Feedback via Ca²⁺-Activated Ion Channels Modulates Endothelin 1 Signaling in Retinal Arteriolar Smooth Muscle

Abstract

Purpose.: To investigate the role of feedback by Ca²⁺-sensitive plasma-membrane ion channels in endothelin 1 (Et1) signaling in vitro and in vivo.

Methods.: Et1 responses were imaged from Fluo-4–loaded smooth muscle in isolated segments of rat retinal arteriole using two-dimensional (2-D) confocal laser microscopy. Vasoconstrictor responses to intravitreal injections of Et1 were recorded in the absence and presence of appropriate ion channel blockers using fluorescein angiograms imaged using a confocal scanning laser ophthalmoscope.

Results.: Et1 (10 nM) increased both basal [Ca²⁺]_i and the amplitude and frequency of Ca²⁺-waves in retinal arterioles. The Ca²⁺-activated Cl⁻-channel blockers DIDS and 9-anthracene carboxylic acid (9AC) blocked Et1-induced increases in wave frequency, and 9AC also inhibited the increase in amplitude. Iberiotoxin, an inhibitor of large conductance (BK) Ca²⁺-activated K⁺-channels, increased wave amplitude in the presence of Et1 but had no effect on frequency. None of these drugs affected basal [Ca²⁺]_i. The voltage-operated Ca²⁺-channel inhibitor nimodipine inhibited wave frequency and amplitude and also lowered basal [Ca²⁺]_i in the presence of Et1. Intravitreal injection of Et1 caused retinal arteriolar vasoconstriction. This was inhibited by DIDS but not by iberiotoxin or penitrem A, another BK-channel inhibitor.

Conclusions.: Et1 evokes increases in the frequency of arteriolar Ca²⁺-waves in vitro, resulting in vasoconstriction in vivo. These responses, initiated by release of stored Ca²⁺, also require positive feedback via Ca²⁺-activated Cl⁻-channels and L-type Ca²⁺-channels.

Introduction

Signaling by the potent constrictor peptide endothelin 1 (Et1) is believed to be important in the regulation of retinal blood flow and to play a potentially pathogenic role in the development of several important ophthalmic diseases, including glaucoma and diabetic retinopathy. ^1–4 Despite this physiological and pathophysiological importance, however, the cell signaling pathways through which Et1 regulates retinal blood flow have received relatively little attention. We have recently shown that Et1 increases both basal [Ca²⁺]_i and the frequency of phasic Ca²⁺-signaling within arteriolar myocytes in isolated retinal arterioles. ⁵ The phasic element of the response was particularly pronounced, with both short duration Ca²⁺-sparks and more prolonged events, which we termed Ca²⁺-oscillations. Previous studies have shown that these “oscillations” are often associated with cell contraction, suggesting they may underpin the constrictor effects of Et1. ⁶ Immunostaining of porcine retinal arterioles reveals both EtA and EtB receptors in the smooth muscle layer, with EtA expression predominating at the protein level, while EtA receptor blockade inhibits most of the Et1-induced constriction in pressurized vessels. ⁷ Consistent with this, we found that Et1-induced increases in Ca²⁺-oscillations within retinal arteriolar myocytes were also mediated via EtA receptors. Et1-induced Ca²⁺-oscillations were also dependent on phospholipase-C activation, with downstream Ca²⁺-release from the sarcoplasmic reticulum involving both inositol 1,4,5-trisphosphate receptors (IP₃Rs) and ryanodine receptors (RyRs) . ⁵  

We have now extended this research to look at the possible contribution of Ca²⁺-activated channels in the plasma membrane to Et1 signaling. Both Ca²⁺-activated Cl⁻-currents (I_ClCa) and large conductance (BK) Ca²⁺-activated K⁺-currents (I_KCa) can be recorded from retinal arteriolar smooth muscle. ^8,9 These represent potential feedback mechanisms, whereby Ca²⁺-release may modulate Ca²⁺-influx via voltage-operated Ca²⁺-channels (VOCCs). ¹⁰ Activation of I_ClCa would result in membrane depolarization, leading to opening of VOCCs and increased Ca²⁺-influx, and so act as a positive feedback mechanism. This has been described in a variety of agonist responses in smooth muscle. ¹¹ Ca²⁺-dependent activation of BK channels, on the other hand, would favor membrane hyperpolarization and closure of VOCCs, acting as a brake or negative feedback mechanism limiting Ca²⁺-influx. ¹²  

