Atherosclerosis 133 (1997) 183–192

Effect of the oxidation state of LDL on the modulation of arterialvasomotor response in vitro.

Nathalie Mougenot a,b, Philippe Lesnik a,c, Juan Fernando Ramirez-Gil a,b, Patrick Nataf a,d,Ulf Diczfalusy e, M. John Chapman a,c,*, Philippe Lechat a,b

a Institut Federatif de Recherche de Physiopathologie et de Genetique Cardio6asculaire, Pitie-Salpetriere Uni6ersity Hospital,75651 Paris Cedex 13, France

b Department of Pharmacology, Pitie-Salpetriere Uni6ersity Hospital, 75651 Paris Cedex 13, Francec Institut National de la Sante et de la Recherche Medicale (INSERM) Unit 321, Hopital de la Pitie-Salpetriere, 75651 Paris Cedex 13, France

d Department of Cardiac Surgery, Hopital de la Pitie-Salpetriere, 75651 Paris Cedex 13, Francee Department of Clinical Chemistry, Huddinge Uni6ersity Hospital, Huddinge, Sweden

Received 29 July 1996; received in revised form 11 April 1997; accepted 5 May 1997

Abstract

Although it is established that highly oxidized LDL modify both vasodilator and vasoconstrictor responses in normal andatherosclerotic arterial tissue, there is a paucity of data on the relationship between the degree of the oxidative modification ofLDL and vasomotor response. We therefore compared the impact of native LDL (Nat-LDL), and of partially (P-oxLDL), ofmoderately (M-oxLDL) and of highly oxidized LDL (H-oxLDL) on the vasomotor response of isolated human internal mammaryartery and of rat thoracic aorta. Copper-mediated oxidative modification for up to 24 h at 37°C was characterised by a progressiveincrease in the net negative electrical charge of LDL, and in the content of oxysterols; by contrast, lipid hydroperoxide andTBARS content peaked in M-oxLDL at 6 h. Neither basal vascular tone nor vasoconstriction induced by KCl (100 mmol/l) weremodified significantly in arterial segments in relation to the degree of LDL oxidation. While Nat-LDL did not modify thecontractile response of rat aorta to norepinephrine, increase in the degree of oxidative modification of LDL progressively andsignificantly shifted the norepinephrine response curve to the right (EC50 values for Nat-LDL, M-oxLDL and H-oxLDL:1.290.5×10−8, 3.591×10−7, 1.390.4×10−6 mol/l respectively) with reduction in the maximal effect (74.5912.2 and100.196.2% for H-oxLDL and M-oxLDL respectively, PB0.05 versus controls). Similar findings were made in human arteriestreated with H-oxLDL (PB0.05 for EC50 and maximal response versus controls). The acetylcholine-induced, endothelial-depen-dent relaxation of rat aortic segments was significantly and progressively impaired with increase in the degree of LDL oxidation,maximal relaxation with H-oxLDL being 3-fold less (PB0.05) than Nat-LDL at the same protein concentration (100 mg/ml).Acetylated LDL was without effect. Our data indicate that the increase in the degree of copper-mediated, oxidative modificationof LDL parallels progressive reduction in the vasomotor response of the arterial wall to norepinephrine-induced contraction andto acetylcholine-induced relaxation subsequent to precontraction. Our data are consistent with the hypothesis that the majoroxysterols (7-ketocholesterol, 7b-hydroxycholesterol) present in Ox-LDL underlie such effects. © 1997 Elsevier Science IrelandLtd.

Keywords: Oxysterol; Copper-oxidized low-density lipoprotein; Endothelium-dependent relaxation; Vasoconstriction; Humaninternal mammary artery; Rat thoracic aorta

1. Introduction

Low density lipoproteins (LDL) constitute the majorvehicle for cholesterol transport in human plasma. Con-

* Corresponding author. Tel.: +33 142177878; fax: +33145828198

0021-9150/97/$17.00 © 1997 Elsevier Science Ireland Ltd. All rights reserved.PII S 0021 -9150 (97 )00124 -X

N. Mougenot et al. / Atherosclerosis 133 (1997) 183–192184

siderable evidence is now available to support the hy-pothesis that oxidative modification of LDL underliesthe atherogenicity of these particles in vivo [1]. Indeed,oxidized forms of LDL and apo B-100, in addition tooxidized lipids and oxysterols, have been identified inatherosclerotic plaques [2]. The biological oxidation ofLDL involves marked modification of its structural andcompositional characteristics, including peroxidation oflipids leading to lipid degradation with formation ofaldehydic products, hydrolysis of phosphatidylcholine toits lyso-derivative, oxysterol formation, fragmentationof apo B-100, increased net negative charge causingelevated electrophoretic mobility and increased density[3–7].

It is now well established that patients presentingatherosclerotic arterial disease often exhibit abnormalvasomotor phenomena such as vasospasm, blood flowreduction, thrombus formation and predisposition tohypertension, which may lead to a variety of clinicalsyndromes including angina, acute myocardial infarc-tion and sudden cardiac death [8,9]. The mechanismswhich underlie impaired vasomotion of the coronaryarteries remain to be clarified. The vascular endotheliumis intimately implicated in the modulation of vasomotortone as a result of its capacity to release several vasoac-tive substances; these factors include endothelium-derived relaxing factors (EDRF) assimilated to nitricoxide (NO), and prostacyclin (PGI2), which may inducerelaxation of underlying smooth muscle cells, as well asvasoactive molecules such as endothelin 1 and an-giotensin II [10,11].

Several mechanisms implicating the action of Ox-LDLhave been proposed to account for the impairment ofendothelial-mediated relaxation which typically occurs inarteries presenting atherosclerotic disease [11,12]. Theseinclude (1) an increased synthesis of the vasoactive factorendothelin 1 [13], a finding which could not be confirmedby Jougasaki et al. [14]; (2) an inhibition of PGI2 releaseby endothelial cells in the presence of Ox-LDL, lipidperoxides or hydrogen radical [15]; (3) reduced synthesisof NO by endothelium [16,17] in response to inhibitionof NO synthase by Ox-LDL [18]; and (4) an inactivationof NO as a result of the production of superoxide anions[19]. In addition, evidence has been provided for theselective inhibition of vascular smooth muscle cell relax-ation in rabbit and human arteries by Ox-LDL. Themechanism of such inhibition involves attenuation of theagonist-induced rise in tissue content of cyclic nucleotides[20]. Moreover, the enrichment of arterial tissue withcholesterol and lipids may influence transmembraneionic movements, and the calcium content of smoothmuscle cells, thereby leading to alteration in the regula-tion of vascular tone [21]. Furthermore, Ox-LDL hasbeen shown to directly induce contraction of arterialrings pretreated with norepinephrine [22] or thrombox-ane mimetics [23,24]. In addition, Ox-LDL potentiates

the contractile response of blood vessels to other contrac-tile agonists [11,22,25]. These vascular effects of Ox-LDLmimic vascular dysfunction observed during atheroscle-rosis and hyperlipidemia, although marked variabilitybetween individuals has been observed [26–29].

