Extracellular calcium sensing promotes human B-cell activation and function

Increased intracellular calcium (Ca_i²⁺) characterizes many cellular responses to external stimuli,^1-3  such as mitogens and receptor ligands.^4,5  These changes in Ca_i²⁺ propagate primary signals from a receptor and arise from internal Ca²⁺ stores or an influx of extracellular calcium (Ca_o²⁺) through voltage-gated or Ca²⁺ release–activated Ca²⁺ channels.^6-8  Along with its role as a second messenger, Ca²⁺ has recently been found to be a receptor ligand. For example, the 36pS cation channel, on hippocampal neurons, elicits action potentials in response to small changes in Ca_o²⁺.⁹  Many cell types respond to Ca_o²⁺ via the G-protein–linked calcium-sensing receptor (CaSR),¹⁰  including cells from endocrine organs,¹¹  bone (osteoblasts and osteoclasts),¹⁰  brain,¹⁰  intestine,¹⁰  kidney,¹⁰  skin,¹²  ovary,¹³  as well as fibroblasts.¹⁴  CaSR is also expressed by normal hematopoietic cells (including red cell and platelet precursors, monocytes, macrophages,¹⁵  and hematopoietic stem cells¹⁶ ) and myeloma cells and cell lines.¹⁷  Despite the evidence for a CaSR on malignant B cells, normal lymphocytes are not known to express CaSR,^15,17  and it is not clear if B cells can respond directly to Ca_o²⁺

In the course of studying human B-cell responses to immunomodulatory agents, we observed that Ca²⁺-containing culture medium was sufficient to cause the release of Ca_i²⁺ stores in B cells. Subsequent experiments suggested that this response to changes in extracellular calcium was mediated by a receptor that is not the previously identified CaSR.

Phosphate-buffered solution (PBS), Hanks balanced salt solution (HBSS), α- and D-modified Eagle medium (αMEM, DMEM), DMEM/F12, RPMI-1640, and fetal bovine serum (FBS), as well as essential and nonessential amino acids, 0.25% trypsin and glutamine, were purchased from Wisent Bioproducts (St-Bruno, QC). AIM-V medium and the calcium indicator, Indo-1 acetoxymethyl (AM), were obtained from Invitrogen Life Technologies (Burlington, ON). Recombinant human albumin (rHA) was from GTC Biotherapeutics (Framingham, MA). Twenty-five percent human serum albumin (HSA) and immunoglobulin (Ig) (Bayer, Toronto, ON), thrombin (Instrumentation Laboratory, Lexington, MA), heparin (Leo Pharma, Thornhill, ON), human insulin (Eli Lilly Canada, Toronto, ON), chlorpromazine (Bioniche Pharma, Pointe Claire, QC), and cyclosporine A (CsA; Sandoz Canada, Dorval, QC), as well as interleukin-2 (IL-2; Chiron, Emeryville, CA), were purchased from the hospital pharmacy. 2-Mercaptoethanol, CaCl₂, MgCl₂, ZnCl₂, NiSO₄, CoCl₂, BaCl₂, thapsigargin, pyruvate, vitronectin, superfibronectin, chondroitin sulfate, hyaluronic acid, insulin growth factor I, actinomycin D, cycloheximide, NPS2390, verapamil, suramin, guanosine 5′-[β-thio]diphosphate, phorbol 12-myristate 13-acetate (PMA), dichlorofluorescein diacetate (DCFH), and β-actin antibodies were purchased from Sigma-Aldrich (St Louis, MO). Caffeic acid phenethyl ester (CAPE), Bay 11–7082, wedelolactone, PD98059, Ly294002, Gö6976, KN-93, KN-92, SB203580, 1,2-bis(o-aminophenoxy)ethane tetraacetate/acetoxymethyl (BAPTA/AM), PP2, SKF96365, econazole, Syk inhibitor, U73122, and U73343 were from Calbiochem (San Diego, CA). F(ab′)₂ goat anti–human IgM was from Jackson Immunoresearch (West Grove, PA). Ethylenediaminetetraacetic acid (EDTA) and genistein were purchased from BioShop (Burlington, ON). Rat brain C6 cells, the Toll-like receptor 7 (TLR-7) agonist, S28690,^18,19  and serum amyloid protein were kindly provided by R. Kerbel (University of Toronto, Toronto, ON), R. Miller (3M Pharmaceuticals, St Paul, MN), and R. Gomer (Rice University, Houston, TX), respectively.

