Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

The endothelium plays an important role in the regulation of vascular tone (Rubanyi, 1993). It synthesizes and releases EDRF, prostacyclin and EDHF, as well as EDCFs such as endothelin and prostaglandin H₂/thromboxane A₂. Of these factors, EDRF has been suggested to be the most important. EDRF is synthesized by endothelium using L-arginine as a substrate and causes vascular relaxation through activation of guanylyl cyclase within the smooth muscle cell. Experimental and clinical data have shown that endothelial function is impaired in hypercholesterolemia and atherosclerosis (Creager et al., 1992; Kolodgie et al., 1990; Lopez et al., 1989). Endothelial dysfunction may be an early marker of atherosclerosis and may occur early in the atherosclerotic process (Zeiher et al., 1991).

Elevated LDL is closely related to the development of atherosclerosis. Recent evidence suggests that ox-LDL is responsible for the atherosclerotic process (Steinberg and Witztum, 1990). ox-LDL has been demonstrated to impair endothelium-dependent vasorelaxation (Galle et al., 1994; Mangin Jr et al., 1993; Plane et al., 1992; Tanner et al., 1991) and to potentiate agonist-induced vasoconstriction (Cox and Cohen, 1996; Galle et al., 1990; Simon et al., 1990), which may contribute to the alteration of vascular reactivity associated with hypercholesterolemia and atherosclerosis.

The mechanisms responsible for endothelial dysfunction induced by ox-LDL are probably multifactorial, including a decrease in synthesis and/or release of EDRF or an increase in the inactivation of EDRF. One possible mechanism for reduced synthesis of EDRF is substrate (L-arginine) deficiency (Tanner et al., 1991). Whether reduced L-arginine availability contributes to impaired endothelial function is controversial. Some studies have shown that impaired endothelial function is improved by an excess supply of exogenous L-arginine (Böger et al., 1995; Creager et al., 1992; Girerd et al., 1990; Tanner et al., 1991). In others however, L-arginine did not improve endothelial dysfunction (Casino et al., 1994; Hayashi et al., 1995; Kugiyama et al., 1990; Pohl et al., 1995).

Although several studies have examined the vascular effects of ox-LDL in animal vessels, no studies have evaluated the direct actions of ox-LDL in human vessels. The present study determined the vasomotor effects of ox-LDL in human saphenous veins and examined whether L-arginine deficiency was involved in the impairment of endothelial function induced by ox-LDL. Because the L-arginine-EDRF pathway exists in the endothelial cells of human saphenous vein (Lüscher et al., 1988; Yang et al., 1991), this blood vessel represents a model with which to examine the vasoactive effects of ox-LDL in humans. Furthermore, because the human saphenous vein is used for coronary artery bypass, and bypass grafts are exposed to the actions of ox-LDL, it was of interest to examine the effects of ox-LDL on this vessel.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Drugs. Acetylcholine chloride, 5-HT, sodium nitroprusside, L-arginine, L-phenylephrine and indomethacin were purchased from Sigma Chemical Company (St. Louis, MO). Serotonin first was dissolved in a 1:1 solution of 1 M HCl and 1 M NaOH to make a 10⁴ M solution and then was diluted in deionized water. Indomethacin was dissolved in 1% Na₂CO₃ to make a 0.06 M stock solution. Other drugs were dissolved in deionized water.

Preparation of blood vessel. All studies were approved by the appropriate institutional review boards. Human saphenous veins were obtained from white males undergoing coronary bypass surgery from University Hospital and the Medical College of Georgia (Augusta, GA). After the vein was removed, it was immediately placed in cold Krebs buffer, previously aerated with 95% O₂-5% CO₂, and transported to our lab. Composition of the Krebs buffer solution was as follows (mM): NaCl 118.0, KCl 4.7, NaHCO₃ 25.0, MgSO₄ 1.2, KH₂PO₄ 1.1, CaCl₂ 2.5, EDTA 0.01, glucose 11.0. The final pH of the Krebs' buffer was 7.4. The vessel was carefully cleaned of fat and connective tissue and cut into 3 to 5-mm rings. Each ring was suspended between stainless steel hooks in a 10-ml tissue bath containing Krebs buffer maintained at 37°C and continuously aerated with 95% O₂-5% CO₂. One hook was connected to a force transducer for recording the isometric tension. The rings were progressively stretched to the optimal resting tension (2 g) and were allowed to equilibrate for 45 min. The buffer was changed every 20 min. All rings were initially contracted twice with KCl (70 mM) to confirm the viability of vessels before the experimental protocols were performed.

