Nitric oxide inhibits capacitative Ca²⁺ entry and enhances endoplasmic reticulum Ca²⁺ uptake in bovine vascular endothelial cells

Abstract

In vascular endothelial cells, elevation of cytosolic free calcium concentration ([Ca²⁺]_i) causes activation of nitric oxide synthase (NOS) and release of nitric oxide (NO). The goal of the study was to characterize the interplay between [Ca²⁺]_i and NO production in this cell type. Simultaneous measurements of [Ca²⁺]_i and intracellular NO concentration ([NO]_i) in cultured bovine vascular endothelial cells (CPAE cell line) with the fluorescent indicators fura-2 and DAF-2, respectively, revealed that Ca²⁺ influx following agonist-induced intracellular Ca²⁺ store depletion (capacitative Ca²⁺ entry, CCE) represents the preferential Ca²⁺ source for the activation of the Ca²⁺-calmodulin-dependent endothelial NOS (eNOS). Exposure to the NO donor sodium nitroprusside (SNP) showed that high NO levels suppressed CCE and had an inhibitory effect on Ca²⁺ extrusion by the plasmalemmal Ca²⁺-ATPase. This inhibitory effect on CCE was mimicked by the membrane-permeant cGMP analogue 8-bromo-cGMP, but was reversed by the NO scavenger haemoglobin and prevented by the inhibitor of the NO-sensitive guanylate cyclase ODQ. Brief exposure to SNP reduced the peak of ATP-induced Ca²⁺ release from the endoplasmic reticulum (ER) and accelerated Ca²⁺ reuptake into the ER. Prolonged incubation with SNP resulted in enhanced Ca²⁺ loading of the ER, as revealed by direct measurements of store content with the ER-entrapped low-affinity Ca²⁺ indicator mag-fura-2. The results suggest that in vascular endothelial cells, NO synthesis is under autoregulatory control that involves NO-dependent [Ca²⁺]_i regulation. Via cGMP-dependent inhibition of CCE and acceleration of Ca²⁺ sequestration into the ER, NO can lower [Ca²⁺]_i and therefore exert an autoregulatory negative feedback on its own Ca²⁺-dependent synthesis.

The vascular endothelium forms a selective permeability barrier between the bloodstream and the vessel wall, which plays a key role in blood pressure regulation and cardiovascular homeostasis (e.g. Moncada et al. 1991). The endothelium responds to neurohumoral and physical stimuli by producing nitric oxide (NO), which relaxes vascular smooth muscle cells (VSMCs) (Carrier et al. 1997), inhibits platelet aggregation (Radomski et al. 1987), modulates leukocyte adhesion (Kubes et al. 1991), and prevents VSMC proliferation (Garg & Hassid, 1989). NO is generated from the amino acid l-arginine by nitric oxide synthase (NOS). Three isoforms of NOS have been identified. The neuronal-type (nNOS or NOS I) and the endothelial-type (eNOS or NOS III) are constitutive and their activation is Ca²⁺-calmodulin-dependent. In contrast, the third isoform (the macrophage-type NOS II or iNOS) is inducible and Ca²⁺ independent (Moncada et al. 1991; Forstermann et al. 1994). Agonists that raise intra-cellular Ca²⁺ concentration ([Ca²⁺]_i) activate constitutively expressed NOS transiently, generating NO in submicromolar concentrations (Forstermann et al. 1994; Blatter et al. 1995; Nakatsubo et al. 1998). In contrast, inducible NOS requires immunological activation of cells by cytokines or endotoxins. Typically, the enzyme is active for prolonged periods of time, producing NO in micromolar concentrations (Forstermann et al. 1994; Brorson et al. 1999). Therefore, inducible NOS appears to be more important in pathophysiological conditions, such as hypertension, atherosclerosis, diabetes, cerebral ischaemia and septic shock (Hobbs et al. 1999; Li & Forstermann, 2000). In target cells, most of the actions of NO are mediated by the activation of the soluble guanylate cyclase and formation of cyclic GMP (cGMP) (Hobbs et al. 1999).

Vascular endothelial cells respond to stimulation by vasoactive agonists with a transient elevation of [Ca²⁺]_i. Agonist-induced elevations of [Ca²⁺]_i result from release of Ca²⁺ from intracellular stores (endoplasmic reticulum, ER) through inositol 1,4,5-trisphosphate (IP₃)-sensitive release channels, followed by activation of capacitative Ca²⁺ entry (CCE) through plasma membrane Ca²⁺ influx channels (Holda et al. 1998). Removal of Ca²⁺ from the cytosol occurs via extrusion by the plasma membrane Ca²⁺-ATPase pump (PMCA) (Klishin et al. 1998; Sedova & Blatter, 1999) and the plasmalemmal Na⁺–Ca²⁺ exchanger (Goto et al. 1996; Sedova & Blatter, 1999) as well as sequestration into intracellular stores by the sarcoplasmic- endoplasmic reticulum Ca²⁺-ATPase (SERCA) (Morgan & Jacob, 1998).

NO and cGMP-activated kinases have been shown to modulate virtually all Ca²⁺ transporting systems in a variety of non-endothelial target cells including VSMCs (Blatter & Wier, 1994; Clementi, 1998). In contrast, the paracrine and autocrine effects of NO on Ca²⁺ homeostasis in endothelial cells themselves have seldom been subject of investigations, leaving a largely incomplete picture. What is known is scant, and the few findings about [Ca²⁺]_i regulation in endothelial cells by NO are contradictory. To shed some light onto the confusing picture regarding autocrine and paracrine effects of NO on Ca²⁺ homeostasis in vascular endothelial cells, we set out to systematically investigate the effect of NO on release and uptake of Ca²⁺ from the ER as well as on Ca²⁺ entry and extrusion in vascular endothelial cells. Our results demonstrate that in pulmonary artery endothelial cells, NO inhibits capacitative Ca²⁺ entry, reduces the peak of ATP-induced ER Ca²⁺ release and PMCA activity, and enhances filling and accumulation of Ca²⁺ into intracellular stores. The results suggest that autocrine and paracrine modulation of [Ca²⁺]_i by NO exerts an important autoregulatory feedback on NO synthesis in vascular endothelial cells.

Part of this work has been presented in abstract form (Dedkova & Blatter, 2001).

METHODS

Cell culture and solutions

Experiments were performed on calf pulmonary artery endothelial (CPAE) cells in non-confluent cultures. The CPAE cell line was obtained as passage 15 from American Type Culture Collection (ATCC CCL−209; Manassas, VA, USA). The cells were cultured in Eagle's minimum essential medium, supplemented with 10% fetal bovine serum (GIBCO, Grand Island, NY, USA) and 2 mm l-glutamine, and kept at 37 °C in an atmosphere of 5% CO₂ and 95% air. Once per week, the cells were dispersed using a Ca²⁺-free (0.1% EDTA) 0.25% trypsin solution, and subcultured onto glass coverslips for later experimentation. Cells were passaged up to six times after they were obtained from ATCC. Cells from all these six passages were used for experimentation. For each series of experiments, cells from typically two to nine different passages were used.

