The mechanism of agonist induced Ca²⁺ signalling in intact endothelial cells studied confocally in in situ arteries

Abstract

In endothelial cells there remain uncertainties in the details of how Ca²⁺ signals are generated and maintained, especially in intact preparations. In particular the role of the sarco-endoplasmic reticulum Ca²⁺-ATPase (SERCA), in contributing to the components of agonist-induced signals is unclear.

The aim of this work was to increase understanding of the detailed mechanism of Ca²⁺ signalling in endothelial cells using real time confocal imaging of Fluo-4 loaded intact rat tail arteries in response to muscarinic stimulation. In particular we have focused on the role of SERCA, and its interplay with capacitative Ca²⁺ entry (CCE) and ER Ca²⁺ release and uptake. We have determined its contribution to the Ca²⁺ signal and how it varies with different physiological stimuli, including single and repeated carbachol applications and brief and prolonged exposures.

In agreement with previous work, carbachol stimulated a rise in intracellular Ca²⁺ in the endothelial cells, consisting of a rapid initial phase, then a plateau upon which oscillations of Ca²⁺ were superimposed, followed by a decline to basal Ca²⁺ levels upon carbachol removal. Our data support the following conclusions: (i) the size (amplitude and duration) of the Ca²⁺ spike and early oscillations are limited by SERCA activity, thus both are increased if SERCA is inhibited. (ii) SERCA activity is such that brief applications of carbachol do not trigger CCE, presumably because the fall in luminal Ca²⁺ is not sufficient to trigger it. However, longer applications sufficient to deplete the ER or even partial SERCA inhibition stimulate CCE. (iii) Ca²⁺ entry occurs via STIM-mediated CCE and SERCA contributes to the cessation of CCE. In conclusion our data show how SERCA function is crucial to shaping endothelial cell Ca signals and its dynamic interplay with both CCE and ER Ca releases.

1. Introduction

Endothelial cells play a variety of important physiological roles, including influencing vascular tone, permeability, inflammation and platelet aggregation [1–5]. There is structural and functional heterogeneity in the endothelium depending on the size and location of the blood vessel it lines. However elevation of intracellular Ca²⁺ in endothelial cells is key to their physiological functions, including regulation of vascular tone, inter-endothelial cell gap formation, angiogenesis, immune defense, gene expression and cell growth [6,7]. In endothelial cells Ca²⁺ can be released from the intracellular store in the ER. Depletion of the ER Ca²⁺ store induces Ca²⁺ entry across the plasma membrane, in a process known as capacitative or store operated Ca²⁺ entry (CCE or SOCE) [8–11]. Coupling of agonists to receptors can open receptor operated channels, through which Ca²⁺ can also enter the cell [12–15]. Details of the interplay between these elements and the characteristics of the resulting Ca²⁺ signals are still being elucidated. For example cytosolic calcium signalling plays an important role in the generation of endothelium derived relaxing factors (NO, EDHF), which contributes to vascular smooth muscle relaxation, and in turn, reduces blood pressure. However, the mechanisms controlling the spatial and temporal characteristics of the Ca²⁺ signalling involved in the activation of eNOS and other processes, are still poorly understood. Recently, the proteins STIM-1 and Orai 1 have emerged as candidate components mediating CCE [16–20]. Stim-1 responds to the depletion of Ca²⁺ stores activating CCE via interaction with ORAI. However, it has also been reported that unlike macrovessels, CCE is not important in microvascular endothelial cell [21,22]. We have recently shown that in ureteric precapillary arterioles (microvessels), Ca²⁺ signalling in endothelial cells does not require Ca²⁺ entry and consists mainly of asynchronous, localized, high frequency Ca²⁺ oscillations [23].

From the above it can be appreciated that the SR/ER Ca-ATPase (SERCA) will be intimately involved in governing the characteristics and kinetics of agonist-induced Ca signals, through binding Ca, influencing Ca releases through IP₃-dependent Ca²⁺ channels and instigating and terminating CCE via effects on luminal Ca levels [24]. However the information concerning SERCA's role in the endothelium is limited. Fierro and Parekh found that SERCA pumps were remarkably effective in RBL-1 cells and could prevent I_CRAC from activating, despite significant Ca²⁺ leakage from the stores [25]. Patch clamp studies in other types of cells [26,27] suggest however that SERCA should be able to terminate CCE as it refills the store.

Thus there is a need to systematically examine all these components in one cell type and endothelium from a single type of blood vessel. Furthermore there is also a lack of clarity around the roles of SERCA and CCE in the Ca²⁺ signals occurring in intact endothelium. While there may be advantages to studying cultured endothelial cells, recent research has highlighted the importance of studying the intact endothelium e.g. [28,29]. In the current study therefore we have determined the temporal and spatial organization of Ca²⁺ signals induced by muscarinic stimulation in situ in Fluo-4 loaded endothelial cells of rat tail artery using confocal microscopy, to better define the mechanisms controlling the characteristics of the Ca signals in response to muscarinic stimulation and the exact role of SERCA during different parameters of stimulations.

