Ca²⁺ Waves Initiate Antigen-Stimulated Ca²⁺ Responses in Mast Cells

Abstract

Ca²⁺ mobilization is central to many cellular processes, including stimulated exocytosis and cytokine production in mast cells. Using single cell stimulation by IgE-specific antigen and high speed imaging of conventional or genetically-encoded Ca²⁺ sensors in RBL and bone marrow-derived rat mast cells, we observe Ca²⁺ waves that originate most frequently from the tips of extended cell protrusions, as well as Ca²⁺ oscillations throughout the cell that usually follow the initiating Ca²⁺ wave. In contrast, antigen conjugated to the tip of a micropipette stimulates local, repetitive Ca²⁺ puffs at the region of cell contact. Initiating Ca²⁺ waves are observed in most RBL cells stimulated with soluble antigen, are sensitive to inhibitors of Ca²⁺ release from ER stores and to extracellular Ca²⁺, but they do not depend on store-operated Ca²⁺ entry. Knockdown of TRPC1 and TRPC3 channel proteins by shRNA reduces the sensitivity of these cells to antigen and shifts the wave initiation site from protrusions to the cell body. Our results reveal spatially encoded Ca²⁺ signaling in response to immunoreceptor activation that utilizes TRPC channels to specify the initiation site of the Ca²⁺ response.

INTRODUCTION

Changes in intracellular Ca²⁺ play significant roles in numerous cellular responses such as secretion, gene expression, and cell migration. These cellular functions require spatial and temporal regulation of cytosolic Ca²⁺ (1, 2). Among these regulated events are Ca²⁺ puffs, waves, and regenerative oscillations that mediate localized cellular responses and support transfer of information across the cell and organelles (3).

Ca²⁺ waves were first characterized in Xenopus oocyte fertilization (4), and they have since been identified in excitable (5) and nonexcitable cell types, including hepatocytes (6), HeLa cells (7), and neutrophils (8). In myocytes, Ca²⁺ waves were shown to initiate from elementary Ca²⁺ events called “Ca²⁺ sparks” (9), and are thought to propagate through the cytosol by calcium-induced calcium release from ER stores (10). Ca²⁺ waves are frequently initiated by activation of plasma membrane receptors that stimulate Ca²⁺-dependent signaling within the cell (11). Similar mechanisms may be involved in stimulating Ca²⁺ puffs and maintaining the propagation of Ca²⁺ waves in non-excitable cells (3, 12). However, Ca²⁺ waves in response to immunoreceptor signaling have not been previously reported.

Ca²⁺ oscillations have been characterized in many cell types, including RBL-2H3 mast cells (13, 14) where they have been temporally correlated with degranulation events (15, 16). Ca²⁺ oscillations are sustained by store-operated Ca²⁺ entry (SOCE), other ion channels, as well as membrane potential (17).

Mast cells play key roles in the inflammatory process in both innate and adaptive immune responses (18). In the latter, binding of multivalent antigen to receptor-associated IgE aggregates this receptor, FcεRI, which causes mast cell activation, resulting in Ca²⁺ mobilization and consequent exocytotic release of mediators of allergy and inflammation (19). RBL-2H3 cells are immortalized mucosal mast cells that have been utilized for extensive biochemical and cell biological investigations of mast cell function (20-22).

In the present study, we used high-speed confocal imaging to investigate cytoplasmic Ca²⁺ dynamics activated via FcεRI in RBL cells and in rat bone marrow-derived mast cells (BMMCs), which are also mucosal in character (23). We find that Ca²⁺ responses to soluble antigen initiate in the form of a wave that begins most frequently at the tip of an extended cell protrusion and propagates throughout the entire cell in several seconds. In contrast, localized delivery of antigen attached to the tip of a micropipette results in repetitive, localized Ca²⁺ puffs that infrequently develop into propagated waves. Our results provide evidence that Ca²⁺ wave initiation from extended protrusions depends on Ca²⁺ influx via TRPC channels leading to the onset of SOCE-dependent Ca²⁺ oscillations and mast cell activation.

MATERIALS AND METHODS

cDNA plasmids

The GCaMP2 construct (24) was provided by Dr. M. Kotlikoff, Cornell University College of Veterinary Medicine. shRNA plasmids targeting TRPC channels (TRPC1, TRPC3, TRPC5, TRPC7 and GFP control) were characterized in RBL cells as previously described (25).

Chemicals and reagents

Fluo4AM and Fluo5FAM were purchased from Invitrogen/Molecular Probes (Eugene, OR). U73122, D-sphingosine, thapsigargin, A23187, 2-aminoethyldiphenyl borate (2-APB), and GdCl₃ were from Sigma-Aldrich (St. Louis, MO). N,N’-dimethylsphingosine (DM-sphingosine) was from Avanti Polar Lipids (Alabaster, AL).

Cells

RBL-2H3 cells (26) were maintained in monolayer culture in Minimum Essential Medium supplemented with 20% fetal bovine serum (Atlanta Biologicals, Norcross, GA, USA) and 10 μg/ml gentamicin sulfate. All tissue culture reagents were obtained from Gibco (Grand Island, NY, USA) unless otherwise noted. Rat BMMCs were differentiated from femur-derived stem cells of Lewis strain rats by culturing for 14-28 days in the presence of rat stem cell factor (50 ng/ml) and rat IL-3 (100ng/ml) as previously described (27). Cells were harvested 3–5 days after passage and plated overnight in MatTek coverslip dishes in complete medium for experiments with Fluo-5F AM loading. For experiments with GCaMP2 as the Ca²⁺ indicator, 5 × 10⁶ RBL-2H3 cells were electroporated in 0.5 ml of cold electroporation buffer (137mM NaCl, 2.7mM KCl, 1mM MgCl₂, 1mg/ml glucose, 20mM Hepes, pH=7.4) with 5-8 μg of plasmid DNA at 280V and 950μF using Gene Pulser X (Bio-Rad), then plated onto MatTek dishes (~500,000 cells/plate). Cells were sensitized with 0.5 μg/ml anti-2,4-dinitrophenyl (DNP) IgE (28) during overnight cell culture incubation. Under these conditions, the cell transfection efficiency with GCaMP2 is typically 20-30%.

