Divergent signalling mechanisms for venous versus arterial contraction as revealed by Endothelin-1

Abstract

Objective

Venous function is underappreciated in its role in blood pressure determination, a physiological parameter normally ascribed to changes in arterial function. Significant evidence points to the hormone endothelin-1 (ET-1) as being important to venous contributions to blood pressure. We hypothesized that the artery and vein should similarly depend on the signaling pathways stimulated by ET-1, specifically phospholipase C (PLC) activation. This produces two functional arms of signaling: diacylglycerol (DAG; protein kinase C activation) and inositol trisphosphate (IP₃) production (intracellular calcium release).

Methods

The model was the male Sprague Dawley rat. Isolated tissue baths were used to measure isometric contraction. Western blot and immunocytochemical analyses measured the magnitude of expression and site of expression, respectively, of IP₃ receptors in smooth muscle/tissue. Pharmacological methods were used to modify phospholipase C activity and signaling elements downstream of phospholipase C (IP₃ receptors, protein kinase C).

Results

ET-1-induced contraction was phospholipase C-dependent in both tissues as the phospholipase C inhibitor U73122 significantly reduced contraction in aorta (86±4% of control, P<.05) and vena cava (49±11% of control, P<.05). However, ET-1-induced contraction was not significantly inhibited by the IP₃ receptor inhibitor 2-APB (100 μM) in vena cava (82±8% of control, P=.23) but was in the aorta (55±4% of control, P<.05). All three IP₃ receptor isoforms were located in venous smooth muscle. IP₃ receptors were functional in both tissues as the novel membrane-permeable IP₃ analogue (Bt-IP₃; 10μM) contracted aorta and vena cava. Similarly, while the PKC inhibitor chelerythrine (10μM) attenuated ET-1-induced contraction in vena cava and aorta (5±2% and 50±5% of control, respectively; P<.05), only the vena cava contracted to the DAG analogue 1-oleoyl-2-acetyl-sn-glycerol (OAG).

Conclusions

These findings suggest that ET-1 activates phospholipase C in aorta and vena cava, but vena cava contraction to ET-1 may be largely IP₃-independent. Rather, DAG – not IP₃ – may contribute to contraction to ET-1 in vena cava, in part by activation of protein kinase C. These studies outline a fundamental difference between venous and arterial smooth muscle and further reinforce a heterogeneity of vascular smooth muscle function that could be taken advantage of for therapeutic development.

Introduction

More attention has been given to the physiology of veins since researchers linked changes in venous capacitance to increases in blood pressure ¹. The role of veins in regulating blood pressure is still largely overlooked, even though it was noted over 25 years ago that human hypertensive patients demonstrated impaired venous distensibility and decreased venous capacitance ^{2, 3}. This change in distensibility could ultimately increase blood pressure by increasing arterial blood volume as the storage capacity of veins decreases. Nonetheless, the physiological and signalling mechanisms regulating venous contraction are largely unexplored. They are assumed to be similar to arteries, but this assumption should be tested; this is the goal of this present study. We test the general hypothesis that contraction to a hormone important to both arterial and venous function uses similar signalling pathways.

We focus on the functions of the hormone endothelin-1 (ET-1) because of the high potency of ET-1 in contracting venous tissue^4–6, and strong evidence that ET-1 supports venous function as a means to elevate blood pressure ^{1, 7, 8}. ET-1 is a 21-amino acid peptide, originally characterized as an endothelium-derived constricting factor in the vasculature ⁹. The physiological responses elicited by ET-1 are attributed to the two G protein-coupled receptors to which ET-1 binds: the ET_A and ET_B receptor ^{10, 11}. As with other Gα_q-coupled receptors, both ET receptors can mediate Ca²⁺ release from intracellular stores by activation of phospholipase C and the subsequent production of IP₃ and diacylglycerol (DAG) ^{12, 13}. While sarcoplasmic reticulum (SR) Ca²⁺ release by ET-1 is primarily IP₃-mediated in arterial smooth muscle, the activation of protein kinase C (PKC) by DAG can inhibit IP₃ production to reduce IP₃-dependent Ca²⁺ release and contraction ^{14, 15}. However, independent of PKC activation, DAG can directly activate several types of non-selective cation channels to increase Ca²⁺ influx and augment contraction¹⁶. This is true in rabbit portal vein, where activation of PKC by DAG is necessary for IP₃-mediated Ca²⁺ signalling to occur ¹⁷. This suggests that the mechanisms regulating Ca²⁺ release and Gα_q-mediated contraction vary widely, and the mechanisms are not wholly identical between arteries and veins.