Until now, studies of the effects of Ox-LDL on arterialvasomotricity have involved the use of lipoprotein prepa-rations which have been extensively oxidized in thepresence of copper ions in vitro [11,16,19,20,22–24,30–35]. Indeed, there is a paucity of data on the relationshipbetween the degree of LDL oxidation and the vasomotorresponse of the arterial wall. The purpose of the presentstudy was therefore to investigate the effects of progres-sive degrees of oxidative modification of human LDL onarterial vasomotricity in both human and rat vessels, andto compare such effects to those of native LDL (Nat-LDL).

2. Materials and methods

2.1. Blood samples

Blood (250 ml) was drawn from healthy normolipi-demic human subjects by venipuncture into sterile plasticsacs containing citrate–dextrose solution at the localBlood Transfusion Center (CNTS, Rungis, France).Plasma was subsequently separated by low speed cen-trifugation at 500×g for 20 min at 4°C.

2.2. Isolation of LDL

The major particle subspecies of LDL was isolatedfrom plasma in the density interval 1.024–1.050 g/mlby sequential preparative ultracentrifugation. EDTA(final concentration, 0.1%) and gentamycin (0.001%)were added to plasma before separation. In order toavoid any contribution of Lp(a) to LDL preparations,and because significant amounts of Lp(a) may bepresent over the density interval corresponding to denseLDL (d=1.050–1.063 g/ml), we chose an upper limit-ing density interval of 1.050 g/ml for LDL isolation.The purity and integrity of LDL and apo B-100 wasestablished on the basis of criteria described by Chap-man et al. [36]. In this way, the potential contaminationof LDL with other lipoproteins and plasma proteinswas excluded. The isolated LDL were dialyzed at 4°Cin ‘Spectrapor’ membrane tubing (Spectrum Medical;exclusion limit, 12 000 to 14 000) against degassed 0.01mol/l phosphate-buffered saline (PBS) (pH 7.4) con-taining gentamicin (0.005%) and subsequently filteredthrough a 0.22 mm filter (Costar) and stored at 4°C.The content of lipid peroxides in Nat-LDL prepara-tions was B0.5 nmol/mg LDL protein as determinedby the iodide assay of El-Saadani et al. [37]. Theprotein content of lipoprotein fractions was determined

N. Mougenot et al. / Atherosclerosis 133 (1997) 183–192 185

by the procedure of Lowry et al. [38] using bovineserum albumin as standard. Before oxidative modifica-tion, LDL were dialyzed against degassed PBS (0.01mol/l) at pH 7.4 and at 4°C in order to remove EDTA.

2.3. Chemical modification of LDL

Oxidative modification of LDL was carried out invitro by incubating Nat-LDL (500 mg protein/ml) inPBS in the presence of 2.5 mmol/l CuCl2 at 37°C forperiods of 0–48 h [39] in order to induce differentdegrees of oxidation. At the end of the incubationperiod, Ox-LDL preparations were divided into twoparts. To the first aliquot, EDTA (1%) was added toprevent further oxidation [40]; this aliquot was thenused for analysis of the degree of LDL oxidation bymeasuring (i) lipid hydroperoxide content (expressed asnmol/mg LDL protein) [37], (ii) substances reactivewith thiobarbituric acid and including lipid-derivedaldehyde content (expressed as nmol malondialdehyde(MDA) equivalent/mg LDL protein following the reac-tion with thiobarbituric acid and determined as TBARS[41]), and (iii) the increase in the relative mobility ofLDL on agarose gel due to an enhanced net negativecharge. To the second aliquot, RPMI (this mediumdoes not facilitate oxidation) and a-tocopherol (0.20mmol/l) were added to avoid any further oxidationduring the experimental procedure (see below). Theselipoprotein preparations were subsequently used in ex-periments on vascular segments in organ bath cham-bers. Under these conditions we observed that LDLwas completely protected from further oxidation asassessed by analyses of markers of oxidation after 6 hof incubation in the organ bath (data not shown).

In order to provide insight as to whether the biologi-cal responses induced by Nat-LDL or Ox-LDL weredependent on the protein or lipid moieties of the parti-cle, we used acetylated LDL (Ac-LDL) as a form ofLDL in which only the protein moiety was modifiedand Ox-LDL as a lipid and protein-modified LDL.Ac-LDL was prepared according to the procedure ofBasu et al. [42], using a ratio of 40 mol acetic anhy-dride/mol lysine in apo B-100.

Oxysterol patterns in copper–Ox-LDL were deter-mined by use of a highly accurate technique based onisotope dilution mass spectrometry with individualdeuterated standards [3]. The net negative electricalcharge on Nat-LDL and on modified LDL was esti-mated by electrophoresis on agarose gel (Corning) ac-cording to Noble [43]. The electrophoretic mobility ofmodified LDL was compared with that of the Nat-LDLfrom which it was derived and expressed as the relativeelectrophoretic mobility (REM). The relative mobilityof Ox-LDL on agarose gel electrophoresis was taken asan index of lipoprotein oxidation.

2.4. Vessel preparation

Experimental studies of arterial reactivity were per-formed on both human internal mammary artery andon thoracic aorta from the Wistar rat (300 g). In bothcases, the vessels were harvested from male donors.Animals were stunned by a blow on the neck, and theintact segments were quickly removed after thoraco-tomy. This experimentation was performed within theframework of the Guide for the Care and Use ofLaboratory Animals published by the US NationalInstitutes of Health (NIH publication No. 85-23, re-vised 1985). Human internal mammary artery segmentswere obtained from patients who had undergone coro-nary bypass surgery (Department of Thoracic and Car-diac Surgery, Pitie-Salpetriere Hospital, Paris, France).Arterial rings were prepared from arterial tissue whichwas excess to requirements upon completion of thesurgical procedure.