In the current study we used two-dimensional (2-D) Ca²⁺-imaging in vitro to test for possible feedback during Et1-signaling using appropriate ion channel inhibitors of I_ClCa and BK channels. The physiological relevance of our findings was further investigated by recording constrictor responses to intravitreal Et1 in vivo in the presence and absence of appropriate channel antagonists. Our findings indicate that positive feedback via I_ClCa plays a crucial role both in Et1 Ca²⁺-signaling and the resulting mechanical responses in retinal arterioles. 

Materials and Methods

All animal use was performed under license from the United Kingdom Home Office under the Animals (Scientific Procedures) Act 1986 and complied with the ARVO Statement for the Use of Animals in Ophthalmic and Visual Research. In vivo experiments led to terminal anesthesia without recovery. 

Ca²⁺-Imaging in Isolated Rat Retinal Arterioles

Techniques used for arteriole isolation have previously been described in detail. ⁶ Retinas dissected from male Sprague-Dawley rats (200–300 g) were mechanically triturated using low-Ca²⁺ (100 μM) Hanks' solution. Following centrifugation, wash, and repeated centrifugation, vascular fragments were incubated for 2 hours with Fluo-4AM (10 μM) in low-Ca²⁺ solution and pipetted into a glass-bottomed organ bath (0.17-mm thick base) on the stage of an inverted microscope (Nikon Eclipse TE2000U; Nikon Instruments, Kingston, UK) . Arteriole segments (outer diameters of 25–50 μm) were visualized using a PlanApo, ×60, 1.4 Numerical Aperture (NA), oil-immersion objective, anchored using tungsten wire slips and superfused with normal Hanks' solution at 37°C . Two-dimensional Ca²⁺-images were obtained using a rotating-disc confocal system (QLC100; Visitech International, Sunderland, UK ). Fluo-4 fluorescence was excited using a 488-nm Light Emitting Diode (LED) laser (Sapphire 488-20; Coherent, Reading, United Kingdom), and emitted light was passed through a 530- to 560-nm band-pass filter . The image was captured at 20 frames per second (fps) using an electron multiplying Charge Coupled Device (CCD) camera (iXon; Andor Technology, Belfast, UK) . Imaging at each site was limited to 100 seconds to minimize photodamage. Fluorescence was background corrected using the mean counts from a tissue-free image region. Background-corrected fluorescence (F) was normalized to the resting fluorescence (F₀) in the same cell, and increases in F/F₀ were interpreted as being due to changes in intracellular [Ca²⁺] ([Ca²⁺]_i).  

In Vivo Imaging of Arteriole Diameter

Changes in retinal arteriolar diameter following intravitreal injections of Et1 with or without various ion channel blockers were recorded using fluorescein angiography with a Heidelberg Retinal Angiograph 2 (HRA2) confocal scanning laser ophthalmoscope (CSLO; Heidelberg Engineering, Heidelberg, Germany) . This was modified by removing both chin and forehead rests and adding an adjustable animal stage. Pigmented male Hooded Lister rats (320–450 g) were used to maximize image contrast. Rats were anesthetized using 0.35 mL per 100 grams of a solution containing ketamine (37.5 mg/mL; Fort Dodge Animal Health, Southampton, UK) and xylazine (5 mg/mL; KVP Pharma und Veterinar-Produkte GmbH, Kiel, Germany) . Depth of anesthesia, assessed using corneal and foot pinch reflexes, was maximal within 10 minutes and lasted up to 1.25 hours, after which top-up doses (0.1 mL/100 g) were introduced if needed. Rats were placed on a heat pad at 25°C (SnuggleSafe; Lenric International, Littlehampton, UK) and the rectal temperature monitored (digital thermometer model 77,020; Medisana, Dusseldorf, Germany).  

