- Email: [email protected]
Plasma levels of antibodies against oxidized LDL are inherited but not associated with HDL-cholesterol level in families with early coronary heart disease
Atherosclerosis 224 (2012) 123e128
Contents lists available at SciVerse ScienceDirect
Atherosclerosis journal homepage: www.elsevier.com/locate/atherosclerosis
Plasma levels of antibodies against oxidized LDL are inherited but not associated with HDL-cholesterol level in families with early coronary heart disease Timo Paavola, Tiia Kangas-Kontio, Tuire Salonurmi, Sanna Kuusisto, Tuija Huusko, Markku J. Savolainen, Sakari Kakko* Department of Internal Medicine, Biocenter Oulu and Clinical Research Center, University of Oulu, Aapistie 5A, 90220 Oulu, Finland
a r t i c l e i n f o
a b s t r a c t
Article history: Received 2 February 2012 Received in revised form 7 June 2012 Accepted 12 June 2012 Available online 28 June 2012
Objective: Oxidized low-density lipoproteins (oxLDL) and antibodies against them (anti-oxLDLs) are thought to play a central role in atherosclerosis. One proposed antiatherosclerotic mechanism for HDL is to prevent oxidation of LDL. This study examined whether plasma HDL-cholesterol (HDL-C) is related to plasma anti-oxLDL levels. Methods: We collected families based on probands with low HDL-C and premature coronary heart disease (CHD). Antibody levels were determined in samples from 405 subjects. Immunoglobulin G, M and A levels against two in vitro models of oxLDL, malondialdehyde-acetaldehyde-modified LDL (MAALDL) and copper oxidized LDL (CuOx-LDL), were measured by ELISA. We carried out heritability estimation of antibody traits and bivariate analyses between HDL-C, LDL-C and antibody traits. Results: All the antibody levels were significantly inherited (p < 0.001), heritability estimates ranging from 0.28 to 0.65. HDL-C exhibited no environmental or genetic cross-correlations with antibody levels. Significant environmental correlations were detected between LDL-C and both IgG levels (rE ¼ 0.40, p ¼ 0.046 and rE ¼ 0.39, p < 0.001). There were no differences in antibody levels between subjects with normal and low HDL-C, or between CHD-affected and non-affected subjects. Conclusion: In this study, low HDL-C level displayed no significant associations with the anti-oxLDL levels measured. The heritability of the anti-oxLDL levels was a novel and interesting finding. Ó 2012 Elsevier Ireland Ltd. All rights reserved.
Keywords: Immunology Lipoproteins Immunoglobulin isotypes Coronary artery disease Genetics Pedigree Bivariate analysis Genetic cross-correlation Environmental cross-correlation
1. Introduction Modified lipoproteins, including oxidized low-density lipoproteins (oxLDL), and humoural immunity against these compounds are thought to play a central role in the initiation and propagation of atherosclerosis [1]. Plasma levels of antibodies against oxLDL (anti-oxLDL) have been associated with atherosclerosis, but these associations have been found to vary in different studies e.g. depending on the antibody type and the model of atherosclerosis being studied. Plasma levels of IgG-type anti-oxLDL (IgG-oxLDL) have been associated positively [2], inversely [3] or not at all with atherosclerosis [4,5]. High plasma levels of IgM-oxLDL have been considered to protect against atherosclerosis in most studies [2,3]. The role of IgA-oxLDL in atherosclerosis has not been studied so far. Low plasma HDL-cholesterol (HDL-C) level is an independent risk factor for atherosclerosis [6] and it is the most common
* Corresponding author. P.O. Box 5000, FIN-90014 Oulu, Finland. Tel.: þ358 8 315 4570; fax: þ358 8 315 4543. E-mail address: sakari.kakko@oulu.fi (S. Kakko). 0021-9150/$ e see front matter Ó 2012 Elsevier Ireland Ltd. All rights reserved. http://dx.doi.org/10.1016/j.atherosclerosis.2012.06.056
dyslipidemia associated with premature and familial coronary heart disease (CHD) [7]. The major antiatherogenic mechanism of HDL-particles is thought to be reverse cholesterol transport, but HDL-particles are also anti-oxidative and anti-inflammatory [8e10]. Few studies have focused on the relationship between plasma levels of HDL-C and anti-oxLDL, although it could be speculated that high HDL-C levels could confer protection against oxidation of LDL and therefore be associated with a reduced oxLDL burden and lower anti-oxLDL levels. Plasma HDL-C has been shown to be inversely related to oxLDL-levels [11]. Low plasma HDL-C levels have been linked with higher total anti-oxLDL [12]. An inverse association between HDL-C and total anti-oxLDL binding to malondialdehyde-modified LDL (MDA-LDL) has been reported [13]. The level of IgM-oxLDL has been found to be positively correlated with HDL-C [14], whereas the level of IgG-MDA-LDL correlated inversely with HDL-C [4]. In addition to HDL-C, anti-oxLDL levels could be associated with other well-known risk factors of atherosclerosis such as a high lowdensity lipoprotein cholesterol (LDL-C) level [4]. Statins, drugs which cause an extensive reduction of plasma LDL-C levels, have been claimed to affect anti-oxLDL levels. Atorvastatin, fluvastatin
124
T. Paavola et al. / Atherosclerosis 224 (2012) 123e128
and pravastin have been shown to decrease the levels of autoantibodies [15e17]. Little is known about the genetic regulation of plasma antioxLDL levels, and the heritabilities of anti-oxLDL levels have not been studied earlier. IgG- and IgM-oxLDL have been reported to be associated with a single nucleotide polymorphism (SNP) in the lipoprotein lipase gene (LPL) and IgM-oxLDL with an SNP in the opiate-like receptor 1 gene (OLR1) [18]. In diabetic subjects, antioxLDL levels have been associated with polymorphisms of peroxisome proliferator-activated receptor gamma2 gene (PPARg2) [19]. The purpose of this study was to investigate whether anti-oxLDL levels are linked to the phenotype of low plasma HDL-C levels predisposing to early CHD. Two in vitro models of oxLDL, namely malondialdehyde-acetaldehyde-modified LDL (MAA-LDL) and copper oxidized LDL (CuOx-LDL), were studied in Finnish families with low HDL-C levels and early onset CHD. The heritabilities of anti-oxLDL levels were studied and in addition, bivariate analysis between anti-oxLDL levels and HDL-C or LDL-C was carried out to examine whether they share a common genetic or environmental background. 2. Materials and methods 2.1. Subjects Probands with premature CHD (i.e. acute myocardial infarction, coronary artery bypass graft operation or percutaneous transluminal coronary angioplasty before the age of 55 years), low HDL-C levels (<1.1 mmol/l) and normal to moderately elevated levels of triglycerides (<3.5 mmol/l) and total cholesterol (<7.0 mmol/l), no diabetes and an entry about family history of CHD in the hospital records were selected from the Oulu University Hospital as described earlier [20]. All the relatives of the proband (independent of their CHD status and HDL-C levels) who were willing to participate were examined and extended pedigrees were recruited to ascertain the genetic background. There were 39 pedigrees with three generations on average (minimum 2 and maximum 5) and with an average pedigrees size of 19 subjects (minimum 5 and maximum 89). A total number of 405 samples (39 families) were available for antibody measurements. From CHD-patients, lipid measurements were taken before or at least three months after myocardial infarction or coronary bypass operation. Information about smoking (recoded as current smoker or not, as pack-years of smoking), average alcohol consumption (grams of alcohol/week), medication and past medical history of the subjects was obtained by using a questionnaire. The study material was collected at 1990s before the era of intensive statin treatment and therefore only fortyeight study subjects were being treated with a statin; simvastatin at a daily dose of 10 mg was used by 50% of statin users and the most effective statin in use was atorvastatin at a daily dose of only 10 mg. Aspirin was used by 106 subjects, beta blocker by 125 subjects, an ACE inhibitor by 44 subjects, an angiotensin II receptor blocker by 5 subjects, a calcium channel blocker by 47 subjects, nitrates by 35 subjects, a hormone replacement therapy by 13 subjects and vitamin supplements by 41 subjects.Informed consent was obtained from all the study subjects. The ethical committee of the Oulu University Hospital approved this study. 2.2. Lipid and lipoprotein measurements Blood samples were obtained after an overnight fast, and plasma was separated by centrifugation at 800 g for 10 min and kept at þ4 C until further analysis. From plasma, VLDLfraction (d < 1.006 g/ml) was separated by ultracentrifugation
in a Kontron TFT 45.6 rotor at 105,000 g for 18 h. Thereafter, the other apoB-containing lipoproteins were precipitated by adding 25 ml of 2.8% (weight/volume) heparin and 25 ml of 2 M MgCl2 into 1 ml of VLDL-free fraction. After precipitation and centrifugation at 1000 g and þ4 C for 30 min, the HDL-C concentration of the supernatant was measured. The plasma LDL-C concentration was calculated by subtracting the cholesterol concentration of the HDL fraction from that of the VLDLfree fraction. The plasma concentrations of cholesterol and triglycerides and the lipoprotein fractions were measured by enzymatic colorimetric methods (Boehringer Mannheim GmbH, Germany) using a Kone Specific Analyser (Kone Instruments, Espoo, Finland). 2.3. Antibody measurements The experimental models of oxidized LDL used in this study were CuOx-LDL and MAA-LDL. MAA-LDL and CuOx-LDL were generated and tested as previously described [21]. LDL for antigen preparation was isolated from three healthy male subjects using no medication. The antigens were from a single batch for each antibody type (IgG-MAA-LDL, IgG-CuOx-LDL etc.). ELISA method was used to measure the levels of antibodies binding to modified LDL from 405 subjects as previously described [3,22]. Briefly, plates were coated with 5 ug/ml of CuOxLDL and MAA-LDL, in PBS-EDTA (PBS with 0.27 mmol/l EDTA), for overnight at þ4 C in white microfluor plates (Thermo electron Corp., Milford, MA, USA) and the PBS-washed plates blocked with PBSeEDTA containing 0.5% gelatin - protein from cold water fish skin (PBS-EDTA-FG) for 1 h at room temperature (RT). To determine the total IgG, IgM and IgA against CuOx-LDL and MAA-LDL, the plasma samples were added on PBS-washed plates and incubated for 1 h at RT. The secondary antibody (alkaline phosphatase-labelled goat anti-human (ALPanti-human) IgG, IgM or IgA (SigmaeAldrich, Saint Louis, Missouri, MO, USA)) diluted into TBS-FG-buffer, was added on washed plates, and incubated for 1 h at RT. As the substrate for the ALP-antihuman, a volume of 25 uL of 30% water solution of LumiPhos 530 (Lumigen, Inc., Southfield, MI, USA) was used. Luminescence was measured after 60 min dark incubation by Victor2 Luminometer (Wallac, PerkineElmer, Boston, MA, USA). The plates were washed three times with PBSeEDTA with an automated plate washer after every incubation before Lumiphos. For IgG determinations plasma samples were diluted into PBS-EDTA-FG as follows: 1:2000 (MAALDL and CuOx-LDL), 1:4000 (MAA-LDL high samples) or 1:8000 (CuOx-LDL high samples). For IgM determinations, 1:1500 (MAALDL), 1:4000 (MAA-LDL high samples), 1 : 2500 (CuOx-LDL) or 1:7000/10,000/500/5000 (CuOx-LDL high and low samples) dilutions, and for IgA determinations, 1:4000 (MAA-LDL) or 1:3000 (CuOx-LDL) dilutions were used. For each plasma sample, triplicate determinations were performed and an average relative luminescence unit (RLU) value was calculated. Each plate contained a triplicate standard of purified immunoglobulin, a zero-sample of pure PBS-EDTA-FG, and two triplicate control samples (‘high’ and ‘low’), which were diluted to cover as wide a range of the standard as possible. Linear equation for the standard curve was determined before the calculation of relative plasma antibody concentrations of the samples. The relative concentration was divided by 1000 and it is expressed as relative units (RU). To detect variation in an assay and between all the assays of an antibody type, intra-assay and inter-assay coefficients of variation (CV) were calculated, respectively, using control samples. The inter-assay coefficients of variation (CVs) were 15% or below in all assay series. The intraassay CVs were 15% or below, except in the following cases: IgG-MAA-LDL high (control) 23.1% and low (control) 17.5% (in one
T. Paavola et al. / Atherosclerosis 224 (2012) 123e128
out of three assays), and IgM-CuOx-LDL high 17.1% and low 28.4% (in two different assays out of two assays).
