|
Advertisement | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Journal of Lipid Research, Vol. 47, 515-520, March 2006 Paraoxonase-1 and serum concentrations of HDL-cholesterol and apoA-I
Clinical Diabetes Unit, Division of Endocrinology, Diabetes, and Nutrition, Medical Faculty, University Hospital, 1211 Geneva 14, Switzerland Published, JLR Papers in Press, December 3, 2005.
1 Present address of M-C. Blatter Garin: Swiss Institute of Bioinformatics, University Medical Center, Medical Faculty, University of Geneva, Geneva, Switzerland.
2 To whom correspondence should be addressed. e-mail: richard.james{at}hcuge.ch
Paraoxonase-1 (PON1) and HDL are tightly associated in plasma, and this is generally assumed to reflect the need for the enzyme to associate with a hydrophobic complex. The association has been examined in coronary cases and age-matched controls. Highly significant (P < 0.0001), positive associations were observed between PON1 activities and concentrations and HDL-cholesterol and apolipoprotein A-I (apoA-I) concentrations in cases and controls. Corrected slopes were significantly different in cases (cases vs. controls: arylesterase, r = 0.19 vs. 0.38, P < 0.02 for apoA-I and r = 0.15 vs. 0.34, P < 0.02 for HDL-cholesterol) such that if PON1 should influence serum HDL, it would be less effective in coronary cases. When examined as a function of the PON1 gene promoter polymorphism C107T, highly significant differences (P < 0.001) in HDL-cholesterol and apoA-I were observed between genotypes for controls, with high expresser alleles having the highest HDL concentrations. This relationship was lost in cases with coronary disease. The coding region polymorphisms Q192R and L55M of the PON1 gene showed no association with HDL. The promoter polymorphism was an independent determinant of HDL concentrations in multivariate analyses. These data are consistent with an impact of PON1 on plasma concentrations of HDL, with detrimental modifications to the relationship in coronary cases.
Supplementary key words atherosclerosis oxidative stress low density lipoprotein high density lipoprotein apolipoprotein A-I
A number of studies have demonstrated the close, physiological association between paraoxonase-1 (PON1) and HDL in plasma. HDL provides a vector that facilitates the secretion of the enzyme by liver (1), essentially by offering a hydrophobic harbor for the retained signal peptide of PON1 and coincidentally stabilizing the enzyme. The lipoprotein also furnishes a hydrophobic environment that could be important for PON1 function (2). In return, the enzyme prevents or limits the oxidation of HDL (3), although the extent to which this provides physiological benefit to HDL in vivo is less well established. The latter is of particular relevance given that HDLs appear to be the primary transporters of oxidized lipids in plasma (4) and that oxidation can compromise HDL function (5). Moreover, a recent study suggested that PON1 could influence HDL concentrations (6). That study, however, was performed under the particular conditions of patients with familial hypercholesterolemia accompanied by a high prevalence of atherosclerosis in the population. Thus, at present, PON1 is suggested to be a primary determinant of the antioxidant capacity of HDL, and the antioxidant function is evoked as the principal source of the beneficial impact of the enzyme on atherosclerotic disease (7). Several independent studies have demonstrated that serum PON1 activity and concentration are decreased in patients with cardiovascular disease and other pathologies that increase the risk of vascular disorders (810). The cause has yet to be established, but reduced PON1 activity may contribute to the disease process. In this context, one prospective study has suggested that decreased serum PON1 status precedes the onset of vascular problems (11). A potential extrapolation of these studies is that PON1 could have a beneficial impact on vascular disease by influencing serum HDL concentrations, the latter being a well-established, powerful cardiovascular disease risk determinant. To gain further insight into this possibility, the present study examined i) whether PON1 was an independent determinant of serum HDL concentrations in subjects without evidence of coronary disease (controls), and ii) whether the relationship between PON1 and HDL concentrations was modified in cases with confirmed cardiovascular disease.
Study population This study was based on a population with confirmed (angiography) coronary artery disease (CAD; n = 710, 79.5% men) and controls (n = 199, 50.8% men) recruited within the 4665 year range to minimize age differences with cases. Briefly, cases were recruited from patients referred consecutively to the Cardiology Division of the University Hospital, Geneva, for coronary angiography. All participants underwent angiography, and the presence of coronary disease was defined as stenosis estimated at >20% in at least one major epicardial artery. Participants with no detectable coronary artery anomaly were classified as coronary disease-negative (controls). The majority of controls were referred for examination of heart valve anomalies, with a minority being referred for investigation of chest pains. The latter were diagnosed as being of noncardiac origin. Each participant completed questionnaires on personal and family medical history with an interviewer and gave a fasting blood sample. All participants gave written, informed consent, and the ethics committee of the University Hospital approved the study. The anthropomorphic and clinical characteristics of these populations have been described in detail in previous publications (9, 12) and are briefly recapitulated in Table 1.