We demonstrate that the initial Ca²⁺ wave and the early Ca²⁺ oscillations originated from the ER by the interplay of repetitive releases and re-uptake of Ca²⁺ by IP₃-dependent Ca²⁺ channels and SERCA pump activity, respectively. In contrast, interplay between Ca²⁺ release, Ca²⁺ influx through CCE channels and SERCA was required to sustain the late Ca²⁺ oscillations and plateau.

2. Materials and methods

2.1. Ca²⁺ measurements

Tail was removed from rat humanely killed by cervical dislocation under CO₂ anesthesia in accordance with Home Office legislation Artery was dissected from the ventral grove, cleaned of fat and kept in physiological saline before use. They were loaded with Fluo-4 acetoxymethyl ester (Invitrogen, UK, 15 μmol/L; dissolved in DMSO with pluronic acid) for 2–3 h at 20 °C and then transferred to indicator-free solution for 30 min. Small segments of Fluo-4 loaded tail artery (3–4 mm in length) were cut open and fixed to the bottom of the chamber by aluminum foil clips, to minimize movement, with endothelium facing down. All experiments were performed at 30 °C. We used a Nipkow disc based, confocal microscope (Perkin Elmer), connected to a sensitive iXon cooled charge-coupled device camera (Andor). Images were collected at 30 frames per second using a 60× water objective (NA 1.20) for best spatial resolution or dry (20×, 0.70 NA) for a larger field of view.

2.2. Solutions

Physiological saline of the following composition was used (mmol/L): NaCl 120.4, KCl 5.9, MgSO₄ 1.2, CaCl₂ 2.0, glucose 8, and HEPES 11. In some experiments, Ca²⁺-free solution (2 mmol/L EGTA) was used. 2-Aminoethoxydiphenyl borate (2-APB), ryanodine, and U-73122 (to inhibit phospholipase C), were from Calbiochem (Nottingham, UK), all other chemicals were from Sigma (UK). Phenylephrine, endothelin-1 (ET-1), ryanodine, and carbachol were dissolved in water; 2-APB, cyclopiazonic acid (CPA, to inhibit SERCA), U-73122 in DMSO.

2.3. Immunohistochemistry

Specimens of dissected tail artery in the absence of stimulation (control) or following two consecutive applications of carbachol in the presence of CPA to deplete Ca²⁺ stores and maximally activate CCE (see Fig. 11 for details), were fixed with 4% paraformaldehyde in 0.1 M phosphate-buffered saline (PBS), pH 7.4 for 1.5 h at 23 °C, and then in 25% sucrose in 0.1 M PBS overnight at 4 °C, cut on a cryostat at 10 μm and mounted onto poly-lysine coated microscope slides. Cryostat sections were rinsed in 0.1 M PBS, permeabilized in alcohol and processed for single labelling immunofluorescence. In brief, tissue sections were pre-blocked in 10% normal donkey serum in 0.1 M PBS and incubated with rabbit polyclonal STIM1 (C-terminal) antibody (ProSci Incorporated, Nottingham, UK; 1:300) overnight at 4 °C. The secondary antibody was Texas red-conjugated (AffiniPure) donkey anti-rabbit IgG (Jackson ImmunoResearch Laboratories; West Grove, PA, USA) and sections of tail artery were mounted under Vectashield containing 4′6-diamidino-2-phenylindole (DAPI) (Vector Laboratories) to preserve fluorescence and stain cell nuclei. Specificity of immunostaining was determined by omitting the primary antibody. Sections were examined using an Axioplan Universal microscope, and images were processed using the Axio Vision 3.0 Imaging system with deconvolution options (Carl Zeiss Vision, Jena, Germany).

Fig. 11.

Effect of combined action of CPA and CCh on distribution of STIM 1 in ECs of intact rat tail artery. (A) Representative images of EC labelled by anti-STIM1 antibody (Texas red) and DAPI (blue) before and after stimulation with CCh in the presence of CPA (10 μM) to maximally deplete the ER. (B) Representative graph showing activation of Ca²⁺ entry induced by combined action of CPA and CCh in the presence of external Ca²⁺. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of the article.)

2.4. Statistics

All analysis was performed using Microcal Origin 8.0 (Massachusetts, USA). Quantification of calcium records from confocal imaging were made by measuring the peak amplitude, frequency and duration of calcium transients. The results are given as percentage of control value unless otherwise stated. Values are means ± SEM, and n is number of vessels or cells. Between 3 and 5 animals were used for each series of experiments. Differences were taken as significant at P < 0.05 in Student's t-test.

3. Results

3.1. Live morphology of the ECs in situ

By taking a series of optical sections in the Z direction it was possible to recreate a three-dimensional image of the endothelium of intact rat tail artery loaded with Fluo 4 (Fig. 1, Video 1). The endothelial cells (EC) appeared as monolayer of cells lining the luminal surface of the elastic lamina with variable length and width (Fig. 1, Video 1).

Fig. 1.

Live morphology of endothelial cells (ECs) of intact rat tail artery in situ. 3-Dimensional confocal image of intact ECs seen in x–y projection observed using 60× objective (Video 1).

