The therapeutic mode of action of dimethyl fumarate (DMF), approved for treating patients with relapsing-remitting multiple sclerosis, is not fully understood. Recently, we and others demonstrated that Ab-independent functions of distinct B cell subsets are important in mediating multiple sclerosis (MS) relapsing disease activity. Our objective was to test whether and how DMF influences both the phenotype and functional responses of disease-implicated B cell subsets in patients with MS. High-quality PBMC were obtained from relapsing-remitting MS patients prior to and serially after initiation of DMF treatment. Multiparametric flow cytometry was used to monitor the phenotype and functional response-profiles of distinct B cell subsets. Total B cell counts decreased following DMF treatment, largely reflecting losses of circulating mature/differentiated (but not of immature transitional) B cells. Within the mature B cell pool, DMF had a greater impact on memory than naive B cells. In keeping with these in vivo effects, DMF treatment in vitro remarkably diminished mature (but not transitional B cell) survival, mediated by inducing apoptotic cell death. Although DMF treatment (both in vivo and in vitro) minimally impacted B cell IL-10 expression, it strongly reduced B cell expression of GM-CSF, IL-6, and TNF-α, resulting in a significant anti-inflammatory shift of B cell response profiles. The DMF-mediated decrease in B cell proinflammatory cytokine responses was further associated with reduced phosphorylation of STAT5/6 and NF-κB in surviving B cells. Together, these data implicate novel mechanisms by which DMF may modulate MS disease activity through shifting the balance between pro- and anti-inflammatory B cell responses.
Multiple sclerosis (MS) is a chronic inflammatory demyelinating disease of the CNS (1). An imbalance between proinflammatory immune effectors and anti-inflammatory immune regulators has been implicated in MS disease pathogenesis, with a traditional focus on the role of particular T cell subsets (1). Oral dimethyl fumarate (DMF) has been recently approved for treating patients with relapsing-remitting MS (RRMS) (2–4), yet the mode of action for DMF is still not fully understood. Although emerging evidence has suggested that DMF can downregulate T cell and myeloid cell proinflammatory responses (5–11), relatively little is known about the impact of DMF on B cell subset responses, which are now strongly implicated in relapsing MS disease activity based on the clinical success of B cell depleting therapy.
Interestingly, B cell depletion using anti-CD20 monoclonal Abs effectively decreased MS disease activities without apparently affecting the abnormal Ab levels in the CSF of MS patients (12–15), suggesting that secreting pathogenic autoantibodies may not be the primary mechanism by which B cells contribute to MS relapses. Indeed, we and others have shown that Ab-independent functions of B cells, such as Ag presentation and production of proinflammatory cytokines by functionally distinct B cell subsets, are important contributors to MS disease activity (16–23).
In this study we examined whether and how DMF may influence both the phenotypes and functional response profiles of distinct B cell subsets. We show that total B cell counts diminish substantially following the initiation of DMF treatment, a decrease that largely reflects the loss of circulating differentiated but not of immature transitional B cells in treated patients. In vitro treatment with DMF mirrored the in vivo effects, directly inducing mature B cell but not transitional B cell apoptosis. The functional analysis further revealed that treatment with DMF (both in vivo and in vitro) decreased B cell expression of proinflammatory (GM-CSF, IL-6, and TNF-α) but minimally impacted anti-inflammatory (IL-10) B cell response profiles, associated not only with preferential apoptosis, but also with reduced phosphorylation of STAT5/6 and NF-κB in surviving B cells. Our study suggests that the capacity of DMF to limit new MS inflammatory disease activity may, in part, relate to its ability to mediate an anti-inflammatory shift in the balance of phenotypically and functionally distinct B cell subsets.