125
3. Results 3.1. Baseline characteristics, antibody levels, sex and age
2.4. Statistical analyses All the statistical analyses (except for heritability estimation and bivariate analyses) were performed using the SPSS Statistics version 17.0 (SPSS, Inc., Chicago, Illinois). Statistical significance was defined as p < 0.05. Evident outlier values were removed (n 7 in each antibody type) and skewed variables were normalised using log-transformation for parametric methods. In order to assess linear associations between continuous variables and antibody levels, partial Pearson correlation analysis was used. To study associations between antibody levels and binomial variables, the logistic binary regression model (forward stepwise regression) was used. Age and sex were related to some antibody levels as specified in Section 3.1 . All analyses were conducted either separately for both sexes or gender was taken as a factor into the partial correlation analysis, logistic binary regression model and heritability analysis as specified in the text. Age was taken as a covariate into the partial correlation analysis, logistic binary regression model and heritability analysis as specified in the text. Statin medicated subjects or subjects without reliable information about medication were excluded from these analyses, except when studying the effect of statin medication. Of these subjects using statins and excluded, 19 were probands. Quantitative genetic models were based on the decomposition of the phenotypic value of an individual to genetic and environmental components [23]. In the simplest model, phenotypic variance can be described as: s2p ¼ s2g þ s2e , where s2p, s2g and s2e are phenotypic, genetic and environmental variance, respectively. Heritability is the proportion of the total phenotypic variance that is explained by genetic factors. Heritability calculations require knowledge of the degree of relatedness between individuals in the study, and the family structure gives us kinship coefficient, which is used in the polygenic modelling to describe the expected genetic sharing between two relatives. The basic variance components approach has been extended to a multivariate framework in which the phenotype covariance can be further decomposed to include correlation between traits [24]. In the bivariate analysis, trait specific means, variance components relating to additive genetic effects, random environmental effects, covariate effects, as well as associated correlations; correlation caused by additive genetic effects and correlation caused by random environmental effects can be estimated simultaneously using maximum-likelihood estimates. Heritability estimation and bivariate quantitative genetic analysis were performed as described in our earlier article [25]. Solarsoftware [26] and the same pedigree data as in the previous article were used, except that only the 39 pedigrees with antibody levels measured could be included in the analyses. Gender- and age-adjusted heritabilities of all the antibody traits were computed in the whole study population, including statin medicated subjects. To investigate shared genetic contributions between antibodies, HDL-C and LDL-C, we carried out a bivariate analysis. Genetic and environmental correlations were estimated between pairs of the studied phenotypes, based on the maximum likelihood ratio and variance component decomposition. The bivariate analysis between antibodies and HDL-C was performed in the whole population adjusting for age and gender by subtracting a constant (6% of the mean HDL-C levels: 0.06 mmol/l for men and 0.09 mmol/l for women) from HDL-C values of the statin users as described previously [25]. The bivariate analysis between antibodies and LDL-C was performed without statin medicated subjects adjusting for age and gender, since the effect of statin on individual LDL-C level is variable and stronger than on HDL-C.
The study population without statin medicated subjects is presented in Table 1. The majority of the subjects were 30e70 years old, their mean age being 47.4 years for men and 48.9 years for women. The numbers of CHD-patients were 50 in men and 15 in women. The measured antibody concentrations are presented in Table 2. No differences in antibody concentrations were detected between subjects with low and normal plasma levels of HDL-C. Females had significantly higher levels of IgG-CuOx-LDL and IgM-oxLDLs than males (Table 2), but no significant differences were detected in other antibody types. The partial Pearson correlations adjusted for age and sex between different antibody types are presented in Table 3. Anti-MAA-LDL- and anti-CuOx-LDL levels exhibited strong positive correlations with each other in all antibody classes, with correlation coefficients ranging from 0.49 to 0.90. IgG- and IgA-levels correlated positively, with correlation coefficients ranging from 0.16 to 0.26, whereas IgM- and IgA-levels had an inverse association, with correlation coefficients ranging from 0.11 to 0.19. Age was positively correlated with IgG-CuOx-LDL and negatively with IgM-CuOx-LDL in females and positively with both IgAlevels in both sexes (Supplementary Table 1). 3.2. Heritabilities and genetic and environmental cross-correlations Heritabilities ranged from 0.28 in IgG-CuOx-LDL up to 0.65 in IgG-MAA-LDL and IgA-MAA-LDL (Table 4) in the whole study population including statin medicated subjects adjusted for sex and age. In the bivariate analysis, there were no genetic or environmental cross-correlations between HDL-C and any of the antibodies in the whole study population, when adjusted for sex, age and manually for statin medication (Table 4, see ‘Statistical analyses’). However, significant environmental correlations were detected between LDL-C and both IgG-MAA-LDL and IgG-CuOx-LDL (Table 4, rE ¼ 0.40, p ¼ 0.046 and rE ¼ 0.39, p < p ¼ 0.001, respectively). No other significant cross-correlations were detected between LDL-C and antibody levels. 3.3. CHD and antibody levels Clinical end points, CHD and early CHD (defined as affected by CHD before the age of 55 in males and the age of 65 in females) Table 1 Baseline characteristics in subjects without statin medication.
n Age, years CHD, n (%) Probands, n Early CHD, n (%) Age at first CHD manifestation, years Smokers, n (%) BMI, kg/m2 T-C, mmol/l HDL-C, mmol/l LDL-C, mmol/l T-TG, mmol/l
Males
Females
166 47.4 (40.3e54.5) 50 (30) 9 31 (19) 49.5 (43.5e55.5) 41 (25) 26.2 (23.9e28.6) 5.45 (4.78e6.13) 1.05 (0.86e1.24) 3.36 (2.76e3.97) 1.44 (0.95e1.94)
161 48.9 (38.2e59.7) 15 (9) 3 14 (9) 51.0 (43.0e59.0) 22 (14) 24.8 (21.7e27.9) 5.38 (4.72e6.04) 1.37 (1.14e1.61) 3.25 (2.66e3.85) 1.23 (0.88e1.58)
The data is expressed as median (interquartile range) or as number of subjects (percentage). Abbreviations: CHD, coronary heart disease; Early CHD, CHD before age 55/65 in males/females; T-C, total cholesterol; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; T-TG, total triglycerides; BMI, body mass index.
