A pH-stabilizing role of voltage-gated proton channels in IgE-mediated activation of human basophils

Abstract

Eosinophils and other phagocytes use NADPH oxidase to kill bacteria. Proton channels in human eosinophils and neutrophils are thought to sustain NADPH oxidase activity, and their opening is greatly enhanced by a variety of NADPH oxidase activators, including phorbol myristate acetate (PMA). In nonphagocytic cells that lack NADPH oxidase, no clear effect of PMA on proton channels has been reported. The basophil is a granulocyte that is developmentally closely related to the eosinophil but nevertheless does not express NADPH oxidase. Thus, one might expect that stimulating basophils with PMA would not affect proton currents. However, stimulation of human basophils in perforated-patch configuration with PMA, N-formyl-methionyl-leucyl-phenylalanine, or anti-IgE greatly enhanced proton currents, the latter suggesting involvement of proton channels during activation of basophils by allergens through their highly expressed IgE receptor (FcεRI). The anti-IgE-stimulated response occurred in a fraction of cells that varied among donors and was less profound than that to PMA. PKC inhibition reversed the activation of proton channels, and the proton channel response to anti-IgE or PMA persisted in Ca²⁺-free solutions. Zn²⁺ at concentrations that inhibit proton current inhibited histamine release elicited by PMA or anti-IgE. Studied with confocal microscopy by using SNARF-AM and the shifted excitation and emission ratioing of fluorescence approach, anti-IgE produced acidification that was exacerbated in the presence of 100 μM Zn²⁺. Evidently, proton channels are active in basophils during IgE-mediated responses and prevent excessive acidification, which may account for their role in histamine release.

The IgE-mediated activation of basophils is a hallmark of allergic reactions. Binding of allergen to allergen-specific IgE on basophils results in cross-linking of high-affinity receptors (FcεRI), thereby activating a signaling pathway that results in the release of histamine and other mediators. Basophils are among several cell types that express voltage-gated proton channels (1). Proton channels are opened by depolarization and/or cytoplasmic acidification and seem to be designed for rapid, efficient acid extrusion from cells (2). In phagocytes, they are thought to enable sustained NADPH oxidase activity by compensating for the electrogenic activity of the oxidase (3–5). Stimuli that activate NADPH oxidase in human eosinophils and neutrophils and in murine osteoclasts greatly enhance the opening of proton channels in these cells studied in perforated-patch configuration (6–10), largely via PKC phosphorylation (11). In contrast, phorbol myristate acetate (PMA) has no clear effect on proton currents in alveolar epithelial cells that lack NADPH oxidase (6). The function of proton channels in basophils is unknown, but must differ from that of phagocytes, because basophils lack NADPH oxidase (12). Here, we show that proton channels in basophils respond vigorously to agents that elicit histamine release. Furthermore, histamine release stimulated by anti-IgE or PMA was inhibited by Zn²⁺ at concentrations that inhibit proton currents, consistent with the idea that proton channel activity is linked to basophil activation. Studies of pH_i using the fluorescent pH sensitive dye SNARF-AM suggest that proton channels limit acidification during basophil responses.

Results

Proton Currents in Human Basophils Respond to PMA.

Human basophils voltage-clamped in the perforated-patch configuration exhibit voltage-gated proton currents (Fig. 1A) that resemble those in human eosinophils (7), both in properties and in amplitude. In basophils at symmetrical pH 7.0 (50 mM NH₄⁺ on both sides of the membrane clamps pH_i near pH_o), the proton conductance, g_H, activates during depolarizing voltage pulses above +20 mV, and activates more rapidly at more positive voltages. The net, leak-subtracted H⁺ current (I_H) at +80 mV was 30 ± 18 pA (mean ± SD) in 21 cells. In similar measurements, I_H in human eosinophils was 48 pA (7).

Fig. 1.

