Sympathetic nerve stimulation induces local endothelial Ca²⁺ signals to oppose vasoconstriction of mouse mesenteric arteries

Abstract

It is generally accepted that the endothelium regulates vascular tone independent of the activity of the sympathetic nervous system. Here, we tested the hypothesis that the activation of sympathetic nerves engages the endothelium to oppose vasoconstriction. Local inositol 1,4,5-trisphosphate (IP₃)-mediated Ca²⁺ signals (“pulsars”) in or near endothelial projections to vascular smooth muscle (VSM) were measured in an en face mouse mesenteric artery preparation. Electrical field stimulation of sympathetic nerves induced an increase in endothelial cell (EC) Ca²⁺ pulsars, recruiting new pulsar sites without affecting activity at existing sites. This increase in Ca²⁺ pulsars was blocked by bath application of the α-adrenergic receptor antagonist prazosin or by TTX but was unaffected by directly picospritzing the α-adrenergic receptor agonist phenylephrine onto the vascular endothelium, indicating that nerve-derived norepinephrine acted through α-adrenergic receptors on smooth muscle cells. Moreover, EC Ca²⁺ signaling was not blocked by inhibitors of purinergic receptors, ryanodine receptors, or voltage-dependent Ca²⁺ channels, suggesting a role for IP₃, rather than Ca²⁺, in VSM-to-endothelium communication. Block of intermediate-conductance Ca²⁺-sensitive K⁺ channels, which have been shown to colocalize with IP₃ receptors in endothelial projections to VSM, enhanced nerve-evoked constriction. Collectively, our results support the concept of a transcellular negative feedback module whereby sympathetic nerve stimulation elevates EC Ca²⁺ signals to oppose vasoconstriction.

sympathetic nerves regulate vascular tone through the corelease of ATP and norepinephrine (NE) at neurovascular junctions (41). NE acts on vascular smooth muscle (VSM) cell (VSMC) G_q-coupled α₁-adrenergic receptors [the predominant α-adrenergic receptor isoform in mesenteric VSM (21, 43)] to activate phospholipase C (PLC) and elevate diacylglycerol and inositol 1,4,5-trisphosphate (IP₃), thereby activating PKC and IP₃ receptors (IP₃Rs) in the sarcoplasmic reticulum. ATP released at sympathetic nerve-muscle junctions activates VSMC ionotropic purinergic P2X receptors, specifically P2X₁ receptors (P2X₁Rs), and causes an influx of extracellular Ca²⁺ and Na⁺ into the cell, creating an excitatory junction potential that, if sufficiently large, activates voltage-dependent Ca²⁺ channels (VDCCs) and promotes contraction (23, 33, 46).

Activation of purinergic and adrenergic receptors induces two temporally distinct Ca²⁺ events in VSMCs: fast, rapid-onset junctional Ca²⁺ transients and slower-onset Ca²⁺ waves (23, 24). Junctional Ca²⁺ transients reflect P2X₁R-mediated Ca²⁺ influx (11, 24, 25), whereas Ca²⁺ waves are propagating elevations in Ca²⁺ caused primarily by IP₃-mediated Ca²⁺ release from the sarcoplasmic reticulum (10, 12, 15, 18, 45) that may also contribute to VSM contraction (31, 46).

In addition to perivascular sympathetic nerves, luminal endothelial cells (ECs) also modulate VSMC function through the production of vasodilators such as nitric oxide (NO) and through EDHF (6, 32). However, more recent evidence has suggested that the converse pathway, signaling from VSMCs to ECs, may be important for the regulation of arterial diameter in response to stimuli, such as KCl and phenylephrine (PE), that cause VSMC contraction (4). This observation is consistent with previous reports (4, 14, 37, 42) showing that adrenergic receptor activation by exogenous agonists not only elevates intracellular Ca²⁺ in VSMCs but also increases Ca²⁺ in the ECs lining these vessels. A previous report (16) has described spontaneous local EC Ca²⁺ events in rat mesenteric arteries after the stimulation of VSM with PE, without further defining the location of these Ca²⁺ events.

The current thinking is that this VSM-to-endothelium signal is communicated via the passage of a factor (e.g., Ca²⁺ or IP₃) through myoendothelial gap junctions, which connect VSMCs to endothelial projections through the internal elastic lamina, to elevate EC Ca²⁺ (4, 16, 22). Dora et al. (4) first postulated that α₁-adrenergic agonist-induced increases in endothelial Ca²⁺ resulted from the movement of Ca²⁺ from VSMCs to ECs through myoendothelial gap junctions. However, other studies (19, 22) have shown that elevating VSM Ca²⁺ by membrane depolarization does not affect endothelial Ca²⁺. Instead, recent reports (16, 22) have shown that inhibition of IP₃ generation within VSMCs or blockade of EC IP₃Rs prevents vasoconstrictor-evoked increases in bulk Ca²⁺ in ECs in intact rat mesenteric arteries, suggesting a role for IP₃ in VSM-to-endothelium communication. Although these studies indicated that bath-applied PE can affect EC Ca²⁺ signaling through the engagement of smooth muscle α-adrenergic receptors, it is not known whether sympathetic nerve stimulation can also influence endothelial function.