Intravitreal injections (10 μL) of relevant solutions (see Results section) were administered using a 29-gauge insulin syringe (Kendall monoject; Tyco Healthcare group, Basingstoke, UK) while viewing the eye through a surgical microscope (Zeiss OPH1 1-H; Carl Zeiss, City, UK) . The intraocular drug concentration was estimated assuming equilibration into a vitreal volume of 15 μL (total volume = 25 μL). ¹³ High intravitreal concentrations (approximately 10 times those typical in vitro) were used, since rapid clearance via the anterior segment may greatly reduce the effective concentration at the retina. ¹⁴ Viscotears (Novartis, Camberley, UK) was applied to prevent corneal drying, and 15 minutes were then allowed to pass before angiography) . Ten percent sodium fluorescein (0.025 mL/100 g body weight; Martindale Pharmaceuticals, Essex, UK) dissolved in sterile water (Braun, Melsungen, Germany) was injected into the tail vein of the rat. Fluorescence was excited at 488 nm, and emitted light was filtered using a 500-nm long-pass filter before passing through the confocal aperture and being recorded by the scanhead photodiode.  

Solutions and Drugs

The Hanks' solution used to superfuse isolated arterioles and as vehicle for intravitreal injections had the following composition (mM): NaCl, 140; KCl, 6; D-glucose, 5; CaCl₂, 2; MgCl₂, 1.3; HEPES, 10; and pH set to 7.4 with NaOH. Low-Ca²⁺ solution used for vessel isolation differed only in that it contained 0.1 mM CaCl₂. Drugs were sourced as follows: endothelin 1 and iberiotoxin (Tocris, Bristol, UK), nimodipine (Alexis, Nottingham, UK), and DIDS (Sigma, Gillingham, UK) . Stock solutions of nimodipine were prepared in dimethyl sulfoxide (DMSO) and then diluted to the final concentration using Hanks' solution, with a final DMSO concentration never more than 0.1% (vol/vol) . All other drugs were dissolved directly in Hanks' solution. Drugs were applied in prewarmed bath solution to isolated arterioles, or injected intravitreally as described above.  

Data Analysis

Phasic Ca²⁺-signals in the smooth muscle layer of isolated arterioles were analyzed in terms of amplitude and frequency. The relatively slow 2-D imaging rate (20 fps) made it impossible to resolve the brief, localized Ca²⁺-sparks previously characterized by us in this tissue using confocal linescanning. ^5,6,15 Prolonged (>500 ms duration) global rises in [Ca²⁺]_i were observed and are the focus of this article. These events are arbitrarily referred to as “oscillations” in our previous reports, since signal propagation along the length of the cell could not be detected in transverse linescans. ^5,6,15,16 However, in the current study, 2-D imaging revealed both propagating Ca²⁺-waves and synchronized Ca²⁺-oscillations (Fig. 1). The time-course of changes in F/F₀, spatially averaged across the area of each cell, was plotted using ImageJ (Shareware; NIH, Bethesda, MD) and the frequency and amplitude of phasic Ca²⁺-signals determined manually . Changes in the baseline [Ca²⁺]_i were assessed from the values of F/F₀ observed between phasic Ca²⁺-signals.  

Figure 1.

 

Et1-stimulates Ca²⁺-waves in retinal arteriolar smooth muscle. (A) Two-dimensional confocal image of Fluo-4–loaded myocytes in the wall of a retinal arteriole. These were imaged at 20 fps before and after the addition of endothelin 1 (Et1; 10 nM). (B) Time series plots of the normalized fluorescence for each of the three cells outlined in A (dashed line). (C) The resulting stack was resampled along the long axis of the same three myocytes, (solid lines in A). The images shown were recorded in the presence of Et1 and demonstrate propagating Ca²⁺-waves. In these cells, increases in [Ca²⁺]_i were initiated at the upper and lower ends of the cells and propagated centrally. The gradient of the Ca²⁺-wavefront is a measure of the propagation velocity. (D) Rescaled plot of the image region outlined on cell 3 in C. Image contrast was manipulated to allow the cell edge to be plotted (dashed line), illustrating that each wave was followed by cell shortening. (E) Summary data (mean ± SEM) for changes in wave frequency and amplitude following addition of Et1 (284 cells in 15 arterioles). ***P < 0.0001 versus control .