3.2. Temporal and spatial characteristics of the Ca²⁺ transients induced by carbachol

Initially we used three concentrations of CCh (0.1, 1 and 10 μM) to assess the sensitivity of the intact endothelium. We observed variation between endothelial cells, ranging from transient localized Ca²⁺ events (puffs) seen mainly with low concentration of [CCh] to complex global Ca²⁺ transients, with an initial spike, then a plateau component, which usually had Ca²⁺ oscillations superimposed on it (Fig. 2, Video 2). The percentage of cells responding with global Ca²⁺ transient to CCh was concentration dependent. Thus about 40–60% of cells responded to 0.1 μM, 90–92% to 1 μM and 99–100% to 10 μM CCh. The initial spike component was a regenerative propagating (14–44 μm/s) Ca²⁺ wave, initiated at one or both ends of the cells (Fig. 2A and C, Video 2). As shown in Fig. 2C, where traces from the images in Fig. 2A have been superimposed, to show how the initial wave from either region 1 or 3, propagates to region 2 in the centre of the cell. The frequency of Ca²⁺ oscillations was also variable and concentration dependent (Fig. 2B and D). Ca²⁺ oscillations at 0.1 μM CCh ranged between 0.02 and 0.08 Hz (n = 51 cells, 5 vessels), at 1 μM they were 0.08–0.33 Hz (n = 63 cells, 7 vessels), and at10 μM CCh they were 0.18–0.36 Hz (n = 77 cells, 11 vessels). Since 1 μM CCh induced reproducible Ca²⁺ transients with distinct Ca²⁺ oscillations seen in the majority of cells, this concentration was used for the remainder of the study.

Fig. 2.

Temporal and spatial characteristics of the Ca²⁺ transient induced by CCh in EC of intact rat tail artery. (A) Stack of pseudo-colour images showing propagating Ca²⁺ wave initiated at two ends of the EC seen as an initial spike component. (B) Graph showing time dependent changes of Fluo-4 fluorescence induced by CCh in three parts of EC where region 1 and 3 are initiation sites. (C) Superimposed traces of initial Ca spike seen as propagating Ca wave initiated in region 1 (top panel) and region 3 (bottom panel) propagating to region 2 (centre of the cell). Arrows in A and C show direction of the wave propagation. (D) Patterns of the typical Ca²⁺ transients induced by 1 μM CCh in intact EC recorded by placing a region of interest in central part of the cell (Video 2). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of the article.)

3.3. Contribution of ER

To identify the ER's role in the complex Ca²⁺ responses induced by CCh, the effects of removal of extracellular Ca²⁺ for variable periods of times on the parameters of the Ca²⁺ transient induced by an application of CCh were studied. We found that the removal of external Ca²⁺ for 1–3 min had little or no effect on the amplitude of the initial spike and the onset of Ca²⁺ oscillations (Fig. 3Ai). However, when the endothelial cells were exposed to Ca²⁺-free solution for 30 min or more, this resulted in suppression of Ca²⁺ oscillations and reduction of all components of the Ca²⁺ transient (Fig. 3Aii). The amplitude of the initial spike component after 30 min exposure to Ca²⁺-free solution was reduced to 78.0 ± 2.5% (n = 27, P < 0.05) and after 60 min to 62 ± 4.2% (n = 35, P < 0.05) of control, respectively (Fig. 3B).

Fig. 3.

Effects of removal of extracellular Ca²⁺ for different periods of time on the Ca²⁺ transient induced by CCh in the ECs of intact rat tail artery. (A(i–ii)) Original records of CCh induced Ca²⁺ transients in control conditions and at different times of exposure (i) 1–3 min, (ii) 30 min to Ca²⁺-free solution containing 2 mM EGTA. B(a–d) Percentage of inhibition of the initial phasic component of the Ca²⁺ transient induced by CCh after exposure of the ECs to 30 s (b), 5 min (c) and 30 min (d). Peak Ca²⁺ transient in control condition was taken for 100% (a). (c and d) P < 0.05, significantly different from control.

We next investigated the ability of the ER to sustain Ca²⁺ signalling responses to CCh in the absence of extracellular Ca²⁺. To do this cells in Ca²⁺-free solution were stimulated by a series of applications of CCh of different durations; 10, 60 or 120 s. The endothelial cells were exposed to Ca²⁺-free solution for 3 min and then stimulated by a series of CCh applications (in 0-Ca²⁺ with 2 mM EGTA solution) and Ca²⁺ transients were recorded from a large number of cells using 20× objective lens.

When EC were stimulated by brief (10 s) applications of CCh, Ca²⁺ transients and oscillations could be observed for 15.0 ± 2.3 min (Fig. 4A). When the CCh application was increased to 60 s, EC quickly lost their ability to generate normal Ca²⁺ signalling after 3–4 CCh applications (Fig. 4B). It can also be seen that the first application of CCh induced a normal Ca²⁺ spike and initial Ca²⁺ oscillations but the sustained component gradually relaxed and late Ca²⁺ oscillations were abolished (Fig. 4B). Calcium transients induced by the 2nd and 3rd applications of CCh showed a marked reduction in the amplitude of the initial spike and plateau, and Ca²⁺ oscillations were not observed (Fig. 4B), In Ca²⁺-free solution and longer application of CCh (2 min), only one response was produced and the Ca²⁺ transients spontaneously relaxed to baseline (Fig. 4C).

Fig. 4.