126
T. Paavola et al. / Atherosclerosis 224 (2012) 123e128
Table 2 Relative antibody concentrations (relative units, RU) in plasma in subjects without statin medication. Males n
All
IgG-MAA-LDL IgGeCuOx-LDL IgM-MAA-LDL IgMeCuOx-LDL IgA-MAA-LDL IgA-CuOx-LDL
73.6 93.4 1.7 3.1 4.1 10.7
Females HDL-C < 1.00 mmol/l
166
67 (54.8e92.5) (73.7e113.2)b (1.0e2.3)c (2.1e4.1)c (3.2e5.0) (8.0e13.4)
78.1 89.3 1.7 3.2 4.2 10.7
(55.6e100.5) (67.6e111.0) (1.1e2.4) (2.2e4.2) (3.2e5.2) (7.9e13.5)
HDL-C 1.00 mmol/l
All
98
161
72.9 93.8 1.6 3.1 4.0 10.7
(53.9e91.9) (74.7e112.9) (1.0e2.3) (2.0e4.1) (3.1e4.9) (7.9e13.5)
80.8 105.2 2.2 4.2 3.6 9.8
(64.6e97.0) (86.0e124.5)b (1.5e3.0)c (3.1e5.3)c (2.3e4.8) (6.9e12.7)
HDL-C < 1.30 mmol/l
HDL-C 1.30 mmol/l
65
92
80.4 108.7 2.5 4.4 3.9 10.3
(59.8e101.1) (87.4e130.0) (1.7e3.2) (3.3e5.6) (2.5e5.2) (6.8e13.7)
80.8 103.2 2.1 3.9 3.3 9.4
(67.0e94.6) (84.1e122.3) (1.5e2.8) (2.8e5.0) (2.1 e4.6) (7.1e11.8)
The data is expressed as median (interquartile range). Abbreviations: MAA-LDL, malondialdehyde-acetaldehyde-modified low density lipoprotein; CuOx-LDL, copper oxidized low density lipoprotein; IgG-MAA-LDL, IgG binding to MAA-LDL; HDL-C, high-density lipoprotein cholesterol. Statistical significance of differences is assessed by logistic binary regression model: sex is predicted by age and antibody level; HDL-C group is predicted by age and antibody level. P-value of an antibody term predicting sex in a statistically significant model: b p < 0.01, c p < 0.001.
were analysed in the whole study population excluding statin medicated subjects in the sexes separately (Supplementary Table 2). Comparisons of the antibody levels between affected and non-affected subjects were adjusted firstly only for age and secondly for age, HDL-C, LDL-C, BMI and smoking status. None of the differences in these end point variables reached statistical significance.
LDL-, IgA-MAA-LDL- and IgA-CuOx-LDL levels differed between non-smoking and smoking female subjects (medians of relative plasma antibody concentrations in non-smoking and smoking females, respectively: 4.1 and 4.5, p ¼ 0.045; 3.7 and 2.6, p ¼ 0.048; 10.1 and 6.9, p ¼ 0.004, data not shown).
3.4. Statin medication and antibody levels
This study investigated the relationship between plasma antioxLDL levels and the phenotype of low plasma HDL-C levels predisposing to early onset CHD. Our hypothesis was that in these patients, HDL-C would be inversely associated with the anti-oxLDL levels, since HDL-particles are important in inhibiting the oxidation of LDL-particles by paraoxonase-mediated anti-oxidative mechanisms [8,9]. Our results indicated that there was no association between the plasma levels of anti-oxLDLs and HDL-C and that antioxLDL levels did not associate with CHD or early CHD. Furthermore, no genetic or environmental correlation was found between antioxLDL levels and HDL-C levels. However, this does not conflict with some of the previous studies, where the associations between anti-oxLDL- and HDL level have been relatively modest [2,5] and also negative results have been published [27,28]. The main significant finding in this work was that in our population the antioxLDL levels were significantly inherited. The main limitation of this study was our relatively small study population. However, our homogenous and unique study sample
Subjects with and without statin medication were compared adjusting for age in the sexes separately. Statin medicated males had lower IgA-MAA-LDL (4.1 vs. 3.5, p ¼ 0.003) and IgA-CuOx-LDL (10.7 vs. 8.8, p < 0.001) levels than those male subjects without statin medication (Supplementary Table 3). No other significant differences were detected in antibody levels. 3.5. Other CHD-risk factors: LDL-C, triglycerides, BMI, cigarette smoking and antibody levels To detect potential confounding factors in our data set, partial Pearson correlations were computed in the whole study population without statin medicated subjects adjusted for age in the sexes separately (Supplementary Table 1). Only a few, relatively weak, significant correlations were detected, which validated our approach to adjust only for sex and age in most of our analyses: there was a positive correlation between LDL-C and IgG-CuOx-LDL (p ¼ 0.048) and a negative correlation between BMI and IgM-CuOxLDL (p ¼ 0.012) in females. Logistic binary regression model was used to predict smoking status by age and antibody level in the whole study population without statin medicated subjects in sexes separately. IgM-CuOx-
4. Discussion
Table 4 Heritabilities (H2) of the studied traits adjusted for sex and age in the whole study population including subjects receiving statin therapy (n ¼ 397). Genetic (rG ) and environmental (rE) cross-correlation coefficients between HDL-C and antibodiesc and between LDL-C and antibodiesd are shown. Trait
Table 3 Partial Pearson correlation coefficients (r) between antibodies in subjects without statin medication (n ¼ 327) adjusted for age and sex.
IgG-MAA-LDL IgG-CuOx-LDL IgM-MAA-LDL IgM-CuOx-LDL IgA-MAA-LDL
IgG-CuOxLDL
IgM-MAALDL
IgM-CuOxLDL
IgA-MAALDL
IgA-CuOxLDL
0.49c e e e e
0.04 0.07 e e e
0.08 0.08 0.76c e e
0.16b 0.19b 0.13a 0.06 e
0.08 0.26c 0.19b 0.11a 0.90c
Abbreviations: MAA-LDL, malondialdehyde-acetaldehyde-modified low density lipoprotein; CuOx-LDL, copper oxidized low density lipoprotein; IgG-MAA-LDL, IgG binding to MAA-LDL. a p < 0.05. b p < 0.01. c p < 0.001.
IgG-MAA-LDL IgG-CuOx-LDL IgM-MAA-LDL IgM-CuOx-LDL IgA-MAA-LDL IgA-CuOx-LDL
H2 SE
0.65 0.28 0.38 0.53 0.65 0.63
0.10b 0.09b 0.10b 0.11b 0.12b 0.11b
rE
rG
HDL-C
LDL-C
HDL-C
LDL-C
0.15 0.03 0.07 0.14 0.07 0.14
0.40a 0.39b 0.10 0.21 0.20 0.28
0.25 0.36 0.01 0.04 0.04 0.04
0.28 0.42 0.37 0.46 0.26 0.29
Abbreviations: MAA-LDL, malondialdehyde-acetaldehyde-modified low density lipoprotein; CuOx-LDL, copper oxidized low density lipoprotein; IgG-MAA-LDL, IgG binding to MAA-LDL; SE, standard error; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol. a p < 0.05. b p < 0.001. c Adjusted for sex and age in the whole population, statin medicated included. HDL-C levels were manually adjusted for statin medication (see ‘Statistical analyses’). d Adjusted for sex and age in subjects without statin medication.
T. Paavola et al. / Atherosclerosis 224 (2012) 123e128
includes a prominent sample of subjects with low HDL-C predisposing to early atherosclerosis without too many subjects receiving confounding statin medication. In addition our study material was concentrated on subjects with low HDL-C levels predisposing them to early CHD and if there were to be a mechanistic link between low HDL-C, anti-oxLDLs and CHD, it would be likely to be identified in this material. The CHD patients in our material had manifest CHD and it is possible that the association between anti-oxLDL, HDL and degree of atherosclerosis is different in subjects with early or no atherosclerosis and this association cannot be found in patients with advanced atherosclerosis. However, most of our study subjects did not have manifest CHD. The strong heritability of the plasma anti-oxLDL levels was a novel and surprising finding without any previous publication. It can be speculated that the antibody response is not only dependent on the amount of antigens in blood and other tissues, but could also reflect a partly inherited individual capacity to produce antibodies. This view is also supported by some previous studies, since significant heritabilities have been reported in antibody responses against hepatitis B virus surface antigen [29] and against Plasmodium falciparum antigens [30]. Previously, anti-oxLDL levels have also been shown to be associated with SNPs at LPL- and OLR1-genes [18], and with polymorphisms of PPARg2 [19]. The anti-oxLDL levels had no common genetic background with HDL-C or LDL-C in this study. In our material, the additive genetic heritability of HDL-C was 43% [20]. Males using statin medication had significantly lower IgAMAA-LDL- and IgA-CuOx-LDL levels than males without statin medication but no differences in other antibody types were observed. In the REVERSAL study, the IgM-MDA-LDL level was significantly decreased after 18 months of statin therapy [17]. Another study claimed that atorvastatin therapy could prevent IgG- and IgM formation against all forms of oxLDL after acute coronary syndrome without ST-elevation [16]. Both prava- and fluvastatin have been shown to decrease the IgG-MDA-LDL level [15]. It should be noted that our study population used several different statins. The role of IgA-oxLDL antibodies in this issue and their reactivity to individual statins remain to be clarified in future studies. The anti-oxLDL levels in our material displayed few associations with other traditional risk factors of CHD, i.e. high LDL-C, high BMI and smoking. There are also previous studies indicating that antibodies have only minor or no associations with total cholesterol [4,13], LDL-C [2,13], BMI [27] or smoking [13]. The lack of associations with factors which have been linked with increased oxidation of LDL-particles, such as low HDL-C [11,27], high LDL-C [18,31], smoking [31], and CHD [32], is evident in this study. Therefore, one must consider whether there is any positive association between the amount of oxLDL in arterial intima and the levels of free anti-oxLDLs in plasma. Indeed, some studies have indicated there is no cross-sectional positive correlation between anti-oxLDL- and oxLDL levels in plasma [27,31]. Authors suggest that immunocomplexes may explain this observation in part, since circulating complexes may exit the circulation more rapidly. However, a temporal association of some form is likely to exist [16]. Thus, it is possible that the levels of free anti-oxLDLs are affected by the cumulated oxLDL burden of the whole lifetime, but significant confounding factors, i.e. cross-reactions and the possibly partly inherited individual capacity to produce antibodies, as shown in our material, may reduce their role as precise markers of lifetime oxLDL exposure at the population level. Environmental cross-correlations between LDL-C and IgG-type antibody levels were significant and it could be speculated that IgG-type antibody levels are related to the life-time ox-LDL burden to a greater extent than the other antibody classes [33].
