Apolipoprotein E levels and apolipoprotein E genotypes in incident cardiovascular disease risk in subjects of the Prevention of Renal and Vascular End-stage disease study

Journal of Clinical Lipidology (2016) -, -–-

Original Article

Apolipoprotein E levels and apolipoprotein E genotypes in incident cardiovascular disease risk in subjects of the Prevention of Renal and Vascular End-stage disease study James P. Corsetti, MD, PhD*, Ron T. Gansevoort, MD, PhD, Stephan J. L. Bakker, MD, PhD, Robin P. F. Dullaart, MD, PhD Department of Pathology and Laboratory Medicine, University of Rochester School of Medicine and Dentistry, Rochester, NY, USA (Dr Corsetti); Department of Nephrology, University of Groningen and University Medical Center Groningen, The Netherlands (Drs Gansevoort, and Bakker); and Department of Endocrinology, University of Groningen and University Medical Center Groningen, The Netherlands (Dr Dullaart) KEYWORDS: Apolipoprotein E; Cardiovascular disease; Lipoproteins; Apolipoprotein E polymorphism; Apolipoprotein B

BACKGROUND: Apolipoprotein E (apoE) is a component of all major lipoprotein classes with multiple functions including clearance of circulating triglyceride-rich lipoprotein particles and hepatic production of triglyceride-rich lipoprotein, thus affording several avenues for apoE involvement in atherosclerosis development. ApoE has 3 isoforms (E2, E3, and E4) based on a common genetic polymorphism. Numerous studies have been performed assessing cardiovascular disease (CVD) risk relative to the 6 resulting genotypes; however, surprisingly, few studies have been performed assessing risk attributable to apoE plasma levels either alone or in addition also taking into account apoE genotypes. OBJECTIVE: To examine the role of apoE levels together with apoE genotypes on incident CVD risk in a large population-based cohort and also to afford preliminary characterization of atherogenic apoE-containing lipoprotein particles. METHODS: Cox multivariable proportional hazards modeling was performed on a cohort of the Prevention of Renal and Vascular End-Stage Disease (PREVEND) study as a function of apoE levels and apoE genotypes adjusted for age, gender, and past history of CVD. Further modeling was performed with single addition of clinical and biomarker parameters to elucidate the nature of apoE-associated risk. RESULTS: High apoE levels were demonstrated to be associated with CVD risk (hazard ratio per apoE standard deviation, 1.20; 95% confidence interval, 1.11–1.31; P , .0001) both overall and within the high-frequency apoE genotype groups (ε2ε3, ε3ε3, and ε3ε4). Only on addition of apoB-containing lipoprotein parameters to models, did apoE levels lose association with risk. CONCLUSIONS: ApoE levels positively associate with incident CVD risk with apoE-associated risk likely residing in apoB-containing lipoproteins. Ó 2016 National Lipid Association. All rights reserved.

* Corresponding author. Department of Pathology and Laboratory Medicine, University of Rochester Medical Center, Box 626, 601 Elmwood Avenue, Rochester, NY 14642, USA. E-mail address: [email protected] Submitted March 4, 2016. Accepted for publication March 5, 2016.

1933-2874/Ó 2016 National Lipid Association. All rights reserved. http://dx.doi.org/10.1016/j.jacl.2016.03.003

2

Introduction Apolipoprotein E (apoE) is a 34 kDa protein comprised of 299 amino acid residues.1 Its encoding gene, APOE, is located on chromosome 19 in close proximity to APOC1, APOC2, and APOC4.2,3 ApoE is a component of all major lipoprotein classes, and as such, it has multiple functions in metabolism and inflammation-related processes including clearance of circulating atherogenic triglyceride-rich lipoproteins (TRLs); hepatic secretion of very low–density lipoprotein (VLDL); stimulation of cholesterol efflux; inhibition of platelet aggregation; inhibition of T-lymphocyte, smooth muscle cell, and endothelial cell proliferation; and inhibition of inflammation and oxidative stress.4–8 The apoE gene demonstrates 3 well-known common alleles, ε2, ε3 (wild type), and ε4, which give rise to 3 protein isoforms: apoE2 (Cys112, Cys158), apoE3 (Cys112, Arg158), and apoE4 (Arg112, Arg158), respectively. The apoE isoforms are known to demonstrate differential functionality relating to clearance of TRL from the circulation based primarily on differences in affinities of the isoforms for various receptors mediating TRL uptake and on differences in affinities of the isoforms for the various lipoprotein classes.7,9–11 The differences in affinities derive from structural differences in apoE isoforms. The apoE N-terminal domain contains the lipoprotein receptor–binding site12–15; whereas the C-terminal domain exhibits the principal lipoprotein-binding region.13,14 The most notable change involving lipoprotein receptor binding is the decreased affinity of apoE2 arising from structural alterations close to the receptor-binding site brought about by the ε2 allele.16,17 With regard to lipoprotein preferences, apoE2 and apoE3 favor association with small phospholipid-rich high-density lipoprotein (HDL), whereas apoE4 favors association with large, triglyceriderich VLDL.16,18,19 In view of functional differences in apoE isoforms relating to lipoprotein metabolism, there has been a great deal of work oriented toward assessing potential associations of apoE isoforms with cardiovascular disease (CVD) risk.20–24 In this regard, recent meta-analyses have generally concluded associations of the ε4 allele with higher risk and the ε2 allele with lower risk.20–24 Specifically, results have demonstrated: (1) a 42% higher coronary risk for carriers of the ε4 allele in comparison to ε3ε3 individuals and no significant risk association for the ε2 allele20; (2) slightly higher risk for ε4 carriers compared with ε3ε3 individuals and 20% lower coronary risk for ε2 carriers21; (3) the ε4 allele to be a risk factor for development of myocardial infarction (MI) and the ε2 allele to be a protective factor for development of MI22; (4) a tentative association of increased risk with the ε4 allele23; and (5) suggestion of increased risk for the ε4 allele and no association with risk for the ε2 allele.24 However, also should be noted are studies not supportive of such results.5,25 In contrast to the extensive work on apoE isoforms and CVD risk, there have been remarkably few studies dealing with associations of apoE levels with risk and even fewer

Journal of Clinical Lipidology, Vol -, No -, - 2016 studies assessing potential roles for interactions of apoE levels with apoE isoforms in the establishment of risk.26–30 Thus, the current work was conducted in a large cohort of the Prevention of Renal and Vascular End-Stage Disease (PREVEND) study to assess the potential association of apoE levels with incident CVD risk, to assess potential associations of interaction of apoE levels with apoE genotypes in risk, and to preliminarily characterize apoEcontaining atherogenic lipoprotein particles.

