NMDA-receptor antagonists block B-cell function but foster IL-10 production in BCR/CD40-activated B cells

Background B cells are important effectors and regulators of adaptive and innate immune responses, inflammation and autoimmunity, for instance in anti-NMDA-receptor (NMDAR) encephalitis. Thus, pharmacological modulation of B-cell function could be an effective regimen in therapeutic strategies. Since the non-competitive NMDAR antagonist memantine is clinically applied to treat advanced Alzheimer`s disease and ketamine is supposed to improve the course of resistant depression, it is important to know how these drugs affect B-cell function. Results Non-competitive NMDAR antagonists impaired B-cell receptor (BCR)- and lipopolysaccharide (LPS)-induced B-cell proliferation, reduced B-cell migration towards the chemokines SDF-1α and CCL21 and downregulated IgM and IgG secretion. Mechanistically, these effects were mediated through a blockade of Kv1.3 and KCa3.1 potassium channels and resulted in an attenuated Ca2+-flux and activation of Erk1/2, Akt and NFATc1. Interestingly, NMDAR antagonist treatment increased the frequency of IL-10 producing B cells after BCR/CD40 stimulation. Conclusions Non-competitive NMDAR antagonists attenuate BCR and Toll-like receptor 4 (TLR4) B-cell signaling and effector function and can foster IL-10 production. Consequently, NMDAR antagonists may be useful to target B cells in autoimmune diseases or pathological systemic inflammation. The drugs’ additional side effects on B cells should be considered in treatments of neuronal disorders with NMDAR antagonists.

Background B cells are important mediators of the adaptive immune response by their ability to provide antigen presentation and costimulation for T cells and to differentiate into antibody secreting plasma cells. B cells are activated through the ligation of their antigen-specific B-cell receptors (BCR) and costimulatory ligands such as CD40, which drive their proliferation, survival and differentiation [1]. In addition, B cells can be stimulated by innate signals like lipopolysaccharide (LPS), a major constituent of the gram-negative bacterial cell wall that binds to Toll-like receptor 4 (TLR4) expressed on B cells [2,3]. TLR4 plays a pivotal role in the initiation of inflammation and is considered as a potent drug target to prevent severe sepsis, the leading cause of death amongst critically ill patients [4][5][6]. Systemic inflammation induced by LPS also seems to affect neuronal pathology, for instance in multiple sclerosis, Alzheimer's and Parkinson's disease [7][8][9][10].
NMDAR antagonists block the activity of ionotropic glutamate receptors of the NMDA type, which play a central role in synaptic transmission, memory formation and neuronal excitotoxicity [45]. NMDAR antagonists like memantine and ketamine are in use or trial to treat neuronal disorders like Alzheimer's disease and resistant depression, respectively [46,47]. The possibility of their oral application and their non-competitive action on the channel pore, but not the glutamate binding site, make those antagonists suitable to control the glutamatergic transmission in the brain in chronic treatments of neurological diseases [48,49].
In view of the implication of B cells as source for antibodies against receptors and ion channels causing neuronal autoimmune diseases, their immune regulatory function [50] and role in LPS-induced inflammation [51], we investigated how non-competitive NMDAR antagonists modulate B-cell function. We found that the drugs impair B-cell migration, BCR-and LPS-induced proliferation and immunoglobulin (Ig) production. For both stimulatory conditions, inhibition was mediated through cross-inhibition of K v 1.3 and K Ca 3.1 potassium channels and attenuated B-cell signaling. However, antagonist ifenprodil could enhance the production of IL-10, fostering an anti-inflammatory B10 phenotype. Hence, non-competitive NMDAR antagonists may be suitable drugs to dampen pathological inflammatory reactions and to modulate B-cell function in autoimmune diseases. The additional effects of NMDAR antagonists on B cells may be beneficial in treating neuronal disorders.