Preparation of ox-LDL. Human LDL (5 mg/ml) was purchased from PerImmune Inc. (Rockville, MD) in NaCl (0.5 M) with 0.15% EDTA (pH 7.2) and stored at 4°C. Native LDL was oxidatively modified according to a modification of the method of Cox and Cohen (1996). Before oxidation, LDL (1-2 ml) was dialyzed at 4°C for 24 h against phosphate-buffered saline (PBS) to remove EDTA. Composition of the PBS was NaCl 137 mM, NaH₂PO₄ · H₂O 10 mM, NaOH 7 mM, pH 7.2. Then LDL (5 mg/ml) was oxidized by incubating it with CuSO₄ (5 µM) for 24 h at 37°C, followed by dialysis at 4°C for 24 h against PBS containing 0.01% EDTA to remove the copper ion. The lipid peroxide content of the copper-oxidized LDL was measured fluorometrically as thiobarbituric acid-reactive substances (Yagi, 1976). Lipid peroxidation was expressed as nanomoles of MDA (malonaldehyde) per milligram of LDL protein, using 1,1,3,3-tetraethoxypropane as the standard. The lipid peroxidation values of the LDL before and after incubation with CuSO₄ (5 µM) for 24 h were 0 and 15.03 ± 3.47 nmol MDA/mg protein (P < 0.01), respectively. The ox-LDL sample was stored at 4°C.

Experimental protocols. After the vessel rings were contracted twice with KCl to confirm the viability of vessels, the presence of functional endothelium was determined. First, vessel rings were preconstricted with phenylephrine (10⁵ M). After achievement of a stable contraction, a concentration-response curve to ACh (10⁹ to 10⁴ M) was constructed to check for the presence of functional endothelium. In human saphenous veins, EDRF production is less than that in arteries (Lüscher et al., 1988; Yang et al., 1991). The maximal value of ACh-induced endothelium-dependent relaxation at 10⁵ M in human saphenous veins is about 20% in the absence of indomethacin and about 40% in the presence of indomethacin (Lüscher et al., 1988; Yang et al., 1991). In the vessel rings without endothelium, ACh does not produce relaxation. Because the saphenous veins were from patients undergoing coronary artery bypass, mechanical damage to endothelium may occur during surgery. In the present study, vessel rings with a relaxation response to ACh (10⁶ M) of 15% or more were considered to have functional endothelium and were chosen to determine the effect of ox-LDL on ACh-induced relaxation. The first relaxation response to ACh was used as control. After washout and return to base line, vessel rings were incubated with ox-LDL (100 µg protein/ml) or ox-LDL plus L-arginine (500 µM) for 45 min. Then the vessel rings were preconstricted with phenylephrine and exposed to cumulative concentrations of ACh as before.