Confocal retinal angiograms used for vessel diameter measurements were captured in high-resolution mode using narrow (15°) fields of view within the superior retina and centered 0.5 to 1.5 disc diameters from the edge of the optic disc. Images were analyzed using custom software written in MATLAB (Mathworks, Natick, MA), using its Image Processing Toolbox, as described in a recent article from our group. ¹⁷ Images were automatically segmented to identify vessels and arteriole diameters determined for each pixel along the centerline of the vessel. These values were combined to give an overall mean diameter along the length of all arterioles within the image. Diameter measurements at branch points or at the edge of the image were excluded. Vessel diameters were estimated using a conversion factor of 1.15 μm/pixel, based on previously published data for the use of the HRA2 CSLO with rats. ¹⁸  

Summary data have been expressed as the mean ± SEM. The distributions of frequency and amplitude for phasic Ca²⁺-signals in retinal arteriolar smooth muscle are skewed, ¹⁵ so the statistical significance of apparent changes was tested using nonparametric tests (Wilcoxon signed-rank test for paired frequency and baseline [Ca²⁺]_i data, and Mann-Whitney U test for unpaired amplitude data). In vivo arteriole diameters following intravitreal injections of specific agents in separate groups of animals were compared using nonparametric one-way ANOVA (Kruskal-Wallis test) with Dunn's post-hoc test to correct for multiple comparisons. In all cases, the acceptable significance level was set at 0.05. 

Results

Et1 Stimulates Ca²⁺-Waves in Retinal Arteriolar Smooth Muscle

Et1 (10 nM) elevated baseline [Ca²⁺]_i and stimulated phasic Ca²⁺-signals in individual smooth muscle cells within isolated retinal arterioles (Figs. 1A, 1B), consistent with our previous findings using one-dimensional (1-D) confocal linescanning. ⁵ In the current study, 2-D confocal Ca²⁺-imaging provided information about the spatial patterning of signals at the cost of reduced temporal resolution. Sampling rates (20 fps) were too slow to allow the brief localized subcellular Ca²⁺-sparks seen in rapid line scans to be recorded. ¹⁶ Et1 typically stimulated Ca²⁺-waves, although the initiation site often varied from wave to wave in the same cell (Fig. 1C). These waves propagated along the long axis of individual myocytes with a mean velocity of 41 ± 4 μm/s (n = 77 events in 11 cells in 3 vessels; Fig. 1C) and were followed by myocyte contraction (Fig. 1D). This finding suggests that Ca²⁺-waves represent an important cellular mechanism underlying arteriolar constriction. 

For summary purposes, signals were averaged over the whole area of each cell and the phasic changes in mean [Ca²⁺]_i quantified. The term “Ca²⁺-waves” is used to refer to these events throughout the rest of this article, but true Ca²⁺-oscillations were occasionally seen, with relatively synchronous rises in [Ca²⁺]_i at different sites in the same cell. Et1 increased wave frequency from 1.26 ± 0.10/20 s under control conditions to 4.29 ± 0.16/20 s in the presence of Et1 (P < 0.0001; n = 284 cells in 15 vessels from 15 animals). Wave amplitude (ΔF/F₀) was also increased, from 0.25 ± 0.01 to 0.53 ± 0.02 (P < 0.0001; Fig. 1E). Superfusion with Et1 raised mean baseline [Ca²⁺]_i from 1.27 ± 0.03 to 1.70 ± 0.08 (P < 0.0001). 

Role of Ca²⁺-Activated Cl⁻- and K⁺-Currents in Cellular Ca²⁺-Signaling Responses to Et1 in Vitro