Effects of removal of extracellular Ca²⁺ on Ca²⁺ transients induced by series of CCh applications for different lengths of time. (A, B and C) Representative traces of the Ca²⁺ transients induced by a series of 10, 60 and 120 s applications of CCh, respectively.

The ability of the EC to respond with regular Ca²⁺ transients for 15–20 min to short applications of CCh in Ca²⁺-free solution (Figs. 4A and 5A), suggests that under these conditions there is a minimal loss of ER Ca²⁺ and, that most of the Ca²⁺ released from the ER is quickly re-sequestered into the store by SERCA activity. To investigate this, the effects of the SERCA inhibitor CPA on the Ca²⁺ transients induced by brief (10 s) applications of CCh in Ca²⁺-free solution, were investigated. Exposure of the EC to CPA (20 μM to maximally inhibit SERCA), produced a large increased in Ca²⁺ which spontaneously relaxed to the baseline. Subsequent applications of CCh were ineffective, suggesting depletion of the Ca²⁺ store (Fig. 5B).

Fig. 5.

Effects of SERCA pump inhibition by cyclopiazonic acid (CPA) on the Ca²⁺ transients induced by brief (10 s) applications of CCh in Ca²⁺-free solution. (A) Ca²⁺ transients induced by a series of 10 s applications of CCh in Ca²⁺ free solution (control experiments); (B) inhibition of CCh-induced Ca²⁺ transients induced by CCh in Ca²⁺ free solution by CPA (20 μM). In these protocol EC were first pre-exposed to Ca²⁺-free solution with 2 mM EGTA for 3 min to fully remove external Ca²⁺.

3.4. Role of phospholipase C and ER release channels

The involvement of phospholipase C (PLC) in the transduction pathway leading to CCh evoked Ca²⁺ transients in EC in Ca²⁺-free solution was studied by using U-73122, an inhibitor of PLC activity. Pretreatment of the EC for 30 min with U-73122 (50 μM) abolished Ca²⁺ transients induced by CCh in Ca²⁺-free solution (n = 7 vessels, Fig. 6A). The possible roles of IP₃ receptor (IP₃R) and ryanodine receptor (RyR) channels in the ER CCh-induced Ca²⁺ release in Ca²⁺-free solution was studied by using the (non-selective) inhibitor of IP₃R, 2-APB and the (selective) inhibitor of RyR channels, ryanodine. In these experiments EC were pretreated with 2-APB (50 μM) or ryanodine (50 μM) for 30 min in Ca²⁺-free solution and then stimulated with CCh. Since 2-APB is known to also block Ca²⁺ influx from the extracellular space [30,31], its effect on Ca²⁺ release was studied in Ca²⁺-free solution i.e. a protocol similar to that used in Fig. 3. Pretreatment of the ECs in Ca²⁺-free solution with 2-APB for 30 min abolished CCh-induced Ca²⁺ transients (Fig. 6B). Application of CCh following 30 min ryanodine pretreatment had little or no effect on CCh-evoked Ca²⁺ transient (n = 11 cells, 3 vessels, Fig. 6C). This suggests that RyR channels are not involved in CCh-induced Ca²⁺ signalling in these EC (Fig. 6C). To confirm this the efficacy of inhibiting RyR channels was tested using caffeine (5 mM) and focusing confocally on the smooth muscle cell layer of the artery. Here caffeine induced a large Ca²⁺ transient associated with vasoconstriction (not shown), which was abolished by the 30 min pre-treatments with ryanodine, suggesting that RyR channels under these conditions are indeed blocked (Fig. 6D).

Fig. 6.

Effects of the PLC inhibitor U-73122, 2-APB and ryanodine on the Ca²⁺ transient induced by CCh in the EC. (A and B) Full inhibition of Ca²⁺ transients induced by CCh in Ca²⁺-free solution by U-73122 (50 μM) or 2-APB (50 μM), respectively; (C and D) effects of ryanodine (50 μM) on Ca²⁺ transients induced by CCh in EC and smooth muscle cells (SMC) of the same intact rat tail artery, respectively. In C and D, CCh and caffeine were applied in the presence of external Ca²⁺, respectively.

3.5. Contribution of extracellular calcium

To study the role of Ca²⁺ influx in the control of Ca²⁺ signalling in the intact EC, we used two (non-selective) blockers of CCE, La³⁺ and Gd³⁺. Lanthanum (10 μM) in the presence of external Ca²⁺ had little or no effect on Ca²⁺ transients induced by a series of short (10 s) applications of CCh, suggesting that Ca²⁺ influx was not required to maintain Ca²⁺ signals, within this period (Fig. 7A,i). In contrast, Ca²⁺ transients induced by 1 min applications of CCh showed a time dependent inhibition by La³⁺ (Fig. 7A,ii); late Ca²⁺ oscillations were inhibited and the plateau was markedly reduced, even in the first CCh application. Subsequent applications of CCh showed inhibition of all components of the Ca²⁺ transient (n = 49 cells, 7 vessels, Fig. 7A,ii). A 2 min application of CCh in the presence of La³⁺ evoked Ca²⁺ transients which spontaneously relaxed to the baseline (Fig. 7A,iii). Again, the initial Ca²⁺ spike and early Ca²⁺ oscillations were not affected but late Ca²⁺ oscillations were suppressed (Fig. 7A,iii) (n = 21 cell, 5 vessels). Similar effects were obtained with Gd³⁺ (data not shown).