This study tests the hypothesis that ET-1-mediated venous contraction, like arterial contraction, depends on phospholipase C activation and both arms of PLC signalling - DAG and IP₃ production. Using pharmacological inhibitors, we first investigated the role of PLC in ET-1-induced contraction in aorta and vena cava. We then examined how IP₃ and DAG can mediate contraction in aorta and vena cava, and how contraction to ET-1 is mediated by IP₃ and DAG. Our results suggest that ET-1-induced contraction of rat aorta is mediated in part by IP₃, but ET-1-induced contraction in rat vena cava primarily involves DAG and not likely IP₃.

Methods

Animal Care and Use

All procedures that involved animals were performed in accordance with the Institutional Animal Care and Use Committee and the Guide for the Care and Use of Laboratory Animals at Michigan State University. Male Sprague-Dawley rats (SD) (250–300 g; 8–12 weeks old) were used. Animals were euthanized with sodium pentobarbital (60 mg/kg IP).

Isometric Contraction and Compound Source

Aorta and vena cava were dissected and cleaned of outer adipose tissue in physiological salt solution (PSS) containing (mM): NaCl, 130; KCl, 4.7; KH₂PO₄, 1.18; MgSO₄•7H₂O, 1.17; NaHCO₃, 14.8; dextrose, 5.5; Na₂EDTA•2H₂O, 0.03; CaCl₂, 1.6; (pH=7.2). Endothelium-intact tissue rings were then mounted in warmed, aerated PSS (37°C; 95/5% O₂/CO₂) in isolated tissue baths (20 mL) for measurement of isometric contractile force using a 750 TOBS Tissue Organ Bath System (Danish Myo Technology, Aarhus, Denmark) and PowerLab for Windows (ADInstruments, Colorado Springs, CO, USA). The tissues were placed under optimum resting tension (1 g for vena cava, 4 g for aorta) ^{11, 18} and initially challenged with 10μM norepinephrine (vena cava) or phenylephrine (aorta) to test for tissue viability. Different agonists were used for the initial challenge because vena cava do not respond to phenylephrine, and to remain consistent with previously published work ^{5, 6, 18, 19}. Endothelium viability was confirmed by relaxation to 1μM acetylcholine after contraction by phenylephrine (aorta) or norepinephrine (vena cava). Tissues were washed every 15 min until they returned to resting tension. Cumulative concentration response curves or responses to single concentrations of agonists were recorded. Antagonists, inhibitors, or their vehicles were incubated with the tissues for 1h prior to addition of agonists. The specific agonists and antagonists, and corresponding solvents, were: 1-Oleoyl-2-acetyl-sn-glycerol (OAG), acetonitrile/ethanol; 2-aminoethoxydiphenylborane (2-APB), dimethyl sulfoxide (DMSO); Bt-IP₃, DMSO; chelerythrine, DMSO; ET-1, dH₂O; norepinephrine, dH₂O; phenylephrine, dH₂O; U-37122, DMSO; and U-73343, DMSO. All compounds were purchased from Sigma-Aldrich Corporation (St Louis, MO USA), with the following exceptions: Bt-IP₃ (SiChem, Bremen, Germany); ET-1 (1–21) (Bachem, Torrance, CA USA); and OAG (Cayman Chemical, Ann Arbor, MI USA).

Protein Isolation and Western Blot Analysis

Endothelium-intact tissues were ground with mortar and pestle under liquid nitrogen in 1 mL of ice-cold homogenation buffer (50mM Tris (pH 7.4), 4% SDS, 20% glycerol, 0.5mM phenylmethylsulfonyl fluoride, 1mM orthovanadate, 10-μg/mL aprotinin, 10-μg/mL leupeptin). Homogenate was vortexed, sonicated, transferred to a plastic centrifuge tube, and spun at 4°C to pellet debris; the supernatant was then kept. A Bicinchoninic Acid (BCA) assay was used to determine protein concentration. Due to the high molecular weight of IP₃R protein (~300 kDa), Western blotting was performed using techniques for high molecular weight proteins as outlined in current literature ^20–22. Samples (4:1 in denaturing sample buffer, boiled for 5 min) were separated on gradient (8–15%) Tris-acetate gels. Proteins were then wet-transferred to nitrocellulose membrane at 30 V for 1 h at 4°C. Membranes were blocked for 3–4 h (phosphate-buffer saline (PBS), 5% Bio-Rad^® milk). Blots were probed for between 1 h to overnight with primary antibody (rocking, at 4°C), rinsed three times in PBS + Tween (0.1%) with a final rinse in PBS and incubated with the appropriate secondary antibody for 1 h at 4°C (rocking). Primary antibodies used included: anti-IP₃R1 (1:1000; Neuromab, Davis, CA USA); anti-IP₃R2 (1:1000; Millipore, Billerica, MD USA); and anti-IP₃R3 (1:1000; Millipore, USA). Methods of detection included standard ECL capture on film or digital capture using a Licor-Fc (Li-Cor, Lincoln NE, USA). Band density was quantified using ImageJ software (NIH, Bethesda MD, USA).