Heparinized blood (30 mL) was collected from consenting healthy volunteers with approval from the Research Ethics Board, Sunnybrook Health Sciences Center. B cells and monocytes were isolated by negative selection (RosetteSep Human B Cell Enrichment Cocktail and RosetteSep Human Monocyte Enrichment Cocktail; StemCell Technologies, Vancouver, BC) as described previously.²⁰ 

Isolated peripheral blood mononuclear cells (PBMCs) or B cells (1.5 × 10⁶ cells/mL) were cultured in the indicated serum-free medium containing 2-mercaptoethanol (5 × 10⁻⁵ M) in 24-well plates or 5-mL round-bottom tubes (BD Labware, Franklin Lakes, NJ) at 37°C in 5% CO₂ for 4 hours, unless indicated otherwise in the figure legends. Signal transduction inhibitors, plasma, albumin, amino acids, pyruvate, glucose, inorganic salts, PMA, and S28690 were added at the indicated concentrations at the beginning of the culture period. Cells were trypsinized by incubation with 0.25% trypsin at 37°C for 30 minutes prior to being washed in PBS and analyzed.

Cells were blocked with human intravenous gamma globulin (Bayer) for 5 minutes prior to being stained for surface immunophenotyping with preoptimized volumes of CD54, CD62L, CD69, CD80, CD83, CD86, 41BBL, and CD19, CD14, or CD3 (or matched isotype controls) (BD Pharmingen, Missisauga, ON) for 20 minutes at room temperature. Cells were then washed and resuspended in PBS and 0.25 μg/sample 7-amino-actinomycin D (7AAD) (BD Pharmingen) prior to flow cytometric analysis. Staining of viable nucleated cells was determined by gating on forward scatter properties and 7AAD⁻ cells. Reactive oxygen species (ROSs) in B cells were analyzed by staining with 5 μM DCFH²¹  and CD19 antibodies for 20 minutes at 37°C, followed by 2 washings with PBS. Ten thousand events were collected per sample with a FACSCalibur (Becton Dickinson, San Jose, CA) and analyzed using FlowJo software (Tree Star, Ashland, OR). Standardization of the flow cytometer was performed before each experiment using Sphero Calibration Particles (BD Pharmingen).

Purified B cells and monocytes were cultured in AIM-V and rat brain C6 cells were cultured in RPMI with 10% FBS, with or without the indicated inhibitors for 1 hour. Cells were then harvested and washed in PBS. Total RNA was extracted using the RNeasy kit (Qiagen, Missisauga, ON) according to the manufacturer's instructions. To remove contaminating genomic DNA, 10 μg of the total RNA preparation was incubated with 10 U RNase-free DNase I (Promega, Madison, WI) for 30 minutes at 37°C. Total RNA concentration was determined in a spectrophotometer at 260 nm.