For experiments with TRPC shRNA constructs, plasmids containing these sequences (Origene Technologies) were co-electroporated with that for GCaMP2 using 20 μg of TRPC shRNA plasmid and 8 μg of GCaMP2 plasmid for 5 × 10⁶ cells. No effect of these plasmids on cell viability was detected.

For characterization of rat BMMCs, differentiated cells were labeled with 2 μg/ml Alexa488-IgE (29) for 2 hr at 37°C, then washed and fixed in 4% (w/v) p-formaldehyde and 0.1% (w/v) glutaraldehye for 20 min at room temperature. Fixed cells were further labeled in 10 mg/ml BSA/PBS/0.01% NaN₃ with 3 μg/ml monoclonal anti-α-galactosyl GD_1b ganglioside AA4 (30), followed by 10 μg/ml Alexa555-goat anti-muIgG₁ (Invitrogen Corp.). Washed, labeled cells were imaged as previously described (31).

Live cell calcium imaging

24 hours after transfection, cells were washed with buffered salt solution (BSS: 135mM NaCl, 5mM KCl, 1.8mM CaCl₂, 1mM MgCl₂, 1 mg/ml glucose, 20mM Hepes, pH=7.2-7.4, 1mg/ml BSA) and imaged using a heated (37°C) 63X (NA=1.4) oil immersion objective on a Zeiss Meta 510 confocal microscope. GCaMP2, Fluo-4 or Fluo-5F were excited using the 488 nm line of a krypton/argon laser and viewed with a 505–530 band-pass filter. In experiments using Fluo-4 AM or Fluo-5F AM, cells were labeled by incubating with 0.5 μM (RBL cells) or 2.5 μM (rat BMMCs) of the indicator in BSS in the presence of 2.5 mM sulfinpyrazone for 30 min in 37°C (5 min for rat BMMCs), then washed and imaged in BSS.

Cells expressing GCaMP2 or loaded with Fluo dyes were approached with a ~5 μm diameter pulled glass capillary, typically positioned within 100 microns from the cell and pre-filled with stimulating solution. Cells were imaged at 10-30 Hz while applying a puff of 10 seconds/5psi from the pipette. Pressure and duration of stimulation was controlled by a Picospritzer III (FMI Medical Instruments, GmbH). Pharmacological reagents were added to the dishes just prior to initiating data collection (see Results). For some experiments, DNP-BSA (Ag) or unmodified BSA was covalently coupled to the micropipette surface using chemical treatments and procedure previously described for covalent attachment of proteins to patterned surfaces (32).

Data analysis

Images of individual cells are representative from multiple experiments. Off line image analysis was carried out using Zeiss LSM image analysis software (Carl Zeiss, Inc) and ImageJ (NIH). Changes in fluorescence (F) were normalized by the initial fluorescence (F₀) and are expressed as ΔF.

Unless stated otherwise, cells that displayed any type of Ca²⁺ elevation including Ca²⁺ puffs upon stimulation were counted as responding cells. Wave occurrence was calculated as the percentage of cells with a detectable Ca²⁺ wave, compared to the total number of responding cells in a given experimental condition. The lag time before Ca²⁺ wave initiation was measured from the initiation of the pulse of Ag or other stimulus. Statistical analysis of Ca²⁺ oscillations was carried out by counting and averaging the oscillatory peaks during the first 2 minutes of stimulation, including the initiating wave. Wave propagation velocity was measured as the time (number of frames x time per frame) for spreading of the wave-front across a cell of measured dimension. Unless otherwise stated, all data are presented as mean ± SEM. For each of the experiments, the number of cells analyzed (n) is presented in the text. Data were processed and plotted using Origin 8 (OriginLab) and Excel (Microsoft Corp.). Statistical comparisons between experiments were preformed using Students t-test.

RESULTS

Ca²⁺ mobilization events in individual RBL mast cells

To observe the earliest steps in the responses of RBL mast cells to stimulation of IgE receptors by multivalent antigen, we delivered a short (10 sec) pulse of antigen (Ag) via a micropipette placed within 100 μm of the cell while imaging changes in cytosolic Ca²⁺ (Figure 1A). For most experiments, we used the genetically encoded Ca²⁺ indicator, GCaMP2, which has a moderate K_d (~350 nM) with rapidly reversible kinetics (33). GCaMP2 is brighter than wild type GFP, maintains a 5-fold dynamic range, displays very low pH sensitivity and is fully functional at 37°C (24, 33).

Figure 1. Visualization of Ca²⁺ dynamics in individual RBL mast cells.

A) Cells were plated at low density on MatTek dishes, and individual cells were stimulated using a pulled capillary (arrow; 0.5-1 μm diameter), filled with stimulating solution (as specified for each experiment) that was puffed at the cell (5psi/10sec pulse) from a distance within 100 μm. B) Image of RBL cell expressing GCaMP2 and sensitized with anti-DNP IgE, stimulated with a puff of 1.7 μg/ml Ag (DNP-BSA). Middle panel shows time line analysis of the stimulated cell and changes in Ca²⁺ concentration occurring over time in the 7 pixel wide line segment defined in the left panel. Changes in Ca²⁺ concentration are indicated by relative changes in brightness, where brighter colors represent higher Ca²⁺ concentration. Right panel expands early time points of the time line analysis and highlights our general observation that Ca²⁺ elevation initiates in cellular protrusions and propagates through the cell body. C) Traces of GCaMP2 intensity changes over time as calculated for the same cell imaged in B (inset). Black and red traces represent the changes in Ca²⁺ concentration in cell body (yellow ROI) or cell protrusion (red ROI), respectively.

Using high imaging rates of 10-30 frames per second, we observe two major types of spatially and temporally resolved Ca²⁺ mobilization dynamics in individual cells as illustrated in Figure 1B and Supplemental Movie 1. Stimulation of cells with a saturating concentration of Ag (1.7 μg/ml in the pipette) resulted in more than 90% of the cells responding with a Ca²⁺ wave that initiated a global Ca²⁺ response. This initiating Ca²⁺ wave followed a lag period of ~20 seconds, on average, after starting the Ag pulse (Table I). As illustrated in Figure 1B and will be discussed below, the Ca²⁺ wave usually initiates from the tip of an extended cell protrusion and propagates down the length of the cell body. The site of wave initiation does not detectably correlate with the location of the micropipette relative to any particular part of the cell. Ca²⁺ waves also initiated most frequently from the tips of extended protrusions in response to bath addition of antigen, but the lag time for wave initiation was more variable under these conditions (data not shown).

