Introduction

Top Abstract Introduction Methods Results Discussion References

Endothelial vasomotor dysfunction accompanies a variety of cardiovascular pathologies and risk factors, including hypercholesterolemia, atherosclerosis, hypertension and reperfusion injury (Stewart, 1995). Disruption of nitric oxide-mediated vasodilation (Gilligan et al., 1994; Panza et al., 1995) along with an increase in vascular oxidant stress (Ohara et al., 1993; Jun et al., 1996) have been implicated in endothelial dysfunction. Recent studies in humans have suggested that the clinical benefit of lipid-lowering therapy (Levine et al., 1995) and antioxidants (Anderson et al., 1995) may be mediated in part by their ability to normalize endothelial function in pathological conditions in vivo. These findings are consistent with previous suggestions that oxidized LDL, which was present in atherosclerotic human blood vessels in vivo (Yla-Herttuala et al., 1989) and selectively enhanced coronary vasoconstriction in vitro (Cox and Cohen, 1996a), plays an important role in endothelial vasomotor dysfunction and the clinical manifestations of atherosclerosis (Steinberg et al., 1989; Cox and Cohen, 1996b)

Lysophosphatidylcholine is an inflammatory lipid responsible for many of the proatherogenic effects of oxidized LDL, including inhibition of endothelium-dependent vasodilation, increased leukocyte adhesion to the endothelium and enhanced expression of cell adhesion molecules on endothelial cells (Steinberg et al., 1989). LysoPC itself exerts chemotactic effects on monocytes (Quinn et al., 1987), is mitogenic for vascular smooth muscle cells (Chai et al., 1996), enhances superoxide anion release from neutrophils (Ginsburg et al., 1989) and isolated blood vessels (Ohara et al., 1994) and enhances mast cell secretion (Marquardt and Walker, 1991). Because the concentration of lysoPC was increased several fold in the blood vessel walls of rabbits (Keaney et al., 1995) and monkeys (Portman and Alexander, 1969) with atherosclerosis, this spectrum of proinflammatory effects documented for lysoPC in vitro may also occur in vivo.

The signal transduction pathways involved in lysoPC-induced endothelial vasomotor dysfunction and other cellular effects remain ill-defined. However, many of the actions of lysoPC can be mimicked by processes that increase phosphatidylcholine-specific PLD activity. PLD catalyzes the hydrolysis of phosphatidylcholine to phosphatidic acid, and has been implicated in cellular secretion, mitogenic signaling and neutrophil superoxide production (Thompson et al., 1993; Cockcroft, 1996). Prompted by the similarity in cellular effects associated with PLD activation and the cellular effects of lysoPC, we hypothesized that this signaling pathway may be involved in lysoPC-induced alterations in endothelial cell function. In support of this hypothesis, lysoPC stimulated PLD activity in a time- and concentration-dependent manner in cultured human coronary artery endothelial cells (Cox and Cohen, 1996c). Although lysoPC can stimulate PLD activity, the relationship between increases in PLD activity and endothelial vasomotor dysfunction remains to be established.

The present study was designed to investigate the role of the PLD signal transduction pathway in endothelial vasomotor dysfunction induced by lysoPC and other stimuli. To relate these functional effects in the same preparation, we have used isolated rabbit aortae to compare the effects of lysoPC and other lysophospholipids on vascular PLD activity and endothelial vasomotor function.