cDNA was synthesized as described previously²⁰  or using a one-step reverse-transcription–polymerase chain reaction (RT-PCR) system (Invitrogen Life Technologies) according to the manufacturer's instructions. The following primers were used: CD83 forward, 5′-GTTATTGGAGGGTGGTGAAGAGAGG-3′; CD83 reverse, 5′-GTGAGGAGTCACTAGCCCTAAATGC-3′; CaSR forward, 5′-GGCTCCATCGTGTTTAAGGA-3′; CaSR reverse, 5′-CTCCGTCCACGACAGAAACT-3′; hypoxanthine-guanine phosphoribosyltransferase (HPRT) forward, 5′-GAGGATTTGGAAAGGGTGTT-3′; HPRT reverse, 5′-ACAATAGCTCTTCAGTCTGA-3′; glyceraldehyde-3-phosphate dehydrogenase (GAPDH) forward, 5′-GAAGGTGAAGGTCGGAGTC-3′; and GAPDH reverse, 5′-GAAGATGGTGATGGGATTTC-3′. These primer sets yield RT-PCR products of 448 bp, 352 bp, 241 bp, and 227 bp for CD83, CaSR, HPRT, and GAPDH, respectively. PCR reactions for CD83 and GAPDH were performed using 3 μL cDNA in the presence of 50 mM KCl, 20 mM Tris-HCl (pH 8.4), 1.5 mM MgCl₂, 0.2 mM dNTP, 1 μM of each primer, and 1 U Taq DNA polymerase (Invitrogen Life Technologies) in a total volume of 50 μL using a Flexigene Thermal Cycler (Techne, Princeton, NJ). Reactions were cycled 35 times after initial denaturation (95°C, 15 minutes) with the following parameters: denaturation at 90°C for 30 seconds, annealing of primers at 60°C for 30 seconds, and extension at 72°C for 30 seconds. Using the one-step RT-PCR kit, reverse transcription of 2 μg CaSR or HPRT RNA was performed at 50°C for 30 minutes and PCR reactions were cycled 40 times with an annealing temperature of 55°C. Final PCR products were electrophoresed on a 1.5% agarose gel.

PBMCs or C6 cells were incubated with Indo-1 AM (5 μM) in serum-free, Ca²⁺-free PBS (unless indicated otherwise) for 30 minutes at 37°C. CD19 antibodies were then added to PBMCs and the cells were incubated for an additional 15 minutes at room temperature. Cells were then washed and resuspended at 1 × 10⁶ cells/mL in Ca²⁺-free PBS.

In response to the indicated stimulus, changes in the concentrations of free Ca_i²⁺ in viable Indo-1 AM–labeled cells maintained at 37°C (using a water bath and tube jacket) were analyzed at rates of approximately 500 events/second using a fixed alignment, 18-color, 4-laser LSRII flow cytometer (Becton Dickinson). The baseline ratio of free Ca²⁺/bound Ca²⁺ was established over the first 45 seconds, after which the tube was removed, the stimulant added, and the tube replaced. Recording continued for up to 400 seconds. The data acquired were analyzed using FlowJo.

Proteins were extracted and immunoblotting with antibodies against phosphorylated signal transducer and activator of transcription 1 (STAT1), STAT3, inhibitor of nuclear factor κB (NF-κB) (IκB), AKT, extracellular signal-regulated kinase (ERK), p38 mitogen-activated protein kinase (MAPK), Ca²⁺/calmodulin (CaM) kinase II (CaMKII), serine-phosphorylated PKC substrates, phosphorylated tyrosine, and total IκB (Cell Signaling Technology, Danvers, MA), as well as CaSR (Abcam, Cambridge, MA), was performed as previously described.²⁰  Supersignal West Pico Luminal Enhancer and Stable Peroxide Solution (Pierce, Rockford, IL) created the chemiluminescent signal that was detected using a Syngene InGenius system (Syngene, Cambridge, United Kingdom).

Isolated B cells were cultured in PBS with or without 250 μM or 500 μM CaCl₂ for 4 hours prior to being washed and resuspended in AIM-V with or without IL-2 and S28690 for 24 hours. Supernatants were then collected and tumor necrosis factor α (TNF-α) secretion was analyzed by Cytometric Bead Array (BD Pharmingen) according to the manufacturer's instructions.

Paired Student t tests were used to determine P values for differences between sample means.