All experiments were carried out at room temperature (20–22 °C) on single cells in non-confluent cultures within 1 week after plating. During the experiments, cells were superfused continuously with a physiological salt solution (standard Tyrode solution) containing (mm): 135 NaCl, 5 KCl, 1 MgCl₂, 2 CaCl₂, 10 glucose and 10 Hepes; titrated to pH 7.3 with NaOH. In nominally Ca²⁺-free solution, CaCl₂ was omitted. In the experiments measuring cellular [NO]_i, 100 μm l-arginine was added to the standard Tyrode solution. In experiments where NOS was inhibited with N^ω-nitro-l-arginine (l-NNA; Sigma Chemical Co., St Louis, MO, USA), l-arginine was omitted. 1−[1,2,4]Oxadiazolo[4,3-a]quinoxalin-1-one (ODQ) was obtained from Tocris Cookson Inc. (Ballwin, MO, USA).

Fluorescence measurements

[Ca²⁺]_i measurements

Spatially averaged [Ca²⁺]_i measurements from single endothelial cells were performed using the fluorescent ratiometric Ca²⁺ indicator fur−2. Cells were loaded with the dye by exposure to 1 ml standard Tyrode solution containing 5 μm fur−2 acetoxymethyl ester (fura-2 AM; Molecular Probes, Eugene, OR, USA) and 5 μl of a Pluronic F-127 stock solution (0.2 g ml⁻¹ Pluronic F-127 dissolved in DMSO) for 20 min at room temperature. Fura-2 was excited by alternately illuminating the cells at 360 nm (F₃₆₀) and 380 nm (F₃₈₀) through a rotating filter wheel. Emitted cellular fluorescence was recorded at 540 nm. Single-cell fluorescence signals were recorded with a photomultiplier tube (model R2693; Hamamatsu Corp.) by masking off individual cells with an iris positioned in the emission path. Changes in [Ca²⁺]_i were expressed as changes of the ratio R = F₃₆₀/F₃₈₀.

[NO]_i measurements

For direct intracellular NO ([NO]_i) measurements, cells were incubated with the fluorescent NO-sensitive dye 4,5-diaminofluorescein (DAF-2; Kojima et al. 1998; Nakatsubo et al. 1998). Cells were exposed to the membrane-permeant DAF-2 diacetate ([DAF-2 DA] = 5 μm; Calbiochem, San Diego, CA, USA) for 30 min at room temperature in 1 ml standard Tyrode solution containing 100 μm l-arginine. Cells were subsequently washed for 10 min. DAF-2 fluorescence was excited at 480 nm (F₄₈₀). Emitted cellular fluorescence was recorded at 540 nm. Changes in cellular DAF-2 fluorescence intensities (F) in each experiment were normalized to the level of fluorescence recorded prior to stimulation (F₀). Changes in [NO]_i are expressed as F/F₀, thus representing percentage increases above basal levels. Endogenous NO production or NO accumulation after exposure to NO donors was quantified by calculating the rate of increase from a linear fit of the DAF-2 F/F₀ signal (d(F/F₀)/dt) under various experimental conditions.

For simultaneous measurement of [Ca²⁺]_i and [NO]_i, CPAE cells were incubated simultaneously with 5 μm fura-2 AM and 5 μm DAF-2 DA for 30 min at room temperature and washed for 10 min to allow sufficient time for de-esterification of the indicators. Fluorescence was excited by alternately illuminating the cells at 360 nm (F₃₆₀) and 380 nm (F₃₈₀) for fura-2 measurements and 480 nm (F₄₈₀) for DAF-2 measurements.

Intra-ER [Ca²⁺] measurements

For measurements of the ER [Ca²⁺], we used an approach based on the ability of the low-affinity fluorescent Ca²⁺ dye mag-fura-2 AM to compartmentalize into cell organelles (Hofer & Machen, 1993). Cells were loaded with mag-fura-2 by exposure to 1 ml standard Tyrode solution containing 2 μm mag-fura-2 acetoxymethyl ester (mag-fura-2 AM; Molecular Probes) and 5 μl of Pluronic F-127 stock solution (see above) for 60 min at 37 °C. During the last 20 min of dye loading, 10 μm BAPTA AM was added to the loading solution. Control experiments (not shown) confirmed that the presence of the Ca²⁺ buffer BAPTA in the cytosol eliminated contamination of the mag-fura-2 signal by changes in cytosolic [Ca²⁺]. Emitted mag-fura-2 fluorescence from single cells was measured at 540 nm in response to alternating illumination with light of 340 nm (F₃₄₀) and 380 nm (F₃₈₀). Under the described experimental conditions, changes of the ratio R = F₃₄₀/F₃₈₀ reflect changes of [Ca²⁺] in the intracellular Ca²⁺ stores (Hofer & Machen, 1993).

Measurement of CCE activity

Mn²⁺ was used as a substitute for Ca²⁺ ions to characterize unidirectional ion flux through the CCE pathway. As established previously (for details see Sedova et al. 2000), the rate of Mn²⁺ entry ([Mn²⁺]_o = 100 μm), and therefore CCE activity, was inferred from the rate of quenching of fura-2 fluorescence excited at the Ca²⁺-insensitive (isosbestic) wavelength of 360 nm (F₃₆₀). Because small cell-to-cell differences of the isosbestic excitation wavelength were observed, we used a linear combination of F₃₆₀ and 5–10% of F₃₈₀ to generate a truly Ca²⁺-insensitive fura-2 signal. This was verified by applying a brief ATP pulse to elicit a [Ca²⁺]_i transient at the beginning of the experiment. After correction, F₃₆₀ remained unaffected by a rise of [Ca²⁺]_i. At the end of each Mn²⁺ quench experiment, cells were permeabilized with digitonin (10 μm). The remaining fluorescence signal was subtracted from the original signal and the F₃₆₀ traces were normalized to the fluorescence level encountered before the addition of Mn²⁺.

Statistical analysis

Statistical differences among the data were determined with the Student's t test for unpaired data and considered significant at P < 0.05. Results are reported as means ± s.e.m. for the indicated number (n) of cells. Each experiment was conducted on a separate cell culture.

RESULTS

Direct measurements of [NO] in CPAE cells

For the direct measurement of endogenous NO production by single endothelial cells, we used the recently developed fluorescent NO-sensitive indicator DAF-2, which has been employed successfully to estimate NO levels in endothelial cells produced by constitutive eNOS (Nakatsubo et al. 1998; Berkels et al. 2000a). Stimulation of CPAE cells with the purinergic agonist ATP (5 μm) caused a gradual increase in DAF-2 fluorescence, which plateaued at about 25% above baseline after about 10 min (Fig. 1A; n = 16 cells). The initial rate of increase of F/F₀ was (0.36 ± 0.05) × 10⁻³ s⁻¹. After 5–10 min, the rise of F/F₀ slowed down considerably (remaining rate of increase (0.09 ± 0.02) × 10⁻³ s⁻¹), reflecting a decrease in endogenous NO production. In the presence of the NOS inhibitor l-NNA, ATP stimulation failed to elevate [NO]_i, suggesting that the elevation of NO level was due to endogenous NO production through activation of eNOS (d(F/F₀)/dt in the presence of l-NNA was (0.02 ± 0.01) × 10⁻³ s⁻¹; n = 8; P < 0.001). The control trace indicates basal NO production in the absence of ATP (n = 5). To confirm that the intracellular DAF-2 indeed sensed NO, cells were treated with the exogenous NO donor sodium nitroprusside (SNP, 100 μm; n = 25), which led to a robust increase in DAF-2 fluorescence (Fig. 1B). The rate of increase of F/F₀ was (3.07 ± 0.44) × 10⁻³ s⁻¹, approximately ninefold faster than observed upon stimulation with ATP (Fig. 1A). The SNP-induced rise of DAF-2 fluorescence was not affected by l-NNA (rate of increase of F/F₀ = (2.58 ± 0.94) × 10⁻³ s⁻¹; not statistically significant), as would be expected from an exogenous source of NO (n = 6).