Endothelial projections to smooth muscle have a distinctive structural architecture containing endoplasmic reticulum elements, gap junctions, IP₃Rs, and intermediate-conductance (IK) Ca²⁺-sensitive K⁺ channels (K_Ca3.1 channels) (27, 30, 36). Using confocal microscopy and connexin (Cx)40^Bac-GCaMP2 mice (27, 39), which express the Ca²⁺ biosensor GCaMP2 specifically in ECs, our laboratory has recently identified and characterized IP₃-mediated Ca²⁺-release events, termed Ca²⁺ pulsars, within endothelial projections of mesenteric arteries to smooth muscle, establishing this approach as a powerful tool for monitoring EC Ca²⁺ signaling at or near the connection points to smooth muscle.

In this study, we tested the hypothesis that brief sympathetic nerve stimulation can engage the vascular endothelium to oppose vasoconstriction. To accomplish this, we determined the effects of sympathetic nerve stimulation on EC Ca²⁺ signaling in intact mesenteric arteries using the GCaMP2 mouse model. Our results are consistent with a model in which NE released from sympathetic nerves activates VSMC α₁-adrenergic receptors to engage EC IP₃Rs in myoendothelial projections and subsequently activate EC IK channels to oppose vasoconstriction.

EXPERIMENTAL PROCEDURES

Mouse models and tissue preparations.

All procedures were approved by the Institutional Animal Care and Use Committee of the University of Vermont. Wild-type C57BL/6 and transgenic Cx40^Bac-GCaMP2 mice were used. The Cx40^Bac-GCaMP2 mouse expresses the Ca²⁺ biosensor GCaMP2 under the control of the Cx40 promoter and thus expresses GCaMP2 only in ECs in the vasculature (27, 39). Mesenteric artery tissue was dissected, and third-order mouse mesenteric arteries were cleaned of connective tissue and removed into HEPES (10 mM)-buffered saline (pH 7.4) at 4°C. For imaging experiments, arteries were slit open longitudinally and mounted on a Sylgard block with the endothelial surface exposed to the bath solution; this en face preparation facilitated the detection and analysis of Ca²⁺ signals.

Solutions.

Experiments using isolated third-order mesenteric arteries were performed in physiological saline solution [containing (in mM) 119 NaCl, 4.7 KCl, 24 NaHCO₃, 1.2 KH₂PO₄, 2 CaCl₂, 1.2 MgCl₂, 7 glucose, and 0.023 EDTA] equilibrated with 95% O₂-5% CO₂ before and during the experiments. Except where indicated, experiments were performed in the presence of capsaicin (1 μM), atropine (10 μM), and MRS-2179 (3 μM) to block sensory nerves, muscarinic receptors, and purinergic P2Y₁ receptors, respectively.

Electrical field stimulation.

Two platinum electrodes (0.25-mm diameter, World Precision Instruments) were placed in parallel on either side of a mesenteric artery (or en face preparation) and connected to a Grass S48 stimulator. After equilibration at 35°C, perivascular nerves of pressurized (80 mmHg) arteries or en face arterial preparations were stimulated by trains of electrical pulses (50–150 V, 0.25-ms pulse, 15 Hz, 5-s train duration). A submaximal voltage was used and was varied initially to some extent until a stable response was achieved. The voltage remained constant throughout both imaging experiments and diameter measurements. The train duration was 5 s to minimize tissue movement for Ca²⁺ measurements in the en face preparation (9, 25).

Diameter measurements of isolated mesenteric arteries.

Arteries were cannulated to size-matched micropipettes on an arteriograph system, pressurized to 80 mmHg using an electronic servo-pressure transducer (Living Systems Instrumentation, Burlington, VT), and continuously perfused with oxygenated, prewarmed (37°C) physiological saline solution. All ion channel blockers were added to the bath solution. Luminal vessel diameter was monitored using a charge-coupled device camera and edge-detection software (IonOptix). Electrical field stimulation (EFS)-evoked constrictions were analyzed using MiniAnalysis (Synaptosoft, Fort Lee, NJ) software. EFS-evoked constrictions were measured as the area under the curve (AUC) and calculated for the region between the baseline and maximum constriction from the point of EFS application to 95% recovery of the diameter to pre-EFS values. Both fast and slow decay phases of constriction were included in the calculations of AUC and thereby provide an index of overall constrictor responses. At least three constrictions per artery were averaged.

Ca²⁺ imaging.