The arterial segments, of both human and rat origin,were immersed in Krebs buffer (with the followingcomposition in mmol/l: NaCl, 113.5; KCl, 4; CaCl2–2H2O, 1.3; NaH2PO4, 1.1; MgCl2–6H2O, 1.2;NaHCO3, 24; glucose, 11.09) at room temperature andthen dissected from the adventitial tissue. These vesselsegments were cut off transversely in rings 5 mm longwith a fully-preserved endothelium. The rings weresuspended on stainless steel hooks in an isolated organbath filled with 20 ml of RPMI solution (pH 7.4) of thefollowing composition in mmol/l: NaCl, 102.7; KCl,5.36; Ca(N03)2–4H2O, 0.42; Na2HP04–7H2O, 5.64;MgS04–7H2O, 0.4; NaHCO3, 23.8; glucose, 11.11; glu-tathione; a vitamin mixture; and amino acids main-tained at 37°C. This solution was gassed continuouslywith a carbogen mixture of 95% O2–5% CO2 in orderto maintain a constant pH of 7.4. The tension devel-oped in the vascular rings was measured by an isomet-ric force transducer (EMKA Technologies,IT1-4501025), which was connected to a Gould trans-ducer-amplifier (13-461550) and to a computerized sys-tem providing on-line tension values. The rings wereprogressively stretched up to an optimal resting tensionof 4 g. After a stabilisation period of 1 h, the contractileresponse of arterial rings to a depolarizing solution of100 mmol/l KCl (modified Krebs buffer: enriched inK+ (100 mEq/l) and impoverish in Na+ (35 mEq/l))was first tested, in order to evaluate their functionalintegrity. Replacement of RPMI solution with a solu-tion of high K+ (100 mmol/l) concentration induced avasoconstriction attributable to an influx of extracellu-lar calcium ions through voltage-operated channels.Once the plateau value for the contractile response wasattained, the vascular rings were rinsed with RPMIsolution, and allowed to equilibrate for an additionalperiod of 30 min. Arterial rings were then incubatedwith 100 mg protein/ml Nat-LDL and Ox-LDL for 90

N. Mougenot et al. / Atherosclerosis 133 (1997) 183–192186

min. The direct effect of LDL on basal vascular tone,and on the contractile response to depolarizing potas-sium (100 mmol/l) was subsequently investigated. Theconcentration-response to norepinephrine was studied(cumulated concentrations from 10−8−3×10−4 mol/l) in the presence or absence of Nat-LDL or Ox-LDL.The effects of cumulative concentrations of the relaxingagents, acetylcholine (10−7–10−4 mol/l) and trinitrine(6.6×10−6, 2×10−5, 8.6×10−5 and 3×10−4 mol/l),were determined in the presence or absence of Nat-LDL or modified LDL in tissues which were precon-tracted with norepinephrine. Each experiment involved6 independent rings for each dose-response curve. Ineach curve, the maximal effect and EC50 were deter-mined. Contractile responses to norepinephrine wereexpressed as a percentage of the effect obtained withthe hyperpotassic (K+ =100 mmol/l) medium. Relax-ations were expressed as a percentage relative to thevalue of precontraction obtained with norepinephrine.

In order to prevent foaming and aggregation of LDLin the organ bath, oxygen bubble size was reducedwithout modifing PO2 (500 mmHg); Antifoam B (10−5

mol/l), a nonionic emulsifier, was added to the organbath before addition of LDL. Preliminary experiments(data not shown) showed that antifoam B (10−5 mol/l)did not modify the contractile state of the arterial rings,nor the vasoconstrictive response to potassic depolari-sation or norepiniphine, nor the vasorelaxant responseto acetylcholine. In addition, the absence of any effectof the antifoaming solution on the physical integrity ofLDL was verified by electrophoresis of Nat-LDL andOx-LDL in agarose gel after incubation for 24 h withantifoaming solution.

2.5. Chemical compounds

Norepinephrine hydrochloride, acetylcholine chlorideand antifoam B were purchased from Sigma (St Louis,MO). All drugs were dissolved in distilled water, andwere diluted with RPMI solution. For preparation ofpotassium-rich Krebs solution, sodium was replacedisotonically by potassium to attain a potassium concen-tration of 100 mmol/l. RPMI-1640 perfusion mediumwas obtained from Bio-Whittaker. Vitamin E (DL toco-pherol, ICN Biomedicals) was dissolved in absoluteethanol and further diluted with RPMI solution. Trini-trine was purchased from Besins-Iscovesco (Paris,France).

2.6. Statistical analysis

The data are expressed as means9S.E.M. Maximaleffect and EC50 were compared using nonparametricanalysis of variance (Kruskal-Wallis test). When a sig-nificant difference was detected between groups, com-parisons were performed with the non-parametric

Mann and Whitney test. Results were considered assignificant when PB0.05. For multiple comparisons ofdata, Bonferroni’s correction was performed.

3. Results

3.1. Characteristics of nati6e and modified LDL

The degree of oxidative modification of LDL wascharacterized by the content of lipid peroxides and thethiobarbituric acid-reactive products of peroxidationwhich include MDA and other aldehydes (Table 1).Lipoperoxides and TBARS were present in partiallyoxidized LDL (P-oxLDL) at levels of 48–72 nmol lipidperoxides/mg LDL protein and 27–57 nmol equivalentsMDA/mg LDL protein, respectively. The content ofthese oxidation products attained a maximum in mod-erately oxidized LDL (M-oxLDL) (24 h oxidation)(242–298 nmol lipid peroxides/mg LDL protein and64–92 nmol equivalents MDA/mg LDL protein), de-creasing thereafter in highly oxidized LDL (H-oxLDL)(Table 1). Ac-LDL were almost free of such oxidationproducts. The REM of Ox-LDL on agarose gel wasincreased up to 4-fold (H-oxLDL) as compared withNat-LDL, whereas P-oxLDL and M-oxLDL displayedREM values which were intermediate between those ofNat-LDL and its highly oxidized derivative (1.2–1.7and 2.1–2.7, respectively, Table l).