Proton currents in human basophils respond vigorously to PMA and GFX. (A, C, and E) Families of currents in a basophil before stimulation (A), after exposure to 60 nM PMA for 4 min (C), and after addition of 2.5 μM GFX (E). In each, the membrane was held at −20 mV, and 8-s pulses were applied in 10-mV increments every 30 s up to +80 mV. (B) Superimposed currents during the transition to PMA, with 4-s pulses to +60 mV applied every 15 s. (D) Test currents after addition of GFX. (F) I–V relationships from the families in A, C, and E. (G) g_H–V relationships from the families in A, C, and E were calculated by using V_rev measured in each condition and the steady-state I_H obtained by extrapolating exponential fits. (H) Voltage dependence of the activation time constant (τ_act) obtained by fitting the currents in A, C, and E to a rising exponential (after a delay). Calibration bars apply to A–E. The bath included Ca²⁺.

The PKC activator PMA is a potent and effective stimulus of histamine release from basophils (13). Nearly every basophil stimulated with PMA exhibited a profound enhancement of proton currents. After addition of PMA to the bath solution, I_H during a test pulse increased progressively over several minutes (Fig. 1B). The same family of pulses in PMA elicited larger currents (Fig. 1C) that activated more rapidly at each voltage. H⁺ current–voltage relationships (Fig. 1F) reveal that the voltage at which the H⁺ current was first activated, V_threshold, was shifted ≈20 mV more negative after stimulation with PMA. The chord conductance (g_H) calculated from the H⁺ current (Fig. 1G) was also shifted −20 mV by PMA, and the maximum H⁺ conductance (g_H,max) was doubled in this cell. The impression that PMA stimulation increased the opening rate of proton channels is confirmed in Fig. 1H. The activation time constant, τ_act, was decreased (activation became faster) by ≈3-fold at all voltages. Average responses of proton currents in basophils and neutrophils to PMA are summarized in Table 1. Consistent with the lack of NADPH oxidase (12), no hint of electron current was seen.

Table 1.

Effects of anti-IgE and PMA on proton currents in human basophils compared with PMA effects in human neutrophils

Parameter	Basophil with anti-IgE	Basophil with PMA	Neutrophil with PMA*
gH,max	2.29 ± 0.16 (31)	2.86 ± 0.21 (40)	1.9
	(anti-IgE/control)	(PMA/control)	(PMA/control)
τact at Vtest	2.17 ± 0.19 (28)	5.04 ± 0.56 (52)	3.7
	(control/anti-IgE)	(control/PMA)	(control/PMA)
τtail at Vhold	1.20 ± 0.04 (33)	1.32 ± 0.07 (64)	5.5
	(anti-IgE/control)	(PMA/control)	(PMA/control)
ΔVthreshold, mV	−11.0 ± 0.9 (33)	−19.0 ± 1.0 (61)	−38.8
	(anti-IgE − control)	(PMA − control)	(PMA − control)

Parameters [mean ± SEM (n)] were measured when the effects of PMA (60–160 nM) or anti-IgE (usually 0.5–1.0 μg/ml) peaked or approached steady state, usually within a few minutes. Only cells that clearly responded are included. Responses are given as ratios where 1.0 means no response, except for V_threshold, for which 0 mV means no response. The g_H,max values were calculated from the largest pulses applied. Gating kinetics, τ_act and τ_tail, respectively, were obtained from exponential fits to currents during test pulses or upon repolarization to the holding potential (V_hold). Data for Ca²⁺ or Ca²⁺-free bath are combined because no parameter differed significantly for either stimulus. All PMA effects except on τ_tail are significantly greater (P < 0.05) than those for anti-IgE.

*Values from a previous study (6).

The enhanced gating mode, or “activation,” of phagocyte proton channels is largely attributable to phosphorylation by PKC of either the channel itself or a closely related protein (11). Fig. 1D shows that the PKC inhibitor GF109203X (GFX) reversed the increase in I_H produced by PMA in basophils. GFX also reversed the shift in the I_H–V relationship (Fig. 1F), the increase in g_H,max (Fig. 1G), and the change in τ_act (Fig. 1H). In this cell, GFX seemed to reverse all effects of PMA. In contrast, GFX reversed the effects of PMA on proton currents in eosinophils only partially (11).

Proton Currents in Human Basophils Respond to Anti-IgE.

