LPL polymorphism (D9N) predicts cardiovascular disease risk directly and through interaction with CETP polymorphism (TaqIB) in women with high HDL cholesterol and CRP

Atherosclerosis 214 (2011) 373–376

Contents lists available at ScienceDirect

Atherosclerosis journal homepage: www.elsevier.com/locate/atherosclerosis

Short communication

LPL polymorphism (D9N) predicts cardiovascular disease risk directly and through interaction with CETP polymorphism (TaqIB) in women with high HDL cholesterol and CRP James P. Corsetti a,∗ , Ron T. Gansevoort b , GerJan Navis b , Charles E. Sparks a , Robin P.F. Dullaart c a b c

Department of Pathology and Laboratory Medicine, University of Rochester School of Medicine and Dentistry, Rochester, NY, USA Department of Nephrology, University of Groningen and University Medical Center Groningen, 9700 RB, Groningen, The Netherlands Department of Endocrinology, University of Groningen and University Medical Center Groningen, 9700 RB, Groningen, The Netherlands

a r t i c l e

i n f o

Article history: Received 21 September 2010 Received in revised form 16 November 2010 Accepted 17 November 2010 Available online 26 November 2010 Keywords: Lipoprotein lipase D9N polymorphism Cholesteryl-ester transfer protein TaqIB polymorphisms C-reactive protein HDL cholesterol Cardiovascular disease

a b s t r a c t Objective: We sought to determine whether concurrently high levels of HDL cholesterol and CRP predict initial cardiovascular events in women, and to assess additional risk involving two genes encoding proteins involved in reverse cholesterol transport. Methods: A graphical approach identified high-risk subgroups in a population-based female cohort. Polymorphism-associated risk was assessed for CETP (TaqIB [rs708272]) and LPL (D9N [rs1801177]) using multivariable analysis adjusted for clinical parameters and biomarkers. Results: A high HDL-C/high CRP high-risk subgroup was identified. Multivariable modeling revealed D9N as predicting subgroup cardiovascular disease risk directly (minor allele-carriers versus major allele homozygotes: HR 5.16, 95% CI 1.43–18.54, p = 0.012) and through interaction with TaqIB (highest risk in minor allele carriers of both polymorphisms). Conclusions: In women with high HDL-C and high CRP levels, an LPL polymorphism associated with risk and interacted with a CETP polymorphism such that the highest risk occurred in subjects with presumably decreased activities of both proteins. © 2010 Elsevier Ireland Ltd. All rights reserved.

1. Introduction Concurrently high levels of HDL cholesterol (HDL-C) and Creactive protein (CRP) have recently been reported to be associated with increased cardiovascular disease (CVD) risk in men initially without CVD [1] and in postinfarction patients [2]. Furthermore, in postinfarction patients, a variant of the cholesteryl-ester transfer protein (CETP) gene associated with decreased CETP mass and activity has been reported to predict risk in patients with high HDLC and CRP levels [2]. This finding is consistent with impaired early phases of reverse cholesterol transport (RCT) as possibly playing a role in the observed risk-association by facilitating dysfunctional transformation of HDL. To investigate this notion further, the current study aimed: to replicate the finding of a similar high HDL-C/high CRP highrisk subgroup in women without CVD and to explore in men and women, initially without CVD, the effect on risk of decreased activ-

∗ Corresponding author at: Department of Pathology and Laboratory Medicine, University of Rochester Medical Center, Box 626, 601 Elmwood Avenue, Rochester, NY 14642, USA. Tel.: +1 585 275 4907; fax: +1 585 273 3003. E-mail address: James [email protected] (J.P. Corsetti). 0021-9150/$ – see front matter © 2010 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.atherosclerosis.2010.11.029