The major oxysterol found in H-oxLDL was 7-oxoc-holesterol (7-ketocholesterol), but in addition, signifi-cant amounts of 7b-hydroxycholesterol, 7a-hydroxy-cholesterol and 5b-6b epoxycholesterol were recovered(Table 2). 24-Hydroxycholesterol, 25-hydroxycholes-terol, 27-hydroxycholesterol and, 5a-6a epoxycholes-terol and cholestane 3b-5a-6b-triol were found asminor components (0.08, 0.44, 0.15, 2.7 and 0.2%,respectively, of total oxysterols). In P-oxLDL, the pro-portion of individual oxysterol species was some 50–100-fold less as compared to that in H-oxLDL.

Table 1Characteristics of Nat-LDL and of its copper-oxidized forms

Lipid REMMalondialdehydehydroperoxides (nmol equivalent

MDA/mg LDL(nmol/mg LDLprotein) protein)

0.290.12.592.5 1Nat-LDL (0 h):1.590.34291560912P-oxLDL (6 h)

M-oxLDL 78914 :2.490.3270928(24 h)

179585931H-oxLDL :490.2(48 h)

Ac-LDL 190.51694 3–4

Nat-LDL were oxidized by incubation at 37°C at a concentration of0.7 mg protein/ml with 2.5 mmol/l CuCl2 during 48 h (see Section 2).

N. Mougenot et al. / Atherosclerosis 133 (1997) 183–192 187

Table 2Effect of the oxidation state on the content of oxysterols in copper-oxidized LDL

Duration of copper oxidationOxysterols (ng oxysterol /mg LDL protein)

24 h (M-oxLDL) 48 h (H-oxLDL)0 h (Nat-LDL) 6 h (P-oxLDL)

347.89190.8 6446.4942357a-OH 33714.398889.225.798354.39175.7 41284.395577.16812.194017.37.8917b-OH

70962 160.7970 1181.49440.4 4650.7914605a,6a-epoxy1042.89565.6 7792.892832.45b,6b-epoxy 136.4991.85 28941.494285

193.6977.829.391 584.3994.822.1953b,5a,6b-triol48.6918.2 58.579224-OH 46.4917.2 194.3930.3

120972.7 877.8951.525-OH 5.794 9.3911538.69950 1700098486.37-oxo 25.794 62432.1925000

98.698.1 129.399.1 294.3932.395.7910.127-OH

Nat-LDL were oxidized by incubation at 37°C at a concentration of 0.7 mg protein/ml with 2.5 mmol/l CuCl2 during 48 h (see Section 2). 7a-OH,7a-hydroxycholesterol; 7b-OH, 7b-hydroxycholesterol; 5a,6a-epoxy, cholesterol-5a,6a-epoxide; 5b,6b-epoxy, cholesterol-5-b,6b-epoxide;3b,5a,6b-triol, cholestane-3b,5a,6b-triol; 24-OH, 24-hydroxycholesterol; 25-OH, 25-hydroxycholesterol; 7-oxo, 7-keto cholesterol; 27-OH, 27-hy-droxycholesterol.Values are mean9S.D.

Only trace amounts of oxysterols were found in Nat-LDL (Table 2).

3.2. Effects of Nat-LDL and Ox-LDL on the basalarterial tone of rat thoracic aorta and human internalmammary artery

The effect of different degrees of copper oxidation ofLDL on basal tone were studied in preparations of ratthoracic aorta. Ox-LDL was either without effect orcaused only weak, nonsignificant vasoconstrictions(1.6% only in rat aorta with H-oxLDL) (results areexpressed as a percentage of developed tension inducedby KCl 100 mmol/l). In human artery, H-oxLDLslowed the initial spontaneous decrease of the devel-oped tension, however, after 90 min, a similar level oftension was recorded with control and Ox-LDL ex-posed rings (data not shown). Nat-LDL was withouteffect on the resting tone of both rat arterial rings andhuman internal mammary artery. These data were ob-tained in studies of 6 separate arterial preparations.

3.3. Effects of Nat-LDL and Ox-LDL onnorepinephrine and KCl-induced contractions of rataorta and human mammary artery.

Depolarization of the arterial rings with 100 mmol/lKCl generated a contractile response which was un-changed for both types of vessels (rat aorta and humanmammary arteries), after exposure to Nat-LDL or Ox-LDL (data not shown). The concentration response tonorepinephrine was studied with M-oxLDL and withH-oxLDL in rat thoracic aorta, but only at the highestdegree of oxidation in internal mammary arteries.Norepinephrine induced a concentration-dependent in-crease in the tension developed in these vessels, with a

maximal effect which was superior to that induced bypotassium depolarization. As shown in Fig. 1A, Nat-LDL did not modify the maximal effect of nore-pinephrine in rat aorta (127.2912.9 versus 12194.6%for control) nor did it induce a significant shift in thedose-response curve. By contrast, H-oxLDL signifi-cantly shifted the norepinephrine curve to the right(EC50=1.390.4×10−6 mol/l with H-oxLDL versus5.990.8×10−8 mol/l for controls (PB0.01) and3.591×10−7 mol/l with M-oxLDL (PB0.05)) andequally significantly decreased the maximal effect ofnorepinephrine (74.5912.2% with H-oxLDL and100.196.2% with M-oxLDL, PB0.05).

Similar results were also observed in human arteriestreated with H-oxLDL (Fig. 1B) in which a significantincrease in EC50 (P=0.05) occurred as compared withthe control, and in which a reduction was also seen inthe maximal effect (PB0.05). Treatment with Nat-LDL 100 mg protein/ml) for 90 min resulted in anon-significant shift to the right of the dose responsecurve (EC50 increase).