	Methods

Top Abstract Introduction Methods Results Discussion References

Measurement of vasomotor activity. Aortae were removed from New Zealand White rabbits (4-6 lbs) with careful attention to retaining the endothelium, cleaned of connective tissue, and cut into ring segments 5 mm in length. Rings were suspended between two stainless steel hooks, one fixed immobile and the other attached to a force transducer, and changes in force were measured in grams with a computerized data acquisition system (MP100, BIOPAC Systems, Inc., Santa Barbara, CA). Tissues were suspended at optimum passive force (5 g) in 10 ml modified Krebs' buffer with the following composition: NaCl (113.2); KCl (4.6); CaCl₂·2H₂O (1.6); KH₂PO₄ (1.2); MgSO₄ (1.2); glucose (10); NaHCO₃ (24.8), continuously oxygenated and maintained at 37°C. After equilibration for 60 min, washing the tissues via buffer replacement every 15 min, a depolarization-induced contraction was induced by KCl (67 mM) to confirm viability of the tissue and provide a reference contraction for normalization. Tissues were returned to base-line force by washing and were re-equilibrated for 30 min. After incubation with test agents as described in the text or figure legends, endothelial vasomotor function was assessed by contracting the tissue with phenylephrine (1 µM) and, on achieving a stable contraction, cumulatively adding increasing concentrations of acetylcholine (1 nM to 10 µM) to induce relaxation.

Assay of PLD activity. PLD activity was assessed in intact tissue by measuring the production of PEt in the presence of ethanol as a specific marker of PLD activity (Hu et al., 1996). Tissue was equilibrated for 30 min at 37°C in HEPES-buffered Krebs' solution of the following composition (mM): NaCl (113.2); KCl (4.6); CaCl₂·2H₂O (1.6); KH₂PO₄ (1.2); MgSO₄ (1.2); glucose (10); NaHCO₃ (24.8), and HEPES (10 mM), bubbled continuously with 5% CO₂ in O₂. Tissues were then transferred to a labeling buffer (7.5 ml/well of 6-well cell culture plate) consisting of Krebs-HEPES + 0.02% bovine serum albumin and [³H]myristic acid (6 µCi/ml). Tissue was incubated in labeling buffer for 3 hr in a 37°C incubator under 5% CO₂, followed by three washes in Krebs-HEPES without bovine serum albumin or label and re-equilibration in the same for 30 min at 37°C. Tissues were then transferred to wells of a 6-well plate (1-2 rings/well) containing Krebs-HEPES supplemented with 1% ethanol and further additions as indicated in the text and figure legends, and incubated as above for 45 min. The reaction was terminated by freezing the tissue in tongs cooled in liquid N₂ and homogenizing in 1 ml choroform/methanol (1:2). Homogenates were centrifuged (1000 × g 10 min), supernatants were transferred to clean tubes and chloroform (1.25 ml) was added, followed by vortexing. After incubation for 10 min at 4°C, phases were split by addition of water (1.25 ml) followed by vortexing and centrifugation (1000 × g 10 min). The upper phase was aspirated and the lower phase was evaporated to dryness under vacuum. Lipid residues were redissolved in 10 µl chloroform along with authentic PEt (75 µg) as a tracer, spotted onto Whatman LK6DF thin-layer chromatography plates, and resolved in a solvent system of chloroform/methanol/acetic acid (65:15:2). Lipid spots were visualized by exposure to iodine vapor. Spots corresponding to PEt were scraped into scintillation vials, extracted for 10 min in 1 ml methanol, and radioactivity determined by liquid scintillation spectroscopy.

Data presentation and statistical analysis. Unless indicated otherwise, data are expressed as the mean ± S.E. for the number of rings shown in parentheses. For all experiments, the data represent tissue from at least four animals. Acetylcholine-induced relaxation is expressed as percent (100% = base-line tone before contraction with phenylephrine, 1 µM). PLD activity is expressed as the amount of [³H] PEt formed normalized as a percentage of the total ³H-labeled phospholipids in each sample. Where appropriate, unpaired Student's t test was used to compare means; P < .05 was considered statistically significant.