While measuring Ca_i²⁺ levels in B cells exposed to a number of immunomodulatory agents, we observed that Ca²⁺-containing media alone caused rapid increases in Ca_i²⁺ (Figure 1A). Addition of CaCl₂ to cells in Ca²⁺-free PBS, or to PBS containing various amounts of CaCl₂, produced similar increases in free Ca_i²⁺ concentrations (Figure 1B) (which could not be replicated by other cations, such as Mg²⁺ [Figure 1C], Zn²⁺, Ni²⁺, Co²⁺, or Ba²⁺ [data not shown], or by stimuli known to signal in a Ca²⁺-independent fashion, such as the TLR7 agonist S28690²²  [Figure 1D]). Of concern was that these observations were due simply to toxicity of Ca_o^2+23,24  causing a calcium leak across damaged plasma membranes or that internal calcium stores had been depleted resulting in open Ca²⁺ release–activated Ca²⁺ channels.²⁵  However, very little cell death was seen when cells were cultured in Ca²⁺-containing media, such as AIM-V, for several days.²⁰  Moreover, stimulation of B cells bathed in Ca²⁺-free or Ca²⁺-containing medium for short periods of time (30-45 minutes) with thapsigargin resulted in Ca_i²⁺ release (4.2 ± 0.6-fold increases in the Indo violet/Indo blue ratio upon thapsigargin treatment [n = 20] compared with 2.3 ± 0.5-fold increases upon CaCl₂ treatment [n = 3]), indicating that internal calcium stores were intact. In addition, the patterns of Ca_i²⁺ changes in Figure 1A-D were more consistent with the rapid, biphasic changes caused by activation of a cell surface receptor, rather than influx through open channels on the plasma membrane.²⁶  Note that addition of 500 μM CaCl₂ to cells bathed in nontoxic levels of calcium (50 or 500 μM) caused the release of Ca_i²⁺ stores (Figure 1B), suggesting that B cells sensed changes in Ca_o²⁺ and not merely the presence or absence of the ion.

Since calcium channel inhibitors such as SKF96365,²⁷  econazole,²⁷  and verapamil²⁸  could not block release of Ca_i²⁺ stores caused by Ca_o²⁺ (data not shown), other signal transduction inhibitors (including G-protein, tyrosine kinase, and phospholipase Cγ [PLCγ] inhibitors) were tested for their ability to block responses to Ca_o²⁺. Strikingly, tyrosine kinase (genistein) and PLCγ (U73122) inhibitors abrogated Ca_i²⁺ release (Figure 2A), suggesting that the response to Ca_o²⁺ may involve a signaling cascade from a Ca²⁺-activated extracellular receptor or receptors. Consistent with this, Ca_o²⁺-induced tyrosine phosphorylation was inhibited by genistein (Figure 2B lanes 2-3 top panel), and PKC substrate phosphorylation (which correlates with PLC activity) was abrogated by U73122 (Figure 2B lanes 2,4 middle panel). Ca_o²⁺-induced Ca_i²⁺ release was also inhibited when cells were pretreated with trypsin (further suggesting that the response was mediated by a cell surface protein or proteins) (Figure 2C). In addition, PMA inhibited Ca_o²⁺-induced Ca_i²⁺ release, suggesting negative regulation by activated PKC^29-31  (Figure 2C).

The Ca_o²⁺-sensing mechanism in B cells was found to be independent of CaSR or other G-protein–coupled receptors (GPCRs). In contrast to rat brain C6 cells or monocytes, B cells did not express CaSR mRNA (Figure 2D) or protein (Figure 2E). The CaSR inhibitor, NPS2390,¹¹  did not inhibit Ca_o²⁺-induced Ca_i²⁺ release in B cells, whereas it was able to block Ca_o²⁺-induced Ca_i²⁺ release in C6 cells that expressed the CaSR (Figure 2F). In addition, Ca_o²⁺-induced Ca_i²⁺ release was not abrogated by the G-protein inhibitors suramin and guanosine 5′-[β-thio]diphosphate (data not shown), further arguing against a role for the CaSR (which is a GPCR) in B-cell Ca_o²⁺ sensing.

