The effect of albumin on copper-induced LDL oxidation

Ž .Biochimica et Biophysica Acta 1344 1997 300–311

The effect of albumin on copper-induced LDL oxidation

Edit Schnitzer a,1, Ilya Pinchuk a, Ariella Bor a, Menahem Fainaru b, Dov Lichtenberg a,)

a Department of Physiology and Pharmacology, Tel-AÕiÕ UniÕersity, Sackler School of Medicine, Ramat AÕiÕ, Tel AÕiÕ, 69978, Israelb Department of Internal Medicine, Beilinson Medical Center, Tel-AÕiÕ UniÕersity, Sackler School of Medicine, Ramat AÕiÕ, Tel AÕiÕ,

69978, Israel

Received 16 July 1996; revised 14 October 1996; accepted 16 October 1996

Abstract

In an attempt to gain deeper understanding of the mechanism or mechanisms responsible for the protective effect of2q Ž .serum albumin against Cu -induced peroxidation of low density lipoprotein LDL , we have examined the influence of the

Ž . 2qconcentrations of bovine serum albumin BSA , Cu and LDL on the kinetics of peroxidation. Since the common methodof monitoring the oxidation by continuous recording of the absorbance of conjugated dienes at 234 nm cannot be used athigh BSA-concentrations because of the intensive absorption of BSA, we have monitored the time-dependent increase ofabsorbance at 245 nm. At this wavelength, conjugated dienes absorb intensely, whereas the background absorbance of BSAis low. Using this method, as well as the TBARS assay for determination of malondialdehyde, over a large range of BSAconcentrations, we show that in many cases the influence of BSA on the kinetics of oxidation can be compensated for byincreasing the concentration of copper. This reconciles the apparent contradiction between previously published data.Detailed studies of the kinetic profiles obtained under different conditions indicate that binding of Cu2q to albumin playsthe major role in its protective effect while other mechanisms contribute much less than copper binding. This conclusion isconsistent with the less pronounced effect of BSA on the oxidation induced by the free radical generator AAPH. It is alsoshown that the copper-albumin complex is capable of inducing LDL oxidation, although the kinetics of the latter process isvery different from that of copper-induced oxidation. Nevertheless, when compared to copper induced oxidation at similarconcentration of the oxidation-promotor, the kinetics of oxidation induced by copper-albumin complex is very different andis consistent with a tocopherol mediated peroxidation, characteristic under low radical flux. Similar kinetics was observedfor copper-induced oxidation only at much lower copper concentrations.

Keywords: LDL; Oxidation; Copper-induced oxidation; Albumin; Atherosclerosis

X Ž .Abbreviations: apoB, apolipoprotein B; AAPH, 2,2 -azo-bis amidinopropane hydrochloride; BHT, butylated hydroxytoluene; BSA,bovine serum albumin; DMSO, dimethylsulfoxide; LDL, low density lipoprotein; MDA, malondialdehyde; Na EDTA, disodium2

ethylenediaminetetraacetate; PUFA, polyunsaturated fatty acids; TBA, 2-thiobarbituric acid; TBARS, thiobarbituric acid reactivesubstances; TCA, trichloroacetic acid; TMP, tocopherol-mediated peroxidation.

) Corresponding author. Fax: q972 3 6409113.1 This work constitutes a part of Edit Schnitzer’s Ph.D. thesis, submitted.

0005-2760r97r$17.00 Copyright q 1997 Elsevier Science B.V. All rights reserved.Ž .PII S0005-2760 96 00154-3

( )E. Schnitzer et al.rBiochimica et Biophysica Acta 1344 1997 300–311 301

1. Introduction

Atherosclerosis is a complex chronically develop-ing disease caused by excessive accumulation of

Ž .lipids mostly cholesterol esters in smooth musclecells and macrophages, which in turn results in the

w xformation of foam cells and cell death 1,2 . Oxida-Ž .tive modification of low density lipoproteins LDL

is believed to be a primary cause of this fatal disease.Specifically, peroxidation of polyunsaturated fatty

Ž .acids PUFA is followed by formation of breakdownproducts of the lipid peroxides, mostly aldehydes.These highly potent substances react with the protein

Ž .moiety of the LDL-apolipoprotein B apoB , thusmodifying it to the extent that it is no longer recog-nized by the LDL receptor. Internalization of theoxidatively modified lipoproteins occurs via the scav-

w xenger receptor pathway 3,4 . This mechanism, unlikethe uptake of the non-oxidized LDL through the LDLreceptor pathway, is not down-regulated and there-

w xfore results in the formation of foam cells 3–6 .Appropriate methods for determination of the sus-

ceptibility of LDL to oxidation in vivo are not avail-able. Much effort has therefore been devoted todeveloping in vitro models for assessment of thesusceptibility of LDL to oxidation, using variouspromotors to initiate oxidation. One of the mostcommonly used models is Cu2q-induced oxidation.Under most conditions studied thus far, the kineticsof this reaction is characterized by a ‘lag-phase’ ofrelatively slow oxidation, which precedes the autoxi-

w xdation of most of the oxidizable lipids 7 . During the‘lag phase’, the naturally occurring LDL-associatedantioxidants, mostly vitamin E, are being oxidizedw x6–8 and under most conditions lipid peroxidation is

w xquite limited 7–10 . Only after all the vitamin Ebecomes oxidized, autoxidation occurs, yielding rapidproduction of conjugated dienic hydroperoxides,which later decompose into aldehydes and other

w xproducts 6,7,10,11 . However, under conditions ofslow production of free radicals, peroxidation duringthe initial stages of the ‘lag phase’ becomes verysignificant. This oxidation, which is probably induced

Ž .by a tocopherol mediated peroxidation TMP mecha-nism, is an increasing function of the vitamin E

w xcontent of the LDL 12 . Under this mechanism,vitamin E can be regarded as being a pro-oxidantw x12–15 , although this conclusion is not generally

w xaccepted 16 . Tocopherol may in fact mediate lipidperoxidation either through direct promotion of oxi-

w x qdation by tocopheroxy radicals 12–15 or by CuŽ 2q.obtained by tocopherol mediated reduction of Cuw x w x17–19 . Under ‘mild oxidative conditions’ 17 , suchas the conditions under which TMP kinetics wereobserved in the present study, the major contributionto the composite TMP is the former.