Table I.

Summary of Ca²⁺ Response Parameters with GCaMP2^a

	Condition	Response (% of Cells)	Ca2+ Waves (% of Cells)	Waves from PRTS (% of Waves)	Oscillationsb	wave velocity (μm/sec)	Laq Time (sec)
Ag	1700ng/ml	96.3	90	65	4.5 ± 0.3	42.6 ± 4	19.4 ± 1.25
	170ng/ml	94	87	54	4.9 ± 0.4	37.3 ± 5.1	19.5 ± 2.1
	17ng/ml	90.5	81	62	5 ± 0.4	18.7 ± 2.2	41.8 ± 4.4
	1.7ng/ml	53.8	78	86	3.7 ± 0.9	15.9 ± 1.6	42.6 ± 6.5
	0[Ca2+}o	96.8	89	31	1.5 ± 0.1	25.8 ± 4.5	30.2 ± 4.2
	50mM KCI	91	73	27.2	5.7 ± 0.5	27.8 ± 4.2	30.2 ± 4.9
	2mM Ba2+	64	78	72	1.6 ± 0.3	46.9 ± 3.5	20.9 ± 0.8
	2mM Sr2+	90	89	75	1.7 ± 0.2	44.4 ± 3.6	19.4 ± 1.5
	TG	76.2	43.7	71	1.7 ± 0.2	30.3 ± 16.4	29.1 ± 4
	A23187	73.9	44.8	30.7	2 ± 0.2	19.9 ± 5.3	17.9 ± 2.7
Ag	U73122 1μM	100	72	35	4 ± 0.5	47.2 ± 11	14.8 ± 2.3
	U73122 2μM	100	58	33	4.2 ± 0.4	47.2 ± 7.5	16.9 ± 2.3
	U73122 2.5μM	30	13	-	1 ± 0.2
	U73122 3.3μM	25	16	-	1 ± 0.3
	U73122 5μM	0	-	-	-
	2-APB	95.8	86	84	1.6 ± 0.2	25.9 ± 4	27.5 ± 2.7
	sphingosine	26.9	15	75	4.4 ± 0.8	27.8 ± 4.7	49.8 ± 5.7
	DM-sphingosine	62.1	15	33	2.8 ± 0.5	41.2 ± 9	29.3 ± 6.3
	Nifedipine	91	93	43	4.6 ± 0.3	26.8 ± 3.8	12.8 ± 0.8
	1μM Gd3+	80	87	69	3.7 ± 0.7	40.1 ± 6.2	22.7 ± 2.9
	10μM Gd3+	80.8	89	50	1.8 ± 0.1	21.4 ± 3	20.8 ± 2.3

^aUnless otherwise stated Ag was used at 1700ng/ml; number of cells in each experiment was 16-80 as specified in text or figure legends

^bDuring first two minutes of response

^*Data are mean ± SEM

In most cells the Ca²⁺ wave is followed by a second manifestation of calcium dynamics – regenerative Ca²⁺ oscillations that occur throughout the cell body as previously described for these cells (13, 14, 16). Although not seen in the cell shown in Figure 1B, Ca²⁺ oscillations often extend throughout the length of the cell, including protrusions. Spatial and temporal resolution of these two processes are shown in Figure 1C as time traces for the tip of the protrusion (red trace and red region of interest (ROI) in inset) and the cell body (black trace, yellow ROI in inset) for the cell in Figure 1B. In some cells, we observed localized Ca²⁺ puffs prior to the initiation of waves; however, these were infrequent under conditions of optimal stimulation (see below).

We used Fluo4 and Fluo5F to confirm our main results for these Ca²⁺ dynamics. Fluo4 and GCaMP2 have similar affinities for Ca²⁺, but the fluorescence response kinetics are approximately an order of magnitude slower for the GFP-based sensor. As shown in Supplemental Figure S1, despite this slower response time for GCaMP2, Ca²⁺ wave velocities measured using GCaMP2 were significantly higher than those measured with Fluo4, indicating that GCaMP2 kinetics are not a limiting factor in our measurements. The percentage of cells exhibiting Ca²⁺ waves detected by Fluo4 is smaller, and the observed waves initiate after a longer lag time and exhibit slower propagation velocities. Interestingly, fewer Ca²⁺ oscillations were detected with both Fluo4 and Fluo5F compared to GCaMP2, possibly due to Ca²⁺ buffering effects of these organic dyes, which are difficult to load into cells at controlled concentrations. Overall, we found GCaMP2 to be most useful for our experiments.

Although there are cell-to-cell differences in appearance and characteristics of Ag-stimulated Ca²⁺ waves and subsequent oscillations in the RBL cells, we found that several characteristics are highly consistent for a particular concentration of Ag. These include the lag time to initiation of the first Ca²⁺ peak, the wave velocity, and the frequency (number observed in 2 minutes) of the oscillations (Table I).

Cellular protrusions are the primary initiation sites for Ca²⁺ wave responses to Ag stimulation in RBL cells

As summarized in Figure 2B, ~90% of the cells responded to Ag (1.7 μg/ml) with a propagating Ca²⁺ wave, whereas for the remaining ~5-10% of the responding cells we detected a non-directional, global elevation in cytoplasmic Ca²⁺. Although stimulated influx of extracellular Ca²⁺ is required for cellular degranulation, we observed wave initiation with similar frequency in the presence and absence of extracellular Ca²⁺, suggesting that Ca²⁺ influx is not necessary for Ca²⁺ wave generation and propagation. However, a clear difference in the spatial localization of wave initiation is observed, depending on the presence or absence of extracellular Ca²⁺. As illustrated in Figure 2A and Supplemental Movie 2, and summarized in Figure 2B, wave initiation in the absence of extracellular Ca²⁺ occurs most frequently at a site in the cell body, with less frequent (<30%) initiation in elongated protrusions. This contrasts with extracellular Ca²⁺ at 2-10mM when the majority of waves stimulated by Ag initiate near the tips of these protrusions (Fig. 1 and Fig. 2B). Similarly, reducing the driving force for Ca²⁺ influx by depolarizing the cells with addition of 50 mM KCl prior to Ag stimulation also decreases the incidence of waves initiating from elongated protrusions to <30% (Table I). These results suggest that initiation of Ca²⁺ waves from elongated protrusions depends on Ca²⁺ influx.

