autoantibodies to oxidized low density lipoprotein: the relationship to low density lipoprotein...

ORIGINAL ARTICLES

Autoantibodies to Oxidized LowDensity Lipoprotein: the Relationship toLow Density Lipoprotein Fatty AcidComposition in DiabetesM.E. Griffin1, D. McInerney1, A. Fraser1, A.H. Johnson2, P.B. Collins2, D. Owens3,G.H. Tomkin1,3

1The Adelaide Hospital, Dublin, Ireland;2Department of Biochemistry, Royal College of Surgeons in Ireland,Dublin, Ireland and3Department of Clinical Medicine, Trinity College, Dublin, Ireland

Autoantibodies to oxidized low density lipoprotein have been shown to be an independentpredictor of the progression of carotid atherosclerosis. This study examines the relationshipbetween low density lipoprotein fatty acid composition and autoantibodies to bothmalondialdehyde-modified and copper-oxidized low density lipoprotein in non-diabeticpatients with (n = 17), and without (n = 18), definite evidence of previous myocardialinfarction. The third group were non-insulin-dependent diabetic patients with no evidenceof atherosclerosis (n = 15) and the fourth group were patients with non-insulin-dependentdiabetes (n = 17) who had definite evidence of previous myocardial infarction. Fatty acidswere measured by gas-liquid chromatography. Antibodies to malondialdehyde-modifiedlow density lipoprotein and copper-oxidized low density lipoprotein were determined byan ELISA method. Autoantibodies to copper-oxidized low density lipoprotein weresignificantly higher in the non-diabetic patients with heart disease when compared to anyother group (p , 0.05). Autoantibodies to malondialdehyde-modified low density lipoproteinwere significantly higher in the non-diabetic subjects with heart disease and in bothdiabetic groups compared to non-diabetic subjects without coronary heart disease(p , 0.05). Linoleic acid (%) in low density lipoprotein did not differ between groups butarachidonic acid (%) was significantly lower in both diabetic and non-diabetic patientswith coronary heart disease (p , 0.05). The diabetic patients with low antibodies had39.6 ± 2.2 % polyunsaturated fatty acids in their low density lipoprotein while diabeticpatients with high antibodies had 46.7 ± 1.2 % polyunsaturates in their low densitylipoprotein (p , 0.01). This study confirms the association between antibodies to oxidizedlow density lipoprotein and coronary heart disease and shows raised low density lipoproteinantibody levels in diabetic patients with and without demonstrable atherosclerosis. In thediabetic patients, those with high antibody levels had high polyunsaturated fatty acidlevels in their LDL suggesting a possible role for dietary intervention. 1997 by JohnWiley & Sons, Ltd.

Diabet. Med. 14: 741–747 (1997)

No of Figures: 2. No of Tables: 2. No of Refs: 34

KEY WORDS NIDDM Atherosclerosis Oxidized low density lipoproteinMalondialdehyde-modified low density lipoprotein Copper-oxidized lowdensity lipoprotein Oxidized low density lipoprotein antibodies Lowdensity lipoprotein fatty acids

Received 22 October 1996; revised 16 April 1997; accepted 16 April 1997

Introduction

The association between low density lipioprotein (LDL)cholesterol and atherosclerosis is well established. There

Abbreviations: LDL low density lipoprotein; CHD coronary heart is, however, controversy about the role of low densitydisease, MDA-LDL malondialdehyde-modified LDL, HbA1c haemo- lipoprotein oxidation in its pathogenesis. The effectivenessglobinA1c, (C-CAD-) no evidence of diabetes mellitus and no evidence of antioxidants in the protection against atherosclerosisof atherosclerosis, (C CAD+) no evidence of diabetes mellitus anddefinite evidence of previous myocardial infarction, (DM CAD-) NIDDM is the subject of current research in many centres.1patients with no evidence of atherosclerosis, (DM CAD+) NIDDM Oxidized LDL has a number of potentially atherogenicpatients with definite evidence of previous myocardial infarction properties such as uptake by the non-LDL receptor* Correspondence to: Professor G.H. Tomkin, 1, Fitzwilliam Square,Dublin 2, Ireland pathway leading to dysregulation of cellular cholesterol

741CCC 0742–3071/97/090741–07$17.50 1997 by John Wiley & Sons, Ltd. DIABETIC MEDICINE, 1997; 14: 741–747