127
As we have shown, different antibody classes may associate inversely with each other, though the concentrations of antibodies against MAA-LDL and CuOx-LDL epitopes were directly associated. It is interesting that the levels of both IgG-oxLDL and IgA-oxLDL antibodies related inversely to the potentially atheroprotective IgM-oxLDL antibodies. This same reciprocality may be present in the genetic and environmental cross-correlations with LDL-C, though only two of the correlations reached statistical significance. These findings could imply that plasma IgG-oxLDL and IgAoxLDL- versus IgM-oxLDL levels were differently regulated. 5. Conclusion In conclusion, this study suggests that the phenotype of low HDL-C predisposing to early CHD displayed no significant associations with the plasma anti-oxLDL levels. We confirmed that there are no environmental or genetic cross-correlations between HDL-C and the anti-oxLDL levels as studied by bivariate analysis. Furthermore, there were no significant differences between the anti-oxLDL levels in CHD-affected subjects compared with nonaffected subjects. The heritability of the anti-oxLDL levels is a novel and interesting finding, which needs to be confirmed in future studies. Acknowledgements We are indebted to Doctor Sohvi Hörkkö for her great help in the antibody measurements. We also acknowledge the excellent technical assistance by Ms. Marja-Leena Kytökangas, Ms. Saara Korhonen, Ms. Sari Pyrhönen, Ms. Sirpa Rannikko, Ms. Marja Veneskoski, PhD student Maria Laitinen and Doctor Antti Nissinen. Doctors Jukka Juvonen, Tuomo Jääskeläinen, Erkki Kiviniemi and Mikko Lehtola (deceased) are greatly acknowledged for their help in collecting the pedigrees. This work has been supported by the Academy of Finland Research Funding and the Academy of Finland SALVE programme for 2009e2012, the Finnish Cardiovascular Research Foundation, the Finnish Cultural Foundation, the Finnish Foundation for Alcohol Studies, the Finnish Medical Foundation, the Finnish Medical Society Duodecim, the Paavo Nurmi Foundation, and the Sigrid Jusélius Foundation. The funding sources had no role in the collection, analysis and interpretation of data; in the writing of the report; and in the decision to submit the article for publication. The authors disclose no actual or potential conflicts of interests. Appendix A. Supplementary data Supplementary data related to this article can be found online at http://dx.doi.org/10.1016/j.atherosclerosis.2012.06.056. References [1] Virella G, Lopes-Virella MF. Atherogenesis and the humoral immune response to modified lipoproteins. Atherosclerosis 2008;200:239e46. [2] Tsimikas S, Brilakis ES, Lennon RJ, et al. Relationship of IgG and IgM autoantibodies to oxidized low density lipoprotein with coronary artery disease and cardiovascular events. J Lipid Res 2007;48:425e33. [3] Karvonen J, Päivänsalo M, Kesäniemi YA, Hörkkö S. Immunoglobulin M type of autoantibodies to oxidized low-density lipoprotein has an inverse relation to carotid artery atherosclerosis. Circulation 2003;108:2107e12. [4] Rossi GP, Cesari M, De Toni R, et al. Antibodies to oxidized low-density lipoproteins and angiographically assessed coronary artery disease in white patients. Circulation 2003;108:2467e72. [5] Wilson PW, Ben-Yehuda O, McNamara J, Massaro J, Witztum J, Reaven PD. Autoantibodies to oxidized LDL and cardiovascular risk: the Framingham offspring study. Atherosclerosis 2006;189:364e8. [6] Gordon DJ, Rifkind BM. High-density lipoproteinethe clinical implications of recent studies. N Engl J Med 1989;321:1311e6.
128
T. Paavola et al. / Atherosclerosis 224 (2012) 123e128
[7] Genest Jr JJ, Martin-Munley SS, McNamara JR, et al. Familial lipoprotein disorders in patients with premature coronary artery disease. Circulation 1992;85:2025e33. [8] Navab M, Hama SY, Cooke CJ, et al. Normal high density lipoprotein inhibits three steps in the formation of mildly oxidized low density lipoprotein: step 1. J Lipid Res 2000;41:1481e94. [9] Navab M, Hama SY, Anantharamaiah GM, et al. Normal high density lipoprotein inhibits three steps in the formation of mildly oxidized low density lipoprotein: steps 2 and 3. J Lipid Res 2000;41:1495e508. [10] Vaisar T, Pennathur S, Green PS, et al. Shotgun proteomics implicates protease inhibition and complement activation in the antiinflammatory properties of HDL. J Clin Invest 2007;117:746e56. [11] Lapointe A, Couillard C, Piche ME, et al. Circulating oxidized LDL is associated with parameters of the metabolic syndrome in postmenopausal women. Atherosclerosis 2007;191:362e8. [12] Lehtimäki T, Lehtinen S, Solakivi T, et al. Autoantibodies against oxidized low density lipoprotein in patients with angiographically verified coronary artery disease. Arterioscler Thromb Vasc Biol 1999;19:23e7. [13] van de Vijver LP, Steyger R, van Poppel G, et al. Autoantibodies against MDALDL in subjects with severe and minor atherosclerosis and healthy population controls. Atherosclerosis 1996;122:245e53. [14] van Bommel EF, van Tits LJ, van den Berg EA, Prins J, Stalenhoef AF. Autoantibodies against oxidized low-density lipoprotein and lipid profile in patients with chronic periaortitis: case-control study. Rheumatol Int 2011;31:201e8. [15] Zhang B, Noda K, Matsunaga A, Kumagai K, Saku K. A comparative crossover study of the effects of fluvastatin and pravastatin (FP-COS) on circulating autoantibodies to oxidized LDL in patients with hypercholesterolemia. J Atheroscler Thromb 2005;12:41e7. [16] Papathanasiou AI, Lourida ES, Tsironis LD, Goudevenos JA, Tselepis AD. Shortand long-term elevation of autoantibody titers against oxidized LDL in patients with acute coronary syndromes. Role of the lipoprotein-associated phospholipase A2 and the effect of atorvastatin treatment. Atherosclerosis 2008;196:289e97. [17] Choi SH, Chae A, Miller E, et al. Relationship between biomarkers of oxidized low-density lipoprotein, statin therapy, quantitative coronary angiography, and atheroma: volume observations from the REVERSAL (reversal of atherosclerosis with aggressive lipid lowering) study. J Am Coll Cardiol 2008;52:24e32. [18] Chen Q, Reis SE, Kammerer C, et al. Association of anti-oxidized LDL and candidate genes with severity of coronary stenosis in the Women’s ischemia syndrome evaluation study. J Lipid Res 2011;52:801e7. [19] Niskanen L, Lindi V, Erkkilä A, et al. Association of the PRO12ALA polymorphism of the PPAR-gamma2 gene with oxidized low-density lipoprotein and cardiolipin autoantibodies in nondiabetic and type 2 diabetic subjects. Metabolism 2003;52:213e7.
[20] Kangas-Kontio T, Kakko S, Tamminen M, et al. Genome scan for loci regulating HDL cholesterol levels in Finnish extended pedigrees with early coronary heart disease. Eur J Hum Genet 2010;18:604e13. [21] Veneskoski M, Turunen SP, Kummu O, et al. Specific recognition of malondialdehyde and malondialdehyde acetaldehyde adducts on oxidized LDL and apoptotic cells by complement anaphylatoxin C3a. Free Radic Biol Med 2011; 51:834e43. [22] Sämpi M, Ukkola O, Päivänsalo M, Kesäniemi YA, Hörkkö S. Early atherosclerosis and IgG2 to bacteria are associated with FcgammaRIIa genotype in non-smokers. Eur J Clin Invest 2009;39:517e26. [23] Amos CI. Robust variance-components approach for assessing genetic linkage in pedigrees. Am J Hum Genet 1994;54:535e43. [24] Almasy L, Dyer TD, Blangero J. Bivariate quantitative trait linkage analysis: pleiotropy versus co-incident linkages. Genet Epidemiol 1997;14: 953e8. [25] Kangas-Kontio T, Huotari A, Ruotsalainen H, et al. Genetic and environmental determinants of total and high-molecular weight adiponectin in families with low HDL-cholesterol and early onset coronary heart disease. Atherosclerosis 2010;210:479e85. [26] Almasy L, Blangero J. Multipoint quantitative-trait linkage analysis in general pedigrees. Am J Hum Genet 1998;62:1198e211. [27] Wang Y, Fang X, Wang S, Feng Y, Yin J. Relation between plasma oxLDL antibodies and oxLDL in the circulation. Inflammation 2007;30:7e13. [28] Fraley AE, Schwartz GG, Olsson AG, et al. Relationship of oxidized phospholipids and biomarkers of oxidized low-density lipoprotein with cardiovascular risk factors, inflammatory biomarkers, and effect of statin therapy in patients with acute coronary syndromes: results from the MIRACL (myocardial ischemia reduction with aggressive cholesterol lowering) trial. J Am Coll Cardiol 2009;53:2186e96. [29] Hohler T, Reuss E, Evers N, et al. Differential genetic determination of immune responsiveness to hepatitis B surface antigen and to hepatitis A virus: a vaccination study in twins. Lancet 2002;360:991e5. [30] Duah NO, Weiss HA, Jepson A, Tetteh KK, Whittle HC, Conway DJ. Heritability of antibody isotype and subclass responses to Plasmodium falciparum antigens. PLoS One 2009;4:e7381. [31] Chen HW, Kuo CL, Huang CS, Kuo SJ, Liu CS. Oxidized low-density lipoproteins, autoantibodies against oxidized low-density lipoproteins and carotid intima media thickness in a clinically healthy population. Cardiology 2008;110: 252e9. [32] Liu KZ, Cuddy TE, Pierce GN. Oxidative status of lipoproteins in coronary disease patients. Am Heart J 1992;123:285e90. [33] Khosravi A, Hommel M, Sayemiri K. Age-dependent antibody response to Plasmodium falciparum merozoite surface protein 2 (MSP-2). Parasite Immunol 2011;33:145e57.
