Glucocorticoid modulation of Ca²⁺ homeostasis in human B lymphoblasts

Abstract

1. We determined the effect of cortisol (200 nm for 48 h) on the intracellular Ca²⁺ concentration ([Ca²⁺]_i) and parameters of Ca²⁺_i signalling in 19 lymphoblastoid cell lines (LCLs).

2. Using the fluorescent dye fura-2, the basal [Ca²⁺]_i in Ca²⁺-containing medium was 63.5 ± 2.4 nm in vehicle (ethanol)-treated LCLs and 55.7 ± 2.6 nm (mean ±s.e.m.) in cortisol-treated LCLs.

3. Ca²⁺_i signalling following platelet-activating factor (PAF, 100 nm) addition was enhanced by cortisol treatment, with LCLs having small PAF responses showing the largest percentage increase after cortisol treatment. Mean peak [Ca²⁺]_i responses to PAF were enhanced 67.0% and 55.7% in Ca²⁺-free and Ca²⁺-containing medium, respectively.

4. The endoplasmic reticulum Ca²⁺-ATPase inhibitor thapsigargin (100 nm) caused a transient increase in [Ca²⁺]_i in Ca²⁺-free medium in which the peak change was increased in cortisol-treated cells (98.5 ± 5.8 vs. 79.8 ± 4.5 nm). Peak changes in the freely exchangeable Ca²⁺ in response to 5 μm ionomycin were also enhanced in cortisol-treated cells (923.7 ± 113.9 vs. 652.2 ± 64.5 nm) and correlated to the PAF-evoked [Ca²⁺]_i response.

5. Cortisol-treated LCLs exposed to thapsigargin to empty intracellular Ca²⁺ stores (10 min treatment in Ca²⁺-free medium) and exposed to CaCl₂ or MnCl₂ had a greater rate of Ca²⁺ entry (18.6 ± 1.8 vs. 13.8 ± 1.5 nm s⁻¹) and higher rate constant for Mn²⁺ entry (0.0345 ± 0.0029 vs. 0.0217 ± 0.0020) than vehicle-treated cells. Peak [Ca²⁺]_i in cells exposed to CaCl₂ was also enhanced (869.4 ± 114.7 vs. 562.6 ± 61.7 nm). Parameters of divalent cation influx were highly correlated to the peak [Ca²⁺]_i elicited by thapsigargin or ionomycin.

6. Inclusion of RU 486 (a glucocorticoid antagonist) with cortisol prevented the decrease in basal [Ca²⁺]_i and stimulatory actions of cortisol on all Ca²⁺_i parameters. RU 486 alone had no apparent effects on basal [Ca²⁺]_i or ${\text{Ca}}_{\text{i}}^{\text{2+}}$ signalling.

7. Based on data obtained over a wide range of responses (in the presence and/or absence of cortisol or RU 486), the results show that cortisol stimulation of glucocorticoid receptors decreases basal [Ca²⁺]_i and enhances PAF-evoked [Ca²⁺]_i signalling, most probably through its effects on intracellular Ca²⁺ stores. In turn, the extent of Ca²⁺ entry via store-operated plasma membrane Ca²⁺ channels is closely linked to the size of the Ca²⁺ stores.

Stimulation of B lymphocytes leads to diverse cellular functions including secretion, proliferation, differentiation or programmed cell death (apoptosis). Inflammatory agents binding or cross-linking surface membrane receptors increase protein kinase/phosphatase activities and activate phospholipase C, generating inositol 1,4,5-trisphosphate (IP₃). IP₃ mobilizes Ca²⁺ from intracellular Ca²⁺ stores and produces a transient rise in cytosolic Ca²⁺ (Ca²⁺₁). The emptying of Ca²⁺ from IP₃-sensitive Ca²⁺ stores also results in enhanced Ca²⁺ influx across the plasma membrane, which is manifested as an enhancement of the Ca²⁺₁ transient, oscillatory increases in Ca²⁺₁, or a secondary, sustained increase in [Ca²⁺]_i. This Ca²⁺ entry pathway, termed store-operated Ca²⁺ entry (SOCE), is responsible for refilling intracellular Ca²⁺ stores (Putney, 1990) and is an important regulator of Ca²⁺-dependent gene expression (Negulescu et al. 1994; Dolmetsch et al. 1998; for review, see Putney & Bird, 1993). In B lymphocytes, elevation of [Ca²⁺]_i serves as a differentiation signal towards antibody-secreting cells (Clevers et al. 1985; Huang et al. 1995) and enables interleukin-2 (IL-2) synthesis (Taira et al. 1987) and cell proliferation (Ambrus et al. 1991). The inhibition of Ca²⁺ influx via the SOCE pathway results in marked decreases in [Ca²⁺]_i and IL-2 (Diegel et al. 1994), supporting the hypothesis that the B cell response is related to the degree of [Ca²⁺]_i elevation (Clark & Lane, 1991). Additional changes in Ca²⁺₁ regulation in anti-IgM-stimulated malignant B-1 cells suggest that dysfunctional Ca²⁺₁ signalling may increase the susceptibility for these cells to undergo apoptosis (Dang et al. 1995). Thus, the regulation of [Ca²⁺]_i represents a key factor in the control of B lymphocyte development and function.

Cortisol and synthetic glucocorticoids such as dexamethasone are important modulators of the immune system (Cupps & Fauci, 1982), and exert both inhibitory and stimulatory actions on lymphocyte development and cellular responses to inflammatory stimuli (Guyre et al. 1988; Wilckens & De Rijk, 1997). In murine precursor B cells, slight elevation of the circulating level of glucocorticoids decreased the number of cycling cells by inducing apoptosis (Garvey et al. 1993). In B cells undergoing lymphopoiesis, negative regulatory effects of glucocorticoids were countered by the presence of (T) lymphocyte-supportive stromal cells (Borghesi et al. 1997). Glucocorticoids produce variable responses in cell surface Ia (sIa) expression, the response being dependent on the duration of exposure and the state (i.e. resting or activated) of the B cell (McMillan et al. 1988). In a report examining cell activation of resting B cells, cortisol acted as an inhibitor of anti-Ig-antibody-mediated IP₃ synthesis, Ca²⁺₁ signalling and entry into the cell cycle (Dennis et al. 1987). These results suggest that, similar to T lymphocytes, cortisol effects on B cell function are dependent on the extent of glucocorticoid exposure and the context in which inflammatory mediators are presented. In light of the significant influence of Ca²⁺ homeostasis on B cell proliferation and differentiation, widely different effects of glucocorticoids on Ca²⁺₁ signalling and B cell physiology could be expected at different stages of B lymphocyte development.