ORIGINAL ARTICLESmetabolism.2 Oxidized LDL is antigenic, stimulates (n = 11): 8 of these subjects were controlled on diet

alone, 4 on sulphonylureas and 1 on sulphonylurea andautoantibody formation, and has been isolated fromatherosclerotic plaques.3 Although it has been postulated metformin, and 2 on metformin alone. The last group

(DM CAD+, n = 17) had definite evidence of previousthat the inflammatory reaction during atherosclerosismay have an immunological basis,4,5 the role of oxidized myocardial infarction assessed by a combination of

clinical evidence and ECG or angiogram and NIDDM.LDL antibodies as mediators of this reaction is at presentspeculative.6 Tests of in vivo LDL oxidation are imprecise Ten of these subjects were controlled with diet alone, 4

with sulphonylurea therapy, 2 on sulphonylurea andbut increased oxidation of LDL in vivo has beendemonstrated both in diabetes and coronary heart disease metformin, and 1 with metformin alone. None of the

subjects had evidence of neuropathy, retinpopathy or(CHD).7,8 These measurements have been supported byin vitro studies showing increased susceptibility to nephropathy. Current smokers, subjects on hypolipida-

emic agents, and subjects with abnormal thyroid or liveroxidation of LDL from the same groups.9,10 Recentlyantibodies to oxidized LDL have been shown to be function tests were excluded. The study was approved

by the Hospital Ethics Committee and all subjects gaveincreased in non-diabetic patients with progressive carotidartery stenosis.11 In non-insulin-dependent (NIDDM) informed consent.patients, antibodies to oxidized LDL were also elevatedcompared to non-diabetic subjects.12 The susceptibility

Carotid Ultrasonographyof LDL to oxidation in vitro depends on fatty acidcomposition.13 We have recently demonstrated a relation-

Carotid ultrasonography was used in addition to clinicalship between linoleic acid, the major unsaturated fattyhistory, resting ECG, and stress ECG to determine thatacid in LDL containing two double bonds, and suscepti-the patients were free of atheromatous disease. All scansbility of LDL to be oxidized in normocholesterolaemicwere performed by one consultant radiologist with adiabetic patients14 and have shown that a diet rich inspecial interest in the project, using a Toshiba duplexoleic acid could alter fatty acid composition15 and reducescanner. The common carotid was examined bilaterallysusceptibility of LDL to be oxidized.16 This led us to3 cm proximal to the bifurcation, with the patient lyingquestion the advisability of a high polyunsaturated fatsupine with the neck slightly extended. Intima-medialdiet for diabetic patients, a concept supported in non-thickness was examined according to the technique ofdiabetic patients by the findings of an increase of v6Pigmole et al.8 Subjects with an intima-media thicknesspolyunsaturated fatty acids in aortic plaques.17 Althoughof # 0.7 mm and no evidence of plaque were consideredit is difficult to measure LDL oxidation accurately in vivo,eligible for the study as disease-negative patients, pro-measurement of autoantibodies against malondialdehyde-vided resting and maximal stress ECG were also negative.modified LDL (MDA-LDL), generated during lipid peroxi-

dation, should reflect in vivo oxidation of LDL. Copperhas been utilized to induce LDL oxidation in vitro and

Methodsit is possible that copper oxidized LDL may detectantibodies to different epitopes from those which bind

Fasting lipids were determined after a 16 h fast byto MDA-modified LDL. The purpose of this study wasstandard laboratory methods. Glucose was analysed into define the relationship between the LDL fatty acidfasting serum using a Beckman manual glucose analysercomposition, antibodies to oxidized LDL and atheroscler-2. Haemoglobin A1c (HbA1c) was determined as aosis in diabetic and non-diabetic subjects.percentage of total haemoglobin by an enzyme linkedimmunoassay where the monoclonal antibody was spe-Subjects and Methodscific for HbA1c (laboratory range for normal = 2.5–4.9 %).

Subjects

LDL IsolationFour groups of patients were recruited from the outpatientdiabetic and cardiology clinics. The first group (C CAD−,n = 17) had no evidence of atherosclerosis on clinical Blood was collected following a 16 h fast and centrifuged

at 4 °C. The antioxidants ethylenediaminetetraacetic acidexamination, resting ECG or coronary angiogram and noevidence of diabetes mellitus by WHO criteria. The (1 g l−1) and butylated hydroxytoluene (4.4 mg l−1) were

added to the plasma. LDL was isolated from the plasmasecond group (C CAD+, n = 18) had definite evidence ofprevious myocardial infarction diagnosed by a combi- by sequential ultracentrifugation at a density range of

1.025–1.063 g ml−1.19 Protein was determined by anation of clinical evidence and ECG or angiogram andno evidence of diabetes mellitus by WHO criteria. The modification of the Lowry method.20 The LDL preparation

gave a single band on SDS gradient gel electrophoresisthird group (DM CAD−, n = 15) were NIDDM patientswith no evidence of atheroslcerosis on clinical examin- when stained with Coomassie blue. The protein was

identified as apo B100 using molecular weight markers.ation and coronary angiogram (n = 4) or maximal stressECG combined with normal carotid Doppler examination LDL was stored frozen at −20 °C for fatty acid analysis.