Figure 1. Effect of NOS inhibition on the ATP- and SNP-induced [NO]_i elevations.

A, ATP (5 μm) induced an increase of [NO]_i that was completely abolished when cells were preincubated with the NOS inhibitor l-NNA (100 μm). The control trace represents basal [NO]_i levels in the absence of ATP. B, effect of exposure to SNP on [NO]_i in the presence and absence of l-NNA. Cells were pretreated with l-NNA for > 40 min before exposure to SNP. DAF-2 fluorescence intensities in each experiment were normalized to the level of fluorescence recorded prior to stimulation.

Simultaneous measurements of [Ca²⁺]_i and [NO]_i: Effect of NO on ATP-induced increase of [Ca²⁺]_i

We simultaneously measured changes in [Ca²⁺]_i and NO production in single CPAE cells in response to ATP stimulation (5 μm) before and after cells were exposed to SNP (100 μm). Stimulation of CPAE cells with ATP induced a typical biphasic [Ca²⁺]_i response (Fig. 2A) consisting of a rapid transient increase due to Ca²⁺ release from IP₃-sensitive internal Ca²⁺ stores, followed by a maintained plateau resulting from Ca²⁺ influx from extracellular space via CCE (Holda et al. 1998). The rise of DAF-2 fluorescence lagged behind the peak of the [Ca²⁺]_i transient and was predominantly observed during the plateau phase of the [Ca²⁺]_i transient, suggesting that Ca²⁺ entering via CCE may be the preferential source for activation of eNOS (see also Discussion). The rate of rise of the DAF-2 F/F₀ signal was (0.38 ± 0.06) × 10⁻³ s⁻¹ (n = 13). In the same experiment, we also attempted to evaluate the effect of high levels of NO on [Ca²⁺]_i. When cells were exposed to SNP later during the plateau phase of the ATP-induced [Ca²⁺]_i transient, there was a robust increase of DAF-2 fluorescence as well as a decrease of [Ca²⁺]_i (Fig. 2A). After exposure to SNP, the rate of rise of DAF-2 F/F₀ increased significantly (approximately 10-fold) from, on the average, (0.38 ± 0.06) × 10⁻³ s⁻¹ to (3.74 ± 0.62) × 10⁻³ s⁻¹ (n = 13; P < 0.0001).

Figure 2. Simultaneous recordings of [Ca²⁺]_i and [NO]_i in CPAE cells.

A, ATP stimulation (5 μm) induced a biphasic [Ca²⁺]_i increase and NO production that became more prominent during the plateau phase of the [Ca²⁺]_i transient. Subsequent SNP (100 μm) exposure significantly increased [NO]_i level and suppressed the sustained [Ca²⁺]_i plateau phase. B, high NO levels, achieved by SNP application 120 s before ATP stimulation, significantly decreased the ATP-induced [Ca²⁺]_i transient.

When cells were exposed to SNP prior to stimulation with ATP (Fig. 2B), the DAF-2 F/F₀ signal rose at a typical rate (d(F/F₀)/dt = (3.31 ± 0.06) × 10⁻³ s⁻¹; n = 6). The ATP-mediated rise of [Ca²⁺]_i, however, was reduced significantly. On average (n = 6), SNP reduced the peak and the plateau phase by 59 ± 8% (P < 0.005) and 53 ± 6% (P < 0.001), respectively.

The initial results demonstrated directly that not only is the endogenous NO production Ca²⁺ dependent, but [Ca²⁺]_i itself is also affected by NO in endothelial cells. We therefore set out to identify the Ca²⁺ source for eNOS activation and the Ca²⁺ regulatory pathways and mechanisms that are affected by NO in this cell type.

Ca²⁺ source for activation of endogenous NO production

The results shown in Fig. 2A suggested that Ca²⁺ entering the cell might be more proficient in activating eNOS than Ca²⁺ released from intracellular stores. The protocol applied in this experiment, however, did not allow a conclusive answer regarding the preferential Ca²⁺ source, because ATP stimulation in the presence of extracellular Ca²⁺ results in Ca²⁺ release and entry that overlap in time. (In separate experiments (n = 7; data not shown), we determined that in the presence of extracellular Ca²⁺, about 61% of the amplitude of the ATP-induced [Ca²⁺]_i transient was due to store release, whereas the remainder resulted from Ca²⁺ entry.) Because of this overlap between release and entry, we applied an experimental protocol that separated Ca²⁺ release from entry in time, but also resulted in [Ca²⁺]_i elevations of similar amplitude (Fig. 3). For that purpose, we exposed cells to the SERCA blocker thapsigargin (Thastrup et al. 1990) in the absence of extracellular Ca²⁺. Exposure to thapsigargin led to a transient increase of [Ca²⁺]_i due to Ca²⁺ leakage from the ER and depletion of the stores due to inhibition of Ca²⁺ reuptake. Subsequent exposure to 2 mm extracellular Ca²⁺ resulted in an increase of [Ca²⁺]_i, which under these conditions reflects Ca²⁺ influx through the CCE pathway (e.g. Holda et al. 1998) and will be referred to here as CCE transient. Simultaneous measurement of [NO]_i showed that [NO]_i increased mainly during activation of CCE (d(F/F₀)/dt = (0.31 ± 0.04) × 10⁻³ s⁻¹) and not during the thapsigargin-induced elevation of [Ca²⁺]_i (d(F/F₀)/dt = (0.05 ± 0.02) × 10⁻³ s⁻¹; n = 6; P < 0.001), even though both [Ca²⁺]_i signals reached similar amplitudes. This result demonstrates that Ca²⁺ entering the cell seems to be more efficient at activating eNOS than Ca²⁺ liberated from intracellular stores (see also Fig. 4).

Figure 3. Ca²⁺ source for eNOS activation.

Simultaneous recordings of [Ca²⁺]_i (bottom trace) and [NO]_i (top trace) from CPAE cells loaded with fura-2 and DAF-2. Intracellular Ca²⁺ stores were depleted by exposure to thapsigargin (TG, 1 μm) in the absence of extracellular Ca²⁺. CCE was elicited by increasing [Ca²⁺]_o to 2 mm.

Figure 4. Inhibition of ATP-induced CCE by NO.

A, simultaneous recordings of [Ca²⁺]_i (bottom trace) and [NO]_i (top trace) from CPAE cells. Cells were exposed to ATP (5 μm) in the absence of extracellular Ca²⁺ to deplete intracellular Ca²⁺ stores. After recovery of the ATP-induced [Ca²⁺]_i transient, 2 mm Ca²⁺ was added to induce CCE. SNP, added 120 s before Ca²⁺ application, significantly inhibited ATP-induced CCE. B, the rate of Mn²⁺ entry ([Mn²⁺]_o= 100 μm; [Ca²⁺]_o= 0) was monitored as the rate of quenching of fura-2 fluorescence excited at the Ca²⁺-insensitive (isosbestic) wavelength of 360 nm (F₃₆₀; normalized). The fast Mn²⁺ influx triggered by ATP stimulation was significantly suppressed by SNP (100 μm). The Mn²⁺ quench F₃₆₀ trace labelled ‘-SNP’ was obtained from a cell stimulated with ATP but not subsequently exposed to SNP.