Drugs and chemicals. Phenylephrine, acetylcholine chloride, sodium nitroprusside, lysoPC (palmitoyl), lysoPS (palmitoyl) and PDBu were purchased from Sigma Chemical Co. (St. Louis, MO). [³H]Myristic acid (1 mCi/ml) was purchased from NEN (Boston, MA). LysoPAF, GF109203X and PEt were purchased from BIOMOL (Plymouth, MA). Lysophospholipids were dissolved in phosphate-buffered saline and sonicated 3 to 4 min immediately before use. PDBu and GF109203X were dissolved in DMSO, and PEt was dissolved in chloroform. All other compounds were made in deionized water.

	Results

Top Abstract Introduction Methods Results Discussion References

Effect of lysoPC on PLD activity and endothelial function. To study the role of PLD activation in lysoPC-induced endothelial dysfunction, the effects of a range of lysoPC concentrations on vascular PLD activity and endothelium-dependent relaxation were compared. LysoPC concentration-dependently (10-30 µM) stimulated PLD activity in isolated rabbit aorta (fig. 1A). PLD activity in untreated segments was unchanged during the same time (45 min) period ([³H]PEt at 45 min = 100.7% of initial level at t = 0; n = 2). Thus, lysoPC stimulated PLD activity in isolated rabbit blood vessels, similar to its effect in cultured human coronary artery endothelial cells (Cox and Cohen, 1996c).

View larger version (22K): [in this window] [in a new window]   Fig. 1.   Effect of lysoPC on (A) PLD activity and (B) acetylcholine-induced relaxation of rabbit aortic rings. Rabbit aortic rings were incubated for 40 min with lysoPC as indicated and then assayed for PLD activity or endothelial function as described under "Methods." Asterisks denote a significant difference from basal as determined by Students' t test (P < .05). Data represent the mean ± S.E. of the number of rings indicated and represent tissue from at least four animals. CONT, control.

Preincubation of blood vessels with lysoPC caused a progressive loss of endothelium-dependent relaxation evoked by acetylcholine in arteries preconstricted with phenylephrine (1 µM) (fig. 1B). LysoPC had no effect on the magnitude of precontraction induced by 1 µM phenylephrine (% 67 mM KCl contraction = 123.4 ± 2.6, 126.7 ± 5.5, 130.5 ± 5.0, 119.8 ± 6.7 and 112.7 ± 9.5 for control, 3 µM, 10 µM, 20 µM, and 30 µM lysoPC, respectively; n = 4-6). The concentrations of lysoPC required to inhibit endothelium-dependent relaxation and stimulate vascular PLD activity were identical.

Effect of other lysophospholipids on PLD activity and endothelial function. Because other lysolipids in addition to lysoPC can inhibit endothelial vasomotor function (Mangin et al., 1993), the effects of lysoPAF and lysoPS on PLD activity and endothelium-dependent relaxation in isolated blood vessels were compared. LysoPAF and lysoPS both stimulated vascular PLD activity (fig. 2, A and B) and inhibited acetylcholine-induced relaxation (fig. 2, C and D) in isolated blood vessels. The concentrations of both lysolipids required to inhibit endothelium-dependent relaxation were identical with those that stimulated vascular PLD activity. Furthermore, the reduced potency of lysoPS on vascular PLD activity (fig. 2B) relative to lysoPC and lysoPAF was reflected in a reduced inhibition of endothelium-dependent relaxation (fig. 2D). Thus, endothelial dysfunction induced by lysophospholipids was closely associated with their ability to stimulate vascular PLD activity.

View larger version (42K): [in this window] [in a new window]   Fig. 2.   Effect of lysoPAF and lysoPS on PLD activity (A, B) and endothelial vasomotor function (C, D). Rabbit aortic rings were incubated with lysophospholipids as indicated for 40 min, followed by assay of PLD activity or endothelial function as described under "Methods." Asterisks denote a significant difference from basal as determined by Students' t test (P < .05). Data represent the mean ± S.E. of the number of rings indicated and represent tissue from at least four animals. CONT, control.