Smooth Muscle Cell Dissociation and Immunofluorescence

Whole aorta and vena cava tissues were isolated, cleaned of perivascular fat, and cut into ~1 mm rings. Rings were transferred to microcentrifuge tubes and incubated with dissociation solution (80mM NaCl, 80mM monosodium glutamate, 5.6mM KCl, 20mM MgCl₂, 10mM HEPES, 10mM glucose, and 1-mg/mL BSA, pH 7.3) with 1-mg/mL dithiothreitol and 0.3-mg/mL papain for 18 min in a 37°C water bath. The solution was removed and replaced with fresh dissociation solution containing 100μM CaCl₂ and 1-mg/mL collagenase and incubated 9 min in a 37°C tissue bath. The solution was removed and cells were re-suspended in dissociation solution by gentle trituration. Cells were transferred to coverslips using a Shandon Cytospin 4 Centrifuge (Thermo Scientific, Waltham, MA, USA). Cells were then fixed in Zamboni’s fixative for 20 min, permeabilized with 1% Triton X-100 in PBS for 20 min, and blocked with 1% goat serum (diluted in PBS) for 1 h at 37°C. Primary antibodies (diluted in blocker) were added to the coverslips, and cells were incubated at 37°C for 1 h. Antibodies used included: mouse anti-IP₃R1, 1:1000 (NeuroMab, Davis CA, USA); rabbit anti-IP₃R2, 1:1000 (Millipore); rabbit anti-IP₃R3, 1:1000 (EMD Millipore, Billerica MA, USA); rabbit anti-α-actin, 1:100 (Abcam, Cambridge MA, USA); and FITC-conjugated mouse anti-α-actin, 1:1000 (Sigma-Aldrich, St. Louis MO, USA). Coverslips were washed briefly 3 times with PBS, and coverslips were incubated in the appropriate secondary antibodies (goat anti-mouse Alexa Fluor 568, 1:1000; goat anti-rabbit 568, 1:1000; and goat anti-rabbit 488, 1:1000, Life Technologies, Grand Island NY, USA) for 1 h at 37°C. Coverslips were washed 3 times with PBS and placed face down onto slides in Prolong Gold with DAPI (Life Technologies). Cells were then imaged using Olympus^® FV1000 confocal system mounted on a Nikon^® inverted microscope.

Statistical Analysis

All data from contractility experiments were normalized to the maximal tissue contraction during initial adrenergic challenge. Mean, standard error and variance was calculated from the normalized calculated data. For comparisons of two samples of equal variance, statistical significance between groups was established using a two-tailed, unpaired Student’s t test (α=0.05). For samples of unequal variance, the Mann-Whitney U test was used (α=0.05). For multiple sample comparisons, two-way ANOVA was used followed by Bonferroni post hoc analysis to compare individual means. Calculations were performed using Microsoft Excel (Microsoft Corporation, USA) and GraphPad Prism (GraphPad Software Inc., USA).

Experimental Clarification

We began studies on aorta versus vena cava over a dozen years ago and, as is done in tissue bath experiments, had to develop a method to normalize responses from experiment to experiment. We originally started with PE as an agonist for normalization for all tissues, but quickly found that the vena cava did not contract to PE, an α1 adrenergic receptor agonist. We decided that understanding adrenergic receptor activation was important, as both blood vessel types are innervated by the sympathetic nervous system. If isolated tissues did not contract to a stimulus of adrenergic receptors to a set magnitude, then the tissue would not be included in experimental outcomes. We choose this over KCl because KCl had the possibility of activating nerves that were intrinsic to the blood vessels, and thus not be as pure a readout of smooth muscle function. One could argue it would be better to use NE for wakeups in both tissues, but we were hesitant to do this because we had significant amounts of historical data in the using NE in vena cava and PE in the aorta. In the vena cava, substances that cause a greater maximum contraction than NE do not wash out well (U46619; ET-1) or confound experimental outcomes (ET-1). As such, adoption of NE for initial challenge in the vena cava was the best compromise.

Results

Phospholipase C-Mediates Contraction to ET-1 in artery and vein

Isolated vessels contracted to their original challenge with the magnitude of ~2400 mg (PE; aorta) and ~60 mg (NE, vena cava). The magnitude of contraction stimulated by a maximum concentration of ET-1 (100 nM) was ~2700 mg in aorta and ~300 mg in vena cava. These data underscore the significant efficacy of ET-1 in the vena cava when compared to an adrenergic stimulus.