3.4. Influence of Nat-LDL and Ox-LDL onendothelium-dependent and -independant 6asodilatationsin aortic and mammary artery rings

For relaxation experiments with acetylcholine, therings were precontracted with a concentration of nore-pinephrine that gave 50% of the maximal contraction.In rat aortic rings (Fig. 2, Table 3), vasodilatation inresponse to variation in the concentration of acetyl-choline (10−7–10−4 mol/l) was not impaired after 90min incubation with Nat-LDL (100 mg protein/ml):72.6915.7 versus 78.599.5% in controls (n=6). Incontrast, endothelial-dependent relaxation induced bycumulative concentrations of acetylcholine (10−7–10−4

N. Mougenot et al. / Atherosclerosis 133 (1997) 183–192188

mol/l) was significantly reduced after 90 min of incuba-tion with H-oxLDL: 21.693.6% (PB0.05) (Table 3).The effect induced by M-oxLDL was intermediate be-tween that of Nat-LDL and H-oxLDL (Fig. 2). Inorder to define the mechanism of the reduced responseto acetylcholine of H-oxLDL-treated rings, we usedAc-LDL as a control for Ox-LDL. After 90 min incu-bation with Ac-LDL, relaxations were not impaired(79.699.3%), and in addition they were significantlydifferent to the response found with H-oxLDL (PB

Fig. 2. Acetylcholine-induced endothelium-dependent relaxation. Infl-uence of native (�), acetylated (2) and moderately oxidized ( ) andhighly oxidized (�) lipoproteins vs. control (�) on intact rat thoracicaorta rings. Rings were precontracted with norepinephrine; LDL ormodified LDL (100 mg protein/ml) were added to the rings in theorgan bath for 90 min before contraction. Relaxations are expressedas a percentage of precontraction. * PB0.05 vs. control, † PB0.05vs. rings pretreated with H-oxLDL. Each curve is the mean of 6experiments. Bars represent S.E.M. See Section 2 for details ofexperimental conditions.

Fig. 1. Influence of native and of oxidized LDL (100 mg protein/ml)on the norepinephrine dose-response curve in rat thoracic aorta (A)and human internal mammary arteries (B). See Section 2 for detailsof experimental conditions. �, Control; �, with Nat-LDL (REM=1); , with M-oxLDL (REM=2); �, with H-oxLDL (REM=4).The data are expressed as percentage of the KCl depolarization-in-duced contraction. Each curve is the mean of 6 experiments. Barsrepresent S.E.M. * PB0.05 vs. controls, † PB0.05 vs. H-oxLDL.Each point represents the mean9S.E.M. n=6 for each group.

0.05) (Fig. 2). Maximal endothelial-independent relax-ation induced by trinitrine (the rings were precontractedwith the maximal concentration of norepinephrine) wasnot impaired with Nat-LDL nor with H-oxLDL (Fig.2, Table 3). A minor tendency to reduce the relaxanteffect (20%) at the initial concentrations used (6.6×10−6 and 2×10−5 mol/l) was observed with H-oxLDL.

In human mammary arteries, no reproducible acetyl-choline-induced relaxation could be obtained. This find-ing is probably related to variation in the degree ofendothelial integrity after the surgical procedure.

4. Discussion

Our present studies demonstrate for the first timethat a progressive increase in the degree of oxidationmodification of LDL tends to lead to a pronouncedalteration in arterial vasomotricity, with significant re-duction in endothelial-dependent acetylcholine-inducedrelaxation and concomitant reduction in nore-pinephrine-induced vasoconstriction. Such observationswere made not only in rat aorta but also in humanmammary artery. Thus, Ox-LDL not only appears toinhibit the ability of blood vessels to relax, but equallyrenders vessels more prone to vasospasm in response tocontractile stimuli. Interestingly, these effects on arte-

N. Mougenot et al. / Atherosclerosis 133 (1997) 183–192 189

rial vasomotricity correlate closely with the increasingcontent of oxysterols in Ox-LDL, suggesting thatboth lysolipids [23,30–32] and oxysterols may con-tribute significantly to the Ox-LDL-induced alterationof vasomotor phenomena.

The potential effect of LDL on resting arterialtone has been variously appreciated in the literature,with some discrepancies. Indeed, some authors de-tected an LDL-induced increase in contractile tension[44]. However in these latter studies, the magnitudeof the modifications was consistently low. Further-more, the vascular reactivity of arteries to native andmodified forms of LDL could differ on the basis oftheir tissular origin. In stimulated pig coronaryartery, in which a major basal relaxant action is ex-erted by endothelium, the addition of Nat-LDL re-sulted in slowly developing contractions, which wereenhanced by Ox-LDL [24]. Neither Nat-LDL (100mg/ml) nor LDL oxidized to different degreesmodified basal arterial tone significantly. A trend toan increase in basal tone in rat aorta was, however,observed with an increase in the degree of oxidationof LDL.

We did not observe any modification in the con-tractile response of either rat or human arteries topotassium-induced depolarization by either Ox-LDLor Nat-LDL at the same protein concentration. Suchcontraction is dependent upon the opening of cal-

cium channels. Our results suggest therefore thatLDL in their native or oxidized state do not modifyeither the number of calcium channels that areopened under depolarization nor modify their kinet-ics. A previous study demonstrated that cholesterolitself sensitizes the artery to calcium ions [45].

Our results are not however consistent with thishypothesis. Thus, such a depolarisation induced con-traction was not modified by 12–72 mmol/l KCl inarterial rings from hypercholesterolemic rabbits [26].The influence of LDL on depolarization-induced con-traction might however vary with the degree of de-polarization. Indeed, Galle et al. [22] did not detectany increase in potassium contraction using a KClconcentration of 130 mmol/l, in agreement with ourpresent findings; nonetheless, these authors observedan enhanced contractile response at low potassiumconcentrations. Such an effect could be secondary toLDL-induced enhancement of smooth muscle cellmembrane permeability to extracellular calcium [44]and indeed a phenomenon of this type could poten-tially play an important role in vivo in directing mi-nor variations in membrane potential and will needfurther investigation.

In studies of normal arteries, we noted a reducedsensitivity to norepinephrine in the presence of Ox-LDL with a significant shift to the right of the doseresponse curve and reduction in the maximal con-tractile effect depending on the degree of LDL oxi-dation. However Nat-LDL at similar lowconcentration (100 mg protein/ml) did not signifi-cantly impair the contractile response to nore-pinephrine. Our experiments clearly demonstratemodulation of arterial vasomotricity by oxidatively-modified LDL as a function of the degree of oxida-tion. The mechanisms of these actions remainunclear and here again, considerable differences inresponse have been reported in the literature accord-ing to the type of vessel used. Differences in bothvascular structure and in the density of specific re-ceptors to specific agonists may possibly explainthese discrepancies. In rabbit femoral arteries, Galleet al. [22] observed an enhanced contractile responseto norepinephrine or to serotonin after 1 h of incu-bation with oxidized human LDL. By contrast, thiseffect was not observed with non-Ox-LDL.