Oxidation of LDL can be inhibited by both water-soluble and lipid-soluble antioxidants. In addition,plasma proteins such as albumin may also inhibit

w x w xcopper-induced LDL oxidation 20,21 . Thomas 20w xand Zawadzki et al. 21 have proposed that albumin

has a protective effect and that this effect is notlimited to the ‘metal sequestering potential’ of albu-min but also relates to its capacity as ‘a radical

w xtrapping protein’ 20 andror to its interaction withw x w xLDL 21 . By contrast, Deigner et al. 22 showed

Žunder conditions of LDL and albumin concentrationsw x.similar to those used by Thomas 20 that oxidation

is not inhibited by albumin.Since albumin may be present in the intima of

arteries, especially due to permeabilization of thew xarterial wall by lipidic oxidation products 10,23,24 ,

the effect of albumin on LDL oxidation carries patho-physiological significance. We therefore tried to rec-oncile the apparently contradictory observations ofthe previous studies. Our results show that the appar-ent discrepancy is due to a difference in Cu2q con-centration used in the previous studies. Furthermore,based on previous data regarding copper binding to

w x w xalbumin 25,26 and LDL 27,28 , we show that overthe studied range of concentrations of Cu2q, albumin,and LDL, the observed effects of serum albuminŽ .both bovine and human are mostly due to copperbinding. Other mechanisms play a much smaller role.

2. Material and methods

2.1. Materials

ŽBoth bovine and human serum albumin essen-. Žtially fatty acid-free , CuCl , a-tocopherol, TBA 2-2

. Ž .thiobarbituric acid , TCA trichloroacetic acid andŽ .MDA malondialdehyde were purchased from Sigma

Ž . Ž X ŽSt. Louis, MO . AAPH 2,2 -azo-bis 2-amidino-

( )E. Schnitzer et al.rBiochimica et Biophysica Acta 1344 1997 300–311302

. .propane hydrochloride , was purchased from PolyŽ .Sciences Warrington, PA . Polyethylene glycol

Ž20 000 was purchased from Fluka Buchs Switzer-.land . TBA reagent was made by dissolving 0.375 g

Ž .TBA in 2.5 ml HCl 10 N and 15 ml TCA. Thesolution was then brought to a final volume of 100ml by adding water. LDL and BSA concentrations

w xwere determined according to Lowry 29 .

2.2. LDL isolation

Human plasma was prepared from the blood ofindividuals fasting for 12 h by centrifugation of the

Žblood in chilled tubes containing Na EDTA 0.1%2.wrv final concentration; pH 7.4 . The successive

procedures involved in LDL isolation were startedimmediately after blood drawing and were all carriedout at 48C. LDL was isolated by sequential flotationof plasma in KBr saline solutions containing 1 mM

Ž .Na EDTA pH 7.4 and varying KBr concentrations2

that correspond to densities of 1.019–1.05 grml.Centrifugation was carried out in an L8-70 Beckman

Žultracentrifuge Beckman instruments, Mountain.View, CA using a 70.1 rotor at 40 000 rpm for 20 h.

The LDL was recovered by tube slicing, re-isolated atthe limiting density under the same conditions andsubsequently dialyzed at 48C in the dark for 24 hagainst four changes of 1 liter of a buffered solutionŽ .pH 7.4 containing 146 mM NaCl, 3.3 mM sodiumphosphate and EDTA. The EDTA concentration inthe dialysis medium was five-fold higher than themolar concentration of LDL, so that after dilution ofthe LDL to 0.2 mM, the medium contained 1mM

w xEDTA 30 . ‘The total copper concentration’ given asC throughout this study given for the various oxida-T

tion experiments is in fact equal to the total copperconcentration minus 1 mM. Prior to being diluted foroxidation studies, the LDL samples were stored at48C for up to four days.

2.3. Vitamin E-enrichment of LDL

LDL was enriched by vitamin E according tow xEsterbauer et al. 31 , using 25 ml of a 10 mM stock

Ž .solution of a-tocopherol in DMSO 250 nmol perml of plasma. Following 6 h of incubation, LDL wasfractionated as above. Control experiments were car-ried out with LDL fractionated after the addition of

25 ml DMSO per ml plasma. This procedure resultedin a 14-fold enrichment of LDL by vitamin E, as

w xdetected by HPLC according to 32 , utilizing a re-verse-phase C-18 column and methanol as a mobilephase.