Effects of NO on capacitative Ca²⁺ entry

In a next series of experiments, we directly measured the effect of NO, through the use of the NO-donor SNP, on capacitative Ca²⁺ entry. To activate CCE, we depleted intracellular Ca²⁺ stores either by stimulation with ATP or exposure to the SERCA blocker thapsigargin.

CCE activation and NO production after ATP-induced store depletion

To deplete ER Ca²⁺ stores, single CPAE cells were stimulated with ATP (5 μm) in Ca²⁺-free medium. Changes in [Ca²⁺]_i and NO production were measured simultaneously in cells loaded with fura-2 and DAF-2 (Fig. 4A). CCE was characterized by analysis of the changes of [Ca²⁺]_i that occurred upon re-addition of 2 mm extracellular Ca²⁺ following store depletion. Single endothelial cells were stimulated twice with ATP in the absence of extracellular Ca²⁺ (Fig. 4A). Application of ATP caused a reproducible transient increase of [Ca²⁺]_i reaching a peak after 10–20 s and returning to the basal level within 1–2 min. No measurable change in [NO]_i was observed during the ATP-induced [Ca²⁺]_i transient. The subsequent addition of 2 mm extracellular Ca²⁺ led to an increase of [Ca²⁺]_i, which under these conditions resulted from Ca²⁺ influx through the CCE pathway. During the CCE transient, NO production was initiated, leading to a gradual increase of the cellular DAF-2 signal (the rate of rise of the DAF-2 F/F₀ signal was (0.31 ± 0.07) × 10⁻³ s⁻¹; n = 5). Consistent with the data shown in Fig. 2A and Fig. 3, these experiments also confirmed that Ca²⁺ entering the cell is more effective for the activation of NO synthesis than Ca²⁺ released from internal stores.

To determine the effect of NO on CCE itself, SNP (100 μm) was applied 2 min before Ca²⁺ re-addition after the second stimulation with ATP of the same cell. Besides the expected robust increase of the DAF-2 fluorescence (d(F/F₀)/dt = (2.48 ± 0.67) × 10⁻³ s⁻¹) signal, the application of SNP significantly decreased both the rate and the amplitude of the CCE transient. In the presence of SNP, on average (n = 5), the rate and the amplitude decreased to 22 ± 4% (P < 0.001) and 23 ± 4% (P < 0.001), respectively, compared with control (see also Fig. 6).

Figure 6. Effect of NO on CCE parameters after store depletion with thapsigargin and ATP.

Summary of the effects of SNP on CCE parameters after store depletion with ATP stimulation (left) and thapsigargin (TG) exposure (right). The columns represent the amplitude of the CCE transient (filled column; measured 1 min after addition of Ca²⁺), the rate of rise of [Ca²⁺]_i during the CCE transient (open column) and the rate Mn²⁺ entry measured as the rate of quenching of fura-2 fluorescence (hatched column). The data are presented as average percentages ±s.e.m. of control, i.e. the same parameters measured in the absence of SNP. Numbers in parentheses indicate the number of individual cells. The rate of rise of [Ca²⁺]_i and the rate of Mn²⁺ entry were measured as the slope of the linear portion of the signals presented in Figs 4 and 5.

We also used the Mn²⁺ quench technique to assess the effect of NO on CCE in the CPAE cells. As established previously (Sedova et al. 2000), Mn²⁺ can act as a surrogate for Ca²⁺ ions to study ion permeation through the CCE pathway. The use of Mn²⁺ has the following advantages: Mn²⁺ permeates the CCE channel similarly to Ca²⁺, is only a poor substrate for Ca²⁺-ATPases, and can be used to estimate CCE activity from the rate of quenching of the fluorescence from cytosolic Ca²⁺ indicators such as fura-2 when measured at a Ca²⁺-independent wavelength. Exposure of CPAE cells to 100 μm Mn²⁺ had little effect on the fura-2 signal measured at the 360 nm excitation wavelength (Fig. 4B). Application of ATP (5 μm) significantly increased the decline of F₃₆₀ as a result of Mn²⁺ entry and quenching of fura-2. Subsequent exposure to SNP in the continued presence of ATP, however, induced a marked decrease in quenching rate indicative of a NO-induced inhibition of CCE. On average (n = 13), in the presence of SNP, the rate of Mn²⁺ entry was only 29 ± 5% of that observed under control conditions (P < 0.001). The second trace in panel B shows that, at the point in time when SNP was applied, quenching of fura-2 continued at an unchanged rate in the absence of SNP. Thus, the change in quenching rate of the top F₃₆₀ trace reflects an effect caused by SNP, and was not the due to the fact that fura-2 quenching might be nearing completion. The data presented in Fig. 4 suggest that Ca²⁺/Mn²⁺ permeable plasma membrane channels and CCE are potential targets for an inhibitory effect of NO.

CCE activation by thapsigargin-induced store depletion

We applied a second protocol to evaluate the role of NO for CCE regulation. Intracellular Ca²⁺ stores were depleted of Ca²⁺ by preincubating CPAE cells for 15 min in Ca²⁺-free solution containing 1 μm thapsigargin. Re-application of 2 mm external Ca²⁺ evoked the typical biphasic CCE transient (Fig. 5A). The CCE transient was characterized by an initial rise of [Ca²⁺]_i that rapidly reached a peak, followed by a slower decline of [Ca²⁺]_i to a sustained plateau of elevated [Ca²⁺]_i. Upon removal of extracellular Ca²⁺, [Ca²⁺]_i decreased quickly to the resting level. As established previously, under these conditions this decay reflects solely Ca²⁺ extrusion that occurs predominantly by the PMCA (Klishin et al. 1998; Sedova & Blatter, 1999). As shown in Fig. 5A, exposure to extracellular Ca²⁺ in the presence of 100 μm SNP also caused an increase in [Ca²⁺]_i, but the rate of Ca²⁺ entry was slower (only 64 ± 2% of control; n = 14; P < 0.001), and the amplitude was lower than in the absence of the NO donor (75 ± 2% compared with control; n = 14; P < 0.001).

Figure 5. Effect of NO on CCE triggered by thapsigargin-induced store depletion.

A, CCE was triggered by exposure of CPAE cells to 2 mm Ca²⁺ (filled bar). Intracellular Ca²⁺ stores were depleted previously (not shown) by exposure to thapsigargin (TG, 1 μm) for 15 min in the absence of extracellular Ca²⁺. Activation of CCE in the presence of SNP inhibited CCE, resulting in a slower rate of rise of [Ca²⁺]_i, a lower amplitude and slower decline of the CCE transient. B, addition of Mn²⁺ (100 μm) to thapsigargin-pretreated cells stimulated a fast Mn²⁺ influx, which was partially suppressed by SNP ([Ca²⁺]_o= 0). The control Mn²⁺ quench F₃₆₀ trace labelled ‘-SNP’ was obtained from a thapsigargin-treated cell not subsequently exposed to SNP.

We also examined the effect of SNP on Mn²⁺ influx after previous store depletion with thapsigargin. The F₃₆₀ signal of a representative experiment is shown in Fig. 5B. The rate of quenching of the fura-2 signal increased after exposure to Mn²⁺ because of Mn²⁺ entry through store-operated cation influx channels. Application of 100 μm SNP resulted in a slowing of the fura-2 quenching rate, suggesting inhibition of CCE, while in the absence of SNP, the quenching rate remained constant. In the presence of SNP, the rate of Mn²⁺ entry after store depletion was only 68 ± 5% (n = 13; P < 0.001) of the quenching rate in the absence of SNP.