Oxidation of LDL may exert distinct effects on arte-rial vasoconstriction induced by different agonistswhich depend on the species, the tissue studied and thenature of the in vitro system used for oxidative modifi-cation. The progressive impairment of norepinephrinecontraction in our in vitro experiments could be relatedto the toxic effects of some oxysterol componentspresent in LDL. Indeed, we found elevated levels offour oxysterols (7a- and 7b-hydroxycholesterol, 7-keto-cholesterol, cholesterol-5b,6b-epoxide) in Ox-LDL

Table 3Effects of LDL on acetylcholine and trinitrine-induced vascular relax-ation in rat thoracic aorta

Pretreatment Agonist

Maximal relaxation withMaximal relaxation withtrinitrine 3.10−4 Macetylcholine 10−4 M

(% relaxation/ (% relaxation/precontraction) precontraction)

−85.693.9Control −78.599.5Nat-LDL −85.695.2−72.6915.7

(REM=1)—M-oxLDL −52.5912.3

(REM=2)−87.694.4H-oxLDL −21.693.6*

(REM=4)Ac-LDL −79.699.3** —

(REM=4)

Nat-LDL, Ox-LDL and Ac-LDL were added to the organ bath at aconcentration of 100 mg protein/ml for 90 min. Vessels were precon-tracted with norepinephrine. The endothelium-dependent and en-dothelium-independent relaxation are expressed as a percentage ofthe pre-contraction value.Results are expressed as mean9S.E.M. of 6 experiments.* PB0.05 vs. control. ** PB0.05 vs. rings pretreated with H-oxLDL.

N. Mougenot et al. / Atherosclerosis 133 (1997) 183–192190

which are particularly cytotoxic for smooth muscle cellsin culture [46–48]. It is equally relevant that Ox-LDLare both cytotoxic and genotoxic for endothelial cells[49]. Such toxicity, which our data suggest is dependenton the degree of LDL oxidation, may also be depen-dent on the presence of oxysterols in Ox-LDL particles[50,51]. These data are equally relevant to the in vivosituation since the oxysterol content of Ox-LDL in ourstudy resembles that found in human atheroscleroticlesions [2,52] and is compatible with levels reported inhuman plasma [7]

The most consistent finding concerning the impact ofLDL oxidation on arterial vasomotricity involves thereduction of endothelial-induced relaxation by Ox-LDL. In addition to the present studies, this effect hasbeen reported in both rabbit aorta [16,27,30–32,53–55]and in pig coronary artery [24,33]. Endothelial cells, aswell as macrophages, possess scavenger receptors formodified LDL [1,56]. Our experiments demonstratethat acetylcholine-mediated relaxation was not im-paired by Ac-LDL. It is improbable therefore that theeffect of Ox-LDL on endothelial-dependent vasodilata-tion is mediated by binding to scavenger receptorsspecific for Ac-LDL.

Using a bioassay, Galle et al. [16] have demonstratedthat Nat-LDL and Ox-LDL inactivated endothelial-derived relaxing factor (EDRF) in rabbit femoral arter-ies after its liberation by endothelial cells, but not itsformation. The inhibition of arterial relaxation inducedby Ox-LDL has been attributed to lysolecithin genera-tion during oxidative modification [23,30–32]. Theselysolipids may therefore accumulate in the atheroscle-rotic arterial wall. By contrast, other authors haveobserved that both free and lipoprotein-boundlysolecithin either lacked or exerted variable effects onendothelial-dependent relaxation [19,33–35] while oth-ers concluded that Nat-LDL could also inactivate en-dothelial-derived relaxing factor [16]. We interpret theseconflicting data to suggest that lysolipids are probablynot the only factor which may contribute to impairedvasomotion. Other constituents of Ox-LDL, such asoxysterols which increase progressively during oxidativemodification and are dramatically elevated in H-oxLDL, may be of greater significance and are knownto have potent cellular activities. Indeed, the study ofDeckert et al. [55] demonstrates the role of cholesterolderivatives in position 7 as potent inhibitors of en-dothelium-dependent relaxation. Such inhibition wasnot observed in the presence of elevated amounts oflipoperoxides or lysophosphatidylcholine. Our studyconfirms that oxysterols delivered to arterial tissue in amore physiological lipoprotein-associated form, dis-played similar actions. The most active oxysterols de-scribed by Deckert et al. [55] were also quantitativelythe most important in our Ox-LDL preparations (7-ke-tocholesterol, 7b-hydroxycholesterol). Taken together,

the results of Deckert et al. [55] and our own argue fora significant role of oxysterols in the alteration ofendothelium-dependent relaxation.

Non-endothelial-dependent vasorelaxation inducedby trinitrine was not modified by either Nat-LDL orOx-LDL, thereby confirming the specific interaction ofOx-LDL with EDRF. This finding is in agreement withthe observation that atherosclerotic arteries in vivoretain their sensitivity to nitrate compounds [57].

In conclusion, our data demonstrate that the progres-sive oxidative modification of LDL leads to a progres-sive degree of alteration in arterial vasomotricity, andthat such mechanisms are not exclusively dependent onthe presence of active lysolipids, but also to that ofbiologically-active oxysterols in Ox-LDL. The mecha-nisms of such perturbation of vasomotor response re-main to be elucidated but our results indicate: (1) analtered coupling between a-adrenoreceptor activationand calcium mobilization which may result from toxiceffects of Ox-LDL on vascular smooth muscle cells;and (2) inactivation of EDRF and/or a reduction in itsrelease by the arterial wall. Furthermore, the effects ofLDL oxidation on vascular reactivity might depend notonly on vascular structure, a parameter which varies asa function of both species and vascular location, butalso on the relative contribution of the distinct bio-chemical pathways to the modulation of vascular tone.Such marked inhomogeneity in the interactions of LDLand of Ox-LDL with vascular tissue could contribute toa marked variation in the development of atheroscle-rotic lesions among arteries at different anatomicallocations in the vasculature.

Acknowledgements

This study was partially supported by the Faculty ofMedicine, University Pierre et Marie Curie (Paris VI)and by INSERM.

References

[1] Steinberg DS, Parthasarathy S, Carew TE, Khoo JD, WitztumJL. Beyond cholesterol: modifications of low density lipoproteinthat increase its atherogenicity. N Engl J Med 1989;320:915.