NMDAR antagonists block B-cell proliferation induced by BCR or LPS stimulation
Splenic B cells were stimulated with anti-IgM (Fab') 2 fragment goat anti-mouse (α-IgM) to mimic BCR triggering by antigens, or with the TLR4 ligand LPS. B-cell proliferation was determined by 3 [H]-Thymidine incorporation at 24 h in the presence or absence of the NMDAR antagonists memantine, an NMDAR open-channel blocker, ifenprodil, a non-competitive inhibitor of the GluN2B subunit of NMDARs, or the competitive NMDAR antagonist D-APV [49,52]. Memantine and ifenprodil inhibited α-IgMas well as LPS-induced DNA synthesis in a concentration dependent manner ( Figure 1A and B). In contrast, the competitive antagonist D-APV had no effect, even at very high doses (300 μM). The proliferative response of B cells activated with PMA and ionomycin (IO) was also inhibited by ifenprodil and memantine, but not by D-APV ( Figure 1C). Costimulation by CD40 Abs enhanced α-IgMand LPS-induced B-cell proliferation, and under these conditions the antagonists only had a weak inhibitory effect, reducing DNA synthesis by 29-32%, respectively, compared to a 72-90% reduction in the absence of CD40 stimulation ( Figure 1D).
The effects of NMDAR antagonists on apoptosis was evaluated on B cells activated for 24 h with α-IgM, α-IgM + CD40, LPS, and LPS + CD40 ( Figure 1E). 5-10% more apoptotic cells were detected in antagonist-treated cultures whereby ifenprodil had stronger effects than memantine, especially on B cells stimulated with α-IgM only. In case of CD40 costimulation, B-cell apoptosis was much lower and both antagonists had no enhancing effect on cell death.

NMDAR antagonists induce membrane depolarization and inhibit K v 1.3 and K Ca 3.1 channels in B cells
We previously reported that protein expression of functional NMDARs in murine T cells is elusive and that NMDAR antagonists inhibit K v 1.3 and K Ca 3.1 channels [53], which are considered as potent targets for immunosuppression [54,55]. These potassium channels are also expressed on B cells and their inhibition was found to differentially influence B-cell proliferation after BCR activation or PMA/IO stimulation [56][57][58][59]. Since K Ca 3.1 and K v 1.3 channel activities influence membrane depolarization and, thereby, the Ca 2+ -flux into the cell [60], we first determined the drugs' effects on the membrane potential. Ifenprodil (20 μM) and memantine (30 μM) reduced the membrane potential of α-IgMor LPS-activated B cells from~−40 mV to~−20 mV and~−10 mV, respectively. Addition of KCl served as a positive control for membrane depolarization ( Figure 2A). Next, we recorded K v 1.   concentrations (300 μM) ( Figure 2D). Thus, K v 1.3 and K Ca 3.1 channels, whose specific blockade abolishes B-cell activation [56,59], are partially inhibited by the non-competitive NMDAR antagonists ifenprodil and memantine.

BCR-and LPS-induced B-cell signaling is attenuated by NMDAR antagonist
Next, we assessed the antagonists' effects on B-cell signaling and Ca 2+ mobilization, which is critical for B-cell activation and proliferation [11,13,61]. Indo-1 AM-labelled B cells showed a concentration-dependent inhibition of BCR-induced Ca 2+ -flux upon treatment with ifenprodil or memantine ( Figure 3A). Furthermore, the levels of phosphorylated Akt, S6 and Erk1/2 were significantly lower in α-IgM-activated B cells in the presence of ifenprodil compared to untreated cells ( Figure 3B, left panel). Notably, B cells stimulated with LPS showed a very similar inhibition of Akt, S6 and Erk1/2 activation by ifenprodil ( Figure 3B, right panel). In long-term stimulation, α-IgMand LPS-activated B cells cultured with ifenprodil exhibited lower levels of pErk1/2 and pS6 in the cytoplasm ( Figure 3C) and a reduced nuclear accumulation of pErk1/2 and NFATc1 ( Figure 3D). Thus, NMDAR antagonists downregulate major signaling events of two distinct B-cell activating receptors that play an important role in innate and antigen-specific B-cell responses [18,62]. Since CD40 costimulation rescued the inhibitory effects of NMDAR antagonists on BCR-induced B-cell proliferation and apoptosis (Figure 1), we analysed pErk1/2 and pS6 expression under costimulatory conditions and found an enhanced activation of both signaling molecules compared to α-IgM stimulation alone ( Figure 3E). However, although addition of ifenprodil reduced pErk1/2 and pS6 levels in α-IgM + CD40-treated B cells, these levels were still above those found after α-IgM treatment. Hence, antagonist-induced attenuated signaling in CD40 costimulated B cells is still above a critical threshold needed for B-cell activation.

