Lipids

, Volume 48, Issue 1, pp 51–61

High Serum Apolipoprotein E Determines Hypertriglyceridemic Dyslipidemias, Coronary Disease and ApoA-I Dysfunctionality

Authors

    • Turkish Society of Cardiology
    • Cerrahpaşa Medical FacultyIstanbul University
  • Günay Can
    • Cerrahpaşa Medical FacultyIstanbul University
  • Ender Örnek
    • Etlik Ihtisas Education Hospital
  • Erkan Ayhan
    • Cardiology Department, Medical FacultyBalıkesir University
  • Nihan Erginel-Ünaltuna
    • Department of Genetics, Institute for Experimental MedicalIstanbul University
  • Sani N. Murat
    • Etlik Ihtisas Education Hospital
Original Article

DOI: 10.1007/s11745-012-3724-8

Cite this article as:
Onat, A., Can, G., Örnek, E. et al. Lipids (2013) 48: 51. doi:10.1007/s11745-012-3724-8
  • 326 Views

Abstract

The relevance of serum apolipoprotein E (apoE) levels to two hypertriglyceridemic dyslipidemias has not been clarified. We explored, in a cross-sectional (and short-term prospective) evaluation, the independent relationship of serum apoE to the atherogenic dyslipidemia, hypertriglyceridemia with elevated apoB (HtgB) and to apoA-I dysfunctionality, previously shown in Turkish adults to be independent of apoE genotype. Serum apoE concentrations were measured by immunonephelometry in 1,127 middle-aged adults. In multivariable regression analysis, apoE concentrations showed log-linear associations with apoB and apoA-I levels, waist circumference, independent of C-reactive protein (CRP), homeostatic model assessment (HOMA) index and other confounders. The likelihood of atherogenic dyslipidemia and of HtgB roughly tripled per 1-SD increment in apoE concentrations, additively to apoE genotype, HOMA, apoA-I, CRP concentrations and waist circumference; yet apoA-I, protective against atherogenic dyslipidemia, appeared to promote HtgB, a finding consistent with apoA-I dysfunctionality in this setting. Each 1-SD increment in the apoE level was moreover, associated in both genders with MetS (at OR 1.5), after adjustment for sex, age, apoB, apoA-I and CRP, or for apoE genotypes. Circulating apoE predicted in both genders age-adjusted prevalent and incident coronary heart disease (CHD), independent of apoE genotype and CRP (OR 1.32 [95 % CI 1.11; 1.58]). To conclude, in a general population prone to MetS, elevated apoE concentrations are strongly linked to HtgB and atherogenic dyslipidemia, irrespective of apoE genotype, are associated with MetS and CHD. Excess apoE reflects pro-inflammatory state and likely autoimmune activation.

Keywords

Apolipoprotein A-IApolipoprotein BApolipoprotein E concentrationsAtherogenic dyslipidemiaCoronary heart diseaseHypertriglyceridemia with elevated apoB

Abbreviation

apo

Apolipoprotein

LDL

Low-density lipoprotein

CHD

Coronary heart disease

Lp(a)

Lipoprotein(a)

CRP

C-reactive protein

MetS

Metabolic syndrome

HDL

High-density lipoprotein

SHBG

Sex hormone-binding globulin

HOMA

Homeostatic model assessment

TARF

Turkish adult risk factor study

HtgB

Hypertriglyceridemia with elevated apoB

VLDL

Very low-density lipoprotein

Introduction

Apolipoprotein E (apoE) is the main ligand for clearance through low-density lipoprotein (LDL) receptor and LDL receptor-related protein of very low-density lipoproteins (VLDL) remnants and chylomicron remnants; it thereby influences concentrations of lipoproteins and lipids [1]. The impact on serum lipids and lipoproteins [2] and on coronary heart disease (CHD) [3] of the six common isoforms of apoE in plasma, encoded by the ε2, ε3 and ε4 alleles, has been well established in the past two decades. ApoE controls cholesterol efflux from cells, activates lecithin-cholesterol acyltransferase (LCAT), has antioxidant properties and plays a role in the regulation of the inflammatory response [4]. Through stimulation of VLDL production and decrease in its clearance, apoE explains 20–40 % of the variability of triacylglycerol concentrations [5]. The clinical relevance of serum apoE concentrations has not been adequately investigated in regard to their relation to the two hypertriglyceridemic dyslipidemias and the metabolic syndrome (MetS). In a pilot study, we had previously reported that serum apoE levels in Turkish adults rendered information additive to genotype relative to hypertriglyceridemic dyslipidemias and the MetS [6]. It was recently documented that changes in the size and composition of high-density lipoprotein (HDL) particles were associated with MetS, concurrently with circulating apoE [7].

Studies documenting the relation of circulating total apoE to cardiovascular disease risk are scarce [8, 9], particularly in diverse populations. In two studies using Dutch cohorts, high serum apoE predicted cardiovascular disease or mortality, in one study [8] it was shown to be independent of the apoE genotype, and the suggestion was made that high apoE concentrations indicate either an adverse lipoprotein profile or a pro-inflammatory response.