Methods Subjects for the current work were participants of PREVEND, a large prospective general population-based study begun in 1997 to investigate CVD and renal disease with focus on albuminuria.31 Briefly, a questionnaire was sent to all inhabitants (28–75 years old, N 5 85,421) of the city of Groningen, the Netherlands, requesting demographic and CVD morbidity data and to supply an early morning urine specimen. Response rate was 47.8%. Included in the study were all subjects with urine albumin levels $10 mg/L and a group of randomly selected subjects with urine albumin level ,10 mg/L. Exclusions included subjects with insulin-using diabetes mellitus and pregnant women after which resulted the PREVEND cohort of 8592 study subjects. CVD outcomes were followed and included cardiovascular mortality and any of the following at hospitalization: fatal and nonfatal MI, ischemic heart disease, percutaneous transluminal coronary angioplasty, and coronary artery bypass grafting. The municipal register was the source of mortality data with cause of death obtained by linkage of death certificate number to primary cause of death (Dutch Central Bureau of Statistics). Hospital morbidity data were from PRISMANT (Dutch national registry of hospital discharge diagnoses). Follow-up time was from initial urine collection in 1997 to date of either first CVD event or study termination (31 December 2008) if no CVD event. For the present study, there were additional exclusions as follows: lack of apoE level or apoE genotyping and lack of metabolic syndrome (MetS) status. This resulted in a study group of 5485 subjects with mean follow-up time of 9.4 years. The PREVEND study has been approved by the Medical Ethics Committee of the University of Groningen. Informed written consent was obtained from all participants. Serum and plasma were prepared from venous blood samples with collection after overnight fast and after 15 minutes of rest with laboratory analyses performed after storage overnight at 220 C. Lipids, lipoproteins, highsensitivity C-reactive protein (CRP), and glucose were determined by standard analytical methods as described previously.32,33 Low-density lipoprotein cholesterol (LDL-C) was determined from the Friedewald equation. ApoE genotyping was performed as described previously.34,35 Clinical variables were defined as follows. Past cardiac history included: before entry into PREVEND

Corsetti et al

ApoE levels and apoE genotypes in coronary risk

hospitalization for MI, revascularization procedures, or obstructive coronary artery disease. MetS was assessed according to the revised National Cholesterol Education Program - Adult Treatment Panel-III criteria36; namely, a subject was categorized as having MetS if 3 or more of the following criteria were met: waist circumference .88 cm for women and .102 cm for men; hypertension (blood pressure $ 130/85 mm Hg or antihypertension drug use); fasting plasma triglyceride level $ 1.70 mmol/ L; HDL-C , 1.30 mmol/L for women and ,1.00 for men; and fasting plasma glucose level $ 5.6 mmol/L. Urinary albumin concentration was measured by nephelometry (BNII; Dade Behring, Marburg, Germany). Two 24-hour urine collections were performed; urinary albumin excretion (UAE) was calculated as the mean value of the urine collections. The serum creatinine–based Chronic Kidney Disease Epidemiology Collaboration equation37 was used to calculate estimated glomerular filtration rate (eGFR). Smoking status entailed a trilevel variable with the following values: never smoker/former smoker/current smoker. Ethanol use was a bilevel variable with values: ,1 drink per day/$1 drink per day. Data and graphical analyses were performed using Statistica 12 (StatSoft, Inc, Tulsa, OK). Results are reported as means 6 standard deviations (SDs) for normally distributed variables and medians (interquartile ranges) for non-normally distributed variables (triglycerides, triglyceride/apolipoprotein B (apoB) ratio, UAE, and CRP). Non-normal variables were logarithmically transformed for subsequent statistical analyses. Kaplan–Meier plots were used to follow event occurrence over time with log-rank statistic for assessing statistical significance. Multivariable Cox proportional hazards modeling was used to follow CVD outcomes over time with continuous independent variables (including the logarithmically transformed variables as cited previously) standardized by transformation to distributions having means of 0 and SDs of 1. Hazard ratios (HRs) for continuous variables are reported per 1 SD unit. Two-sided P values ,.05 were considered statistically significant.

Results Table 1 summarizes clinical and laboratory parameters in the study population as a function of the 6 genotypes resulting from the 3 common allelic variants (ε2, ε3, and ε4) of APOE. There were no significant differences in demographic variables across the various APOE genotype groups. All the lipid and lipoprotein markers demonstrated significant differences with regard to apoE genotypes. Except for HDL-C, triglycerides, apolipoprotein A-I (apoA-I), and cholesterol/HDL-C ratio, lipid and lipoprotein markers demonstrated either generally smooth increases across subjects in Table 1 from ε2ε2 to ε4ε4 (cholesterol, non–HDL-C, LDL-C, apoB, and apoB/apoAI ratio) or generally smooth decreases (apoE and