To examine the glucocorticoid regulation of Ca²⁺ homeostasis in activated B lymphocytes, we studied cortisol-induced alterations of Ca²⁺ signalling in Epstein- Barr virus (EBV)-transformed B lymphoblasts. In vitro transformation of peripheral blood cells by EBV leads to partial differentiation and immortalization of cells of the B lymphocyte lineage, lending to their characterization as a model of activated B cells. EBV-generated lymphoblastoid cell lines (LCLs) are often used where stable, proliferating cultures of pure B cells are required. Their growth in tissue culture surmounts obstacles often associated with freshly isolated cell populations, including low cell number, deleterious effects of cell isolation chemicals or procedures, and exposure to leucocyte-derived cytokines or hormones. Several groups have suggested that LCLs exhibit signalling and membrane ion transport properties common to non-transformed cells (Schulam et al. 1990; Kuruvilla et al. 1993; Kojima et al. 1994; Siffert et al. 1995), supporting the idea that independent of viral transformation, LCLs retain membrane signalling properties representative of B lymphocytes. Thus, to study the role of cortisol on Ca²⁺ homeostasis in activated B cells, we examined parameters of receptor-mediated [Ca²⁺]_i signalling, intracellular Ca²⁺ stores and Ca²⁺ influx across the plasma membrane (SOCE) in LCLs.

METHODS

Chemicals

Unless otherwise specified, the chemicals used in this study were obtained from Sigma. Fura-2-acetoxymethylester (fura-2 AM) was from Molecular Probes; ionomycin and platelet-activating factor (PAF) were from Calbiochem; thapsigargin (TG) was from Alexis Corporation; mifepristone (RU 486) was from Biomol Research Laboratories, Inc.

Culture of B lymphoblasts

LCLs from the Centre d'Etude du Polymorphisme Humain (CEPH) collection (Coriell Institute for Medical Research, Camden, NJ, USA) were cultured as previously described (Brzustowicz et al. 1997) in RPMI 1640 culture medium containing 2 mM L-glutamine, 100 u ml⁻¹ penicillin, 100 μg l⁻¹ streptomycin and 15 % heat-inactivated fetal bovine serum (Irvine Scientific). Cells were passaged 2-3 times per week, and cell density was maintained between 0.5 × 10⁶ and 1 × 10⁶ cells ml⁻¹. Cells used for experiments were derived from fresh cultures thawed from frozen stocks at 3-4 month intervals. Cells were routinely monitored for mycoplasma contamination with the Mycotrim TC Triphasic Culture System (Irvine Scientific).

In preparation for experimental protocols, 10⁷ cells (0.5 × 10⁶ ml⁻¹) were resuspended for 48 h in fresh culture medium containing glucocorticoid (200 nM cortisol), glucocorticoid receptor (GR) antagonist (RU 486: Gagne et al. 1985) or vehicle (ethanol, final concentration of 0.04 %). Cell numbers were determined with a Coulter Counter ZM.

Measurements of [Ca²⁺]_i

Lymphoblasts were washed at 37°C with Hepes-buffered solution (HBS) comprising (mM): 140 NaCl, 5 KCl, 1 CaCl₂, 1 MgCl₂, 20 N-2-hydroxyethyl-piperazine-N′-2-ethanesulfonic acid (Hepes), 10 glucose (pH 7.4), and 0.1 % fatty acid-free bovine serum albumin. Cells (2 × 10⁷ ml⁻¹) were then incubated for 30 min at 37°C with 2 μm fura-2 AM and 0.125 mM sulfinpyrazone (an anionic transport inhibitor used to retard fura-2 extrusion: Di Virgilio et al. 1988) as previously described (Dang et al. 1995; Brzustowicz et al. 1997). Aliquots of lymphoblasts (2 × 10⁶ (100 μl)⁻¹) were centrifuged to remove extracellular fura-2, washed once with HBS and resuspended in a cuvette containing 3 ml of either Ca²⁺-free HBS (0.3 mM ethylene glycol-bis (β-aminoethyl ether)-N,N,N′,N′-tetraacetic acid (EGTA) substituted for CaCl₂) or HBS. [Ca²⁺]_i measurements were performed in a CM3 spectrophotometer equipped with stirring and temperature control (Instruments SA, Inc.). Excitation wavelengths were set at 340 nm/380 nm and emission wavelength at 505 nm. [Ca²⁺]_i was calculated from the 340 nm/380 nm fluorescence ratio (R) according to the equation described by Grynkiewicz et al. (1985). Calibration was achieved by treating cells with 100 μm digitonin in the presence of 1 mM CaCl₂ (R_max), followed by 15 mM EGTA (pH 8.0; R_min). In some experiments, TG, a sarco(endo)plasmic reticulum Ca²⁺-ATPase (SERCA) inhibitor (Lytton et al. 1991), was used to inhibit Ca²⁺ reuptake by intracellular Ca²⁺ stores. Autofluorescence of unloaded cells in HBS was subtracted from the cellular fluorescence of the dye prior to all calculations for [Ca²⁺]_i.