The contribution of I_ClCa to the Et1 response was tested using DIDS (1 mM; Fig. 2A). DIDS inhibits over 80% of I_ClCa in retinal arteriolar smooth muscle at negative membrane potentials close to the resting potential. ⁹ Ca²⁺-wave frequency was reduced from 3.40 ± 0.21/20 s in the presence of Et1 alone, to 0.62 ± 0.10/20 s when DIDS was also present (P < 0.0001; n = 89 cells in four arterioles; Fig. 2C). Wave amplitude was not significantly altered, with mean values of 0.57 ± 0.03 in the presence of Et1 alone and 0.40 ± 0.04 when both Et1 and DIDS were present (NS). Addition of 9-anthracene-carboxylic acid (9AC, 1 mM), which also blocks Cl⁻-currents in retinal arteriolar smooth muscle, ⁹ inhibited the Et1-induced Ca²⁺-waves (Fig. 2D). Frequency was reduced from 3.01 ± 0.19/20 s in the presence of Et1, to 1.10 ± 0.23 when 9AC was also added (P < 0.0001; n = 80 cells in four arterioles). Wave amplitude also fell from 0.79 ± 0.05 with Et1 alone to 0.38 ± 0.03 following addition of 9AC (P < 0.0001). Neither DIDS nor 9AC reversed the Et1-induced elevation of baseline [Ca²⁺]_i. Mean baseline F/F₀ was 2.05 ± 0.14 with Et1 alone and 2.01 ± 0.14 with both Et1 and DIDS present (NS). In similar experiments using 9AC, mean baseline values were 1.51 ± 0.14 when only Et1 was present and 1.44 ± 0.17 when 9AC was also added (NS). 

Figure 2.

 

Effects of ion channel blockers on Et1-stimulated phasic Ca²⁺-signals. (A) (i) Two-dimensional confocal image of Fluo-4–loaded myocytes in the wall of a retinal arteriole. These were imaged at 20 fps, and the change in fluorescence for each cell was used as a measure of changes in [Ca²⁺]_i. (ii) Time series showing the normalized fluorescence (F/F₀) for the cell outlined in white in i. Addition of Et1 (10 nM) stimulated Ca²⁺-waves/oscillations, but these were rapidly inhibited by addition of DIDS (1 mM). (B) Confocal image of arteriolar myocytes (i) and fluorescence time series (ii) for the outlined myocyte. Iberiotoxin (IbTx; 100 nM) increased the amplitude of phasic Ca²⁺-signals stimulated by Et1. (C) Summary data showing the effects of DIDS (1 mM) on signal frequency and amplitude in presence of Et1 (89 cells in four arterioles). (D) Summary data showing the effects of 9-AC (1 mM) on signal frequency and amplitude (80 cells in four arterioles). (E) Summary data showing the effects of iberiotoxin (100 nM) on signal frequency and amplitude (51 cells in four arterioles). (F) Summary data showing the effects of nimodipine (10 μM) on signal frequency and amplitude (64 cells in three arterioles). **P < 0.0005; ***P < 0.0001 versus Et1 alone .

These results suggest that positive feedback via I_ClCa plays an important role in the stimulation of phasic Ca²⁺-waves by Et1 at the cellular level. Indeed, when the Cl⁻-channel blockers were applied to Et1-stimulated arterioles, Ca²⁺-wave frequency fell back to the control levels seen in untreated vessels. Wave frequency was 0.62 ± 0.1/20 s in the presence of Et1 and DIDS, as compared with 1.23 ± 0.18/20 s under control conditions in the same cells (NS). Similarly, mean frequency was 1.10 ± 0.23/20 s for 9AC and Et1, compared with 0.68 ± 0.13/20 s during the control period (NS). 

We tested for negative feedback via activation of BK current by applying iberiotoxin (100 nM), a selective BK-channel blocker, to Et1-treated vessels. ¹⁹ This increased the amplitude but not the frequency of Ca²⁺-waves (Figs. 2B, 2D). Mean wave amplitude rose from 0.24 ± 0.01 in the presence of Et1 alone to 0.33 ± 0.02 in vessels superfused with Et1 and iberiotoxin (P < 0.0001). Mean frequency was 6.86 ± 0.45/20 s and 6.37 ± 0.44/20 s for Et1 alone and Et1 plus iberiotoxin, respectively (NS, n = 51 cells in four arterioles). Once again, baseline [Ca²⁺]_i wasn't affected, with mean values of 1.54 ± 0.06 with Et1 alone and 1.48 ± 0.05 when Et1 and iberiotoxin were present simultaneously (NS). 

