratio between groups showed increased Tfh: Tfr cell ratios of both CD25^- and CD25^int cells at W96 compared to W52 in cladribine-treated patients. Likewise, both PD-1⁺CD25^- and PD-1⁺CD25^int Tfh: Tfr cell ratios were higher at W72 compared to W52. However, at W96 of cladribine therapy we found a shift in PD-1⁺CD25^int Tfh: Tfr cell ratios with a significantly lower ratio compared to W52 ( Figure 2H ).

Analyzing the long-term cladribine-induced effects on non-follicular effector T cells (CD4⁺CD127⁺CXCR5^-), we measured absolute counts of the functionally distinct effector T cells subsets Th1, Th17 and Th17.1 characterized according to their expression of CXCR3 and CCR6 ( Figure 3A ). After W52 of cladribine treatment we found a significant and suggestively significant lower number of Th17 and Th17.1 cells, respectively, compared to untreated patients with RRMS. ( Figure 3B ). After the second cladribine treatment cycle we found a suggestively significant decrease in Th1 cells at W60 compared to W52, however, at W72 and W96 Th1 cell numbers were comparable to W52. ( Figure 3B ). Likewise, numbers of Th17.1 were significantly lower at W60 and W72 but reconstituted at W96 to levels comparable to W52 ( Figure 3B ).

We further subdivided the Tfh cell subsets into the effector subpopulations Tfh1 (CXCR3⁺CCR6^-), Tfh2 (CXCR3-CCR6-), Tfh17 (CXCR3^-CCR6⁺) and Tfh17.1 (CXCR3⁺CCR6⁺) ( Figure 3A ). Here we found that for both CD25^-, PD-1⁺CD25^- and CD25^int Tfh cell populations the number of Tfh1 and Tfh2 cells was significantly lower at W52 in cladribine-treated patients compared to untreated patients ( Figures 3C–F ). Conversely, numbers of Tfh17 cells were higher in both the CD25^int and the PD-1⁺CD25^int Tfh cell populations, and Tfh17.1 cells were significantly higher in CD25^- and PD-1⁺CD25^int Tfh cells at W52 in cladribine-treated patients compared to untreated ( Figures 3C–F ). After the second treatment cycle numbers of CD25^-, PD-1⁺CD25^- and PD-1⁺CD25^int Tfh17 cells were lower at W60 or W72 compared to W52 but were reconstituted to levels comparable to W52 at W96 ( Figures 3C–F ). Conversely, numbers of CD25^int and PD-1⁺CD25^int Tfh1 and Tfh2 cells were significantly lower at both W60, W72 and W96 compared to W52 of cladribine treatment ( Figures 3C–F ).

To analyze the long-term effects of cladribine treatment on B cell cytokine responses, we stimulated cryopreserved PBMCs from 20 healthy controls (median age 42 years; range 24-69 years), 20 untreated patients with RRMS (median age 37 years; range 26-62 years) and 20 patients treated with cladribine at W52 and W96 with or without CpG for 2 days and evaluated the IL-10, TGF-β, and LTα profile ( Figures 4A, C, E ) of B cells after a short re-stimulation. ( Figures 4B, D, F ) Without CpG-stimulation we found a significantly lower frequency of IL-10, TGF-β, and LTα producing B cells at W52 compared to untreated patients, however, frequencies of both IL-10 and TGF-β producing B cells were reconstituted to levels comparable to untreated patients at W96. Following CpG stimulation there were no differences between frequencies of IL-10 producing B cells in any of the groups, but frequencies of TGF-β were suggestively significantly decreased compared to untreated at W52 but not at W96 ( Figures 4B, D ). Conversely, the frequency of LTα producing B cells remained lower in cladribine-treated than in untreated patients at W52 and W96 with and without CpG stimulation ( Figure 4F ).

Cryopreserved PBMCs from 20 healthy controls, 20 untreated patients with RRMS and 20 patients treated with cladribine at W52 and W96 were stimulated for 48 h with the autoantigens MOG, MBP, CRYAB (for 8 patients from each group), or RASGRP2 together with co-stimulatory anti-CD28 or medium with anti-CD28 alone as control. Following stimulation, IFN-γ, IL-17, IL-13 and IL-21 spot forming units (SFU) at the single cell level were measured. Untreated patients with MS had a suggestively higher number of IL-13 and IL-17 SFUs than healthy controls after stimulation with MOG and CRYAB respectively ( Figures 5B, C ).

IFN-γ and IL-13 SFUs were lower after stimulation in medium with anti-CD28 at W96 and for IFN-γ SFUs also at W52 compared to untreated patients ( Figures 5A, B ). IFN-γ SFUs were significantly lower at W96 in response to MBP, MOG and RASGRP2 and the IFN-γ response to RASGRP2 was also suggestively significantly lower at W52 compared to untreated ( Figure 5A ). IL-17 SFUs were significantly lover at W96 in response to MOG and CRYAB compared to untreated patients ( Figure 5C ) IFN-γ⁺IL-17⁺ SFUs were significantly lower in response to CRYAB at W52 and furthermore suggestively significant lower at W96 in response to both CRYAB and MOG compared to untreated patients ( Figure 5E ). There were no difference in IL-21 response between groups to any of the autoantigens ( Figure 5D ).