2.4. Monitoring of oxidation

Oxidation was monitored at 378C by continuousrecording of absorbance at several wavelengths fol-lowing the addition of either Cu2q or AAPH and

Ž .using a Kontron Uvikon 930 double-beam spectro-photometer equipped with a 12 position automatedsample changer. Measurements were carried out in1.5 ml quartz cuvettes containing LDL, with or with-out BSA, at different Cu2q concentrations, as indi-cated for each experiment. Time-courses of the ab-sorbance were measured against a reference contain-ing the same constituents but without CuCl .2

The most commonly used spectroscopic methodfor monitoring LDL oxidation is based on continuousrecording of the absorbance at 234 nm. The ab-sorbance at this wavelength, usually attributed toconjugated dienes, is in fact a sum of absorbancies ofvarious contributors, mostly conjugated dienic hydro-

w xperoxides and 7-keto cholesterol 33 . This methodcannot be used when the system contains relativelyhigh albumin concentrations because albumin absorbsintensely at any wavelength below 242 nm. As an

Ž .example, the OD of 1 mg BSArml 15 mM at 234Ž .nm is about 3 OD units not shown , and with this

high background, any determination of additionalabsorbance by either conjugated dienic hydro-peroxides andror 7-keto cholesterol is very question-able. Yet, spectroscopic monitoring of the accumula-tion of these oxidation products is possible at longerwavelength because the ratio between the absorbanceat 245 nm and the absorbance at 234 nm is 0.6 for

w xconjugated dienic hydroperoxides 34 , 0.7 for 7-ketow xcholesterol 35 , but only 0.15 for BSA. As a result,

the accumulation of absorbing products formed uponoxidation of 0.2 mM LDL can be monitored at 245

Ž .nm even in the presence of up to 60 mM 4 mgrmlBSA. In the range of BSA concentrations where thekinetics can be also studied at 234 nm, monitoringthe oxidation at 245 nm yielded the same kinetic

Ž .profiles as those observed at 234 nm not shown . Infact, during LDL oxidation the absorbance at 245 nm


increased linearly with the absorbance at 234 nm,both in the absence of albumin and in the presence of

Ž 2 .up to 6 mM BSA r G0.998 . The slope of thisdependence was 0.7, in reasonable agreement with

w xthe spectra of the oxidation products 34,35 . More-over, the lag period evaluated from measurement at245 nm was identical to the lag determined by moni-

Ž .toring the oxidation at 234 nm not shown . Based onthis relationship, it can be concluded that monitoringat 245 nm is as reliable as at 234 nm. Hence, in thepresent work, we present the kinetics of oxidation asmonitored at 245 nm.

ŽThe oxidative modification of LDL both in the.absence and in the presence of albumin was also

determined by measuring spectrophotometrically theŽ .amount of thiobarbituric acid TBA according to the

w xmethod of Buege and Aust 36 . Specifically, a solu-Ž .tion of 50 mM CuCl in saline 150 mM NaCl was2

Ž .added to two LDL solutions 0.2 mM; 15 ml onlyŽone of which contained albumin 1 mgrmls15

.mM . Aliquots of 0.5 ml were taken from these twosolutions at 30, 60, 120 and 250 min and after 24 h ofincubation at 378C. In each sample, the oxidation wasimmediately terminated by the addition of 24 mM

Ž .EDTA and 20 mM BHT butylated hydroxytoluene .Ž .TBA reagent 1 ml was then added and the solutions

were boiled at 1008C for 20 min. The solutions weresubsequently cooled to room temperature and cen-trifuged at 8000 rpm for 10 min and the absorbanceof the supernatants was then determined at 532 nm.The amount of TBARS was determined from stan-dard curves generated from known concentrations ofmalondialdehyde in the absence and presence of BSA.

In a few experiments, oxidation was induced byAAPH, which is a water-soluble radical-initiator thatgenerates radicals at a constant rate through thermal

w xdecomposition 37 . Oxidation was monitored by con-tinuous recording of absorbance after AAPH additionusing the same procedure as described above forCuCl .2

2.5. Reproducibility of kinetic measurements of oxi-dation

Under any given conditions, the variability of theresults observed for each LDL preparation, as tested

Žby measurements in duplicates or, in several case,.triplicates was up to "9% with respect to the lag,

"5% with respect to maximal rate and "2% withrespect to the rate of initial stages of oxidation via a

Ž .TMP mechanism see Section 4 . The effect of BSAon the kinetics of oxidation, as measured for variousLDL preparations, varied by up to 10% with respectto the lag, 25% with respect to the maximal rate, and

Ž20% with respect to the initial burst of oxidation see.Section 4 .

2.6. Viscosity measurements

Viscosity was measured with Cannon–Fenske rou-Ž .tine viscometer Induchem size 50.

3. Theoretical consideration

The kinetics of copper-induced oxidation of LDLis determined by the concentration of copper bound

w xto LDL particles 28 ; i.e., by the ratio C rL, whereL

C is the concentration of bound copper and L is theL

LDL concentration. In the presence of BSA, theconcentration of bound copper is lowered by binding

Žof copper to BSA. C sC yC yC where C ,L T W A T

C and C are the total Cu2q concentration, theW A

concentration of free Cu2q and the concentration of.albumin-bound copper, respectively .