Effect of lysophospholipids on sodium nitroprusside-induced relaxation. Vasorelaxation induced by sodium nitroprusside, an endothelium-independent vasodilator, was unaffected by these lysophospholipids in the highest concentration studied (30 µM) (fig. 3). These data indicate that the inhibition of acetylcholine-induced vasorelaxation observed with the lysophospholipids described above was mediated via an effect on endothelial function and was not a nonspecific effect on the ability of the blood vessel to relax.

View larger version (23K): [in this window] [in a new window]   Fig. 3.   Effect of lysophospholipids on sodium nitroprusside-induced relaxation. Rabbit aortic rings were incubated 40 min with each lysophospholipid (30 µM) and contracted with phenylephrine (1 µM), followed by cumulative addition of sodium nitroprusside as shown. Data represent the mean ± S.E. of the number of rings indicated in parentheses and represent tissue from at least 4 animals. CONT, control.

Role of vascular PKC activity in lysoPC-induced endothelial dysfunction. Because some studies have implicated PKC activation in the mechanism of lysoPC-induced endothelial vasomotor dysfunction (Ohgushi et al., 1993), the ability of GF109203X, a highly selective inhibitor of PKC (Toullec et al., 1991), to prevent lysoPC-induced inhibition of endothelium-dependent relaxation was tested. GF109203X (3 µM), at a concentration that maximally inhibited PKC activity in isolated enzyme preparations (Toullec et al., 1991), inhibited PDBu-induced aortic contraction by 91.2 ± 2.4% (fig. 4). These data confirm that GF109203X (3 µM) effectively inhibited PKC activity in intact rabbit aorta. In contrast, preincubation of rabbit aorta with GF109203X (3 µM) did not affect the ability of lysoPC (20 µM) to inhibit acetylcholine-induced relaxation (fig. 5). These data suggest that lysoPC inhibited endothelium-dependent relaxation in rabbit aorta via a mechanism independent of PKC, consistent with a possible involvement of PLD activation.

View larger version (16K): [in this window] [in a new window]   Fig. 4.   Effect of GF109203X on PDBu-induced contraction. Original force recordings of rabbit aortic rings incubated 30 min with GF109203X (3 µM) or an equivalent volume of DMSO (0.3%), followed by addition of PDBu (1 µM). Summary data are shown in the inset and represent the mean ± S.E. of the number of rings indicated in parentheses and represent tissue from at least four animals. CONT, control

View larger version (21K): [in this window] [in a new window]   Fig. 5.   Effect of GF109203X on lysoPC-induced inhibition of acetylcholine-mediated relaxation. Rabbit aortic rings were incubated 30 min with GF109203X (3 µM) or an equivalent volume of DMSO (0.3%), followed by addition of lysoPC (20 µM) for an additional 40 min. After wash-out, phenylephrine (1 µM) was added to induce contraction, followed by cumulative addition of acetylcholine (1 nM to 3 µM). Data represent the mean ± S.E. of the number of rings indicated in parentheses and represent tissue from at least four animals.

Effect of orthovanadate on PLD activity and endothelial function. To determine whether the association between PLD activation and endothelial vasomotor dysfunction was a phenomenon particular to lipid mediators or was a more general feature of interventions that stimulate PLD activity, the effects of orthovanadate on vascular PLD activity and endothelium-dependent relaxation were compared. Sodium orthovanadate is a tyrosine phosphatase inhibitor that stimulated PLD activity in many systems by increasing the level of cellular tyrosine phosphorylation (Natarajan et al., 1996). Sodium orthovanadate (0.1-3 mM) stimulated PLD activity (fig. 6, circles) and inhibited acetylcholine-induced relaxation (fig. 6, triangles) in a parallel manner. Thus, stimulation of vascular PLD activity with a mediator unrelated to lysophospholipids was also closely associated with endothelial dysfunction in isolated blood vessels.