Figure 2. Dependence of stimulated Ca²⁺ waves and oscillations on extracellular Ca²⁺and stimulus.

A) Representative cell sensitized with IgE and stimulated with 1.7μg/ml Ag in the absence of extracellular Ca²⁺. Right panel shows time line analysis for early time points and indicated segment in left panel image. B) Histogram showing percentage of cells responding to stimulus with Ca²⁺ waves as averaged over multiple experiments for 0mM Ca²⁺ (n=29), 2mM Ca²⁺ (n=80); 10mM Ca²⁺ (n=28). Bar height shows total percentage of waves; orange portion represents the percentage of waves originating in protrusions (PRTS). C) Average number of oscillations within 2 min of stimulation under specified conditions. D) Representative traces of cells sensitized with anti-DNP IgE and stimulated with puff from pipette containing 1.7 μg/ml DNP-BSA (Ag), 1μM thapsigargin (TG), 20μM A23187 or BSS buffer only (vehicle). E) Percentage of cells responding to stimulus with Ca²⁺ waves averaged over multiple experiments and represented as in (B) with indicated conditions: Ag (n=80); TG (n=16); A23187 (n=29). F) Average number of oscillations within 2 min of stimulation under specified conditions. The first oscillation corresponds to the originating Ca²⁺ wave (dotted line). Error bars correspond to standard error of the mean (SEM).

Extended protrusions are common morphological features of RBL mast cells that are readily observed at low cell densities. They develop several hours after cell attachment and define an axis of cell migration (J. Lee and D. Holowka, unpublished results). Similar protrusions are frequently observed with mucosal mast cells imaged in vivo (for examples see: (34-37).

We found that substitution of extracellular Ca²⁺ with equimolar concentrations of Ba²⁺ or Sr²⁺ prior to stimulation did not markedly alter either the percentage of cells with Ag-stimulated waves or initiation from elongated protrusions (Table I). The velocity of wave propagation was reduced by ~40% in the absence of extracellular Ca²⁺, whereas substitution of Ca²⁺ by Ba²⁺, which is not taken up into stores by SERCA pumps (38), does not alter wave velocity (Table I). These results are consistent with wave initiation from protrusions by a Ca²⁺ influx component and propagation by Ca²⁺-dependent Ca²⁺ release from intracellular stores (10, 39). Interestingly, we find that the lag time for wave appearance following Ag stimulation was longer in the absence of extracellular Ca²⁺ (Table I), but was not changed when Ca²⁺ was substituted with Ba²⁺ or Sr²⁺. This delay could be attributed to a reduced “ignition” capacity of the calcium response in the absence of a Ca²⁺ influx, and is further discussed below.

In addition to the observed effects on Ca²⁺ waves, the absence of extracellular Ca²⁺ substantially decreases the frequency of subsequent Ca²⁺ oscillations (Figure 2C), as expected from the dependence of these oscillations on sustained, store-operated Ca²⁺ influx (17). Substitution of Ba²⁺ or Sr²⁺ for extracellular Ca²⁺ markedly reduces the frequency of oscillations (Table I), consistent with the high selectivity of CRAC channels to Ca²⁺ over Ba²⁺ or Sr²⁺ (40, 41).

We compared the dynamic features of Ag-stimulated Ca²⁺ responses in RBL cells to those observed upon stimulation with either the SERCA inhibitor thapsigargin (TG) or the Ca²⁺ ionophore, A23187. As illustrated by individual time traces in Figure 2D and summarized in figure 2F, the oscillatory dynamics of Ag responses are largely absent from responses to TG and A23187. For TG at 1μM in the pipette, a transient increase in cytoplasmic Ca²⁺ was typically observed, followed by a more sustained increase that depends on extracellular Ca²⁺. As summarized in Figures 2E, Ca²⁺ waves are less frequently observed with TG stimulation, but, when observed, waves usually initiate from extended protrusions. For stimulation by A23187 at 20 μM in the pipette, Ca²⁺ responses initiate as a wave in only about 40% of the cells, and this occurs very infrequently from extended protrusions (Figure 2E). Furthermore, when the response to A23187 does initiate as a Ca²⁺ wave, the wave velocity is substantially slower than observed in cells stimulated by high concentration of Ag or TG (see Table I). Similar to TG, A23187-mediated elevation in cytoplasmic Ca²⁺ does not cause oscillatory behavior in the majority of the stimulated cells (Figures 2D and F).

In summary, elevations in cytoplasmic Ca²⁺ caused by either ionophore or TG-induced release from stores can also initiate Ca²⁺ waves from discrete sites along the plasma membrane, but with lower probability. These results further indicate that cytoplasmic Ca²⁺ oscillations such as those initiated by Ag can be uncoupled from Ca²⁺ wave initiation caused by these non-receptor means of cell activation.

Characterization of the Ca²⁺ response in rat BMMCs

To evaluate whether the characteristics of Ca²⁺ responses that we observe for RBL-2H3 cells are general for mast cells, BMMCs were differentiated from Lewis rats as previously described (27). When plated overnight on glass slides in MatTek dishes, these cells attach firmly and frequently exhibit extended protrusions, similar in morphology to RBL cells. We confirmed their identity by labeling with Alexa 488-IgE and AA4 mAb that labels a mast cell specific ganglioside (42, 43; supplemental figure S4) as well as by staining granules with Alcian blue ((44); data not shown). BMMCs sensitized with IgE and labeled with Fluo-5F were stimulated as described above for RBL cells. As shown in Figure 3, we find that the majority of responding BMMCs exhibit initiating Ca²⁺ waves that often begin in extended protrusions, similar to Ca²⁺ waves observed in RBL cells. Also similar to Ca²⁺ responses in RBL cells, oscillations throughout the cytoplasm often follow the initial, propagating Ca²⁺ wave (Fig. 3b). These findings indicate that antigen-stimulated Ca²⁺ responses in rat BMMCs, including Ca²⁺ waves that initiate from specific morphological features, are similar to those characterized in cultured RBL mast cells.