To discern the effects of BSA caused by Cu2q-bi-nding from effects mediated by other mechanisms, itis necessary to evaluate C rL. This, in turn, requiresL

knowledge of the binding constants and numbers ofCu2q-binding sites on both LDL and BSA. Copperbinding to BSA has been previously investigated andtwo independent detailed studies resulted in close

w xagreement 25,26 . Both of these studies revealed oneŽ 11 y1.very ‘tight’ binding site K)2=10 M . In

addition, the experimental data of both of these stud-ies are consistent with the existence of one ‘weaker’binding site of a binding constant K s5.2=106

Ay1 w x 2qM 26 . For Cu -binding to LDL, the literature

w xvalues are not as unequivocal 17,27,28 . Kuzuya etw xal. 27 interpreted their equilibrium dialysis data in

terms of two types of binding sites: two extremelystrong binding sites, which bind copper even in thepresence of excess EDTA but do not play a role in

w x ŽLDL oxidation 27 , and about eight weaker but.effective binding sites whose binding constant was

not evaluated. Kontush et al. interpreted the results of


their ‘precipitation experiments’ in terms of 96 bind-4 y1 w xing sites of a binding constant of about 10 M 17 .

Another, less direct approach that can be used forevaluation of copper binding to LDL is to study thedependence of LDL oxidation on copper concentra-tion. Using this approach, Gieseg and Esterbauer

w xhave recently proposed 28 that copper binding toLDL can be described in terms of 10–38 binding

Ž .sites, characterized by a binding constant of 1–2 =

106 My1. No indication was found in this study for‘strong but ineffective’ binding sites. Given the ex-

w xperimental conditions of the latter study 28 as wellas of our recent experiments, carried out at low

ŽcopperrLDL molar ratios Pinchuk et al., unpub-.lished data , the occurrence of such ‘ineffective bind-

ing’ is quite questionable. Furthermore, in the presentŽ .study as in most previous studies strong binding of

copper to such sites would have had a very smallŽeffect on C rL where C denotes the concentrationL L

.of copper bound to ‘effective’ binding sites . Thebinding of copper to LDL, as evaluated from ourstudies, carried out over a wide range of Cu2q andLDL concentrations, is consistent with 17 binding

Ž . 6 y1sites ns17 of a binding constant K s10 M .L

These values were used in all the computations givenbelow.

Given the strong binding of copper to the tightŽbinding sites of albumin, whenever C -A where AT

.is the molar albumin concentration , essentially allthe copper is bound to albumin. In most of ourexperiments, the molar concentration of albumin wassmaller than C . Under these conditions, all of theT

tight binding sites of albumin are occupied. Thismeans that C sC qAqC X qC where C X is theT W A L A

concentration of copper ions that are bound to the‘weak’ binding sites of albumin. C X correlates withA

Ž .the free copper concentration C such thatw

X X 6C r C = AyC s5.2=10Ž .A W A

Hence,

ArC X s1r 5.2=106 =C q1Ž .A W

and therefore

C X sA=5.2=106C r 1q5.2=106CŽ .A W W

Similarly,

C s17L=106C r 1q106CŽ .L W W

Hence,6 6C sAqC 1q 5.2=10 A r 1q5.2=10 CŽ . Ž .T W W

6 6q 17=10 L r 1q10 CŽ . Ž .W

From this equation C can be computed for eachW2q Ž .solution of known concentrations of Cu C , al-T

Ž . Ž .bumin A and LDL L . This allows for the compu-tation of the concentration of the ‘oxidizing complex’Ž .C , and of the average number of copper ionsL

Ž .bound to LDL C rL . Thus, when albumin is addedL

to a solution containing LDL and Cu2q, it is possibleto compute how much copper has to be added tocompensate for the effect of added albumin on theC rL. Fig. 1 presents the results of such computa-L

tions: for each BSA concentration, the computedŽ .relative occupancy C rnL is depicted in this figureL

as a function of the total copper concentration C .T

The concentration of copper that is required for com-pensation for copper binding to albumin can be drawnfrom this figure.

4. Results and discussion

4.1. The effect of albumin on the kinetics of copper-induced lipid-peroxidation

Fig. 2 depicts the time course of oxidation of LDLŽ .0.2 mM by 4 mM or 9 mM CuCl in the presence2

Ž .Fig. 1. Computed dependencies of occupancy C rnL on totalL

copper concentration. Computations were performed as describedin the text for 0.2 mM LDL, in the presence of varying BSA

Žconcentrations as indicated in the figure in mM units; 1 mMs.66.6 mgBSArml .


Ž .of 0–10.5 mM BSA as given next to the curves .Comparison of the different kinetic profiles leads tothe following conclusions:

4.1.1. The effect of albumin depends on the concen-tration of copper

2q Ž .At 4 mM Cu ,3 mM BSA 200 mgrml inhibitedŽ .the oxidation Fig. 2A , similar to the observation of

w x 2qThomas 20 , whereas at 9 mM Cu , the same BSAconcentration had no significant effect on the kinetics

Ž .of oxidation Fig. 2B , similar to the observation ofw xDeigner et al. 22 . Hence, the apparent contradiction

between these previous investigations is a result ofthe different copper concentrations used in these two

Ž .studies 5 mM in Thomas, 10 mM in Deigner et al. .In fact, increasing the concentration of copper com-pensated for most of the effects of BSA on LDLoxidation. At relatively high Cu2qrBSA molar ratioŽ 2q .e.g., at 9 mM Cu and 3 mM BSA , when theoccupancy of the LDL copper binding sites remains

Ž .relatively high C rnLs0.6; Fig. 1 , the kinetics ofL

oxidation was not altered significantly by the pres-ence of BSA. In contrast, at lower copper concentra-

Ž .tions e.g., 4 mM , the presence of the same BSA

Ž .concentration 3 mM resulted in a marked decreaseŽ .of C rnL to about 0.1 and the oxidation wasL

markedly inhibited. It thus appears that increasing theconcentration of copper can compensate for the pres-ence of BSA. The results obtained with human serum

Ž .albumin HSA were qualitatively similar to thoseŽ .presented in Fig. 2 not shown , indicating that,

similar to BSA, the effect of HSA is also mediatedmainly through copper-binding. In fact, at a suffi-ciently high concentration of copper, BSA had almostno effect on the kinetics of oxidation, even at muchhigher BSA concentrations. For example, at 50 mM

2q Ž . Ž .Cu , 15 mM 1 mgrml BSA C rnLs1 hadL

only a slight effect on the oxidation of LDL, asmeasured either by continuous monitoring at 245 nm

Ž .or by determination of TBARS Fig. 3 .