The characteristic response of human basophils occurs when allergens bind to IgE, resulting in cross-linking of high-affinity receptors (FcεRI) that activates a signaling pathway that results in the release of histamine and other mediators. Cross-linking IgE with goat anti-human IgE antibody (anti-IgE) elicited a distinct proton channel response (Fig. 2) in a fraction of cells. More than half of the cells from the most responsive donors responded to anti-IgE, but most cells from certain other donors failed to respond. Five cells each from three donors failed to respond to anti-IgE but responded subsequently to PMA. In studies of highly purified (≈98%) basophils (in Methods), 60% of cells (6/10) responded to anti-IgE, confirming that some nonresponding cells were basophils. Donor-to-donor variability is typically observed in histamine release by basophils in response to anti-IgE (14, 15).

Fig. 2.

Proton currents in human basophils are enhanced by anti-IgE stimulation in a GFX-sensitive manner. (A–C) Currents in response to the same applied family of pulses before stimulation (A), after stimulation by 0.8 μg/ml anti-IgE (B), and in the presence of 3 μM GFX (C). (D) The time course of these responses. Anti-IgE was added to the bath incrementally, and then the bath stirred. The bath was Ca²⁺ free. (E–H) Time course of anti-IgE responses in four other basophils. Horizontal arrows indicate zero current. In each, test pulses to +40 mV (E–G) or +60 mV (H) were applied every 30 s, and at the arrow, 0.6 μg/ml anti-IgE was applied to the bath followed by stirring, except for G, where 6.6 μg/ml was applied. In H, “wash” indicates a bath exchange with the same solution as a control; in this experiment anti-IgE was applied by complete bath exchange. The bath included 2 mM Ca²⁺, except H, which was Ca²⁺-free.

Anti-IgE responses in Fig. 2 D–H illustrate the variety of time courses observed. Depolarizing pulses were applied every 30 s to elicit proton currents. Typically, there was a delay of 1 min to several minutes before the onset of a response. I_H then began to increase. In some cells, I_H peaked and immediately began to decline (Fig. 2 E and G), although not to its original amplitude. In other cells, I_H remained elevated for prolonged periods. These responses were observed after stimulation with anti-IgE at concentrations ranging from 0.2 to 10 μg/ml. At higher concentrations, the response seemed more rapid and transient, but there were exceptions. Most measurements were performed by using 0.5–1.0 μg/ml anti-IgE.

In the experiment in Fig. 2D, anti-IgE was added incrementally, with a profound response occurring at 0.8 μg/ml. In families of currents recorded before and after the addition of anti-IgE (Fig. 2 A and B), the H⁺ current was larger and activated faster at each voltage after anti-IgE. In Table 1, the parameter values during the peak of the anti-IgE response are given. In cells responding to anti-IgE, g_H,max doubled, and channel opening was faster (smaller τ_act), but closing kinetics (τ_tail) was not affected. The proton conductance–voltage (g_H–V) relationship characterized as V_threshold was shifted −11 mV by anti-IgE and −20 mV by PMA. Proton channel responses to anti-IgE and PMA in human basophils were qualitatively identical, but the anti-IgE responses were consistently smaller.

In basophils that exhibited a sustained increase in proton current in response to anti-IgE, GFX reversed the anti-IgE effects distinctly but not completely (Fig. 2 C and D). In cells that responded to anti-IgE with a transient increase in I_H (e.g., Fig. 2G), GFX added after the response had subsided but remained above the initial level, further reduced the current (data not shown).

Roughly half the cells studied did not respond to anti-IgE but did respond to a subsequent application of PMA, confirming that they had not entered whole-cell configuration due to spontaneous patch rupture, which abolishes the PMA response in eosinophils (16). Addition of PMA to cells that had already responded to anti-IgE invariably produced a greater response that (compared with their initial state) was indistinguishable from the response of cells stimulated only with PMA (Table 1 and not shown). GFX reversed most of the combined effects of anti-IgE and PMA (data not shown).

fMLF Also Activates Proton Channels.