3 triglyceride/apoB ratio). CRP levels were significantly different with smooth decreases exhibited separately for ε2 carriers and ε3 carriers. Glucose levels, eGFR, and UAE were not significantly different as a function of apoE genotypes. Analyses were performed to determine effects of MetS and diabetes on apoE levels and the triglyceride/apoB ratio. Results (median, interquartile range, P value) demonstrated for subjects with MetS in comparison to cases without MetS higher levels of apoE (g/L; 0.0429, 0.0346–0.0536 vs 0.0358, 0.0287–0.0444; P , .0001) and higher values of the triglyceride/apoB ratio (mmol/g; 1.48, 1.14–2.07 vs 1.12, 0.88–1.49; P , .00001). Similar results were found in subjects with diabetes for apoE (g/L; 0.0435, 0.0349– 0.0566 vs 0.0367, 0.0293–0.0455; P , .00001) and the triglyceride/apoB ratio (mmol/g; 1.58, 1.23–2.25 vs 1.16, 0.90–1.55; P , .00001). In addition, further analyses of apoE levels and the triglyceride/apoB ratio were performed as a function of apoE genotypes for MetS (Table 2) and diabetes (Table 3). Results generally showed increased apoE and triglyceride/apoB ratios in MetS and diabetic subjects irrespective of apoE genotypes. We next elucidated potential relationships between apoE levels and CVD risk and between apoE genotypes and CVD risk. For apoE levels, 2 Cox proportional hazards models were run. Model 1 (adjusted for age, gender, and past history of CVD) revealed HR (per SD), 95% confidence interval, and P value as follows for apoE levels: 1.13, 1.05 to 1.22, and .002. Corresponding results for model 2 (adjusted for age, gender, past history of CVD, and apoE genotype) were: 1.20, 1.11 to 1.31, and ,.0001 (Table 4). To elucidate the nature of the link of CVD risk with increasing levels of apoE, Cox proportional hazards modeling was performed as a function of apoE levels with single addition to the model of clinical and laboratory parameters. Two models were run. Model 1 was adjusted for age, gender, and past history of CVD; model 2 was adjusted for age, gender, past history of CVD, and apoE genotype. HR results for apoE levels demonstrated continued statistical significance with single addition of all the clinical parameters as well as UAE, eGFR, CRP, and glucose. Of note, apoE significance was maintained after adjustment for both apoA-I and HDL-C. However, with single addition of all lipid and lipoprotein markers associated with apoB (apoB, cholesterol, non–HDL-C, LDL-C, triglycerides, cholesterol/HDL-C ratio, and apoB/apoA-I ratio), statistical significance of the apoE HR was lost. This was strongly suggestive of dependence of apoE risk on levels of apoBcontaining lipoproteins. In addition, analyses were extended to explore effects on risk focusing on potential interactions of apoE levels with apoE genotypes. Thus, proportional hazards modeling was performed separately for the higher frequency genotype groups (ε2ε3, ε3ε3, and ε3ε4) as a function of apoE levels (continuous variable) with models adjusted for age, gender, and past history of CVD. Results for apoE levels (HR, 95% confidence interval, P value) in the 3 genotype

4

Table 1

Clinical covariates and biomarker levels (mean 6 standard deviation or median [interquartile range]) in the study population according to apoE genotypes Total population

ε2ε2

N (%) Age (y) Females (%) Cardiac events (%) BMI (kg/m2) Systolic BP (mm Hg) Diastolic BP (mm Hg) Past cardiac history (%) Diabetes mellitus (%) MetS (%) Statins (%) Antihypertensives (%) Smoking status (%) Never Former Current Ethanol use (%) Cholesterol (mmol/L) Non–HDL-C (mmol/L) LDL-C (mmol/L) HDL-C (mmol/L) Triglycerides (mmol/L) ApoE (g/L) ApoB (g/L) ApoA-I (g/L) Cholesterol/HDL-C ApoB/ApoA-I Triglyceride/apoB UAE (mg/24 h) eGFR (mL/min/ 1.73 m2) CRP (mg/L) Glucose (mmol/L)

5485 50.2 53.8 7.8 26.21 130 74

46 (0.8) 50.8 6 50.0 13.0 26.22 6 130 6 74 6

6 12.7 6 4.20 6 21 6 10

13.6

4.93 19 10

ε2ε3

ε2ε4

699 (12.7) 50.8 6 12.5 57.5 4.6 26.39 6 4.28 130 6 21 75 6 10

134 (2.4) 49.6 6 49.3 6.0 25.83 6 127 6 74 6

12.9

3.81 22 10

ε3ε3

ε3ε4

ε4ε4

3121 (56.9) 50.3 6 12.7 53.2 8.3 26.25 6 4.19 130 6 21 74 6 10

1345 (24.5) 49.7 6 12.8 54.2 8.6 26.12 6 4.22 129 6 20 74 6 10

140 (2.6) 48.5 6 51.4 5.7 25.49 6 127 6 73 6

P* 11.6

4.06 18 10

P† .28 .29 .008 .18 .52 .55

.35 .67 .28

5.2

9.1

3.5

1.5

5.4

5.2

4.4

.09

3.9

2.2

4.5

2.3

4.2

3.2

2.2

.38

17.9 4.7 16.6

19.6 4.4 19.6

18.5 3.3 16.3

16.4 2.2 15.7

17.9 4.9 16.4

18.5 5.4 17.8

10.0 3.6 12.1

.24 .21 .57 .99

28.8 38.0 33.2 24.9 5.67 6 1.14 4.33 6 1.22 3.69 6 1.06 1.33 6 0.40 1.16 (0.85–1.69)

0.8 0.8 0.9 34.8 5.32 6 2.08 3.94 6 2.22 3.02 6 1.63 1.38 6 0.43 1.48 (0.87–2.38)

12.4 12.6 13.1 24.4 5.32 6 1.04 3.94 6 1.17 3.28 6 0.99 1.38 6 0.42 1.18 (0.86–1.69)

2.5 2.6 2.2 29.9 5.43 6 1.25 4.12 6 1.40 3.39 6 1.20 1.31 6 0.43 1.23 (0.87–1.73)

57.4 56.9 56.6 24.2 5.69 6 1.12 4.35 6 1.19 3.72 6 1.04 1.34 6 0.40 1.15 (0.84–1.66)