Role of Voltage-Operated Ca²⁺-Channels in Cell Signaling Responses to Et1

Alteration of membrane potential leading to downstream changes in the activation of voltage-operated Ca²⁺-channels represents an important mechanism through which Ca²⁺-sensitive ion channels may modulate global Ca²⁺-signaling. ¹⁰ The L-type Ca²⁺-channel blocker nimodipine (10 μM) inhibits stimulation of Ca²⁺-waves by Et1 (Fig. 2E). ²⁰ Frequency was reduced from 4.78 ± 0.30/20 s in the presence of Et1 alone to 0.93 ± 0.14/20 s when nimodipine was also added (P < 0.0001; n = 71 cells in four arterioles). Mean wave amplitude also decreased from 0.48 ± 0.02 with Et1 alone to 0.31 ± 0.02 in the presence of Et1 and nimodipine (P < 0.0005). Unlike inhibitors of Ca²⁺-activated Cl⁻- and K⁺-currents, however, nimodipine reduced baseline [Ca²⁺]_i back to control levels seen prior to addition of Et1, from a mean value of 1.60 ± 0.10 in the presence of Et1 to 1.00 ± 0.03 with Et1 and nimodipine (P < 0.0001). 

Role of Ca²⁺-Activated Conductances in Et1 Vasoconstrictor Responses in Vivo

The physiological significance of Ca²⁺-activated Cl⁻- and K⁺-currents during Et1-induced vasoconstriction was tested in vivo using fluorescein angiography and confocal imaging with a scanning laser ophthalmoscope. Vessel diameters after intravitreal drug injections were compared with diameters recorded after injection of the same volume (10 μL) of vehicle (Hanks' solution). Et1 (estimated final intraocular concentration = 100 nM) produced a marked vasoconstriction (Fig. 3). Mean arteriolar diameter was reduced from 37.3 ± 2.0 μm after injection of Hanks' solution to 27.7 ± 0.8 μm following injection of Et1 (P < 0.01). Arteriole diameters following injection with Hanks' solution alone were very similar to those in uninjected eyes, indicating that constrictor responses were not a mechanical artifact. 

Figure 3.

 

In vivo changes in arteriolar diameter in response to Et1 and ion channel blockers. (A) Confocal fluorescein angiograms imaged in anesthetized rats with a 15° scan angle. (Scale bars = 100 μm.) Angiograms recorded (i) before and (ii) 15 minutes after intravitreal injection of Et1 are compared. Arterioles are indicated with an arrow. Although similar regions of the eye were imaged, the removal and replacement of the animal necessary for intravitreal injection made it difficult to identify any given arteriole before and after Et1 treatment, particularly as Et1-induced constriction changed vessel landmarks and capillary filling patterns. (iii) Angiogram recorded from a different animal 15 minutes after intravitreal injection of Et1 and DIDS. (B) Column chart plotting summary data for mean arteriolar diameter (±SEM; 15–44 arterioles in 7–10 animals per group). Measurements were made before (control) or after intravitreal injection of 10 μL Hanks' solution or the indicated drug combination. Intraocular drug concentration was estimated from the concentration of the injectate using an assumed vitreal volume of 15 μL. ¹³ Statistically significant differences are marked both for comparisons with injection of Hanks' solution alone (*P < 0.05, **P < 0.01), and with Et1 alone (†P < 0.05) .

Arteriolar diameters following intravitreal injections of Et1 plus DIDS or iberiotoxin were compared with those following injection of Et1 alone. DIDS completely inhibited Et1's constrictor effect, with a mean arteriolar diameter of 40.7 ± 2.6 μm (P < 0.05 vs. Et1 alone; Fig. 3). Neither iberiotoxin nor penitrem A, both BK-channel blockers, ^8,19 produced any statistically significant changes in arteriole diameter when compared with Et1 alone. 