Figure 3. Ca²⁺ dynamics in individual rat BMMCs.

A) Confocal image of a representative rat BMMC sensitized with anti-DNP IgE and loaded with Fluo5F (left panel), and time line analysis of indicated ROI for this cell stimulated at t=0 with a puff of 1.7 μg/ml DNP-BSA (right panel). Changes in Ca²⁺ concentration in response to stimulation are depicted, where brighter colors represent higher Ca²⁺ concentration. Time line analysis highlights the Ca²⁺ elevation initiating in the tip of the cellular protrusion and propagating through the cell body. B) Traces of Fluo5F intensity changes over time as calculated for the same cell imaged in (A). Black and red traces in right panel represent changes in Ca²⁺ concentration in cell body (yellow ROI) and in cell protrusion (red ROI), respectively, in the left panel.

Antigen dose dependence and spatial localization

To assess further the nature and function of the Ca²⁺ waves, we characterized the dose dependence of Ag-stimulated Ca²⁺ responses in RBL cells detected with GCaMP2 (Supplemental Figure S2). More than 90% of the cells respond to antigen doses in the range of 17-1700 ng/ml in the pipette, and more than 50% of the cells respond at the lowest dose tested, 1.7 ng/ml (Figure S2A). Ca²⁺ waves were observed in a similarly high percentage of the responding cells (≥80%) at all doses tested, but Ca²⁺ waves initiate more frequently at extended protrusions at the lowest Ag dose (Figure S2C). Wave velocity was found to be dependent on the concentration of Ag (Figure S2B), and longer lag times between antigen addition and the first Ca²⁺ response were observed at lower Ag doses (Figure S2D), but the frequency of Ca²⁺ oscillations changes little over the four orders of magnitude of Ag concentration (Figure S2E).

Ca²⁺ puffs are localized and transient elevations in cytoplasmic Ca²⁺ that may serve as a priming mechanism for wave initiation and subsequent oscillations (39, 45, 46). Stimulated Ca²⁺ puffs are observed less often than Ca²⁺ waves at higher doses of antigen, but appear more often at the lowest concentration of Ag tested (Figure S2F). Consistent with these observations, Ca²⁺ puffs may represent sub-threshold activation of the cells and production of insufficient amounts of IP₃ to generate a propagated Ca²⁺ elevation and a global Ca²⁺ response.

Interestingly, when antigen is covalently coupled to the tip of a micropipette and delivered locally by cell contact, we usually observe a series of stimulated Ca²⁺ puffs that frequently occur near the point of contact. As illustrated in Figure 4, two different responses are typical: In one case, repetitive puffs are observed at relatively regular time intervals, often over several minutes of contact, but they fail to propagate as waves or global cellular Ca²⁺ elevation for more than several microns (Fig. 4A). Alternatively, Ca²⁺ puffs initiated by local contact of the antigen-coated pipette occasionally propagate as waves that traverse the length of the cell (Fig. 4B). Under these conditions of stimulation, contact at a cell protrusion did not significantly increase the probability of wave propagation. Ca²⁺ puffs or waves were observed in more than 85% of cells contacted by DNP-BSA coated tips in both the presence and absence of extracellular Ca²⁺ (Fig. 4C) and only rarely in cells contacted by unconjugated micropipettes or those conjugated with unmodified BSA (Fig. 4C). Moreover, Ca²⁺ puffs were just as frequent when cells were locally stimulated with DNP-BSA conjugated micropipettes in the absence of extracellular Ca²⁺ (Fig. 4C). These results suggest that the localized responses are mediated by release of Ca²⁺ from internal stores, rather than by Ca²⁺ influx. Moreover, localized activation of FcεRI can stimulate spatially restricted Ca²⁺ responses, and, in this situation, the site of Ca²⁺ response initiation is determined by the location of receptor stimulation.

Figure 4. Stimulation of RBL cells by contact with Ag-coated micropipette.

A and B) Representative cells expressing GCaMP2, sensitized with IgE and stimulated with DNP-BSA conjugated micropipettes. A) Micropipette (indicated by yellow arrowhead) contacting the cell at the tip of a protrusion elicits a train of spatially restricted Ca²⁺ puffs, each traveling no more than 30μm along the protrusion. B) Contact stimulation at the cell body results repetitive Ca²⁺ puffs in the cell body that sometimes propagate as a wave to the protrusion. Left panels show images with ROIs defined; Right panels show Ca²⁺ concentration changes in ROIs of corresponding color. Black arrow indicates initiation of contact between micropipette and cell. C) Histogram showing percentage of cells responding with local Ca²⁺ puffs only (light blue) or more global Ca²⁺ elevation (dark blue) due to contact with DNP-BSA-conjugated micropipettes in the presence (n=28) or absence (n=18) of extracellular Ca²⁺, or with unmodified BSA-conjugated micropipettes (n=18).

These results show that the localized Ca²⁺responses can be initiated by localized receptor stimulation, and these responses are apparently independent of Ca²⁺ influx. Overall, these results show that Ca²⁺ responses to soluble Ag begin most frequently as Ca²⁺ waves that are propagated from initiation sites near the tips of extended protrusions, whereas those stimulated by locally applied micropipette-conjugated antigen are initiated at the site of stimulation, and are more commonly manifested as repetitive Ca²⁺ puffs that less frequently evolve into Ca²⁺ waves.

Pharmacologic characterization of Ca²⁺ dynamics

In initial efforts to identify molecular components of Ag-stimulated waves and oscillations, we investigated the effects of several different inhibitors of IP₃ production and modulators of Ca²⁺ influx. As summarized in figure 5 and in Table I, the phospholipase C (PLC) inhibitor, U73122, reduces the percentage of cells showing Ca²⁺ responses with a sharp dose dependence that is half maximal between 2 and 2.5 μM. U73122 also reduces the percentage of cells with an initiating Ca²⁺ wave with a similar dose-dependence, suggesting that Ag-stimulated Ca²⁺ waves depend on IP₃ and/or diacylglycerol production (Figure 5). In this dose range, U73122 is relatively selective for PLC isoforms in these cells, but at higher doses other proteins are also affected by this maleimide derivative (47). As summarized in Figure 5, Ag-stimulated Ca²⁺ oscillations are inhibited with a similar dose-dependence, consistent with IP₃-dependent intracellular Ca²⁺ increase as a necessary precursor to SOCE (48).