We were first to show in a general population that the enhanced pro-inflammatory state associated with high-density lipoprotein (HDL) dysfunction is a prominent feature of the cardiometabolic risk profile among Turkish adults [10, 11]. High serum apoA-I levels conferred a risk of type-2 diabetes in this population [10, 12]. This paradoxical association proved in women to be independent of apoE genotype and apoB concentrations [13], but the association with apoE concentrations warrants further investigation. Evidence of dysfunction of apoA-I and HDL particles is not confined to this population sample. Though not designated as HDL dysfunction, thousands of participants in the trials PROSPER [14] and PROVE IT-TIMI 22 [15] displayed a lack of association between baseline levels of these lipoproteins and the primary end-point, which strongly suggests operation of impaired atheroprotective properties of these lipoproteins.

It was therefore attractive to explore in a representative population sample of Turks, the relation of serum apoE concentrations with (1) homeostatic model assessment (HOMA) and/or inflammatory biomarkers, (2) the atherogenic (high triacylglycerol/low HDL) dyslipidemia, and the hypertriglyceridemia with elevated apoB (HtgB), (3) apoA-I dysfunctionality and (4) the likelihood of MetS and CHD. This analysis shed light on the link between apoE levels and hypertriglyceridemia and the presence of dysfunctional apoA-I.

Subjects and Methods

Sample Population

This study sample was recruited randomly from participants of the 2005–2008 follow-up surveys of the Turkish Adult Risk Factor Study, a prospective study on cardiac disease and its risk factors in a representative sample of adults in Turkey carried out biennially since 1990 in 59 communities throughout seven geographical regions of the country [16]. Details of the overall sampling were described previously [16]. The study was approved by the Ethics Committee of the Medical Faculty, Istanbul University. Written informed consent for participation was obtained from all individuals. Partial logistical support was provided by the Turkish Ministry of Health. Data were obtained by history of the past years via a questionnaire, physical examination of the cardiovascular system and sampling of blood.

Measurement of Risk Factors

Blood pressure (BP) was measured with an aneroid sphygmomanometer (Erka, Germany) in the sitting position on the right arm, and the mean of two recordings 3 min apart was recorded. Waist circumference was measured—with the subject standing and wearing only underwear, at the level midway between the lower rib margin and the iliac crest. Nonsmokers, former smokers and current smokers formed the categories of cigarette smoking. Anyone consuming alcohol once a week or more was considered as an alcohol user.

Blood samples were collected, after an overnight fast >11 h in 95 % of the sample, spun at 1,000 g and shipped on cooled gel packs to Istanbul to be stored in deep-freeze at −75 °C, until analyzed in a central laboratory. Serum concentrations of apoE, apoB, apoA-I and CRP were measured by nephelometry (BN Prospec, Behring Diagnostics, Westwood, MA). Within-run and day-to-day coefficients of variation for apoE assays were 1.74–3.06 %, respectively. Serum concentrations of triacylglycerol and glucose, measured in the fasting state, as well as of total, LDL and HDL cholesterol (LDL-C, HDL-C plus second generation, direct quantification) were determined using enzymatic kits from Roche Diagnostics with a Hitachi 902 autoanalyzer. Fasting concentrations of sex hormone-binding globulin (SHBG) and insulin were assayed by the electrochemiluminescence immunoassay ECLIA on Roche Elecsys 2010 immunauto-analyzer (Roche Diagnostics, Mannheim, Germany). The existence of the ApoE ε2/ε3/ε4 alleles (ApoE polymorphisms Cys112Arg and Arg158Cys) was analysed by means of the TaqMan technology (ABI 7900HT, Applied Biosystems, UK), as described in detail previously [13].

Definitions

The dyslipidemia consisting of triacylglycerol ≥1.7 mmol/L and HDL-cholesterol ≤1.03 in men and ≤1.30 mmol/L in women was designated as atherogenic dyslipidemia, while hypertriglyceridemia with apoB >1.2 g/L is abbreviated as HtgB. MetS was identified when three out of the five criteria of the National Cholesterol Education Program ATP-III were met, modified for prediabetes (fasting glucose 5.56–6.95 mmol/L) [17] and further for male abdominal obesity using as cutout point ≥95 cm, as assessed in the Turkish Adult Risk Factor study [18]. Diabetes was diagnosed with the criteria of the American Diabetes Association [19], namely by self report or when plasma fasting glucose was ≥7 mmol/L or when 2 h postprandial glucose was >11.1 mmol/L. HOMA was calculated with the following formula [20]: insulin (mIU/L)* glucose (in mmol/L)/22.5. Usage of any type of statin drugs for lipid lowering for at least 3 months duration was designated as statin usage which served among adjustments in regression analyses.

Information on the mode of death was obtained from first-degree relatives and/or health personnel of local health office. Cause of death was assigned with the consideration also of pre-existing clinical and laboratory findings elicited during biennial surveys. CHD death comprised death from heart failure of coronary origin and fatal coronary event. Non-fatal CHD was identified by presence of angina pectoris, a history of myocardial infarction with or without accompanying Minnesota codes of the ECG [21] (Q, QS pattern codes 1.1–2) or a history of myocardial revascularization. Typical angina and, in women, age >45 years were prerequisite for a diagnosis when angina was isolated. ECG changes of “ischemic type” of greater than minor degree (codes 1.1–2, 4.1–2, 5.1–2, 7.1) were considered as myocardial infarct sequelae or myocardial ischemia, respectively.