Fig. 7.

Effects of La³⁺ on Ca²⁺ transients induced by CCh and readmission of external Ca²⁺ to Ca²⁺-free solution following ER Ca²⁺ depletion. (A) Effects of La³⁺ (10 μM) on Ca²⁺ transients induced by CCh applied for 10 (i), 60 (ii) or 120 (iii) s, respectively. (B) Effects of La³⁺ (10 μM) on Ca²⁺ transient induced by readmission of external Ca²⁺ to Ca²⁺-free solution following ER Ca²⁺ depletion by either 20 μM CPA (i) or 1 μM CCh (ii) applied in Ca²⁺ free solution.

Using either a CPA 20 μM (17 cells, 5 vessels) or long exposure to CCh in Ca²⁺-free solution to deplete the store, produced a substantial sustained rise of intracellular Ca²⁺ upon readmission of external Ca²⁺. This Ca²⁺ influx was inhibited by 10 μM La³⁺ (n = 16 cells, 5 vessels, Fig. 7Bi,ii) or 1 μM Gd³⁺ (data not shown).

3.6. Effects of low [CPA] on Ca²⁺ transients induced by brief applications of carbachol

As brief applications of CCh in Ca²⁺-free solution or in La³⁺ produced regular Ca²⁺ transients, it suggests that the endothelial cells under these conditions maintain normal Ca²⁺ signalling, as long as SERCA is functional (see Figs. 5A and 7A,i). Thus with brief applications of CCh in Ca²⁺-free solution, either the amount of Ca²⁺ released from the ER is small and/or SERCA activity is high enough so that most of the Ca²⁺ released is taken back into the ER, keeping luminal level of Ca²⁺ above the threshold of activation of CCE. To test this hypothesis in the next series of experiments the effects of short application of CCh on Ca²⁺ transients in the presence of low concentrations of CPA (5 μM) were studied, in the presence and the absence of external Ca²⁺ and in the presence of CCE blockers. Fig. 8 shows Ca²⁺ transients produced by these conditions, with (Fig. 8A) and without (Fig. 8B) external Ca²⁺. As was shown previously (e.g. Fig. 4) and in Fig. 8A, applications of CCh produced Ca²⁺ transients the duration of which correlated well with the duration of the CCh pulse. The EC were then exposed to 5 μM CPA for 30 s and then in the continuous presence of CPA, stimulated further by brief applications of CCh. We found that even after this brief exposure to CPA the first application of CCh evoked larger Ca²⁺ transients (Fig. 8A). With CPA the duration of the Ca²⁺ transient was no longer limited by the time of application of CCh. After removal of the agonist Ca²⁺ remained elevated for 30–40 s before relaxing to a new steady state baseline (Fig. 8A). In the presence of CPA the duration of the plateau, measured at 50% of the peak amplitude, was increased 2.1 ± 0.4 times and its amplitude to 2.2 ± 0.44 times (n = 31, 5 vessels, P < 0.05) relative to control (Fig. 8A). The amplitude of the spike component evoked by the first application of CCh in the presence of CPA was also increased 1.2 ± 0.1 times, P < 0.05. These data show that a large part of the sustained component of the CCh-induced Ca²⁺ transient was observed after removal of CCh from the media, which in the absence of CPA produced a quick relaxation of the Ca²⁺ transient (Fig. 8A). This suggests that this part of the Ca²⁺ transient could be due to activation of CCE, which remains open for longer period of times due to the slowed operation of the SERCA. To test this we performed the same experiments but in Ca²⁺-free solution or in the presence of CCE blockers. Fig. 8B shows typical traces of the Ca²⁺ transients induced by brief applications of CCh in Ca²⁺-free solution (n = 27 cells, 5 vessel) or in the presence of 10 μM La³⁺ (n = 18 cells, 4 vessels, Fig. 8C). It illustrates that in the absence of external Ca²⁺ or in the presence of La³⁺, CCh stimulation of the EC in the presence of CPA, produced brief spike-like Ca²⁺ transients with no plateau component (Fig. 8B, C). Also, EC now were unable to respond with Ca²⁺ transients when stimulated by a series of brief CCh applications. It was also noticeable that the amplitude of the spike component evoked by the first application of CCh in the presence of CPA, was still increased, 1.23 ± 0.12 times in Ca²⁺ free solution (n = 21 cells, 3 vessels, P < 0.05) and 1.37 ± 0.21 times in the presence of La³⁺ (n = 33 cells, 7 vessels, P < 0.05).

Fig. 8.

Effects of low concentration of CPA (5 μM) on temporal characteristics of Ca²⁺ transients induced by brief applications of CCh under different experimental conditions. (A and B) Effects of CPA (5 μM) on Ca²⁺ transients induced by applications of CCh in the presence and the absence of external Ca²⁺, respectively; (C) effects of La³⁺ (10 μM) on Ca²⁺ transients induced by CCh in the presence of CPA in normal physiological solution.