Materials and Methods
Subjects and study design
A total of 13 patients (11 females, 2 males) with McDonald criteria–confirmed RRMS, mean age 41 (range 20–60), were prospectively followed at a single center in Montreal, Canada, prior to and following treatment initiation with DMF. Patients were assessed every 3 mo with clinical review, physical examination, and expanded disability status score. At study entry, patients had an average expanded disability status score of 2.5 (range 1.0–4.0), preceding annualized relapse rate of 0.8 (0–2), and disease duration of 9.6 y (range 1–27 y). Of the 13 patients, 11 had previously been treated with either IFN or glatiramer acetate, one had received a single dose of ofatumumab 18 mo prior to recruitment, and one was treatment-naive. All participants, including healthy control (HC) volunteers providing blood for in vitro studies, were recruited at the Montreal Neurological Institute and Hospital after providing informed consent as approved by the Montreal Neurological Institute and Hospital ethics review board.
Sample processing and cell culture
High-quality PBMCs were isolated from all HC participants and from RRMS patients prior to and following treatment initiation with DMF. All steps of sample procurement, handling, PBMC isolation (by density centrifugation using Ficoll; GE Healthcare), cryopreservation, and subsequent thawing followed the identical standard operating procedures developed and validated by the experimental therapeutics program of the Montreal Neurological Institute. Cryopreserved PBMC from individual patients, collected serially, were thawed and cultured in batch, thereby eliminating interassay variability. Where indicated, magnetic bead sorting (Miltenyi Biotec) was used to positively select CD19+ B cells from fresh PBMC with purities routinely >98% as confirmed by flow cytometry. Isolated B cells were plated in U-bottom 96-well plates at 2 × 105/well in a total volume of 200 μl of serum-free x-vivo medium (Lonza), and stimulated with soluble CD40L (1 μg/ml; Enzo Life Sciences), goat anti-human BCR cross-linking Ab (Xab) (10 μg/ml; Jackson ImmunoResearch) with or without IL-4 (20 ng/ml; R&D Systems) for 48 h, at which time supernatants were collected and frozen (−70) for subsequent quantification of cytokine secretion by ELISA and the cells were analyzed by flow cytometry (as described below). B cells were cultured in parallel wells in either medium alone, vehicle (DMSO), monomethyl fumarate (MMF), or DMF (Sigma Aldrich, Oakville, ON, Canada). MMF and DMF were added to individual wells at concentrations of 50 μM, with DMSO control added at the equivalent concentration.
Abs for flow cytometry analysis and intracellular cytokine staining
B cell immunophenotyping panels are listed in the Supplemental Table. Abs used to phenotype B cells were directed against: CD11c (B-ly6), CD20 (2H7), CD24 (ML5), CD27 (L128), CD38 (HB7), CD43 (1G10), CD80 (B Montreal Neurological Institute B1), CD83 (HB15e), CD86 (2331(Fun-1)), HLA-DR (G46-6), IgD (1A6-2), IgG (G18-145), IgM (G20-127) and appropriate isotype controls, all purchased from BD Bioscience. IgA (IS11-8E10) was purchased from Miltenyi Biotec. Abs for intracellular cytokine staining (ICS) targeted: IL-6 (MQ2-6A3), IL-10 (JEF3-19F1), TNF-α (MAb11), and GM-CSF (BVD2-21C11), as well as appropriate isotype controls, all from BD Bioscience. ICS involved 4 h stimulation with PMA (20 ng/ml; Sigma-Aldrich), Ionomycin (500 ng/ml; Sigma-Aldrich), and GolgiStop (Monensin; BD Bioscience) was followed by cell surface staining, then two washes and addition of a fixation/permeabilization buffer (Cytofix/Cytoperm; BD Bioscience). Cells were then washed with the ICS washing buffer (BD Bioscience) and ICS Abs (noted above) were added to the cell suspensions followed by two additional washes with the ICS washing buffer. For apoptosis assays, cells were stained with Annexin V and propidium iodide (PI; BD Biosciences) following the cell surface staining. All flow phenotyping was carried out by a single operator who was blinded to the sample source and followed the same standardized immune phenotyping protocol (Fig. 1), using an LSR Fortessa flow cytometer (BD Biosciences) and FlowJo software analysis (Tree Star).