Data Analysis

Descriptive parameters were shown as means ± standard deviations (SD) or in percentages. Due to a skewed distribution, values derived from log-transformed (geometric) means were used for apoE, CRP, insulin and SHBG. Two-sided t-tests and Pearson’s Chi-square tests were used to analyze the differences between means and proportions of groups. Spearman correlations served to analyze univariate correlations. Multiple linear regression analyses were performed with continuous parameters. These were expressed in terms of an increment of 1 SD. Likelihood estimates (OR) and 95 % confidence intervals (CI) were obtained by use of logistic regression analyses in models that adjusted for sex, age and relevant confounders, again expressed in terms of 1-SD increment. A brief follow-up (mean 1.4 ± 0.8 year) was available, newly developing CHD cases during, which were evaluated as well. A value of p < 0.05 on the two-tail test was considered statistically significant. Statistical analyses were performed using SPSS-10 for Windows.

Results

The study sample consisted of 1,127 adults aged 55.4 ± 11 years. MetS was identified in 266 (50.4 %) men and in 331 (55.4 %) women. Isolated atherogenic dyslipidemia was present in 22 %, isolated HtgB in 10 % and both combined in 10 % of the sample.

Distribution of apoE Concentrations

Relatively large waist girths, high serum triacylglycerol, low HDL-cholesterol concentrations, and normal LDL-cholesterol were characteristics of the sample displayed in Table 1. Geometric mean apoE concentrations were 3.89 ± 1.5 in men and 4.07 ± 1.396 mg/dL among women (p = 0.041). Median (interquartile range) values were 3.88 (3.16–4.85) mg/dL. In 71 participants, apoE concentrations were measured in each survey 2 years apart; correlation of the paired samples was 0.51 (p < 0.001) and the means were similar (p = 0.18).
Table 1

Baseline characteristics of the sample population, by gender

 

n

Men (n = 528)

Women (n = 599)

p*

Mean

SD

Mean

SD

Age, year

1,127

55.7

10.7

55.2

10.4

0.48

Apolipoprotein Ea, (mg/dL)

1,127

3.89

1.5

4.07

1.39

0.041

Waist circumference, (cm)

1,125

97.6

11

95.4

12.9

0.002

Systolic blood pressure, (mmHg)

1,123

122.2

19.7

125.8

21.6

0.004

Diastolic blood pressure, (mmHg)

1,123

76.6

10.7

77.8

11.1

0.063

Fasting glucose, (mmol/L)

1,002

5.65

1.94

5.61

2.23

0.77

Total cholesterol, (mmol/L)

1,127

4.81

1.01

5.15

1.05

<0.001

HDL-cholesterol, (mmol/L)

1,126

1.02

0.26

1.22

0.40

<0.001

LDL-cholesterol, (mmol/L)

1,097

2.89

0.81

3.10

0.85

<0.001

Fasting triacylglycerol, (mmol/L)

1,086

1.97

1.19

1.81

1.07

0.02

Apolipoprotein A-I, (g/L)

1,047

1.35

0.24

1.49

0.27

<0.001

Apolipoprotein B, (g/L)

1,058

1.00

0.29

1.00

0.27

0.93

C-reactive proteina, (mg/L)

1,071

2.09

3.01

2.56

2.89

0.002

HOMA indexa

935

2.02

2.44

2.21

2.15

0.091

SHBGa, (nmol/L)

857

38.5

1.61

48.9

1.66

<0.001

Physical activity grade I–IV

1,123

2.65

0.82

2.18

0.76

<0.001

Currently smoking n, (%)

1,064

199

40.0

83

14.7

<0.001

Use of alcohol n, (%)

1,068

74

14.7

6

1.1

<0.001

Use of statins n, (%)

1,123

36

6.8

49

8.2

 

aGeometric mean and SD values

*p value between men and women

HOMA homeostasis model assessment, SHBG sex hormone-binding globulin

Covariates of apoE Concentrations

Lipids, lipoproteins and inflammation-related proteins were compared in groups of normal or elevated serum triacylglycerol and apoB (Table 2). This demonstrates that in individuals with high apoB (>1.2 g/L), (a) apoE concentrations are 20 % higher (p < 0.001) than in those without hyperapoB, regardless of triacylglycerol elevation or low HDL-C; (b) apoA-I concentrations are by 10 % higher (0.14 g/L; p < 0.001) than in individuals without hyperapoB, regardless of triacylglycerol elevation or low HDL-C; (c) fasted triacylglycerol are 52 % higher (0.53 mmol/L; p < 0.001) than in individuals without hyperapoB, regardless of HDL-cholesterol category. ApoE genotype distribution was similar in all four groups.
Table 2

Mean values of HDL-cholesterol and inflammation-related apoproteins in conditions with normal and elevated triglyceride and apoB levels

 