Contents lists available at SciVerse ScienceDirect
Atherosclerosis journal homepage: www.elsevier.com/locate/atherosclerosis
Plasma levels of antibodies against oxidized LDL are inherited but not associated with HDL-cholesterol level in families with early coronary heart disease Timo Paavola, Tiia Kangas-Kontio, Tuire Salonurmi, Sanna Kuusisto, Tuija Huusko, Markku J. Savolainen, Sakari Kakko* Department of Internal Medicine, Biocenter Oulu and Clinical Research Center, University of Oulu, Aapistie 5A, 90220 Oulu, Finland
a r t i c l e i n f o
a b s t r a c t
Article history: Received 2 February 2012 Received in revised form 7 June 2012 Accepted 12 June 2012 Available online 28 June 2012
Objective: Oxidized low-density lipoproteins (oxLDL) and antibodies against them (anti-oxLDLs) are thought to play a central role in atherosclerosis. One proposed antiatherosclerotic mechanism for HDL is to prevent oxidation of LDL. This study examined whether plasma HDL-cholesterol (HDL-C) is related to plasma anti-oxLDL levels. Methods: We collected families based on probands with low HDL-C and premature coronary heart disease (CHD). Antibody levels were determined in samples from 405 subjects. Immunoglobulin G, M and A levels against two in vitro models of oxLDL, malondialdehyde-acetaldehyde-modified LDL (MAALDL) and copper oxidized LDL (CuOx-LDL), were measured by ELISA. We carried out heritability estimation of antibody traits and bivariate analyses between HDL-C, LDL-C and antibody traits. Results: All the antibody levels were significantly inherited (p < 0.001), heritability estimates ranging from 0.28 to 0.65. HDL-C exhibited no environmental or genetic cross-correlations with antibody levels. Significant environmental correlations were detected between LDL-C and both IgG levels (rE ¼ 0.40, p ¼ 0.046 and rE ¼ 0.39, p < 0.001). There were no differences in antibody levels between subjects with normal and low HDL-C, or between CHD-affected and non-affected subjects. Conclusion: In this study, low HDL-C level displayed no significant associations with the anti-oxLDL levels measured. The heritability of the anti-oxLDL levels was a novel and interesting finding. Ó 2012 Elsevier Ireland Ltd. All rights reserved.
Keywords: Immunology Lipoproteins Immunoglobulin isotypes Coronary artery disease Genetics Pedigree Bivariate analysis Genetic cross-correlation Environmental cross-correlation
1. Introduction Modified lipoproteins, including oxidized low-density lipoproteins (oxLDL), and humoural immunity against these compounds are thought to play a central role in the initiation and propagation of atherosclerosis [1]. Plasma levels of antibodies against oxLDL (anti-oxLDL) have been associated with atherosclerosis, but these associations have been found to vary in different studies e.g. depending on the antibody type and the model of atherosclerosis being studied. Plasma levels of IgG-type anti-oxLDL (IgG-oxLDL) have been associated positively [2], inversely [3] or not at all with atherosclerosis [4,5]. High plasma levels of IgM-oxLDL have been considered to protect against atherosclerosis in most studies [2,3]. The role of IgA-oxLDL in atherosclerosis has not been studied so far. Low plasma HDL-cholesterol (HDL-C) level is an independent risk factor for atherosclerosis [6] and it is the most common
* Corresponding author. P.O. Box 5000, FIN-90014 Oulu, Finland. Tel.: þ358 8 315 4570; fax: þ358 8 315 4543. E-mail address: sakari.kakko@oulu.fi (S. Kakko). 0021-9150/$ e see front matter Ó 2012 Elsevier Ireland Ltd. All rights reserved. http://dx.doi.org/10.1016/j.atherosclerosis.2012.06.056
dyslipidemia associated with premature and familial coronary heart disease (CHD) [7]. The major antiatherogenic mechanism of HDL-particles is thought to be reverse cholesterol transport, but HDL-particles are also anti-oxidative and anti-inflammatory [8e10]. Few studies have focused on the relationship between plasma levels of HDL-C and anti-oxLDL, although it could be speculated that high HDL-C levels could confer protection against oxidation of LDL and therefore be associated with a reduced oxLDL burden and lower anti-oxLDL levels. Plasma HDL-C has been shown to be inversely related to oxLDL-levels [11]. Low plasma HDL-C levels have been linked with higher total anti-oxLDL [12]. An inverse association between HDL-C and total anti-oxLDL binding to malondialdehyde-modified LDL (MDA-LDL) has been reported [13]. The level of IgM-oxLDL has been found to be positively correlated with HDL-C [14], whereas the level of IgG-MDA-LDL correlated inversely with HDL-C [4]. In addition to HDL-C, anti-oxLDL levels could be associated with other well-known risk factors of atherosclerosis such as a high lowdensity lipoprotein cholesterol (LDL-C) level [4]. Statins, drugs which cause an extensive reduction of plasma LDL-C levels, have been claimed to affect anti-oxLDL levels. Atorvastatin, fluvastatin
124
T. Paavola et al. / Atherosclerosis 224 (2012) 123e128
and pravastin have been shown to decrease the levels of autoantibodies [15e17]. Little is known about the genetic regulation of plasma antioxLDL levels, and the heritabilities of anti-oxLDL levels have not been studied earlier. IgG- and IgM-oxLDL have been reported to be associated with a single nucleotide polymorphism (SNP) in the lipoprotein lipase gene (LPL) and IgM-oxLDL with an SNP in the opiate-like receptor 1 gene (OLR1) [18]. In diabetic subjects, antioxLDL levels have been associated with polymorphisms of peroxisome proliferator-activated receptor gamma2 gene (PPARg2) [19]. The purpose of this study was to investigate whether anti-oxLDL levels are linked to the phenotype of low plasma HDL-C levels predisposing to early CHD. Two in vitro models of oxLDL, namely malondialdehyde-acetaldehyde-modified LDL (MAA-LDL) and copper oxidized LDL (CuOx-LDL), were studied in Finnish families with low HDL-C levels and early onset CHD. The heritabilities of anti-oxLDL levels were studied and in addition, bivariate analysis between anti-oxLDL levels and HDL-C or LDL-C was carried out to examine whether they share a common genetic or environmental background. 2. Materials and methods 2.1. Subjects Probands with premature CHD (i.e. acute myocardial infarction, coronary artery bypass graft operation or percutaneous transluminal coronary angioplasty before the age of 55 years), low HDL-C levels (<1.1 mmol/l) and normal to moderately elevated levels of triglycerides (<3.5 mmol/l) and total cholesterol (<7.0 mmol/l), no diabetes and an entry about family history of CHD in the hospital records were selected from the Oulu University Hospital as described earlier [20]. All the relatives of the proband (independent of their CHD status and HDL-C levels) who were willing to participate were examined and extended pedigrees were recruited to ascertain the genetic background. There were 39 pedigrees with three generations on average (minimum 2 and maximum 5) and with an average pedigrees size of 19 subjects (minimum 5 and maximum 89). A total number of 405 samples (39 families) were available for antibody measurements. From CHD-patients, lipid measurements were taken before or at least three months after myocardial infarction or coronary bypass operation. Information about smoking (recoded as current smoker or not, as pack-years of smoking), average alcohol consumption (grams of alcohol/week), medication and past medical history of the subjects was obtained by using a questionnaire. The study material was collected at 1990s before the era of intensive statin treatment and therefore only fortyeight study subjects were being treated with a statin; simvastatin at a daily dose of 10 mg was used by 50% of statin users and the most effective statin in use was atorvastatin at a daily dose of only 10 mg. Aspirin was used by 106 subjects, beta blocker by 125 subjects, an ACE inhibitor by 44 subjects, an angiotensin II receptor blocker by 5 subjects, a calcium channel blocker by 47 subjects, nitrates by 35 subjects, a hormone replacement therapy by 13 subjects and vitamin supplements by 41 subjects.Informed consent was obtained from all the study subjects. The ethical committee of the Oulu University Hospital approved this study. 2.2. Lipid and lipoprotein measurements Blood samples were obtained after an overnight fast, and plasma was separated by centrifugation at 800 g for 10 min and kept at þ4 C until further analysis. From plasma, VLDLfraction (d < 1.006 g/ml) was separated by ultracentrifugation
in a Kontron TFT 45.6 rotor at 105,000 g for 18 h. Thereafter, the other apoB-containing lipoproteins were precipitated by adding 25 ml of 2.8% (weight/volume) heparin and 25 ml of 2 M MgCl2 into 1 ml of VLDL-free fraction. After precipitation and centrifugation at 1000 g and þ4 C for 30 min, the HDL-C concentration of the supernatant was measured. The plasma LDL-C concentration was calculated by subtracting the cholesterol concentration of the HDL fraction from that of the VLDLfree fraction. The plasma concentrations of cholesterol and triglycerides and the lipoprotein fractions were measured by enzymatic colorimetric methods (Boehringer Mannheim GmbH, Germany) using a Kone Specific Analyser (Kone Instruments, Espoo, Finland). 2.3. Antibody measurements The experimental models of oxidized LDL used in this study were CuOx-LDL and MAA-LDL. MAA-LDL and CuOx-LDL were generated and tested as previously described [21]. LDL for antigen preparation was isolated from three healthy male subjects using no medication. The antigens were from a single batch for each antibody type (IgG-MAA-LDL, IgG-CuOx-LDL etc.). ELISA method was used to measure the levels of antibodies binding to modified LDL from 405 subjects as previously described [3,22]. Briefly, plates were coated with 5 ug/ml of CuOxLDL and MAA-LDL, in PBS-EDTA (PBS with 0.27 mmol/l EDTA), for overnight at þ4 C in white microfluor plates (Thermo electron Corp., Milford, MA, USA) and the PBS-washed plates blocked with PBSeEDTA containing 0.5% gelatin - protein from cold water fish skin (PBS-EDTA-FG) for 1 h at room temperature (RT). To determine the total IgG, IgM and IgA against CuOx-LDL and MAA-LDL, the plasma samples were added on PBS-washed plates and incubated for 1 h at RT. The secondary antibody (alkaline phosphatase-labelled goat anti-human (ALPanti-human) IgG, IgM or IgA (SigmaeAldrich, Saint Louis, Missouri, MO, USA)) diluted into TBS-FG-buffer, was added on washed plates, and incubated for 1 h at RT. As the substrate for the ALP-antihuman, a volume of 25 uL of 30% water solution of LumiPhos 530 (Lumigen, Inc., Southfield, MI, USA) was used. Luminescence was measured after 60 min dark incubation by Victor2 Luminometer (Wallac, PerkineElmer, Boston, MA, USA). The plates were washed three times with PBSeEDTA with an automated plate washer after every incubation before Lumiphos. For IgG determinations plasma samples were diluted into PBS-EDTA-FG as follows: 1:2000 (MAALDL and CuOx-LDL), 1:4000 (MAA-LDL high samples) or 1:8000 (CuOx-LDL high samples). For IgM determinations, 1:1500 (MAALDL), 1:4000 (MAA-LDL high samples), 1 : 2500 (CuOx-LDL) or 1:7000/10,000/500/5000 (CuOx-LDL high and low samples) dilutions, and for IgA determinations, 1:4000 (MAA-LDL) or 1:3000 (CuOx-LDL) dilutions were used. For each plasma sample, triplicate determinations were performed and an average relative luminescence unit (RLU) value was calculated. Each plate contained a triplicate standard of purified immunoglobulin, a zero-sample of pure PBS-EDTA-FG, and two triplicate control samples (‘high’ and ‘low’), which were diluted to cover as wide a range of the standard as possible. Linear equation for the standard curve was determined before the calculation of relative plasma antibody concentrations of the samples. The relative concentration was divided by 1000 and it is expressed as relative units (RU). To detect variation in an assay and between all the assays of an antibody type, intra-assay and inter-assay coefficients of variation (CV) were calculated, respectively, using control samples. The inter-assay coefficients of variation (CVs) were 15% or below in all assay series. The intraassay CVs were 15% or below, except in the following cases: IgG-MAA-LDL high (control) 23.1% and low (control) 17.5% (in one
T. Paavola et al. / Atherosclerosis 224 (2012) 123e128
out of three assays), and IgM-CuOx-LDL high 17.1% and low 28.4% (in two different assays out of two assays).