Figure 6 summarizes the results regarding inhibition of CCE by NO. When CCE was activated by thapsigargin depletion of internal stores, several parameters of CCE activity, i.e. the amplitude of the CCE transient (filled column), the rate of rise of the CCE transient (open column), as well as the rate of Mn²⁺ entry (hatched column), were reduced by approximately 25–35% in the presence of SNP. The effect of NO was much more pronounced when CCE was activated by ATP-dependent store depletion. Under these circumstances, all three parameters decreased by 75–80% in the presence of NO. Under the premise that both protocols led to store depletion and CCE activation, this discrepancy in the ability of NO to inhibit CCE suggested that other Ca²⁺ regulatory mechanisms are involved and affected by NO. This, as well as the mechanism through which NO acted on CCE, was further investigated in the following series of experiments.

Inhibition of CCE by SNP is due to NO

To confirm that the observed effects of SNP on CCE were indeed the result of the action of NO itself, we performed two types of control experiments. In a set of experiments (n = 4), we established that a different NO donor (S-nitroso-N-acetyl-penicillamine or SNAP, 100 μm) had the same inhibitory effect on CCE as SNP (data not shown), suggesting that the effects of SNP were probably due to its ability to produce NO and not due to some non-specific action of this substance.

To further establish this point, we treated cells with SNP in the presence of the NO scavenger haemoglobin (10 μm), which is known to effectively bind NO (Moncada et al. 1991; Blatter & Wier, 1994). As shown in Fig. 7A, the effects of SNP on amplitude (104 ± 2% of control) and kinetics (rate of entry 98 ± 2% and rate of decay 102 ± 3% of control) of the CCE transient were reversed in the presence of haemoglobin (n = 4; all parameters not significantly different from control). The effect of haemoglobin indicated that inhibition of CCE by SNP was mediated by NO.

Figure 7. Inhibition of CCE by SNP is mediated by NO and cGMP.

CCE transients elicited with 2 mm extracellular Ca²⁺ after store depletion with thapsigargin (1 μm). Changes of [Ca²⁺]_i evoked by thapsigargin itself are not shown. A, pretreament of CPAE cells with the NO scavenger haemoglobin (Hb) abolished the inhibitory effect of SNP on CCE. Hb (10 μm) was added 100 s before SNP application. B, preincubation with the membrane-permeant cGMP analogue 8-bromo-cGMP (300 μm) before addition of 2 mm Ca²⁺ inhibited CCE. C, CCE transient in the absence and presence of SNP (100 μm) and 8-bromo-cGMP (300 μm). D, effect of SNP on CCE in the presence of the NO-sensitive guanylate cyclase inhibitor ODQ (10 μm). Cells were preincubated for 40 min.

Inhibition of CCE by NO is mediated by cGMP

There is ample evidence that many effects of NO in various target cells are mediated by activation of soluble guanylate cyclase and elevation of cGMP levels (e.g. Xu et al. 1994; Kwan et al. 2000). To test whether the observed SNP effects on CCE in CPAE cells might also be mediated by increased cGMP levels, we used the exogenous cell-permeant cGMP analogue 8-bromo-cGMP (300 μm). For these experiments, we applied the thapsigargin protocol to deplete intracellular stores and to activate CCE. Figure 7B shows CCE transients recorded in the absence and presence of 8-bromo-cGMP treatment. Similar to SNP (Fig. 5A), 8-bromo-cGMP altered the shape of the CCE transient by decreasing the rate of Ca²⁺ entry (65 ± 3% of control; n = 6; P < 0.001), its amplitude (72 ± 3% of control; n = 6; P < 0.001) and decay kinetics after removal of extracellular Ca²⁺ (72 ± 4% of control; n = 6; P < 0.001). When 8-bromo-cGMP and SNP were applied simultaneously (Fig. 7C), the effect on amplitude (69 ± 4% of control; n = 3), rate of rise (57 ± 6%) and decay (76 ± 5%) of the CCE transient was very similar and not significantly different from the effect of cGMP alone. Thus, maximally active concentrations of cGMP and NO did not have additive effects on the kinetics and amplitude of the CCE transient.

Further evidence for the involvement of the NO-cGMP signalling pathway was established by the use of 1H-[1,2,4]oxadiazolo[4,3-a]quinoxalin-1-one (ODQ), a rather selective inhibitor of the NO-sensitive guanylate cyclase (Fig. 7D). In the presence of 10 μm ODQ (n = 5), the effect of SNP was completely abolished. With ODQ, the amplitude as well as the rise and decay kinetics of the CCE transient were not significantly different with and without SNP or compared with control.

The results presented in Fig. 7 strongly suggested that store-operated Ca²⁺ entry in endothelial cells is modulated via an NO-mediated increase in cGMP levels.

Inhibition of PMCA by NO

We have established previously a protocol that can be used to investigate the kinetics of PMCA activity and its regulation in CPAE cells (Klishin et al. 1998). To do so, we use the thapsigargin depletion protocol and subsequent re-exposure to 2 mm extracellular Ca²⁺ to elicit a CCE transient. We have demonstrated that under these circumstances, [Ca²⁺]_i in CPAE cells is determined solely by Ca²⁺ entry via CCE and extrusion across the cell membrane. As shown previously (Sedova & Blatter, 1999), Ca²⁺ extrusion is carried by the PMCA, and Ca²⁺ removal via Na⁺-Ca²⁺ exchange becomes relevant only when the PMCA is blocked. Consequently, under control conditions, changes of [Ca²⁺]_i during the CCE transient that are observed after removal of extracellular Ca²⁺ reflect Ca²⁺ removal by the PMCA.

Figure 8A shows two CCE transients that were elicited in thapsigargin-depleted cells. In contrast to Fig. 7, in this experiment, cells were exposed to 2 mm extracellular Ca²⁺ only for a brief period of time (< 1 min) and Ca²⁺ was removed again during the rising phase of the CCE transient. The experiment clearly demonstrates that SNP not only slowed CCE, resulting in a significantly lower amplitude of the CCE transient, but it also slowed the decay of the CCE transient after removal of Ca²⁺. The slow decline of the CCE transient resulted from NO-dependent inhibition of the PMCA. This result is similar to the CCE transients presented in Figs 5A, 7B and 7C, which also revealed a significant slowing of the decay kinetics upon removal of extracellular Ca²⁺ in the presence of SNP (Fig. 5A), 8-bromo-cGMP (Fig. 7B) or both compounds (Fig. 7C). Figure 8B shows the decay kinetics of normalized CCE transients after removal of extracellular Ca²⁺ in the presence and absence of SNP (left) and 8-bromo-cGMP (right). In both cases, the kinetics of the decay of the CCE transient, expressed as the time constant, τ, of a monoexponential fit to the [Ca²⁺]_i signal, were significantly slowed by SNP and 8-bromo-cGMP, respectively. In fact, on average (Fig. 8E), τ of the decline of the CCE transient was only 66 ± 4% (n = 7; P < 0.001) in the presence of SNP, and only 72 ± 4% (n = 6; P < 0.001) of control with 8-bromo-cGMP present.

Figure 8. Inhibition of the PMCA by NO.