3.7. Role of SERCA in control of Ca²⁺ spike and Ca²⁺ oscillations

The data obtained in the presence of low concentration of [CPA] showed that the amplitude of the spike component induced by the first application of CCh, in the presence or the absence of external Ca²⁺, was always increased and Ca²⁺ oscillations abolished (Fig. 8). These data suggest that SERCA acts as a negative feedback mechanism controlling the characteristics of Ca²⁺ spikes and oscillations, by quickly buffering cytoplasmic Ca²⁺ released from the ER. To address this important point we studied the effects of SERCA inhibition on Ca²⁺ oscillations induced by longer (30 s) applications of CCh in the presence and the absence of external Ca²⁺. Data are presented in Fig. 9.

Fig. 9.

Effects of CPA on Ca²⁺ transients induced by 30 s applications of CCh in the presence and the absence of external Ca²⁺. (A) Typical records of Ca²⁺ transients induced by 30 s application of CCh in the presence of external Ca²⁺ in two EC with high (Cell 1) and low (Cell 2) frequency of Ca²⁺ oscillations recorded before (left panel) and 30 s after (right panel) addition of CPA (5 μM); (B) typical records of Ca²⁺ transients induced by 30 s application of CCh in EC exposed to Ca²⁺ free solution for 2 min in the absence (Cell 1) and the presence of 5 μM CPA (Cell 2) added 30 s before application of an agonist. Note that CPA fully blocked Ca²⁺ oscillations in the presence and the absence of external Ca²⁺.

As already described (Figs. 2 and 4) CCh produced Ca²⁺ transients with oscillations superimposing on the plateau component, in the presence and absence of external Ca²⁺. Fig. 9A shows that in the presence of external Ca²⁺ brief (20–30 s) pre-treatment of the EC with 5 μM CPA produced an increase in the amplitude and duration of the initial Ca²⁺ spike and plateau and inhibition of Ca²⁺ oscillations. The amplitude of the initial component was increased 1.2 ± 0.1times control and plateau 2.9 ± 0.4 times (n = 37 cells, 7 vessels, P < 0.05). Also, as was the case with brief applications of CCh (Fig. 8A), its removal did not terminate the Ca²⁺ transient; these lasted for an additional 30–50 s before gradually relaxing to new baseline (Fig. 9A). As already discussed, EC pre-exposed to Ca²⁺-free solution for 1–3 min retained their ability to generate early Ca²⁺ oscillations (Fig. 9B, top trace). However, if EC were pretreated for 30 s with CPA before CCh stimulation, Ca²⁺ oscillations in Ca²⁺-free solution were also abolished (n = 45 cells, 5 vessels, Fig. 9B, bottom trace). Again the amplitude of the spike component was increased 1.19 ± 0.12 times (n = 17 cells, 3 vessels, P < 0.05).

In order to study the effects of SERCA on the parameters of the Ca²⁺ spike, EC were stimulated in Ca²⁺-free solution with brief (3–5 s) applications of CCh. Although the number of cells responding to CCh was decreased, we could still identify cells producing Ca²⁺ transients consisting of initial Ca²⁺ spikes (Fig. 10A). The advantage of these experiments was that in the absence of external Ca²⁺, EC could respond with regular Ca²⁺ spikes for 20–25 min with minimal decrease in their amplitude (Fig. 10A). Addition of 5 μM CPA to Ca²⁺-free solution 30 s before the next brief application of CCh, increased the amplitude and duration of the Ca²⁺ spike (Fig. 10A and B). The duration of the Ca²⁺ spike measured at 50% of the peak amplitude was increased by CPA 3.4 ± 0.3 times (n = 9 cells, 2 vessels, P < 0.05). The duration of the Ca²⁺ spike evoked by the second application of CCh, in the presence of CPA, was increased further (Fig. 10A). Fig. 10B shows superimposed records of the last and the first Ca²⁺ spikes recorded before and after CPA, on an expanded time scale. It can be clearly seen that the amplitude and duration of the Ca²⁺ spike induced by the brief application of CCh in Ca²⁺-free solution, in the presence of CPA, were increased.

Fig. 10.

Effect of CPA on the temporal characteristics of the initial Ca²⁺ spike induced by brief application of CCh in Ca²⁺ free solution. (A) A series of Ca²⁺ spikes induced by a series of brief (3 s) application of CCh in Ca²⁺ free solution before and after addition of 5 μM CPA; (B) superimposed records of Ca²⁺ spike extracted from A recorded 30 s before (i) and 30 s after (ii) addition of CPA to Ca²⁺ free solution. Note that CPA increased the amplitude and duration of initial Ca²⁺ spike measured at 50% of peak amplitude to 3.4 ± 0.3 times (n = 9, 2 vessels, P < 0.05, significantly different from control).

3.8. Evidence for the presence of STIM 1

Our data suggest involvement of a Ca²⁺ release/entry coupling mechanism in the control of Ca²⁺ signalling in EC of tail artery. Although we did not undertake a systematic study on the physiological conditions involving STIM/Orai interaction, we used antibodies against STIM 1 for immunohistochemistry. We found (Fig. 11A) that STIM 1 is present in the EC of these vessels. It was homogenously distributed in the EC before stimulation (Fig. 11A, left hand side), but appears as puncta following depletion of the ER by a combined action of CPA and CCh (Fig. 11A, right hand side). The combined action of CCh and CPA (bottom panel) was used in these experiments to maximally deplete the store and activate Ca entry via SOC (Fig. 11B).