24.6 24.5 24.4 25.5 5.82 6 1.13 4.51 6 1.19 3.85 6 1.04 1.31 6 0.39 1.18 (0.88–1.74)

2.4 2.6 2.8 29.3 .24 5.94 6 1.22 ,1026 ,1026 26 4.63 6 1.32 ,10 ,1026 26 3.98 6 1.20 ,10 ,1026 1.31 6 0.35 .005 .007 1.24 (0.90–1.68) .001 ,1023

0.0394 6 0.0159 0.0950 6 0.0373 0.0503 6 0.0163 0.0484 6 0.0170 0.0380 6 0.0129 0.0354 6 0.0147 0.0293 6 0.0111 1.04 6 0.29 0.59 6 0.19 0.92 6 0.26 0.97 6 0.32 1.05 6 0.29 1.10 6 0.29 1.13 6 0.30 1.40 6 0.29 1.45 6 0.35 1.44 6 0.29 1.40 6 0.30 1.40 6 0.29 1.38 6 0.28 1.38 6 0.27 4.65 6 1.78 4.28 6 2.33 4.24 6 1.67 4.69 6 2.25 4.66 6 1.74 4.85 6 1.81 4.91 6 1.94 0.77 6 0.26 0.42 6 0.16 0.66 6 0.23 0.72 6 0.28 0.77 6 0.26 0.82 6 0.26 0.85 6 0.28 1.17 (0.91–1.57) 2.69 (2.07–3.53) 1.33 (1.05–1.80) 1.35 (1.01–1.76) 1.14 (0.89–1.52) 1.15 (0.88–1.53) 1.12 (0.90–1.50) 9.65 (6.38–18.67) 7.53 (5.67–14.6) 9.63 (6.25–19.8) 9.66 (6.11–17.1) 9.81 (6.46–18.8) 9.41 (6.38–18.1) 8.21 (5.65–21.0) 69.7 6 13.9 71.8 6 15.8 69.3 6 14.7 69.9 6 14.2 69.6 6 13.8 69.8 6 13.8 72.7 6 13.7 1.39 (0.60–3.09) 4.91 6 1.25

1.82 (0.87–3.86) 1.45 (0.68–3.50) 1.06 (0.55–2.10) 1.52 (0.64–3.33) 1.14 (0.50–2.64) 0.82 (0.35–1.77) 4.63 6 0.61 4.96 6 1.27 4.87 6 1.19 4.91 6 1.23 4.91 6 1.34 4.77 6 1.11

,1026 ,1026 ,1023 ,1026 ,1026 ,1026 .61 .15 .004 .35

,1026 ,1026 ,1026 ,1026 ,1026

.001

.66 .43 .011 .34

ApoB, apolipoprotein B; apoE, apolipoprotein E; apoA-I, apolipoprotein A-I; BMI, body mass index; BP, blood pressure; CRP, C-reactive protein; eGFR, estimated glomerular filtration rate; HDL-C highdensity lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; MetS, metabolic syndrome; non–HDL-C, non–high-density lipoprotein cholesterol; UAE, urinary albumin excretion. *Chi-square testing for categorical variables and raw analysis of variance (ANOVA) for continuous variables. †ANOVA adjusted for gender and age.

Journal of Clinical Lipidology, Vol -, No -, - 2016

Parameter

Corsetti et al

ApoE levels and apoE genotypes in coronary risk

5

Table 2 Medians and interquartile ranges (IQRs) for apoE concentration and triglyceride/apoB (trig/apoB) ratio for subjects without past history of CVD as a function of apoE genotype for subjects with and without MetS along with corresponding comparisons (MannWhitney U) Metabolic syndrome Genotype

N

ApoE (g/L) ε2ε2 9 ε2ε3 121 ε2ε4 21 ε3ε3 492 ε3ε4 226 ε4ε4 14 Trig/ApoB (mmol/g) ε2ε2 9 ε2ε3 121 ε2ε4 21 ε3ε3 485 ε3ε4 225 ε4ε4 14

Without metabolic syndrome

Median

IQR

N

Median

IQR

P

0.1270 0.0549 0.0578 0.0426 0.0385 0.0361

0.1011–0.1435 0.0478–0.0660 0.0502–0.0802 0.0352–0.0518 0.0319–0.0475 0.0277–0.0421

33 554 111 2464 1050 120

0.0843 0.0470 0.0453 0.0355 0.0317 0.0263

0.0617–0.1005 0.0384–0.0556 0.0376–0.0520 0.0290–0.0426 0.0259–0.0396 0.0204–0.0346

.045 ,.0001 .001 ,.0001 ,.0001 .006

2.79–5.00 1.34–2.37 1.41–3.23 1.12–1.96 1.05–1.86 1.03–1.50

32 547 110 2446 1043 120

2.52 1.25 1.25 1.08 1.09 1.09

1.83–3.15 1.02–1.67 0.96–1.61 0.85–1.41 0.85–1.44 0.88–1.47

.065 ,.0001 ,.0001 ,.0001 ,.0001 .19

3.06 1.67 1.98 1.47 1.39 1.26

ApoB, apolipoprotein B; apoE, apolipoprotein E; CVD, cardiovascular disease; MetS, metabolic syndrome.

groups demonstrated a tendency toward increased risk with increasing apoE levels as follows: ε2ε3 (1.22, 0.90–1.66, 0.21), ε3ε3 (1.22, 1.09–1.36, 0.00035), and ε3ε4 (1.15, 0.98–1.35, 0.091). Figure 1 presents corresponding Kaplan–Meier plots for the ε2ε3 (Fig. 1A), ε3ε3 (Fig. 1B), and ε3ε4 (Fig. 1C) genotype groups as a function of apoE levels (tertiles). Hence, both regression and survival analysis results demonstrated increasing risk with increasing apoE levels for each of the 3 genotype groups, and as a consequence, there was no suggestion

of interaction of apoE levels with apoE genotypes in generating risk. To address whether enrichment of the PREVEND cohort with albuminuric subjects affected findings of the current work, sensitivity analyses studies were performed on the group of subjects remaining (N 5 2144) after removal of albuminuric subjects initially added to the cohort for enrichment. Results demonstrated findings similar to those of the full study group. First, regarding apoE levels and CVD risk, results again demonstrated significant positive