Normal TAG and apoB

Elev TAG/low HDL

Elev apoB

Elev TAG/low HDL and apoB

Mean

SD

Mean

SD

Mean

SD

Mean

SD

 

n = 626

n = 247

n = 108

n = 109

ApoE conc.c (mg/dL)

3.53

1.36

4.45a

1.42

4.80a

1.50

5.41a,b

1.42

ApoB (g/L)

0.87

0.18

0.96a

0.15

1.36a,b

0.24

1.42a,b

0.25

ApoA-I (g/L)

1.44

0.27

1.29a

0.20

1.59a,b

0.27

1.45b

0.21

HDL-cholest. (mmol/L)

1.22

0.30

0.90a

0.19

1.23b

0.27

0.95a

0.17

LDL-cholest. (mmol/L)

2.83

0.70

2.77

0.70

3.80a,b

0.85

3.75a,b

0.83

NonHDL-cholest. (mmol/L)

3.64

0.80

4.10

0.78

4.80a,b

0.98

5.29a,b

0.91

Fast. triacylglycerol (mmol/L)

1.24

1.46

2.53

1.41

160a,b

1.58

2.85

1.42

ApoE genotyped E2/E3/E4 groups (%)

12; 76; 12

14; 72; 14

9; 78; 13

3; 82; 15

ap < 0.001 from the normotriglyceridemic group

bp < 0.001 from the high triglyceride/low HDL-C group

cLog-transformed values

dApoE genotype was available in 689 participants, and its distribution among the four categories listed was not significant in either gender (p = 0.49 and 0.32)

Dyslipidemia of hyperapoB in this study comprised elevated TAG and apoB levels with or without low HDL-C

Cutoffs for fasting triacylglycerol 1.70 mmol/L, for apoB 1.20 g/L

Table 3 depicts gender-stratified associations of covariates of serum apoE levels in highly significant (p < 0.001) linear regression models with 11 independent variables. ApoB, apoA-I, waist circumference in both genders and, (inversely) age in men, were significant covariates, independent of CRP, HOMA index, systolic BP and lifestyle confounders.
Table 3

Multivariable linear regression analysis for independent covariates of apoE concentrations (mg/dL), by gender

 

Total (n = 882)b

Men (n = 409)

Women (n = 473)

β coeff.c

SE

p

β coeff.

SE

p

β coeff.

SE

p

Gender, F

0.994

1.03

0.82

      

Age, 11 years

0.988

1.00

0.35

0.96

1.01

0.032

1.025

1.01

0.13

ApoB, 0.3 g/L

1.16

1.00

<0.001

1.16

1.00

<0.001

1.15

1.00

<0.001

ApoA-I, 0.24/0.27 g/L

1.055

1.00

<0.001

1.075

1.00

<0.001

1.042

1.00

0.003

Waist circumfer., 11/13 cm

1.04

1.00

0.006

1.05

1.01

0.029

1.03

1.01

0.042

C-reactive proteina, (mg/L)

1.028

1.03

0.25

1.037

1.04

0.33

1.018

1.02

0.57

HOMA indexa

1.03

1.04

0.33

1.04

1.05

0.47

1.02

1.04

0.62

Physical activity grade IV/I

0.95

1.01

0.44

0.90

1.03

0.24

1.04

1.02

0.61

Systolic BP, 25 mmHg

0.992

1.00

0.59

0.99

1.00

0.57

0.995

1.00

0.76

Current versus never smokers

1.006

1.06

0.84

1.03

1.03

0.95

1.02

1.02

0.60

Alcohol usage, y/n

0.986

1.04

0.75

0.96

1.05

0.48

1.02

1.16

0.92

 Explained apoE variance

 

20 %

  

19 %

  

21 %

 

aLog-transformed values

bAll 11 variables were available only in 78 % of the sample

cTo determine the corresponding change in apoE level, the β coefficient (SE) shows the factor to be multiplied for each increment in the independent variables

Each model significant (p < 0.001). Significant values are highlighted in bold

HOMA homeostasis model assessment

In univariate correlations in genders combined, triacylglycerol (r = 0.53), LDL-cholesterol (r = 0.21), and SHBG (r = −0.16) were strongly correlated (each p < 0.001) with apoE as further variables.

Association of apoE Concentrations with Hypertriglyceridemic Dyslipidemias

Table 4 shows the cross-sectional association of apoE concentrations with two types of dyslipidemias, adjusted for sex, age, waist circumference, apoA-I, CRP, HOMA and statin use. ApoE was positively and apoA-I inversely associated with dyslipidemia of high triacylglycerol/low HDL-cholesterol in both genders, whereas HOMA was further independently associated in women. The likelihood of this dyslipidemia was 2.6-fold per 1-SD increment in apoE concentrations in each gender (95 %CI 2.12; 3.13 in the whole sample). A second model in which apoE genotype was added strengthened the association of circulating apoE, particularly in women.
Table 4

Cross-sectional multivariable logistic regression associations for triacylglycerol/HDL dyslipidemia and hypertriglyceridemic apoB of apoE concentrations, adjusted for sex, age, apoE genotype, HOMA and inflammation-related parameters

 