ity of CETP and lipoprotein lipase (LPL), proteins both involved in HDL particle remodeling. To do this, we used outcome event mapping, a graphical exploratory data analysis tool for identifying high-risk subgroups [1–3], and Cox multivariable proportional hazards regression in conjunction with activity probes consisting of genetic polymorphisms of CETP (TaqIB [rs708272]) [4] and LPL (D9N [rs1801177]) [5] in male and female cohorts of the Prevention of Renal and Vascular End-Stage Disease (PREVEND) population study [6]. 2. Methods 2.1. Study population The study group [1] comprised a cohort (N = 8592) of PREVEND [6], a prospective longitudinal study assessing albuminuria in predicting cardiovascular [7] and renal [8] disease. PREVEND was approved by the medical ethics committee of the University of Groningen, The Netherlands. PREVEND excluded insulin-using diabetics and pregnant women. Further exclusions for the present work included history of diabetes mellitus, renal disease, previous CVD, and incomplete laboratory results. Subjects with CRP levels ≥10 mg/L were excluded (to avoid confounding by inter-current ill-

374

J.P. Corsetti et al. / Atherosclerosis 214 (2011) 373–376

ness); there was no exclusion based on HDL-C levels. Sub-cohorts of men (N = 3405) and women (N = 3619) resulted (see Supplement – Methods). 2.2. Blood markers and genotyping

Table 1 Significant (p < 0.10) univariate Cox regression results for clinical parameters, biomarkers, and polymorphisms (TaqIB of CETP and D9N of LPL) in the high-risk subgroups of female subjects. Marker

Cox regression results Hazard ratioa

Overnight fasted blood samples collected after 15 minutes of rest were stored at −20 ◦ C prior to analysis [9] (see Supplement – Methods). Genotyping of TaqIB (CETP) was performed as described previously [9]. Genotyping of D9N (LPL), was performed using TaqMan-MGB probes, primer, SDS 2.0 genotype calling software, and ABI 7900HT apparatus (Applied Biosystems, Applera Nederland, Nieuwerkerk aan de IJssel, The Netherlands). 2.3. Outcomes Outcomes were combined incidence of CVD mortality and post-baseline hospitalisation [1]. Cardiovascular events included: acute MI, acute and subacute ischemic heart disease, coronary artery bypass grafting, and percutaneous transluminal coronary angioplasty. Median follow-up was 7.6 years (see Supplement – Methods). 2.4. Statistical analyses Outcome event mapping [1–3] was used to identify high-risk subgroups (see Supplement – Methods). Statistica 9.0 (StatSoft, Inc., Tulsa, OK) was used for statistical and graphical analyses. Statistical tests (p < 0.05) included Pearson correlation, Mann–Whitney U, Kruskal–Wallis with Bonferroni correction, chisquare, Kaplan–Meier with log-rank statistic, and multivariable Cox proportional hazards regression (biomarkers treated as continuous variables; polymorphisms treated as minor allele carriers versus major allele homozygotes). 3. Experimental results 3.1. Study population Qualitative identification in female subjects of two highrisk subgroups was reported previously [1]. Subsequent analysis (Supplement – Experimental results) revealed 879 subjects in a low HDL-C/high CRP high-risk subgroup (henceforth HR1); 508 subjects in a high HDL-C/high CRP subgroup (henceforth HR2); and 2232 subjects in a lower risk background subgroup (henceforth BG). Subjects in HR1 versus HR2 (Supplement – Table 1) were older, more hypertensive, have more MS and less alcohol use. Additionally, they demonstrated higher levels of cholesterol, triglycerides, nonHDL-C, glucose, apoB and apoB/apoA1 ratio, and lower HDL-C, apoA1, and HDL-C/apoA1 ratio, a rough measure of HDL particle size. Correlation between HDL-C and triglyceride levels was used to indirectly assess HDL particle remodeling. Correlation coefficients (all significant) were: BG, −0.24; HR1, −0.29; and HR2, −0.15. Results of pairwise comparisons were: BG/HR1 (p = 0.16), BG/HR2 (p = 0.057), and HR1/HR2 (p = 0.0074). Thus in HR2, the magnitude of inverse correlation between HDL-C and triglycerides was significantly less than in HR1. 3.2. Genotype distributions The TaqIB polymorphism (CETP) in females was in Hardy–Weinberg equilibrium; the D9N polymorphism (LPL) was not (for details and likely explanation, see Supplement – Experiment results). Significant differences (p < 0.01) in blood markers (HDL-C, apoA1, triglyceride, apoB, and total cholesterol) according to the polymorphisms (minor allele carriers versus