9

4

9

4

9

4

Data analysis. Data are expressed as the mean ± S.E.M. 5-HT-induced contractions were expressed as a percentage of the maximal contraction evoked by KCl (70 mM). Relaxation responses were calculated as a percent decrease in tension of the stable contraction produced by phenylephrine. EC₅₀ was defined as the concentration of an agonist at which 50% of the maximal response was obtained. Student's t test was used to test the significance of the differences between the two groups. Comparisons among more than two groups were performed using one-way ANOVA. When differences between groups were indicated, a Newman-Keuls' post-hoc comparison was used. The differences were considered to be significant when P < 0.05.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Effect of ox-LDL on ACh-induced endothelium-dependent relaxation. Phenylephrine was used to preconstrict human saphenous veins. Preconstriction before and after ox-LDL treatment was not significantly different (1.7 ± 0.2 vs. 1.4 ± 0.1 g, control vs. ox-LDL group). ACh produced concentration-dependent relaxation in human saphenous vein with endothelium that was significantly attenuated by ox-LDL pretreatment at concentrations between 10⁸ and 10⁴ M (fig. 1). In preliminary studies, consecutive relaxation responses to ACh were elicited to ensure that ACh-induced relaxation did not decrease with time. ox-LDL alone produced either no or very weak contractions. Only 5 of 19 patients showed minimal contraction of 15 ± 3% of KCl (70 mM) contraction. Endothelium-independent relaxation to sodium nitroprusside was not significantly affected by preincubation with ox-LDL (fig. 2).

View larger version (19K): [in this window] [in a new window]   Fig. 1.   Effect of ox-LDL (100 µg protein/ml) on the ACh-induced endothelium-dependent relaxation of human saphenous vein. The rings were preconstricted with 10 µM phenylephrine. Values represent the mean ± S.E.M. of vessels from nine patients. * P < 0.05, ** P < 0.01, *** P < 0.001.

View larger version (18K): [in this window] [in a new window]   Fig. 2.   Effect of ox-LDL (100 µg protein/ml) on the sodium nitroprusside-induced endothelium-dependent relaxation of human saphenous vein. The rings were preconstricted with 10 µM phenylephrine. Values represent the mean ± S.E.M. of vessels from four patients.

Effect of ox-LDL on 5-HT-induced vasoconstriction. ox-LDL preincubation increased the vascular contractile response to 5-HT in the presence of endothelium. The differences were significant at concentrations between 3 × 10⁶ and 10⁴ M (fig. 3). The maximal contractile response to 5-HT was significantly increased from 95 ± 9% of KCl (70 mM) contraction in the control group to 124 ± 12% of KCl contraction in the ox-LDL group (table 1). However, there was no significant difference between the EC₅₀ values (table 1). In the absence of endothelium, ox-LDL did not produce a significant effect on 5-HT-induced contraction (data not shown). This suggests that ox-LDL may interfere with endothelial function and thus increase the vascular reactivity to 5-HT.

View larger version (20K): [in this window] [in a new window]   Fig. 3.   Effect of ox-LDL (100 µg protein/ml) on the contractile response to 5-HT in human saphenous vein with endothelium. Values represent the mean ± S.E.M. of vessels from 10 patients. * P < 0.05, ** P < 0.01, *** P < 0.001.

View larger version (20K): [in this window] [in a new window]   Fig. 4.   The contractile responses to 5-HT in human saphenous vein both with and without endothelium. Values represent the mean ± S.E.M. of vessels from 11 patients. * P < 0.05, ** P < 0.01, *** P < 0.001.

Effect of ox-LDL plus L-arginine on ACh-induced endothelium-dependent relaxation. Phenylephrine-induced preconstriction before and after ox-LDL or ox-LDL plus L-arginine treatment was not significantly different (1.7 ± 0.2 g in control, 1.4 ± 0.1 g in the ox-LDL group and 1.2 ± 0.1 g in the ox-LDL plus L-arginine group). ox-LDL significantly impaired the ACh-induced relaxation between the concentrations of 3 × 10⁸ and 10⁴ M. When the vein was incubated with ox-LDL plus L-arginine, endothelium-dependent relaxation was still impaired (fig. 5). L-arginine did not improve the impairment of ACh-induced relaxation produced by ox-LDL. There were no significant differences in the impairment of ACh-induced endothelium-dependent relaxation between ox-LDL-treated vessels and ox-LDL plus L-arginine-treated vessels. L-arginine alone had no effect on ACh-induced endothelium-dependent relaxation.