Table 3 Medians and interquartile ranges (IQRs) for apoE concentration and triglyceride/apoB (trig/apoB) ratio for subjects without past history of CVD as a function of apoE genotype for subjects with and without diabetes mellitus along with corresponding comparisons (Mann-Whitney U) Diabetes Genotype

N

ApoE (g/L) ε2ε2 1 ε2ε3 27 ε2ε4 3 ε3ε3 102 ε3ε4 41 ε4ε4 3 Trig/ApoB (mmol/g) ε2ε2 1 ε2ε3 27 ε2ε4 3 ε3ε3 100 ε3ε4 41 ε4ε4 3

Without diabetes Median

IQR

N

Median

IQR

P

0.0713 0.0577 0.0578 0.0431 0.0391 0.0270

— 0.0478–0.0706 0.0537–0.0802 0.0357–0.0527 0.0304–0.0491 0.0136–0.0379

40 644 128 2813 1218 129

0.0912 0.0480 0.0460 0.0362 0.0325 0.0269

0.0650–0.1228 0.0395–0.0565 0.0383–0.0553 0.0294–0.0438 0.0265–0.0406 0.0209–0.0358

— .006 .063 ,.0001 .002 .76

4.20 1.85 1.44 1.58 1.46 1.69

— 1.36–2.39 0.88–3.23 1.27–2.14 1.16–2.27 1.43–1.80

39 637 127 2790 1210 129

2.66 1.32 1.33 1.12 1.12 1.11

1.98–3.39 1.04–1.74 1.01–1.73 0.87–1.49 0.87–1.50 0.90–1.45

— ,.0001 .72 ,.0001 ,.0001 .48

Note that only 1 ε2ε2 subject had diabetes.

Journal of Clinical Lipidology, Vol -, No -, - 2016

6

Table 4 Cox proportional hazards modeling results for apoE concentration (HRs of apoE per apoE SD, 95% CI, and P values) with single addition in each case of the listed parameter Model 1

Model 2

Parameter

ApoE HR

95% CI

P

ApoE HR

95% CI

P

Crude Diabetes mellitus (%) MetS (%) Statins Antihypertensives (%) Smoking status Ethanol use BMI Systolic BP Diastolic BP (mm Hg) ApoA-I HDL-C ApoB Cholesterol Non–HDL-C LDL-C Triglycerides (mmol/L) Cholesterol/HDL-C ApoB/ApoA-I UAE (mg/24 h) eGFR (mL/min/1.73 m2) CRP Glucose

1.13 1.12 1.11 1.14 1.12 1.13 1.13 1.11 1.11 1.11 1.15 1.10 1.06 1.02 0.99 1.03 1.01 1.00 1.06 1.11 1.13 1.11 1.11

1.05–1.22 1.03–1.21 1.03–1.20 1.05–1.22 1.04–1.21 1.04–1.22 1.05–1.22 1.03–1.20 1.03–1.20 1.03–1.20 1.07–1.24 1.01–1.19 0.97–1.16 0.94–1.12 0.90–1.08 0.93–1.14 0.91–1.11 0.92–1.09 0.97–1.16 1.03–1.20 1.05–1.22 1.02–1.20 1.02–1.20

.002 .006 .009 .001 .004 .002 .002 .009 .008 .008 ,.001 .02 .20 .60 .82 .59 .91 .95 .68 .006 .002 .17 .01

1.20 1.19 1.17 1.21 1.20 1.19 1.20 1.18 1.18 1.18 1.22 1.16 1.09 1.08 1.02 1.11 1.06 1.04 1.08 1.18 1.20 1.18 1.17

1.11–1.31 1.10–1.30 1.08–1.27 1.12–1.31 1.10–1.30 1.10–1.30 1.11–1.31 1.08–1.28 1.08–1.28 1.09–1.28 1.13–1.33 1.06–1.26 0.98–1.21 0.97–1.20 0.92–1.14 0.97–1.27 0.95–1.19 0.94–1.14 0.97–1.19 1.08–1.27 1.11–1.31 1.08–1.28 1.07–1.28

,.0001 ,.0001 .0002 ,.0001 ,.0001 ,.0001 ,.0001 .0002 .0001 ,.0001 ,.0001 .0007 .11 .17 .69 .11 .27 .49 .15 ,.0001 ,.0001 .0001 .0003

ApoB, apolipoprotein B; apoE, apolipoprotein E; apoA-I, apolipoprotein A-I; BMI, body mass index; BP, blood pressure; CI, confidence interval; CRP, C-reactive protein; eGFR, estimated glomerular filtration rate; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; MetS, metabolic syndrome; non–HDL-C, non–high-density lipoprotein cholesterol; SD, standard deviation; UAE, urinary albumin excretion. Model 1 was adjusted for age, gender, and past history of cardiovascular disease (CVD); model 2 was adjusted for age, gender, past history of CVD, and apoE genotype.

association of apoE levels and risk. Second, analyses to examine the nature of apoE-containing particles were repeated using Cox proportional hazards modeling as a function of apoE levels with single addition of clinical and laboratory parameters to models. Results were generally similar to those of the full cohort in that again only addition of apoB lipids and lipoproteins to models resulted in loss of significant association of apoE levels with risk. Third, regarding potential interactions of apoE levels with apoE genotypes in affecting risk, models were again run separately for ε2ε3, ε3ε3, and ε3ε4 genotype subjects as a function of apoE levels with adjustments for age, gender, and past history of CVD. Results again agreed closely with those of the full cohort in that each genotype group demonstrated positive association of apoE levels with risk.