4.1.2. BSA-copper complex can promote LDL oxida-tion

ŽAt a CurBSA molar ratio lower than 1 e.g., 4.5mM BSA at 4 mM Cu2q or 10.5 mM BSA at 9 mM

2q.Cu , all the copper can be expected to be bound toŽtight binding sites in BSA see theoretical considera-

. Žtions . Yet, oxidation occurred Fig. 2A and 2B,

2q 2q Ž . ŽFig. 2. Effect of albumin on Cu -induced LDL oxidation. At time zero, Cu 4 mM in A, 9 mM in B and C was added to LDL 0.2. Ž .mM in the absence and presence of BSA, at concentrations given in the figure. Panel C describes the initial 60 min time-course of LDL

oxidation by 9 mM Cu2q.


Fig. 3. Effect of 15 mM albumin on Cu2q-induced oxidation.2q Ž . Ž .Cu 50 mM was added to LDL 0.2 mM in the absence and

Ž .presence of BSA 15 mM . Panel A depicts oxidation as evalu-ated by determination of MDA concentration. Panel B presentsthe continuous recording of absorbance at 245 nm.

.respectively , indicating that BSA-bound copper is aninducer of oxidation. Furthermore, the initial rate ofoxidation induced by the copper-albumin complexwas higher than the rate observed in the absence ofBSA. This can be clearly seen in Fig. 2C for 10.5mM BSA. As in many other examples, the time-de-pendence of the increase of absorbance was charac-terized by an ‘initial burst’ of increased absorbancefollowed by a slow down and a subsequent increase

Žin rate after a much longer time Fig. 2B; 10.5 mM. 2BSA . Similar kinetic profiles were previously ob-

2 We have considered and ruled out the possibility that this‘initial burst’ is due to scattering of light by LDL aggregates thatcould have been formed upon oxidation of LDL rather than tochanges in absorption. In fact, the size of LDL particles, as

Žmeasured by quasi-elasting light scattering using Malvern’s pho-ton correlation spectrometer model 4700 equipped with an Argon

.laser of wavelength of 488 nm , remained constant at 27.65"1.62Žnm through the time course of the initial stages of oxidation 3

.h .

served by many investigators under conditions ofw xslow rate of production of free radicals 12,28,30 and

were interpreted in terms of a tocopherol-mediatedŽ . w xperoxidation mechanism TMP 12,13 . The observa-

tion of an initial burst of oxidation promoted by theŽ .albumin-copper complex Fig. 2B is consistent with

this interpretation under the assumption that the rateof production of free radicals due to this complex ismuch slower than the rate of production of radicalsduring oxidation promoted by free copper ions. ATMP mechanism is also consistent with the findingthat the initial rate of LDL oxidation by the copper-

Ž 2qalbumin complex at 9.5 mM Cu and 10.5 mM.BSA was markedly enhanced when the LDL was

Ž .enriched by Vitamin E Fig. 4A : in the non-enrichedLDL, limited peroxidation occurred after a short ‘lag’,followed by subsequent peroxidation after muchlonger time, whereas peroxidation of the vitaminE-enriched LDL was not preceded by any ‘lag’, ascould have been expected for oxidation via a TMPmechanism in vitamin E-enriched LDL at a low rate

w xof free radical production 12 .

Fig. 4. Effect of albumin on Cu2q-induced oxidation of vitamin2q Ž .E-enriched and non-enriched LDL. Cu 9.5 mM was added to

Ž . Ž .LDL solid line and to vitamin E-enriched LDL broken line inŽ . Ž .the presence of 10.5 mM BSA A or 7.5 mM BSA B . The

Žconcentration of vitamin E in the LDL was 0.84 mM 0.36. Žmgrml and in the vitamin E-enriched LDL 12.05 mM 5.19.mgrml .


Under conditions where only part of the copperŽwas bound to BSA e.g., 3 mM BSA in Fig. 2A or 6.mM BSA in Fig. 2B , the kinetic profiles had mid-

way characteristics between those observed for oxida-tion induced by copper and oxidation induced by acopper-albumin complex. Specifically, under suchconditions LDL oxidation is probably promoted byboth these inducers, as indicated by the complexnature of the kinetics of oxidation. An example forsuch complex kinetics is that of the oxidation of

non-enriched LDL by 9.5 mM Cu2q in the presenceŽ .of 7.5 mM BSA Fig. 4B . Under theses conditions

Ž .C rnLs0.05 the initial stages of oxidation wereL

quite similar to those observed in the presence ofŽ .10.5 mM BSA Fig. 4A , but at later stages the rate

increased markedly, probably due to Cu2q-inducedoxidation.