HR1 (low HDL-C/high CRP) 2.82 Ageb 1.35 BMIb b 7.22 Hypertension HDL-C 0.68 HDL-C/apoA1 0.37 Triglycerides 1.39 apoB 1.33 1.71 Cholesterolb Glucose 1.29 TaqIb (CETP) 0.99 HR2 (high HDL-C/high CRP) 2.69 Ageb 1.59 BMIb 4.85 Hypertensionb CRP 1.51 Triglycerides 1.51 apoB 1.69 apoB/apoA1 1.88 Cholesterol 1.67 Glucose 1.45 TaqIb (CETP) 2.62 D9N (LPL) 6.04 a b

95% CI

p

1.84–4.33 1.02–1.80 2.53–20.62 0.50–0.93 0.21–0.68 1.12–1.72 1.04–1.71 1.26–2.31 1.09–1.52 0.47–2.07

<0.001 0.039 <0.001 0.014 0.001 0.003 0.024 <0.001 0.003 0.98

1.49–4.86 1.07–2.37 1.37–17.18 0.95–2.41 1.10–2.05 1.25–2.29 1.30–2.73 1.30–2.13 1.14–1.83 0.59–11.69 1.69–21.67

0.0011 0.022 0.014 0.082 0.01 <0.001 <0.001 <0.001 0.002 0.21 0.0057

Hazard ratio for continuous variables expressed per SD unit. Parameters adjusted for in multivariable models.

major allele homozygotes) were for TaqIB: in HR1 and HR2, none; in BG, higher HDL-C and apoA1. For D9N, there were none. 3.3. Univariate biomarker and polymorphism risk within subgroups Table 1 gives univariate significant Cox regression results for the high-risk subgroups. For TaqIB, the hazard ratio in both subgroups was non-significant; although for HR2, there was a tendency towards significance. For D9N, in HR2 the hazard ratio was significant; in HR1, Cox analysis was not applicable as minor allele carriers had no outcomes. 3.4. Polymorphism multivariable risk within subgroups Multivariable modeling adjusted for significant clinical parameters and biomarkers (HR1 – age, BMI, hypertension, and cholesterol; HR2 – age, BMI, and hypertension; Supplement – Experimental results) was performed as a function of the polymorphisms. In HR1, TaqIB was non-significant; D9N analysis was excluded as noted above. In HR2, D9N remained significant with greater risk for minor allele-carriers (hazard ratio 5.16, 95% CI 1.43–18.54, p = 0.012); TaqIB tended towards significance (hazard ratio 2.15, 95% CI 0.48–9.73, p = 0.32). To assess potential effects on HR2 of enrichment in PREVEND with albuminuric subjects, UAE and serum creatinine were forced into multivariable models. The D9N polymorphism remained significant. Additionally, a sensitivity analysis was performed on a PREVEND female sub-cohort not enriched with albuminuric subjects. High-risk subgroups similar to HR1 and HR2 resulted. In the corresponding HR2 subgroup, D9N tended towards significance (hazard ratio 5.02, 95% CI 0.59–42.70, p = 0.14). 3.5. Polymorphism interaction analysis within subgroups To assess risk-associated interaction between the polymorphisms, Cox models were run including an interaction term.

J.P. Corsetti et al. / Atherosclerosis 214 (2011) 373–376

Proportion without Events

1.00 0.95 0.90 0.85 0.80 0.75 0

500

1000

1500

2000

2500

3000

Time (Days) Fig. 1. Kaplan–Meier plots for female subjects in the high HDL-C/high CRP high-risk subgroup (HR2) showing outcome events over time. Significant interaction between the D9N polymorphism of LPL and the TaqIB polymorphism of CETP (p = 0.0002) is demonstrated in that carriers of the minor alleles of both polymorphisms (light line) show greater outcome events than remaining subjects (heavy line).