Stimulation of basophils with 10 μM N-formyl-methionyl-leucyl-phenylalanine (fMLF), another agonist of histamine release (15), enhanced proton currents in 10 of 23 cells studied (data not shown). The response to fMLF resembled that to anti-IgE and was augmented by subsequent addition of PMA (n = 8). GFX partially reversed the effect of fMLF alone (n = 1) or fMLF and PMA together (n = 4).

Proton Channel Responses Do Not Require Ca²⁺ Influx.

Basophil proton channels responded to PMA or anti-IgE whether or not Ca²⁺ was present in the bath. Considering only cells from two responsive donors, 14 of 22 cells (64%) responded to anti-IgE with ≈0.5 mM free Ca²⁺ (1.5 mM CaCl₂ and 1 mM EGTA) in the bath, and 13 of 20 cells (65%) responded to anti-IgE in a Ca²⁺-free (EGTA-containing) bath solution. The presence or absence of Ca²⁺ in the bath did not detectably affect the magnitude of any proton channel response to PMA or anti-IgE [supporting information Tables S1 and S2]. Proton channels in basophils exposed to 20 μM 1,2-bis(2-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid, tetraacetoxymethyl ester (BAPTA-AM) in Ca²⁺-free solutions responded to stimulation with anti-IgE (n = 3) or PMA (n = 2), suggesting that elevated [Ca²⁺]_i is not required for a response.

Proton Channel Response Precedes Degranulation.

If histamine storage granules expressed proton channels, their fusion with the plasma membrane might increase g_H owing to channel insertion. However, the capacitance of a sample of basophils did not change significantly after proton channel responses to anti-IgE (mean ± SE increase 10 ± 9%, n = 16) or PMA (mean decrease 6 ± 11%, n = 10). Histamine release is slow, especially with PMA as a stimulus (17) and likely did not occur during the proton channel response.

Zinc Inhibition of Proton Currents in Human Basophils.

Typically, voltage-gated proton currents are inhibited potently by Zn²⁺ and other polyvalent metal cations, but Zn²⁺ sensitivity of proton currents in basophils has not been reported. Fig. 3 illustrates the effects of Zn²⁺ on proton currents in a human basophil. Currents during pulses in 10-mV increments up to +60 mV are shown for a basophil in whole-cell configuration in the absence of Zn²⁺ and in the presence of 1, 10, and 100 μM Zn²⁺. As in other cells (2), Zn²⁺ slowed the activation of H⁺ current at each voltage and shifted the g_H–V relationship positively (Fig. 3E). Zn²⁺ seems to reduce g_H,max, but g_H was calculated from I_H at the end of 6- to 10-s pulses. Because Zn²⁺ slows activation, I_H at the end of the pulse increasingly underestimates the steady-state value at higher [Zn²⁺]. The shift of the g_H–V relationship, estimated at ≈10% of g_H,max, was 27 ± 4 mV (mean ± SEM, n = 5) at 1 μM Zn²⁺, 51 ± 8 mV (n = 5) at 10 μM Zn²⁺, and 106 ± 7 mV (n = 3) at 100 μM Zn²⁺. These shifts are close to the expectation of Zn²⁺ effects at pH 7.4 in rat alveolar epithelial cells, where analogous data exist (18), suggesting similar affinity of Zn²⁺ for H⁺ channels in both cells.

Fig. 3.

Effects of Zn²⁺ on proton currents in a human basophil at pH_o 7.4 (Ringer's solution without EGTA) and pH_i 5.5. (A–D) Families of currents in response to depolarizing pulses applied from a holding potential of −60 mV to +60 mV in 10-mV increments. The Zn²⁺ concentration was 0 (A), 1 μM (B), 10 μM (C), and 100 μM (D). (E) The g_H–V relationships from this experiment, calculated by using the current at the end of 6-s pulses (control and wash) or 10-s pulses (all Zn²⁺ concentrations). Symbols in the order of the experiment are ● (control), ■ (1 μM Zn²⁺), ▴ (10 μM Zn²⁺), ▾ (100 μM Zn²⁺), and ○ (after washout). Because Zn²⁺ slows activation, the current amplitude at the end of the pulse increasingly underestimates the steady-state value at higher [Zn²⁺]. Pulses to larger voltages are included in E.