Discussion The present study was undertaken to investigate associations of apoE levels, apoE genotypes, and apoE levels together with apoE genotypes with incident CVD risk in a large population of healthy individuals using Cox

proportional hazards multivariable modeling. Results revealed highly significant positive association of apoE levels with incident CVD risk in models adjusted for age, gender, past history of CVD, and apoE genotype; and in this regard, the apoE-associated risk was largely attributable to apoBcontaining lipoproteins independent of significant risk associations of HDL-C and CRP. In addition, results indicated no association with risk for the interaction of apoE levels with apoE genotype in ε2ε3, ε3ε3, or ε3ε4 subjects. Finally, relationships of apoE levels and triglyceride/apoB ratio with apoE genotypes as a function of MetS and diabetes were examined. Results revealed decreasing levels of apoE across apoE genotypes (ε2ε2–ε4ε4) and higher levels of apoE in both MetS and diabetic subjects. Results were similar for the triglyceride/apoB ratio. Increasing apoE levels in the study population were found to be positively associated with CVD risk in models adjusted for age, gender, past history of CVD, and apoE genotype. Regarding the few reported studies in this area,26–30 increasing apoE levels were also found to be positively associated with CVD mortality: in old age in models adjusted for gender26; with stroke in old age in models adjusted for gender, body mass index,

Corsetti et al

ApoE levels and apoE genotypes in coronary risk

Proportion without Events

A 1.00

0.95

0.90

0.85

Proportion without Events

B

Proportion without Events

1000

2000

3000

4000

Time (Days) 1.00

0.95

0.90

0.85

C

0

0

1000

2000

3000

4000

Time (Days) 1.00

0.95

0.90

0.85

0

1000

2000

3000

4000

Time (Days) Figure 1 Kaplan–Meier plots for apoE genotype subjects (ε2ε3, ε3ε3, and ε3ε4) as a function of apoE tertiles. (A) ε2ε3 subjects (P 5 .10, log-rank); (B) ε3ε3 subjects (P 5 .015, log-rank); and (C) ε3ε4 subjects (P 5 .0002, log-rank). Curves corresponding to apoE tertiles were represented as follows: tertile 1—dashed line, tertile 2—dotted line, and tertile 3—solid line.

hypertension, diabetes, and levels of triglycerides, HDL-C, and LDL-C27; and with incident CVD in Turkish adults.28 Furthermore, apoE levels have been demonstrated to be associated with incident CVD risk32,38 in women but not men having concurrently high levels of HDL-C and CRP. In addition, there have been several studies

7 investigating apoE levels in connection with CVD risk as a function of specific lipoprotein classes. Such findings include: levels of apoE in HDL to be associated with high recurrent CVD risk39; and plasma levels of apoE to be elevated in MetS patients in connection with high levels of large VLDL and low-density lipoprotein particles.34 On the other hand, apoE levels were not found to be higher in MI patients in a case-control study29 and were not significantly associated with degree of atherosclerosis in a coronary artery angiography study.30 In addition, apoE levels in apoC-III–containing VLDL and lowdensity lipoprotein particles have been shown to be associated with lower CVD risk.40 The present study also revealed that apoE-associated risk was closely linked to apoB-containing lipoproteins and not connected with the independent risk associations of HDL-C and CRP. The notion of apoE-associated risk closely connected with apoB-containing lipoproteins is consistent with findings of a coronary angiography casecontrol study assessing atheroma involvement as a function of apoE-enriched and apoE-poor TRLs.41 Results demonstrated, in comparison to controls, higher levels of apoE-enriched TRLs in cases and significant association of apoE-enriched TRLs with extent of involvement. In addition, the idea that apoE apparently plays little role in the independent risk association of HDL-C was supported by findings of a case-control study assessing the role of apoE content of lipoprotein classes in myocardial infarction in which no difference in apoE content of HDL between cases and controls was observed.40 As to the issue of risk in terms of apoE levels together with apoE genotypes, our results indicated no suggestion for the interaction of apoE levels and apoE genotypes as playing a role in the generation of CVD risk in ε2ε3, ε3ε3, or ε3ε4 subjects. This finding was similar to results of earlier studies that also demonstrated positive associations of apoE levels with CVD risk independent of apoE genotype.26,28 Regarding studies of apoE levels in MetS and diabetes, our results showing higher levels of apoE are consistent with results of several earlier studies.34,42–47 Of relevance, the current work extended investigations to include determination of apoE levels as a function of apoE genotypes. Results in both MetS and in diabetes demonstrated higher levels of apoE irrespective of apoE genotype that, in view of our aforementioned demonstrated association of apoE levels with CVD risk, is consistent with the known higher CVD risk in MetS and diabetes. Furthermore, the triglyceride/apoB ratio was assessed as a function of apoE genotypes in MetS and diabetic subjects. In both cases, the ratio decreased across apoE genotypes (ε2ε2–ε4ε4), and it was higher in MetS and diabetes. These findings are suggestive that apoB-containing lipoproteins are larger and more enriched with apoE in MetS and diabetic subjects of the present study in comparison to subjects free of these conditions.24 Our findings are consistent with the higher CVD risk associated with diabetes and MetS in view of

8 studies reporting the promotion of atherosclerosis by accumulation of apoE-enriched, large TRLs.34,41 There were strengths and limitations in the current work. A strength of the study focused on apoE as a key participant in multiple facets of lipoprotein metabolism. Specifically, the current work added to the few studies assessing the role of circulating apoE levels and apoE genotypes as potentially associated with CVD risk; and notably, this was performed on a large number of subjects in a populationbased cohort followed for mean follow-up time of 9.4 years. However, we did not measure particle size and apoE content according to lipoprotein subclasses, especially TRLs, and relied instead on the triglyceride/apoB ratio as an estimate of large triglyceride-rich apoB-containing lipoproteins. In summary, results of the present study in a large population-based cohort demonstrated positive association of apoE levels with incident CVD; and moreover, the association prevailed individually within the 3 highfrequency apoE genotype groups (ε2ε3, ε3ε3, and ε3ε4). Evidence was also presented linking apoE-associated risk with apoB-containing lipoproteins enriched with apoE. Additional analyses in MetS and diabetes subjects as a function of apoE genotypes demonstrated, for both conditions, that findings of generally higher levels of apoE for each apoE genotype group are consistent with the known higher levels of CVD risk in MetS and diabetes.