742 M.E. GRIFFIN ET AL.

Diabet. Med. 14: 741–747 (1997) 1997 by John Wiley & Sons, Ltd.

ORIGINAL ARTICLESreaction was stopped with 50 ml of 1 mol l−1 NaOH.Low Density Lipoprotein Fatty AcidsPlates were read at 405 nm using a Titertek MultiskanMCC/340 (EFLAB Helsinki, Finland). Antibody levelsHeptadecanoic acid (100 mg) was added as an internal

standard to LDL (1 g l−1 protein) and the lipids were were calculated as a ratio (absorbance with MDA-LDL/absorbance with native LDL). Antibodies to cardioli-extracted by a modification of the method of Folch et

al.21 The organic fraction was dried with anhydrous pin have been shown to cross-react with MDA-LDL24

so pooled serum from patients with systemic lupussodium sulphate. The samples were dried under nitrogenand transmethylated.22 Following transmethylation the erythematosus was included on all plates as a positive

control and used to evaluate the reproducibility of thefatty acids methyl esters were extracted into hexane,dried under nitrogen, and reconstituted in iso-octane assay. All antibody determinations in this study were

made on one day, using the same batch of modifiedimmediately prior to determination by gas liquid chroma-tography. The fatty acids were analysed in a Shimadzu LDL and the same batch of reagents. The greatest

sensitivity was obtained using serum dilutions of 1/8 andGC-14A gas chromatograph (Kyoto, Japan) equippedwith a capillary fused silica Permabond FFAP-DF-0.1 1/16 and the assay was linear around this range. Ratio

of MDA-LDL/native at 1/8 and 1/16 dilutions for control(25 m × 0.25 mm internal diameter) column (Durren,Germany). Individual fatty acid concentration was serum (n = 7) was 2.1 ± 0.4 and 1.1 ± 0.5, diabetic serum

(n = 7) 4.7 ± 0.9 and 2.3 ± 0.5, and for SLE serum (n = 7)expressed as a percentage of total fatty acids or asmg fatty acid g−1 LDL protein by relating fatty acid 11.1 ± 3.7 and 5.8 ± 1.9, respectively. The inter-assay

variation for the method was 17%. Antibody levels didconcentration to the internal standard. The intra- andinter-assay variations were 1.8 and 2.6 %, respectively. not vary when serum samples were stored at −20 °C for

up to 6 months.Antibodies to Low Density Lipoprotein

Statistical AnalysisMalondialdehyde-modified LDL (MDA-LDL) was pre-pared by incubating freshly isolated LDL with Results were expressed as mean ± standard deviation.

Statistical analysis was performed using the Student’s t-0.5 mmol l−1 MDA (obtained by acid hydrolysis of MDA-bisdimethyl-acetal) using 1 ml of MDA g−1 LDL protein.4 test and ANOVA analysis of variance. Non-parametric

results were shown as median and range and were logUnbound MDA was removed by dialysis. Approximately65 % of the «-amino group in the lysine residues transformed prior to analysis. Correlation coefficients

were determined by linear and multiple regressionwere derivatized as determined using trinitrobenezenesulphonic acid.23 Copper-oxidized LDL was prepared by analysis using Statworks software on an Apple Macintosh

computer. Values of p , 0.05 were considered to beincubating LDL (1 g l−1 protein) with copper sulphate(10 mmol l−1) for 16 h at 37 °C. This procedure led to statistically significant.modification of 53 % of lysine residues. LDL was dialysedfor 16 h in PBS to remove copper. Results

Antibodies to MDA-LDL and copper-oxidized LDLwere measured by an ELISA method.24 Microtitre plates Subject details are given in Table 1. The non-diabetic

patients with atheromatous disease (C CAD+) were sig-were coated with 50 ml of freshly isolated native LDL,MDA-LDL or copper-oxidized LDL at a concentration of nificantly older than the subjects in the other three

groups (p , 0.05). Total cholesterol, LDL cholesterol,5 mg l−1 in PBS containing EDTA 0.27 mmol l−1 and BHT20 ml l−1. Following a 16 h incubation at 4 °C each well and HDL cholesterol were not significantly different in

any group. Serum triglycerides were higher in diabeticwas washed four times with PBS containing 0.02 %NaN3, 0.05 % Tween 20, and 0.001 % aprotinin. The patients without atheromatous disease (DM CAD−) com-

pared to the other 3 groups (p , 0.05) but diabeticremaining bindings sites were blocked by incubating for2 h at room temperature using 2 % human serum albumin control was similar in the two diabetic groups.