Ca²⁺ imaging was performed on third-order mesenteric artery en face preparations using an Andor Technology Nipkow spinning-disk confocal system coupled to a Nikon Eclipse E600 FN upright microscope with a ×60 water-dipping objective (numerical aperture: 1.0); images were acquired at 15–30 images/s. Cx40^Bac-GCaMP2 mice were used to detect pulsars, which are local, stationary, IP₃R-mediated events (27). Pulsar area, measured at 50% peak amplitude, was found to be ∼14 μm², as previously observed (27). GCaMP2 was excited at 488 nm with a solid-state laser, and fluorescence emission was collected using a 527.5/49-nm band-pass filter. Ca²⁺ pulsars were measured offline by detecting an increase in the fractional fluorescence (F/F_o) that was significantly above background noise (F/F_o > 1.2) using custom-designed software. F/F₀ was evaluated in 9 × 9-pixel regions of interest in the collected image positioned at points corresponding to peak pulsar amplitude; F₀ was obtained from the same region of interest in 10 images without activity. The kinetic properties of pulsars were analyzed using our custom software, which was written by A. D. Bonev (27), and pulsar frequency was determined as number of pulsar events over time per field of view. All events that occurred during EFS, including those that occurred during tissue movement, are presented in histograms. Because some images exhibited movement during EFS, for the sake of consistency, all events that occurred during the 5-s stimulation were excluded in the statistical analysis shown in Fig. 2, which compares the 20-s time bracket after EFS (starting at the end of the EFS pulse) to the 20-s time bracket before EFS for each pharmacological intervention.

Fig. 2.

Inhibition of α-adrenergic receptors prevents EFS-induced increases in Ca²⁺ pulsar frequency. A: summary of the effects of various agents on EFS-induced Ca²⁺ pulsar frequency, comparing the 20-s interval starting immediately after the end of the EFS pulse with the 20-s interval before EFS (control). Events that occurred during EFS were excluded from this analysis because some images showed excessive movement during EFS. EFS significantly increased pulsar activity compared with unstimulated controls. This increase in EC Ca²⁺ pulsars was prevented by pretreatment with 1 μM TTX (n = 8), 500 nM prazosin/10 μM αβ-meATP (n = 6), or 500 nM prazosin alone (n = 18), which reduced pulsar frequency to levels that were not significantly different from unstimulated controls. EFS-induced pulsar activity was not significantly changed by treatment with 10 μM αβ-meATP (n = 10), 10 μM ryanodine (n = 6), or 5 μM nifedipine (n = 5). ns, Not significant. #P < 0.01 vs. unstimulated controls; *P < 0.05, **P < 0.001, and ***P < 0.0001 vs. EFS alone. B: representative fluorescence (F/F₀) traces illustrating the lack of effect of EFS on Ca²⁺ pulsar activity in the presence of prazosin (500 nM). C: summary graph depicting the twofold elevation in pulsar frequency induced by picospritzing 10 μM ATP directly on endothelial cells (ECs) in an en face arterial preparation (n = 3, *P < 0.05). ATP had no effect in the presence of 3 μM MRS-2179 (n = 3), which was routinely included in bath solutions. Picospritzing 100 μM phenylephrine (PE; n = 3) on ECs in an en face preparation did not increase Ca²⁺ pulsar frequency, indicating that neurally released norepinephrine (NE) does not act directly on the endothelium. D: effects of drug treatments on basal pulsar activity. αβ-meATP (10 μM) + prazosin (500 nM), TTX (1 μM), nifedipine (5 μM), ryanodine (10 μM), and blebbistatin (50 μM) had no significant effects on basal pulsar activity (n = 3–6). Application of 10 μM U-73122 significantly reduced basal pulsar activity (n = 3–4; *P < 0.05).

Picospritzing.

ATP or PE was directly applied to the endothelial surface of an en face mesenteric artery preparation by picospritzing for 5 s at a pressure of 0.3 bar. Stimulation/involvement of VSM signaling mechanisms was avoided by adjusting the picospritzer to a distance of ∼10 μm above the endothelium.

Data analysis and statistics.

Prism (GraphPad Software, La Jolla, CA) and OriginPro7.5 software were used for statistical tests and preparation of graphs. Data are expressed as means ± SE unless otherwise noted. For Fig. 1, F and G, means ± SE were calculated using the descriptive “statistics on rows” tool in OriginPro 7.5. For Figs. 2–4, two-tailed paired t-tests were used for the comparison of paired experiments, and two-way repeated-measures ANOVAs followed by Bonferroni post tests were used for the comparison of two or more groups, as appropriate. One-sample t-tests were used to compare a single sample to a hypothetical value. P values of <0.05 were considered statistically significant. Each experimental treatment condition was performed at least three times on at least three separate animals. Two-sample independent t-tests were applied to compare biophysical characteristics (amplitude, duration, and decay and rise times) of pulsars at existing sites before nerve stimulation to those at newly recruited sites after nerve stimulation.

Fig. 1.