We first examined the effects of PLC inhibition on ET-1-induced contraction using the PLC inhibitor U-73122 (1–10μM) and its inactive analogue U-73343 (1–10μM). U-73122 (1μM) significantly inhibited maximal contraction to ET-1 in aorta and vena cava (fig. 1a, b). Inhibition of maximal contraction to ET-1 was significantly greater in vena cava (61.54±8.72% reduction) as compared to aorta (12.55±3.62% reduction). In both tissues, 1μM U-73343 had no effect (fig. 1c, d). Increasing the concentration of U-73122 from 1 to 10μM caused a more robust inhibition of ET-1-induced contraction in aorta (70.83±7.05% reduction), and vena cava (91.04±2.97% reduction) (fig. 2a, b). However, ET-1-induced contraction was also inhibited by 10μM U-73343 in aorta and vena cava (fig. 2c, d). While this suggested some non-specific inhibition by U-73122, the effects of U-73122 were still significantly greater than the effects of the inactive analogue U-73343 in rat aorta. These findings were also not due to non-specific effects of U-73122 on sarcoplasmic reticulum calcium ATPase, since the irreversible calcium ATPase inhibitor thapsigargin (1μM) had no effect on ET-1-induced contraction in either aorta or vena cava (data not shown). These data suggest that ET-1-induced contraction is PLC-dependent in both aorta and vena cava, but ET-1-induced contraction in vena cava is more sensitive to PLC inhibition than aorta.

Figure 1. Effects of the phospholipase-C inhibitor U-73122 (1μM) on endothelin-1-induced contraction in aorta and vena cava.

Measurement of endothelin-1-induced contraction in rat aorta and vena cava in the presence or absence of 1μM U-73122 (a, b) or its inactive analogue U-73343 (1μM) (c, d). Vehicle (0.01% DMSO) or antagonists were incubated with tissue for 1 h prior to agonist exposure. Points represent mean ± SEM for the N indicated in parentheses. ET-1=endothelin-1; NE = norepinephrine; PE = phenylephrine. * = P<.05 versus vehicle.

Figure 2. Effects of the phospholipase-C inhibitor U-73122 (10μM) on ET-1-induced contraction in aorta and vena cava.

Measurement of endothelin-1-induced contraction in rat aorta and vena cava exposed to vehicle, 10μM U-73122 (a, b) or its inactive analogue U-73343 (10μM) (c, d). Vehicle (0.1% DMSO) or antagonists were incubated with tissue for 1 h prior to agonist exposure. Points represent mean ± SEM for the N indicated in parentheses. ET-1=endothelin-1; NE = norepinephrine; PE = phenylephrine. * = P<.05 versus vehicle.

IP₃ Receptor Proteins are present in both arterial and venous smooth muscle cells

Whole-tissue homogenates of rat aorta and vena cava were used to investigate IP₃R protein expression by Western blot (fig. 3). When probed with antibodies against each of the three IP₃R subtypes, a ~270 kDa band was present in all vascular tissues (fig. 3a–c, black arrows) that coincided with bands expressed in positive controls for IP₃R1 (brain), IP₃R2 (brain and liver) and IP₃R3 (brain) at the same molecular weight. The extra bands seen in Fig. 3b and Fig. 3c have been attributed to cross-linking reactions and/or IP₃R protein degradation ²³. As such, the quantification of IP₃R protein expression in aorta and vena cava samples (figs. 3d, e) may not be entirely representative of actual protein expression. Because of this, the presence of IP₃R protein in smooth muscle was also investigated using immunofluorescence and confocal microscopy in freshly dissociated smooth muscle cells from aorta and vena cava. Aortic smooth muscle cells showed positive red immunofluorescence for IP₃R-1 (fig. 4a), IP₃R-2 (fig. 4b) and IP₃R-3 (fig. 4c). Cells that positively expressed smooth muscle alpha-actin (fig. 4d–f) were also positive for IP₃R protein (fig. 4g–i), indicating that these cells were smooth muscle cells and not another cell type. The fluorescence for both IP₃R and alpha-actin was significantly greater than in samples using secondary antibody alone (fig. 4j–l). Vena cava smooth muscle cells also showed positive red immunofluorescence for all three IP₃R subtypes (fig. 5a–c) in smooth muscle cells expressing alpha-actin (fig. 5d–f). As in aorta, fluorescence for both IP₃R and alpha-actin in vena cava smooth muscle cells (fig. 5g–i) was significantly greater than in samples using secondary antibody alone (fig. 5j–l).

Figure 3. Representative Western blot analysis of IP₃ receptor protein expression.

(a–c) Representative Western blot analysis of IP₃ receptor protein expression from 50-μg whole-tissue homogenate from rat aorta (RA) and vena cava (RVC). Homogenate from rat brain (Br) and liver (L) are also included as controls. Blots were probed using antibodies against IP₃R-1 (a), IP₃R-2 (b) and IP₃R-3 (c), as well as β-actin (loading control). Representative of more than 3 experiments for each receptor. (d, e) Summary bar graphs densitometry of IP₃ receptor densitometry in aorta (d) and vena cava (e), normalized to β-actin (loading control). * = P<.05 (one-way ANOVA and Tukey’s post hoc comparison).