Levels of secreted cytokines (GM-CSF, TNF-α, IL-6, and IL-10) within the frozen culture supernatants were quantified by OptEIA ELISA kit (BD Bioscience) based on the manufacturer’s protocols. Briefly, ELISA plates were coated with capture Ab at least 12 h in advance. After 1-h blocking with blocking buffer (10% FCS, PBS), supernatant samples were added to the plate and incubated for 2 h at room temperature. Then, detection Ab was added for 1 h at room temperature. Plates were washed with ELISA washing buffer (0.05% Tween 20, PBS) between each step. Hydrogen peroxide and 3,3′,5,5′-tetramethylbenzidine (BD Bioscience) were then added and the reaction was stopped by 0.01N H2SO4. The plates were then read by a Bio-Rad microplate reader (Model 550; Bio-Rad).
A Student paired t test was used for statistical comparisons between two groups and one-way ANOVA was used for statistical comparisons between more than two groups, as indicated in the figure legends. GraphPad Prism 6 was used to perform all the statistical analyses. A p value ≤ 0.05 was considered statistically significant.
DMF treatment preferentially impacts mature and differentiated B cells in vivo
To assess the impact of in vivo DMF treatment on B cell subsets, absolute counts of surface-defined B cell subsets (gated as in Fig. 1) were quantified by flow cytometry within PBMC obtained from patients with RRMS (n = 12) pretreatment and up to 12 mo after initiation of DMF treatment. Total circulating B cell counts were substantially reduced (by ∼45%) in DMF-treated MS patients (Fig. 2A, p = 0.01), reflecting decreased counts of both naive (Fig. 2B, p = 0.04) and memory (Fig. 2C, p = 0.0011) B cell subsets. The memory B cells appeared more affected, resulting in a small increase of the naive/memory B cell ratio following treatment (Fig. 2D, p = 0.028). Among memory B cells, reduced counts were seen for all subsets, including class-switched memory B cells, non–class-switched memory B cells, and double-negative memory B cells (Fig. 2E–G, p < 0.05). In contrast, immature transitional (CD24high CD38high) B cells were largely not impacted (Fig. 2H), which, together with reduced mature (CD24+/int CD38−/int) B cell numbers (Fig. 2I, p = 0.01), resulted in an increase of the transitional/mature B cell ratio (Fig. 2J, p = 0.02). These data indicate that DMF preferentially impacts mature (especially memory) B cells in vivo. All these changes were observed by the first (3 mo) posttreatment assessment, and at 12 mo on treatment (Supplemental Figs. 1, 2), pointing to an early and persistent differential effect of DMF treatment on distinct B cell subsets.
DMF modulates the balance between proinflammatory and anti-inflammatory B cell responses in treated patients
Prior work has shown that B cells of untreated MS patients can express abnormally high levels of IL-6, TNF-α, and GM-CSF as well as deficient levels of IL-10 (20). We therefore considered whether and how DMF treatment might impact such MS disease–implicated cytokine-defined B cell subsets in treated patents. We found that DMF treatment resulted in substantial decreases in the frequencies of B cells expressing the proinflammatory cytokines IL-6, TNF-α, and GM-CSF (Fig. 3A–C). The counts of IL-10–expressing B cells in these treated patients also tended to decrease, although to a lesser extent (Fig. 3D), resulting in diminished proinflammatory B cell cytokine profiles as indicated by the decreased ratios of proinflammatory/anti-inflammatory B cell subsets (Fig. 3E–G). In keeping with this, DMF treatment also resulted in a substantially reduced expression of the B cell surface molecules CD11c, CD43, CD80, and CD83, known to play important roles in mature B cell:T cell interactions (Fig. 3H–K, p < 0.001).