125
3. Results 3.1. Baseline characteristics, antibody levels, sex and age
2.4. Statistical analyses All the statistical analyses (except for heritability estimation and bivariate analyses) were performed using the SPSS Statistics version 17.0 (SPSS, Inc., Chicago, Illinois). Statistical significance was defined as p < 0.05. Evident outlier values were removed (n 7 in each antibody type) and skewed variables were normalised using log-transformation for parametric methods. In order to assess linear associations between continuous variables and antibody levels, partial Pearson correlation analysis was used. To study associations between antibody levels and binomial variables, the logistic binary regression model (forward stepwise regression) was used. Age and sex were related to some antibody levels as specified in Section 3.1 . All analyses were conducted either separately for both sexes or gender was taken as a factor into the partial correlation analysis, logistic binary regression model and heritability analysis as specified in the text. Age was taken as a covariate into the partial correlation analysis, logistic binary regression model and heritability analysis as specified in the text. Statin medicated subjects or subjects without reliable information about medication were excluded from these analyses, except when studying the effect of statin medication. Of these subjects using statins and excluded, 19 were probands. Quantitative genetic models were based on the decomposition of the phenotypic value of an individual to genetic and environmental components [23]. In the simplest model, phenotypic variance can be described as: s2p ¼ s2g þ s2e , where s2p, s2g and s2e are phenotypic, genetic and environmental variance, respectively. Heritability is the proportion of the total phenotypic variance that is explained by genetic factors. Heritability calculations require knowledge of the degree of relatedness between individuals in the study, and the family structure gives us kinship coefficient, which is used in the polygenic modelling to describe the expected genetic sharing between two relatives. The basic variance components approach has been extended to a multivariate framework in which the phenotype covariance can be further decomposed to include correlation between traits [24]. In the bivariate analysis, trait specific means, variance components relating to additive genetic effects, random environmental effects, covariate effects, as well as associated correlations; correlation caused by additive genetic effects and correlation caused by random environmental effects can be estimated simultaneously using maximum-likelihood estimates. Heritability estimation and bivariate quantitative genetic analysis were performed as described in our earlier article [25]. Solarsoftware [26] and the same pedigree data as in the previous article were used, except that only the 39 pedigrees with antibody levels measured could be included in the analyses. Gender- and age-adjusted heritabilities of all the antibody traits were computed in the whole study population, including statin medicated subjects. To investigate shared genetic contributions between antibodies, HDL-C and LDL-C, we carried out a bivariate analysis. Genetic and environmental correlations were estimated between pairs of the studied phenotypes, based on the maximum likelihood ratio and variance component decomposition. The bivariate analysis between antibodies and HDL-C was performed in the whole population adjusting for age and gender by subtracting a constant (6% of the mean HDL-C levels: 0.06 mmol/l for men and 0.09 mmol/l for women) from HDL-C values of the statin users as described previously [25]. The bivariate analysis between antibodies and LDL-C was performed without statin medicated subjects adjusting for age and gender, since the effect of statin on individual LDL-C level is variable and stronger than on HDL-C.
The study population without statin medicated subjects is presented in Table 1. The majority of the subjects were 30e70 years old, their mean age being 47.4 years for men and 48.9 years for women. The numbers of CHD-patients were 50 in men and 15 in women. The measured antibody concentrations are presented in Table 2. No differences in antibody concentrations were detected between subjects with low and normal plasma levels of HDL-C. Females had significantly higher levels of IgG-CuOx-LDL and IgM-oxLDLs than males (Table 2), but no significant differences were detected in other antibody types. The partial Pearson correlations adjusted for age and sex between different antibody types are presented in Table 3. Anti-MAA-LDL- and anti-CuOx-LDL levels exhibited strong positive correlations with each other in all antibody classes, with correlation coefficients ranging from 0.49 to 0.90. IgG- and IgA-levels correlated positively, with correlation coefficients ranging from 0.16 to 0.26, whereas IgM- and IgA-levels had an inverse association, with correlation coefficients ranging from 0.11 to 0.19. Age was positively correlated with IgG-CuOx-LDL and negatively with IgM-CuOx-LDL in females and positively with both IgAlevels in both sexes (Supplementary Table 1). 3.2. Heritabilities and genetic and environmental cross-correlations Heritabilities ranged from 0.28 in IgG-CuOx-LDL up to 0.65 in IgG-MAA-LDL and IgA-MAA-LDL (Table 4) in the whole study population including statin medicated subjects adjusted for sex and age. In the bivariate analysis, there were no genetic or environmental cross-correlations between HDL-C and any of the antibodies in the whole study population, when adjusted for sex, age and manually for statin medication (Table 4, see ‘Statistical analyses’). However, significant environmental correlations were detected between LDL-C and both IgG-MAA-LDL and IgG-CuOx-LDL (Table 4, rE ¼ 0.40, p ¼ 0.046 and rE ¼ 0.39, p < p ¼ 0.001, respectively). No other significant cross-correlations were detected between LDL-C and antibody levels. 3.3. CHD and antibody levels Clinical end points, CHD and early CHD (defined as affected by CHD before the age of 55 in males and the age of 65 in females) Table 1 Baseline characteristics in subjects without statin medication.
n Age, years CHD, n (%) Probands, n Early CHD, n (%) Age at first CHD manifestation, years Smokers, n (%) BMI, kg/m2 T-C, mmol/l HDL-C, mmol/l LDL-C, mmol/l T-TG, mmol/l
Males
Females
166 47.4 (40.3e54.5) 50 (30) 9 31 (19) 49.5 (43.5e55.5) 41 (25) 26.2 (23.9e28.6) 5.45 (4.78e6.13) 1.05 (0.86e1.24) 3.36 (2.76e3.97) 1.44 (0.95e1.94)
161 48.9 (38.2e59.7) 15 (9) 3 14 (9) 51.0 (43.0e59.0) 22 (14) 24.8 (21.7e27.9) 5.38 (4.72e6.04) 1.37 (1.14e1.61) 3.25 (2.66e3.85) 1.23 (0.88e1.58)
The data is expressed as median (interquartile range) or as number of subjects (percentage). Abbreviations: CHD, coronary heart disease; Early CHD, CHD before age 55/65 in males/females; T-C, total cholesterol; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; T-TG, total triglycerides; BMI, body mass index.