Stimulation of nerves increases Ca²⁺ pulsar frequency in the endothelium of third-order mesenteric arteries from connexin40^Bac-GCaMP2 mice. A and B: en face arterial preparations. The plus signs indicate pulsar sites in a field over 20 s before (A) and after (B) electrical field stimulation (EFS). C and D: pseudocolor images showing the initiation sites of Ca²⁺ pulsars (corresponding to the composite images in A and B) 20 s before (C) and after (D) EFS. The color-coded scale represents the fractional fluorescence (F/F₀). E: representative fluorescence (F/F₀) traces illustrating the effect of EFS on Ca²⁺ pulsars originating from pulsar sites. F and G: histograms of before, during, and after EFS under control conditions (n = 16 fields, 8 arteries; F) and in the presence of 1 μM TTX (n = 9 fields, 5 arteries; G). Each bar includes SEs and corresponds to pulsars per field during 4 s. The dashed lines indicate the mean frequency before stimulation.

Fig. 3.

Stimulation of sympathetic nerves induces the recruitment of new EC Ca²⁺ pulsar sites. A: representative traces of the time course of Ca²⁺ pulsar activity at newly recruited sites after EFS. B and C: histograms of Ca²⁺ pulsars at newly recruited pulsar sites (n = 11 fields, 5 arteries; B) and preexisting sites (n = 34 fields, 15 arteries; C). D: summary graph showing a significantly increased number of new Ca²⁺ pulsar sites per field over the course of 20 s after EFS (n = 13, P < 0.0001) compared with the time control in the absence of EFS (n = 3). The number of new Ca²⁺ pulsar sites after EFS was also significantly increased in the presence of 10 μM αβ-meATP (n = 27, *P < 0.0001) but not in the presence of 1 μM TTX (n = 12) or 500 nM prazosin (n = 21).

Fig. 4.

Intermediate-conductance (IK) channels oppose nerve-induced constriction via a mechanism that is dependent on vascular smooth muscle cell α-adrenergic receptor activation and independent of nitric oxide synthase/cyclooxygenase. A–E: representative diameter changes in pressurized (80 mmHg, 166 ± 25-μm maximal diameter and 129 ± 23-μm resting diameter, n = 34 arteries) third-order mesenteric arteries in response to EFS (indicated by arrows). Gray areas indicate area under the curve, which was calculated as an integral of the constriction. A: application of 100 nM charybdotoxin (ChTx) in the presence of 1 μM paxilline (Pax) increased constriction to identical nerve stimulation by twofold (n = 6 arteries). B: application of 300 nM apamin did not cause vasoconstriction (1.4 ± 1.0%, n = 5 arteries) and did not significantly alter nerve-induced vasoconstriction (P > 0.05, n = 5 arteries). This increase in EFS-induced constriction produced by inhibition of IK channels with ChTx was not significantly effected by pretreatment with 100 μM N-nitro-l-arginine methyl ester (l-NAME)-10 μM indomethacin-1 μM Pax (n = 5 arteries; P > 0.05; C) or 10 μM αβ-meATP-1 μM Pax (n = 5 arteries, P > 0.05; D) but was abolished by pretreatment with 500 nM prazosin-1 μM Pax (n = 5 arteries, *P < 0.05; E). Although 10 μM αβ-meATP completely abolished peak constriction in response to EFS, its effects on the area under the curve were minimal. F: summary graph illustrating the changes in luminal diameter upon EFS. For all experiments with ChTx, arteries were first incubated with 1 μM Pax to block smooth muscle large-conductance channels. Control corresponds to pretreatment of vessels with the indicated pathway inhibitors (in the presence of Pax).

RESULTS

Stimulation of sympathetic nerves acts through VSM α-adrenergic receptors to promote Ca²⁺ signaling in ECs.

Intracellular Ca²⁺ was measured in ECs in third-order mesenteric arteries from Cx40^Bac-GCaMP2 mice, which express the genetically encoded Ca²⁺ biosensor GCaMP2 specifically in the vascular endothelium (27, 39). En face preparations were used in these experiments; the field of view was ∼120 × 136 μm and contained ∼14 ECs. Inhibitors of muscarinic receptors, purinergic P2Y receptors, and sensory nerves were included in the bath solution (see experimental procedures).

In the absence of nerve stimulation, mesenteric artery ECs exhibited a basal level of IP₃-mediated Ca²⁺ signals that manifested as transient, stationary IP₃R-mediated Ca²⁺ events (Ca²⁺ pulsars) primarily at holes in the internal elastic lamina (Fig. 1; see also Ref. 27). Nerve stimulation by EFS (15 Hz, 5-s train) increased the frequency of Ca²⁺ pulsars about twofold for about 20 s after EFS (Figs. 1, A–F, and 2A). EC pulsar frequency increased within 2–5 s after the start of the stimulation pulse and returned to basal levels ∼40 s after the cessation of nerve stimulation (Fig. 1F), consistent with the time course of smooth muscle Ca²⁺ signaling observed in third-order mesenteric arteries after EFS (23). Approximately 81% percent of pulsars were located in the vicinity (<2.5 μm) of holes in the internal elastic lamina.