An alternative explanation for the apparent activation of a specific calcium-sensing receptor on B cells could be that the 3-dimensional conformations of other cell surface receptors or proteins were altered by binding calcium, either causing signal transduction directly or disrupting the binding of negative regulatory molecules and producing signal transduction and Ca_i²⁺ release. Candidate receptors on B cells that might cause Ca_i²⁺ release in this manner upon binding Ca_o²⁺ include the B-cell receptor (BCR),³²  TLRs,³³  and integrins.³⁴  As shown in Figure 1D, the activation of at least one TLR (TLR7) on B cells did not lead to the release of Ca_i²⁺ stores, suggesting that TLR modification by Ca_o²⁺ was not causing Ca_o²⁺-induced Ca_i²⁺ release. Similarly, BCR cross-linking is known to induce the release of Ca_i²⁺ stores (Figure 2G) via the activation of PLCγ through Syk.³²  However, a Syk inhibitor,³⁵  which is otherwise capable of inhibiting BCR cross-linking–induced Ca_i²⁺ release, failed to prevent Ca_o²⁺-induced Ca_i²⁺ release (Figure 2G), suggesting that the binding of Ca_o²⁺ did not simply increase tonic BCR signaling.³⁶  Other candidate receptors, such as integrins, which signal through Src,³⁴  were excluded, since the Src inhibitor PP2 did not affect Ca_o²⁺-induced Ca_i²⁺ release (data not shown). It remains possible that other receptors on B cells are conformationally altered and activated by binding Ca_o²⁺, but the results are consistent with the presence of a specific calcium receptor on B cells, which is not the CaSR.

To determine whether Ca_o²⁺-induced Ca_i²⁺ release caused phenotypic changes in B cells, the expression of a number of surface molecules was studied by flow cytometry. Immediately after isolation, B cells expressed low surface levels of CD54, CD62L, CD69, CD80, CD83, and CD86. Upon exposure to Ca_o²⁺, high CD83 surface expression was noted within 4 hours (Figure 3A bottom panels). The percentage of B cells exhibiting CD83 at 4 hours ranged from 57.5% to 93.8% (mean of 74.2%; n = 30). In contrast to B cells, CD83 did not increase on CD3⁺ T cells or CD14⁺ monocytes cultured under similar conditions (Figure 3A top and middle panels). Expression of the activation marker CD69 also increased on B cells, while CD54, CD62L, CD80, or CD86 did not change (data not shown).

A number of commercially available tissue culture media were studied to determine whether Ca_o²⁺-mediated CD83 expression required factors other than Ca²⁺. Media containing various levels of Ca²⁺, Mg²⁺, Na⁺, K⁺, amino acids, vitamins, glutamine, glucose, pyruvate, and phenol red (AIM-V, DMEM, αMEM, DMEM + F12, and RPMI-1640) all induced high expression of CD83 on the cell surface, whereas media containing little or none of these components (Ca²⁺/Mg²⁺-free PBS and HBSS) did not (Figure 3B). Further testing revealed that only the presence of Ca²⁺ (which was supplied concomitantly with Mg²⁺ in the media) was associated with increased CD83 expression, whereas phenol red, amino acids, pyruvate, and glucose appeared to have little or no effect on CD83 (Table 1).

To provide further evidence for a role of Ca_o²⁺-induced Ca_i²⁺ release in changing CD83 expression, the effects of RPMI ± Ca²⁺, EDTA, and the Ca_i²⁺ chelator BAPTA/AM were tested. Only RPMI containing Ca²⁺ induced CD83 expression, whereas Ca²⁺ chelation with EDTA and BAPTA/AM abrogated CD83 expression (Figure 3C). As a control, the TLR7 agonist S28690 (which signals through calcium-independent pathways²²  [Figures 1D and 7]) induced strong CD83 expression on B cells in the presence of EDTA, suggesting that the effects of EDTA on Ca²⁺-induced CD83 expression were specifically due to chelation of Ca_o²⁺ (Figure 3D). Ca_o²⁺ induced CD83 expression in a dose-dependent manner, whereas MgCl₂ did not (Figure 3E). Taken together, these results suggested that exposure to increased levels of Ca_o²⁺ was responsible for the release of Ca_i²⁺ and the expression of CD83 on the B-cell surface.

To determine whether the increased expression of CD83 on B cells caused by Ca_o²⁺ was transcriptionally regulated, CD83 mRNA levels were measured by RT-PCR. CD83 message was expressed within 30 minutes and peaked after 1 hour of exposure to Ca_o²⁺ (Figure 4A).

CD83 message was abrogated by the transcriptional inhibitor, actinomycin D, but not the translational inhibitor, cycloheximide (Figure 4Bi). Cell surface expression of CD83 protein was also blocked by actinomycin D (Figure 4Bii), suggesting that the spontaneous increase in CD83 was due to de novo RNA transcription.