Analyses Plasma lipids, lipoproteins, and apolipoproteins were quantified as described (13, 14). PON1 enzyme activities were analyzed with phenylacetate [arylesterase activity (ARE)] and paraoxon [basal paraoxonase activity (Ponb)] as substrates (15, 16). PON1 protein concentrations and genotypes were analyzed as detailed previously (15, 16). Correlations between activities and PON mass for controls and cases were as follows: ARE, 0.83 and 0.79; Ponb, 0.72 and 0.69, respectively. Genotype frequencies did not depart from Harvey-Weinberg equilibrium (17).
Statistical analyses
There were significant correlations between PON1 mass and activities and serum concentrations of apoA-I and HDL-cholesterol for subjects free of CAD (Table 2). PON1 gene expression is greatly influenced by the C107T promoter polymorphism. It accounts for up to 25% of intraindividual variations in serum concentrations and activities of the enzyme (16). This was confirmed in the present study, where in controls the promoter genotype accounted for 23.6% and 7.7% of variations in ARE and Ponb (20.7% and 8.8%, respectively, for the cases). Corresponding variations in these activities associated with other genotypes were as follows: for Q192R, 0.1% and 68.2% (controls) and 0.5% and 73.7% (cases); for L55M, 4.9% and 24.4% (controls) and 9.8% and 22.7% (cases), respectively.
Highly significant differences in serum HDL-cholesterol and apoA-I concentrations were observed as a function of the polymorphism, with carriers of the homozygous, high-expresser allele (107CC) having significantly increased concentrations of both variables (Table 3). A similar analysis based on the coding region polymorphisms of PON1 (Q192R and L55M) showed no association of the genotypes with serum concentrations of apoA-I or HDL-cholesterol (Table 3).
When cases with angiographically proven coronary disease were analyzed, univariate analyses showed highly significant, positive correlations between PON1 arylesterase and paraoxonase activities (Table 2) and concentration (data not shown) and apoA-I and HDL-cholesterol (P <0.0001 for all correlation coefficients). However, the slopes, corrected for covariates (Table 1), were significantly different between controls and cases for apoA-I (P < 0.02) and HDL-cholesterol (P < 0.02) (Table 2). This is illustrated in Fig. 1, which combines correlation plots for controls and cases. Similarly, slopes of HDL with paraoxonase activities (corrected for the position 192 polymorphism effects) were significantly different in cases (apoA-I, P < 0.01; HDL-cholesterol, P < 0.02) (Table 2). Analyses of HDL and apoA-I as a function of the promoter polymorphism C107T revealed the 107CC homozygotes to be no longer associated with significantly higher concentrations of these parameters (Table 3). CAD cases showed no significant differences in HDL-cholesterol or apoA-I as a function of the coding region polymorphisms (Table 3).
To ascertain whether PON1 parameters were determinants of HDL, multivariate analysis was performed using the control populations, with apoA-I or HDL-cholesterol as the dependent phenotype. The results are illustrated in Table 4 for HDL-cholesterol. Variables well known to influence serum HDL-cholesterol concentrations were significant, independent determinants, including gender, triglycerides, BMI, and smoking status. The promoter polymorphism C107T was also a significant, independent determinant of HDL-cholesterol concentrations (Table 4). This model accounted for 36% of intraindividual variations in serum concentrations of HDL-cholesterol. When PON1 activity (arylesterase) was introduced into the model in place of the promoter polymorphism (model 2), it was also an independent determinant of HDL-cholesterol. This model accounted for 37% of the variation of HDL-cholesterol. Similar results were obtained when apoA-I replaced HDL-cholesterol in the multivariate analysis. Both the promoter polymorphism, C107T (P < 0.001), and PON1 activity (P < 0.0001) were determinants of apoA-I concentrations.
After submission of this study for review, a publication by Rozek et al. (18) prompted us to extend our analyses to LDL. We observed highly significant, positive correlations between LDL-cholesterol and ARE (r = 0.17, P < 0.0001) and PON1 concentration (r = 0.25, P <0.0001) for cases. Correlation coefficients were lower and nonsignificant for the control group (LDL-cholesterol vs. arylesterase,r = 0.12, P = 0.08; LDL-cholesterol vs. PON1 concentration, r = 0.12, P = 0.10). Elimination of patients on lipid-lowering therapy or diabetic patients did not modify these conclusions. Table 5 shows LDL-cholesterol concentrations as a function of PON1 gene polymorphisms. The C107T promoter polymorphism in cases showed no significant difference in LDL-cholesterol concentrations between genotypes, with CC homozygotes having the highest concentrations. There were no significant differences in LDL-cholesterol concentrations between promoter genotypes in the control group, in which CC homozygotes tended to be associated with lower LDL-cholesterol concentrations. Analysis of the Q192R genotypes revealed a significantly higher level of LDL-cholesterol in the RR subgroup of cases (Table 5). A similar, albeit nonsignificant, trend was observed in the control subjects.