[2] Carpenter KLH, Taylor SE, Ballantine JA, Fussel B, Hallivell B,Mitchinson MJ. Lipids and oxidised lipids in human atheromaand normal aorta. Biochem Biophys Acta 1993;1167:121.

[3] Dzeletovic S, Babiker A, Lund E, Diczfalusy U. Time course ofoxysterol formation during in vitro oxidation of low densitylipoprotein. Chem Phys Lipids 1995;78:119.

[4] Esterbauer H, Dieber-Rotheneder M, Waeg G, Striegl G, Jur-gens G. Biochemical, structural, and functional properties ofoxidized low-density lipoprotein. Chem Res Toxicol 1990;3:77.

[5] Aviram M. Modified forms of low density lipoprotein andatherosclerosis. Atherosclerosis 1993;98:1.

N. Mougenot et al. / Atherosclerosis 133 (1997) 183–192 191

[6] Cazzolato G, Avogaro P, Bittolo-Bon G. Characterization of amore electronegatively charged LDL subfraction by ion ex-change HPLC. Free Radic Biol Med 1991;11:247.

[7] Hodis HN, Kramsch DM, Avogaro P, et al. Biochemical andcytotoxic characteristics of an in vivo circulating oxidized lowdensity lipoprotein. J Lipid Res 1994;35:669.

[8] Zeiher AM, Schachinger V. Coronary endothelial vasodilatatordysfunction: clinical relevance and therapeutic implications. ZKardiol 1994;83:7.

[9] Kalsner S. Coronary artery spasm: multiple causes and multipleroles in heart disease. Biochem Pharmacol 1995;49:859.

[10] Borsum T. Biochemical properties of vascular endothelial cells.Virchows Archiv Cell Pathol 1991;60:279.

[11] Cox DA, Cohen ML. Effects of oxidized low-density lipoproteinon vascular contraction and relaxation: clinical and pharmaco-logical implications in atherosclerosis. Pharmacol Rev 1996;48:3.

[12] Flavahan NA. Atherosclerosis or lipoprotein-induced endothelialdysfunction. Potential mechanisms underlying reduction ofEDRF/nitric oxide activity. Circulation 1992;85:1927.

[13] Boulanger CM, Tanner FC, Bea ML, Hann AWA, Werner A,Luscher TF. Oxidized low density lipoproteins induce mRNAexpression and release of endothelin from human and porcineendothelium. Circ Res 1992;70:1191.

[14] Jougasaki M, Kugiyama K, Saito Y, Nakao K, Imura H, YasueH. Suppression of endothelin-1 secretion by lysophosphatidyl-choline in oxidized low density lipoprotein in cultured vascularendothelial cells. Circ Res 1992;71:614.

[15] Thorin E, Hamilton CA, Dominiczak MH, Reid JL. Chronicexposure of cultured bovine endothelial cells to oxidized LDLabolishes prostacyclin release. Arterioscler Thromb 1994;14:453.

[16] Galle J, Mulsch A, Busse R, Bassenge E. Effects of native andoxidized LDL on formation and inactivation of endothelium-derived relaxing factor. Arterioscler Thromb 1991;11:198.

[17] Casino PR, Kilcoyne CM, Quyyumi AA, Hoeg JM, Panza JA.The role of nitric oxide in endothelium-dependent vasodilatationof hypercholesterolemic patients. Circulation 1993;88:2541.

[18] Liao LX, Shin WS, Lee WY, Clark SL. Oxidized LDL decreasesthe expression of endothelial nitric oxide synthase. J Biol Chem1995;270:319.

[19] Plane F, Bruckdorfer KR, Kerr P, Steur A, Jacobs M. Oxidativemodificaton of low-density lipoproteins and the inhibition ofrelaxation mediated by endothelium-derived nitric oxide in rab-bit aorta. Br J Pharmacol 1992;38:216.

[20] Galle J, Banersachs J, Busse R, Bassenge E. Inhibition of cyclicAMP and cyclic GMP-mediated dilatations in isolated arteriesby oxidised LDL. Effect of native and oxidized LDL on forma-tion of endothelium-derived relaxing factor. Arterioscler Thromb1992;12:180.

[21] Gleason M, Medow M, Tulenko T. Excess membrane choles-terol alters calcium movements, cytosolic calcium levels, andmembrane fluidity in arterial smooth muscle cells. Circ Res1991;69:216.

[22] Galle J, Bassenge E, Busse R. Oxidized LDL potentiate vasocon-strictions to various agonists by direct interaction with vascularsmooth muscle. Circ Res 1990;66:1287.

[23] Murohara T, Kugiyama K, Ohgushi M, Sugiyama S, Ohta Y,Yasue H. LPC in oxidized LDL elicits vasoconstraction andinhibits endothelium-dependent relaxation. Am J Physiol1994;267:H2441.

[24] Simon BC, Cunningham LC, Cohen RA. Oxidized low densitylipoproteins cause contraction and inhibit endothelium-depen-dent relaxation in the pig coronary artery. J Clin Invest1990;86:75.

[25] Niu XL, Liu LY, Hu ML, Chen X. Some similarities in vasculareffects of oleil acid and oxidized low density lipoproteins onrabbit aorta. J Mol Cell Cardiol 1995;27:531.

[26] Maggi E, Manzini S, Evangelista S, Meli A. High-cholesteroldiet induces altered responsiveness of rabbit arterial smoothmuscle to noradrenaline. Pharmacology 1985;30:273.

[27] Verbeuren TJ, Jordaens FH, Zonnekeyn LL, Van Hove CE,Coene MC, Herman AG. Effect of hypercholesterolemia onvascular reactivity in the rabbit. Endothelium-dependent andendothelium-independent contractions and relaxations in iso-lated arteries of control and hypercholesterolemic rabbits. CircRes 1986;58:552.

[28] Heistad D, Armstrong M, Marcus M, Piegors D, Mark K.Augmented responses to vasoconstrictor stimuli in hypercholes-terolemic and atherosclerotic monkeys. Circ Res 1984;54:711.

[29] Ginsburg R, Bristow M, Davies K, Dibiase A, Billingham M.Quantitative pharmacologic response of normal and atheroscle-rotic isolated human epicardial coronary arteries. Circulation1984;69:430.