NMDAR antagonists impair B-cell migration and Ig production
The migratory response of B cells within the activating lymphoid environment or at inflammatory sites is a key feature for their differentiation and function. We investigated whether NMDAR antagonists affect chemokineinduced migration and found a strong reduction in the migratory response of B cells to the chemokines SDF-1α and CCL21 in the presence of ifenprodil ( Figure 4A). Antibody secretion is the major effector function of B cells. In order to determine the impact of ifenprodil on IgM and IgG production, B cells were stimulated with LPS or LPS + IL-4. Ifenprodil was added at days 1, 2 or 3 and ELISA was performed at day 4. As shown in Figure 4B, the blockade of IgM and IgG secretion was most efficient after addition of ifenprodil at day 1. With increasing time, the inhibitory effect of ifenprodil declined but was still detectable. Hence, NMDAR antagonists not only inhibit B cell proliferation and migration but also antibody secretion.

NMDAR antagonists modulate IL-10 production
Since several B-cell responses were negatively regulated by NMDAR antagonists, we asked whether the druginduced attenuated signaling would influence the production of IL-10, the immunosuppressive cytokine made by B10 cells [18,32,33,35]. Mitogenic stimulation of B cells with PMA and IO leads to the induction of IL-10 mRNA [24,31,63]. Thus, we stimulated B cells with these mitogens for 16 h in the absence or presence of ifenprodil. Drug treatment lead to a strong repression of IL-10 transcripts compared to untreated B cells ( Figure 5A). We then asked whether ifenprodil has effects on B cells that were pre-activated with α-IgM + CD40 Abs, LPS or agonistic CD40 Abs, which are known to give rise to regulatory B10 cells [25,28,35,64]. Ifenprodil was added at day 1, and IL-10 and IFN-γ production were determined at day 2 or 3 ( Figure 5B). IFN-γ production was not altered by ifenprodil. Low levels of IL-10 production were induced in 8% of α-IgM + CD40-stimulated and in 19-27% of CD40-or LPS-activated B cells, which showed low to high levels of IL-10. Ifenprodil had either no effect or lowered the percentage of IL-10 producing B cells in CD40-or LPS-stimulated cultures. In contrast, addition of ifenprodil to α-IgM + CD40-activated B cells increased the frequency of IL-10 producers 1.5-2-fold, although absolute IL-10 expression levels remained low. Experiments with B cells from IL-10-GFP knock-in tiger mice [65] supported these results. α-IgM + CD40-activated B cells, with ifenprodil treatment started after 21-25 h, showed a 50% increase in the percentage of IL-10-GFP + B cells at day 2 ( Figure 5C) and when measured at day 4 a 6-fold increase ( Figure 5D). Therefore, ifenprodil can foster the generation of an IL-10 producing phenotype.