In multivariate analyses, PON1 activity (arylesterase and paraoxonase) and concentration were independent determinants of serum LDL-cholesterol concentration in cases (data not shown). The Q192R genotypes were also independent determinants of LDL-cholesterol concentrations, whereas the PON1 promoter polymorphism C107T was not.
The three salient observations of the present analysis are that PON1 promoter genotype is associated with plasma HDL concentrations, that promoter genotypes (and PON1 activity/concentration) are independently associated with serum HDL-cholesterol and apoA-I concentrations, and, finally, that the relationship between PON1 and HDL is modified in coronary cases. An intriguing interpretation of these data is that PON1 is a determinant of serum levels of apoA-I and HDL-cholesterol. This effect would be compromised in patients with coronary disease, such that the enzyme would be less able to promote higher concentrations of this lipoprotein species. The positive correlation between serum PON1 and HDL has been known for a number of years (19). The general assumption is that PON1 is the essential beneficiary of the association, with HDL determining serum levels of the enzyme. Our recent studies have provided a mechanistic justification for this dependency by showing the importance of HDL for PON1 release from cells and maintenance of enzyme activity (1). Moreover, the role of HDL in serum PON1 transport has been elegantly demonstrated by Sorenson et al. (20). This one-sided interpretation, however, is difficult to reconcile with the observation that PON1 promoter genotype is associated with differences in serum concentrations of HDL-cholesterol and apoA-I. There is no known linkage between the PON1 gene and other genes influencing HDL, which could explain such an association, although linkage disequilibria exist within the PON1 gene [between the promoter polymorphism, C107T, and the L55M coding region polymorphism, and between the latter and the Q192R polymorphic site (16)]. There are reports, however, of genetic linkage between variations in serum HDL concentrations and the chromosomal region 7q21.3, which contains the PON gene cluster (21, 22). Moreover, the promoter has a strong influence on gene expression, and the high-expresser PON1 allele was associated with increased HDL and apoA-I concentrations. The genotype-HDL relationship may thus reflect an impact of PON1 activity on HDL and is consistent with the independent associations of genotype and activity with HDL in multivariate analyses. A recent study has also reported a significant association of the promoter genotype with serum HDL in patients with familial hypercholesterolemia (6), suggesting a beneficial influence on the lipoprotein. Our studies indicate that these observations are not limited to particular conditions that may prevail in familial hypercholesterolemia but can extend to an older population without coronary disease. A number of arguments can be marshaled to explain a beneficial influence of PON1 on HDL. Lipoprotein oxidation, which PON1 can limit, directly accelerates HDL catabolism (23). In addition, oxidation of apoA-I impairs its ability to remove cholesterol from cells (5, 24), which is part of the process of HDL maturation, leading to an increase in size. An inability to increase HDL size causes more rapid catabolism of the lipoprotein (25). Of relevance in this context is a recent study suggesting that PON1 can increase the capacity of HDL to efflux cholesterol from macrophages (26). The association of the promoter genotype with apoA-I and HDL is notable for the higher concentrations of both in PON1 107CC homozygotes. One explanation could be that it reflects the need for higher levels of activity for an impact of PON1 on serum HDL concentrations, presumably to counter the consequences of HDL oxidation. In this context, oxidative stress/lipoprotein oxidation is known to be more prevalent in patients with cardiovascular disease. The second intriguing observation provided by these analyses is that the relationship between PON1 and HDL is altered in coronary cases. The association is significantly weaker in cases, such that PON1 would be associated with a less efficient HDL-increasing capacity. The lower PON1 activity reported in cardiovascular patients (810) would be a contributory factor to any reduced impact of the enzyme on HDL. However, the altered association between PON1 and HDL would appear to be of greater relevance. This suggests that factors other than absolute PON1 activity modulate the relationship between the enzyme and HDL. This may include an inability to fully exert the available PON1 antioxidant capacity (2). Our results are in broad agreement with the principal conclusion of a recent publication by Rozek et al. (18) that the association of PON1 with lipoprotein subfractions differs between controls and coronary cases. PON1 activity correlated more strongly with HDL in controls but correlated more strongly with LDL-cholesterol in coronary cases. However, the present study differs from the former study in that Rozek et al. (18) found no correlation between PON1 genotypes and serum lipid concentrations in either controls or carotid disease cases, in contrast with our observations. Patient selection and/or patient numbers could contribute to the differences between the two studies. An extrapolation of these observations is that one aspect of the association of PON1 with vascular disease may involve an inability to promote higher concentrations of HDL. The data are consistent with the proposal that PON1 and HDL derive mutual benefit from their association in plasma. Our results, together with recent studies (6, 26), suggest that the relationship between HDL and PON1 status merits more detailed analysis.
The authors are particularly grateful to Professor A. Morabia (Clinical Epidemiological Division, University Hospital, Geneva) for advice on statistical analyses. This study was supported by grants from the Swiss National Research Foundation (Grants 32-043088.95 and 31-105310.01). Manuscript received July 5, 2005 and in revised form August 23, 2005 and in re-revised form November 22, 2005.
This article has been cited by other articles:
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Advertisement | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||