126
T. Paavola et al. / Atherosclerosis 224 (2012) 123e128
Table 2 Relative antibody concentrations (relative units, RU) in plasma in subjects without statin medication. Males n
All
IgG-MAA-LDL IgGeCuOx-LDL IgM-MAA-LDL IgMeCuOx-LDL IgA-MAA-LDL IgA-CuOx-LDL
73.6 93.4 1.7 3.1 4.1 10.7
Females HDL-C < 1.00 mmol/l
166
67 (54.8e92.5) (73.7e113.2)b (1.0e2.3)c (2.1e4.1)c (3.2e5.0) (8.0e13.4)
78.1 89.3 1.7 3.2 4.2 10.7
(55.6e100.5) (67.6e111.0) (1.1e2.4) (2.2e4.2) (3.2e5.2) (7.9e13.5)
HDL-C 1.00 mmol/l
All
98
161
72.9 93.8 1.6 3.1 4.0 10.7
(53.9e91.9) (74.7e112.9) (1.0e2.3) (2.0e4.1) (3.1e4.9) (7.9e13.5)
80.8 105.2 2.2 4.2 3.6 9.8
(64.6e97.0) (86.0e124.5)b (1.5e3.0)c (3.1e5.3)c (2.3e4.8) (6.9e12.7)
HDL-C < 1.30 mmol/l
HDL-C 1.30 mmol/l
65
92
80.4 108.7 2.5 4.4 3.9 10.3
(59.8e101.1) (87.4e130.0) (1.7e3.2) (3.3e5.6) (2.5e5.2) (6.8e13.7)
80.8 103.2 2.1 3.9 3.3 9.4
(67.0e94.6) (84.1e122.3) (1.5e2.8) (2.8e5.0) (2.1 e4.6) (7.1e11.8)
The data is expressed as median (interquartile range). Abbreviations: MAA-LDL, malondialdehyde-acetaldehyde-modified low density lipoprotein; CuOx-LDL, copper oxidized low density lipoprotein; IgG-MAA-LDL, IgG binding to MAA-LDL; HDL-C, high-density lipoprotein cholesterol. Statistical significance of differences is assessed by logistic binary regression model: sex is predicted by age and antibody level; HDL-C group is predicted by age and antibody level. P-value of an antibody term predicting sex in a statistically significant model: b p < 0.01, c p < 0.001.
were analysed in the whole study population excluding statin medicated subjects in the sexes separately (Supplementary Table 2). Comparisons of the antibody levels between affected and non-affected subjects were adjusted firstly only for age and secondly for age, HDL-C, LDL-C, BMI and smoking status. None of the differences in these end point variables reached statistical significance.
LDL-, IgA-MAA-LDL- and IgA-CuOx-LDL levels differed between non-smoking and smoking female subjects (medians of relative plasma antibody concentrations in non-smoking and smoking females, respectively: 4.1 and 4.5, p ¼ 0.045; 3.7 and 2.6, p ¼ 0.048; 10.1 and 6.9, p ¼ 0.004, data not shown).
3.4. Statin medication and antibody levels
This study investigated the relationship between plasma antioxLDL levels and the phenotype of low plasma HDL-C levels predisposing to early onset CHD. Our hypothesis was that in these patients, HDL-C would be inversely associated with the anti-oxLDL levels, since HDL-particles are important in inhibiting the oxidation of LDL-particles by paraoxonase-mediated anti-oxidative mechanisms [8,9]. Our results indicated that there was no association between the plasma levels of anti-oxLDLs and HDL-C and that antioxLDL levels did not associate with CHD or early CHD. Furthermore, no genetic or environmental correlation was found between antioxLDL levels and HDL-C levels. However, this does not conflict with some of the previous studies, where the associations between anti-oxLDL- and HDL level have been relatively modest [2,5] and also negative results have been published [27,28]. The main significant finding in this work was that in our population the antioxLDL levels were significantly inherited. The main limitation of this study was our relatively small study population. However, our homogenous and unique study sample
Subjects with and without statin medication were compared adjusting for age in the sexes separately. Statin medicated males had lower IgA-MAA-LDL (4.1 vs. 3.5, p ¼ 0.003) and IgA-CuOx-LDL (10.7 vs. 8.8, p < 0.001) levels than those male subjects without statin medication (Supplementary Table 3). No other significant differences were detected in antibody levels. 3.5. Other CHD-risk factors: LDL-C, triglycerides, BMI, cigarette smoking and antibody levels To detect potential confounding factors in our data set, partial Pearson correlations were computed in the whole study population without statin medicated subjects adjusted for age in the sexes separately (Supplementary Table 1). Only a few, relatively weak, significant correlations were detected, which validated our approach to adjust only for sex and age in most of our analyses: there was a positive correlation between LDL-C and IgG-CuOx-LDL (p ¼ 0.048) and a negative correlation between BMI and IgM-CuOxLDL (p ¼ 0.012) in females. Logistic binary regression model was used to predict smoking status by age and antibody level in the whole study population without statin medicated subjects in sexes separately. IgM-CuOx-
4. Discussion
Table 4 Heritabilities (H2) of the studied traits adjusted for sex and age in the whole study population including subjects receiving statin therapy (n ¼ 397). Genetic (rG ) and environmental (rE) cross-correlation coefficients between HDL-C and antibodiesc and between LDL-C and antibodiesd are shown. Trait
Table 3 Partial Pearson correlation coefficients (r) between antibodies in subjects without statin medication (n ¼ 327) adjusted for age and sex.
IgG-MAA-LDL IgG-CuOx-LDL IgM-MAA-LDL IgM-CuOx-LDL IgA-MAA-LDL
IgG-CuOxLDL
IgM-MAALDL
IgM-CuOxLDL
IgA-MAALDL
IgA-CuOxLDL
0.49c e e e e
0.04 0.07 e e e
0.08 0.08 0.76c e e
0.16b 0.19b 0.13a 0.06 e
0.08 0.26c 0.19b 0.11a 0.90c
Abbreviations: MAA-LDL, malondialdehyde-acetaldehyde-modified low density lipoprotein; CuOx-LDL, copper oxidized low density lipoprotein; IgG-MAA-LDL, IgG binding to MAA-LDL. a p < 0.05. b p < 0.01. c p < 0.001.
IgG-MAA-LDL IgG-CuOx-LDL IgM-MAA-LDL IgM-CuOx-LDL IgA-MAA-LDL IgA-CuOx-LDL
H2 SE
0.65 0.28 0.38 0.53 0.65 0.63
0.10b 0.09b 0.10b 0.11b 0.12b 0.11b
rE
rG
HDL-C
LDL-C
HDL-C
LDL-C
0.15 0.03 0.07 0.14 0.07 0.14
0.40a 0.39b 0.10 0.21 0.20 0.28
0.25 0.36 0.01 0.04 0.04 0.04
0.28 0.42 0.37 0.46 0.26 0.29
Abbreviations: MAA-LDL, malondialdehyde-acetaldehyde-modified low density lipoprotein; CuOx-LDL, copper oxidized low density lipoprotein; IgG-MAA-LDL, IgG binding to MAA-LDL; SE, standard error; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol. a p < 0.05. b p < 0.001. c Adjusted for sex and age in the whole population, statin medicated included. HDL-C levels were manually adjusted for statin medication (see ‘Statistical analyses’). d Adjusted for sex and age in subjects without statin medication.
T. Paavola et al. / Atherosclerosis 224 (2012) 123e128
includes a prominent sample of subjects with low HDL-C predisposing to early atherosclerosis without too many subjects receiving confounding statin medication. In addition our study material was concentrated on subjects with low HDL-C levels predisposing them to early CHD and if there were to be a mechanistic link between low HDL-C, anti-oxLDLs and CHD, it would be likely to be identified in this material. The CHD patients in our material had manifest CHD and it is possible that the association between anti-oxLDL, HDL and degree of atherosclerosis is different in subjects with early or no atherosclerosis and this association cannot be found in patients with advanced atherosclerosis. However, most of our study subjects did not have manifest CHD. The strong heritability of the plasma anti-oxLDL levels was a novel and surprising finding without any previous publication. It can be speculated that the antibody response is not only dependent on the amount of antigens in blood and other tissues, but could also reflect a partly inherited individual capacity to produce antibodies. This view is also supported by some previous studies, since significant heritabilities have been reported in antibody responses against hepatitis B virus surface antigen [29] and against Plasmodium falciparum antigens [30]. Previously, anti-oxLDL levels have also been shown to be associated with SNPs at LPL- and OLR1-genes [18], and with polymorphisms of PPARg2 [19]. The anti-oxLDL levels had no common genetic background with HDL-C or LDL-C in this study. In our material, the additive genetic heritability of HDL-C was 43% [20]. Males using statin medication had significantly lower IgAMAA-LDL- and IgA-CuOx-LDL levels than males without statin medication but no differences in other antibody types were observed. In the REVERSAL study, the IgM-MDA-LDL level was significantly decreased after 18 months of statin therapy [17]. Another study claimed that atorvastatin therapy could prevent IgG- and IgM formation against all forms of oxLDL after acute coronary syndrome without ST-elevation [16]. Both prava- and fluvastatin have been shown to decrease the IgG-MDA-LDL level [15]. It should be noted that our study population used several different statins. The role of IgA-oxLDL antibodies in this issue and their reactivity to individual statins remain to be clarified in future studies. The anti-oxLDL levels in our material displayed few associations with other traditional risk factors of CHD, i.e. high LDL-C, high BMI and smoking. There are also previous studies indicating that antibodies have only minor or no associations with total cholesterol [4,13], LDL-C [2,13], BMI [27] or smoking [13]. The lack of associations with factors which have been linked with increased oxidation of LDL-particles, such as low HDL-C [11,27], high LDL-C [18,31], smoking [31], and CHD [32], is evident in this study. Therefore, one must consider whether there is any positive association between the amount of oxLDL in arterial intima and the levels of free anti-oxLDLs in plasma. Indeed, some studies have indicated there is no cross-sectional positive correlation between anti-oxLDL- and oxLDL levels in plasma [27,31]. Authors suggest that immunocomplexes may explain this observation in part, since circulating complexes may exit the circulation more rapidly. However, a temporal association of some form is likely to exist [16]. Thus, it is possible that the levels of free anti-oxLDLs are affected by the cumulated oxLDL burden of the whole lifetime, but significant confounding factors, i.e. cross-reactions and the possibly partly inherited individual capacity to produce antibodies, as shown in our material, may reduce their role as precise markers of lifetime oxLDL exposure at the population level. Environmental cross-correlations between LDL-C and IgG-type antibody levels were significant and it could be speculated that IgG-type antibody levels are related to the life-time ox-LDL burden to a greater extent than the other antibody classes [33].