A, CCE transients elicited by brief (< 50 s) exposure to 2 mm extracellular Ca²⁺ in the absence and presence of SNP (100 μm). Extracellular Ca²⁺ was removed again during the rising phase of the CCE transient. Intracellular Ca²⁺ stores were previously depleted by thapsigargin treatment. B, normalized decaying phases of the CCE transients (taken from Figs 5A and 7B) after removal of extracellular Ca²⁺ under control conditions and in the presence of SNP and 8-bromo-cGMP (cGMP), respectively. The [Ca²⁺]_i signals were normalized to the level of [Ca²⁺]_i encountered at the time of removal of extracellular Ca²⁺. C, inhibition of the PMCA by lanthanum ([La³⁺]= 1 mm). La³⁺ was applied when Ca²⁺ was removed from the extracellular solution. D, inhibition of the PMCA by La³⁺ in the presence of SNP. E, average normalized time constants (τ) of the decay phase of the CCE transients under control conditions (100 %) and in the presence of SNP, 8-bromo-cGMP (cGMP), La³⁺ and La³⁺+ SNP. τ was determined by a monoexponential fit to the declining phase of the CCE transient. The numbers in parentheses indicate the number of cells for each group.

When the PMCA was blocked with lanthanum ([La³⁺] = 1 mm), the decay of the CCE transient after removal of extracellular Ca²⁺ was slowed significantly (37 ± 1% compared with control; n = 4; P < 0.001) and remained incomplete (Fig. 8C and E); however, SNP had no additional inhibitory effect (37 ± 2% of control; n = 6; P < 0.001; Fig. 8D and E). These data suggested that the observed effects of SNP and cGMP on the decay phase of the CCE transient were indeed due to inhibition of the PMCA.

In summary, these results indicate that NO inhibits the PMCA in CPAE cells and that this inhibition is mediated by a signalling mechanism involving cGMP.

NO enhances Ca²⁺ uptake into intracellular Ca²⁺ stores

As demonstrated so far (summarized in Fig. 6), NO inhibited CCE, independently of whether CCE was activated by ATP- or thapsigargin-induced store depletion; however, the effect of SNP on the thapsigargin-induced cation influx was significantly smaller. Bearing in mind that the difference between the two experimental protocols was the fact that the SERCA pump was blocked (by thapsigargin) or not, we focused on the possibility that NO may influence refilling of intracellular Ca²⁺ stores.

In a first set of experiments, we explored to what extent NO affects the filling status of Ca²⁺ stores from which Ca²⁺ could be released by ATP stimulation. We used the Ca²⁺ ionophore ionomycin to estimate releasable Ca²⁺ store content. Exposure of unstimulated CPAE cells to ionomycin (1 μm) in the absence of extracellular Ca²⁺ led to a rapid and substantial increase of [Ca²⁺]_i (Fig. 9A), giving an estimate of releasable Ca²⁺ from intact filled Ca²⁺ stores. The transient nature of the ionomycin-induced rise of [Ca²⁺]_i was due to subsequent removal of Ca²⁺ from the cell. When cells were stimulated with ATP (5 μm) in Ca²⁺-free medium to release Ca²⁺, subsequent ionomycin exposure released approximately three times less Ca²⁺ from the internal stores, suggesting a substantial depletion by the preceding ATP stimulation (Fig. 9B). The amount of releasable Ca²⁺ by ionomycin under these experimental conditions could be due to resequestration of Ca²⁺ after ATP stimulation and/or liberation of Ca²⁺ from an ionomycin-sensitive store that is not depleted by ATP. To investigate whether NO affects the filling of Ca²⁺ stores in ATP-activated cells, SNP (100 μm) was applied before ATP (Fig. 9C). Pretreatment with SNP resulted in the following differences: SNP slightly attenuated the amplitude of the ATP-induced [Ca²⁺]_i response (compared with panel B); it accelerated the decay of the ATP-induced [Ca²⁺]_i transient; and most notably, it increased the remaining amount of Ca²⁺ in the stores that could be released by ionomycin (Fig. 9C). On average, the amplitude of the ionomycin-induced [Ca²⁺]_i signal in the presence of SNP was 54 ± 13% higher (P < 0.005; n = 8) than in the absence of SNP (Fig. 9B). Similar results were obtained when SNP was added not prior to but at the peak of the ATP-activated [Ca²⁺]_i response. In this case, the ionomycin-induced release was 44 ± 8 % (n = 3) higher than in the absence of SNP (not shown).

Figure 9. Effect of SNP on the amount of releasable Ca²⁺ from ionomycin-sensitive Ca²⁺ stores.

Ionomycin (Iono, 1 μm) was applied in the absence of extracellular Ca²⁺ at the end of each experiment, to estimate the amount of releasable Ca²⁺ from intracellular stores. A, control [Ca²⁺]_i transient due to ionomycin-induced Ca²⁺ release. B, stimulation with ATP (5 μm) prior to ionomycin exposure decreased the releasable amount of Ca²⁺. C, ionomycin treatment following ATP stimulation in the presence of SNP (100 μm, applied 200 s prior to ATP stimulation). The presence of SNP reduced the amplitude of the ATP-induced [Ca²⁺]_i transient, accelerated its recovery and augmented the amount of store Ca²⁺ releasable by ionomycin exposure. D, ionomycin-releasable [Ca²⁺]_i in a CPAE cell with disabled ER. ER Ca²⁺ release was abolished by preceding exposure to thapsigargin (TG, 1 μm for 15 min) and verified by the lack of a response to ATP stimulation. E, ionomycin-induced [Ca²⁺]_i signal in a thapsigargin-treated cell in the presence of SNP. Panels A to E are from different cells.

When the ER Ca²⁺ store was eliminated (Fig. 9D) by prolonged pretreatment with thapsigargin (substantiated by the lack of a response to ATP stimulation), exposure to ionomycin resulted in a small elevation of [Ca²⁺]_i (n = 4). This small (< 20 % of control) increase of [Ca²⁺]_i was consistent with the liberation of Ca²⁺ from intracellular organelles other than the ER that are capable of sequestering Ca²⁺ (e.g. mitochondria). As shown in Fig. 9E, however, preincubation with SNP had no effect on the ionomycin-induced response in thapsigargin-treated cells (n = 5), suggesting that NO produced by 100 μm SNP did not affect sequestration of Ca²⁺ into non-ER organelles (such as mitochondria) in a relevant fashion.

To further distinguish whether this effect of SNP was due to an activation of SERCA or an inhibition of Ca²⁺ release from the stores, the following experiments were performed. Cells were stimulated twice with 5 μm ATP in the absence of extracellular Ca²⁺. Between the two ATP stimulations, 2 mm extracellular Ca²⁺ was re-introduced to replenish the Ca²⁺ stores. As shown in Fig. 10A, two subsequent exposures to 5 μm ATP resulted in [Ca²⁺]_i transients with virtually identical amplitudes and time courses. Exposure to SNP 1 min prior to the second stimulation with ATP led to a small decrease in the amplitude of the ATP-induced [Ca²⁺]_i transient (Fig. 10B), possibly due to some inhibitory effect of NO on Ca²⁺ release. In contrast, when cells (n = 5) were incubated in SNP during the entire period of store reloading (> 20 min; Fig. 10C), the ATP-induced [Ca²⁺]_i transient was significantly larger in amplitude (190 ± 39 % of control; P < 0.05) and the rate of decay was accelerated significantly (227 ± 63 % of control; P < 0.05). These results suggested that prolonged exposure to SNP resulted in an NO-mediated enhanced filling of intracellular Ca²⁺ stores.

Figure 10. Effect of SNP on Ca²⁺ reuptake into the stores.