Total

Men

Women

OR

95 %CI

OR

95 %CI

OR

95 %CI

For atherogenic dyslipidemia

280/880a

132/405a

148/475a

 Gender, F

1.31

0.93; 1.85

  

 Age, 11 years

0.97

0.81; 1.15

0.91

0.70; 1.17

1.07

0.84; 1.35

 ApoE concentr., 1.44-fold

2.58

2.12; 3.13

2.57

1.94; 3.39

2.58

1.95; 3.41

 ApoA-I, 0.24/0.27 g/L

0.39

0.32; 0.48

0.38

0.27; 0.51

0.40

0.31; 0.53

 HOMA, twofold

1.16

1.05; 1.29

1.08

0.93; 1.24

1.28

1.10; 1.50

 Statin usage, y/n

2.34

1.34; 4.08

3.52

1.37; 9.07

1.82

0.91; 3.67

Model with apoE genotype

193/632a

91/296a

102/336a

 ApoE concentr., 1.44-fold

3.17

2.44; 4.11

2.87

2.04; 4.10

3.73

2.50; 5.58

 ApoA-I, 0.24/0.27 g/L

0.33

0.25; 0.40

0.32

0.22; 0.47

0.31

0.22; 0.46

 HOMA, twofold

1.16

1.02; 1.32

1.01

0.85; 1.20

1.41

1.14; 1.75

 ApoE2 group

2.87

1.45; 5.68

4.15

1.47; 11.7

2.34

0.89; 6.17

 ApoE4 group

3.09

1.34; 7.12

5.72

1.67; 19.6

1.96

0.59; 6.52

 Waist circumference, 11/13 cm

1.18

0.94; 1.49

1.48

1.04; 2.10

0.97

0.69; 1.36

 Statin usage, y/n

1.39

0.67; 2.91

2.41

0.65; 8.91

1.10

0.44; 2.79

For hypertriglyceridemic apoB

175/882a

73/409a

102/473a

 Gender, F

0.92

0.63; 1.36

    

 Age, 11 years

1.26

1.03; 1.52

1.06

0.80; 1.41

1.41

1.08; 1.84

 ApoE concentr., 1.44-fold

1.98

1.65; 2.38

1.61

1.29; 2.02

2.62

1.95; 3.52

 ApoA-I, 0.24/0.27 g/L

1.50

1.23; 1.83

1.81

1.30; 2.45

1.34

1.06; 1.75

 HOMA, twofold

1.12

1.001; 1.25

1.08

0.92; 1.27

1.15

0.98; 1.35

 Statin usage, y/n

1.79

1.03; 3.11

1.03

0.38; 2.83

2.53

1.27; 5.07

Model with apoE genotype

126/632a

50/296a

76/336a

 Age, 11 years

1.44

1.15; 1.82

1.28

0.920; 1.80

1.54

1.10; 2.15

 ApoE concentr., 1.44-fold

2.47

1.94; 3.15

1.72

1.31; 2.26

4.56

2.93; 7.10

 ApoA-I, 0.24/0.27 g/L

1.46

1.14; 1.83

1.90

1.30; 2.81

1.24

0.90; 1.71

 HOMA, twofold

1.10

0.95; 1.26

1.11

0.92; 1.34

1.10

0.88; 1.36

 ApoE2 group

4.50

1.91; 10.6

3.20

1.06; 9.72

9.21

2.53; 7.10

 ApoE4 group

5.99

2.21; 16.3

3.71

0.97; 14.2

14.9

3.32; 67.0

 Statin usage, y/n

1.40

0.69; 2.81

1.60

0.44; 5.79

1.45

0.59; 3.57

aNumber of participants with dyslipidemia/total sample

Significant values are highlighted in bold

88 participants with both dyslipidemias combined were included in each model

ORs for continuous variables are expressed in terms of 1-SD increment

Models were also adjusted for C-reactive protein, waist circumference and gender (p in all non-significant). The females had a p = 0.084 for atherogenic dyslipidemia for which age was absolutely unrelated in each sex

HOMA homeostasis model assessment

Regarding HtgB, circulating apoE (OR 1.98 [95 %CI 1.65; 2.38]), apoA-I (OR 1.50 [95 %CI 1.23; 1.83]) and CRP were positively associated (OR 1.13 [95 %CI 1.00; 1.27] in both genders. When the apoE genotype was further added to the adjustment in these models, the associations both of circulating apoE and apoA-I were only slightly attenuated and persisted in being significant.

Prospective multivariable logistic regression analysis for the new development of hypertriglyceridemias (not shown in detail) demonstrated HOMA and (inversely) serum apoA-I to be significant determinants of atherogenic dyslipidemia, after adjustments for age, waist girth and concentrations of apoE and CRP. ApoA-I protected against the development of atherogenic dyslipidemia (RR 0.22 [95 %CI 0.33; 0.90]); yet it tended to predict the risk of development of HtgB in 17 out of 371 women, additively to apoE concentrations.