The transient increase in EC Ca²⁺-signaling responses to EFS was prevented by TTX (1 μM, 5-min incubation), a blocker of neuronal Na⁺ channels (Figs. 1G and 2A). Stimulation of sympathetic nerves leads to the release of NE and ATP, which activate α₁-adrenergic and P2X₁Rs, respectively, in VSMCs (41). Prazosin, an inhibitor of α₁-adrenergic receptors, prevented the EFS-induced increase in EC Ca²⁺ signaling (Fig. 2, A and B). In contrast, the P2X₁R agonist/antagonist αβ-meATP (10 μM), which transiently activates and then chronically desensitizes P2X₁Rs, had no effect on EFS-induced increases in EC Ca²⁺ signaling (Fig. 2A). These results are consistent with the idea that the activation of VSMC α₁-adrenergic receptors by nerve-released NE leads to an increase in EC Ca²⁺ signaling.

It is nonetheless conceivable that NE and ATP released by sympathetic nerves might directly affect EC Ca²⁺ signaling. To test this, we directly picospritzed ATP (10 μM) or PE (100 μM) onto the endothelium of an en face arterial preparation and measured changes in EC Ca²⁺ signals (Fig. 2C). PE did not affect EC Ca²⁺ signaling when picospritzed onto the endothelium, suggesting that the third-order mesenteric artery ECs used in this study do not express the molecular machinery necessary to respond directly to NE, a result consistent with a previous report (14). However, picospritzed ATP did cause a significant elevation in EC Ca²⁺ signaling, an effect that was blocked by preincubation with the P2Y₁ receptor antagonist MRS-2179 (3 μM, 10-min incubation; Fig. 2C). Thus, because MRS-2179 is routinely included in all bath solutions (see experimental procedures) (44) and ECs are unresponsive to PE, it is unlikely that neurally derived NE or ATP acts directly on the vascular endothelium under the experimental conditions used.

Recent studies (17, 22) have indicated that VSM can communicate with the endothelium through myoendothelial gap junctions in an IP₃-dependent manner. To explore this possibility, we tested the effects of U-73122 (10 μM), an inhibitor of PLC. Inhibition of PLC greatly reduced basal Ca²⁺ pulsar activity, as previously reported (27) (see also Fig. 2D), and EFS had no effect.

The lack of effect of P2X₁R inhibition on EFS-induced EC Ca²⁺ signaling suggests that P2X₁R-mediated Ca²⁺ influx and activation of VDCCs by excitatory junction potentials are not required for EFS-induced EC Ca²⁺ signaling, further suggesting that VSMC intracellular Ca²⁺ is not the primary mediator of VSM-to-endothelium signaling. To further explore this issue, we tested the effects of antagonists of VDCCs and ryanodine receptors (RyRs) on nerve-induced increases in EC Ca²⁺ signaling. Pretreatment with the VDCC blocker nifedipine (5 μM) minimized movement in response to EFS but did not affect the increase in EC Ca²⁺ pulsar activity induced by sympathetic nerve stimulation (Fig. 2A). Blebbistatin (50 μM), a selective inhibitor of myosin-actin interactions, also minimized contraction without affecting EC Ca²⁺-signaling responses to EFS (2.3-fold increase in pulsar frequency after EFS, P < 0.05, n = 4 arteries). Likewise, the RyR blocker ryanodine (10 μM) failed to block the increase in EC Ca²⁺ pulsar activity induced by the stimulation of sympathetic nerves (Fig. 2A). Basal EC Ca²⁺ signaling was also unaffected by nifedipine, blebbistatin, ryanodine, prazosin, and αβ-meATP (Fig. 2D). Collectively, these results suggest that neither increases in smooth muscle global Ca²⁺ nor elevated RyR-mediated Ca²⁺ spark activity makes a major contribution to communicating sympathetic nerve input to the endothelium, implying a potential role for IP₃ in transducing this signal.

Sympathetic nerve stimulation recruits new EC Ca²⁺-signaling sites.

We (27) have previously shown that the endothelial-dependent vasodilator ACh enhances Ca²⁺ pulsar activity by inducing new sites and by increasing pulsar frequency at existing sites. Thus, EFS could increase EC Ca²⁺ signaling by affecting activity at existing sites, by recruiting new sites, or both. Here, we found that EFS considerably increased the number of new pulsar sites, with the transient increase in pulsar activity peaking at 4 sites/field per 4 s after EFS, as measured over 20 s (Fig. 3, A and B). The amplitude, duration, and rise and decay times of Ca²⁺ pulsars were not significantly changed upon nerve stimulation (Table 1). Interestingly, in contrast to stimulation with ACh, EFS did not increase the frequency of pulsars at existing sites (Fig. 3C and Table 1). Therefore, as shown in Fig. 3D, the increase in frequency is due to the appearance of new sites. The residual number of new sites in the presence of TTX (Fig. 3D) provides a measure of the frequency of spontaneously occurring new sites over the duration of the experiment. The addition of prazosin reduced the number of new sites to basal levels (i.e., that observed in the presence of TTX), indicating that inhibition of α₁-adrenergic receptors prevented the recruitment of new pulsar sites. αβ-meATP did not affect EFS induction of new sites (Fig. 3D). Collectively, these observations indicate that the recruitment of new pulsar sites after sympathetic nerve stimulation is mediated by VSMC adrenergic signaling pathways.