Figure 4. Representative immunohistochemical staining for each of the three IP₃ receptor subtypes in freshly dissociated smooth muscle cells from rat aorta.

(a–c) Red fluorescence indicates the presence of punctate fluorescent staining (inset) for IP₃R-1 (a), IP₃R-2 (b) and IP₃R-3 (c) protein. (d–f) Green fluorescence indicates staining for smooth muscle α-actin. (g–i) Overlay of IP₃R (red), smooth muscle α-actin (green) and DAPI nuclear stain (blue). (j–l) Negative controls, where primary antibodies were absent. Representative of 3 experiments.

Figure 5. Representative immunohistochemical staining for each of the three IP₃ receptor subtypes in freshly dissociated smooth muscle cells from rat vena cava.

(a–c) Red fluorescence indicates the presence of punctate fluorescent staining (inset) for IP₃R-1 (a), IP₃R-2 (b) and IP₃R-3 (c) protein. (d–f) Green fluorescence indicates staining for smooth muscle α-actin. (g–i) Overlay of IP₃R (red), smooth muscle α-actin (green) and DAPI nuclear stain (blue). (j–l) Negative controls, where primary antibodies were absent. Representative of 3 experiments.

IP₃ analog stimulated contraction in both arteries and veins

To test the relationship between IP₃ and smooth muscle function, isometric contraction of aorta and vena cava was measured in the presence of the membrane permeable IP₃ analogue, Bt-IP₃. The vehicle bars quantify the drift in tissue baseline that occurred over the long time of Bt-IP₃ incubation. Bt-IP₃ (10μM) caused a significant and prolonged contraction in aorta (fig. 6a) and vena cava (fig. 6b) that was significantly greater than with vehicle (fig. 6c, d), suggesting that IP₃ can induce contraction in both aorta and vena cava.

Figure 6. Rat aorta and vena cava contract to the membrane permeable IP₃ analogue Bt-IP₃, and effects of 2-APB on ET-1-induced contraction.

(a, b) Representative tracings of rat aorta (a) and vena cava (b) contraction during exposure to Bt-IP₃, a membrane permeable analogue of IP₃. Shown are responses from tissues incubated with vehicle (0.1% DMSO) (a and b, left) and 10μM Bt-IP₃ (a and b, right). Representative of 7 experiments. (c, d) Summary Bar graph of contraction in aorta and vena cava to 10μM Bt-IP₃. Black bars represent maximal contraction to 10μM Bt-IP₃. White bars represent maximum contraction to vehicle (1% DMSO/0.1% Pluronic). Bars represent mean ± SEM for the number of animals indicated in parentheses. (e, f) Contractile response to increasing concentrations of ET-1 in rat aorta (e) and vena cava (f), in the presence or absence of the IP₃ receptor antagonist 2-APB (100μM). Vehicle or antagonists were incubated with tissue for 1 h prior to ET-1 exposure. NE = norepinephrine; PE = phenylephrine. * = P<.05 versus vehicle.

IP₃ Receptor Inhibition blocked arterial but not venous ET-1-induced contraction

Because functional IP₃R were present in both aorta and vena cava, we next investigated the role of IP₃ receptor activation during ET-1-induced contraction. While having no effect on resting tension in either tissue, the IP₃ receptor antagonist 2-APB (100μM) significantly attenuated ET-1-induced contraction in aorta (fig. 6e) but not vena cava (fig. 6f). This was not due to the known non-specific effects of 2-APB on non-selective cation channels or voltage-gated Ca²⁺ channels ^{24, 25}, since several inhibitors of non-selective cation channels and voltage-gated Ca²⁺ channels had no effect on ET-1-induced contraction in either tissue (Table I). This suggests that IP₃ is an important regulator of ET-1-induced contraction in aorta but not vena cava.

Table I.

Measurement of endothelin-1 potency and efficacy, as derived from isometric contractility concentration response data.

AORTA
		Maximum (% PE)		log(EC50)
Drug	Conc. (μM)	Vehicle	Exposed	Vehicle	Exposed
Nifedipine	1	140±7%	132±25%	−8.46±0.06	−8.29±0.14
Diltiazem	10	149±13%	114±12%	−7.97±0.05	−8.02±0.04
SKF-96365	10	144±14%	113±20%	−7.92±0.11	−7.75±0.42
LOE-908	10	160±3%	142±13%	−8.15±0.03	−8.28±0.05
Nif/SKF/LOE	0.05/10/10	171±30%	111±15%*	−8.06±0.05	−8.21±0.06