Association of apoE Concentrations with MetS and CHD

In a logistic regression model for MetS, serum apoE and each of the 4 other independent variables examined (age, apoB, apoA-I and CRP) were associated significantly with MetS likelihood in either gender (Table 5). Serum apoE proved to be associated with MetS independently of the apoE genotypes in a second model. The OR per 1-SD of apoE in sexes combined was 1.43 (95 %CI 1.22; 1.67). CHD, developing newly in 24 of 618 individuals, was predicted by serum apoE only in women after adjustment for age and CRP (Table 5). Combined prevalent and incident CHD was associated with apoE concentrations, when adjusted to apoE genotype and CRP. Whereas, the apoE2 and E4 groups did not reach statistically significant associations compared to the ε3 homozygotes, apoE concentrations were significantly and additively associated with CHD at an OR of 1.32 (95 % CI 1.11; 1.58) per 1-SD increment.
Table 5

Multivariable logistic regression analysis of serum apoE and certain covariates for prevalent metabolic syndrome, and incident and prevalent coronary heart disease

 

OR

95 % CI

OR

95 % CI

OR

95 % CI

For metabolic syndrome

Total

542/1,040b

Men

249/496b

Women

293/544b

 Gender, female

1.72

1.28; 2.30

    

 Age, 11 years

1.52

1.31; 1.76

1.40

1.14; 1.71

1.62

1.30; 2.00

 ApoE conc.a, 1.44-fold

1.52

1.31; 1.76

1.57

1.29; 1.91

1.43

1.14; 1.80

 ApoB, 0.3 g/L

1.92

1.61; 2.36

1.76

1.35; 2.29

2.10

1.61; 2.73

 ApoA–I, 0.24/0.27 g/L

0.46

0.39; 0.55

0.54

0.44; 0.68

0.40

0.31; 0.50

 C-reactive proteina, threefold

1.18

1.08; 1.29

1.11

0.98; 1.25

1.26

1.10; 1.44

Model 2

n = 386/710b

n = 165/335b

n = 221/375b

 Gender, female

1.37

1.00; 1.87

    

 Age, 11 years

1.40

1.19; 1.64

1.23

0.99; 1.54

1.56

1.22; 2.00

 ApoE conc.a, 1.44-fold

1.43

1.22; 1.67

1.50

1.21; 1.85

1.32

1.04; 1.68

 ApoE2 group

1.53

0.91; 2.58

1.75

0.88; 3.47

1.27

0.56; 2.86

 ApoE4 group

1.56

0.82; 2.99

1.57

0.67; 3.70

1.41

0.51; 3.91

 C-reactive proteina, threefold

1.22

1.10; 1.35

1.09

0.95; 1.25

1.36

1.17; 1.59

For incident CHD

24/618b,c

11/288b

13/330b

 Gender, female

0.90

0.39; 2.09

    

 Age, 11 years

1.94

1.26; 3.00

1.80

0.98; 3.34

1.88

0.99; 3.58

 ApoE conc.a, 1.44-fold

1.32

0.94; 1.85

1.04

0.60; 1.79

1.88

1.07; 3.31

 C-reactive proteina, threefold

1.20

0.91; 1.59

1.08

0.73; 1.59

1.17

0.95; 1.44

For prevalent and incid. CHD

n = 142/710b

n = 71/335b

n = 71/375b

 Gender, female

0.82

0.55; 1.22

    

 Age, 11 years

2.40

1.94; 2.97

2.24

1.69; 3.00

2.50

1.80; 3.44

 ApoE conc.a, 1.44-fold

1.32

1.11; 1.58

1.27

1.01; 1.60

1.45

1.07; 1.94

 ApoE2 group

1.34

0.69; 2.59

1.09

0.47; 2.51

1.83

0.62; 5.41

 ApoE4 group

1.24

0.55; 2.83

0.69

0.23; 2.08

2.58

0.71; 9.33

 C-reactive proteina, threefold

1.11

0.98; 1.25

0.97

0.81; 1.15

1.30

1.07; 1.59

aLog-transformed values, ORs for continuous variables are expressed in terms of 1-SD increment

bNumber of participants with MetS or CHD/number at risk

cOnly participants with follow-up are included

Significant values are highlighted in bold

Discussion

In the current study on a large middle-aged general population sample of the Mediterranean area, serum apoE concentrations reflected pro-inflammatory properties by demonstrating linear associations with apoB, apoA-I, waist circumference in both genders independent of CRP, HOMA index, systolic BP and lifestyle confounders. The likelihood of atherogenic dyslipidemia and of the HtgB was 2½-fold increased per 1-SD increment in apoE concentrations in each gender, additively to apoE genotype, HOMA, apoA-I, CRP concentrations and waist circumference. A further salient finding was that apoA-I, inversely related to atherogenic dyslipidemia, was associated positively with the HtgB implicating the presence of proinflammatory, dysfunctional apoA-I. Serum apoE concentrations were significantly associated with both CHD and MetS in either gender, when adjusted for apoE genotype and CRP.

Independent Associations of apoE Levels with Inflammation Markers

Serum apoE was associated log-linearly—beyond fasting triacylglycerol—with apoB, apoA-I and waist circumference. In women, hypertriglyceridemia with elevated apoB was independently, additively and strongly associated with both apoE concentrations and apoε4 allele. ApoE is known to be involved in the regulation of immune and inflammatory responses [22]; it clearly presented pro-inflammatory features in this study. The stated close link between concentrations of apoE and apoB is consistent with the prediction by apoB in Turks of cardiometabolic events, additive to markers of central obesity and inflammation [23].