In striking contrast to the observations in serum-free media, CD83 expression did not increase when cells were cultured in the presence of 10% FBS (data not shown) or 30% plasma (Figure 4C). A number of serum and plasma components were tested to determine the factor responsible for preventing the Ca_o²⁺-induced increase in CD83 expression. Heparin, thrombin, vitronectin, superfibronectin, chondroitin sulfate, hyaluronic acid, insulin, insulin growth factor I, Ig, and serum amyloid protein did not inhibit Ca_o²⁺-induced CD83 expression (data not shown). However, albumin (both HSA and rHA) strongly abrogated CD83 message and surface protein expression in AIM-V (Figure 4C) and PBS containing Ca²⁺ (data not shown), consistent with its well-known ability to bind and buffer Ca_o²⁺.^37-39 

Signaling pathways connecting the putative Ca_o²⁺-sensing receptor on B cells to subsequent transcription of CD83 were then investigated. The CD83 promoter contains 4 SP1-binding sites, and one NF-κB element.⁴⁰  The NF-κB inhibitors CAPE, Bay 11-7082, and wedelolactone^41-43  inhibited Ca_o²⁺-induced expression of CD83 mRNA (Figure 5Ai) and protein (Figure 5Aii), in contrast to the SP1 inhibitor, mithramycin^44,45  (data not shown), suggesting that Ca_o²⁺ sensing was mediated by NF-κB.

In the canonical NF-κB pathway,⁴⁶  components of the IκB kinase (IKK) complex are phosphorylated and activated by a number of kinases. Activated IKK then phosphorylates IκB, causing it to be ubiquinated and degraded by the proteasome, releasing NF-κB dimers that then translocate to the nucleus to act as transcription factors. Analysis of several signaling pathways (by immunoblotting for phosphorylated intermediates) revealed that AKT, ERK, p38 MAPK, PKC, and CaMKII all became strongly activated prior to the phosphorylation and degradation of IκB (Figure 5B). CaM and calcineurin (CN) appeared to be especially important mediators as CD83 expression was strongly inhibited by the CaM inhibitor, chlorpromazine, and the CN inhibitor, CsA (Figure 5C). Furthermore, the inhibitory effects were additive, suggesting that CaM and CN acted in parallel. Inhibition of CaMKII by KN-93 completely abrogated CD83 expression (Figure 5C). Inhibitors of the classical PKC isozymes (Gö6976), ERK (PD98059), p38 MAPK (SB203580), and PI3K (Ly294002) also decreased Ca_o²⁺-mediated CD83 expression, although not to the same extent as chlorpromazine and CsA, or KN-93 (Figure 5C). Jun N-terminal kinase, mammalian target of rapamycin, and Bruton tyrosine kinase were found to not be significantly involved in controlling Ca_o²⁺-mediated CD83 expression (data not shown). By using these inhibitors in combination, CaM was found to signal in parallel with PI3K, ERK, and PKC, as additive effects were noted on the inhibition of CD83 expression (Figure 5D). PI3K and ERK signaled sequentially (since the effects of Ly294002 and PD98059 were the same as PD98059 alone), as did PKC and ERK (since the effects of Gö6976 and PD98059 were the same as Gö6976 alone) (Figure 5E). Furthermore, PKC was found to signal upstream of AKT (since Gö6976 inhibited phosphorylation of AKT [Figure 5F: compare lanes 2 and 4, 3rd panel from top]), PKC and AKT were upstream of ERK (since Gö6976 and Ly294002 inhibited phosphorylation of ERK [Figure 5F: compare lanes 2-4, 4th panel from top]), and CaMKII had a significant effect on ERK activation (Figure 5F: compare lanes 2 and 6, 4th panel from top).

In addition to increasing CD83 expression, B cells engage in immune responses by becoming activated through receptors (such as for antigen, cytokines, and pathogen-associated molecular patterns⁴⁷ ) and altering gene expression to produce factors such as cell surface proteins and cytokines. Accordingly, the effects of exposing B cells to increased levels of Ca_o²⁺ on some of these functions were examined.