Mn²⁺ uptake assay

Mn²⁺ was used as a Ca²⁺ surrogate to study divalent cation entry (Hallam & Rink, 1985). Fura-2-loaded cells were centrifuged, resuspended in 500 μl microfuge tubes in Ca²⁺-free HBS and treated with vehicle (DMSO; 0.5 % final volume) or TG (500 nM) for 10 min. Cells were recentrifuged and suspended in cuvettes containing Ca²⁺-free medium; after 10 s, 0.35 mM MnCl₂ was added. Fluorescence was monitored at excitation and emission wavelengths of 360 nm and 505 nm. Data were normalized using 360 nm values obtained immediately prior to addition of MnCl₂. Autofluorescence of Mn²⁺-treated (1 mM), digitonin-permeabilized cells and buffer for each experiment was subtracted from the fluorescence record prior to normalization.

Data analysis and statistics

Mn²⁺ quenching of fura-2 (percentage fluorescence quenching at times 0, 10, 30 and 60 s compiled from individual experiments of vehicle- and TG-treated cells) was best-fitted to the monoexponential function:

where A is the cellular pool at time t, a₁ is the fura-2 fluorescence before treatment with Mn²⁺, and k₁ is the rate constant. To obtain an estimate of the TG-sensitive rate of Mn²⁺ influx, Mn²⁺ quenching in DMSO-treated cells was subtracted from quenching in TG-treated cells and fitted to the equation:

in which parameters were defined as above and a constraint on the maximal difference between DMSO- and TG-treated cells (45 %, cf. Fig. 4B) was incorporated.

Figure 4. Cortisol treatment increases Mn²⁺ entry in TG-treated cells.

A, representative experiments are depicted in which fura-2-loaded cells were centrifuged and pretreated with 500 nM TG (TG-labelled traces) or vehicle (D-labelled traces, for 0.5 % DMSO) in Ca²⁺-free medium for 10 min. Cells were then centrifuged and the pellet resuspended in 3 ml Ca²⁺-free medium; traces show the normalized response of fluorescence decline immediately after addition of 0.35 mM MnCl₂ (arrows). B, data used to calculate the rate constant for TG-sensitive quenching of fura-2 in vehicle (left panel) and cortisol-treated (right panel) LCLs (n= 18) as described in Methods and legend to Table 1. Shown are means ±s.e.m. of LCLs evaluated for Mn²⁺ entry; s.e.m.s smaller than the symbol are not shown. Mn²⁺ entry in vehicle- (□, DMSO) and TG-treated cells (○); TG-sensitive Mn²⁺ influx is given by (▵), and curve-fitted lines describing the exponential decay of fluorescence are given for each condition. C, results of weighted least-squares analysis of k₁ (rate constant, see also Table 1) indicated a P value < 0.01 between vehicle- and cortisol-treated LCLs.

Repeated measures analysis of variance (ANOVA) was performed with an IBM-compatible PC using the general linear model procedure (PROC GLM, Statistical Analysis Systems (SAS) Institute, Cary, NC, USA). This analysis controlled between-cell line variability, which, as shown for PAF-evoked Ca²⁺₁ transients and other parameters of Ca²⁺ signalling, varies widely between LCLs from different individuals. Duncan's multiple range test was used to assess differences between treatment groups. Correlation analysis was performed using SAS PROC CORR. Non-linear regression analysis for computations of parameters of Mn²⁺ quenching of fura-2 quench experiments utilized SAS REG. Statistical analyses of the rate constants derived from curve-fitting the Mn²⁺ data were performed using weighted least squares according to the method described by Johnson & Milliken (1983). P values < 0.05 were taken as statistically significant. Data are reported as mean values ±s.e.m. for n LCLs.

RESULTS

Effects of cortisol on basal [Ca²⁺]_i and receptor-mediated [Ca²⁺]_i transient

To examine Ca²⁺₁ signalling in LCLs, we used platelet-activating factor (PAF), a potent cytokine that has diverse actions throughout the immune system (Snyder, 1990). In B cells, this lipid mediator alters several immunomodulatory functions including nuclear-factor-κ B-mediated immunoglobulin synthesis and secretion (Smith & Schearer, 1994; Rosskopf et al. 1995), cytokine expression (e.g. TNFα and IL-2; Smith et al. 1994) and Ca²⁺-dependent expression of cell cycle genes and proliferation (Mazer et al. 1991, 1994; Rosskopf et al. 1995). The PAF receptor is expressed in both resting and activated cells (Nguer et al. 1992, 1996); its activation results in phosphatidylinositol bisphosphate hydrolysis, IP₃ generation and a transient Ca²⁺₁ response (Mazer et al. 1991). Previous work by the present authors and Siffert's group showed that the PAF-evoked [Ca²⁺]_i response in lymphoblasts displays marked variability between cell lines established from different individuals (Siffert et al. 1995; Brzustowicz et al. 1997). Based on widely differing [Ca²⁺]_i responses to PAF, we selected a panel of 19 LCLs from the CEPH collection to study the effects of cortisol on Ca²⁺₁ homeostasis. For these experiments, the effects of cortisol on the PAF-evoked [Ca²⁺]_i response, intracellular Ca²⁺ store size and SOCE were assessed. In addition, we determined whether modulation of Ca²⁺₁ occurred through the glucocorticoid-GR complex by performing the study in the presence of RU 486.

Preliminary experiments showed that 200 nM cortisol added 1 or 15 min prior to 100 nM PAF did not alter basal [Ca²⁺]_i or the PAF-evoked Ca²⁺₁ response. Furthermore, enhancement of the PAF responses was observed only after 12-24 h treatment with this concentration of cortisol (data not shown). We therefore performed all experiments after 48 h of exposure of cells to 200 nM glucocorticoid or inhibitor. Figure 1A shows examples of the PAF-evoked Ca²⁺₁ response in cortisol- or vehicle-treated cells suspended in HBS. Cells treated with glucocorticoid exhibited larger Ca²⁺₁ transients over the entire range of agonist concentration, with no apparent alteration of sensitivity or time course of response. As peak Ca²⁺₁ signals at 100 nM PAF are maximal and highly reproducible for a given LCL (Brzustowicz et al. 1997), this concentration of agonist was used to evaluate the influence of cortisol on Ca²⁺₁ signalling. In Ca²⁺-containing medium, basal [Ca²⁺]_i in cortisol-treated cells was slightly decreased (55.7 ± 2.6 nM) compared with vehicle-treated cells (63.5 ± 2.4 nM, P < 0.05, n= 19). LCLs treated with RU 486, or the combination of RU 486 and cortisol, exhibited a basal [Ca²⁺]_i that was not significantly different from that of ethanol-treated cells.