Associations with the Two Hypertriglyceridemic Dyslipidemias, and their Distinctive Properties

Independent associations of apoE levels with the two hypertriglyceridemic dyslipidemias were similar in magnitude, disclosing roughly doubling or trebling ORs, and the females showed somewhat stronger association to the hypertriglyceridemia with hyperapoB. Sniderman and associates [24] firstly drew attention to this atherogenic dyslipoproteinemia in subjects with type-2 diabetes. We now demonstrate that it may develop in the population at large beyond the diabetic state and that its main distinction from the commonly observed atherogenic dyslipidemia is a positive (not inverse) association of apoA-I with HtgB (Tables 2, 5). These associations are consistent with apoE being involved in a proinflammatory process in which elevated levels of apoA-I, apoB and triacylglycerol are involved as well, suggesting an autoimmune activation, and reflect oxidative impaired apoA-I function in HtgB, likely via binding to the pro-inflammatory Lp(a) (as further discussed below). Current findings indicate the existence of an interaction for elevated apoE between circulating low HDL-C and high apoB as surrogates of a heightened proinflammatory state. Minor differences from the atherogenic dyslipidemia include a lower prevalence in the community and an age-dependency in females.

Elevated apoA-I levels conferred a risk of HtgB, in contrast to protecting against atherogenic dyslipidemia. In a population sample harboring impaired glucose tolerance or oxidative stress, apoA-I particles are expectedly heterogeneous, in most people functioning normally, displaying in others (severely) impaired function. In the context of atherogenic dyslipidemia, higher compared to lower levels of normally functioning apoA-I particles will show protective features, whereas in the analysis for individuals with HtgB, the dysfunctional high (compared to normal or low) levels of apoA-I particles (due to enhanced low-grade inflammation) are likely to impart increased risk for HtgB. In an enhanced inflammatory milieu mediated by apoE, shown to function as an activator for phospholipid transfer protein [25], apoA-I may lose atheroprotective properties by binding to Lp(a) (which carries apoB) and promote the HtgB.

Does apoE Mediate Subclinical Inflammation and Promote HDL Dysfunction?

Alterations leading to dysfunctional HDL in diabetes and inflammatory states have been recognized [26]. Using shotgun proteomics to investigate the composition of HDL isolated from healthy subjects and subjects with CHD, it was demonstrated that HDL3 in CHD patients was selectively enriched in proteins that play critical roles in macrophage biology, lipid metabolism and inflammatory response and that apoE was the protein most significantly enriched [27]. Investigators suggested that redistribution of apoE from HDL2 to HDL3 might impair cholesterol efflux and promote the formation of macrophage foam cells in vivo. In a limited number of Korean patients with MetS, Park and co-workers [28] found that, compared with control subjects, parameters reflecting a proinflammatory state such as serum triacylglycerol, apoC-III and uric acid were enriched particularly in HDL2 to HDL3 fractions rather than in VLDL, an enrichment that was correlated with the loss of antioxidant and anti-atherogenic activities.

The current study suggests that circulating apoE, which in men independently links to dysfunctional apoA-I concentrations and the two dyslipidemias, mediates in women abdominal obesity as risk factors for hypertriglyceridemia. We have previously reported in a larger sample of the TARF cohort that high apoA-I tertile conferred a two-fold risk for diabetes compared with participants in the low apoA-I tertile in the general adult population [12]. Furthermore, genetically elevated apoA-I levels have recently been shown to be associated not with decreased (but rather with tendency to increased) risk of CHD or myocardial infarction [29].

Excess circulating apoE, secreted by the liver, is herein clearly associated with hypertriglyceridemic dyslipidemias, MetS [11] and occasionally with coronary disease [9, 13, 30].

Current interpretation is in agreement with the report on 85-year-old Dutch adults in whom high plasma apoE concentrations preceded chronic inflammation and thereby associated with increased cardiovascular mortality [8] as well as, with a report on another Dutch cohort in whom apoE, interacting with apoA-II, was strongly linked to cardiovascular disease [9]. The group of Brenner indicated that apoE readily binds lipid antigens (apoE-lipid-antigen complexes) via CD1 molecules and targets them to stimulate systemic immune responses [31]. They considered it likely that apolipoproteins might provide a pathway for the delivery of self-lipid antigens and contribute to inflammatory diseases such as atherosclerosis. Lp(a) lipoprotein may assume antigenic properties in oxidative stress and serve as a target for both apoA-I and apoE. Indeed, apoA-I has recently been reported to be combined during oxidation to LDL (apoAI-LDL), high levels of which could mark in a cross-sectional study coronary artery disease more accurately than CRP [32]. Furthermore, increased circulating haptoglobin, a polymorphic glycoprotein, may bind apoA-I during pro-inflammatory state and decrease the amount of free apoA-I for enzyme stimulation impairing cholesterol esterification [33]. As indicated in Table 6, our observations on apoA-I tending to be associated with an increased risk for the development of HtgB are consistent with dysfunctionality of high apoA-I concentrations we previously documented to be diabetogenic [12] (or atherogenic [10]).
Table 6

Distinctive features of the two hypertriglyceridemic dyslipidemias

 

Atherogenic dyslipidemia

Hypertriglyceridemic hyperapoB

Men

Women

Men

Women

Prevalencea (%)

32.5

31

18

21.5

Age

c

c

c

Independent

Statin usage

c

c

c

Independent

Waist circumference

Slt. indep.

c

Slt. indep.

c

HOMA

c

Independent

c

Slt. indep.

