Natalizumab promotes activation and pro-inflammatory differentiation of peripheral B cells in multiple sclerosis patients

The enormous success of B cell-depleting anti-CD20 antibodies in treatment of multiple sclerosis (MS) [1] corroborates that B cells play an important role in its pathogenesis. Antigen-activated, differentiated B cells most likely act as potent antigen-presenting cells for the activation of T cells and as providers of pro-inflammatory cytokines [2]. Hereby, B cells are considered to be crucial for the development of new inflammatory central nervous system (CNS) lesions and acute MS relapses.

Based on the earlier assumption that the pathophysiology of MS is mainly mediated by T cells, the majority of established MS medications have been designed and trialed to target pro-inflammatory T cell properties. As a consequence, relatively little is known on how these MS drugs may influence B cells, and if so, how this may contribute to their therapeutic efficacy. Therefore, we investigated in a parallel approach how the four commonly prescribed medications dimethyl fumarate (DMF), fingolimod (FTY), glatiramer acetate (GA), and natalizumab (NAT) affect peripheral B cells with a focus on B cell activation, differentiation, and cytokine production, a procedure allowing us to directly compare treatment effects without inter-study discrepancies.

First, we investigated the effect of treatment with FTY, NAT, DMF, and GA on the overall abundance of B cells in the blood. Compared to untreated MS patients, the FTY group showed a reduced B cell frequency, NAT treatment resulted in a significant increase of B cells, and both DMF and GA had no detectable effect (Fig. 1a, b). DMF and even more so FTY raised the relative abundance of immature transitional B cells, while the frequency of differentiated memory B cells was correspondingly lower in both groups. DMF treatment was furthermore associated with a reduced frequency of CD27⁺ antigen-experienced B cells, while NAT treatment resulted in a substantial rise of this mature B cell population. Lastly, the proportion of plasmablasts was elevated upon FTY treatment (Fig. 1c). Using this set of parameters, GA treatment exerted no detectable effect on B cells.

Analyzing the effect of NAT in detail, we found that NAT treatment was associated with an incline of the total number of leukocytes (Fig. 2a left), including lymphocytes, monocytes, eosinophils, and basophils (Fig. 2a right). Within the enriched fraction of lymphocytes, the frequency of B cells was proportionally increased, while CD4⁺ T cells were compensatory reduced (Fig. 2b left). However, when calculating absolute numbers, all investigated immune cell populations were significantly elevated upon NAT treatment (Fig. 2b right). Furthermore, B cells from NAT-treated patients were distinctly enriched in antigen-activated and memory B cells (Fig. 2c), showed a higher production of pro-inflammatory cytokines (Fig. 2d), and presented with an elevated frequency and total number of activated CD25⁺, CD69⁺, and CD95⁺ cells (Fig. 2e). While the expression of MHC-II remained unchanged, the level of the co-stimulatory molecules CD40, CD80, and CD86 was increased on B cells upon NAT treatment (Fig. 2f). In conjunction, this finding revealed a rise in all investigated immune cell subpopulations and pointed towards a predominant increase in pro-inflammatory B cell subsets upon NAT treatment. In line with this data, the longitudinal analysis of interrelated samples collected prior and after NAT treatment initiation confirmed our observations regarding memory, antigen-experienced (CD27⁺), CD40⁺, CD95⁺, and TNF⁺ B cells (Fig. 3) and consolidated the assumption that NAT treatment triggers the activation and pro-inflammatory differentiation of B cells. In addition, NAT treatment was associated with an increase in the production of pro-inflammatory TNF and IL-6 by CD14⁺ myeloid cells (Fig. 2g) as well as an enhanced expression of the activation marker CD150, an effect which was even more pronounced when absolute cell numbers were analyzed. Only minor effects could be detected regarding the assessed parameters on T cells, where the number of MHC-II⁺ CD8⁺ T cells was slightly elevated in NAT-treated patients (Fig. 2h).

In order to study the immune-stimulating effect of NAT mechanistically, we cultured whole PBMC of healthy donors with increasing concentrations of NAT (0–120 μg/ml) in vitro, which were chosen according to reported serum levels in treated patients [4]. In line with the patient’s ex vivo data, in vitro NAT exposure enhanced the production of pro-inflammatory TNF and IL-6 by B cells and CD14⁺ myeloid cells (Fig. 4a, b) and upregulated the expression of CD40, CD69, and CD95 on B cells (Fig. 4c). To explain mechanistically how the aforementioned enrichment of B cells in NAT-treated patients may occur, we investigated the proliferation and apoptosis behavior of these cells upon NAT treatment. Within the chosen culture period, NAT exposure to PBMC or purified B cells exerted an effect on neither B cell frequency (no antibody 11.1 ± 0.7%, NAT 12.0 ± 0.9%; isotype 11.8 ± 1.1%), nor their apoptosis rate (Fig. 5a, b) or proliferation (Fig. 5c-d).