Figure 1. Cortisol treatment enhances PAF-evoked ${\text{Ca}}_{\text{i}}^{\text{2+}}$ signalling.

A, example of ${\text{Ca}}_{\text{i}}^{\text{2+}}$ transients in an LCL exposed to 1, 10 or 100 nM PAF in HBS. 10 × 10⁶ cells were treated with vehicle (ethanol) or cortisol (200 nM) for 48 h and processed for ${\text{Ca}}_{\text{i}}^{\text{2+}}$ measurements as described in Methods. Basal [Ca²⁺]_i was determined as the mean [Ca²⁺]_i value immediately preceding addition of agonist, and the PAF-evoked [Ca²⁺]_i response was calculated as the difference between basal [Ca²⁺]_i and peak [Ca²⁺]_i after PAF addition. For this LCL, PAF addition (arrow) to cortisol-treated cells resulted in peak [Ca²⁺]_i increases of 123 % (1 nM), 163 % (10 nM) and 158 % (100 nM) compared with vehicle-treated cells. B and C, PAF-evoked [Ca²⁺]_i responses in LCLs suspended in Ca²⁺-free and Ca²⁺-containing HBS. Left panels: percentage stimulation of PAF-evoked [Ca²⁺]_i response in cortisol-treated cells (Δ[Ca²⁺]_i(C)) to vehicle-treated cells (Δ[Ca²⁺]_i(V)) as a function of control response. Right panels: mean changes in PAF-evoked [Ca²⁺]_i in cells treated for 48 h with vehicle, cortisol (200 nM), RU 486 (200 nM), or cortisol (200 nM) and RU 486 (200 nM) (n= 19). ANOVA indicated overall P values < 0.0001 for both Ca²⁺-free and Ca²⁺-containing HBS, with significant differences between cortisol and all other treatments.

Figure 1B and C shows the extent of stimulation following cortisol treatment on peak PAF-evoked [Ca²⁺]_i responses, in Ca²⁺-free and Ca²⁺-containing HBS. In the left panels, the range of responses in vehicle-treated cells is graphed as the dependent variable, and the percentage stimulation following cortisol treatment is shown on the ordinate. LCLs exhibiting small PAF responses demonstrated the greatest percentage stimulation following glucocorticoid treatment. This relation was maintained for individual LCLs in Ca²⁺-free or Ca²⁺-containing HBS, as analysis of the percentile stimulation of the cortisol-enhanced PAF responses in the two conditions showed a high degree of correlation (r= 0.8890, P < 0.0001, n= 19). Additional experiments showed the lack of cortisol responsiveness in LCLs with inherently large responses was not a result of the high dose of agonist used; one cell line that showed no enhancement following cortisol treatment similarly failed to show increased [Ca²⁺]_i responses at lower (1-10 nM) PAF concentrations (data not shown). The right panels of Fig. 1B and C show that the mean PAF-evoked [Ca²⁺]_i response for 19 LCLs was significantly increased by cortisol treatment (by 67.0 and 55.7 % in Ca²⁺-free and Ca²⁺-containing HBS, respectively). RU 486 had no discernible effect on the PAF response, and the stimulatory actions of cortisol were absent in cells treated with equimolar concentrations of RU 486 and glucocorticoid.

Effects of cortisol on intracellular Ca²⁺ stores

PAF responses in cortisol-treated cells were increased in the absence of extracellular Ca²⁺ influx (Fig. 1B). This result suggests that the size of the IP₃-releasable Ca²⁺ pool(s) was altered by glucocorticoid treatment. We examined whether cortisol increased intracellular Ca²⁺ stores by challenging LCLs with TG and ionomycin, two agents that mobilize Ca²⁺ from Ca²⁺ stores independently of IP₃. TG, a sarco(endo)-plasmic reticulum (SER) Ca²⁺-ATPase (SERCA) inhibitor (Lytton et al. 1991), inhibits SERCA-mediated Ca²⁺ uptake into the SER. Its addition to lymphoblasts in Ca²⁺-free HBS results in a transient increase in Ca²⁺₁, owing to an endogenous leak of Ca²⁺ from these stores and subsequent extrusion of Ca²⁺ across the plasma membrane. Alternatively, addition of the Ca²⁺ ionophore ionomycin to cells results in its insertion into intracellular membranes and the electroneutral exchange of Ca²⁺ for H⁺. Addition of 100 nM ionomycin to cells in Ca²⁺-free HBS results in a rapid increase in [Ca²⁺]_i that serves as an estimate of the freely exchangeable Ca²⁺ (FEC). Cortisol treatment (48 h, 200 nM) resulted in significant increases in the TG-mediated rise in [Ca²⁺]_i (Fig. 2A) and the FEC (Fig. 2B). The left panels illustrate the protocols used to obtain these parameters, and the right panels indicate the mean values for the four treatment groups. Treatment with the glucocorticoid antagonist had no effects of its own (on either parameter), and treatment of cells with RU 486 eliminated cortisol-induced increases in the TG-mediated rise in [Ca²⁺]_i and FEC. The TG-evoked rise in [Ca²⁺]_i appeared to reflect the FEC measurement, as comparison of these two parameters gave a correlation coefficient (r) of 0.5622 (P < 0.0001, n= 71). These data suggest that cortisol enhances SERCA-regulated and freely exchangeable intracellular Ca²⁺ stores in LCLs.

Figure 2. Cortisol treatment enhances intracellular Ca²⁺ stores in LCLs.