Interaction was significant in HR2 but not in HR1. Fig. 1 illustrates the interaction in HR2 through Kaplan–Meier curves for minor allele carriers of both polymorphisms versus remaining subjects (p = 0.0002). 3.6. Polymorphism analyses in male PREVEND subjects Characterization of male PREVEND subjects and identification of high-risk subgroups similar to females have been reported previously [1]. Similar analyses (Supplement – Experimental results) as above in males revealed non-significance for both polymorphisms in the corresponding high-risk subgroups. 4. Discussion Results of the current study demonstrate in women initially without CVD a high-risk subgroup at high levels of HDL-C and CRP, similar to findings in men [1] and postinfarction patients [2]. Furthermore, investigations using multivariable risk-modeling within the subgroup with genetic markers to probe effects of decreased activities of CETP and LPL demonstrated D9N as a significant predictor of risk, and a trend towards significance for TaqIB. Moreover, interaction between D9N and TaqIB was significant with highest risk occurring in minor allele carriers of both polymorphisms, in each case, the allele reportedly associated with decreased protein activity [4,5]. An additional finding in women, as in men [1], was identification of a second high-risk subgroup at low HDL-C and high CRP levels enriched with metabolic syndrome subjects. This subgroup demonstrated a higher level of inverse correlation of HDL-C with triglycerides than the high HDL-C subgroup, a finding potentially indicative of less efficient HDL particle remodeling in the high HDL-C subgroup. Corresponding findings were not seen in men. This may be related to the overall higher event rate in men (7.74%) versus women (2.95%) that when analyzed specifically in the context of event rates for D9N revealed similar values for minor allele carriers (men – 12.50%; women – 14.29%), but dissimilar values for major allele homozygotes (men – 7.62%; women – 2.43%). Thus, it appears that lack of D9N associated risk in men stems not from differences in the higher outcome rates of minor allele carriers but rather from the relatively higher risk in major allele homozygotes of men versus women. LPL [10,11] is involved with triglyceride-rich lipoprotein metabolism and lipoprotein remodeling including HDL [12,13].

375

Consistent with our study, the reduced-activity minor allele of D9N is generally associated with increased CVD risk [5]. D9Nassociated risk, especially in conjunction with TaqIB, suggests a role for altered HDL remodeling in pro-atherogenic transformation of HDL, a process likely potentiated by the inflammatory environment defining the subgroup. In this regard, high CRP may play a direct role in this transformative process. Recent reports [14] detail direct effects of CRP potentially relating to HDL transformation including myeloperoxidase release from human neutrophils, superoxide anion production via up-regulation of NAD(P)H oxidase in human foam cells; and superoxide release from rat macrophages. These effects could potentially link CRP to HDL dysfunction given recent reports of loss of HDL anti-atherogenic functionality through myeloperoxidase-mediated apoA1 oxidation [15]. Moreover like LDL [14], HDL outer layer phospholipids could be oxidized by CRPmediated elaboration of oxidative species generating CRP binding sites on HDL particles that, again like oxidized LDL [14], could result in subsequent macrophage uptake of CRP-bound HDL. Limitations in the current study included study population over-sampling with high-UAE subjects; however, findings regarding D9N-associated risk appeared little affected when UAE and creatinine were forced into multivariable models and when sensitivity analysis of a sub-cohort without albuminuric enrichment also gave similar results. Further study limitations related to the lack of evidence for the causal role of inflammation and oxidative stress in HDL dysfunctional transformation and lack of evidence for decreased activity of CETP and LPL in minor allele carriers. Future studies should be designed to address these issues. The current study demonstrated in a population of women initially without CVD a high-risk subgroup defined by concurrently high levels of HDL-C and CRP. This subgroup was similar to highrisk subgroups previously identified in men initially without CVD [1] and in postinfarction patients [2]. Among high-risk women, carriage of the less active allele of D9N (LPL) was found to associate with risk and with simultaneous carriage of the less active allele of TaqIB (CETP) to intensify risk. We conclude that compromised early stages of RCT involving alterations in HDL particle remodeling likely play a role in establishing risk in women with high levels of HDL-C and CRP. Sources of funding Grant 2001.005 (Netherlands Heart Foundation). Disclosures None. Acknowledgements Contributions of R.T. Gansevoort, R.P.F. Dullaart, and G.J. Navis were made on behalf of the PREVEND study group. We are indebted to all PREVEND collaborators. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.atherosclerosis.2010.11.029. References [1] Corsetti JP, Gansevoort RT, Sparks CE, Dullaart RPF. HDL protection against primary cardiac risk is lost with inflammation. Eur J Clin Invest 2010;40:483–9. [2] Corsetti JP, Ryan D, Rainwater DL, Moss AJ, Zareba W, Sparks CE. Decreased CETP activity associates with recurrent risk in postinfarction patients with high HDL cholesterol and high CRP. Atheroscler Thromb Vasc Biol 2010;30:1657–64.