VENA CAVA
		Maximum (% NE)		log(EC50)
Drug	Conc. (μM)	Vehicle	Exposed	Vehicle	Exposed
Nifedipine	1	303±33%	307±90%	−8.87±0.11	−8.83±0.25
Diltiazem	10	612±96%	581±33%	−8.14±0.08	−8.13±0.06
SKF-96365	10	538±108%	438±77%	−8.01±0.18	−8.03±0.13
LOE-908	10	551±71%	478±75%	−8.18±0.09	−8.36±0.09
Nif/SKF/LOE	0.05/10/10	748±41%	516±93%*	−8.29±0.03	−8.36±0.09

Maximum response to endothelin-1 is shown as percent of phenylephrine contraction (aorta, %PE) or norepinephrine contraction (vena cava, %NE). Potency (EC₅₀[M]) data are given as log(EC₅₀) to allow for standard error calculation and statistical comparison. L-VGCC = L-type voltage-gated Ca²⁺ channel; NSCC = non-selective cation channel; SOCC = store-operated Ca²⁺ channel; Nif = nifedipine; SKF = SKF-96365; LOE = LOE-908. For each experiment, N > 5.

^*P<.05 versus control.

DAG-Mediated Contraction is important in veins but not arteries

To test the role of DAG, the other arm of PLC signalling, and smooth muscle function, we examined the ability of the membrane permeable DAG analogue 1-oleoyl-2-acetyl-sn-glycerol (OAG) to cause contraction in aorta and vena cava (fig. 7a, b). OAG caused a significant and concentration-dependent contraction in vena cava (fig. 7c) that was absent in aorta (fig. 7d). This contraction was completely reversed by the protein kinase C (PKC) cell permeable, catalytic domain inhibitor chelerythrine (10μM), indicating that OAG-mediated contraction was due to PKC activation (fig. 7e). Neither OAG nor chelerythrine impacted tissue viability, as both aorta and vena cava were able to maximally contract to 10μM phenylephrine or 10μM norepinephrine at the end of the experiment (data not shown). These data suggest that DAG, by activation of PKC, may facilitate contraction in vena cava but not aorta.

Figure 7. OAG-induced contraction in aorta and vena cava and PKC-dependence.

(a,b) Representative tracings of aorta and vena cava contractions to increasing concentrations of OAG (0.01μM – 100μM, marked by arrows). (c,d) Measurement of OAG-induced contraction in aorta and vena cava. (e) Summary Bar graph representing relaxation of OAG-induced contraction in aorta and vena cava in the presence of chelerythrine (10μM). Black bars represent maximal contraction to OAG (100μM). White bars represent maximum contraction to OAG after addition of chelerythrine (10μM). (f, g) Contractile responses to increasing concentrations of ET-1 in rat aorta (f) and vena cava (g) in the presence or absence of the protein kinase C antagonist chelerythrine (10μM). Vehicle (0.1% DMSO) or antagonists were incubated with tissue for 1 h prior to ET-1 exposure. In (c-f), points and bars represent mean ± SEM for the number of animals indicated in parentheses. Chel = chelerythrine; NE = norepinephrine; PE = phenylephrine. * = P<.05 versus vehicle.

PKC Inhibition profoundly reduces venous ET-1-induced contraction

To test the effects of PKC inhibition on ET-1-induced contraction, isometric contraction of aorta and vena cava to ET-1 was measured in the presence or absence of chelerythrine (10μM). While chelerythrine significantly attenuated ET-1-induced contraction in aorta (fig. 7f), ET-1-induced contraction in vena cava was nearly abolished by the same concentration of chelerythrine (fig. 7g). These data suggest that ET-1-induced contraction in vena cava is PKC-dependent, and reinforces our finding that DAG is an important regulator of contraction in vena cava.

Discussion

Veins have been of interest in our laboratory because of strong evidence that suggests they can support blood pressure, in particular when considering the potential role of ET-1. Different from arteries, veins do not easily desensitize to ET-1 (1–21)⁶, are significantly sensitive to reactive oxygen species⁵, and contain highly active xanthine oxidase²⁶. More generally, veins have faster kinetics of agonist-induced contraction⁴. The role of ET-1 in elevating venomotor tone to support high blood pressure has been clarified in the deoxycorticosterone salt rat model of hypertension^27–30. Collectively, these reasons were the impetus for studying an important contractile pathway in arteries versus veins, given that its function supports many of the events described above. The principal and novel finding of this study is that ET-1-induced contraction in vena cava involves PLC-dependent production of DAG, and not IP₃. Vena cava are more sensitive to PLC inhibition than aorta, but ET-1-induced contraction in vena cava is unaffected by the IP₃ receptor antagonist 2-APB. Instead, contraction to ET-1 in vena cava is due in large part to the actions of PKC. These findings highlight the differential activation of a pathway thought to be common for ET receptors in arteries and veins, as illustrated in figure 8. As this is one of the first studies of its kind, comparison between our findings and those of other is difficult. An extensive literature search yielded only two other papers comparing ET-1-induced signaling in arteries versus veins^{31, 32}. Thus, the hypotheses put forth in this study have been relatively untested prior to now.