The potency of Zn²⁺ in inhibiting proton currents in basophils activated with PMA in perforated-patch configuration was also examined. A previous study reported that Zn²⁺ inhibited proton current in activated eosinophils with 20-fold greater potency than in unstimulated cells (5), although this conclusion likely reflected the experimental design rather than a genuine difference in efficacy (7). By the same type of analysis performed for whole-cell studies, the Zn²⁺ sensitivity was similar in basophils studied in perforated-patch configuration after PMA stimulation. The average shift of the g_H–V relationship was 25 ± 6 mV (n = 4) at 1 μM Zn²⁺, 45 ± 4 mV (n = 5) at 10 μM Zn²⁺, and 67 ± 8 mV (n = 3) at 100 μM Zn²⁺. The slightly lower efficacy in these measurements may reflect the lower pH_o of 7.0 compared with pH_o 7.4 for the whole-cells studies. Competition between H⁺ and Zn²⁺ reduces the apparent potency of Zn²⁺ at lower pH_o (18). Thus, the affinity of externally applied Zn²⁺ for proton channels in basophils is similar before or after stimulation.

Histamine Release Is Inhibited by Zn²⁺.

To explore whether proton channel activity is involved in basophil function, we measured histamine release in the presence of the proton channel inhibitor, Zn²⁺. Zn²⁺ inhibits histamine release stimulated by several agonists (19), but PMA stimulation had not been explored. Fig. 4 shows that Zn²⁺ inhibited histamine release elicited by anti-IgE (●) or by PMA (□). Inhibition was more potent with anti-IgE as a stimulus (IC₅₀ ≈ 20 μM) than with PMA (IC₅₀ ≈ 90 μM). Zn²⁺ also inhibited histamine release in basophils stimulated with PMA in a nominally Ca²⁺-free solution (1 mM Mg²⁺ and no added Ca²⁺), by 76 ± 12% (n = 3) at 100 μM Zn²⁺.

Fig. 4.

Inhibition of histamine release by Zn²⁺. Histamine release induced by 0.3 μg/ml anti-IgE (●) or 60 nM PMA (□) in the presence of various concentrations of ZnCl₂, normalized to release in the absence of ZnCl₂. Each value is the mean ± SEM of measurements in three to five experiments using blood from different donors. Spontaneous release was 10.0 ± 0.9% (mean ± SEM, n = 5) of total histamine content (range 7.7–12.9%), the PMA stimulated release was 65 ± 7% (42–82%, n = 5 except n = 3 at 10 μM Zn²⁺), and the anti-IgE stimulated release was 24 ± 6% (13–34%, n = 3). The values for the two stimuli at 30 μM and 100 μM differ significantly (P < 0.001 and P < 0.05, respectively, by Student's t test).

Proton Channels Extrude Acid Generated During the Anti-IgE Response.

If proton channels are active in stimulated basophils, Zn²⁺ should prevent acid extrusion. Fig. 5 shows the anti-IgE response of several basophils examined by fluorescence imaging using confocal microscopy with SNARF labeling and the shifted excitation and emission ratioing of fluorescence (SEER) method (20). The graph shows the average time course of [H⁺]_i in many individual cells. In Ringer's solution, anti-IgE resulted in acidification (○) that was more profound in the presence of 100 μM Zn²⁺ (●). Zn²⁺ had no consistent acute effect on unstimulated cells. In 5 other experiments at temperatures ranging from 20°C to 30°C, anti-IgE produced greater acidification with less tendency to recover in the presence of Zn²⁺. Evidently, a Zn²⁺ sensitive proton extrusion mechanism in the plasma membrane is active during IgE-mediated responses.

Fig. 5.

Average [H⁺]_i in basophils (99% pure) stimulated with 1 μg/ml anti-IgE in the absence (○) or presence of 100 μM Zn²⁺ (●) at ≈30°C and imaged by using confocal microscopy and SEER. The mean ± SEM of 25 control cells and 46 cells in Zn²⁺ is shown, with all data pairs after the star significantly different by Student's t test (P < 0.05). Pseudocolor images indicating [H⁺]_i in trios of basophils from this experiment taken after stimulation with anti-IgE at the indicated time points are shown in two rows; the top row is control, the lower in the presence of Zn²⁺.