The left panels show representative examples of the TG-evoked [Ca²⁺]_i response (A) and the ionomycin-TG-evoked rise in [Ca²⁺]_i (FEC) (B) in LCLs treated as described in the legend to Fig. 1. Arrows indicate addition of 100 nM TG (A) or 5 μm ionomycin-100 nM TG (B) to cells suspended in Ca²⁺-free medium; TG was added to prevent reuptake of Ca²⁺ and diminution of the [Ca²⁺]_i transient. The peak [Ca²⁺]_i response minus the basal [Ca²⁺]_i was used to estimate TG-releasable Ca²⁺ pools defined by the SER and FEC. The right panels show the results from 19 (A) and 17 (B) LCLs. ANOVA indicated overall P values < 0.0001 (TG-evoked [Ca²⁺]_i) and < 0.0025 (ionomycin-TG-evoked [Ca²⁺]_i), with significant differences between cortisol and all other treatment conditions.

Effects of cortisol on parameters of SOCE

The size, or capacity, of intracellular Ca²⁺ stores is tightly linked to the control of Ca²⁺ entry from the external medium. As cortisol enhanced intracellular Ca²⁺ stores, we examined whether a similar upregulation of store-operated Ca²⁺ entry (SOCE) was present in cortisol-treated cells. For these experiments, we measured both the rate of Ca²⁺ entry (change in [Ca²⁺]_i during the first 10 s) and peak change in [Ca²⁺]_i following Ca²⁺ reintroduction to cells with depleted Ca²⁺ stores (10 min of TG treatment in Ca²⁺-free medium). Figure 3A shows an example of an experiment used to determine the rate and peak [Ca²⁺]_i after CaCl₂ was introduced into the medium (final extracellular [Ca²⁺]= 0.15 mM). Parameters of Ca²⁺ influx following Ca²⁺ readdition to TG-treated cells showed considerable variability between LCLs with regard to rate of Ca²⁺ entry (3.1-39.3 nM s⁻¹) and peak [Ca²⁺]_i levels (233.1-2233 nM). Within individual experiments, the correlation coefficient for the rate of Ca²⁺ entry and the peak [Ca²⁺]_i level was 0.7464 (P < 0.0001, n= 71), indicating a high degree of consistency between the two measurements. Figure 3 also shows that cortisol treatment significantly increased both the rate of Ca²⁺ entry (Fig. 3B) and peak [Ca²⁺]_i (Fig. 3C). These results suggest glucocorticoids enhance the Ca²⁺ entry pathway activated by intracellular Ca²⁺ store depletion.

Figure 3. Cortisol enhancement of Ca²⁺ entry and peak [Ca²⁺]_i following Ca²⁺ addition to TG-treated cells.

In representative experiments (A), cells challenged with 100 nM TG in Ca²⁺-free medium for 10 min (not shown) had 0.45 mM CaCl₂ added to the cuvette (final [Ca²⁺]_o= 0.15 mM). The first 10 s of the increase in [Ca²⁺]_i were used to estimate the rate of Ca²⁺ influx (nM s⁻¹, B), and peak changes in [Ca²⁺]_i were determined as the difference between [Ca²⁺]_i immediately prior to CaCl₂ addition and the maximal [Ca²⁺]_i (C). Control experiments with cells treated with DMSO (vehicle, 0.2 %) for 10 min had low (< 0.05 nM s⁻¹) rates of Ca²⁺ influx after CaCl₂ addition, hence no correction for non-TG-mediated increases in ${\text{Ca}}_{\text{i}}^{\text{2+}}$ was performed. ANOVA indicated overall P values < 0.0001 for both analyses, with significant differences between cortisol and all other treatment conditions.

Estimates of rates of Ca²⁺ influx, or peak changes in [Ca²⁺]_i following Ca²⁺ readdition to cells, may be affected by Ca²⁺-binding proteins and intracellular or plasma membrane Ca²⁺ transport mechanisms. We therefore assessed fura-2 quenching by Mn²⁺ as an indicator of store-operated divalent cation entry in TG-treated cells. Mn²⁺ has been used previously to characterize Ca²⁺ entry and SOCE in a variety of cells (Hallam & Rink, 1985; Missiaen et al. 1990). It has substantially greater affinity than Ca²⁺ to fura-2, and is a poor substrate for SERCA and other Ca²⁺ transport systems (Gomes da Costa & Madeira, 1986; Missiaen et al. 1990). In addition, quenching is monitored at the isosbestic point of fura-2, enabling changes in the rate of Mn²⁺ entry to be monitored independently of changes in [Ca²⁺]_i. Figure 4A illustrates a representative experiment showing fluorescence records of TG- and vehicle-elicited Mn²⁺ influx for the four treatment conditions. To quantify the rate of Mn²⁺ entry, data obtained with cells treated with TG, vehicle or the difference between these curves (i.e. TG-sensitive Mn²⁺ entry) were best-fitted to a monoexponential function (Fig. 4B, and see Methods and Table 1 for additional details). Rates of Mn²⁺ quenching of intracellular fura-2 showed considerable variability between LCLs; e.g. from 0.11 to 1.79 % s⁻¹ for vehicle-treated cells and 0.14 to 2.46 % s⁻¹ for cortisol-treated cells. Rate constants derived for each treatment indicated cortisol-treated cells had a significantly greater mean rate constant for TG-sensitive Mn²⁺ entry than control or RU 486-treated cells (Fig. 4C and Table 1). The parameters describing rates of TG-sensitive Ca²⁺ entry and Mn²⁺ entry (quenching at 30 s) were highly correlated with each other (r= 0.5935, P < 0.0001, n= 69). These experiments, coupled with the findings of Fig. 3, suggest that measurements of rates of Ca²⁺ entry and peak changes in [Ca²⁺]_i in TG-treated lymphoblasts are largely unaffected by Ca²⁺ transport or sequestration mechanisms. Importantly, the data indicate that LCLs treated with cortisol have enhanced SOCE.

Table 1.