Figure 8. Cartoon depicting differences in arterial vs venous ET-1 signalling.

In aorta, ET-1-induced contraction is largely dependent phospholipase C activation and IP₃-mediated Ca²⁺ release. In vena cava, however, IP₃ receptors are minimally involved in ET-1 signalling. Instead, DAG, either directly or by activating protein kinase C, may increase the influx of calcium through calcium–permeable ion channels and initiate contraction. However, the exact mechanisms by which DAG and protein kinase can influence contraction remain relatively unknown and may be independent of calcium channel activation. CC = Ca²⁺-permeable ion channel; DAG = diacylglycerol; ET-1 = endothelin-1; ETR = endothlin-1 receptor; IP₃ = (1,4,5) inositol trisphosphate; PIP₂ = phosphatidylinositol bisphosphate; PKC = protein kinase C; PLC = phospholipase C; RyR = ryanodine receptor; SERCA = smooth endoplasmic reticulum Ca²⁺ ATPase.

PLC mediates ET-1-induced contraction in both artery and vein

PLC is most proximal to the ET receptor in this signaling pathway. As expected, both aorta and vena cava contraction to ET-1 was markedly attenuated by the PLC inhibitor U-73122 (10μM). However, inhibition by a lower concentration of U-73122 (1μM) was significantly greater in vena cava as compared to aorta. This suggests that contractile ET receptors in veins signal qualitatively through a similar Gα_q-mediated pathway as is seen in arteries ^{10, 33}. However, the marked difference in sensitivity to U-73122 and U-73343 in vena cava as compared to aorta suggests that differences do exist after PLC activation and the formation of IP₃ and DAG. The difference in sensitivity could also be due to an increased importance of Ca²⁺ influx during ET-1-induced contraction of aorta, as mibefradil-sensitive Ca²⁺ channels mediate a substantial portion of ET-1-induced contraction in the rat thoracic aorta³⁴. While direct measures of IP₃ and DAG from these tissues could reinforce these findings, the inability to separate smooth muscle cell IP₃ and DAG concentrations from that produced in other cell types in the whole tissue hinders the meaningfulness of these experiments. These data, combined with the lack of inhibition of ET-1-induced contraction by IP₃ receptor inhibition, suggest that DAG, and not IP₃, may mediate venous contraction to ET-1.

IP₃ receptor expression and IP₃-mediated contraction occur in both arteries and veins

Smooth muscle from both aorta and vena cava expresses all three IP₃ receptor subtypes. Unfortunately, no comparisons could be reliably drawn between the quantities of IP₃ receptor protein expression in aorta versus vena cava, since aorta have substantially more smooth muscle than vena cava ^{4, 6}. In a different approach, we used immunofluorescent labeling of IP₃ receptor in freshly dissociated smooth muscle cells to compare IP₃ receptor expression in aorta versus vena cava. Both aorta and vena cava smooth muscle contain all 3 IP₃ receptor subtypes, showing that the differences in total IP₃ receptor expression we saw in Western blotting experiments were not indicative of differences in smooth muscle IP₃ receptor expression between aorta and vena cava. Immunocytochemistry allowed us to pinpoint IP₃ receptors to alpha-actin positive cells, which we interpreted as smooth muscle expression. In those experiments, we could locate all three isoforms of the IP₃ receptor in smooth muscle cells from both aorta and vena cava. The Westerns used a homogenate of the aorta and vena cava. Because smooth muscle is a small percentage of those cells expressed in the vena cava, especially relative to the aorta which is predominantly smooth muscular, it was unfair to use the Westerns for IP₃ receptor comparison in smooth muscle. It is fair to state that it is possible that the relative lower expression of IP₃ receptors in the vena cava observed in the Westerns could be reflective of the tissue overall, and IP₃ receptors may not participate in ET-1-induced contraction in the vena cava because of this reason. These experiments were followed with functional experiments. IP₃ receptors appear to be functionally coupled to contraction in both tissues, evidenced by the gradual and sustained contraction caused by the membrane-permeant IP₃ analog, Bt-IP₃ (10μM). It is important to note that, similar to acetoxymethyl ester-linked Ca²⁺ dyes (e.g. Fluo4-AM), Bt-IP₃ is inactive until it undergoes esterase-dependent cleavage inside the cell. As such, development of contraction to Bt-IP₃ is limited by the rate at which this cleavage occurs and is not necessarily representative of the rate at which IP₃ is produced normally via PLC. Taken together, these data are consistent with the idea that IP₃R are expressed in venous and arterial smooth muscle and that IP₃, presumably by activating IP₃ receptors, can cause contraction in vena cava as well as aorta.