376

J.P. Corsetti et al. / Atherosclerosis 214 (2011) 373–376

[3] Corsetti JP, Ryan D, Moss AJ, Zareba W, Sparks CE. NAD(P)H oxidase polymorphism (C242T) and high HDL-C associate with recurrent coronary events in post infarction patients. Atherosclerosis 2007;196:461–8. [4] Dullaart RPF, Sluiter W. Common variation in the CETP gene and the implications for cardiovascular disease and its treatment: an updated analysis. Pharmacogenomics 2008;9:747–63. [5] Sagoo GS, Tatt I, Salanti G, et al. Seven lipoprotein lipase gene polymorphisms, lipid fractions, and coronary disease: a HuGE association review and metaanalysis. Am J Epidemiol 2008;168:1233–46. [6] Pinto-Sietsma SJ, Janssen WMT, Hillege HL, Navis G, de Jong PE. Urinary albumin excretion is associated with renal functional abnormalities in a non-diabetic population. J Am Soc Nephrol 2000;11:1882–8. [7] Lambers Heerspink HJ, Brantsma AH, de Zeeuw D, Bakker SJ, de Jong PE, Gansevoort RT. Albuminuria assessed from first-morning-void urine samples versus 24 hour urine collections as a predictor of cardiovascular morbidity and mortality. Am J Epidemiol 2008;168:897–905. [8] Verhave JD, Gansevoort RT, Hillege HL, Bakker SJ, de Zeeuw D, de Jong PE. An elevated urinary albumin excretion predicts de novo development of renal function impairment in the general population. Kidney Int 2004;66: S18–21.

[9] Borggreve SE, Hillege HL, Wolffenbuttel BHR, et al. An increased coronary risk is paradoxically associated with common cholesteryl ester transfer protein gene variations that relate to higher high-density lipoprotein cholesterol: a population-based study. J Clin Endocrinol Metab 2006;91:3382–8. [10] Stein Y, Stein O. Lipoprotein lipase and atherosclerosis. Atherosclerosis 2003;170:1–9. [11] Wang H, Eckel RH. Lipoprotein lipase: from gene to obesity. Am J Physiol Endocrinol Metab 2009;297:E271–88. [12] Murdoch SJ, Breckenridge WC. Influence of lipoprotein lipase and hepatic lipase on the transformation of VLDL and HDL during lipolysis of VLDL. Atherosclerosis 1995;118:193–212. [13] Borggreve SE, De Vries R, Dullaart RPF. Alterations in high-density lipoprotein metabolism and reverse cholesterol transport in insulin resistance and type 2 diabetes mellitus: role of lipolytic enzymes, lecithin:cholesterol acyltransferase and lipid transfer proteins. Eur J Clin Invest 2003;33:1051–69. [14] Devaraj S, Singh U, Jialal I. The evolving role of C-reactive protein in atherothrombosis. Clin Chem 2009;55:229–38. [15] Shao B, Oda MN, Oram JF, Heinecke JW. Myeloperoxidase: an oxidative pathway for generating dysfunctional high-density lipoprotein. Chem Res Toxicol 2010;23:447–54.