View larger version (23K): [in this window] [in a new window]   Fig. 5.   Effect of ox-LDL (100 µg protein/ml) or ox-LDL plus L-arginine (500 µM) on the ACh-induced endothelium-dependent relaxation of human saphenous vein. The rings were preconstricted with 10 µM phenylephrine. Values represent the mean ± S.E.M. of vessels from 5 to 9 patients. * P < 0.05, ** P < 0.01 between control and ox-LDL group;  P < 0.05,  P < 0.01 between control and ox-LDL plus L-arginine group.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

This is the first study to evaluate the effects of ox-LDL on endothelium-dependent relaxation and 5-HT-induced vasoconstriction in human saphenous veins. ox-LDL inhibited ACh-induced endothelium-dependent relaxation but did not affect endothelium-dependent relaxation to the NO donor sodium nitroprusside. L-arginine did not prevent the impairment of endothelium-dependent relaxation induced by ox-LDL. ox-LDL also enhanced the vasoconstrictor responses to 5-HT, which were endothelium-dependent.

Hypercholesterolemia, particularly elevated plasma LDL levels, represents one of the most important risk factors for the development of atherosclerosis. ox-LDL is thought to be the atherogenic form of LDL (Steinberg and Witztum, 1990). LDL must first undergo oxidative modification in order to be taken up by macrophages to generate foam cells. ox-LDL is a potent chemoattractant for circulating monocytes and an inhibitor of the motility of resident macrophages, and it is potentially cytotoxic for endothelial cells. It has also been shown to impair endothelium-dependent relaxation (Galle et al., 1994; Mangin Jr. et al., 1993; Plane et al., 1992; Tanner et al., 1991) and to potentiate agonist-induced vasoconstriction in animal blood vessels (Cox and Cohen, 1996; Galle et al., 1990; Simon et al., 1990). Recent investigations have demonstrated that oxidation of native LDL occurs in vivo, and ox-LDL has been identified in human atherosclerotic arteries (Ylä-Herttuala et al., 1989). Furthermore, antioxidant administration improves endothelial function and retards the progression of atherosclerosis (Anderson et al., 1995; Guarnieri et al., 1996; Keaney et al., 1993). And in addition, single LDL apheresis improved impaired endothelial function in hypercholesterolemic humans, which correlated with the decrease in plasma ox-LDL levels (Tamai et al., 1997). Collectively, these studies support the interpretation that ox-LDL is an important atherogenic factor.

In the present study, ox-LDL enhanced the contractile response to 5-HT, which was endothelium-dependent. Our results are consistent with the study of Cox and Cohen (1996) in pig coronary artery and suggest that ox-LDL may induce endothelial dysfunction to produce its vasomotor effect. We also examined the influence of endothelium on 5-HT-induced contraction. Endothelial removal significantly increased the vascular contractile response to 5-HT, an observation that is in agreement with the results reported by Valentin et al. (1996) in rabbit saphenous vein. Thus the effect of ox-LDL on the vascular contractile response to 5-HT mimicked the response elicited by endothelial removal. Several studies have shown that ox-LDL inhibits 5-HT-induced EDRF-mediated relaxation in pig coronary artery (Cox and Cohen, 1996; Simon et al., 1990; Tanner et al., 1991). Data concerning human vessels have not been published. Golino et al. (1991) demonstrated that 5-HT had a vasodilating effect on normal human coronary arteries in vivo. However, the normal vasodilator response to 5-HT was reversed to vasoconstriction in atherosclerotic coronary arteries. This suggests that the enhanced vasoconstriction in response to 5-HT in human coronary artery may be associated with endothelial dysfunction. Therefore, it is possible that ox-LDL increases the contractile responsiveness to 5-HT in human saphenous vein by producing a functional impairment in 5-HT-induced endothelium-dependent relaxation. Another possibility is that ox-LDL also interferes with the basal release of EDRF and thus increases the contractile response to 5-HT. However, there is very little EDRF release under basal conditions in human saphenous vein (Yang et al., 1991). This mechanism does not seem to be important, an interpretation supported by our finding in this study that ox-LDL alone causes no or minimal vasoconstriction.