Activation of receptors on B cells leads to signal transduction that culminates in gene transcription and also mitochondrial activation to support the energy-dependent activated state.⁴⁸  Activated mitochondria make ROSs, which can be measured by flow cytometry with dyes such as DCFH.²¹  Accordingly, the magnitude of DCFH staining early after activation of a receptor can be taken as a measure of the strength of signaling through the receptor. As shown in Figure 6A, B cells exposed to Ca_o²⁺ for 4 hours exhibited higher baseline levels of ROSs and much higher increases in ROS production in response to BCR signaling (mimicked with Ig cross-linking antibodies) or a combination of IL-2 and the TLR7 agonist, S28690 (shown previously to cause strong cytokine production by human B cells).⁴⁷  These results suggested that pulses of Ca_o²⁺ may enhance subsequent responses of B cells to immunogenic signals.

Consistent with increased responsiveness to immunogenic signals, changes in expression of the costimulatory molecules CD86 and 41BBL, 4 hours after stimulation with either Ig cross-linking antibodies or IL-2 and S28690, were higher if the cells were first exposed to Ca_o²⁺ (Figure 6B). Strikingly, despite having no effect on cell numbers 24 hours later, this brief exposure to Ca_o²⁺ was sufficient to significantly enhance phosphorylated STAT1 and STAT3 levels in response to IL-2 and S28690 (Figure 6C), which has been shown to reflect the magnitude of cytokine production.^19,47  Consistent with this, TNF-α levels in the culture supernatants were increased by the brief exposure to Ca_o²⁺ (Figure 6D).

The observations in this paper suggest that increases in Ca_o²⁺ cause Ca²⁺ to be released from intracellular stores in human B cells (Figure 1) and activate signaling cascades that involve PLCγ, CaM, CaMKII, CN, PKC, PI3K, ERK, and NF-κB (Figure 5). This response to Ca_o²⁺ results in increased CD83 expression (Figure 3) and responsiveness to cytokine, TLR, and BCR signals as measured by ROS production, costimulatory molecule expression, and cytokine production (Figure 6). Taken together, these results suggest the presence of a receptor-mediated Ca_o²⁺-sensing mechanism on B cells that may regulate their behavior in vivo.

CD83 has been shown to be rapidly and spontaneously expressed on adherent monocytes⁴⁹  cultured in serum-containing medium. However, this “spontaneous” CD83 expression on monocytes is controlled by posttranscriptional mechanisms,⁴⁹  in contrast to our findings that “spontaneous” (ie, Ca_o²⁺-induced) CD83 up-regulation on B cells was transcriptionally regulated (Figure 4).

Ca_o²⁺ induced a biphasic Ca_i²⁺ response, marked by a Ca_i²⁺ transient followed by a sustained plateau (Figures 1–2), which is typical of receptor-mediated increases in Ca_i²⁺.⁵⁰  The putative calcium-sensing receptor on B cells does not seem to be the well-known CaSR found on many other cell types¹⁰  since CaSR was not expressed by B cells (Figure 2D,E) and its specific inhibitor (NPS2390)¹¹  did not block Ca_o²⁺-induced Ca_i²⁺ release or CD83 expression (Figure 2F and data not shown). Moreover, Mg²⁺, Zn²⁺, Co²⁺, and Ni²⁺ could not substitute for Ca_o²⁺ (Figures 1C, 3E and data not shown), providing further evidence that Ca²⁺ was not signaling through the CaSR or via calcium channels that can be activated by other cations.⁵¹  The role of calcium channels in Ca_o²⁺ sensing in B cells was further excluded as channel inhibitors, such as SKF96365, econazole, and verapamil, did not inhibit Ca_o²⁺-induced Ca_i²⁺ release or CD83 expression (data not shown). In addition, the presence of equimolar concentrations of BaCl₂ in lieu of CaCl₂ did not induce changes in the fluorescence of Indo-1 AM, further excluding a role for calcium channels in B-cell calcium sensing (as Ba²⁺ is known to enter cells via calcium channels and bind Indo-1 AM)^25,52  (data not shown). However, increases in free Ca_i²⁺ concentration and CD83 expression were inhibited by tyrosine kinase and PLCγ inhibitors (Figure 2A and data not shown), and Ca_i²⁺ release was abrogated when cells were pretreated with trypsin (which would cleave extracellular portions of cell surface receptors) or with PMA (which inhibits signaling through some cell surface receptors)^29-31  (Figure 2C), suggesting that a receptor on B cells was involved in sensing Ca_o²⁺.