Discussion

Endothelin 1 Stimulates Propagating Ca²⁺-Waves in Retinal Arteriolar Myocytes

We have previously reported that stimulation of retinal arteriolar smooth muscle with Et1 results in an increased frequency of phasic Ca²⁺-signals at the cellular level. ⁵ However, the use of transverse line scanning in these earlier studies meant that the spatial characteristics of these signals could not be determined, so the term Ca²⁺-oscillation was arbitrarily adopted to describe these events. In the current study, 2-D confocal Ca²⁺-imaging revealed that Et1 mainly stimulates propagating Ca²⁺-waves in individual myocytes (Fig. 1C), although synchronized Ca²⁺-oscillations were occasionally seen. These waves were associated with cell shortening, consistent with the hypothesis that they play an important role in arteriolar constriction. ^6,15 Since Ca²⁺-waves were asynchronous in adjacent myocytes, the phasic signaling and mechanical activity observed at the cellular level could explain the tonic constrictor responses to Et1 seen at whole vessel level. ²¹ Similar Ca²⁺-waves have also been described in pulmonary arteriolar and bronchial smooth muscle both in response to constrictor agonists and depolarization by high [K⁺] solutions, suggesting that phasic Ca²⁺-signaling is a common feature of excitation–contraction coupling in smooth muscle. ^22,23 At 42 μm/s, the speed of wave propagation was higher in retinal arteriole myocytes than that reported for pulmonary arterioles (12 μm/s²²) and vasopressin-induced Ca²⁺-waves in cultured vascular smooth muscle (16 μm/s²⁴), but was more similar to that seen in Et1-treated vena cava (29 μm/s²⁵) and bronchiolar smooth muscle (36 μm/s²²). Differences in propagation velocity presumably reflect variations in cellular architecture and the distribution of Ca²⁺-signaling molecules. There was considerable variability in the frequency and amplitude of Et1-stimulated Ca²⁺-waves in different groups of vessels (Fig. 2). Considerable variability in Et1 responses has previously been reported (e.g., in retinal and pulmonary arterioles.) ^{5, 22} Although the mechanisms underlying such variability are unclear, this observation emphasizes the benefit of the “repeated-measures” experimental design to maximize statistical power in such studies. It is also worth noting that Et1 may also increase the Ca²⁺-sensitivity of the contractile machinery via Rho-kinase activation. ²⁶  

Positive Feedback via Ca²⁺-Activated Cl⁻-Current

Several lines of evidence suggest that activation of I_CaCl plays an important role in the vasoconstrictor response to Et1. Two I_CaCl blockers, DIDS and 9AC, inhibited Et1 Ca²⁺-signaling (Fig. 2), reducing wave frequency to resting values seen in the absence of Et1. The physiological significance of this finding was confirmed in vivo, since intravitreal injection of DIDS along with Et1 completely inhibited the vasoconstriction seen with Et1 alone (Fig. 3). DIDS also inhibits the Et1-induced constriction of pressurized retinal arterioles ex vivo. ⁹ Although DIDS and 9AC are two of the most frequently used I_ClCa blockers, they have been shown to block a variety of different Cl⁻-channels as well as other cation channels (see Ref. 11 for review). In control experiments on retinal arteriolar myocytes, however, DIDS had no effect on any of the other currents we have previously identified in rat retinal arteriolar smooth muscle at the concentration used in this study, suggesting a specific action on I_ClCa (data not shown). ⁹ The fact that two structurally distinct blockers produced similar effects on Et1-induced Ca²⁺-signaling also suggests that this reflected a common action on I_ClCa, rather than disparate off-target effects. We believe, therefore, that our results constitute strong evidence that stimulation of Ca²⁺-waves and contraction in arteriolar smooth muscle by Et1 are dependent on activation of I_ClCa . These results also provide indirect evidence that the mechanical response in arteriolar smooth muscle is driven by phasic rather than tonic Ca²⁺-signaling, since DIDS, which completely blocked the stimulation of Ca²⁺-waves and constriction by Et1, failed to inhibit the Et1-induced increase in baseline [Ca²⁺]_i. Previous reports on stimulation of bronchiolar and pulmonary arterial smooth muscle by Et1 and a range of other agonists have come to similar conclusions, again emphasizing the importance Ca²⁺-waves and oscillations in stimulus–contraction coupling. ^22,23,27,28  

