Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

It is now well established that elevated plasma contents of LDL are a major risk factor for accelerated atherogenesis and subsequent cardiovascular problems such as coronary heart disease (Castilli and Anderson, 1986; Flavahan, 1992; Cox and Cohen, 1996a). Several lines of evidence indicate that o-LDL is an important causative factor for this effect (Steinberg et al., 1989). o-LDL has been demonstrated to accumulate in atherosclerotic lesions of arteries (Yla-Herttuala et al., 1989). Various components of the vascular wall, including endothelial cells and macrophages, are involved in the oxidative modification of LDL (Henriksen et al., 1981). In a cell-free system, the oxidation of LDL has been shown to be mimicked in solutions containing certain metallic ions such as Cu⁺⁺ (Liu et al., 1994). Functionally, o-LDL has been demonstrated to cause potentiation of vascular smooth muscle contraction and impairment of endothelium-dependent relaxation (Simon et al., 1990; Kugiyama et al., 1990; Cox and Cohen, 1996b). These effects are similar to the impaired functional responsiveness of atherosclerotic arteries (Kugiyama et al., 1990).

The purine nucleoside, adenosine, plays an important role in the metabolic regulation of coronary blood flow (Berne, 1980). It is a powerful coronary vasodilator, the formation of which is sensitive to changes in tissue metabolic state. The vasodilatory effect of adenosine has been shown to be mediated via extracellular A₂ receptors located both on the smooth muscle and endothelium of the coronary artery (DesRosiers and Nees, 1987; Olsson and Pearson, 1990; Abebe et al., 1994; Schiele and Schwabe, 1994). On the basis of the responsiveness to different adenosine analogs, the coronary adenosine A₂ receptors appear to be of different subtypes, at least in some species. In this regard, we recently observed that in porcine coronary artery while the vasorelaxant effects of the 5-uronamide adenosine analogs, NECA and CGS-21680, were partially endothelium-dependent, stimulating both smooth muscle and endothelial A₂ receptors, those of the C-2 substituted analogs, including CAD, were endothelium-independent, indicating activation of solely smooth muscle cell A₂ receptors (Abebe et al., 1994, 1995). The endothelium-dependent relaxations evoked by NECA and CGS-21680 have been found to involve the release of nitric oxide (Abebe et al., 1995).

The influence of atherosclerosis on the vascular action of adenosine has been assessed by three groups of investigators. Reports by Toda et al. (1988) indicate that atherosclerosis attenuates the relaxations of monkey mesenteric arteries to adenosine. However, Sellke et al. (1990) and Lopez et al. (1989) have reported that vasorelaxant responses of coronary, femoral and iliac arteries from monkeys to adenosine are not altered by atherosclerosis. These studies, however, in addition to being contradictory, do not offer details of the effects of atherosclerosis on adenosine receptor-mediated responses of the blood vessels. Moreover, the relationships between atherosclerosis and the vascular actions of adenosine receptor activation in other species are not known. The objective of our study was, therefore, to investigate in some detail the effect of LDL, a known causative factor for atherosclerosis, on adenosine receptor-mediated vasorelaxant responses of porcine coronary artery. This coronary artery has been shown to resemble that of the human in several aspects including its responsiveness to adenosine receptor activation (Makujina et al., 1990) and the effect that atherosclerosis elicits on it (Cox and Cohen, 1996b). The study was performed by measuring relaxations and nitrite production in coronary artery rings in response to NECA, CGS-21680 and/or CAD in the presence and absence of LDL. As demonstrated previously in our laboratory (Abebe et al., 1995), nitrite measurement allows the assessment of nitric oxide release from the coronary artery endothelium. Further, the oxidation of LDL was evaluated by TBARS assays (Schuh et al., 1978; Hessler et al., 1983) in order to determine the role of o-LDL in the responsiveness of the artery to the adenosine agonists. We have also examined the effect of the antiatherogenic lipoprotein, HDL, (Steinberg, 1978) on the oxidation and vascular effects of LDL. To our knowledge, this is the first attempt investigating the effects of LDL on adenosine receptor-mediated functional responses of coronary artery.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Preparation of vascular segments. Male pig hearts were obtained from a local slaughterhouse within 30 min of exsanguination and transported to the laboratory in ice-cold Krebs-Henseleit solution (composition in mM: 118 NaCl, 4.8 KCl, 1.2 MgSO4, KH2O4, 25 NaHCO3, 2.5 CaCl2 and 11 glucose; pH 7.4). Left circumflex coronary arteries were dissected from the hearts and cleaned of adhering fat and connective tissue. Ring segments of 3 mm in length were cut, with special care taken to minimize damage to the endothelium. The endothelium was removed from some of the rings by gently abrading the lumen with tips of forcepts (Abebe et al., 1994).