Table 1.

Comparison of Ca²⁺ pulsars from the mesenteric endothelium before and after nerve stimulation

	Ca2+ Pulsars
Parameters	Before EFS	New sites after EFS
Amplitude, F/F0	1.37 ± 0.02	1.37 ± 0.02
Rise time, ms	194 ± 10	189 ± 8
Duration, ms	278 ± 12	251 ± 10
Half-time for decay, ms	175 ± 11	153 ± 8
Frequency, Hz	0.08 ± 0.006	0.07 ± 0.003

The Ca²⁺ pulsar parameters shown are peak amplitude [in fractional fluorescence (F/F₀)], rise time (measured as the 10%-to-90% signal interval), duration (measured as 50% peak signal width), and half-time for decay of the signal. Kinetic characteristics of Ca²⁺ pulsars were calculated from 43 and 96 pulsars recorded in 5 fields over 20-s time courses before and after nerve stimulation, respectively. EFS, electrical field stimulation. “After EFS” refers to Ca²⁺ pulsars originating from new, EFS-induced pulsar sites.

EC IK channels regulate nerve-evoked constriction.

IK channels and IP₃Rs are colocalized to endothelial projections (27, 35), suggesting the potential for local communication between Ca²⁺ pulsars and Ca²⁺-sensitive IK channels. To test the hypothesis that EFS-induced IP₃R-mediated Ca²⁺ signals in ECs activate IK channels to oppose nerve-evoked constriction, we exploited a combination of pharmacological agents with selectivity against the K⁺ channels expressed in ECs and smooth muscle cells. The bee venom toxin apamin was used to selectively block small-conductance (SK) Ca²⁺-sensitive K⁺ channels (K_Ca2.3 channels) (2, 28, 40), which are also expressed in ECs. IK channels were blocked using the synthetic inhibitor Tram-34 and the scorpion toxin charybdotoxin (ChTx), which also blocks large-conductance (BK) Ca²⁺-sensitive K⁺ channels, but not SK channels, with high affinity (3, 28). In experiments using ChTx, the potential effects of BK channels, which are expressed in VSMCs but not in normal ECs (8, 20, 26, 34), were eliminated using the selective blocker paxilline (13, 28, 29). Elevation of intravascular pressure to 80 mmHg induced myogenic tone, constricting these mesenteric arteries by 22.7 ± 0.8% (n = 34). Block of IK channels with Tram-34 (10 μM) or ChTx (100 nM, in the presence of paxilline) caused a small, but significant, constriction when added alone (6.3 ± 0.9% and 4.8 ± 0.6% for Tram-34 and ChTx, respectively, n = 5 arteries; Fig. 4). Block of IK channels also increased EFS-induced vasoconstriction, measured as an integral of the constriction (AUC; Fig. 4, A and F), by approximately twofold, but did not significantly change peak amplitude (n = 5 arteries, P > 0.05). In contrast, application of the SK channel blocker apamin (300 nM) did not cause vasoconstriction (1.4 ± 1.0%, n = 5 arteries) and did not significantly alter nerve-induced vasoconstriction (P > 0.05, n = 5 arteries; Fig. 4B). These results suggest that EC IK channels tonically oppose pressure-induced constriction as well as EFS-induced constriction.

In addition to SK/IK channel-dependent EDHF, the endothelium releases the vasoactive factors NO and prostanoids through the action of NO synthase (NOS) and cyclooxygenase (COX), respectively (5, 32). Block of NOS with N-nitro-l-arginine methyl ester [l-NAME (100 μM)] and COX with indomethacin (10 μM) did not significantly affect nerve-induced vasoconstriction, suggesting that NOS/COX does not play a significant role in opposing nerve-induced vasoconstriction in these mesenteric arteries (P > 0.05, n = 5 arteries). Consistent with this interpretation, block of IK channels in the presence of l-NAME (100 μM)/indomethacin (10 μM) resulted in an increase in nerve-induced vasoconstriction comparable to that observed under control conditions (Fig. 4, C and F). This suggests that IK channels regulate nerve-evoked constriction in small mesenteric arteries independent of NOS/COX activity.