[30] Kugiyama K, Kerns SA, Morrisett JD, Roberts R, Henry PD.Impairment of endothelium-dependent arterial relaxation bylysolecithin in modified low-density lipoproteins. Nature1990;344:160.

[31] Mangin EL, Kugiyama K, Nguy JH, Kerns SA, Henry PD.Effects of lysolipids and oxidatively modified low-density lipo-protein on endothelium-dependent relaxation of rabbit aorta.Circ Res 1993;72:161.

[32] Yokoyama M, Hirata K, Miyake R, Akita H, Ishikawa Y,Fukuzaki H. Lysophosphatidylcholine: essential role in the inhi-bition of endothelium-dependent vasorelaxation by oxidized lowdensity lipoprotein. Biochem Biophys Res Commun1990;168:301.

[33] Tanner FC, Noll G, Boulanger CM, Luscher TF. Oxidized lowdensity lipoproteins inhibits relaxations of porcine coronaryartery: role of scavenger receptor and endothelium-derived nitricoxide. Circulation 1991;83:2012.

[34] Hayashi T, Ishikawa T, Kuzuya M, et al. Chelation of copperreduces inhibition by oxidized low density lipoprotein of en-dothelium-dependent relaxation in porcine coronary arteries.Heart Vessels 1994;9:283.

[35] Plane F, Kerr P, Bruckdorfer KR, Jacobs M. Inhibition ofendothelium-dependent relaxation by oxidized low density lipo-proteins. Biochem Soc Trans 1990;18:1177.

[36] Chapman MJ, Laplaud PM, Luc G, et al. Further resolution ofthe low-density lipoprotein spectrum in normal human plasma:physicochemical characteristics of discrete subspecies separatedby density gradient ultracentrifugation.. J Lipid Res 1988;29:442.

[37] El Saadani MH, Esterbauser H, El Sayed M, Goher M, NassarAY, Jurgens G. A spectrophotometric assay for lipid peroxidesin the serum lipoproteins using a commercially availablereagents. J Lipid Res 1989;30:627.

[38] Lowry OH, Rosebrough NJ, Farr AL, Randall RJ. Proteinmeasurement with the Folin phenol reagent. J Biol Chem1951;193:265.

[39] Esterbauer H, Striegl G, Puhl H, Rotheneder M. Continuousmonitoring of in vitro oxidation of human low-density lipo-protein. Free Radic Res Commun 1989;6:67.

[40] Lesnik P, Dentan D, Vonica A, Moreau M, Chapman MJ.Tissue factor pathway inhibitor associated with LDL is inati-vated by cell- and copper-mediated oxidation. ArteriosclerThromb Vasc Biol 1995;15:1121.

[41] Buege JA, Aust SD. Microsomal lipid peroxidation. MethodsEnzymol 1976;52:302.

[42] Basu SK, Goldstein JL, Anderson RGW, Brown MS. Degrada-tion of cationized low-density lipoprotein and regulation ofcholesterol metabolism in homosygous familial hypercholes-terolemia fibroblasts. Proc Natl Acad Sci USA 1976;73:3178.

[43] Noble RP. Electrophoretic separation of plasma lipoproteins inagarose-gel. J Lipid Res 1968;9:693.

N. Mougenot et al. / Atherosclerosis 133 (1997) 183–192192

[44] Sachinidis A, Locher R, Mengden T, Steiner A, Vetter W.Vasoconstriction: a novel activity for low-density lipoprotein.Biochem Biophys Res Commun (1989) 315.

[45] Yokoyama M, Henry P. Sensitization of isolated canine coro-nary arteries to calcium ions after exposure to cholesterol. CircRes 1979;45:479.

[46] Peng SK, Tham P, Taylor CB, Mikkelson B. Cytotoxicity ofoxidation derivatives of cholesterol on cultered aortic smoothmuscle cells and their effect on cholesterol biosynthesis. Am JClin Nutr 1979;32:1033.

[47] Hughes H, Mathews B, Lenz ML, Guyton JR. Cytotoxicity ofoxidized LDL to porcine aortic smooth muscle cells is associatedwith the oxysterols 7-ketocholesterol and 7-hydroxycholesterol.Arterioscler Thromb 1994;14:1177.

[48] Zhou Q, Smith TL, Kummerow FA. Cytotoxicity of oxysterolson cultered smooth muscle cells from human umbilical arteries.Proc Soc Exp Biol Med 1994;202:75.

[49] Esterbauser H. Cytotoxicity and genotoxicity of lipid oxidationproducts. Am J Clin Nutr 1991;57:423.

[50] Pettersen KS, Boberg KM, Stabursvik A, Pryds H. Toxicity ofoxygenated cholesterol derivatives toward cultered human um-bilical vein endothelial cells. Arterioscler Thromb 1991;11:423.

[51] Sevanian A, Berliner J, Peterson H. Uptake, metabolism and

cytotoxicity of isomeric cholesterol-5,6 epoxides in rabbit aorticendothelial cells. J Lipid Res 1991;32:147.

[52] Bjorkhem I, Anderson O, Diczfalusy U, et al. Atherosclerosisand steril 27-hydroxylase: evidence for a role of this enzyme inelimination of cholesterol from human macrophages. Proc NatlAcad Sci USA 1994;91:8592.

[53] Takahashi M, Yui Y, Yasumoto H, et al. Lipoproteins areinhibitors of endothelium-dependent relaxation of rabbit aorta.Am J Physiol 1990;258:1.

[54] Andrews HE, Bruckdorfer KR, Dunn RC, Jacobs M. Lowdensity lipoproteins inhibit endothelium-dependent relaxation inrabbit aorta. Nature 1987;327:237.

[55] Deckert V, Persegol L, Viens L et al. Inhibitors of arterialrelaxation among components of human oxidized low densitylipoproteins. Cholesterol derivatives oxidized in position 7 arepotent inhibitors of the endothelium-dependent relaxation. Cir-culation 1997;95:723.

[56] Sparrow CP, Parthasarathy S, Steinberg D. A macrophage re-ceptor that recognised oxidezes LDL but not acetylated LDL. JBio Chem 1989;264:2599.

[57] Verdiere C, Drobinski G, Lechat P, Mezghani M, GrosgrogeatY. Effets de la trinitrine injectable sur les stenoses coronairesatheromateuses. Arch Mal Coeur 1985;6:853.

..