Tissue bath studies. Coronary artery rings, with and without endothelium, were mounted individually between two stainless steel hooks in isolated tissue baths containing gassed (95%0₂ + 5% CO₂) Krebs-Henseleit solution maintained at 37°C. Changes in isometric tension were measured with Grass FT.03D force transducers connected to sensormedics R611 dynographs or Grass 7E polygraphs. Tissues were equilibrated for 90 min under a resting tension of 2 × g and subsequently challenged with KCl (50 mM) until reproducible contractile responses were obtained. In our previous studies (Abebe et al., 1994, 1995), this resting tension was found to be suitable for the experiments performed. The integrity of the endothelium was evaluated as described by Abebe et al. (1994) in KCl- (30 mM) contracted rings using bradykinin (100 nM) as an endothelium-dependent relaxant agent. After this, the tissues were equilibrated and then incubated with n-LDL (100 µg/ml, 200 µg/ml) for 20 min and 4 hr in the absence or presence of copper sulfate (5 µM) and/or HDL (100 µg/ml). Some tissues were also equilibrated for similar times in the absence of lipoproteins with or without copper sulfate to be used as controls. The coronary rings were then contracted with PGF₂ (20 µM) and relaxant responses to NECA, CGS-21680 and/or CAD determined in a cumulative fashion (Abebe et al., 1994). All experiments were performed in the presence of indomethacin (10 µM) to minimize the possible effects of endogenous prostanoids (Abebe et al., 1995).

Determination of nitrite. The effect of LDL on adenosine receptor-mediated nitrite generation was determined using NECA as an agonist, following the assay methodology we used previously (Abebe et al., 1995). In brief, endothelium-intact and denuded coronary artery rings were equilibrated for 90 min in isolated tissue baths containing Krebs-Henseleit solution (95% O₂ + 5% CO₂, 37°C). The vessels were subsequently treated with indomethacin, KCl and bradykinin in the manner described above. The rings were then incubated for 20 min or 4 hr with or without n-LDL (100 µg/ml) in the presence or absence of HDL (100 µg/ml). After this, the coronary artery rings were transferred to 24-well dishes containing gas (95% O₂ + 5% CO₂)-saturated Krebs-Henseleit solution and incubated with NECA (10 µM) for 30 min at 37°C. Other vessels used to assess basal release of nitrite were treated similarly but without NECA. The incubation solutions were removed and assayed for nitrite contents using the Griess reaction, as determined previously in our laboratory (Abebe et al., 1995). Briefly, the media were first treated with acidified sulfanilamide and N-(1-naphthyl)-ethylenediamine solution for 30 min at room temperature. Samples of the resulting solutions were then read spectrophotometrically at 540 nm and compared with standard sodium nitrite solutions. As shown in our earlier study (Abebe et al., 1995), in positive control experiments, bradykinin elicited a marked increase in the release of nitrite in endothelium-intact but not denuded coronary artery rings (data not shown).

Determination of LDL oxidation. The extent of LDL oxidation was assessed by determination of TBARS as described by Schuh et al. (1978) with a slight modification. Specifically, n-LDL (100, 200 µg/ml) was incubated in gassed (95% O₂ + 5% CO₂) Krebs-Henseleit solution for 20 min and 4 hr at 37°C in the presence or absence of copper sulfate (5 µM) or HDL (100 µg/ml). The lipoprotein-containing supernatant was removed and mixed with 1 ml each of 1% 2-thiobarbituric acid and 20% trichloroacetic acid and heated to 95°C in boiling water for 1 hr. After cooling, the solution was centrifuged at 1000 × g for 10 min. The absorbance of the supernatant was immediately determined spectrophotometrically at 532 nm. As control, the absorbance of "unoxidized" LDL (i.e., LDL used directly without being incubated in Krebs-Henseleit solution) was measured after being treated with reagents in the above manner. Freshly diluted trimethoxypropane was employed as a standard.