Quenching of intracellular fura-2 by Mn²⁺: k₁ (s⁻¹)

Treatment	Vehicle	TG	TG sensitive
Control	0.446 ± 0.030	1.392 ± 0.077	2.169 ± 0.203
Cortisol	0.441 ± 0.047	1.731 ± 0.090*	3.450 ± 0.293**
RU 486	0.495 ± 0.044	1.483 ± 0.074	2.169 ± 0.178
Cortisol + RU486	0.489 ± 0.040	1.413 ± 0.079	2.005 ± 0.198

Cells were treated with vehicle, cortisol, RU 486 or cortisol and RU 486 (as described in text and legend to Fig. 4) and assessed for Mn²⁺ influx (cf. Fig. 4). Values are means ± asymptotic S.E.M. of k₁ (× 10⁻²), the rate constant. Percentage fluorescence quenching values at 0, 10, 30 and 60 s were compiled for vehicle- and thapsigargin (TG)-treated cells and were best fitted by the monoexponential function A=a₁e^−k₁t, described in Methods. TG-sensitive quenching of fura-2 describes the exponential fit of Mn²⁺ quenching of TG-treated cells minus the quenching in vehicle-treated cells. a₁ values ranged from 96.99 ± 1.44 to 99.18 ± 0.05 in vehicle- and TG-treated cells, and 44.37 ± 1.56 to 44.97 ± 4.53 for curve-fitted data describing TG-sensitive Mn²⁺ influx, and were not significantly different within treatments. n= 18 for each treatment. Significantly different at

^*P < 0.05

^**P < 0.01.

These results indicate that glucocorticoid treatment enhances Ca²⁺ homeostasis in LCLs. Correlation of the extent of cortisol stimulation of PAF-evoked [Ca²⁺]_i signals (cf. Fig. 1B and C, left panels) with intracellular Ca²⁺ store or SOCE parameters could suggest potential determinants effecting the cortisol-induced stimulation. However, these analyses resulted in weak correlations showing high dependence on one or two data points. We therefore performed a correlation analysis of the entire data set of Ca²⁺ and Mn²⁺ parameters to examine the relation between PAF-evoked increases in [Ca²⁺]_i and Ca²⁺ store size and parameters of SOCE. Although this analysis does not specifically examine the cortisol-induced enhancement of the PAF response, it examines parameters of Ca²⁺₁ signalling throughout the range of cortisol-modulated Ca²⁺₁ responses. The correlation coefficients for these comparisons are indicated in Table 2. In Ca²⁺-free medium, PAF-evoked [Ca²⁺]_i responses were positively correlated to Ca²⁺ stores and parameters describing SOCE. Similar correlations were obtained when the PAF response in Ca²⁺-containing HBS was used. The strongest correlation occurred with FEC, and is shown in Fig. 5A. As intracellular Ca²⁺ stores can regulate Ca²⁺ entry, we also compared the relation between store size and parameters of SOCE (Table 3). In general, correlation coefficients for these comparisons were higher than those for PAF-mediated [Ca²⁺]_i responses and intracellular Ca²⁺ store size or SOCE. Two examples of the strong association are given by the correlations between the peak [Ca²⁺]_i in TG-treated cells and the peak [Ca²⁺]_i achieved following TG addition (Fig. 5B), and TG-sensitive Mn²⁺ entry and FEC (Fig. 5C). These results suggest that the magnitude of the PAF-evoked Ca²⁺₁ signal is dependent on the size of the intracellular Ca²⁺ stores, which in turn are related to Ca²⁺ entry through the SOCE pathway(s).

Table 2.

Correlation analysis of PAF-evoked [Ca²⁺]_i response with intracellular Ca²⁺ store size and parameters of SOCE

	Parameter	r	n	P
PAF-evoked [Ca2+]i in Ca2+-free HBS	TG-evoked [Ca2+]i	0.2713	74	0.0194
	FEC	0.3544	72	0.0023
	Rate of Ca2+ entry	0.2701	74	0.0199
	Peak [Ca2+]i	0.2425	72	0.0401
	Mn2+ quenching	0.2703	70	0.0236

Analyses were performed for the PAF-evoked [Ca²⁺]_i response obtained in Ca²⁺-free HBS vs. parameters of intracellular Ca²⁺ store size (TG-evoked [Ca²⁺]_i and FEC), rate of Ca²⁺ entry and and peak [Ca²⁺]_i in TG-treated cells, and the degree of TG-sensitive Mn²⁺ quenching at 30 s (obtained as described in legends to Figs 2–4).

Figure 5. Correlations between PAF-evoked [Ca²⁺]_i response and intracellular Ca²⁺ store size (A) and intracellular Ca²⁺ store size and parameters of SOCE (B and C).

A, correlation of PAF-evoked [Ca²⁺]_i responses (in Ca²⁺-free HBS) and FEC; parameters were obtained as described in legends to Figs 1 and 2. B, correlation of TG-evoked [Ca²⁺]_i and peak [Ca²⁺]_i levels achieved after 10 min TG treatment and readdition of CaCl_2. TG-evoked [Ca²⁺]_i was obtained as described in legend to Fig. 2, and peak [Ca²⁺]_i values were obtained as in legend to Fig. 3. C, intracellular Ca²⁺ store size (FEC) and TG-sensitive Mn²⁺ entry. FEC was determined as in the legend to Fig. 2, and Mn²⁺ entry was obtained from TG-sensitive decreases in fluorescence at 30 s after addition of MnCl₂ (cf. Fig. 4). Correlations were obtained with data from the four treatment conditions for 17-19 LCLs (for A, n= 72; for B, n= 71; for C, n= 67).

Table 3.

Correlation analysis of measurements of intracellular Ca²⁺ store size and parameters of SOCE

	Parameter	r	n	P
TG-evoked [Ca2+]i	Rate of Ca2+ entry	0.4433	73	0.0001
	Peak [Ca2+]i	0.6905	71	0.0001
	Mn2+ quenching	0.2969	69	0.0132
FEC	Rate of Ca2+ entry	0.3551	71	0.0024
	Peak [Ca2+]i	0.4134	72	0.0003
	Mn2+ quenching	0.4804	67	0.0001

Analyses were performed for TG-evoked [Ca²⁺]_i and FEC and the rate of Ca²⁺ entry and peak [Ca²⁺]_i in TG-treated cells and the degree of TG-sensitive Mn²⁺ quenching at 30 s (obtained as described in legends to Figs 2–4).