To confirm that the VSMC α₁-adrenergic receptor pathway, as described above (see Fig. 2), is functionally linked with IK-dependent attenuation of nerve-evoked constriction, we preincubated pressurized mesenteric arteries with prazosin (500 nM) or the purinergic receptor blocker αβ-meATP (10 μM) and measured changes in EFS-evoked constrictions in the presence and absence of the IK channel blocker ChTx (100 nM). Prazosin reduced the sustained component of EFS-evoked constrictions, measured as the AUC, by 37.9 ± 2.4% (n = 5 arteries, P < 0.05). The initial rapid constriction, which did not make a major contribution to the total AUC, was eliminated by αβ-meATP (n = 4 arteries; Fig. 4D), confirming that this transient contraction represents the response to an excitatory junction potential caused by the rapid influx of Na⁺/Ca²⁺ through P2X₁Rs. In arteries pretreated with αβ-meATP (in the presence of 1 μM paxilline), ChTx induced an increase in vasoconstriction comparable to that observed in the absence of αβ-meATP (Fig. 4, D and F). In contrast, the increase in vasoconstriction due to IK channel block was abolished in the presence of prazosin (500 nM; Fig. 4, E and F), indicating that the VSMC α₁-adrenergic receptor signaling pathway is the primary contributor to the local increase in Ca²⁺ pulsar events at myoendothelial projections, subsequent activation of proximate IK channels, and regulation of vascular tone.

DISCUSSION

Here, we propose the novel concept that brief stimulation of sympathetic nerves induces an increase in IP₃-mediated Ca²⁺ pulsars in the endothelial projections to smooth muscle. Our data support a model in which sympathetic nerve stimulation does not activate EC Ca²⁺ directly but instead induces heterocellular communication from VSMCs to ECs, leading to increased EC Ca²⁺ pulsar activity at myoendothelial projections (Fig. 5). Sympathetic nerve-induced Ca²⁺ pulsars signal locally to activate EC IK channels, engaging a NOS/COX-independent EDHF mechanism that opposes nerve-induced vasoconstriction. Each of the sympathetic nerve-evoked responses (increased EC Ca²⁺ pulsar activity, IK channel activation, and regulation of vasoconstriction) is mediated by VSMC α₁-adrenergic signaling.

Fig. 5.

Model of negative feedback regulation by the endothelium in response to sympathetic nerve stimulation. Activation of sympathetic nerves leads to the corelease of ATP and NE, which activate vascular smooth muscle cell purinergic receptors (P2X₁Rs) and α₁-adrenergic receptors [α₁-G protein-coupled receptors (α₁-GPCR)], respectively. The subsequent elevation of smooth muscle inositol 1,4,5-trisphosphate (IP₃) by the activation of α₁-adrenergic receptors is transmitted through myoendothelial gap junctions to IP₃ receptors (IP₃Rs) in the endothelial projections to increase intracellular Ca²⁺ (pulsars). This, in turn, activates colocalized EC IK channels, providing a hyperpolarizing influence on the smooth muscle to oppose vasoconstriction. VDCC, voltage-dependent Ca²⁺ channels; PLC, phospholipase C; SR, sarcoplasmic reticulum; IEL, inner elastic lamina; ER, endoplasmic reticulum.

Our finding that VSMC α-adrenergic receptor activation increases EC Ca²⁺ signaling is consistent with previous reports (4, 14, 37, 42) showing that VSMC adrenergic receptor activation by bath-applied PE increases Ca²⁺ in the ECs lining these vessels. Activation of α₁-adrenergic receptors elevates the levels of both smooth muscle IP₃ and intracellular Ca²⁺, which could permeate myoendothelial gap junctions to activate IP₃Rs in the endothelial projections and thereby increase Ca²⁺ pulsar activity. Our data do not support the idea that the movement of Ca²⁺ underlies this transcellular signal. Sympathetic stimulation leads to the release of ATP and NE. ATP signaling through VSMC P2X₁Rs does not appear to be involved in this transcellular communication, indicating that the elevation of VSMC Ca²⁺ via extracellular Ca²⁺ influx is not sufficient to drive this process. The failure of inhibitors of RyRs and VDCCs to prevent the signaling of sympathetic nerves to the endothelium, as reported here, suggests that neither increases in Ca²⁺ spark activity nor rises in global intracellular Ca²⁺ significantly contribute to this signaling cascade, strongly suggesting that communication from VSM to the endothelium is mediated by VSMC IP₃ diffusing through myoendothelial gap junctions. Moreover, the diffusion gradient between VSMCs and ECs for Ca²⁺ is likely to be very modest compared with the IP₃ gradient. The idea that IP₃ acts as the intercellular currency in this context is consistent with the dramatically lower mobility of Ca²⁺ (∼13 μm/s) compared with IP₃ (∼280 μm/s), which, in turn, reflects the strong intracellular buffering of Ca²⁺ by Ca²⁺-binding proteins (1). Earlier reports (4, 16, 37) that stimulation of smooth muscle by either PE or activation of VDCCs by membrane depolarization with high K⁺ increased endothelial Ca²⁺ led to the suggestion that Ca²⁺ is the transcellular signaling entity. However, other studies (19, 22) have reported that activation of VDCCs has no effect on endothelial Ca²⁺. The consensus of recent studies (17, 22, 35) supports the concept that IP₃ is the currency of communication after the stimulation of VSMC adrenergic receptors with PE. Our data reinforce the importance of adrenergic signaling in this mechanism and further define the nature of the EC Ca²⁺ signal, showing that NE transiently released from sympathetic nerves acts through VSMC α₁-adrenergic receptor signaling pathways to increase EC Ca²⁺ pulsar activity. Furthermore, our data suggest that activation of EC IK channels should attenuate the NE-induced smooth muscle membrane potential depolarization.