Figure 5. Effects of inhibitors on intracellular Ca²⁺ waves and oscillations.

A) Percentage of cells from multiple experiments showing a measurable Ca²⁺ response to stimulation by 1.7 μg/ml Ag in the presence of the following reagents in the extracellular buffer: 1-5 μM U73122 (n>18 each), 20mM 2-APB (n=23), 8μM D-sphingosine (n=26); 8μM DM-sphingosine (n=31), 10μM nifedipine (n=15), 1 or 10 μM GdCl₃ (Gd³⁺; n>20 each) B) Percentage of cells responding to stimulus with Ca²⁺ waves as averaged over multiple experiments with indicated conditions. Bar height shows total % waves; orange portion represents % waves originating in protrusions. C) Average number of oscillations within 2 min of stimulation under specified conditions. The first oscillation corresponds to the originating Ca²⁺ wave (dotted line). Error bars correspond to SEM.

2-APB was initially described as a non-competitive antagonist of IP₃ receptor (IP₃R)-mediated Ca²⁺ release by microsomal preparations (49), and it was subsequently shown to inhibit SOCE in T cells at concentrations in the range of 10-50μM (50). In RBL cells, 2-APB at concentrations up to 40μM has no effect on Ag-stimulated Ca²⁺ release from stores in the absence of extracellular Ca²⁺, indicating that it fails to inhibit IP₃R-mediated Ca²⁺ release from stores in these cells (D. Holowka, unpublished results). As summarized in Figure 5, addition of 20μM 2-APB prior to Ag stimulation does not significantly reduce the percentage of responding cells (Figure 5A), or the percentage of cells that exhibit wave initiation (Figure 5B). In contrast, 20μM 2-APB strongly inhibits Ca²⁺ oscillations subsequent to wave initiation (Figure 5C), consistent with its inhibitory effect on SOCE that contributes to Ca²⁺ oscillations in these cells (51-53)

A previous study by Penner and colleagues (54) identified both D-sphingosine and N,N-dimethylsphingosine (DMS) as effective inhibitors of IP₃ and TG-elicited Ca²⁺ release-activated Ca²⁺ (CRAC) channel activation in RBL cells. As summarized in Figure 5B and Table I, we found that 8 μM D-sphingosine or 8 μM DMS strongly inhibits Ca²⁺ wave initiation. Earlier studies suggested that DMS prevents Ca²⁺ responses by inhibition of sphingosine kinase (55, 56). However, this mechanism is unlikely for the inhibition we observe because D-sphingosine is a substrate of this enzyme family. In other studies, we have evidence that these basic amphiphiles inhibit Ca²⁺ responses by electrostatically neutralizing negatively charged phospholipids at the inner leaflet of the plasma membrane ((57); N. Smith et al., manuscript in preparation). The small percentage of cells that do respond in the presence of D-sphingosine appear to have oscillation frequencies similar to untreated cells (Figure 5C), suggesting that this aspect of SOCE is less potently inhibited than wave initiation. These results identify these sphingosine derivatives as effective inhibitors of the earliest manifestations of Ca²⁺ responses to Ag in these cells.

We further tested the effects of inhibitors of dihydropyridine-sensitive Ca²⁺ channels on Ag-stimulated Ca²⁺ wave initiation in RBL cells (58). As summarized in Figure 5, we observed a modest (~30%) reduction of wave initiation from extended protrusions by 10 μM nifedipine, but no effects of other L channel inhibitors such as Bay K (+) were detected (data not shown), suggesting that the effect of nifedipine observed may be due to non-specific inhibition of K⁺ channels resulting in cell depolarization (59, 60). To investigate further a possible contribution of CRAC conductance to wave generation, we examined the effects of Gd³⁺ on Ca²⁺ responses to Ag. Pre-addition of 10 μM Gd³⁺ strongly inhibits Ca²⁺ oscillations in response to Ag (Fig. 5C), consistent with inhibition of SOCE (61), but it causes only a small reduction in the percentage of waves initiating from extended protrusions (Figure 5B). These results are consistent with those obtained with 2-APB that indicate little or no contribution of CRAC channels to Ag-stimulated wave initiation. Because Ca²⁺ influx contributes to the spatial localization of Ca²⁺ wave initiation from protrusions (Fig. 2), these results suggest that Ca²⁺ channels other than CRAC or voltage gated channels contribute to this process.

Involvement of TRPC channels in the initiation of Ca²⁺ waves from protrusions

Expression of TRPC channels in RBL cells was demonstrated by Pizzo et al. (62) and more recently by Ma et al (25). TRPC channels have been identified as receptor activated Ca²⁺ channels, and we investigated the contribution of TRPC channels to wave initiation by IgE-FcεRI in RBL mast cells using shRNA sequences targeting TRPC1, TRPC3, TRPC5 and TRPC7 to knock down expression of these proteins (25). For these experiments, RBL cells were co-transfected with the Ca²⁺ indicator GCaMP2 and one of the TRPC targeting shRNA plasmids as described in Materials and Methods. We found that transfection with TRPC1 shRNA and, to a lesser extent, with TRPC3 shRNA reduces the responsiveness of the cells to a low concentration of Ag compared to cells expressing a mock shRNA sequence, and these effects are particularly notable if cells responding with Ca²⁺ puffs are not included in the count (Figure 6A). As shown in Supplemental Figure S3, we observe ~50% reduction in TRPC1 expression due to shRNA transfection in these cells, similar to previous results (25). This value represents a lower limit to the knockdown of TRPC1 cell expressing GCaMP2, which is co-transfected with the shRNA constructs. Under these knock down conditions, we did not observe a significant reduction in the percentage of cells responding to Ag in cells transfected with TRPC5 or TRPC7. In addition to reduction in cell responsiveness, we found that transfection with TRPC1 or TRPC3 shRNA prolongs the lag time from Ag addition to the onset of the Ca²⁺ response compared to cells transfected with TRPC5, TRPC7, or the control shRNA (Fig 6B). These results suggest that TRPC1 and TRPC3 play a significant role in the initiating Ca²⁺ wave phase of the cellular response to Ag stimulation.