Discussion
Here, we show that non-competitive NMDAR antagonists attenuate adaptive (BCR) as well as innate (LPS/ TLR4) B-cell signaling. The drugs inhibited IgM and IgG secretion, B cell migration and impaired B-cell proliferation and viability, which were partially overcome by CD40 costimulation. Since the non-competitive antagonists ifenprodil and memantine, but not the competitive antagonist D-APV, blocked the activity of K v 1.3 and K Ca 3.1 channels, the used non-competitive antagonists seem to act mainly via inhibition of those K + channels, which maintain a favourable electrochemical gradient that is required for a sustained Ca 2+ -entry through Ca 2+ -release activated channels (CRAC) [15,66]. Due to the similar action of the non-competitive NMDAR antagonists on K v 1.3 and K Ca 3.1 channels as described for the action of specific K + channel inhibitors [57][58][59], a differential modulation seems excluded and is probably the cause for their effects reported here. In line with the inhibition of both K + channel types in B cells, BCR-induced Ca 2+ -flux was reduced and BCR-and TLR4-induced downstream activation of Erk1/2, Akt, S6, and NFATc1 was dampened. Ifenprodil added to B cells pre-activated with BCR/CD40 Abs fostered IL-10 production, but when added at the beginning of B-cell stimulation reduced IL-10 transcripts. Thus, the enhancement of IL-10 production seems to depend on the timing and concentration of drug application, although further experimentation is required for identification of the underlying mechanisms. IL-10 expression is regulated by Ca 2+ -level [63], Bruton's tyrosine kinase (Btk), the adaptor protein BLNK/SLP65, CamKII, Erk1/2, and transcription factors like CREB, STAT3, NFκB or NFAT [13,18,[67][68][69][70][71][72]. The activation of Erk1/2 in B cells in turn is dependent on Ca 2+ -flux and PI-3K activation [73,74]. IL-10 secretion is differentially controlled depending on the activation stimulus and availability of IL-21 [26,75]. Furthermore, IL-10 production by B10 cells seems to be transient [76]. We found that ifenprodil impairs BCR/CD40-induced Erk1/2 and Akt activation and thus speculate that upon ifenprodil treatment subtle differences in Ca 2+ -level [77] and the activity of Erk1/2, Akt and other signaling molecules favour IL-10 production. Interestingly, genetic deletion or inhibition of K v 1.3 channels in T cells was found to increase IL-10 production in T cells along with an amelioration of experimental autoimmune encephalomyelitis and allergic asthma [78,79]. Since NMDAR antagonists block K v 1.3 and K Ca 3.1 channels in B cells, the increase of IL-10 production in BCR/CD40-activated B cells may result from similar mechanisms. Our data suggest that application of non-competitive NMDAR antagonists during chronic treatments of neurological disorders like Alzheimer`s disease may not only involve neuronal NMDARs, but may also have additive (See figure on previous page.) Figure 3 Effects of NMDAR antagonists on B-cell signaling. A) Reduced Ca 2+ -flux in BCR-activated B cells in the presence of NMDAR antagonists. Indo-1 AM-labelled B cells were stimulated with α-IgM (10 μg/ml) in the presence or absence of ifenprodil (left) or memantine (right) and Ca 2+ -flux was determined by flow cytometry. Corresponding graphs show the mean + SD relative ΔCa 2+ -flux of three experiments. B-E) NMDAR antagonists attenuate BCR-and LPS-induced activation of important signaling molecules. B cells were activated with α-IgM (10 µg/ml) or LPS (10 µg/ml) or α-IgM plus CD40 Abs (5 µg/ml) in the presence or absence of ifenprodil (30 μM) in B) short-term and C-E) long-term stimulation. Activation of the indicated signaling proteins in B) total, C, E) cytoplasmic and D) nuclear protein extracts was analyzed by Western blot. β-Actin and Lamin B expression served as controls for protein loading. Indicated numbers give the relative protein expression after quantification and normalization to controls. Data are the representative of two (E) and three (B-D) independent experiments. side-effects by targeting B cells, which are assumed to contribute to these disorders [7,9,10]. Given that the drugs impaired several B-cell functions, but enhanced IL-10 production in BCR/CD40-stimulated B cells, their employment in systemic inflammation or autoimmune diseases, for instance in sepsis or anti-NMDAR encephalitis, appears promising. Here, antagonists may limit B-cell hyper-reactivity and antibody production or mediate immunomodulation or suppression through an enhanced frequency of IL-10 secreting B cells. IL-10 producing B cells also target T cells as they induce IL-10 producing CD4 + T cells, suppress Th1 cell differentiation and increase the number of CD4 + CD25 + Foxp3 + regulatory T cells in vivo [29]. Furthermore, although action of noncompetitive NMDAR antagonists on memory B cells is not investigated, pharmacological modulation of memory Bcell differentiation or secondary B-cell responses can be envisaged. Since specific blockade of K v 1.3 and K Ca 3.1 channels results in immunosuppression of T and B cells [54,59] and non-competitive NMDAR antagonists block these two K + channels in B cells, application of NMDAR antagonists may also be useful to treat acute and chronical allograft rejections driven by B cells. Memantine, which passed clinical trials and is in use to treat advanced Alzheimer`s disease, might show similar effects as the specific K v 1.3 and K Ca 3.1 blockers Shk and TRAM-34 in treating allograft vasculopathy or kidney allograft rejection [80,81]. However, further studies are required to determine the drug's suitability for in vivo treatment of these immune disorders.