Discussion

Proton Currents in Human Basophils Are Enhanced by PMA, Anti-IgE, and fMLF.

Proton channel gating in human basophils was enhanced profoundly by PMA stimulation. As occurs in neutrophils (6), PLB-985 cells (8), and eosinophils (5, 7), I_H and g_H,max increased, τ_act decreased, and V_threshold was shifted negatively (Table 1). All of these changes tend to increase proton current at any given voltage. Although other changes were generally similar, the shift in V_threshold was only half as large in basophils as in human neutrophils studied under identical conditions, −20 mV compared with −39 mV, respectively, and there was no clear slowing of τ_tail in basophils, in contrast to the dramatic slowing seen in neutrophils and other phagocytes. The response pattern of basophils is strikingly reminiscent of that in phagocytes that lack gp91^phox expression, including cells from chronic granulomatous disease patients and gp91^phox knockout PLB-985 cells (8). In contrast, PMA has no clear effect on proton currents in rat alveolar epithelial cells (6) or in HEK-293 or COS-7 cells transfected with the human or mouse proton channel gene (21). There appears to be a full response in phagocytes (neutrophils, PLB-985 cells, and eosinophils), a partial response in leukocytes (including basophils) lacking NADPH oxidase, and little or no response in other cells. Perhaps this is coincidental, but the pattern is striking nonetheless.

Proton current responses in human basophils speak against the hypothesis that the gp91^phox component of NADPH oxidase is a functioning proton channel (5, 22). That large proton currents are present in basophils (1) that lack detectable cytochrome b₅₅₈, which contains gp91^phox (12) establishes that gp91^phox is not the proton channel in unstimulated basophils. In the present study, we show that PMA increased g_H,max as much in basophils that lack gp91^phox as it did in eosinophils and neutrophils with ample gp91^phox. Therefore, the enhanced proton conductance in phagocytes does not reflect the activity of gp91^phox.

More surprising than the PMA response was the finding that stimulation of basophils through receptor-mediated pathways by anti-IgE or fMLF also enhanced proton currents. Qualitatively, the anti-IgE response was identical to that elicited by PMA, but it was less profound and was often not sustained. In cells with a clear anti-IgE response, subsequent addition of PMA invariably produced an augmented response. There are striking similarities between the stimulation of proton channels and histamine release from basophils. There was marked variability among donors in the responsiveness of their basophils to anti-IgE, with proton channel response rates ranging from 0 to 69%. Nearly all cells from all donors responded to PMA. Donor-to-donor variability in histamine release in response to anti-IgE (14, 15) mirrors the variability in proton channel responsiveness. In the general population, the distribution of responsiveness of basophils to activation by anti-IgE is broad, ranging 100-fold, but basophils respond well to PMA even in individuals whose basophil response to anti-IgE is not detectable (14). In addition, the proton channel response occurred after a delay reminiscent of the distinct delay that precedes the increase in [Ca²⁺]_i (23, 24) and the release of histamine in basophils stimulated with anti-IgE (17) or ragweed pollen antigen (25).

Pathways That Activate Histamine Release and Proton Channels in Basophils Are Not Identical and Vary According to Agonist.

Histamine release from basophils elicited by PMA is inhibited by the selective PKC inhibitor GFX. In contrast, anti-IgE stimulated histamine release is insensitive to GFX, and inhibition by several other PKC inhibitors is attributable to their effects on other kinases (15, 26). PKC activity has been implicated in both activating and de-activating signaling in both rat basophilic leukemia cells and human basophils. Therefore, inhibitors of the broader family of PKC isozymes might produce a weak effect on secretion, due to competing inhibition of both activating and de-activating mechanisms. GFX reversed the activation of proton channels by PMA or anti-IgE, supporting a role for phosphorylation in proton channel activation by either agonist. The anti-IgE response was only partially reversed, suggesting involvement of an additional pathway. The lack of capacitance changes during proton channel responses together with previous studies of the time course of histamine release (17) indicate that proton channel enhancement likely precedes degranulation.