127
As we have shown, different antibody classes may associate inversely with each other, though the concentrations of antibodies against MAA-LDL and CuOx-LDL epitopes were directly associated. It is interesting that the levels of both IgG-oxLDL and IgA-oxLDL antibodies related inversely to the potentially atheroprotective IgM-oxLDL antibodies. This same reciprocality may be present in the genetic and environmental cross-correlations with LDL-C, though only two of the correlations reached statistical significance. These findings could imply that plasma IgG-oxLDL and IgAoxLDL- versus IgM-oxLDL levels were differently regulated. 5. Conclusion In conclusion, this study suggests that the phenotype of low HDL-C predisposing to early CHD displayed no significant associations with the plasma anti-oxLDL levels. We confirmed that there are no environmental or genetic cross-correlations between HDL-C and the anti-oxLDL levels as studied by bivariate analysis. Furthermore, there were no significant differences between the anti-oxLDL levels in CHD-affected subjects compared with nonaffected subjects. The heritability of the anti-oxLDL levels is a novel and interesting finding, which needs to be confirmed in future studies. Acknowledgements We are indebted to Doctor Sohvi Hörkkö for her great help in the antibody measurements. We also acknowledge the excellent technical assistance by Ms. Marja-Leena Kytökangas, Ms. Saara Korhonen, Ms. Sari Pyrhönen, Ms. Sirpa Rannikko, Ms. Marja Veneskoski, PhD student Maria Laitinen and Doctor Antti Nissinen. Doctors Jukka Juvonen, Tuomo Jääskeläinen, Erkki Kiviniemi and Mikko Lehtola (deceased) are greatly acknowledged for their help in collecting the pedigrees. This work has been supported by the Academy of Finland Research Funding and the Academy of Finland SALVE programme for 2009e2012, the Finnish Cardiovascular Research Foundation, the Finnish Cultural Foundation, the Finnish Foundation for Alcohol Studies, the Finnish Medical Foundation, the Finnish Medical Society Duodecim, the Paavo Nurmi Foundation, and the Sigrid Jusélius Foundation. The funding sources had no role in the collection, analysis and interpretation of data; in the writing of the report; and in the decision to submit the article for publication. The authors disclose no actual or potential conflicts of interests. Appendix A. Supplementary data Supplementary data related to this article can be found online at http://dx.doi.org/10.1016/j.atherosclerosis.2012.06.056. References [1] Virella G, Lopes-Virella MF. Atherogenesis and the humoral immune response to modified lipoproteins. Atherosclerosis 2008;200:239e46. [2] Tsimikas S, Brilakis ES, Lennon RJ, et al. Relationship of IgG and IgM autoantibodies to oxidized low density lipoprotein with coronary artery disease and cardiovascular events. J Lipid Res 2007;48:425e33. [3] Karvonen J, Päivänsalo M, Kesäniemi YA, Hörkkö S. Immunoglobulin M type of autoantibodies to oxidized low-density lipoprotein has an inverse relation to carotid artery atherosclerosis. Circulation 2003;108:2107e12. [4] Rossi GP, Cesari M, De Toni R, et al. Antibodies to oxidized low-density lipoproteins and angiographically assessed coronary artery disease in white patients. Circulation 2003;108:2467e72. [5] Wilson PW, Ben-Yehuda O, McNamara J, Massaro J, Witztum J, Reaven PD. Autoantibodies to oxidized LDL and cardiovascular risk: the Framingham offspring study. Atherosclerosis 2006;189:364e8. [6] Gordon DJ, Rifkind BM. High-density lipoproteinethe clinical implications of recent studies. N Engl J Med 1989;321:1311e6.
128
T. Paavola et al. / Atherosclerosis 224 (2012) 123e128
[7] Genest Jr JJ, Martin-Munley SS, McNamara JR, et al. Familial lipoprotein disorders in patients with premature coronary artery disease. Circulation 1992;85:2025e33. [8] Navab M, Hama SY, Cooke CJ, et al. Normal high density lipoprotein inhibits three steps in the formation of mildly oxidized low density lipoprotein: step 1. J Lipid Res 2000;41:1481e94. [9] Navab M, Hama SY, Anantharamaiah GM, et al. Normal high density lipoprotein inhibits three steps in the formation of mildly oxidized low density lipoprotein: steps 2 and 3. J Lipid Res 2000;41:1495e508. [10] Vaisar T, Pennathur S, Green PS, et al. Shotgun proteomics implicates protease inhibition and complement activation in the antiinflammatory properties of HDL. J Clin Invest 2007;117:746e56. [11] Lapointe A, Couillard C, Piche ME, et al. Circulating oxidized LDL is associated with parameters of the metabolic syndrome in postmenopausal women. Atherosclerosis 2007;191:362e8. [12] Lehtimäki T, Lehtinen S, Solakivi T, et al. Autoantibodies against oxidized low density lipoprotein in patients with angiographically verified coronary artery disease. Arterioscler Thromb Vasc Biol 1999;19:23e7. [13] van de Vijver LP, Steyger R, van Poppel G, et al. Autoantibodies against MDALDL in subjects with severe and minor atherosclerosis and healthy population controls. Atherosclerosis 1996;122:245e53. [14] van Bommel EF, van Tits LJ, van den Berg EA, Prins J, Stalenhoef AF. Autoantibodies against oxidized low-density lipoprotein and lipid profile in patients with chronic periaortitis: case-control study. Rheumatol Int 2011;31:201e8. [15] Zhang B, Noda K, Matsunaga A, Kumagai K, Saku K. A comparative crossover study of the effects of fluvastatin and pravastatin (FP-COS) on circulating autoantibodies to oxidized LDL in patients with hypercholesterolemia. J Atheroscler Thromb 2005;12:41e7. [16] Papathanasiou AI, Lourida ES, Tsironis LD, Goudevenos JA, Tselepis AD. Shortand long-term elevation of autoantibody titers against oxidized LDL in patients with acute coronary syndromes. Role of the lipoprotein-associated phospholipase A2 and the effect of atorvastatin treatment. Atherosclerosis 2008;196:289e97. [17] Choi SH, Chae A, Miller E, et al. Relationship between biomarkers of oxidized low-density lipoprotein, statin therapy, quantitative coronary angiography, and atheroma: volume observations from the REVERSAL (reversal of atherosclerosis with aggressive lipid lowering) study. J Am Coll Cardiol 2008;52:24e32. [18] Chen Q, Reis SE, Kammerer C, et al. Association of anti-oxidized LDL and candidate genes with severity of coronary stenosis in the Women’s ischemia syndrome evaluation study. J Lipid Res 2011;52:801e7. [19] Niskanen L, Lindi V, Erkkilä A, et al. Association of the PRO12ALA polymorphism of the PPAR-gamma2 gene with oxidized low-density lipoprotein and cardiolipin autoantibodies in nondiabetic and type 2 diabetic subjects. Metabolism 2003;52:213e7.
[20] Kangas-Kontio T, Kakko S, Tamminen M, et al. Genome scan for loci regulating HDL cholesterol levels in Finnish extended pedigrees with early coronary heart disease. Eur J Hum Genet 2010;18:604e13. [21] Veneskoski M, Turunen SP, Kummu O, et al. Specific recognition of malondialdehyde and malondialdehyde acetaldehyde adducts on oxidized LDL and apoptotic cells by complement anaphylatoxin C3a. Free Radic Biol Med 2011; 51:834e43. [22] Sämpi M, Ukkola O, Päivänsalo M, Kesäniemi YA, Hörkkö S. Early atherosclerosis and IgG2 to bacteria are associated with FcgammaRIIa genotype in non-smokers. Eur J Clin Invest 2009;39:517e26. [23] Amos CI. Robust variance-components approach for assessing genetic linkage in pedigrees. Am J Hum Genet 1994;54:535e43. [24] Almasy L, Dyer TD, Blangero J. Bivariate quantitative trait linkage analysis: pleiotropy versus co-incident linkages. Genet Epidemiol 1997;14: 953e8. [25] Kangas-Kontio T, Huotari A, Ruotsalainen H, et al. Genetic and environmental determinants of total and high-molecular weight adiponectin in families with low HDL-cholesterol and early onset coronary heart disease. Atherosclerosis 2010;210:479e85. [26] Almasy L, Blangero J. Multipoint quantitative-trait linkage analysis in general pedigrees. Am J Hum Genet 1998;62:1198e211. [27] Wang Y, Fang X, Wang S, Feng Y, Yin J. Relation between plasma oxLDL antibodies and oxLDL in the circulation. Inflammation 2007;30:7e13. [28] Fraley AE, Schwartz GG, Olsson AG, et al. Relationship of oxidized phospholipids and biomarkers of oxidized low-density lipoprotein with cardiovascular risk factors, inflammatory biomarkers, and effect of statin therapy in patients with acute coronary syndromes: results from the MIRACL (myocardial ischemia reduction with aggressive cholesterol lowering) trial. J Am Coll Cardiol 2009;53:2186e96. [29] Hohler T, Reuss E, Evers N, et al. Differential genetic determination of immune responsiveness to hepatitis B surface antigen and to hepatitis A virus: a vaccination study in twins. Lancet 2002;360:991e5. [30] Duah NO, Weiss HA, Jepson A, Tetteh KK, Whittle HC, Conway DJ. Heritability of antibody isotype and subclass responses to Plasmodium falciparum antigens. PLoS One 2009;4:e7381. [31] Chen HW, Kuo CL, Huang CS, Kuo SJ, Liu CS. Oxidized low-density lipoproteins, autoantibodies against oxidized low-density lipoproteins and carotid intima media thickness in a clinically healthy population. Cardiology 2008;110: 252e9. [32] Liu KZ, Cuddy TE, Pierce GN. Oxidative status of lipoproteins in coronary disease patients. Am Heart J 1992;123:285e90. [33] Khosravi A, Hommel M, Sayemiri K. Age-dependent antibody response to Plasmodium falciparum merozoite surface protein 2 (MSP-2). Parasite Immunol 2011;33:145e57.