DMF exposure in vitro induces selective B cell subset apoptosis that mirrors the preferential in vivo depletion of mature B cell subsets
To explore how DMF treatment may lead to preferential decreases of mature B cells in vivo, we exposed freshly isolated human HC peripheral B cells to either DMF or MMF and assessed B cell subset survival and apoptosis. We found that DMF exposure resulted in reduced survival of total B cells, which was due to apoptotic cell death (Fig. 4A, 4B). These cell losses largely reflected apoptosis of mature B cell subsets, whereas, in contrast, survival of transitional B cells was minimally affected (Fig. 4C–F). MMF also induced B cell apoptosis but to a much lesser extent compared with DMF (Fig. 4A–F). Substantial DMF-induced apoptosis of B cells was also seen when DMF was added to B cells that were activated by combined stimulation through the BCR, CD40, and IL-4 (a combination known to enhance B cell survival; Fig. 4G, 4H). Together, these results further support the concept that DMF treatment mainly impacts mature B cell survival, and that the mechanism underlying their preferential losses reflects at least in part the enhanced susceptibility of these subsets to DMF-induced apoptotic cell death.
DMF decreases proinflammatory B cell cytokine responses partially through a pSTAT5/6 and NF-κB dependent mechanism
Based on our earlier observation that in vivo DMF treatment reduces the proinflammatory cytokine response profiles of B cells in patients, we assessed the impact of in vitro exposure to DMF and MMF on cytokine responses of activated B cells. We found that DMF exposure substantially reduced B cell secretion of multiple cytokines as measured by ELISA and recapitulated the selective effects that in vivo treatment had on distinct cytokine-defined B cell subsets, namely, causing marked reductions in the proinflammatory B cell cytokine responses (IL-6, TNF-α, and GM-CSF), and a lesser reduction in IL-10 responses (Fig. 5A–D). These differential effects again resulted in substantially decreased proinflammatory/anti-inflammatory cytokine ratios expressed by the B cells exposed to DMF (Fig. 5E–G). As with the survival data, MMF exposure in vitro had little or no effect on B cell cytokine responses (Fig. 5A–G).
We considered that the substantial reduction in proinflammatory cytokine secretion by B cells exposed to DMF was greater than might be expected based on the degree of apoptotic cell death previously noted. This raised the possibility that, in addition to inducing apoptotic loss of B cells, DMF may modulate the cytokine expression profile of viable cells. To address this, we examined the effects of DMF and MMF on B cell cytokine expression using ICS, and gated on live cells only. Indeed, exposure to DMF (and to a lesser extent to MMF) decreased proinflammatory B cell cytokine expression in non-apoptotic cells (Fig. 5I). Prior work has indicated that DMF may downregulate dendritic cell responses through NF-κB signaling (24, 25) and our own work recently showed that STAT5 and STAT6 signaling in B cells contributes to induction of proinflammatory GM-CSF expression (20). We therefore assessed whether DMF exposure may impact either of these signaling pathways. We observed that B cell exposure to DMF strongly decreased both NF-κB (Fig. 6A) and STAT5/STAT6 (Fig. 6B, 6C) phosphorylation in live activated B cells.
Oral DMF has been recently approved for treating patients with relapsing MS (3, 4, 26–29). The mode of action of DMF has not been fully elucidated as it relates to MS disease mechanisms. Increasing evidence, particularly the success of B cell depleting therapy in limiting new MS disease activity (12–14), has pointed to B cells as relevant mediators of disease relapses. In this study we tested whether and how DMF may impact the B cell compartment in the context of MS. Using a series of iterative in vivo and in vitro studies, we first monitored the effect of DMF on B cell subset phenotypes in treated patients and observed that mature B cells (particularly memory B cells), but not transitional B cells, are preferentially lost in the circulation. This suggested that DMF does not prevent the output of immature B cells from bone marrow, but preferentially impacts more mature B cell subsets. A series of ex vivo and in vitro experiments further indicated that DMF may downregulate proinflammatory cytokine production from B cells, both by preferential induction of their apoptotic cell death as well as by limiting proinflammatory cytokine responses of viable B cells, through inhibition of both pSTAT5/6 and NF-κB.