The proton channel inhibitors Zn²⁺ and La³⁺ inhibit histamine release from basophils stimulated with anti-IgE or A23187 (19, 27), raising the possibility that proton channels are involved. However, both metal cations also inhibit calcium release activated calcium (CRAC) currents (28) and Na⁺/Ca²⁺ exchange (29). Hypothetically, either transporter, if present in basophils, might mediate the Ca²⁺ influx that is required for histamine release in response to anti-IgE (30). Although CRAC currents have not been recorded in basophils, they exist in human neutrophils (31), and CRACM1 protein is expressed in basophils (D.W.M., unpublished data). However, Zn²⁺ inhibits histamine release from rat basophilic leukemia cells more potently than it prevents Ca²⁺ influx, suggesting that it acts by a different mechanism (32). Surprisingly, Ca²⁺ influx stimulated by anti-IgE in human basophils was not inhibited by up to 300 μM Zn²⁺ (D.W.M., unpublished data). This result suggests that in human basophils, CRAC channels either have low Zn²⁺ sensitivity, or they do not mediate the anti-IgE induced Ca²⁺ influx required for histamine release.

To isolate proton channel involvement in histamine release, we examined the Zn²⁺ sensitivity of histamine release induced by PMA, which occurs without Ca²⁺ influx (13) or [Ca²⁺]_i increases (26, 33); in fact, PMA inhibits CRAC currents (34). The proton channel responses to either PMA or anti-IgE occurred in Ca²⁺-free solutions and were indistinguishable from responses in Ca²⁺-containing solutions and hence are independent of Ca²⁺ influx. That histamine release stimulated by PMA was inhibited by Zn²⁺ (Fig. 4) suggests that proton channel activity may be required. The inhibition by Zn²⁺ seemed to be more potent for anti-IgE (IC₅₀ ≈ 20 μM) than for PMA-stimulated (IC₅₀ ≈ 90 μM) responses, consistent with a previous study in which histamine release elicited by anti-IgE and fMLF was inhibited by Zn²⁺, with IC₅₀ 10 μM and 40 μM, respectively (19). The variable sensitivity to Zn²⁺ was attributed to each agonist activating histamine release by a different mechanism. Because the anti-IgE response requires Ca²⁺ influx, but the responses to fMLF and PMA do not (30, 33), the anti-IgE response might normally involve a combination of proton channels and CRAC channels (or whatever mediates Ca²⁺ influx), whereas the PMA and fMLF responses may reflect Zn²⁺ acting exclusively on proton channels.

The Zn²⁺ data suggest that proton channels are required for histamine release. Proton channels might facilitate histamine release by compensating charge, as they do in phagocytes (3, 4), or by modulating pH. Outward H⁺ current could compensate for electrogenic Ca²⁺ influx, which is required for anti-IgE-stimulated histamine release. However, during the PMA response, which does not involve Ca²⁺ influx, proton channels must act by a different mechanism. Histamine release consumes metabolic energy (13, 25, 27) and thus likely generates cytoplasmic acid. Consistent with this interpretation, the results in Fig. 5 showed acidification after anti-IgE stimulation, which has not been reported previously. Importantly, Zn²⁺ unmasked occult proton efflux during the response, thereby indicating that proton channels are most likely responsible. Other plasma membrane proton extrusion mechanisms (Na⁺/H⁺-antiport and H⁺-ATPase) are insensitive to Zn²⁺ (35, 36). Together, these results indicate that voltage-gated proton channels in basophils may help keep pH_i in an optimal range for histamine secretion.

Methods

Isolation of Human Basophils.