Given the above results, it appears that the majormechanism responsible for the protective effect ofalbumin against copper-induced oxidation is binding

Fig. 5. Effect of albumin on LDL oxidation induced by low Cu2q concentration. Cu2q at concentrations given in the figure was added toŽ . Ž . Ž .LDL 0.2mM in the absence of BSA broken line and in the presence of 4.5 mM BSA solid lines . The inset describes the initial stages

Ž .of oxidation 25 min .


of copper to albumin. In an attempt to study whetherthis is the sole mechanism, we have studied the effect

Žof BSA at a much lower copper concentration 0.5.mM , where the oxidation is more sensitive to the

concentration of copper. In addition, under theseconditions an ‘initial burst’ can be seen in the ab-

Ž .sence of BSA broken line in Fig. 5 , probablyreflecting oxidation mediated by a TMP mechanismw x12–14 , as in other conditions of slow radical flux.In the presence of 4.5 mM BSA, no oxidation wasobserved unless the copper concentration was raisedmarkedly. Based on our theoretical considerations,we expected to be able to compensate for the effect

Žof BSA on the value of C rnL which was equal toL.0.11 in the absence of albumin by increasing the

copper concentration to 6.85mM. In fact, the kineticprofile became quite similar to that observed in the

Ž 2q.absence of BSA at 0.5 mM Cu only at 8 mM2q ŽCu , the maximal rate as evaluated from first

. 2qderivatives was recovered at 7.5 mM Cu but inŽ .the initial stages inset similar kinetics required 9

mM or more.It thus appears that copper binding to BSA may be

responsible for most of the effects of BSA on oxida-tion of LDL. Nonetheless, compensating for the vari-ous effects requires different supplementation by cop-per, indicating that other factors must also be in-volved. This conclusion is supported by comparisonof the oxidation induced by 9 mM Cu2q at 10.5 mM

Ž .and 15 mM BSA Fig. 2B . In the presence of 10.5mM BSA, all the copper is bound to BSA and theobserved LDL oxidation must be attributed to thecopper-albumin complex. Further addition of BSAŽ .e.g., to 15 mM cannot affect the reaction via cop-per-binding, since all of the copper was alreadybound to BSA at 10.5 mM BSA. Yet, LDL oxidationis inhibited, indicating that, in addition to copperbinding, albumin has other effects, as previously

w xproposed 20,21 .

4.2. The effect of BSArLDL ratio on the kinetics ofcopper-induced oxidation

Copper binding to the tight binding sites of albu-min is several orders of magnitude tighter than toLDL. Hence, for any given copper concentration, the

Ž .binding of copper to LDL i.e., the ratio C rnL isL

critically dependent on the concentration of albumin

w x w x 2qFig. 6. Effect of the BSA r LDL ratio on Cu -induced LDL2q Ž .oxidation. Cu 40mM was added at time zero to LDL in the

Ž . Ž .absence of solid lines and presence broken line of BSA. LDLand BSA concentrations were as given in the figure. The weight

Žratio between BSA and LDL was kept constant BSArLDL s.40 .

and is only slightly dependent on the LDL concentra-tion. Our conclusion, that the major effect of BSA onthe oxidation is due to copper binding, thereforeimplies that the ratio of BSArLDL cannot be re-garded as relevant to oxidation, in contrast to the

w xconcept expressed in several previous studies 22,38 .This conclusion is supported by the results of the

Cu2q-induced oxidation experiments depicted in Fig.6. These experiments were conducted at a constantalbuminrLDL ratio, varying the concentrations ofboth LDL and albumin. The time-course of oxidation


Ž .of 0.2 mM LDL 100 mg LDL proteinrml in theŽ .presence of 60 mM BSA 4 mgrml is given in Fig.

6A. Fig. 6B and C depict the oxidation in solutionsmade from that of Fig. 6A by serial two-fold dilutionsteps in a medium containing a constant copper con-

Ž .centration 40 mM . As obvious from these time-courses, dilution resulted in a reduction of the protec-tive effect of BSA. This result is consistent with thedilution-induced increase of C rnL, which accordingL

to Fig. 1 is close to 1 in the absence of BSA, aboutŽ .zero for the most concentrated sample Fig. 6A , 0.1

for the sample of Fig. 6B and about 0.9 for Fig. 6C.Hence, at high Cu2q concentration, a highalbuminrLDL ratio is not sufficient to ensure that theLDL is protected against Cu2q-induced oxidation. Infact, even when the concentration ratio was as high as

Ž .in the plasma, as in the experiments of Fig. 6Ž .oxidation may occur at very low albumin and LDL

w xconcentrations 39,40 .

4.3. LDL oxidation by AAPH

In an attempt to evaluate the possible involvementand the importance of mechanisms other thancopper-binding in the protective effect of BSA, wehave studied the kinetics of peroxidation induced bythe water soluble free radical generator AAPH. Under

Žthe conditions used in our experiments 0.2 mM.LDLq4 mM AAPH , the kinetics of oxidation was

Žsimilar to that observed at 9 mM copper above the.saturating copper concentration . Under the latter

conditions, 15 mM BSA drastically modified copperŽ .induced oxidation Fig. 2B . By contrast, the AAPH-

Ž .induced oxidation was affected much less Fig. 7 .Interestingly, the initial stages of oxidation were en-hanced by BSA, as could have been expected forTMP upon decreasing the rate of production of freeradicals. Nonetheless, the maximal rate of oxidationwas slower than in the absence of BSA. This mayalso result from reduction of the rate of production offree radicals, but a decrease in the rate of propagationis also possible.