Studies performed in animal vessels have shown the inhibitory effect of ox-LDL on EDRF-mediated relaxation produced by various agonists (Galle et al., 1994; Mangin Jr. et al., 1993; Plane et al., 1992; Tanner et al., 1991). In the present study, ox-LDL also inhibited the ACh-induced relaxation in human saphenous vein. Lüscher et al. (1988) and Yang et al. (1991) indicated that ACh-induced endothelium-dependent relaxation in human saphenous vein was mediated by EDRF. Thus ox-LDL could conceivably interact with and inhibit the EDRF pathway.

Oxidation of LDL has been demonstrated to occur in vivo (Ylä-Herttuala et al., 1989). Serum LDL levels are about 1 mg/ml and 2 mg/ml in normal and hypercholesterolemic humans, respectively (Casino et al., 1994; Creager et al., 1992). About 5% of plasma LDL from monkeys and humans is modified (Cazzolato et al., 1991; Hodis et al., 1994). Thus 100 µg protein/ml (the concentration employed in the present study) of ox-LDL represents a pathophysiologically relevant concentration that may occur in human atherosclerotic lesions. The concentrations of ox-LDL in atherosclerotic lesions may be even higher. If the vasoactive effects of ox-LDL in vitro occur in coronary artery in vivo, then ox-LDL may contribute to the clinical events observed in atherosclerosis. Additionally, because saphenous veins are often used for coronary artery bypass grafts, decreased function of saphenous vein grafts associated with the development of atherosclerosis may due to the vasoactive effects of ox-LDL.

The mechanisms responsible for the impaired endothelium-dependent relaxation are not yet clear. ox-LDL may interfere with multiple steps and hence have many effects, including decreased substrate (L-arginine) availability for EDRF formation (Tanner et al., 1991), altered transmembrane signaling transduction (Ohgushi et al., 1993), decreased expression of NO synthase (Liao et al., 1995), increased production of endothelium-derived contracting factors (Boulanger et al., 1992), increased degradation or inactivation of EDRF (Mügge et al., 1991; Ohara et al., 1993) and decreased response of vascular smooth muscle to EDRF (Schmidt et al., 1991).

A possible mechanism for reduced synthesis of EDRF is decreased availability of L-arginine, the natural substrate for EDRF synthesis. Impaired endothelium-dependent vasorelaxation can be improved by intravascular infusion or diet supplementation of L-arginine in hypercholesterolemic animals and humans (Böger et al., 1995; Creager et al., 1992; Girerd et al., 1990; Tanner et al., 1991), which suggests that endothelial dysfunction in hypercholesterolemia may be caused by a reduction in intracellular L-arginine availability or metabolism. ox-LDL may interfere with receptor-operated release of L-arginine from intracellular stores or with the synthesis of L-arginine (Tanner et al., 1991). However, other studies do not support the interpretation that L-arginine deficiency is responsible for the impaired endothelial function (Casino et al., 1994; Hayashi et al., 1995; Kugiyama et al., 1990; Pohl et al., 1995). Because administration of exogenous L-arginine does not affect endothelium-dependent relaxation in normal vessels, the amount of endogenous L-arginine appears to be sufficient for EDRF formation and is not a rate-limiting factor for EDRF synthesis in normal vessels (Creager et al., 1992; Girerd et al., 1990). Most authors found no difference in L-arginine concentrations between normal and hypercholesterolemic animals and humans (Bode-Böger et al., 1996; Hayashi et al., 1995; Pasini et al., 1992), although decreased plasma L-arginine levels in hypercholesterolemia have also been reported (Jeserich et al., 1992). Moreover, Minor et al. (1990) has suggested that NO synthesis is not impaired, but rather is actually increased, in diet-induced hypercholesterolemic and atherosclerotic rabbits.