Acknowledgments Authors’ contributions: J.P.C. designed the study, performed data analysis, and drafted the article. R.T.G. and S.J.L.B. was involved with data collection and helped with study design and article review. R.P.F.D. designed the study, performed data analysis, and drafted the article. All authors read and approved the final version of the article. The contributions of R.T.G, S.J.L.B, and R.P.F.D. were made on behalf of the PREVEND study group, and we are indebted to all PREVEND collaborators. Dade Behring (Marburg, Germany) is acknowledged for supplying equipment (Behring Nephelometer II) to determine apolipoproteins and urinary albumin.

Financial disclosures This work is supported by grant Netherlands Heart Foundation (2001.005). The authors have no financial disclosures.

References 1. Rall SC, Weissgraber KH, Mahley RW. Human apolipoprotein E. The complete amino acid sequence. J Biol Chem. 1982;257: 4171–4178. 2. Myklebost O, Rogne S. A physical map of the apolipoprotein gene cluster on human chromosome 19. Hum Genet. 1988;78:244–247.

Journal of Clinical Lipidology, Vol -, No -, - 2016 3. Allan CM, Walker D, Segrest JP, Taylor JM. Identification of a new human gene (APOC4) in the apolipopotein E, C-I, and C-II gene locus. Genomics. 1995;28:291–300. 4. Curtiss LK, Boisvert WA. Apolipoprotein E and atherosclerosis. Curr Opin Lipidol. 2000;11:243–251. 5. Greenow K, Pearce NJ, Ramji DP. The key role of apolipoprotein E in atherosclerosis. J Mol Med. 2005;83:329–342. 6. Getz G, Reardon CA. Apoprotein E as a lipid transport and signaling protein in the blood, liver, and artery wall. J Lipid Res. 2009;50: S156–S161. 7. Mahley RW, Weisgraber KH, Huang Y. Apolipoprotein E: structure determines function, from atherosclerosis to Alzheimer’s disease to AIDS. J Lipid Res. 2009;50:S183–S188. 8. Jofre-Monseny L, Minihane AM, Rimbach G. Impact of apoE genotype on oxidative stress, inflammation and disease risk. Mol Nutr Food Res. 2008;52:131–145. 9. Luc F, Ducimetriere P, Bard JM, et al. Distribution of apolipoprotein E between apo B- and non apo B-containing lipoproteins according to apo E phenotype. Atherosclerosis. 1997;131:257–262. 10. Phillips MC. Apolipoprotein E isoforms and lipoprotein metabolism. IUBMB Life. 2014;66:616–623. 11. Hatters DM, Peters-Libeu CA, Weisgraber KH. Apolipoprotein E structure: insights into function. Trends Biochem Sci. 2006;31: 445–454. 12. Mahley RW. Apolipoprotein E: cholesterol transport protein with expanding role in cell biology. Science. 1988;240:622–630. 13. Weisgraber KH. Apolipoprotein E: structure-function relationships. Adv Protein Chem. 1994;45:249–302. 14. Mahley RW, Rall SC. Apolipoprotein E: far more than a lipid transport protein. Annu Rev Genomics Hum Genet. 2000;1:507–537. 15. Morrow JA, Arnold KS, Dong J, Balestra ME, Innerarity TL, Weisgraber KH. Effect of arginine 172 on the binding of apolipoprotein E to the low density lipoprotein receptor. J Biol Chem. 2000;275: 2576–2580. 16. Wilson C, Wardell MR, Weisgraber KH, Mahley RW, Agard DA. Three-dimensional structure of the LDL receptor-binding domain of human apolipoprotein E. Science. 1991;252:1817–1822. 17. Dong LM, Parkin S, Trakhanov SD, et al. Novel mechanisms for defective receptor binding of apolipoprotein E2 in type III hyperlipoproteinemia. Nat Struct Biol. 1996;3:718–722. 18. Dong LM, Weisgraber KH. Human apolipoprotein E4 domain interactions. Arginine 61 and glutamic acid 255 interact to direct the preference for very low density lipoproteins. J Biol Chem. 1996;271: 19053–19057. 19. Dong LM, Wilson C, Wardell MR, et al. Human apolipoprotein E. Role of arginine 61 in mediating the lipoprotein preferences of the E3 and E4 isoforms. J Biol Chem. 1994;269:22358–22365. 20. Song Y, Stampfer MJ, Liu S. Meta-analysis: apolipoprotein E genotypes and risk for coronary heart disease. Ann Intern Med. 2004; 141:137–147. 21. Bennet AM, Di Angelantonio E, Zheng Y, et al. Association of apolipoprotein E genotypes with lipid levels and coronary risk. JAMA. 2007;298:1300–1311. 22. Xu H, Li H, Liu J, et al. Meta-analysis of apolipoprotein E gene polymorphism and susceptibility of myocardial infarction. PLoS One. 2014;9:e104608. 23. Yin YW, Sun QQ, Zhang BB, et al. Association between apolipoprotein E gene polymorphism and the risk of coronary artery disease in Chinese population: evidence from a meta-analysis of 40 studies. PLoS One. 2013;8:e66924. 24. Zhang M, Gu W, Qiao S, Zhu E, Zhao Q, Lv S. Apolipoprotein E gene polymorphism and risk for coronary heart disease in the Chinese population: a meta-analysis of 61 studies including 6634 cases and 6393 controls. PLoS One. 2014;9:e95463. 25. Liu S, Ma J, Ridker PM, Breslow JL, Stampfer MJ. A prospective study of the association between APOE genotype and the risk of myocardial infarction among apparently healthy men. Atherosclerosis. 2003;166:323–329.