Antibodies to the copper-oxidized LDL were signifi-in PBS and the wells were washed four times as above.Serum (50 ml per well) was added at dilutions of 1/8 cantly higher in the non-diabetic subjects with heart

disease when compared to any other groups [C CAD+]and 1/16 and incubated for 16 h at 4 °C. Followingwashing of the wells, alkaline phosphatase-conjugated 2.3 ± 0.4 vs [C CAD−] 1.5 ± 0.2, [DM CAD+] 1.5 ± 0.2

and [DM CAD−] 1.5 ± 0.2 (p , 0.05). Autoantibodies torabbit anti-human IgG (200 ml per well) (Jackson Immuno-Research, West Grove, PA, USA) was added at a dilution MDA-LDL were significantly higher in the non-diabetic

patients with CHD [C CAD+] at 3.5 ± 0.5 compared toof 1 in 1500 in PBS containing Tween (0.05 %) andhuman serum albumin (0.2 %) and incubated at room the non-diabetic subjects without CHD [DM−, CAD−]

(1.5 ± 0.1 p , 0.005) (Figure 1). Sixty-six per cent of thesetemperature for 1 h. Plates were washed and 200 ml ofp-nitrophenyl phosphate added to each well at a patients had antibody levels above the median value of

2.25, whereas only one control patient without CADconcentration of 1 g l−1 in 50 mmol l−1 carbonate buffercontaining 1 mmol l−1 MgCl (pH 9.8). Following a 30- had levels greater than 2.25. The diabetic patients without

CAD had significantly higher MDA-LDL antibodies atmin incubation in the dark at room temperature the

743AUTOANTIBODIES TO OXIDIZED LDL AND LDL FATTY ACID COMPOSITION

1997 by John Wiley & Sons, Ltd. Diabet. Med. 14: 741–747 (1997)

ORIGINAL ARTICLESTable 1. Subjects characteristics

Control Control Diabetic DiabeticCAD− CAD+ CAD− CAD+

Number 17 19 15 17M/F 7/10 12/7 7/8 10/7Total cholesterol (mmol l−1) 6.4 ± 0.3 5.6 ± 0.3 6.1 ± 0.2 5.8 ± 0.3LDL cholesterol (mmol l−1) 4.1 ± 0.2 3.8 ± 0.2 4.0 ± 0.2 3.9 ± 0.3HDL cholesterol (mmol l−1) 1.4 ± 0.1 1.1 ± 0.1 1.1 ± 0.1 1.0 ± 0.1Triglycerides (mmol l−1)

range 0.88–4.73 1.51–11 0.57–5.08 0.73–6.62median 1.8 2.2 1.62 1.79

Fasting glucose 4.9 ± 0.2 4.8 ± 0.2 8.6 ± 0.9 8.8 ± 0.8HbA1c (%) – – 6.4 ± 0.5 6.2 ± 0.4Age (years) 61 ± 2.3 70.5 ± 1.8a 64.4 ± 1.5 66 ± 1.5

Mean ± SD.ap , 0.05 different from other groups.

Figure 1. Antibodies to MDA-LDL were measured by an ELISAtechnique in the plasma of non-diabetic patients without Mand with CAD and in NIDDM patients without and with

Figure 2. Antibodies to MDA-LDL and to copper-oxidized LDLm CAD. Results are expressed as absorbance in the presenceofwere measured in serum by an ELISA technique. Results weremodified LDL/absorbance with native LDL. Both diabetic groupsexpressed as absorbance with modified LDL/absorbance withand the non-diabetic patients with CAD had significantly highernative LDL. There was a positive correlation between the twoantibody levels; *p , 0.05 compared to non-diabetic patientsmethods (r = 0.325, p , 0.05).without heart disease