If ox-LDL-induced impairment of endothelial function involves L-arginine deficiency, then administration of L-arginine would reverse the impairment. However, the results of the present study do not support this hypothesis. L-Arginine pretreatment did not prevent ox-LDL-induced impairment of endothelium-dependent relaxation. The present study evaluated short-term effects of ox-LDL in vitro, which did not fully reflect the condition in hypercholesterolemia. Recent investigations suggested that the plasma concentration of ADMA, an endogenous inhibitor of NO synthase (Vallance et al., 1992), is increased in hypercholesterolemic animals, resulting in decreased NO formation (Yu et al., 1994). The increased L-arginine/ADMA ratio produced by exogenous L-arginine supplementation would competitively overcome the inhibition of NO synthase produced by the increased ADMA level and enhance NO production in hypercholesterolemic rabbits (Bode-Böger et al., 1996). The different levels of ADMA may explain the various effects of L-arginine administration on impaired endothelium-dependent relaxation in hypercholesterolemia.

The inhibitory effect of ox-LDL on endothelium-dependent relaxation may involve alterations in muscarinic receptors and/or in the receptor signal transduction pathway. This effect seems concentration-dependent. In rabbit aorta, low concentrations of ox-LDL (50 µg protein/ml) selectively inhibited endothelium-dependent relaxation evoked by the receptor-dependent agonist ACh. However, high concentrations of ox-LDL (>50 µg protein/ml) also inhibited endothelium-dependent relaxation evoked by the receptor-dependent agonist A23187 (Kugiyama et al., 1990). The effect of ox-LDL on NO synthase appears to be both time- and concentration-dependent (Hirata et al., 1995; Liao et al., 1995). Galle et al. (1991) showed that EDRF formation was not attenuated by a 1-h incubation with the potentially cytotoxic ox-LDL (1 mg protein/ml). ox-LDL seems less likely to influence NO synthase in our study because of the short incubation with ox-LDL. ox-LDL (100 µg protein/ml) had no effect on endothelium-independent relaxation in response to sodium nitroprusside in our study, so the capacity of vascular smooth muscle to relax in response to EDRF was unaffected by ox-LDL. Another possible explanation for the impaired endothelium-dependent relaxation is related to the release of EDCF, which could oppose the effects of EDRF. Because our buffer contained indomethacin, we can exclude the influence of vasoconstrictor prostanoids. Whether ox-LDL increases endothelin release is controversial (Boulanger et al., 1992; He et al., 1996). He et al. (1996) showed that different degrees of oxidation of LDL had different effects on endothelin production. Extensively oxidized LDL (24-h exposure to copper) inhibited endothelin secretion from cultured endothelial cells. In our study, LDL is oxidized through a 24-h exposure to CuSO₄. Thus, increased endothelin release is unlikely to contribute to the vasomotor effects of ox-LDL.

Endothelium-dependent relaxation may be mediated by both EDRF and EDHF. Recently, Shoemaker et al. (1996) reported that EDHF was involved in ACh-induced relaxation in human saphenous vein. Therefore, the possibility that ox-LDL also inhibits EDHF-mediated relaxation cannot be totally excluded in our study. Another mechanism that cannot be excluded is the increased inactivation of EDRF by ox-LDL or superoxide anion. Hypercholesterolemia and ox-LDL have been reported to increase superoxide production (Maeba et al., 1995; Ohara et al., 1993). ox-LDL may also directly inactivate EDRF after its release from endothelium (Chin et al., 1992; Galle et al., 1991).

In summary, ox-LDL inhibited endothelium-dependent relaxation and increased endothelium-dependent contraction evoked by 5-HT in human vessels. ox-LDL-induced impairment of endothelium-dependent relaxation, at least in human saphenous vein, is probably not due to a deficiency of L-arginine for EDRF formation. Because ox-LDL does produce endothelial dysfunction in human saphenous vein, it may influence the consequences of coronary bypass surgery and saphenous vein graft function. These results also indicate that human saphenous vein can be used as a model to study the vasomotor effects of ox-LDL.