The role of IP₃ during ET-1-induced contraction is different in arteries versus veins

Having determined that ET-1-induced contraction was dependent on PLC and that functional IP₃R were present in artery and vein, it was logical to next test the ability of an IP₃R antagonist to block ET-1-induced contraction. The IP₃R antagonist 2-APB (100μM) significantly attenuated ET-1-induced contraction in aorta. However, 2-APB had no significant effect on vena cava contraction to ET-1, suggesting that contraction to ET-1 is not highly dependent on IP₃ receptor activation in vena cava. This experiment points to a significant difference in how ET-1 signals in arteries versus veins. There are, however, limitations to be considered.

We used a concentration of 2-APB that maximally inhibits IP₃ receptors with the fewest possible interactions with other transient receptor potential (TRP) channels expressed in smooth muscle. Several non-specific effects of 2-APB are documented that complicate the interpretation of these results. 2-APB can also act as both an activator and an inhibitor of TRP channels at concentrations similar to those used here ^{24, 25}. However, several other inhibitors of Ca²⁺ channels and TRP channels had no effect on ET-1-induced contraction in either aorta or vena cava (table I). While it cannot be ruled out entirely that the inhibition of ET-1-induced contraction by 2-APB in rat aorta is due to non-specific effects of 2-APB on ion channels other than IP₃ receptors, our findings represent another stark pharmacological difference between aorta and vena cava in terms of ET-1-induced contraction.

DAG reveals differential signalling in arteries versus veins

DAG can both negatively and positively affect cytosolic Ca²⁺ by its actions as an activator of protein kinase C or several different TRP cation channels in the plasma membrane ^{35, 36}. Our experiments did not examine the mechanisms by which DAG regulates venous contraction to ET-1 beyond activation of PKC, but they did investigate the ability of DAG to cause contraction. The DAG analogue OAG caused significant contraction in vena cava but not aorta, a contraction reversed by the PKC inhibitor chelerythrine (10μM) (fig. 7e). Strengthening the idea that PKC was particularly important to venous contraction was the ability of chelerythrine to reduce profoundly ET-1-induced contraction. Chelerythrine is considered a nonselective inhibitor of PKC, and would inhibit several PKC isoforms that are sensitive to DAG activation as well as other non-DAG sensitive isoforms ³⁷. In this way, our findings are internally consistent as it suggests that DAG PKC isoforms may be more important in the vena cava versus aorta, but PKC, in general, is important in mediating ET-1-induced contraction in both tissues. These data illustrate another pharmacological difference between aorta and vena cava. The role of DAG as a positive regulator of agonist-induced contraction in veins is a viable and interesting mechanism in need of further investigation.

Limitations, Conclusions and Clinical Relevance

Limitations to this study should be noted. First, we have used ET-1 as an illustrative agonist and present no other data using a different agonist. Thus, our conclusions have to be circumscribed to ET-1- signalling. Second, we have used one artery and vein pair – the aorta and vena cava- in the rat as models. Large arteries and veins do not have strictly identical physiological functions to smaller arteries and veins. ET contracts smaller arteries and veins in the mesentery not only from the rat but also the mouse ³⁸, suggesting that the present work may apply to arteries and veins generally.

Our findings suggest that, in both aorta and vena cava, ET-1 activates PLC and likely the production of IP₃ and DAG. However, while ET-1-induced contraction in aorta involves IP₃, ET-1-induced contraction in vena cava is instead more dependent upon DAG. Our experimental evidence suggests that ET-1-induced contraction in the vena cava may be largely independent of the actions of IP₃. Furtherermore, pharmacological differences exist between aorta and vena cava, as shown by the differences in OAG-induced contraction and the different effects of U-73122, U-73343, 2-APB and chelerythrine on ET-1-induced contraction. We interpret these pharmacological differences to imply that DAG may be the primary regulator of ET-1-induced contraction in vena cava, perhaps through activation of PKC. These studies outline a new and fundamental difference between venous and arterial smooth muscle, in terms of excitation-contraction coupling and Ca²⁺ mobilization during ET-1-induced contraction, and further reinforce the heterogeneity of vascular smooth muscle. Since changes in venous capacitance are associated with a multitude of medical conditions, including syncope, hemorrhage, shock, heat stroke and congestive heart failure, these findings also present new potential therapeutic targets (specifically DAG interference) specific to veins.

Clinical Relevance.

These studies outline a new and fundamental difference between venous and arterial smooth muscle, in terms of excitation-contraction coupling and Ca²⁺ mobilization during ET-1-induced contraction, and further reinforce the heterogeneity of vascular smooth muscle. Since changes in venous capacitance are associated with a multitude of medical conditions, including syncope, hemorrhage, shock, heat stroke and congestive heart failure, these findings also present new potential therapeutic targets (specifically DAG interference) specific to veins.