4. Discussion

The intact endothelium of tail artery was amenable to study with confocal microscopy and produced reproducible Ca²⁺ signal for 2–3 h. Clear Ca²⁺ signals were obtained without contamination from myocytes. Given the caveats that have arisen about using isolated or cultured endothelial cells for studying Ca²⁺ signalling [32–35], we suggest this is a good model for studying intact endothelium of macrovessels.

Our data reveal that, in the intact rat tail artery preparation, many aspects of the complex Ca²⁺ signal in EC produced in response to agonist stimulation are generated without the need for Ca²⁺ entry, as continuous cycles of IP₃-induced ER Ca²⁺ release and SERCA uptake occur. We discuss our findings in the context of previous data.

4.1. Importance of SERCA

Our major finding concerns the large role of SERCA in governing the responses of the endothelium to CCh and switching CCE on and off. SERCA binds and hence buffers cytoplasmic Ca²⁺ and acts as a powerful negative feedback mechanism controlling the amplitude and duration of the initial Ca²⁺ spike and early Ca²⁺ oscillations. Even brief exposure of EC to low concentration of the SERCA pump inhibitor CPA, markedly increases the amplitude of the Ca²⁺ spike and blocks Ca²⁺ oscillations. This buffering by SERCA prevents activation of CCE during brief applications of CCh, allowing EC to sustain Ca²⁺ oscillations in Ca²⁺-free solution for 15–20 min. A similar conclusion was reached by Kwan et al., 1990 in lacrimal acinar cells [49]. However, longer application of CCh (>20 s) causes partial depletion of the ER, resulting in activation of CCE. Inactivation of CCE is fast as SERCA takes up Ca²⁺, as also shown in patch clamp studies in other types of cells [26,27]. SERCA effectively buffers this Ca²⁺ entry by binding (and then transporting it into the ER lumen) [36] so that when the agonist is removed, CCE is terminated and there is rapid relaxation of the Ca²⁺ signalling. When SERCA activity is decreased by CPA, stimulation of EC, even with brief applications of CCh, now results in activation of CCE, and the channels remain open after removal of CCh.

4.2. SERCA and CCE

The finding that low [CPA] potentiated the initial spike and blocked Ca²⁺ oscillations suggests that Ca²⁺ is continuously cycling across the ER. The Ca²⁺ efflux through IP₃R is most likely countered by reuptake via SERCA and thus short applications of CCh are not enough to reach the threshold for CCE activation. Our results indicate that it is SERCA activity that prevents CCE activation. It might seem surprising that SERCA can maintain a sufficient Ca²⁺ store content to prevent CCE activation in the face of continuous Ca²⁺ release via open IP₃R. Our data are again in good agreement with those of Parekh and his group who found that SERCA pumps were remarkably effective in RBL-1 cells and could prevent I_CRAC from activating, despite significant Ca²⁺ leakage from the stores and a cytosolic Ca²⁺ binding ratio of between 2000 and 5000 (more than two orders of magnitude greater than within the stores) [25].

Small changes in ambient luminal and cytosolic Ca²⁺ could also have significant effects on SERCA pump activity as was shown in pancreatic acinar cells [37]. Small (two fold) rises of cytosolic Ca²⁺ stimulated SERCA activity almost 10 fold [38,39]. Thus, changes in SERCA-mediated uptake will have quite dramatic effects on the activation of CCE. Our data are in good agreement with finding that SERCA is the most important component in the control of the kinetics of Ca²⁺ wave propagation in Xenopus oocytes, where it was shown that the frequency of IP₃-induced Ca²⁺ waves is controlled by SERCA [40].

4.3. Calcium oscillations

The Ca²⁺ oscillations in the intact EC in tail artery seen in Ca²⁺-free solution are best explained as follows: the IP₃ receptor channels need to be sensitized to release Ca²⁺ periodically by cyclical fluctuations of Ca²⁺ within the lumen of the ER. Each pulse of released Ca²⁺ leads to a decrease in luminal Ca²⁺ which has to be replenished before the IP₃R are re-sensitized to deliver the next pulse of Ca²⁺. It is this loading of the ER that explains why the early and the late Ca²⁺ oscillations were quickly abolished even by low [CPA]. Entry of external Ca²⁺, which occurs mainly through CCE channels, combined with SERCA activity, determines the rate of store re-loading responsible for adjusting the sensitivity of the IP₃ receptors to maintain late Ca²⁺ oscillations. The expression and distribution of different isoforms of IP₃R and SERCA, as well as possible sub-divisions within the ER, are probably important for the fine tuning and kinetics of these events [41–44], but were not part of the present study.