Figure 6. Participation of TRPC channels in Ag stimulated initiation of the Ca²⁺ response.

RBL cells were co-transfected with GCaMP2 and shRNA plasmids targeting TRPC1, 3, 5, 7 or mock sequence (for each condition n>29). Cells sensitized with anti-DNP IgE were stimulated with puff from pipette containing 1.7 ng/ml DNP-BSA in BSS. A) Percentage of cells from multiple experiments showing a measurable Ca²⁺ response corresponding to elevation of intracellular Ca²⁺ cells. Dark blue bars include cells that only include Ca²⁺ puffs in the count, and light blue bars exclude this population. B) Average lag time before initiation of Ca²⁺ wave with specified shRNA. Error bars correspond to SEM. *P < 0.05, **P < 0.01 vs. cells expressing shGFP. C) Percentage of cells responding to stimulus with Ca²⁺ waves with indicated shRNA. Bar height shows total % waves; orange portion represents % waves originating in protrusions. D) Model for initiation of Ca²⁺ mobilization in extended protrusions. Ag crosslinks IgE/FcεRI complexes, activating Ca²⁺ influx via TRPC1 channels to cause local increase in cytoplasmic Ca²⁺ at the tip of a cell protrusion. This potentiates IP₃-mediated Ca²⁺ release from the nearby ER store that propagates via Ca²⁺-induced Ca²⁺ release to generate a Ca²⁺ wave that initiates Ca²⁺ oscillations and cell activation. Two possible mechanisms for wave initiation at these protrusions are indicated.

Most notably, expression of TRPC1 and TRPC3 shRNA results in decreased frequency of Ca²⁺ wave initiation from cell protrusions compared to expression of mock shRNA (Figure 6C). Expression of TRPC7 shRNA also decreases the frequency of Ca²⁺ wave initiation from cell protrusions, even though this construct does not significantly alter the capacity of the cells to respond or the lag time of the response. Measurements of Ca²⁺ wave velocity show only marginal differences for cells expressing the TRPC shRNA constructs, suggesting that these channels are not involved in Ca²⁺ wave propagation (data not shown). These results also indicate that reduction in the capacity of cells to respond to Ag stimulation after the TRPC1 or TRPC3 shRNA expression is caused by perturbation of wave initiation of the Ca²⁺ response, rather than in later steps involved in propagating the Ca²⁺ signal, which have been shown to be inhibited by TRPC5 shRNA (25).

DISCUSSION

IgE receptor-mediated Ca²⁺ mobilization plays key roles in mast cell activation and consequent exocytotic release of inflammation and allergy-related mediators. Previous studies characterized dramatic Ca²⁺ oscillations stimulated by Ag in these cells (15) that correlate temporally to exocytotic events (13, 15). In the present study we used high-speed confocal imaging to characterize rapid, early events in stimulated Ca²⁺ mobilization with high spatial resolution. Our principal new finding is that the initial increase in cytoplasmic Ca²⁺ in response to Ag stimulation in RBL mast cells is highly spatially organized, initiating at specific regions of the cell and propagating through the cytosol as a Ca²⁺ wave. The existence and nature of Ca²⁺ waves in hematopoietic cells has been controversial, and the study characterizing high-speed, circumferential waves in neutrophils (63) could not be confirmed (64). Lee and Oliver previously described evidence for a Ca²⁺ influx pathway in RBL cells that precedes Ca²⁺ release from intracellular stores (65). The Ca²⁺ waves stimulated via FcεRI that we have characterized in the present study are similar in several aspects to Ca²⁺ waves observed in certain excitable and other non-hematopoietic cells (6, 7, 66), but they have not been previously detected in Ca responses stimulated by multichain immune recognition receptors.

In certain cell types, elementary Ca²⁺ events such as Ca²⁺ puffs can lead to initiation and propagation of a Ca²⁺ wave, and this has been shown to be regulated by a number of factors, including spatial organization of intracellular Ca²⁺ release sites (67-70). In RBL mast cells, the high incidence of wave initiation from elongated cellular protrusions suggests that these protrusions contain specialized structures that ignite the cellular response. Interestingly, the percentage of stimulated Ca²⁺ waves that initiate from protrusions increases at weak stimulation indicating that protrusions are particularly sensitive to Ag. A survey of the literature indicates that such protrusions are a common feature of mucosal mast cells in tissue sites (34, 36, 37, 71), consistent with an important role for these protrusions in mast cell physiology. The high sensitivity of wave initiation in protrusions that we observe could result from locally enhanced coupling of the IgE receptor and the machinery responsible for Ca²⁺ release from stores. We do not detect enhanced expression of FcεRI in these protrusions, but we find that initiation of Ca²⁺ waves in protrusions is selectively reduced in the absence of extracellular Ca²⁺, suggesting a facilitating Ca²⁺ influx pathway at these protrusions. As discussed below, our results point to Ca²⁺- dependent potentiation of IP₃ receptors close to sites of Ca²⁺ influx at protrusions, rendering these receptors more sensitive and therefore more probable sites for Ca²⁺ wave initiation.

Ca²⁺ oscillations following initial waves exhibit a global, non-directional increase in Ca²⁺ concentration that contrasts with directional Ca²⁺ elevation seen for the waves. Because of this spatial distinction and the pharmacological differences, we conclude that waves and oscillations are two discrete but coupled Ca²⁺ events in IgE receptor-mediated signaling. We suggest that Ca²⁺ waves are the initiating event of mast cell activation, established by the unique spatial distribution of Ca²⁺ influx pathways in these cells, involving TRPC channels that may be preferentially expressed in the elongated protrusions, rendering these to be more sensitive to Ag stimulation. Subsequent to the waves are Ca²⁺ oscillations that are thought to result from IP₃-induced release of Ca²⁺ from ER (72). These reoccurring Ca²⁺ elevations were suggested to play a role in prolonging the duration of the Ca²⁺ signal, enhancing its efficacy (16) and also in providing temporal encoding for downstream events such as degranulation (15). Interestingly, we could occasionally detect Ca²⁺ oscillations in cells that did not demonstrate a Ca²⁺ wave upon Ag stimulation, suggesting that the propagation of Ca²⁺ through the cytoplasm during the wave period is not required for the onset of Ca²⁺ oscillations.