DISCUSSION

This report characterizes several novel aspects of glucocorticoid action on cellular Ca²⁺ homeostasis in human B lymphoblasts. Forty-eight hour treatment of cells with cortisol decreased basal [Ca²⁺]_i and increased the PAF-evoked [Ca²⁺]_i transient. Cortisol increased the size of intracellular Ca²⁺ stores and increased the rate of TG-sensitive Ca²⁺ and Mn²⁺ entry, indicating enhanced SOCE. These results identify links among the size of the Ca²⁺ stores, Ca²⁺ mobilization from these stores, and the pathways that refill the Ca²⁺ stores. As the stimulatory effects of cortisol on Ca²⁺₁ signalling occur in conjunction with increases in intracellular Ca²⁺ stores and SOCE, the results indicate a major regulatory role of glucocorticoids on lymphoblast Ca²⁺ homeostasis.

Cortisol-induced changes in Ca²⁺ homeostasis could result via modulation of transcriptional events, or through a non-genomic mode of action initiated at the plasma membrane. In the former, more traditional, mechanism of steroid action, the triggering of genomic events is mediated by cortisol binding to a glucocorticoid receptor (GR) and subsequent modulation of nuclear transcription through the glucocorticoid-GR complex (Boumpas et al. 1993). Alternatively, a non-genomic mechanism of action is characterized as a rapid (< 2 min) onset, membrane-mediated response that occurs independently of DNA transcription or protein synthesis inhibition (for review, see Wehling, 1997). Non-genomic responses usually involve ion transport systems and are presumed to occur through a membrane-resident receptor. For example, cortisol both acutely and chronically affects cAMP concentrations and ⁴⁵Ca²⁺ accumulation in cichlid fish pituitary cells (Borski et al. 1991), and acutely hyperpolarizes guinea-pig neurons, an effect inhibited by GR antagonists (Hua & Chen, 1989). However, we failed to observe rapid effects of cortisol on Ca²⁺₁ signalling. Furthermore, the long (12 h) latency for enhancement of the Ca²⁺₁ signal is more consistent with a transcriptional/translational mechanism of action. As LCLs contain GRs (Tomita et al. 1985; Sinclair et al. 1994), we performed experiments with RU 486, a competitive antagonist of the cytosolic GR (Agarwal et al. 1987). The results of these experiments showed complete inhibition of cortisol-induced changes in all measured parameters of Ca²⁺ homeostasis and Mn²⁺ influx, suggesting activation of endogenous GR is critical for the modulatory effects of cortisol. Taken together, these results do not support a non-genomic action of cortisol and instead suggest the primary effect of cortisol occurs through the classically described, genomic mechanism of glucocorticoid action.

The present observations showing glucocorticoid-stimulated decreases in basal [Ca²⁺]_i and enhanced Ca²⁺ signalling in a model of activated lymphocytes are, with one exception, quite different from those reported in the literature for lymphocytes or lymphoid cells. That is, in response to increased cortisol, activated T cells and resting B cells exhibit reduced Ca²⁺₁ signalling (Dennis et al. 1987; Baus et al. 1996) whereas thymocytes demonstrate an increase in basal [Ca²⁺]_i (McConkey et al. 1989). S49 lymphoma cells exhibit increased basal [Ca²⁺]_i and decreased intracellular Ca²⁺ stores (TG-evoked [Ca²⁺]_i changes in Ca²⁺-free medium: Lam et al. 1993; Bian et al. 1997) in response to dexamethasone (18-24 h) treatment. The discrepancies between glucocorticoid effects in these cells and immortalized lymphoblasts suggest that Ca²⁺₁ signalling pathways downstream of glucocorticoid-GR activation differ between resting and activated lymphocytes, subsets of lymphocytes, and lymphoma cells. These differences are likely to be key factors responsible for some of the well-described effects of glucocorticoids on lymphocyte function, i.e. inhibition of cell proliferation (Bowen & Fauci, 1984) and induction of apoptosis (McConkey et al. 1989; Lam et al. 1993; Bian et al. 1997). Interestingly, one recent report showed dexamethasone enhanced receptor-mediated [Ca²⁺]_i transients and other parameters of Ca²⁺ signalling in U937 and erythroleukaemia cells (Willmott et al. 1997). Although these investigators were unable to associate alterations in Ca²⁺ homeostasis with an increase in a glucocorticoid-regulated protein (lipocortin 1), their work lends support to the finding that glucocorticoids enhance Ca²⁺-signalling pathways through modulation of intracellular Ca²⁺ stores. As this phenomenon is observed in monocytic, erythroid and lymphocytic cell lines, and occurs independently of viral transformation, glucocorticoid modulation of Ca²⁺ homeostasis may reflect differentiation-specific responses of haematopoietic cells to elevated levels of cortisol.

In cortisol-treated LCLs, the increase in the PAF-evoked [Ca²⁺]_i response was observed in Ca²⁺-free medium and was strongly correlated to the FEC. The TG-evoked [Ca²⁺]_i response and FEC, two markers of the size of intracellular Ca²⁺ stores, were similarly enhanced by cortisol treatment. As PAF-evoked responses in Ca²⁺-free HBS are approximately 70 % of the response in Ca²⁺-containing medium (Brzustowicz et al. 1997), it is apparent that not only the PAF-induced [Ca²⁺]_i signal is highly dependent on the size of the Ca²⁺ stores, but cortisol alters the size of these Ca²⁺ stores. Enhanced Ca²⁺ sequestration capabilities of these compartments could also explain the decrease in basal [Ca²⁺]_i. The simplest explanation of the cortisol-induced enhancement of Ca²⁺₁ signalling would postulate an upregulation of SERCA Ca²⁺ pumps, SER Ca²⁺ binding proteins (e.g. calreticulin), or modifiers of these systems, assuming glucocorticoid-GR signalling events exist for enhancing the expression or activity of these proteins. However, assigning the involvement of a particular Ca²⁺ system as a target of glucocorticoid action, in the absence of specific experiments, is highly speculative. For instance, overexpression of a plasma membrane Ca²⁺ pump in endothelial cells downregulated SERCA Ca²⁺ pump and IP₃-channel expression, and enhanced store-dependent Ca²⁺ entry (Liu et al. 1996). This example illustrates the highly interdependent nature of Ca²⁺ signalling on multiple Ca²⁺ transport and buffering systems. We have ruled out the potential involvement of the Na⁺-Ca²⁺ exchanger as a modifier of enhanced Ca²⁺ responses, as LCLs fail to exhibit measurable Na⁺-Ca²⁺ exchange activity under conditions optimized to detect its activity (Balasubramanyam et al. 1996; Gardner & Balasubramanyam, 1996).