A, fura-2-loaded CPAE cells were stimulated twice with ATP (5 μm) in the absence of extracellular Ca²⁺. Between ATP stimulations, Ca²⁺ stores were allowed to refill in the presence of 2 mm extracellular Ca²⁺ for 20 min. B, ATP stimulation in the presence of SNP. SNP was introduced for 1 min prior to the second ATP stimulation; 20 min was allowed for refilling of the stores between ATP stimulations. C, ATP stimulation after store loading in the presence of SNP. SNP was present during the entire 20 min reloading period between the two ATP stimulations.

This hypothesis was confirmed by direct measurements of Ca²⁺ store content. For this purpose, we used the membrane-permeant form of the low-affinity Ca²⁺ indicator mag-fura-2, which was used to monitor the changes in free [Ca²⁺] in the internal Ca²⁺ stores without cell permeabilization (Hofer & Machen, 1993). Figure 11 shows typical recordings of free luminal [Ca²⁺] in control experiments (Fig. 11A) and in the presence of SNP (Fig. 11B). Single CPAE cells were stimulated with 5 μm ATP for 100 s in a Ca²⁺-containing medium. As shown in Fig. 11A, exposure to ATP caused a fast decrease of the mag-fura-2 signal as a result of Ca²⁺ release from the stores. After removal of ATP, the mag-fura-2 signal increased to previous levels due to reuptake of Ca²⁺. Twenty minutes were allowed for sufficient refilling of the stores prior to the second stimulation with ATP. In contrast, when Ca²⁺ stores refilled in the presence of SNP (Fig. 11B), the magnitude of Ca²⁺ mobilization from the stores by ATP stimulation was higher (147 ± 13 %; n = 5; P < 0.01), and the rate of Ca²⁺ reuptake into the stores was significantly faster (246 ± 27 %; n = 5; P < 0.001). Moreover, the free [Ca²⁺] in the ER was higher after prolonged incubation with SNP. Finally, even though the magnitude of Ca²⁺ release was larger in the presence of SNP, the absolute level of store Ca²⁺ was still higher during ATP stimulation compared with control conditions.

Figure 11. NO-dependent enhancement of Ca²⁺ loading of the ER revealed by direct measurements of store content with the low-affinity indicator mag-fura-2.

CPAE cells were loaded with mag-fura-2 AM and BAPTA AM. Changes in store [Ca²⁺] are expressed as changes of the fluorescence ratio F₃₄₀/F₃₈₀, reflecting Ca²⁺ movements in and out of intracellular Ca²⁺ stores. A, stimulation of cells with ATP (5 μm) in Ca²⁺-containing medium evoked release of Ca²⁺ from the stores, resulting in a rapid decline of the mag-fura-2 ratio. Removal of ATP initiated reuptake of Ca²⁺ into stores. After a 20 min refilling period, accumulated Ca²⁺ was released by a second stimulation with ATP. B, SNP (100 μm) application during the 20 min refilling period after the first stimulation with ATP resulted in enhanced store loading. A subsequent second challenge with ATP resulted in enhanced Ca²⁺ release and acceleration of Ca²⁺ reuptake.

We conclude from these experiments that increasing NO levels accelerates reuptake of Ca²⁺ into the intracellular stores. This allows more rapid refilling of the stores, and, as a consequence results in a de-activation of CCE into CPAE cells.

DISCUSSION

Ca²⁺ dependence of NO synthesis in vascular endothelial cells

Although it is well established that eNOS is regulated in a Ca²⁺-calmodulin-dependent manner, recent findings provided interesting evidence that the source of Ca²⁺ appears to be relevant for the efficient activation of NOS. Several studies have shown that Ca²⁺ entering via CCE is the preferential source for eNOS activation (Wang et al. 1996; Lin et al. 2000). Our simultaneous measurements of [Ca²⁺]_i and [NO]_i provided direct evidence for this idea (Figs 2, 3 and 4). The rise of [NO]_i clearly lagged behind the peak [Ca²⁺]_i increase when CPAE cells were stimulated with ATP. [NO]_i started to rise during the plateau phase of the ATP-induced [Ca²⁺]_i transient (Fig. 2A), during which cytosolic [Ca²⁺] remains elevated due to Ca²⁺ entry via CCE, and not because of release from internal stores. A similar temporal dissociation between peak [Ca²⁺]_i and maximal NO production was observed previously in bradykinin-stimulated vascular endothelial cells (Blatter et al. 1995). To clearly distinguish between the efficacy of released Ca²⁺versus Ca²⁺ entering the cell to activate NO production, and to further exclude the possibility that the ATP-induced [Ca²⁺]_i transient in the absence of extracellular Ca²⁺ (Fig. 4A) was just too short to activate eNOS, we also measured [NO]_i during exposure to thapsigargin (Fig. 3). The transient elevations of [Ca²⁺]_i during thapsigargin exposure are the result of liberation of Ca²⁺ from the stores, and resemble the CCE transients in amplitude and time course (Holda et al. 1998; Klishin et al. 1998). Under these conditions as well, NO production was observed predominantly during activation of CCE and not during release of Ca²⁺ from the stores. Thus, it appears that in endothelial cells, Ca²⁺ entering across the surface membrane is more efficient in activating NOS. This is consistent with previous studies showing that eNOS is mainly associated with the plasma membrane in endothelial cells from numerous vascular beds (Hecker et al. 1994; Garcia-Cardena et al. 1996), from where it can be translocated to various intracellular targets upon receptor stimulation (Prabhakar et al. 1998). In addition, recent evidence showed that eNOS is associated with caveolae, plasma membrane structures that have been implicated in important cellular signalling pathways (Garcia-Cardena et al. 1996; Igarashi et al. 1999).

Autoregulation of NO synthesis in endothelial cells

NO levels in stimulated endothelial cells typically do not exceed micromolar levels (Blatter et al. 1995; Nakatsubo et al. 1998), suggesting regulatory mechanisms that limit and terminate NO synthesis under physiological conditions. This is particularly important because excessive NO levels are detrimental to cells and have been implicated in various pathological conditions of the cardiovascular system. For example, high NO levels have been suggested to play a role in the hyporesponsiveness to vasoconstrictors observed in animals with septic shock (Julou-Schaeffer et al. 1990). An increased production of nitrogen oxides was also observed in vascular endothelium during early stages of hypercholesterolaemia concomitant with an upregulation of eNOS (Kanazawa et al. 1996).

Ca²⁺ plays a key role in the activation of eNOS and affects Ca²⁺ homeostasis in target cells of endothelial NO release (such as VSMCs). This raises the question of whether endothelial NO also influences [Ca²⁺]_i regulation in endothelial cells themselves and thereby exerts autocrine regulation of its own synthesis. Few data are available on the effect of NO on endothelial [Ca²⁺]_i regulation and the results are inconclusive. For example, it was shown that NO (applied in the form of SNP) could stimulate Ca²⁺ influx in porcine aortic endothelial cells (Berkels et al. 2000b). In contrast, NO dose-dependently decreased endothelin 1-stimulated mobilization of Ca²⁺ in microvascular endothelium from human brain (Chen et al. 1999), a response that was mimicked by cGMP and abrogated by inhibitors of guanylate cyclase or cGMP-dependent protein kinase. Furthermore, neither elevation of NO levels (by exposure to the NO donor SNP) and concomitant increase of cGMP nor a reduction of endogenous [NO] by treatment with the NOS inhibitor l-NNA had any effect on the histamine-induced [Ca²⁺]_i increases in human umbilical vein endothelial cells (Bolz & Pohl, 1997), whereas the NOS inhibitor N^ω-nitro-l-arginine methyl ester (l-NAME) enhanced the histamine-induced [Ca²⁺]_i elevations in primary cultures of human cerebral microvascular endothelial cells (Li et al. 1999). In addition, high doses of l-arginine, the substrate for NOS, slightly reduced [Ca²⁺]_i in endothelial cells and markedly decreased the [Ca²⁺]_i in VSMCs in co-culture, demonstrating both autocrine and paracrine effects of NO (Wang et al. 1996). There is also evidence that in vascular endothelial cells, endogenous NO is capable of enhancing CCE, but possibly also of activating the PMCA (Chen et al. 2000). Because NO is capable of influencing Ca²⁺ homeostasis in endothelial cells, it may also affect the Ca²⁺-dependent NO synthesis in NO-producing cells. There is indeed evidence to indicate that NO may have an autoregulatory role via a negative feedback inhibition of NOS activity in rat cerebellum (Rogers & Ignarro, 1992), murine macrophages (Assreuy et al. 1993) and vascular endothelial cells (Buga et al. 1993; Ma et al. 1996).