Previous studies (4, 14, 16, 22, 35, 37) have examined the steady-state effects of continuous PE exposure on endothelial Ca²⁺. Here, we examined the effects of nerve stimulation for 5 s on EC Ca²⁺ signaling associated with endothelial projections to smooth muscle. We found that Ca²⁺ pulsar activity was elevated ∼3 s after the onset of EFS and remained elevated for ∼40 s (Fig. 1). Consistent with the idea that IP₃ derived from stimulated smooth muscle causes a delayed activation of IP₃-mediated Ca²⁺ pulsars in endothelial projections, Wier and colleagues (23) found that smooth muscle IP₃-mediated Ca²⁺ waves remained elevated for 60 s after the cessation of EFS. Our data on nerve-evoked constrictions are also consistent with this view in that block of IK channels affected the duration, measured as the AUC, but not the initial, purinergic-mediated amplitude (Fig. 4). Ca²⁺ pulsars return to baseline within 1 min after stimulation, whereas the constriction lasts longer. This lag presumably reflects the time course of membrane repolarization, reduction in smooth muscle Ca²⁺, and ultimately force. Collectively, these findings are in accord with reports suggesting a role for IP₃ rather than Ca²⁺ in VSM-to-endothelium communication.

The architectural features of endothelial projections through the elastic lamina create a microdomain that forms the basis for the signaling mechanism described here. These endothelial projections are home to evaginations of the endoplasmic reticulum containing IP₃Rs (the source of Ca²⁺ pulsars) and make direct contact with VSMCs via myoendothelial gap junctions, which are composed of Cxs (Cx37 and Cx40), that allow the diffusion of small molecules between VSMCs and ECs (7, 30, 36). Even though endothelial NOS has been shown to colocalize at myoendothelial gap junctions (38), our data suggest that neither NOS nor COX activity plays a significant role in opposing sympathetic nerve-induced vasoconstriction in third-order mesenteric arteries (Fig. 4). Importantly, IK channels have been shown to localize to the myoendothelial projections (27, 36), placing them in a position to respond to local elevations of Ca²⁺ generated by Ca²⁺ pulsar activity. Our data suggest that EC IK channels tonically oppose pressure-induced, as well as nerve-induced, constrictions in small resistance arteries (Fig. 4), similar to their regulation of parenchymal arteriolar diameter and cortical cerebral blood flow (8). Interestingly, the increase in Ca²⁺ pulsar activity observed upon nerve stimulation reflected the recruitment of new Ca²⁺ pulsar sites in ECs rather than an increase in the activity of existing sites (Figs. 1 and 3). Although our experiments did not address this directly, this outcome is precisely what one would predict if the “goal” was to create a system with maximal gain, assuming that IK channels in the vicinity of existing pulsar sites are at least partially activated under basal conditions. The effect of EFS on pulsars was somewhat different from that of ACh, which predominantly recruited new pulsar sites but also increased the activity of existing sites (27). This difference could reflect the source of IP₃ generation (endothelium vs. smooth muscle) and the degree to which IP₃Rs are saturated with IP₃ and Ca²⁺.

Taken together, our data elucidate a novel signaling mechanism in which the stimulation of sympathetic nerves activates VSMC adrenergic receptors to increase EC Ca²⁺ pulsar activity in endothelial projections, ultimately activating nearby IK channels to oppose nerve-induced vasoconstriction and regulate sympathetic tone. Thus, the endothelium actively responds to sympathetic nerve stimulation to regulate vessel diameter and sympathetic tone.

GRANTS

This work was supported by American Heart Association Postdoctoral Fellowship 10POST3870008 (to L. W. M. Nausch), National Heart, Lung, and Blood Institute Grants P01-HL-095488-01 and R01-HL-44455, and the Totman Trust for Medical Research.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

Author contributions: L.W.N. and M.T.N. conception and design of research; L.W.N. and A.D.B. performed experiments; L.W.N. and A.D.B. analyzed data; L.W.N., A.D.B., T.J.H., and M.T.N. interpreted results of experiments; L.W.N. and A.D.B. prepared figures; L.W.N. drafted manuscript; L.W.N., A.D.B., T.J.H., and M.T.N. edited and revised manuscript; L.W.N., A.D.B., T.J.H., Y.N.T., M.I.K., and M.T.N. approved final version of manuscript.