1.44-foldb increment in serum apoE

Triples or doubles the risk

ApoA-I concentration and properties

Low, functional and protective

High, dysfunctional and risk factor

aData based on findings in Table 3

bEquivalent to 1 SD

cNot independently related to dyslipidemia

Gender and C-reactive protein were not independently related to dyslipidemia

HOMA homeostasis model assessment

Hypothesis on the Clinical Significance of apoE Concentrations

We postulate that high serum apoE concentration (which may mediate the presentation of Lp(a) as lipid antigen to the immune system) is associated with elevated apoB secondary to itself or apoA-I binding to Lp(a), rendering the latter unassayable and setting free apoB. This further contributes to oxidative stress, prominent in the females, and promotes the HtgB. Both the HtgB and the induced apoA-I and HDL particle dysfunctionality impart increased cardiometabolic risk (Fig. 1).
https://static-content.springer.com/image/art%3A10.1007%2Fs11745-012-3724-8/MediaObjects/11745_2012_3724_Fig1_HTML.gif
Fig. 1

Schematic representation of the link of elevated apoE concentrations to atherogenic dyslipidemia (AthD) and hypertriglyceridemia with elevated apoB (HtgB), the latter being enhanced by dysfunction of apoA-I interacting with the female sex and further promoting HDL dysfunction. AthD, a combination between Htg and selectively low levels of high-density lipoprotein (HDL) and often of apoA-I, is determined primarily by insulin resistance and its related inflammation. HtgB is linked to elevated concentrations of both apoA-I and apoE, added to that of apoB. The postulated underlying determinants are enhanced subclinical inflammation inducing and being mediated by excess oxidized phospholipids associated with elevated lipoprotein [Lp](a). This process is aggravated by the presumed aggregation of the anti-inflammatory apolipoproteins A-I and E and [Lp](a) lipoprotein, predominantly valid in females. Enrichment of apoE in HDL subfractions is an element in the development of dysfunctional HDL. Dysfunctional apoA-I and HDL is associated not with low levels, and the pro-inflammatory state is largely independent of apoE genotype, abdominal obesity and elevated C-reactive protein. Cardiometabolic outcomes of metabolic syndrome (MetS), diabetes (DM) and coronary heart disease (CHD) are strongly affected by these processes. Please, refer to text for further details. PLTP phospholipase transfer protein

Due to the link of serum apoE concentrations to atherogenic dyslipidemia, these levels were significantly associated in both genders also with the MetS likelihood additively to apoE genotype, age, apoB, apoA-I and CRP concentrations. Hence, recognition of elevated apoE concentrations may indicate the presence of atherogenic dyslipidemia and/or MetS.

In essential agreement with the report of Corsetti et al. [9.], we found that prevalent and incident CHD was associated with excess serum apoE, beyond apoE genotype and CRP elevation. Analysis of purely incident CHD provided similar results though the sample size was limited. It is readily conceivable that cardiometabolic risk is affected by a threshold of elevated apoE (perhaps beyond 5 mg/dl) rather than linearly.

Limitations and Strengths

The essentially cross-sectional design of the study may limit a cause-effect relationship between apoE and the dyslipidemias or apoA-I level or the cardiometabolic risk, but a converse relationship of apoA-I in regard to the two specific dyslipidemias on prospective analysis supports the stipulation. Residual confounders such as socioeconomic status and alcohol consumption might possibly modify the findings but not substantially. The study sample representing a general population, adjustment for confounders including apoE polymorphism, CRP and HOMA index, availability of apoB values and the robustness of the associations form the strengths of the study.

Conclusion

Probably via the related remnant lipoproteins, excess serum apoE represents a marked pro-inflammatory state concurrently with apoA-I dysfunctionality. An increment of 1-SD in apoE concentrations more than doubles the risk of the atherogenic dyslipidemia and confers MetS risk independently. Serum apoE is further log-linearly positively associated with apoA-I, independently and additively to apoB and waist circumference, thereby leading to HtgB, that likely represents an autoimmune state and is a main determinant of CHD in many populations.

Acknowledgments

We thank the Turkish Society of Cardiology and the pharmaceutical companies AstraZeneca and Servier, İstanbul, which have financially supported the Turkish Adult Risk Factor surveys and thank the Turkish Ministry of Health for logistic support. We appreciate the input of G. Hergenç, Ph.D., in biochemical analyses and the dedicated works of G. Çiçek, MD, and Mr. M. Özmay in the survey teams.

Conflict of interest

No conflict of interest is declared by any author.

Copyright information

© AOCS 2012