Corsetti et al

ApoE levels and apoE genotypes in coronary risk

26. Mooijaart SP, Berbee JF, van Heemst D, et al. ApoE plasma levels and risk of cardiovascular mortality in old age. PLoS Med. 2006;3: e176. 27. van Vliet P, Mooijaart SP, de Craen AJ, Rensen PC, van Heemst D, Westendorp RG. Plasma levels of apolipoprotein E and risk of stroke in old age. Ann NY Acad Sci. 2007;1100:140–147. 28. Onat A, Can G, Ornek E, Ayhan E, Erginel-Unaltuna N, Murat SN. High serum apolipoprotein E determines hypertriglyceridemic dyslipidemias, coronary disease and apoA-I dysfunctionality. Lipids. 2013; 48:51–61. 29. Martinelli N, Olivieri O, Shen GQ, et al. Additive effect of LRP8/APOER2 R952Q variant to APOE ε2/ε3/ε4 genotype in modulating apolipoprotein E concentration and the risk of myocardial infarction: a case-control study. BMC Med Genet. 2009;10:41. 30. Lima LM, Carvalho MG, Ferreira CN, et al. Atheromatosis extent in coronary artery disease is not correlated with apolipoprotein-E polymorphism and its plasma levels, but associated with cognitive decline. Curr Alzheimer Res. 2010;7:556–563. 31. Pinto-Sietsma SJ, Janssen WM, Hillege HL, Navis G, De Zeeuw D, De Jong PE. Urinary albumin excretion is associated with renal functional abnormalities in a nondiabetic population. J Am Soc Nephrol. 2000;11:1882–1888. 32. Corsetti JP, Gansevoort RT, Bakker SJ, Navis G, Sparks CE, Dullaart RP. Apolipoprotein E predicts incident cardiovascular disease risk in women but not men with concurrently high levels of highdensity lipoprotein cholesterol and C-reactive protein. Metab Clin Exp. 2012;61:996–1002. 33. Corsetti JP, Gansevoort RT, Sparks CE, Dullaart RP. Inflammation reduces HDL protection against primary cardiac risk. Eur J Clin Invest. 2010;40:483–489. 34. Gruppen EG, Dallinga-Thie GM, Bakker SJ, Dullaart RP. Plasma apoE is elevated in metabolic syndrome: importance of large very low density and low density lipoprotein particles. J Mol Biomark Diagn. 2015;6:1. 35. Blaauwwiekel EE, Beusekamp BJ, Sluiter WJ, Hoogenberg K, Dullaart RP. Apolipoprotein E genotype is a determinant of lowdensity lipoprotein cholesterol and its response to a low-cholesterol diet in type 1 diabetic patients with elevated urinary albumin excretion. Diabet Med. 1998;15:1031–1035. 36. Grundy SM, Cleeman JI, Daniels SR, et al. Diagnosis and management of the metabolic syndrome: an American Heart

9

37.

38.

39.

40.

41.

42.

43.

44.

45.

46.

47.

Association/National Heart, Lung, and Blood Institute Scientific Statement. Circulation. 2005;112:2735–2752. Levey AS, Stevens LA, Schmid CH, et al. CKD-EPI (Chronic Kidney Disease Epidemiology Collaboration): a new equation to estimate glomerular filtration rate. Ann Intern Med. 2009;150:604–612. Corsetti JP, Bakker SJ, Sparks CE, Dullaart RP. Apolipoprotein A-II influences apolipoprotein E-linked cardiovascular disease risk in women with high levels of HDL cholesterol and C-reactive protein. PLoS One. 2012;7:e39110. Sacks FM, Alaupovic P, Moye LA, et al. VLDL, apolipoproteins B, CIII, and E, and risk of recurrent coronary events in the cholesterol and recurrent events (CARE) trial. Circulation. 2000;102:1886–1892. Mendivil CO, Rimm EB, Furtado J, Sacks FM. Apolipoprotein E in VLDL and LDL with apolipoprotein C-III is associated with a lower risk of coronary heart disease. J Am Heart Assoc. 2013;2:e000130. Barbagallo CM, Rizzo M, Noto D, et al. Accumulation of apoEenriched triglyceride-rich lipoproteins in patients with coronary artery disease. Metab Clin Exp. 2006;55:662–668. Tan KC, Shiu SW, Wong Y, Tan S. Plasma apolipoprotein E concentration is an important determinant of phospholipid transfer protein activity in type 2 diabetes mellitus. Diabetes Metab Res Rev. 2006;22: 307–312. Fielding CJ, Castro GR, Donnner C, Fielding PE, Reaven GM. Distribution of apolipoprotein E in the plasma of insulin-dependent and noninsulin-dependent diabetics and its relation to cholesterol net transport. J Lipid Res. 1986;27:1052–1061. Olivieri O, Martinelli N, Bassi A, et al. ApoE ε2/ε3/ε4 polymorphism, ApoC-III/ApoE ratio and metabolic syndrome. Clin Exp Med. 2007;7: 164–172. Dullaart RP, Groen AK, Dallinga-Thie G, de Vries R, Sluiter WJ, van Tol A. Fibroblast cholesterol efflux to plasma from metabolic syndrome subjects is not defective despite low high-density lipoprotein cholesterol. Eur J Endocrinol. 2008;158:53–60. Soderlund S, Watanabe H, Ehnholm C, Jauhiainen M, Taskinen MR. Increased apolipoprotein E level and reduced high-density lipoprotein mean particle size associate with low high-density lipoprotein cholesterol and features of metabolic syndrome. Metab Clin Exp. 2010;59: 1502–1509. Dullaart RP, Kwakernaak AJ, Dallinga-Thie GM. The positive relationship of serum paraoxonase-1 activity with apolipoprotein E is abrogated in metabolic syndrome. Atherosclerosis. 2013;230:6–11.