Intracellular Ca²⁺ oscillations commonly results from lower (physiological) concentrations of agonists in many kinds of cells [45–48]. These oscillations depend upon complex mechanisms of regenerative intracellular signalling and generally do not persist in the absence of extracellular Ca²⁺, leading to the conclusion that entry of Ca²⁺ across the plasma membrane plays a significant role in either their initiation or maintenance. In this study, we have demonstrated that early Ca²⁺ oscillations in ECs of intact rat tail artery are not driven by Ca²⁺ entry, but rather depend upon Ca²⁺ release/Ca²⁺ uptake coupling while late Ca²⁺ oscillations require Ca²⁺ entry for their maintenance. The present study demonstrates that CCE is the underlying mechanism of calcium entry that is required to sustain late CCh-induced [Ca²⁺]_i oscillations in ECs of intact rat tail artery. The observation that CCh-induced Ca²⁺ spike and early Ca²⁺ oscillations can be observed in Ca²⁺-free solution suggests that the mechanisms triggering initial spike and early calcium oscillations are intrinsic to the intracellular milieu and that the role of calcium entry is primarily to replenish any ions lost from the cytoplasm. Late Ca²⁺ oscillations are more complex and involve interplay between Ca²⁺ release and Ca²⁺ entry, working in series with SERCA. However, with long applications of CCh EC lose their ability to sustain late Ca²⁺ oscillations which suggests that under these conditions Ca²⁺ influx is required to maintain normal Ca²⁺ signalling. In agreement with other studies we find CCE is the only physiological mechanism of calcium entry supporting late Ca²⁺ oscillations in intact ECs of large arteries [49,50].

4.4. Capacitative calcium entry

These data indicate that in the presence of low [CPA], a Ca²⁺ influx pathway is activated by brief applications of CCh and remained functional for a considerable time after agonist removal, and thus suggests it is not directly regulated by a receptor operated channel. Selective inhibition of the CCE pathway by La³⁺ or Gd³⁺ relative specific blockers of CCE in EC [51], suggest that the sustained component of the Ca²⁺ transient, seen after removal of CCh from the media was due to Ca²⁺ influx via CCE. We found that with low [CPA], EC in Ca²⁺-free solution could generate only 3–4 Ca²⁺ transients. This suggests that some SERCA activity remained, but was not high enough to counter Ca²⁺ release and Ca²⁺ leak from the ER, and thus eventually the ER Ca²⁺ store becomes depleted. The ability of EC to regularly respond to short applications of CCh in the presence of low [CPA] and external Ca²⁺, suggest that the ER or some of its portions are refilled and could be released by subsequent applications of an agonist. Because there is continued partial inhibition a new steady state base line is set which reflects the new equilibrium between Ca²⁺ leak and Ca²⁺ uptake. It is interesting to also note that TRCP4, one of the main candidates ROC channel in EC [52] is potentiated by La³⁺ [53]. This discrepancy with the inhibition of Ca signals which we saw, may be due to differences in preparations, as we used intact endothelium, but may also suggest that in our preparation, with muscarinic activation, most of the Ca entry is via CCE.

4.5. STIM

Recent data indicate that activation of CCE (CRAC) channels by store depletion involves the redistribution of the ER Ca²⁺ sensor, stromal interaction molecule 1 (STIM1), to peripheral sites where it co-clusters with the channel subunit, Orai1 [18–20]. After ER Ca²⁺ store depletion, STIM clusters are found in the ER, close enough to the cell surface to allow for direct interaction with plasma membrane proteins. The CRAC channel activation involves physical migration of ER-resident STIM to ER–plasma membrane junctions and subsequent aggregation of the Ca²⁺ influx channel. So STIM serves as a Ca²⁺ sensor and messenger at the same time as it also organizes Orai subunits into adjacent plasma membrane clusters [54]. The clusters of messenger STIM and corresponding Orai channel proteins have been termed the ‘elementary unit’ of CCE and CRAC channel activation, as Ca²⁺ influx through CRAC channels occurs precisely at STIM–Orai clusters [1]. In ECs CCE has been reported for several different vascular beds [51,55,56], although as with many other cell types, recording of their inward current (I_CRAC) are scarce, due presumably to its small size. Although there is still some debate over the molecular identity of CCE channels, we were able to demonstrate expression and redistribution of STIM1 in native EC. This is consistent with the recent findings of [57] in proliferating EC who established the requirement of STIM1/Orai1 for CCE. Levels of Stim1 were reported by this group to quite low, but they were able to also measure I_CRAC.

5. Summary

In conclusion we have been able to study Ca²⁺ signalling in response to agonist in intact endothelium from tail artery. We demonstrate alterations in the components of the Ca²⁺ signals depending upon concentration and duration of agonists and describe the mechanisms involved in generating the different components, spike, oscillations and plateau, of the Ca²⁺ signal. Importantly we have found that SERCA activity plays an important role in shaping the Ca²⁺ signals but also in turning CCE on and off. Thus agonist modulation of SERCA activity may be a mechanism of affecting endothelium function in vivo.

Conflict of interest

The authors do not have any conflict of interest.

Appendix A. Supplementary data

Video 1

Live 3-dimensional image showing EC of intact rat tail artery loaded with Fluo-4 in situ. EC are seen as monolayer of cells lining up elastic lamina.

Video 2

Real time confocal imaging of Ca²⁺ signalling in Fluo-4 loaded EC of intact artery stimulated with 1 μM CCh. Images are displayed at a speed of 30 fps.