Several studies have recently shown that DMF treatment in patients reduces circulating T cell (especially CD8+ T cell) counts (6–9). Decreases in total B cell counts have also been noted in DMF-treated patients (7, 9, 11). In vivo, the generation of mature B cells is partially dependent on T cells, such that decreased mature B cells with DMF treatment might indirectly reflect the known effect of DMF on T cells. However, our in vitro observations indicate that DMF can have important direct effects on B cell survival as well as modulation of the response profiles of surviving B cells. The prior T cell studies also noted that memory T cells are more affected by DMF than naive T cells (6–9), which is similar to our observation with B cells, suggesting that a common signaling pathway (susceptible to DMF), may regulate the survival of both memory B cells and memory T cells.
Our implication of B cells as a relevant target for DMF’s ability to limit CNS inflammation is supported by recent results indicating that DMF can decrease disease severity in a B cell–dependent model of experimental autoimmune encephalomyelitis (30). Our data from treated patients and using human-derived B cells indicates that DMF treatment could mediate an anti-inflammatory shift of B cell responses at several levels. First, we note that DMF has relatively little impact on transitional B cells although it substantially impacts survival of mature B cells. Previous studies have shown that transitional B cells can exhibit immune modulatory functions through the secretion of anti-inflammatory IL-10 (31). Our findings suggest that transitional B cells and mature B cells use different machinery to maintain their survival, an insight that may have useful implications for developing more selective B cell targeting therapy. We also note that DMF results in preferential losses of memory versus naive B cells. Compared with naive B cells, memory B cells are known to express higher levels of costimulatory molecules and proinflammatory cytokines (16–18, 20, 32) — which can both more efficiently induce proinflammatory T cell responses — so the increase in naive/memory B cell ratio after DMF treatment may be associated with dampened ability of the B cell compartment to mediate T cell activation. Finally, we show that DMF does not just affect the proportions of phenotypically defined B cell subsets (e.g., transitional versus mature; naive versus memory), but also mediates an anti-inflammatory shift in the remaining B cell cytokine responses. Of particular interest in this regard is our observation that, both in vitro and in vivo, DMF treatment results in an increased ratio of IL-10+ B cells to GM-CSF+ B cells, a consequence of both preferential apoptotic cell death of proinflammatory B cells as well as downmodulation of proinflammatory cytokine expression by surviving B cells (involving pSTAT5/6 and NF-κB signaling). Recently, a subpopulation of STAT5/6–dependent human B cell subset that expresses GM-CSF was found to be over-represented in the circulation of patients with MS (20). An imbalance in these patients between GM-CSF+ B cells and IL-10+ B cells was found to be associated with a proinflammatory shift in their myeloid cell responses. Our current study suggests that the differential effects of DMF on functionally distinct B cell subsets, and in particular an anti-inflammatory shift in the cytokine responses of remaining B cells, may contribute to its ability to limit new MS relapses.
In conclusion, we have demonstrated that DMF particularly targets proinflammatory mature B cell subsets through preferential induction of apoptotic cell death as well as downregulation of pSTAT5/6 and NF-κB in surviving B cells. In the context of MS, these findings extend our understanding of DMF’s putative mode of action in patients as potentially limiting new disease relapses through modulation of the B cell compartment. Of a broader implication, our results point to different utilization of intracellular signaling and survival pathways by functionally distinct human B cell subsets, which may help to guide development of therapies that more selectively target particular B cell subsets of interest.
We appreciate the important input of the Experimental Therapeutics Program at the Montreal Neurological Institute for judicious handling of patient and control samples.
This work was supported by the research foundation of the Multiple Sclerosis Society of Canada (A.B.-O.), Banque National Fellowship (R.L.), and National Natural Science Foundation (Grant 81430035, R.L. and H.L.).
The online version of this article contains supplemental material.
Abbreviations used in this article:
intracellular cytokine staining
The authors have no financial conflicts of interest.