Lipoprotein particles of differing sizes play distinct roles

Transcription

1 Genetic Control of Coordinated Changes in HDL and LDL Size Phenotypes David L. Rainwater, Lisa J. Martin, Anthony G. Comuzzie Abstract We investigated the correlation of high density lipoprotein (HDL) and low density lipoprotein (LDL) particle size distributions in samples from 1100 participants in the San Antonio Family Heart Study. By use of analyses of individual correlations of each HDL fraction with each LDL fraction, we devised new metrics for particle size phenotype, termed HDL and LDL, to optimally reflect the size correlations. Confirming previous studies, we found that the 2 size phenotype variables were positively correlated (r 0.51). Quantitative genetic analysis indicated that nearly half (44%) of the variance in HDL and in LDL was explained by the additive effects of genes. Bivariate genetic analyses indicated that a positive genetic correlation ( G 0.56) exists between them and suggested that the pleiotropic effects of a gene or group of genes account for 31% [ie, 2 G (0.56) ] of the genetic variance in the 2 traits. Triglyceride (TG) levels were negatively related to HDL and LDL, with phenotypic correlations of 0.48 and 0.58, respectively, and genetic correlations of 0.45 and 0.76, respectively, suggesting that genes exert significant pleiotropic effects on the covariation of TGs with each of the size variables. Finally, we evaluated a bivariate model for HDL and LDL in which TG level was included as a covariate. This analysis indicated that a small but significant genetic correlation remains for HDL and LDL, even after accounting for the effects of TGs. Thus, our study demonstrates that the phenotypic correlation of HDL and LDL sizes results in part from the pleiotropic actions of a set of genes, some of which also influence TG levels and some of which do not. (Arterioscler Thromb Vasc Biol. 2001;21: ) Key Words: LDL size distributions HDL size distributions genetics triglycerides Mexican Americans Lipoprotein particles of differing sizes play distinct roles in lipoprotein metabolism and the process of atherosclerosis. The predominance of small dense LDLs was identified as a risk factor for cardiovascular disease (CVD), 1 and prospective studies now have confirmed this initial observation. 2 4 Several mechanisms have been proposed to explain this association 5 : (1) small dense LDLs are more readily converted to highly atherogenic oxidized LDLs; (2) they more readily bind to surface components in the vessel wall; and (3) their size permits a more efficient infiltration into the vessel wall. There is growing evidence that HDL subpopulations differ in their abilities to protect against CVD. 6,7 However, whether current methods of measuring HDL subpopulations provide a substantial increment in CVD-predictive power is more controversial. For example, 2 studies reported significant associations of small dense HDLs (ie, HDL 3, separated by electrophoresis and density fractionation) and CVD, but they differed as to whether HDL 3 improved the prediction. 8,9 Using nuclear magnetic resonance methodology, Freedman et al 10 found that lipoprotein subpopulation measures, for HDL and LDL, significantly increased the proportion of variation in the coronary artery occlusion score that could be predicted. A number of studies have reported a positive correlation between LDL and HDL size phenotypes Although triglyceride (TG) concentrations are associated with LDL and HDL size variation, recent results suggest that at least a portion of the correlation between LDL and HDL sizes is independent of TG levels. 15 These observations suggest the existence of a metabolic mechanism(s) that governs the coordinated covariation in the relative sizes of several classes of lipoproteins. The purpose of the present study was to investigate the nature of the relationship between LDL and HDL size phenotypes. To facilitate this investigation, we have developed new metrics for lipoprotein size, HDL and LDL, which reflect the correlation of HDL and LDL size fractions, respectively. HDL and LDL are size distribution variables that are based on subtracting the absorbance in smaller particles from the absorbance in larger particles. With these new size variables, we have tested the hypothesis that there are lipoprotein size genes responsible for the coordinated covariation between LDL and HDL. Methods Subjects and Samples The San Antonio Family Heart Study investigates risk factors for CVD and diabetes in families of Mexican Americans living in and Received August 8, 2001; revision accepted August 29, From the Department of Genetics, Southwest Foundation for Biomedical Research, San Antonio, Tex. Correspondence to David L. Rainwater, PhD, Department of Genetics, Southwest Foundation for Biomedical Research, PO Box , San Antonio, TX david@darwin.sfbr.org 2001 American Heart Association, Inc. Arterioscler Thromb Vasc Biol. is available at