The suggestion that I_ClCa is involved in Et1 signaling in retinal arterioles is consistent with electrophysiological studies from our own and other laboratories. G-protein coupled agonist responses in a variety of smooth muscles are believed to be crucially dependent on activation of I_ClCa. ^11,29 Et1 stimulates transient depolarizations in a range of vascular smooth muscle, including that in retinal arterioles, and these events are inhibited by Cl⁻-channel blockade. ^30,31 Et1 also promotes repetitive inward I_ClCa currents in rabbit choroidal smooth muscle. Depletion of intracellular Ca²⁺-stores blocks these currents, suggesting they are triggered by Ca²⁺-release. ³² Transient depolarizations resulting from repetitive activation of I_ClCa would be expected to promote Ca²⁺-influx via activation of voltage-operated Ca²⁺-channels, amplifying Ca²⁺-release to generate phasic waves or oscillations. ^10,11  

Negative Feedback via Ca²⁺-Activated K⁺-Current

It is well established that activation of large conductance Ca²⁺-activated K⁺-currents (BK currents) can inhibit contractility in vascular and other smooth muscle through membrane hyperpolarization and deactivation of L-type Ca²⁺-channels. ^33–35 We obtained some evidence that BK channels exerted negative feedback on Et1-induced Ca²⁺-waves in retinal arteriolar smooth muscle (Fig. 2). Only wave amplitude was affected by the BK-selective inhibitor iberiotoxin, however, with no change in either wave frequency or baseline [Ca²⁺]_i. Intravitreal injection of either iberiotoxin or penitrem A, another BK-blocker, along with Et1 produced no measurable increase in the Et1-induced vasoconstriction (Fig. 3). This is consistent with the hypothesis that Et1-induced vascular constriction is primarily regulated by the frequency, rather than the amplitude, of phasic Ca²⁺-signals. ^5,22 It should be noted, however, that the estimated Et1 concentration within the vitreous may have approached 100 nM, which can generate maximal vascular constriction in vitro. ³⁶ The functional relevance of BK-mediated negative feedback cannot, therefore, be entirely dismissed on the basis of these experiments. 

Role of Other Ion Channels in Et1-Mediated Signaling

It is well established that Ca²⁺-influx via L-type channels plays an important role in agonist stimulation of vascular smooth muscle . ^11–37 Nimodipine reduced both the amplitude and frequency of Et1-induced Ca²⁺-waves in retinal arterioles (Fig. 2), entirely consistent with the hypothesis that membrane depolarization secondary to activation of I_ClCa is an important mechanism in phasic Ca²⁺-signaling. Unlike I_ClCa inhibitors, however, nimodipine also blocked the Et1-induced elevation in baseline [Ca²⁺]_i. This suggests that depolarization by a mechanism independent of I_ClCa is responsible. Electrophysiological studies also show that Et1 produces a tonic decrease in resting membrane potential in retinal arteriolar smooth muscle and that this is not affected by inhibition of I_ClCa. ³¹ One may speculate that Et1 activates a nonspecific cation current, possibly conducted via Transient Receptor Potential Canonical (TRPC) channels, as observed in coronary myocytes . ³⁸  

Conclusion

We propose that activation of I_ClCa represents an important positive feedback mechanism during Et1 signaling in retinal arteriolar myocytes. Stimulation of Et_A receptors by Et1 promotes Ca²⁺-release from the sarcoplasmic reticulum (SR) via both IP₃ and ryanodine gated receptors (IP₃Rs and RyRs). ⁵ This triggers activation of Ca²⁺-sensitive Cl⁻-channels, with the resulting depolarization opening L-type voltage-operated Ca²⁺-channels. Previous studies on smooth muscle have suggested that positive feedback via calcium-induced calcium release from the intracellular store provides an adequate mechanistic explanation for agonist-stimulated Ca²⁺-waves. ²⁴ In retinal arterioles, however, additional Ca²⁺-influx seems to be required.  

Footnotes

 Supported by Wellcome Trust Grant No. 074648/Z/04 (JGM), Juvenile Diabetes Research Foundation Fellowship Grant No. 2-2003-525 (TMC), and European Social Fund Studentship (PB).

Footnotes

¹  These authors contributed equally to this work and therefore should be regarded as equivalent authors.

Footnotes

 Disclosure: M. Stewart, None; M. Needham, None; P. Bankhead, None; T.A. Gardiner, None; C.N. Scholfield, None; T.M. Curtis, None; J.G. McGeown, None