In this manuscript, we aimed to investigate in a parallel approach how commonly used MS medications affect B cells, a procedure allowing us to compare treatment effects directly without inter-study discrepancies. Despite the fact that all investigated agents have proven efficacy in large clinical trials, our data showed that DMF, FTY, GA, and NAT exerted differential, in parts even opposing effects on the composition and properties of peripheral B cells. Confirming previous studies, our data showed that GA treatment has no detectable effect on B cell maturation and differentiation [5], while both DMF [6] and FTY [7] suppressed the prevalence and function of mature B cells and increased the relative frequency of their immature phenotypes.

NAT treatment on the contrary was associated with a substantial expansion of all leukocytes in blood, but most strikingly of mature B cell subsets. In an attempt to investigate whether this enrichment may ascribe to changes in cell survival, we challenged B cells in vitro with NAT, but detected neither effects on apoptosis nor proliferation. Various other explanations have been proposed to explain the rise of B cells in blood. Based on the observation that particularly the number of CXCR3⁺ B cells is increased in this compartment [8], Saraste et al. suggested that NAT treatment detains B cells with a high migratory capacity in the blood stream, which would—without treatment—extravasate into the inflamed tissue. Others though claimed that NAT treatment perturbs the homing of mature B cell subsets into secondary lymphoid organs, which increases their prevalence in the circulation [9]. Our in vitro observation that NAT exposure enhances neither apoptosis nor proliferation of B cells may add an additional explanation why B cells are enriched upon NAT treatment.

Besides these assumed effects on cellular compartmentalization, we observed that in vivo and in vitro treatment with NAT resulted in B cells with an elevated production of pro-, but not anti-inflammatory cytokines and enhanced expression of activation marker and co-stimulatory molecules. Especially the latter setting indicated that NAT exerted a direct stimulating effect on B cells, possibly mediated by the bidirectional signaling effect upon binding of the NAT to CD49d [10, 11], a mode of action suggested to increase IL-2, interferon-γ, and IL-17 production of purified CD4⁺ T cells in vitro [12]. In this context, it was shown that NAT therapy modulates microRNA and pro-inflammatory cytokine expression in T cells of MS patients [13, 14], suggesting underlying epigenetic processes. In comparison with the described stimulating impact of NAT on T cells, we observed a much stronger effect on B cells in our study. This may be explained by the higher expression of CD49d on B cells than T cells [15]. Furthermore, CD49d is as well expressed on CD14⁺ myeloid cells [16], a fact possibly explaining why the production of pro-inflammatory cytokine by these cells is also enhanced upon NAT treatment.

We assume that the here described increase in B cell number and activation upon NAT treatment is not just an epiphenomenon but has clinical implications. This assumption is fueled by the observation that high B cell frequencies after NAT treatment are associated with ongoing disease activity [17]. Along the same lines, a recent report showed that B cells are enriched in the CNS parenchyma of patients experiencing relapses after NAT therapy cessation [18] placing focus on the question whether a prompt therapeutic follow-up with B cell suppressive agents can decrease the risk of such relapses. Indeed, first investigations showed that patients receiving the B cell-inhibiting agent FTY as follow-up treatment after NAT experience less frequent relapses than patients receiving GA [19] or no ensuing treatment [20]. B cell depletion with rituximab was even more effective than FTY as a post-NAT treatment [21], underscoring the presumed key role of B cells in relapses after NAT cessation.

Moreover, the development of extra nodal large B cell lymphomas that has been described in some NAT-treated patients [22] may be linked to our observations regarding B cell activation and proliferation.

Lastly, the expansion of peripheral B cells with presumed pathogenic properties plausibly explains why NAT worsens neuromyelitis optica [23, 24]. In this regard, it needs to be evaluated how patients with other CNS demyelinating diseases, which show a prominent B cell contribution, such as myelin oligodendrocyte glycoprotein antibody-associated encephalomyelitis [25], or MS patients solely and consistently responding to plasmapheresis as relapse intervention [26] respond to NAT.

Established MS therapeutics exert fundamentally opposing effects on B cells, reaching from their inhibition (DMF, FTY) to substantial activation (NAT). Possible clinical consequences of these complex alterations yet need to be established.