2 1830 Arterioscler Thromb Vasc Biol. November 2001 around San Antonio, Tex. 16 Probands were selected at random, and all first-, second-, and third-degree relatives were invited to participate in the present study. At a clinic visit, participants provided information on lifestyle variables, including current medications, and they were subjected to a physical examination; subjects who were currently taking lipid-lowering medications ( 2%) were excluded from the study. An oral glucose tolerance test was administered, and the presence of diabetes (yes/no) was determined by World Health Organization (WHO) criteria. 17 In addition, a blood sample was obtained for lipoprotein analyses; plasma was isolated by low-speed centrifugation and stored at 80 C in single-use aliquots. 18 Clinic procedures were reviewed by the Institutional Review Board for the University of Texas Health Science Center at San Antonio, and all subjects gave written informed consent. Measurement of Lipoprotein Phenotypes Cholesterol and TG concentrations were measured enzymatically by using kits from Boehringer-Mannheim Diagnostics and Stanbio, respectively. HDL cholesterol (HDL-C) concentrations were determined 19 in the supernatant after precipitating apob-containing lipoproteins with dextran sulfate Mg 2. Interassay coefficients of variation for control products in these assays were 2.1% for total plasma cholesterol, 4.6% for HDL-C, and 3.7% for TGs. LDL cholesterol (in millimoles per liter) was estimated as follows: total plasma cholesterol [HDL-C (0.42TG)], which has been validated for TG values 8 mmol/l. 20,21 Composite gradient gels, which enable the evaluation of LDL and HDL phenotypes in the same lane, were made in the laboratory. 22 Samples were subjected to electrophoresis at 3000 V h, and cholesterol was stained with Sudan black B. 23 Lipoprotein size distributions of cholesterol stain were measured with an LKB- Ultroscan XL laser densitometer. Calibrators, run on each gel, included albumin (diameter 7.1 nm), lactate dehydrogenase (8.16 nm), ferritin (12.2 nm), and thyroglobulin (17.0 nm) in the Pharmacia high molecular weight standards kit, 2 LDL bands in a lyophilized plasma standard (26.6 and 27.5 nm), 24 and carboxylated polystyrene microspheres (38 nm). The first 4 standards were used as calibrators for HDLs, and the last 4 standards were used as calibrators for LDLs. To assess gel-to-gel variation, we determined LDL and HDL median diameters (half the absorbance is on larger particles and half is on smaller particles 22 ) for a lyophilized plasma sample run on each gel; coefficients of variation were as follows: 2.59% for HDL (mean diameter was 9.69 nm, n 629 runs) and 0.54% for LDL (mean diameter was nm, n 625 runs). To enable comparisons across gels, absorbance profiles were converted from units of migration distance to units of particle diameter (at 0.1-nm increments), as suggested previously, 25 and the data were expressed in terms of fractional absorbance 1000 for HDLs (7.2 to 13 nm) and for LDLs (21 to 29 nm). Lipoprotein size distributions were estimated as absorbance in large size particles minus absorbance in small size particles (see text above). Thus, we calculated HDL as [(fractional absorbance, 9.6 to 10.7 nm) (fractional absorbance, 7.7 to 8.2 nm)] 1000 and LDL as [(fractional absorbance, 27.2 to 28.4 nm) (fractional absorbance, 24.4 to 26.0 nm)] Repeatability for split duplicates was estimated to be 94.2% for HDL and 88.9% for LDL (n 77 samples duplicated). Values subjected to analyses represented the average of all accepted runs of a sample (average 2.1 runs per sample). Quantitative Genetic Analyses We used well-established statistical genetic methods 26,27 to partition the phenotypic correlations among HDL, LDL, and TG levels into correlations due to shared genetic effects (ie, pleiotropy) and those due to random environmental effects (designated as G and E, respectively); total phenotypic correlations ( P ) were calculated as follows: P [ h 12 h 22 G ] [ (1 h 12 ) (1 h 22 ) E ], where h 12 and h 22 are heritabilities of traits 1 and 2, respectively. We have described implementation of these methods in detail previously. 28,29 Results Development of 2 Lipoprotein Size Variables, HDL and LDL Initially, a subset of samples (n 482) was evaluated for correlations between HDLs and LDLs. Each absorbance Figure 1. Average absorbance profile for a subset of 482 samples. Each absorbance profile was expressed as fractional absorbance 1000 at each 0.1-nm increment of particle size for HDLs and for LDLs before averaging. profile was converted into units of particle size at increments of 0.1 nm, and Figure 1 shows the average absorbance profile for these subjects. A correlation coefficient was calculated for the relationship of each HDL fraction (n 58) with each LDL fraction (n 80), and Figure 2 shows a contour plot for these 4640 correlations. The data indicated the existence of 2 size classes of LDL, which are strongly and positively correlated with 2 size classes of HDL: large LDLs (27.2 to 28.4 nm) with large HDLs (9.6 to 10.7 nm) and small LDLs (24.4 to 26.0 nm) with small HDLs (7.7 to 8.2 nm). As might be expected, large size lipoproteins were inversely correlated with small size lipoproteins. These results suggest that 2 relatively narrow size intervals in each lipoprotein class are largely responsible for the significant correlation of LDL and HDL size phenotypes. The large and small size intervals tended to be on the leading and trailing edges of the main absorbance peak of the average LDL and HDL absorbance profiles (Figure 1). Thus, as HDL (or LDL) patterns trend toward larger particles, the large size interval increases relative to the small size interval (and vice versa). Therefore, to generate a single variable representing coordinated size variation in each class of lipoproteins (ie, HDL and LDL), we subtracted absorbance in smaller particles from that in larger particles. When these variables were determined for all samples in the study, we found a strong positive phenotypic correlation between HDL and LDL Figure 2. Two-dimensional contour plot of the matrix of correlations for each LDL fraction with each HDL fraction ( correlations calculated in all) for a subset of 482 absorbance profiles. The correlations ranged from 0.48 (at 27.8, 8.0) to 0.46 (at 27.6, 9.8); contour lines are drawn at r 0.05 increments.