Basophil-containing mononuclear cell fractions were isolated by density gradient centrifugation from freshly drawn venous blood of healthy adult donors according to a protocol approved by the Institutional Review Board of Rush University Medical Center. Basophils were enriched by negative selection using immunomagnetic beads (Basophil Isolation Kit; Miltenyi Biotec, Auburn, CA) and the miniMacs magnetic cell separation system (Miltenyi Biotec) according to the manufacturer's instructions; cells enriched by this method are 59–88% basophils (37). For some experiments basophils were further purified by countercurrent elutriation and two-step Percoll density gradient followed by negative selection (24), resulting in 98–99% purity by Alcian blue staining. The basophils were suspended in PBS with 2 mM EDTA and 0.5% BSA. Freshly isolated cells and cells maintained for up to 2 days at 4°C were used for electrophysiologic measurements.

Electrophysiology.

Standard whole-cell or perforated-patch recording was performed as described previously (16). Spherical, nonadherent cells 7–8 μm in diameter (1.68 ± 0.49 pF, mean ± SD, n = 147) were selected. Seals were formed with Ringer's solution (in mM: 160 NaCl, 4.5 KCl, 2 CaCl₂, 1 MgCl₂, 5 Hepes, pH 7.4) in the bath, and the potential zeroed after the pipette was in contact with the cell. For perforated-patch recording, the pipette and bath solutions contained 110 mM tetramethylammonium methanesulfonate, 50 mM NH₄⁺ in the form of 25 mM (NH₄)₂SO₄, 2 mM MgCl₂, 10 mM BES buffer, and 1 mM EGTA and were titrated to pH 7.0 with tetramethylammonium hydroxide. A Ca²⁺-containing bath solution was identical, but with 1.5 mM CaCl₂ added. The NH₄⁺ in bath and pipette solutions “clamps” pH_i near pH_o (6, 38). The pipette solution included ≈500 μg/ml solubilized amphotericin B (≈45% purity; Sigma); the pipette tip was dipped briefly into amphotericin-free solution. For whole-cell recording, bath and pipette solutions contained 100–200 mM buffer, 1–2 mM CaCl₂ or MgCl₂ (pipette solutions were Ca-free), 1–2 mM EGTA, and TMAMeSO₃ to adjust the osmolality to ≈300 mOsm, titrated with tetramethylammonium hydroxide or methanesulfonate. No liquid junction potential or leak correction has been applied. Experiments were performed at room temperature (20–25°C). Bioactive substances (PMA, anti-IgE, or GFX) were introduced by complete bath changes or were added directly to the bath with subsequent stirring. Nominal concentrations were calculated according to approximate bath volume, and stirring may not have always been complete; thus, stated concentrations are somewhat approximate. Basophils were stimulated with affinity purified goat anti-human IgE (Immunology Consultants Laboratory).

Histamine Release.

Basophil-containing mononuclear cell fractions were isolated from venous blood of healthy adult donors as described above. The cells were suspended in Ringer containing 5 mM glucose, and 0.003% BSA was added directly to each sample before incubation. Contaminating metals in Millipore purified water had been removedby using Chelex 100 Resin (Bio-Rad Laboratories) before adding salts and glucose. The cells (20,000–45,000 basophils as determined by Alcian blue staining) were incubated with the indicated concentrations of anti-IgE or PMA alone and in the presence of the indicated concentrations of ZnCl₂ for 45–60 min at 37°C in round-bottom polystyrene tubes. Total incubation volume was 1 ml. Reactions were stopped by ice. After centrifugation, histamine content in the supernatants was measured by an automated fluorometric assay (39). Spontaneous histamine release was measured by using cells incubated in the absence of stimulus. Total histamine content was measured in supernatants of cells lysed with 1.4% perchloric acid. Results are expressed as histamine released above the spontaneous value.

Dynamic Imaging of the Cytosolic pH of Individual Basophils by SEER Imaging.

Human basophils were allowed to adhere to glass coverslips, incubated with 10 μM 5-(and 6-)carboxy SNARF-1 in HBSS for 15–20 min at room temperature, and washed. SEER imaging (20) was performed by simultaneously acquiring 2 confocal images: F₁, excited at 514 nm and emitted at 500–604 nm and F₂ (excited at 594 nm and emitted at 620–715 nm). The ratio F₁/F₂ monitors [H⁺] with a dynamic range of ≈150.