While the biochemical properties of this putative B-cell calcium receptor have not yet been characterized, we propose a model (Figure 7) whereby Ca_o²⁺ binds to a receptor, inducing the activation of PLCγ and PI3K, presumably resulting in the production of diacylglycerol and inositol triphosphate, and release of Ca_i²⁺ from internal stores. These secondary messengers then activate PKC, AKT, and CaM in parallel, resulting in the activation of CaMKII or ERK, all of which lead to phosphorylation and degradation of IκB, activation of NF-κB, and transcription of CD83. p38 MAPK also plays a role in controlling the expression of CD83 (Figure 5B,C).⁵³  Consistent with this model, AKT has been suggested to stimulate NF-κB transcriptional activity through direct mechanisms involving IKK and NF-κB nuclear translocation,^54,55  as well as by phosphorylating the transactivation domain of p65.^56,57  Recent findings also show that AKT can be phosphorylated in a PKC-dependent fashion, thereby modulating its activity.⁵⁸ 

The physiological relevance of this response of B cells to Ca_o²⁺ in vitro is not yet clear. Why should B cells have a receptor designed specifically to sense the presence of Ca_o²⁺? At present, answers to this question are necessarily speculative. Free ionized calcium levels are tightly controlled in vivo, although they vary considerably from approximately 1 mM in serum and less in interstitial fluids,^59-61  to as high 10 to 15 mM in the epidermis⁶²  and 40 mM near resorbing osteoclasts in bone.¹⁶  It may be that the absolute levels of Ca_o²⁺ are not as important to B cells as sudden changes in concentration, which may reflect tissue damage and the need to activate immune defenses. For example, calcium is known to be released by injured and/or dying cells at sites of inflammation, producing calcium concentrations 1.5- to 3-fold higher than in serum.^63-65  Ca_o²⁺ levels may also rise near B cells when neighboring cells are activated (for example, T cells stimulated by antigen in a lymph node). When activated cells release Ca²⁺ from their internal stores, they subsequently pump cytosolic calcium into the extracellular environment to restore homeostatic Ca_i²⁺ levels (20-100 nM).^59,61  These extracellular calcium fluctuations might be sensed by nearby B cells and prepare them to interact with T cells. In other words, local changes in calcium (which were mimicked in vitro by adding Ca²⁺ to PBS or tissue culture media [Figure 1B] in the absence of buffers such as albumin [Figure 4C]) may be interpreted as danger signals that cause B cells to increase their ability to effectively present antigen (through enhanced expression of CD83, CD86, and 41BBL)²⁰  (Figures 3,6) and make cytokines (Figure 6).

The findings in this paper suggest that Ca_o²⁺ may have a previously uncharacterized role in B-cell responses, in addition to its well-known function as a second messenger. Further studies of this putative calcium-sensing mechanism may lead to a better understanding of B-cell physiology and pathology.

This work was supported by grants from the Ontario Cancer Research Network (OCRN), Canadian Institutes of Health Research (CIHR), and a program project grant from the National Cancer Institute of Canada (NCIC) with funds from the Terry Fox Foundation (D.E.S.), and studentships from the Natural Sciences and Engineering Research Council of Canada (NSERC) (C.M.H.) and NCIC (J.T.).

We thank Drs R. Miller, R. Gomer, and R. Kerbel for S28690, serum amyloid protein, and C6 cells, respectively.

Contribution: C.M.H. designed and performed the research, analyzed the data, and wrote the paper; D.W., J.T., and Y.S. performed research; D.E.S. designed the research and wrote the paper.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: David Spaner, Division of Molecular and Cellular Biology, Sunnybrook Research Institute, S-116A, Research Bldg, Sunnybrook Health Sciences Centre, 2075 Bayview Ave, Toronto, ON, M4N 3M5, Canada; e-mail: david.spaner@swri.ca.