Conclusions
Through their nonspecific action on K v 1.3 and K Ca 3.1 potassium channels, non-competitive NMDAR antagonists are potent modulators of LPS/TLR4-and BCRinduced proliferation, migration, Ig production and anti-inflammatory IL-10 production by B cells. Thus, they may be useful to target B cells under pathological inflammatory conditions. They may also have beneficial side effects during chronic treatments of neurological disorders like Alzheimer's disease.

Methods
Mice C57BL/6 mice were used at the age of 6-10 weeks. IL-10-GFP knock-in mice, designated interleukin-ten ires gfp-enhanced reporter (tiger) mice [65] were 8 or 28 weeks old and kindly provided by J. Hühn, HZI Braunschweig, Germany. All animal work performed was in compliance with the German and local guidelines for the Use of Experimental Animals.

Apoptosis measurement
Apoptosis was determined with the Apoptosis detection kit from BD Pharmingen (Heidelberg, Germany). 2×10 5 splenic B cells were left untreated or were activated with α-IgM (10 μg/ml) or LPS (10 μg/ml) without or with costimulation by CD40 Abs (5 μg/ml, Biolegend, San Diego, CA, USA) in the presence or absence of ifenprodil (30 μM, Tocris Biosciences). At 24 h of culture cells were harvested, stained with Annexin V-FITC (BD Pharmingen) and propidium iodide (PI, Sigma-Aldrich) according to manufacturer's protocol and analyzed by flow cytometry using a FACSFortessa and Cell Quest software (BD Biosciences). The percentage of viable cells was determined by gating on AnnexinV − PI − cells.

Migration assay
Splenocytes (4×10 6 ) were left untreated or were incubated with ifenprodil (30 μM) for 30 min in D-MEM medium (Biochrom) supplemented with 0.1% BSA and 10 mM HEPES pH7.4. Cells were transferred unto fibronectincoated (6.5 μg/ml, Roche Diagnostics, Basel, Switzerland) transwell chambers (3.0 μm pore size, Corning Costar, Tewksburry, MA, USA) and SDF-1α (100 ng/ml) or CCL21 (300 ng/ml, both from PeproTech, Hamburg, Germany) was added. Migration was performed for 150 min at 37°C and stopped with 0.1 M EDTA. Migrated cells were stained with rat anti-mouse B220-FITC Ab (RA3-6B2, BD Pharmingen) and measured for 30 s at a FACS Fortessa. The number of B cells migrated in the presence of chemokine (set as 1.0) was divided by the number of cells migrated in the absence of chemokine to obtain the relative migration values.

Intracellular cytokine staining and IL-10-GFP induction
Splenic B cells were stimulated with α-IgM (10 μg/ml) plus CD40 Abs (5 μg/ml), CD40 Abs alone (5 μg/ml), or LPS (10 μg/ml) for 48 h or 72 h. Ifenprodil (10 μM) was added once and at day 1. Before harvest, cells were treated with IO (800 ng/ml) and PMA (500 ng/ml) for 4 h in the presence of Brefeldin A (3 μg/ml, all from Calbiochem). Thereafter, cells were fixed and stained with IL-10-PE and IFN-γ-FITC Abs using IgG2b-PE/FITC isotype controls (all from eBiosciences, Frankfurt, Germany) and the FoxP3 staining buffer kit (eBiosciences) according to manufacturer's protocol. Cells were analyzed by flow cytometry and the percentage of live cells producing IL-10 or IFN-γ was determined with Cell-Quest Pro. B cells isolated from IL-10-GFP tiger mice were activated with α-IgM/CD40 or LPS and cultured in 96-well plates for 2 or 4 days. Ifenprodil (10 and 20 μM) was added at 21-25 h and cells were harvested either at day 2 or day 4. Before harvest, cells were re-stimulated with PMA (100 ng/ml) and IO (800 ng/ml) and monensin (10 μg/ml) for 4 h. IL-10-GFP expression was analyzed on gated live cells with flow cytometry.

Statistical analysis
Data are given as mean values ± standard deviation (SD). Student's t test was used to determine statistical significances, with *p < 0.05, **p < 0.01 and ***p < 0.001.