2.25 ± 0.4 compared to the control subjects without not have atheromatous disease (p , 0.05). Differencesin individual fatty acids between the groups were similarCAD 1.4 ± 0.1. The mean MDA-LDL antibody level of

the diabetic patients with CAD was 24% higher at when fatty acids were compared as mg g−1 LDL protein.There was no correlation between antibodies to copper-2.8 ± 0.4 than that for diabetic patients without heart

disease and significantly higher than the control patients oxidized LDL and LDL polyunsaturated fatty acids inany of the groups. The percentage of polyunsaturatedwithout heart disease (p , 0.001). We found only a

weak correlation between antibodies to copper-oxidized fatty acid in LDL in the diabetic patients and controlsubjects with high antibodies to copper-oxidized LDLLDL and antibodies to MDA-LDL in the group as a

whole (r = 0.32, p , 0.01) (Figure 2). and those with low antibody levels was similar.Analysis using fatty acid content in mg mg−1 LDLThe fatty acid composition of the LDL is shown in

Table 2. The percentage of oleic acid (18:1) was highest protein demonstrated a significant positive correlationbetween MDA-LDL antibodies and the polyunsaturatedin the C CAD+ subjects (p , 0.01). Levels of oleic acid

were also higher in diabetic patients with heart disease fatty acids in the LDL (r = 0.82, p , 0.01) in the diabeticpatients but not in non-diabetic subjects. There wasthan in diabetic subjects without heart disease (p , 0.05).

Percentage linoleic acid (18:2) did not differ between no correlation between MDA-LDL antibodies and thepercentage of polyunsaturated fatty acids in LDL fromthe groups. Percentage arachidonic acid (20:4) was lower

in both the diabetic and non-diabetic CAD+ groups, non-diabetic or diabetic subjects.Examination of percentage of polyunsaturated fattycompared to diabetic and non-diabetic subjects who did



ORIGINAL ARTICLESTable 2. LDL fatty acid composition (% major fatty acids)

Control Control Diabetic DiabeticCAD− CAD+ CAD− CAD+

16:0 Palmitic acid 26.9 ± 2.0 25.4 ± 0.3 26.5 ± 2.1 24.7 ± 0.616:1 Palmitoleic acid 3.6 ± 0.4 4.2 ± 0.4 4.5 ± 0.9 3.2 ± 0.318:0 Stearic acid 10.2 ± 1.3a 6.7 ± 0.5a 8.5 ± 0.6 8.2 ± 0.718:1 Oleic acid 17.6 ± 1.1 22.4 ± 0.7b 16.4 ± 1.2 19.3 ± 0.9a

18:2 Linoleic acid 34.2 ± 2.4 33.4 ± 1.2 35 ± 1.9 36.6 ± 1.320:4 Arachidonic acid 7.3 ± 2.9 5.8 ± 0.2a 8.9 ± 1.4 6.6 ± 0.4a

Mean ± SD.ap , 0.05 different from CAD−;bp , 0.01 different from CAD− groups.

acids in non-diabetic subjects with MDA-LDL antibody modified to a greater extent than the copper oxidizedLDL. This may make it a more sensitive ligand forlevels below or above 2.25 demonstrated no significant

difference in mean levels of polyunsaturated fatty acids. detecting autoantibodies and could account for theincreased number of subjects reacting to MDA-LDLThe percentage of polyunsaturated fatty acids in LDL

from control subjects with antibody levels below 2.25 compared to copper-oxidized LDL. This is in keepingwith the findings of Salonen et al.11 in their study onwas 42.5 ± 1.4 and in LDL from those with antibody

levels greater than 2.25 was 40.7 ± 2.0 %. In the diabetic autoantibodies to oxidized LDL and progression ofatherosclerosis. They found that the titre of autoantibodiespatients with antibody levels less than 2.25, LDL polyun-

saturates were 39.6 ± 2.2 %, whereas the diabetic patients to MDA-LDL was an independent predictor of theprogression of atherosclerosis whereas there was nowith high antibody levels had mean LDL polyunsaturates

of 46.7 ± 1.2 % (p , 0.01). Mean HbA1c for diabetic significant difference in the titre to copper-oxidized LDLbetween patients and controls. They suggested that thispatients with antibodies . 2.25 was 6.3 ± 0.4 % and

those with antibodies , 2.25 was the same at HbA1c reflected a low density of immuno-dominant epitopes ofoxidized-LDL eliciting autoantibodies in his patients.6.3 ± 0.5 %.