The latter possibility is consistent with our previ-ous results regarding copper-induced oxidation upon

w xsequential exposure of LDL to copper ions 41 .ŽGiven the finding that oxidation products probably

.hydroperoxides formed in LDL particles enhance therate of oxidation in other LDL particles, it is likely

Fig. 7. Effect of albumin on AAPH-induced LDL-oxidation. AtŽ . Ž .time zero, AAPH 4 mM , was added to LDL 0.2 mM . The

solid line presents the time-course observed in the absence ofBSA, the broken line is the time-course observed in presence of

Ž .BSA 15 mM and the dotted line is the time-course of ab-sorbance in the presence of 0.2 wt% polyethylene glycol 20000dalton.

that binding of such products to albumin contributesto its inhibitory effect. Yet, the alternative explana-tion, based on the effect of BSA on the rate of freeradical production by AAPH, can not be ruled out.

In any event, although the rate of free radicalproduction by AAPH is affected by the viscosity ofthe medium, which BSA may increase, the effect ofBSA on AAPH-induced oxidation is not due toBSA-induced changes in the medium viscosity. Thisis evident from the kinetic profiles given in Fig. 7,which show that increasing the viscosity to 1.048

Ž .mPaPs cp by 0.2 wt% PEG of a molecular weightof 20 000 dalton had smaller effects than those of 15mM BSA, whose effect on the viscosity was much

Ž .smaller hs1.012 mPaPs .

5. Conclusions

In an attempt to gain better understanding of theapparently controversial issue of the effect of albu-


min on copper-induced LDL-oxidation, we have stud-ied the kinetics of oxidation as a function of theconcentrations of LDL, copper and albumin. Basedon the results of these systematic studies we conclude

Ž .that: i Although the copper-albumin complex canpromote oxidation, its oxidation potency is much

Ž .lower than that of copper ions and ii Most, althoughnot all of the effects of albumin on the kinetics ofcopper-induced peroxidation result from copper-bind-ing.

All of our experimental data are consistent withthese conclusions. Specifically:1. Since the copper-albumin complex probably pro-

duces free radicals at a much slower rate than freecopper, the kinetics of LDL oxidation by thiscomplex is similar to that observed previously atlow concentrations of the free radical generator

w xAAPH 12–14 . Specifically, when the molar con-centration of copper is lower than that of albumin,Ž .i.e., all of the copper is bound to albumin ,oxidation occurs via a tocopherol mediated per-oxidation. Under these conditions, oxidation ispreceded by a short ‘lag’ and only at later stages,probably after all the vitamin E is consumed,non-inhibited oxidation takes place. This interpre-tation of the results accords with the enhancedinitial rate of oxidation obtained under such condi-tions in vitamin E-enriched LDL.

2. Since the oxidative potency of copper is muchhigher than that of the copper-albumin complex,as long as the system contains relatively high freecopper concentrations, the kinetics of oxidation isdetermined mostly by the copper-induced oxida-

Žtion i.e., oxidation occurs under conditions ofhigh radical flux and the kinetic profiles reveal noclear evidence for a tocopherol-mediated peroxida-

.tion . Under these conditions, the major determi-nant of the kinetics is the ratio between copper

Ž .bound to LDL and LDL C rnL , which of courseL

is an increasing function of the total concentrationof copper and a decreasing function of the albu-min concentration. Given the saturation depen-dence of C rnL on copper concentration, theL

oxidation may appear to be independent of albu-w xmin, as previously claimed by Deigner et al. 22

on the basis of their experiments with an almostsaturating copper concentration.

3. When the free copper concentration is lower than

its ‘saturating level’, the oxidation kinetics is gov-erned by the occupancy of copper binding sites on

Ž . ŽLDL C rnL . Thus, similar although not identi-L.cal kinetics profiles are observed for any given

value of this ratio. The effect of other factors suchas oxidation by the copper-albumin complex ismore pronounced at lower occupancies than athigher C rnL ratios. Other mechanisms may alsoL

be involved in the protective effect of albuminthus result in different kinetics under conditions ofa constant occupancy. One such mechanism maybe related to stimulation of oxidation in non-oxidized LDL particles by the hydroperoxidesformed in other LDL particles. Binding of hydro-peroxides to albumin may result in reduction ofthe overall oxidation process induced by eitherAAPH or copper ions. Nonetheless, even thoughsuch mechanisms may contribute to the protectiveeffect of albumin against copper-induced oxida-tion, their contribution is smaller than that medi-ated by copper-binding.

4. Since Cu2q-binding to albumin is much strongerthan to the binding sites responsible for LDLoxidation, the inhibitory effect of albumin is es-sentially a function of the albumin concentrationrather than of the BSArLDL ratio. This finding isof importance with respect to the possibility ofassaying the peroxidation in diluted non-fractionated plasma if the albumin concentration is

w xsufficiently low 39 .

Acknowledgements

We thank the Chief Scientist of the Israel Ministryof Health, the Ministry of Absorbtion and the Meer-baum Fund for financial support. Helpful discussionswith Dr. A. Hermetter are appreciated. Analysis ofvitamin E by Prof. Weissman’s Lab in Ichilov Hospi-tal is greatly appreciated.

References

w x Ž .1 Ross, R. 1986 N. Eng. J. Med. 314, 488–500.w x Ž .2 Ross, R. 1993 Nature 362, 801–809.w x3 Steinberg, D., Parthasarathy, S., Carew, T.E., Khoo, J.C.