View larger version (19K): [in this window] [in a new window]   Fig. 6.   Effect of orthovanadate on vascular PLD activity and endothelial-dependent relaxation. Rabbit aortic rings were incubated 60 min with orthovanadate as indicated and contracted with phenylephrine (1 µM for control and 0.1 mM orthovanadate; 0.8 µM for 1 and 3 mM orthovanadate), followed by cumulative addition of acetylcholine (1 nM to 3 µM). Relaxation is expressed as percent relaxation of phenylephrine contraction attained at 1 µM acetylcholine. Data represent the mean ± S.E. of the number of rings indicated in parentheses and represent tissue from three to four animals.

Correlation between PLD activity and endothelial dysfunction. To examine the quantitative relationship between PLD stimulation and endothelial dysfunction, the extent of inhibition of acetylcholine (1 µM)-induced relaxation was plotted versus the vascular PLD activity determined for each concentration of lysophospholipid or orthovanadate tested in this study (fig. 7). Although these measures were derived in separate assays under similar but not identical conditions, this analysis revealed a positive correlation (r² = 0.88) between increases in vascular PLD activity and the inhibition of endothelium-dependent relaxation in isolated blood vessels. These data indicate that stimulation of vascular PLD activity was quantitatively associated with endothelial vasomotor dysfunction in isolated blood vessels and that low activity of vascular PLD was necessary for optimal vasomotor functioning of the endothelium.

View larger version (26K): [in this window] [in a new window]   Fig. 7.   Correlation between vascular PLD activity and endothelial vasomotor dysfunction. Each data point represents the inhibition of maximal acetylcholine (1 µM)-induced relaxation of phenylephrine-contracted tissue (%) plotted versus the corresponding level of PLD activity (% of basal) resulting from various concentrations of lysophospholipid or orthovanadate. The data were fit with the linear equation y = mx + b (m = 0.25; b = 0.23; r2 = 0.88).

	Discussion

Top Abstract Introduction Methods Results Discussion References

The ability of lysoPC, either independently or as a component of oxidized LDL, to inhibit endothelial-dependent vasorelaxation is well established (Cowan and Steffen, 1995; Freeman et al., 1996). The effect of lysoPC to impair endothelium-dependent relaxation is generalized to a variety of endothelium-dependent vasodilators, including acetylcholine (Kugiyama et al., 1990), 5-hyroxytryptamine (Cox and Cohen, 1996a), thrombin (Murohara et al., 1994) and calcium ionophore A23187 (Mangin et al., 1993). However, the cellular pathways affected by lysoPC that ultimately result in endothelial vasomotor dysfunction remain unclear. LysoPC was recently documented to stimulate PLD activity in cultured human endothelial cells (Cox and Cohen, 1996c), although the role of this effect in the vasomotor actions of lysoPC was not addressed. The present study has demonstrated the ability of lysoPC to stimulate vascular PLD activity in isolated blood vessels and has documented a close association between the ability of lysoPC and other mediators to stimulate vascular PLD activity and inhibit endothelium-dependent vasorelaxation. This is the first study to directly correlate changes in the activity of a signal transduction pathway with the development of endothelial vasomotor dysfunction in isolated blood vessels.

Activation of PKC has been implicated in some of the cellular effects of lysoPC. Indeed, lysoPC stimulated PKC purified from porcine brain (Oishi et al., 1988) and in cultured endothelial cells (Ohgushi et al., 1993), and a role for PKC in lysoPC-induced endothelial vasomotor dysfunction was suggested in both isolated porcine coronary arteries and rabbit aortae (Ohgushi et al., 1993). However, more recent studies failed to confirm the ability of PKC inhibitors or the down-regulation of PKC to affect lysoPC-induced endothelial dysfunction (Cowan and Steffen, 1995). In part, these previous studies used early inhibitors of PKC, such as staurosporine and calphostin C, that are known to exert effects unrelated to PKC inhibition (Kageyama et al., 1991; Hartzell and Rinderknecht, 1996). For these reasons, we tested the ability of a highly selective PKC inhibitor, GF109203X (Toullec et al., 1991), to inhibit vascular PKC activity and to affect lysoPC-induced endothelial vasomotor dysfunction in isolated rabbit aortae. The ability of GF109203X (3 µM) to block PDBu (1 µM)-induced contraction while having no effect on lysoPC-induced inhibition of acetylcholine-mediated relaxation argued against a role for PKC in this effect, and it was consistent with previous reports dissociating PKC from lysoPC-induced endothelial vasomotor dysfunction (Cowan and Steffen, 1995; Freeman et al., 1996).