This study demonstrates higher levels of autoantibodiesto copper-oxidized LDL in non-diabetic subjects withDiscussionCAD when compared to those without CAD. Examinationof this group using MDA-LDL again showed significantlyOxidative modification of LDL has been shown to occur

in vivo.25 Antigenic epitopes elicit the production of higher antibody levels compared with the other threegroups. In the diabetic patients, there was no significantanti-oxidized-LDL autoantibodies. We used two methods

to modify LDL in order to detect antibodies to oxidized difference in antibody levels to either epitope betweenpatients with and without CAD although the MDA-LDLLDL since the nature of the epitopes is unknown and

may be distributed among multiple apo B fragments. antibody mean level was 25 % higher in the CAD+patients. This may reflect the insensitivity of the detectionMDA-lysine is one of the major end-products of lipid

peroxidation and lysine-4-hydroxy-nonenal which is of atherosclerosis in our diabetic patients since in diabetesthe distribution of atherosclerosis tends to be moreformed during copper-induced LDL oxidation is

another.13 The presence of autoantibodies against these peripheral. Autoantibodies against oxidized LDL areassociated with atherosclerosis in non-diabetic patientsepitopes is an indicator that LDL oxidation took place

in vivo before the blood was taken. The use of a single in some studies.3,27 However other authors have beenunable to demonstrate an association,28,29 the most recentbatch of modified LDL and other reagents for the study

ensured that antigen variation did not influence the being a study from the Netherlands30 where there wasno significant difference in titres to MDA-modified LDLresults. The relatively low correlation between antibodies

to copper-oxidized LDL and MDA-LDL in our study when a group of patients with severe coronary stenosiswas compared to a hospital control group with minorsuggests that antibodies to different epitopes of oxidized

LDL are present in the circulation. However, all the stenosis and a healthy population. There is no goodindication as to why the literature is so controversial.subjects with raised copper antibodies also had raised

MDA antibodies. Mironova et al.26 isolated anti-oxidized Some workers have used MDA-LDL and some have usedcopper-oxidized LDL. The study by Mironova et al.26LDL antibodies by affinity chromatography using copper-

oxidized LDL crosslinked sepharose. These antibodies might suggest that the cross-reactivity of antibodies maybe the reason. In our study we found only a weakwere predominantly IgG antibodies of low affinity which

were specific for copper-oxidized LDL but did cross- correlation between the two methods (r = 0.32, p , 0.01).Bellomo et al.12 reported autoantibodies against oxidat-react to some extent with MDA-modified LDL suggesting

that the recognition sites are not entirely specific for a ively modified LDL in NIDDM. They found that IgGantibody titres against MDA-LDL were significantly higherparticular epitope. In our study the MDA-LDL was



ORIGINAL ARTICLESin NIDDM patients—results similar to ours. They also patients with coronary artery disease. In the diabetic

patients there was a significant relationship with LDLexamined antibodies against copper-oxidized glycatedLDL and again found raised antibodies in the diabetic fatty acid composition. The study suggests that a reduction

in polyunsaturated fatty acids in the diet might decreasepatients. Their method utilizes in vitro glycation of LDLprior to oxidation, which is not comparable to our LDL oxidation and consequently antibodies to oxidized

LDL which may reduce the autoimmune atherogenicsystem in which we oxidized non-glycated LDL fromnon-diabetic subjects. Our results suggest that perhaps factor.copper-oxidized LDL, in the absence of glycation, is nota sensitive test in diabetic patients.

The purpose of this study was to examine the Acknowledgementrelationship between LDL fatty acid composition andantibodies to oxidized LDL since it is the polyunsaturated The authors wish to thank R. Conroy for statistical advice.fatty acids in LDL which are particularly susceptible tooxidation. Analysis of the relationship between LDL fattyacids and MDA-LDL antibodies showed that diabetic Referencespatients with low antibody levels had significantly lowerpolyunsaturates in their LDL. Examination of copper- 1. Steinberg D. Clinical trials of antioxidants in atheroscler-

osis: are we doing the right thing? Lancet 1995; 346: 36–38.oxidized LDL antibodies showed higher levels in C CAD+2. Steinberg D, Parthasarathy S, Carew TE, Khoo JC, Witztumpatients compared to the other groups, but levels were

JL. Beyond cholesterol: modifications of LDL whichnot higher in the DM CAD+ patients. The spread ofincrease its atherogenicity. N Engl J Med 1989; 320:

diabetic control in our patients was small so that the 915–923.lack of correlation between diabetic control and MDA- 3. Yla-Herttuala, Palinski W, Rosenfeld ME, Parthasarathy S,

Carew TE, Butler S, Witztum JL, Steinberg D. EvidenceLDL antibodies is not surprising. Bellomo et al.12

for the presence of oxidatively modified low densityinvestigated the possible role of glycation in sensitizinglipoprotein in atherosclerotic lesions of rabbit and man.LDL to oxidation. The found an increase in IgG antibodiesJ Clin Invest 1989; 84: 1086–1095.