Ž .and Witztum, J.L. 1989 N. Engl. J. Med. 320, 915–924.


w x Ž .4 Parthasarathy, S., Steinberg, D., and Witztum, J.L. 1992Annu. Rev. Med. 43, 219–225.

w x Ž .5 Steinbrecher, U.P., Zhang, H. and Lougheed, M. 1990Free Rad. Biol. Med. 9, 155–168.

w x Ž .6 Esterbauer, H., Gebicki, J., Puhl, H. and Jurgens, G. 1992Free Rad. Biol. Med. 13, 341–390.

w x Ž .7 Esterbauer, H. and Jurgens, G. 1993 Curr. Opin. Lipidol.4, 114–124.

w x8 Esterbauer, H., Striegel, G., Puhl, H. and Rotheneder, M.Ž .1989 Free Rad. Res. Commun. 6, 67–75.

w x9 Steinberg, D. and workshop participants, Antioxidants in theprevention of human atherosclerosis. Summary of the pro-ceedings of a National Heart, Lung, and Blood Institute

Ž .Workshop, September 5–6 1991, Bethesda, MD 1992Circulation 85, 2337–2344.

w x Ž .10 Esterbauer, H. 1992 Nutr. Metab. Cardiovasc. Dis. 2,55–57.

w x Ž .11 Halliwell, B. 1995 Am. J. Clin. Nutr. 61, 670S–677S.w x Ž .12 Bowry, V.W. and Stocker, R. 1993 J. Am. Chem. Soc.

115, 6029–6044.w x Ž .13 Stocker, R. 1994 Curr. Opin. Lipidol. 5, 422–433.w x Ž .14 Witting, P.K., Bowry, V.W. and Stocker, R. 1995 FEBS

Lett. 375, 45–49.w x Ž .15 Waldeck, A.R. and Stocker, R. 1996 Chem. Res. Toxicol.

9, 954–964.w x Ž .16 Abuja, P.M. and Esterbauer, H. 1995 Chem. Res. Toxicol.

8, 753–763.w x17 Kontush, A., Meyer, S., Finckh, B., Kohlschutter, A. and

Ž .Beisiegel, U. 1996 J. Biol. Chem. 271, 11106–11112.w x18 Maiorino, M., Zamburlini, A., Roveri, A. and Ursini, F.

Ž .1993 FEBS Lett. 330, 174–176.w x Ž .19 Yoshida, Y., Tsuchiya, J. and Niki, E. 1994 Biochim.

Biophys. Acta 1200, 85–92.w x Ž .20 Thomas, C.E. 1992 Biochim. Biophys. Acta 1128, 50–57.w x Ž .21 Zawadzki, Z., Milne, R.W. and Marcel, Y.L. 1991 J. Lipid

Res. 32, 243–250.w x22 Deigner, H.P., Friedrich, E., Sinn, H. and Diesel, H.A.

Ž .1992 Free Rad. Res. Commun. 16, 239–246.w x Ž . Ž .23 Esterbauer, H. 1993 Am. J. Clin. Nutr. 57 Suppl , 779S–

786S.

w x Ž .24 Halliwell, B. and Chirico, S. 1993 Am. J. Clin. Nutr. 57Ž .Suppl , 715S–725S.

w x25 Masuoka, J., Hegenauer, J., Van Dyke, B.R. and Saltman, P.Ž .1933 J. Biol. Chem. 268, 21533–21537.

w x Ž .26 Zgirski, A and Frieden, E. 1990 J. Inorg. Biochem. 39,137–148.

w x27 Kuzuya, M., Yamada, K., Hayashi, T., Funaki, C., Naito,Ž .M., Asci, K. and Kuzuya, F. 1992 Biochim. Biophys. Acta

1123, 334–341.w x Ž .28 Gieseg, S.P. and Esterbauer, H. 1994 FEBS Lett. 343,

188–194.w x29 Lowry, O.H., Rosenbrough, N.J., Farr, A.L. and Randall,

Ž .R.J. 1951 J. Biol. Chem. 193, 265–270.w x30 Kleinveld, H.A., Hak-Lemmers, H.L.M., Stalenhoff, A.F.H.

Ž .and Demacker, P.N.M. 1992 Clin. Chem. 38, 2066–2072.w x31 Esterbauer, H., Dieber-Rotheneder, M., Striegel, G. and

Ž .Waeg, G. 1991 Am. J. Clin. Nutr. 53, 314S–321S.w x32 Esterbauer, H., Jurgens, G., Quehenberger, O., and Koller,

Ž .E. 1987 J. Lipid Res. 28, 495–509.w x Ž .33 Kritharides, L., Jessup, W., Gifford, J. and Dean, R. 1993

Anal. Biochem. 213, 79–89.w x Ž .34 Chan, H.W.S. and Levett, G. 1977 Lipids 12, 99–104.w x Ž .35 Pinchuk, I. and Lichtenberg, D. 1996 Free Rad. Res. 24,

351–360.w x Ž .36 Buege, J.A. and Aust, S.D. 1978 Methods Enzymol. 52,

302–310.w x Ž .37 Frei, B. and Gaziano, M. 1993 J. Lipid Res. 34, 2135–

2144.w x38 Dobrian, A., Mova, R., Simionescu, M. and Simionescu, N.

Ž .1993 Biochem. Biophys. Acta 1169, 12–24.w x39 Schnitzer E., Pinchuk, I., Fainaru, M., Schafer, Z. and

Ž .Lichtenberg, D. 1995 Biochem. Biophys. Res. Commun.216, 854–861.

w x40 Regnstrom, J., Strom, K., Moldeus, P. and Nilsson, J.Ž .1993 Free Rad Res. Commun. 19, 267–278.

w x Ž .41 Schnitzer, E., Fainaru, M. and Lichtenberg, D. 1995 FreeRad. Res. 23, 137–149.

Documents

The effect of albumin on copper-induced LDL oxidation