Our own results revealed that Ca²⁺ entering via CCE activates NOS, but higher levels of NO in turn inhibit CCE. Inhibition of CCE occurs either by direct blockade of the entry pathway or by enhancing and accelerating reuptake of Ca²⁺ into depleted intracellular stores by a cGMP-dependent mechanism. Our results lend support for both mechanisms acting in parallel. This conclusion is drawn from the following observations. (1) Exogenous NO from the NO donor SNP significantly suppressed the ATP-induced capacitative Ca²⁺ entry and Mn²⁺ influx (Fig. 4). (2) The inhibitory effect of SNP on CCE and Mn²⁺ entry was smaller when CCE was activated by inhibition of SERCA with thapsigargin rather than with ATP stimulation (Fig. 6). This observation is not unprecedented. Loss of the inhibitory effect of NO on capacitative Ca²⁺ entry upon SERCA blockade was shown previously in platelets (Okamoto et al. 1995; Trepakova et al. 1999), human embryonic kidney (HEK 293) cells (Bischof et al. 1997), rat megakaryocytes (Uneyama et al. 1998) and VSMCs (Cohen et al. 1999). This is consistent with the idea that with inhibition of SERCA, refilling of the stores is prevented completely, and therefore the signal to activate CCE is likely to remain strong and thereby counteracts the inhibitory effect of NO on CCE. Furthermore, acceleration of Ca²⁺ uptake and store refilling by NO inhibits CCE through inactivation of this pathway. (3) In thapsigargin-pretreated cells, SNP not only failed to activate, but slightly inhibited, the PMCA (Fig. 5A and Fig. 8). This argues against the possibility that the acceleration of the decline of the ATP-induced [Ca²⁺]_i transient in the presence of NO (Fig. 9 and Fig. 10) was the result of enhanced Ca²⁺ extrusion from cells, because the PMCA appears to be inhibited by NO (Fig. 8). Measurement of Mn²⁺ influx revealed directly that NO inhibited Ca²⁺ entry and might act on the Ca²⁺-permeable plasma membrane channels, an observation made previously for voltage-gated Ca²⁺ channels in VSMCs (Blatter & Wier, 1994). It is conceivable, however, that NO could affect other membrane ion channels, possibly leading to membrane depolarization. Depolarization would result in a diminished driving force for Ca²⁺ during CCE and therefore reduce the magnitude of the CCE transient. This alternative explanation is unlikely because SNP did not affect the membrane potential significantly in CPAE cells as measured with the voltage-sensitive dye Di-8-ANEPPS (data not shown). (4) We have clearly shown that NO affected Ca²⁺ store content. Estimating store content by exposure to ionomycin in the presence and absence of SNP revealed a higher degree of store filling in the presence of NO. Stimulation with ATP after brief preincubation with SNP caused [Ca²⁺]_i transients with slightly smaller amplitudes, suggesting that NO also affected release. This would be in line with data demonstrating inhibition of IP₃-induced Ca²⁺ release by NO in VSMCs (Ji et al. 1998) and platelets (Cavallini et al. 1996). Our data, however, provide strong evidence that the predominant action of NO on ER Ca²⁺ handling in endothelial cells is an enhancement of Ca²⁺ uptake. Our notion was based on the experimental data showing that SNP, when applied during the 20 min recovery period after the first addition of ATP, evoked a Ca²⁺ release with a significantly larger amplitude during the second application of ATP in the same cell (Fig. 10C). The significantly higher Ca²⁺ content of the ER and the accelerated rate of refilling were directly demonstrated in the experiments with mag-fura-2-loaded cells. As shown in Fig. 11, prolonged exposure to NO led to a larger degree of filling and an enhanced release upon stimulation, followed by acceleration of subsequent refilling of the stores.

cGMP is involved in the autocrine action of NO in endothelial cells

The effect of NO appears to be mediated via increased cGMP levels and the subsequent activation of cytosolic G kinase. In our study, the effect of NO was abolished in the presence of the NO-sensitive guanylate cyclase inhibitor ODQ (Fig. 7D), and the membrane-permeant analogue of cGMP, 8-bromo-cGMP, mimicked the effect of SNP on CCE. The involvement of the cGMP-G kinase signalling cascade in mediating NO effects has been shown in a variety of tissues including VSMCs (Cornwell & Lincoln, 1989), platelets (Geiger et al. 1992), PC12 cells and fibroblasts (Clementi et al. 1995). For example, it was shown that cGMP increases Ca²⁺ sequestration into VSMC Ca²⁺ stores (Twort & van Breemen, 1988) due to phosphorylation of phospholamban (Cornwell et al. 1991) and by isolated ER vesicles from the bovine main pulmonary artery (Raeymaekers et al. 1990).

Conclusion

The effect of NO on Ca²⁺ regulation in endothelial cells is controversial. For example, in contrast to our data, it has been shown that endogenous NO can lower basal [Ca²⁺]_i levels due to activation of Ca²⁺ extrusion by the PMCA in transfected endothelial cells with blocked IP₃-dependent Ca²⁺ release and potentiate CCE in intact cells (Chen et al. 2000). This discrepancy may be explained by the concentration dependence of the NO effect on the Ca²⁺ homeostasis in vascular endothelial cells, where low NO concentrations may have a stimulatory effect on CCE, whereas at higher concentrations, NO becomes inhibitory. Similar concentration-dependent effects have been demonstrated for pancreatic acinar cells (Gukovskaya & Pandol, 1994; Xu et al. 1994) and colonic epithelial cells (Bischof et al. 1995).

In conclusion, there is converging evidence that Ca²⁺ entering via the CCE pathway is the preferential source for activation of the endothelial Ca²⁺-calmodulin-dependent NOS to elevate NO levels. As [NO]_i levels increase, NO in turn is capable of lowering [Ca²⁺]_i by inhibition of CCE and acceleration of reuptake of Ca²⁺ into intracellular stores. Through this mechanism, cellular NO levels exert an autoregulatory negative feedback by regulating [Ca²⁺]_i itself and therefore the Ca²⁺-dependent eNOS. This inhibition occurs at different levels: (1) through direct inhibition of CCE; (2) through clearance of cytosolic Ca²⁺ via enhanced reuptake of Ca²⁺ into intracellular stores; and (3) through de-activation of CCE as a result of enhanced refilling of the stores. Thus, this mechanism has a dual function: it serves as a protection against excessive [Ca²⁺]_i levels and detrimentally high NO production in the vascular endothelium.