PAF-evoked [Ca²⁺]_i transients demonstrated cell line-specific responses with regard to the extent of stimulation by cortisol (Fig. 1B and C). The negative relation describing the percentage increase in the PAF response in cortisol- and vehicle-treated cells could be the result of several factors, including (but not limited to) a decreased ability of cortisol to effect GR activation, variable upregulation of GR in transformed LCLs, and/or a finite capacity of receptor-mediated Ca²⁺ signalling to evoke Ca²⁺ mobilization. With regard to the first possibility, direct measurements of GR levels and their extent of activation in the presence of cortisol in individual LCLs may address this issue. In addition, several cultured cell types retain the metabolic pathways that convert active glucocorticoid to inactive metabolite (and vice versa: Nath et al. 1993; Reeves, 1995; Rajan et al. 1996; Brem et al. 1998). We have observed that LCLs treated with carbenoxolone (an inhibitor of 11β-hydroxysteroid dehydrogenase; Edwards et al. 1996) exhibit enhanced PAF responses, suggesting LCLs may express the cortisol ‘inactivating’ form of this isoenzyme (authors’ unpublished observations). The second possibility, i.e. that certain LCLs are more responsive to cortisol because of increased GR levels, is supported by reports that EBV-transformed LCLs exhibit increased levels of GR protein compared with EBV-positive cells (e.g. Burkitt's lymphoma cells; Sinclair et al. 1994) or human mononuclear leucocytes (Tomita et al. 1985). This is in keeping with idea that LCLs are representative of cytokine-activated B cells, which also express enhanced glucocorticoid receptors (Falus et al. 1995). However, it is difficult to reconcile glucocorticoid responsiveness with the ability of an LCL to respond to an agonist such as PAF. Instead, it is more likely that LCLs exhibit a maximal capacity for receptor-evoked Ca²⁺₁ transients (with or without cortisol stimulation). This conclusion is supported by our study with a larger panel of CEPH LCLs (Brzustowicz et al. 1997), in which the maximal PAF-evoked [Ca²⁺]_i response in HBS was ∼400 nM. The limits of receptor-mediated Ca²⁺ mobilization could be bound by the size of the intracellular Ca²⁺ stores, although the influence of IP₃ metabolic pathways would appear to be of major importance.

The PAF-evoked [Ca²⁺]_i transient and its dependence on the intracellular Ca²⁺ store-SOCE axis is supported by several converging lines of evidence: the significant correlation of the PAF response with intracellular Ca²⁺ stores, the correlation between intracellular Ca²⁺ store size and SOCE, and the positive, modulatory events of cortisol on upregulation of the PAF response and the SOCE pathway and capacity of intracellular Ca²⁺ stores. Recently, Ca²⁺ entry via a store-operated channel (I_CRAC) was reported to be tightly linked to the filling status of intracellular Ca²⁺ stores (Hofer et al. 1998). Although our data do not address the time course of the activation of Ca²⁺ entry with Ca²⁺ depletion from intracellular stores, the results obtained with this limited number of LCLs suggest that over a range of intracellular Ca²⁺ store sizes, the magnitude of the Ca²⁺ entry process is coupled to the size of the Ca²⁺ store.

We observed that cortisol enhances Ca²⁺₁ signalling in a subset of minimally responsive cells. Physiological consequences of increased circulating glucocorticoid may thus serve to augment certain functions attributed to activated B cells, e.g. immunoglobulin secretion. This idea is derived from observations in B lymphocytes showing glucocorticoid-induced potentiation of IL-1 and IL-6 secretion (Emilie et al. 1988) and of IL-4-induced IgE formation (Nusslein et al. 1992). In this regard, immunoglobulin formation is highly dependent on endoplasmic reticulum (ER) synthesis, sorting and trafficking; these functions in turn are partially dependent on ER Ca²⁺ levels (Shachar et al. 1994). Furthermore, low-secreting variants of rat basal leukaemia cells have defective antigen-mediated [Ca²⁺]_i signals (Bingham et al. 1994), and LCLs with enhanced IgM and IgG secretion in response to PAF also have increased PAF-evoked [Ca²⁺]_i responses (Rosskopf et al. 1995). Based on this information and the findings in this report, we suggest intracellular Ca²⁺ stores and Ca²⁺₁-signalling pathways could serve as a major physiological target for glucocorticoid action.

The major finding in this report is that glucocorticoids significantly affect parameters of Ca²⁺ homeostasis in B lymphoblasts. In addition to previously discussed Ca²⁺ transport or IP₃ metabolic pathways, several other candidate systems could be invoked to explain the mechanisms through which glucocorticoid produce its effects. Besides describing the modulatory effects of cortisol, the present study characterizes parameters of Ca²⁺ homeostasis in LCLs representative of the general population. The possibility exists that lymphoblasts may be used to examine the intrinsic regulation of numerous Ca²⁺ parameters, including G-protein-coupled receptors, SOCE and intracellular Ca²⁺ stores. These studies may be accomplished through the comparison of known or postulated modulators of Ca²⁺ metabolism in LCLs exhibiting markedly different responses. Alternatively, they may be performed through phenotypic expressions of Ca²⁺₁ signalling, which, when coupled with non-parametric methods of linkage analysis (Brzustowicz et al. 1997), can be used to identify genetic components that mediate or regulate a specific Ca²⁺ parameter.