In contrast to the lack of association with PKC activation, inhibition of endothelial vasomotor function by lysoPC and other lysophospholipids was closely associated with a stimulation of vascular PLD activity. This association was not restricted to lysophospholipids, because stimulation of vascular PLD activity by orthovanadate was also associated with an inhibition of endothelial-dependent relaxation. Indeed, if maintenance of low vascular PLD activity is a prerequisite for normal endothelial function, then we reasoned that any intervention that stimulates PLD activity should induce endothelial vasomotor dysfunction. In a search of the literature for interventions that increase PLD activity, we discovered that many proinflammatory agents that stimulated PLD activity in cultured cells also induced endothelial vasomotor dysfunction in isolated blood vessels (table 1). Although these data represent independent studies with different experimental systems, there is a remarkably excellent quantitative correlation in the concentrations required to observe each effect. Thus, we propose that stimulation of PLD activity may be a common mechanism by which proatherogenic and proinflammatory agents mediate deleterious effects on vascular endothelial function.

The mechanism by which the PLD signaling pathway may be associated with endothelial dysfunction remains unclear. In leukocytes, activation of PLD and the subsequent generation of phosphatidic acid was closely associated with stimulation of the NADPH oxidase complex and production of superoxide anion (Bonser et al., 1989). Vascular superoxide production was enhanced in hypercholesterolemia (Ohara et al., 1993), atherosclerosis (Keaney et al., 1995) and certain forms of hypertension (Jun et al., 1996), and is thought to be involved in endothelial vasomotor dysfunction via rapid chemical inactivation of nitric oxide (Gryglewski et al., 1986). In addition, lysoPC stimulated superoxide production in both neutrophils (Ginsburg et al., 1989) and isolated blood vessels (Ohara et al., 1994), and reactive oxygen species were implicated in the cellular effects mediated by lysoPC and oxidized LDL (Stiko et al., 1996). In light of recent studies suggesting that the source of vascular superoxide anion production is an NADPH/NADH oxidase system (Pagano et al., 1995), perhaps similar to the oxidase system in leukocytes, it is tempting to speculate that stimulation of vascular PLD activity by lysoPC and possibly other inflammatory lysophospholipids may be an important part of the signaling mechanism by which vascular oxidative stress is increased and endothelial vasomotor function is attenuated in these pathologies.

Although the correlation between vascular PLD activity and endothelial vasomotor dysfunction (r² = 0.88) was compelling, proof that PLD activation is a necessary component of endothelial dysfunction awaits the development of selective PLD inhibitors or genetically altered animal strains lacking PLD activity in vivo. Furthermore, at least two isoforms of PLD have been distinguished biochemically (Morris et al., 1996), and additional studies are required to establish whether one or both of these isoforms is responsible for the activity present in isolated blood vessels. However, the recent cloning and expression of the first mammalian PLD gene from a human cell line (hPLD1) (Hammond et al., 1995) should hasten the identification of selective PLD enzyme inhibitors, as well as provide the molecular tools necessary to develop animal strains lacking one or more isoforms of PLD. Nevertheless, the direct correlation between vascular PLD activity and inhibition of endothelial vasomotor function documented in the present study, along with data from the literature suggesting a coupling between these two events in response to many diverse interventions, strongly support a role for this signal transduction pathway in the etiology of endothelial vasomotor dysfunction.