to both glycated and glyco-oxidized LDL in NIDDM 4. Yla-Herttiala S, Palinski W, Butler SW, Picard S, Steinbergpatients. D, Witztum JL. Rabbit and human atherosclerotic lesions

contain IgG that recognises epitopes of oxidised LDL.The relationship between oxidized LDL antibodies andArterioscler Thromb 1994; 14: 32–40.LDL size was recently investigated by Jansen et al.31 in

5. Orekhov AN, Tertov VV, Kabakov AE, Yu I, Pokrovskynon-diabetic patients. These authors subdivided theirSN, Smirnov VN. Autoantibodies against modified lowpatients into LDL subclass B or A. They demonstrated density lipoprotein: nonlipid factor of blood plasma that

that small LDL was associated with higher antibody stimulates foam cell formation. Arterioscler Thromb 1991;11: 316–326.levels. They did not examine LDL composition. It

6. Hansson GK. Immunological markers of atheroslcerosis.has been shown in both non-diabetic and diabeticLancet 1993; 341: 278.subjects32,33 that small dense LDL contains more choles-

7. Velazquez E, Winocour PH, Kesteven P, Alberti KGMM,teryl esters and triglyceride. It has also been shown that Laker MF. Relation of lipid peroxides to macrovascularmore diabetic patients have small dense LDL.34 Our disease in Type 2 diabetes. Diabetic Med 1991; 8:

752–758.previous findings of an increase in the percentage of8. Nishigaki I, Hagihara M, Tsunekawa H, Maseki M, Yagilinoleic acid in the cholesteryl ester fraction of diabetic

K. Lipid peroxide levels of serum lipoproteins of diabeticLDL, which was associated with increased susceptibilitypatients. Biochem Med 1991; 25: 378–383.

of LDL to oxidation, gives some support to the concept 9. Babiy AV, Gebicki JM, Sullivan DR, Willey K. Increasedof an association between oxidised LDL antibodies and oxidisability of plasma lipoproteins in diabetic patients

can be decreased by probucol therapy and is not due toLDL composition. We found a correlation between theglycation. Biochem Pharmacol 1992; 43: 995–1000.amount of polyunsaturated fatty acids in LDL and MDA-

10. BowieA, Owens D, Collins P, Johnson A, Tomkin GH.LDL antibodies only in the diabetic patients. The reasonGlycosylated low density lipoprotein is more sensitive to

for this is not clear but it is possible that glycation of oxidation: implications for the diabetic patient? Athero-the particle might influence the facility for oxidation, sclerosis 1993; 102: 63–67.

11. Salonen JT, Yla-Herttuala S, Yamamoto R, Butler S,whereas in the non-diabetic the free radicle stress mayKorpela H, Salonen R, Nyyssonen K, Palinski W, Witztumbe more the determining factor. This was a cross-JL. Autoantibody against oxidised LDL and the progressionsectional study. It is not known how long LDL antibodiesof atherosclerosis. Lancet 1992; 339: 883–887.remain in the circulation and what effect alteration in 12. Bellomo G, Maggi E, Poli M, Agosta FG, Bollati P, Finardi

diet has on antibody formation. The finding of significantly G. Autoantibodies against oxidatively modified low densitylipoproteins in NIDDM. Diabetes 1995; 44: 60–66.higher percentage of polyunsaturated fatty acids in the

13. Esterbauer H, Jurgens G, Quenhenberger O, Koller E.diabetic patients with high antibodies does suggest thatAutooxidation of human low density lipoprotein: loss ofit may be useful to reduce polyunsaturated fatty acidspolyunsaturated fatty acids and vitamin E and generation

in the diet. of aldehydes. J Lipid Res 1987; 28: 495–509.In conclusion the study demonstrated an increase in 14. Dimitriadis E, Griffin M, Owens D, Johnson A, Collins P,

Tomkin GH. Oxidation of low density lipoprotein inantibodies to oxidized LDL in diabetic and non-diabetic



ORIGINAL ARTICLESNIDDM: its relationship to fatty acid composition. Diab- oxidised low density lipoprotein and to cardiolipin in

systemic lupus erythmatoses. Lancet 1993; 341: 923–925.etologia 1995; 38: 1300–1306.15. Griffin M, Dimitriadis E, Lenehan K, Collins P, Johnson 25. Palinski W, Rosenfeld ME, Yla-Herttuala S, Gurtner GC,