Data analysis. Relaxant responses of coronary artery rings to NECA, CGS-21680 and CAD were calculated as percentages of reduction in tension generated by PGF₂. The concentrations of adenosine agonists eliciting EC₅₀, which were used to assess tissue sensitivity to the drugs, were determined from dose-response curves. Nitrite contents were calculated as µM/mg tissue weight. For LDL oxidation, the amounts of TBARS were expressed in terms of MDA nmolar equivalents and effects of various treatment conditions were evaluated as %MDA with respect to control. All data are presented as mean ± S.E.M. Statistical differences between groups were assessed as described previously (Abebe et al., 1994, 1995) using Student's t test, and a 95% confidence level (P < .05) was accepted significant.

Drugs and chemicals. Human n-LDL and HDL were purchased from Biomedical Technologies (Stoughton, MA) and stored at 4°C for no more than the expiration date specified by the company. NECA, CGS-21680 and CAD were supplied by Research Biochemicals Inc. (Wayland, MA). PGF_2a, bradykinin, indomethacin, sulfanilamide, N-(1-naphthyl)-ethyldiamine, sodium nitrite and 2-thiobarbituric acid were obtained from Sigma Chemical Co. (St. Louis, MO). MDA bis (dimethyl acetal) (i.e., 1,1,3,3,-tetramethoxypropane) was from Aldrich Chemical Co. (Milwaukee, WI).

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Effect of LDL on NECA, CGS-21680 and CAD-induced relaxations. As reported previously (Abebe et al., 1994), NECA, CGS-21680 and CAD in the concentration range of 10⁹ to 10⁴ M produced dose-dependent relaxations of porcine coronary artery rings contracted with PGF₂ (20 µM), both in the absence and presence of endothelium (figs. 1 and 2). In endothelium-denuded rings, the dose-response curves for NECA and CGS-21680, but not for CAD, were shifted markedly to the right relative to the corresponding responses in intact tissues, as determined by the EC₅₀ values of the agonists (table 1). This observation is reflected by comparison of corresponding curves for the agonists in figures 1 and 2. Preincubation of endothelium-intact coronary artery preparations with n-LDL (100, 200 µg/ml) for 4 hr in the absence or presence (in case of NECA) of copper sulfate (5 µM) resulted in a dose-related inhibition of the relaxations produced by NECA and CGS-21680 but not by CAD (fig. 1). Assessment of the effect of copper sulfate revealed that the compound by itself did not produce alterations in the responsiveness of the vessels to NECA. However, LDL was without effect on the vasorelaxant responses generated by any of the adenosine agonists used in endothelium-denuded rings (fig. 2). The inhibitory effect of LDL on NECA-induced relaxation of the intact tissues was prevented by the concomitant incubation of the preparations with HDL (100 µg/ml) (fig. 3). However, HDL per se, when incubated for a similar duration of time, neither inhibited nor potentiated the responses produced by NECA. However, a 20-min preincubation of coronary artery rings with n-LDL elicited no inhibition of the relaxation evoked by NECA in the presence of endothelium (fig. 4).

View larger version (20K): [in this window] [in a new window]   Fig. 1.   Effect of LDL (100, 200 µg/ml) and/or copper sulfate (5 µM) on NECA-, CGS-2168- and CAD-induced relaxations of PGF2- (20 µM) contracted porcine coronary artery rings in presence of endothelium. Rings were preincubated with LDL and/or copper sulfate for 4 hr before adding the adenosine analogs. Each point shown is mean ± S.E.M. of data from 12 to 24 rings obtained from 4 to 8 pigs. *Significantly different from corresponding control (P < .05).

View larger version (14K): [in this window] [in a new window]   Fig. 2.   Effect of LDL (100 µg/ml) on NECA-, CGS-2168- and CAD-induced relaxations of PGF2- (20 µM) contracted porcine coronary artery rings in absence of endothelium. Rings were preincubated with LDL for 4 hr before adding the adenosine analogs. Each point shown is mean ± S.E.M. of data from 12 to 24 rings obtained from 4 to 8 pigs.

View this table: [in this window] [in a new window]   TABLE 1 EC50 values (µM) for adenosine analogs producing relaxations of endothelium-intact (+) and denuded () porcine coronary artery rings

View larger version (19K): [in this window] [in a new window]   Fig. 3.   Inhibitory effect of HDL (100 µg/ml) on the action of LDL (100 µg/ml) on NECA-induced relaxations of PGF2- (20 µM) contracted porcine coronary artery rings in presence of endothelium. Rings were preincubated with HDL and LDL for 4 hr before adding NECA. Each point shown is mean ± S.E.M. of data from 12 to 24 rings obtained from 4 to 8 pigs. *Significantly different from corresponding control (P < .05).