In addition to Ca²⁺ waves and oscillations, we detect Ca²⁺ puffs in cells stimulated by antigen under certain conditions. Of particular interest are the repetitive Ca²⁺ puffs that are elicited by local application of antigen with a DNP-BSA coated micropipette, demonstrating that the molecular machinery capable of initiating local Ca²⁺ responses is distributed throughout the cell and is not restricted to the protrusions. Ca²⁺ puffs stimulated by Ag-coated micropipette contact are similarly frequent in the absence of extracellular Ca²⁺, indicating that the generation of spatially restricted Ca²⁺ elevations in these cells involves Ca²⁺ release from stores and does not require Ca²⁺ influx. However, the periodic nature of these puffs suggests a complex mechanism for their regulation that may be related to the periodic nature of more global Ca²⁺ oscillations elicited by soluble antigen. It will be interesting to determine whether these localized Ca²⁺ puffs are sufficient to elicit localize exocytosis in these cells.

Pharmacologic characterization of IgE receptor-mediated Ca²⁺ waves yield evidence for PLC activation in this mechanism, whereas store-operated Ca²⁺ entry does not appear to play an important role in this early phase of the Ca²⁺ response. PLCγ-dependent IP₃ production is important for generating Ca²⁺ responses to Ag in mast cells (73), and it is possible that a specific IP₃ receptor family member is involved in Ca²⁺ wave initiation. The greater sensitivity of IP₃R-3 to Ca²⁺ release by IP₃ (74) suggests this family member as a strong candidate for initiation of an IP₃-dependent Ca²⁺ wave at protrusions. Ca²⁺ waves can be triggered by alternate means of Ca²⁺ mobilization, including Ca²⁺ ionophore and thapsigargin, but these agents cause less frequent Ca²⁺ waves (Table I). Furthermore, initiation of Ca²⁺ waves by Ca²⁺ ionophore occurs only infrequently at extended protrusions. Together, these results suggest that Ca²⁺ responses initiated by the IgE receptor are more spatially regulated than those caused by non-receptor-mediated cell activation.

The recent study by Ma et al. provided evidence that TRPC channels contribute to stimulated Ca²⁺ influx in RBL cells (25), suggesting their possible involvement in the Ag-stimulated Ca²⁺ influx pathway that initiates Ca²⁺ waves from extended protrusions. Our results with shRNA-mediated knockdown of TRPC proteins provide evidence that decreased expression of TRPC1, and, to lesser extent, TRPC3, reduces the responsiveness of these cells to low concentration of Ag (Figure 6A, B) and decreases the frequency of wave initiation from extended protrusions (Figure 6C). At higher concentrations of Ag, TRPC 1 channel knock down is less effective in decreasing the responsiveness of the cells (data not shown), indicating that this channel plays an important role in Ca²⁺ responses only under more physiological conditions of IgE receptor stimulation. Other TRPC channels may also contribute to Ca²⁺ wave initiation, and our failure to detect significant effects of TRPC5 knock down, or the smaller effects of TRPC3 and 7 knock down, may be limited by the efficiency of their reduction under our experimental conditions.

These results suggest that Ca²⁺ influx via TRPC channels, particularly TRPC1, is likely to participate in the initial steps of mast cell activation by enhancing the sensitivity of these cells to threshold Ag stimulation. Although the mechanism for this enhancement is not yet understood, it is likely that a Ca²⁺ influx-dependent increase in IP₃R activation is involved. Two possible mechanisms are indicated in Figure 4D: In one, local accumulation of TRPC at protrusions could enhance receptor-stimulated influx at this region; Alternatively, enhanced physical coupling of the TRPC channels to IP₃R (75), PLCγ (76), or FcεRI in this region could mediate local Ca²⁺ influx to initiate wave propagation.

Our observations that rat BMMCs exhibits similar morphologies as RBL mast cells and initiate Ca²⁺ responses to antigen in the form of waves that begin in extended protrusions show that these stimulated waves are a general feature of FcεRI-mediated Ca²⁺ signaling. Both rat BMMCs (23) and RBL mast cells (20) have marker characteristics of mucosal mast cells, and it is not yet clear whether this wave response is a specific feature of this subset of mast cells. Mouse BMMCs undergo chemotatic responses on vitronectin (77), and it may be possible to observe stimulated Ca²⁺ waves under these conditions of cell attachment.

Our data establish Ca²⁺ waves originating from a specific cellular site as a common feature of the Ca²⁺ response to antigen in RBL mast cells and rat BMMCs, but the physiological function of these waves and the molecular basis for their site of initiation are not completely understood. Regulated Ca²⁺ signaling across the cell is a fundamental mechanism for information transfer between organelles or cellular regions that conveys directional and temporal information. In this regard, we have preliminary evidence that direction of RBL cell migration correlates with the site of Ag-stimulated wave initiation (R. Cohen and J. Lee, unpublished results), suggesting a strategic localization of the Ca²⁺ signaling machinery to communicate directional information in these cells. Localized degranulation from extended protrusions might also be mediated by this machinery under certain physiological conditions.

In summary, we show that the response of RBL and rat BMMC to Ag stimulation initiates with a propagating Ca²⁺ wave, most often beginning at the tips of cellular protrusions in a Ca²⁺ influx-dependent manner. These waves initiate with TRPC1-dependent Ca²⁺ influx, then propagate by Ca²⁺-dependent Ca²⁺ release from stores, followed by extracellular Ca²⁺-dependent oscillations. Spatial localization of wave initiation sites to extended protrusions suggests a link between this signaling response and polarized cell motility. Our results indicate that when mast cells encounter physiological levels of antigen that are insufficient to directly activate global Ca²⁺ mobilization, activation of Ca²⁺ influx mediated by TRPC1 channels in extended cell protrusions can ignite a response that is propagated and amplified. In this manner, the initiation of a physiological cellular response is both spatially and temporally controlled.