Socher SS, Butler S, et al. Low density lipoproteinA, Owens D, Tomkin GH. Dietary monounsaturated fattyacids alter low density lipoprotein composition and undergoes oxidative modification in vivo. Proc Natl Acad

Sci 1989; 88: 1372–1378.function in Type 2 diabetes mellitus. Q J Med 1996; 89:211–216. 26. Mironova M, Virella G, Lopes-Virella MF. Isolation and

characterisation of human antioxidised LDL autoanti-16. Dimitriadis E, Griffin M, Owens D, Johnson A, Collins P,Tomkin GH. Lipoprotein composition in NIDDM. The bodies. Arterioscler Thromb Vasc Biol 1996; 16: 222–229.

27. Bergmark C, Wu R, de Faire U, Lefvert AK, Swedenborgeffect of dietary oleic acid on composition and oxidisabilityand function of low and high density lipoproteins. J. Patients with early-onset peripheral vascular disease

have increased levels of autoantibodies against oxidisedDiabtologia 1996; 39: 667–676.17. Felton CV, Crook D, Davies MJ, Oliver MF. Dietary LDL. Arterioscler Thromb Vasc Biol 1995; 15: 441–445.

28. Schumacher M, Eber B Tatzber F, Kaufman P, Esterbauerpolyunsaturated fatty acids and composition of humanplaques. Lancet 1994; 344:1195–1196. H, Klein W. LDL oxidation and coronary atherosclerosis.

Lancet 1992; 4340; 123.18. Pigmole P, Tremoli E, Poli A, Oreste P, Paoletti R. Intimalplus medial thickness of the arterial wall: a direct 29. Virella G, Virella I, Leman RB, Lopes-Virella MF. Antioxi-

dised low density lipoprotein antibodies in patients withmeasurement with ultrasound imaging. Circulation 1986;6: 1399–1406. coronary heart disease and normal healthy volunteers.

Int J Clin Lab Res 1993; 23: 95–102.19. Havel RA, Eder HM, Brayden JH. The distributionand chemical composition of ultracentrifugally separated 30. van de Vijver LPJ, Steyger R, van Poppel G, Boer JMA,

Kruijssen DA, Seidell JC, Princen MG. Autoantibodieslipoproteins in human serum. J Clin Invest 1955; 34:1345–1353. against MDA-LDL in subjects with severe and minor

atherosclerosis and healthy populations controls. Athero-20. Markwell MAK, Haas SM, Bieber LL, Telbert NE. Amodification of the Lowry procedure to simplify protein sclerosis 1996;122: 245–253.

31. Jansen H, Ghanem H, Kuypers JH, Birkenhager JC.determination in membrane and lipoprotein samples.Anal Biochem 1978; 87: 206–210. Antibodies against malondialdehyde-modified LDL are

elevated in subjects with an LDL subclass B pattern.21. Folch J, Lees M, Sloane Stanley GH. A simple methodfor the isolation and purification of total lipids from Atherosclerosis 1995; 255–262.

32. McNamara JR, Jenner JL, Li Z, Wilson PWF, Schaefer EJ.animal tissues. J Biol Chem 1957; 226: 497–509.22. Sattler W, Puhl H, Kostner GM, Esterbauer H. Determi- Change in LDL particle size is associated with change in

plasma triglyceride concentration. Arterioscler Thrombnation of fatty acids in the main lipoprotein classesby capillary gas chromatography: BF3/Methanol trans- 1992; 12: 1284–1290.

33. Latdenpera S, Syvanne M, Kahri J, Taskinen M-R. Regu-esterification of lyophilised samples instead of Folchextraction gives higher yields. Anal Biochem 1991; 198: lation of low density lipoprotein particle size and distri-

bution in NIDDM and coronary artery disease: importance184–190.23. Habeeb AFSA Determination of free amino groups in of serum triglycerides. Diabetologia 1996; 39: 453–461.

34. Feingold KR, Grunfeld C, Pang M, Doerrler W, Kraussproteins by trinitrobenzenesulfonic acid. Anal Biochem1966; 14: 328–334. RM. LDL subclass phenotypes and triglyceride metabolism

in non-insulin-dependent diabetes. Arterioscler Thromb24. Vaarala O, Alfthan G, Jauhiainen M, Leirisalo-Repo, AhoK, Palosuo T. Cross reaction between antibodies to 1992; 12:1496–1502.



autoantibodies to oxidized low density lipoprotein: the relationship to low density lipoprotein...

Documents