View larger version (14K): [in this window] [in a new window]   Fig. 4.   Effect of LDL (100 µg/ml) on NECA-induced relaxations of PGF2- (20 µM) contracted porcine coronary artery rings in presence of endothelium. Rings were preincubated with LDL for 20 min before adding NECA. Each point shown is mean ± S.E.M. of data from 15 rings obtained from 5 pigs.

Effect of LDL on NECA-induced nitrite production. Consistent with our earlier work (Abebe et al., 1995), NECA (10 µM) induced enhancement of nitrite release in endothelium-intact coronary artery rings relative to the amount generated under basal conditions (fig. 5). The nitrite produced by NECA was inhibited to basal level with 4-hr preincubation of the tissues with n-LDL (100 µg/ml) (fig. 5). However, a 20-min preincubation with the lipoprotein did not result in a significant change in the levels of nitrite generated by NECA. The inhibitory effect of LDL on NECA-induced nitrite production was prevented by the concomitant incubation of the blood vessels with HDL (100 µg/ml). Nevertheless, HDL per se with 4-hr incubation did not significantly change the effect of NECA on nitrite generation. In addition, neither LDL, HDL nor a combination of both affected the basal production of nitrite in the coronary artery rings (fig. 5).

View larger version (15K): [in this window] [in a new window]   Fig. 5.   Effect of LDL and/or HDL (100 µg/ml) on NECA- (10 µM) induced production of nitrite in endothelium-intact porcine coronary artery rings. Rings were pretreated with LDL and HDL for 20 min or 4 hr before processing for nitrite assays. Rings were incubated with NECA for 30 min before assaying nitrite levels. Each bar represents mean ± S.E.M. of data from 12 to 18 rings obtained from 4 to 6 pigs. 1, Basal nitrite production (control); 2, effect of NECA; 3, effect of 4-hr preincubation with LDL on NECA-induced nitrite production; 4, effect of 4-hr preincubation with combination of LDL and HDL on NECA-induced nitrite production; 5, effect of 4-hr preincubation with HDL on NECA-induced nitrite production; 6, effect of 20-min preincubation with LDL on NECA-induced nitrite production; 7, effect of 4-hr preincubation with LDL on basal nitrite production; 8, effect of 4-hr preincubation with HDL on basal nitrite production; 9, effect of 4-hr preincubation with combination of LDL and HDL on basal nitrite production. *Significantly different from basal or control (1) (P < .05).

LDL oxidation. In figure 6 is shown the oxidation status of LDL under various experimental conditions. n-LDL (100, 200 µg/ml) incubated in Krebs-Henseliet solution for 4 hr exhibited a concentration-dependent increase in magnitude of oxidation compared to the value obtained in control (i.e., n-LDL assessed without preincubation). The presence of copper sulfate (5 µM) in the medium did not affect the oxidation status of the lipoprotein. The enhancement in LDL oxidation observed with 4-hr incubation was suppressed to control levels by concomitant incubation with HDL (100 µg/ml). However, the oxidation of LDL incubated in Krebs-Henseliet medium for 20 min was not significantly different from that of control.

View larger version (13K): [in this window] [in a new window]   Fig. 6.   Oxidation of LDL (100, 200 µg/ml) incubated in Krebs-Henseleite solution for 20 min and 4 hr in the presence or absence of copper sulfate or HDL relative to the oxidation value of LDL used without prior incubation. Each bar represents mean ± S.E.M. of four to six observations. 1, 100 µg/ml LDL used immediately without preincubated (control); 2, 100 µg/ml LDL preincubated for 20 min; 3, 100 µg/ml LDL preincubated for 4 hr; 4, 100 µg/ml LDL preincubated for 4 hr in the presence of copper sulfate (5 µM); 5, 200 µg/ml LDL preincubated for 4 hr; 6, 100 µg/ml LDL plus 100 µg/ml HDL preincubated for 4 hr. *Significantly different from control (1) (P < .05).

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

The data presented in this investigation show that the coronary vasorelaxant effects of the adenosine agonists, NECA and CGS-21680, were partially endothelium-dependent and this was associated with increased production of nitrite. These results are consistent with our previous observations (Abebe et al., 1994) and further confirm that NECA and CGS-21680 produce vasorelaxation in part by causing the release of nitric oxide from the coronary endothelium (Abebe et al., 1995). Preincubation of porcine coronary artery rings with n-LDL for 4 hr resulted in a selective inhibition of the endothelium-dependent relaxations and nitrite production induced by NECA and CGS-21680. This suggests that a 4-hr exposure of coronary artery to LDL causes impairment of nitric oxide-mediated endothelium-dependent vasorelaxant responses evoked by adenosine receptor stimulation. Although the impairment of endothelium-dependent vasorelaxation by LDL has been reported by a number of other investigators (reviewed in Flavahan, 1992; Cox and Cohen, 1996a), such a documentation of attenuation of adenosine receptor-mediated vascular responses by the lipoprotein is the first of its kind. In addition, our study, for the first time, has demonstrated the inhibitory effect of LDL on nitric oxide contents more directly. Earlier studies performed by other workers have implied LDL-induced inhibition of endothelial nitric oxide based on indirect methods using nitric oxide synthase inhibitors in vasorelaxation experiments and by measuring cGMP (Flavahan, 1992; Cox and Cohen, 1996b). In view of the positive correlation we established previously between the release of nitric oxide and vessel relaxation (Abebe et al., 1995), the decreased nitric oxide contents measured in the current investigation in the presence of LDL is believed to contribute to the reduced endothelium-dependent responses recorded in response to NECA and CGS-21680. In contrast to the 4-hr preincubation period, a 20-min preincubation of the coronary vessels with n-LDL did not produce changes in the relaxant responses and nitrite levels elicited by the adenosine agonists. This indicates that a relatively short period of exposure of the coronary tissue to LDL does not have a deleterious effect on its function. Previous studies from different laboratories have suggested that oxidatively modified LDL rather than n-LDL plays a more important role in altering vascular reactivity, including the inhibition of endothelium-dependent relaxation (Simon et al., 1990; Mangin et al., 1993; Tanner et al., 1991; Cox and Cohen, 1996b). We, therefore, hypothesized that the differing vascular effect of LDL with time is related to its state of oxidation. This possibility was confirmed in our study by the recording of augmented oxidation of LDL incubated in Krebs-Henseleit buffer for 4 hr but not for 20 min. This implies that o-LDL but not n-LDL caused the impaired endothelium-mediated responses documented in response to NECA and CGS-21680. Our finding of enhanced oxidation of LDL with 4-hr incubation in buffer also provides evidence that the lipoprotein is susceptible to oxidative modification when used under in vitro conditions and this could occur both in the presence and absence of a biological system (i.e., coronary artery in our situation). The oxidation of LDL in a cell-free system has been suggested to occur in the presence of trace metallic ion catalysts, such as Cu⁺⁺ (Simon et al., 1990; Dreugnot et al., 1994; Guyton et al., 1995; Cox and Cohen, 1996b). However, this was not the case in our investigation as the magnitude of oxidation of the lipoprotein was similar both in the presence and absence of copper sulfate. In agreement with this, the addition of copper sulfate in the medium did not result in a significant change in the effect of LDL on the relaxant responses of the coronary artery to NECA. In our preliminary experiments, we also observed that the effect of LDL on NECA-induced nitrite production was not influenced by copper sulfate (data not shown). This leads to the conclusion that oxidation of LDL could be induced in the absence of copper sulfate under the experimental conditions employed here. It is possible that under this condition, all the LDL was fully oxidized and no further oxidation could be catalyzed by copper sulfate.

In our experiments, we demonstrated that LDL did not produce an effect on the responses of endothelium-denuded preparations to NECA and CGS-21680. The lipoprotein was also without effect on the coronary artery responses to CAD which produced endothelium-independent relaxation (our study; Abebe et al., 1994, 1995). These results support the notion that LDL/o-LDL does not have a direct effect on the responsiveness of the coronary artery smooth muscle to adenosine receptor agonists. The lack of such an effect on vascular smooth muscle has also been reported by several other workers using other types of drugs and/or preparations (Galle et al., 1995; Tomita et al., 1990; Cox and Cohen, 1996b). The reason for the differential effect of LDL on vascular smooth muscle and endothelial cells is not known, but the data suggest variations in the susceptibility of the two types of cells to the lipoprotein. It may be interesting to investigate if this variation is related to the adenosine receptor subtypes present on the endothelial and smooth muscle cells of the coronary artery (Abebe et al., 1994). It also remains to be investigated if this observation holds true with prolonged (>4 hr) contact of the coronary tissue with LDL. However, reports by Galle et al. (1990) and Niu et al. (1995) have further demonstrated o-LDL-induced enhancement of contractile effects of some other agonists in rabbit femoral artery and aorta in the absence of endothelium indicating that the lipoprotein may have additional (direct) effect on vascular smooth muscle in certain cases. Presently, the reason for the discrepancy between our data and those of others regarding the smooth muscle effect of LDL is not readily apparent. It, however, appears that this could be related to differences in the species of animals (porcine vs. rabbit) and/or types of tissues (coronary vs. femoral artery or aorta) used in these studies. Nevertheless, the porcine coronary artery data seem to be more relevant to the human condition (Makujina et al., 1990; Cox and Cohen, 1996b).

The underlying mechanisms for the inhibitory effect of o-LDL on the functions of vascular endothelial cells are not well understood. However, there is the suggestion that o-LDL contains lysophosphatidylcholine that accumulates in endothelial cells contributing to this effect by different mechanisms including inhibition of agonist-induced production of inositol 1,4,5-trisphosphate and release of intracellular calcium (Yla-Herttuala et al., 1989, Steinbrecher et al., 1984; Yokoyama et al., 1990; Hirata et al., 1991). In addition, the phospholipid, being amphiphilic, has been proposed to alter the physical properties and thus functions of endothelial cell membranes (Portman and Alexander, 1969; Eisenberg et al., 1969). Another possibility is the direct inhibitory effect of LDL or lysophosphatidylcholine on the expression and/or activity of constitutive nitric oxide synthase which is present in endothelial cells (Cox and Cohen, 1996a; Liao et al., 1995). Our observation of lack of effect of LDL on basal nitrite contents supports the view that the lipoprotein is effective only during agonist-induced activation of the signaling system.

Our results also reveal that the inhibitory effects of LDL on endothelium-mediated vasorelaxation as well as on the production of nitrite induced by NECA and/or CGS-21680 were attenuated by HDL. This indicates the antagonistic effect of HDL against LDL at the level of vascular tissue, further confirming its anti-atherogenic action in atherosclerosis (Steinberg, 1978; Matusuda et al., 1983). It has been proposed that HDL could produce this effect by preventing the oxidation of LDL and by facilitating the removal of lysophosphatidylcholine from o-LDL and endothelial cells (Matusuda et al., 1983; Parthasarathy et al., 1990). The former proposal is substantiated by our observation of reduced oxidation of LDL in the presence of HDL. The fact that HDL by itself was without action on the relaxation and release of nitrite provides evidence that its effect was specific against that of LDL on agonist-mediated responses. Such a documentation of effect of HDL represent the first report of its kind.

Our findings could be relevant to the pathogenesis of hypercholesterolemia or atherosclerosis in several aspects. Although endothelium-dependent relaxation has been shown to be impaired in coronary arteries from pigs and humans under these conditions (Simon et al., 1990; Flavahan, 1992; Mangin et al., 1993; Tanner et al., 1991), the scenario for adenosine receptor-mediated responses has not been adequately described. The information presented in this communication contributes to the understanding of this problem and this is particularly important in view of the significant role that adenosine plays in the regulation of coronary function. If adenosine receptor-mediated responses are attenuated in hypercholesterolemia and atherosclerosis under in vivo conditions, this may lead to the development of impaired coronary blood flow and may impact on the function of the heart. Furthermore, if LDL is responsible for this impairment, HDL may play a beneficial role through its "anti-LDL" action. The recognition of the roles of o-LDL and nitric oxide in this mechanism may also provide further understanding for the development of therapeutic interventions.

In summary, the data presented in this report demonstrate that LDL, with a sufficient period of incubation, selectively inhibits the endothelium-dependent relaxant responses and production of nitrite induced by NECA and CGS-21680 in porcine coronary artery. These effects were associated with augmented oxidation of the lipoprotein. HDL attenuated the inhibitory effects of LDL on muscle relaxation, nitrite generation as well as oxidation of the lipoprotein. It is concluded that o-LDL impairs coronary artery relaxation mediated via adenosine receptors by inhibiting nitric oxide-mediated endothelial responses, and this effect is prevented by HDL. Such modulation may be important in hypercholesterolemia and in the development of atherosclerosis.