3 Rainwater et al Genetic Control of HDL and LDL Size 1831 Figure 3. Scatterplot of HDL vs LDL values for samples in the study. Pleiotropic Effects on HDL and LDL Bivariate genetic analyses were used to evaluate the phenotypic correlations between HDL and LDL. The phenotypic correlation between the 2 traits was estimated to be 0.51 in the genetic analyses, which is similar to the simple correlation given in Figure 3, which did not account for the relatedness among individuals. At 0.56 (Table 2, model 1), the genetic correlation ( G ) between the 2 traits was highly significant, implying that 31% of the genetic variance in the 2 traits is due to the pleiotropic actions of a gene or group of genes, ie, G2 (0.56) The remainder of covariation was due to factors not specified in the model, which could include unmeasured covariates as well as residual genetic effects. (Figure 3). The phenotypic correlation coefficient for HDL and LDL was stronger than the correlation coefficient for HDL and LDL median diameters (r 0.54 versus r 0.46, respectively; data not shown), an alternate particle distribution metric used previously. Quantitative Genetic Analyses of HDL and LDL HDL and LDL were subjected to quantitative genetic analyses to determine whether a genetic component contributed to their variation. Table 1 gives the results of the analyses of samples from 1157 individuals in the present study. Covariates that satisfied an entry criterion of P 0.1 in preliminary univariate analyses included sex, age, diabetes status, contraceptive use, and hypertension medications, and their coefficients are given in Table 1. Covariates accounted for 14% and 12% of total variation in HDL and LDL, respectively. Heritabilities of both traits were highly significant (P ); the additive effects of genes accounted for 45% and 44% of residual variation (39% of total variation) in HDL and LDL, respectively. TG level, a trait previously shown to be correlated with lipoprotein size, was also highly heritable (h ), and covariates explained 16% of total variation (Table 1). Effect of TG Concentrations on the Correlation of HDL and LDL Each of the lipoprotein size traits was strongly correlated with TG levels, with phenotypic correlations of 0.48 and 0.58 and genetic correlations of 0.45 and 0.76 for HDL and LDL, respectively (Table 2, models 2 and 3). Thus, shared genes accounted for 20% and 58%, respectively, of the genetic variance in each pair of traits, and this raised the possibility that the genetic correlation between HDL and LDL might be due exclusively to their common correlation with TG levels. After including TG level in the model as a covariate, we found that the phenotypic correlation decreased from 0.51 to 0.31 but that there remained a small, but significant, genetic correlation between HDL and LDL after the adjustment (Table 2, model 4). Discussion In the present study, we have devised new metrics for lipoprotein size distribution phenotype. The variables HDL and LDL were based on inspection of the pattern of correlations between HDL and LDL fractions. Potentially, the range of values can be 1000 (all absorbance in the small size interval) to 1000 (all absorbance in the large size interval). LDL more closely approached that potential than TABLE 1. Results of Quantitative Genetic Analyses of HDL, LDL, and Log e TG (millimolar), Including for Each Trait Mean, SD, h 2,, and Proportion of Trait Variance Explained by Covariates Parameter HDL (n 1157) LDL (n 1157) Log e TGs (n 1135) Mean (male) SD h Sex (female) Age Age Diabetes status Contraceptive use Hypertension medications Variance explained by covariates, % of total indicates covariate effects. Other covariates that were considered but did not satisfy requirement for entry into the model included several dietary variables (such as alcohol and fat intake), usage of medications (such as for diabetes and hypertension), smoking, and menopause status.

4 1832 Arterioscler Thromb Vasc Biol. November 2001 TABLE 2. Results of Bivariate Quantitative Genetic Analyses of HDL, LDL, and Log e TG, Including Residual Heritabilities for Each Trait, Genetic and Environmental Correlations, and Overall Phenotypic Correlations Calculated as Given in the Text Residual h 2 Phenotypic Correlation Genetic Correlation Environmental Correlation Model Traits n Trait 1 Trait 2 P P * G P * E P * 1 HDL, LDL HDL, TG LDL, TG HDL, LDL( TG) Covariates are same as in Table 1. *P values were estimated by comparison with the likelihood of a nested model in which either G or E was fixed at zero (for G and E, respectively), or both were fixed at zero (for P ). Log e TG was included in the model as a covariate. did HDL ( 683 to 863 versus 424 to 356, respectively). Previously, we have used median diameter, the diameter, for which half the HDL (or LDL) absorbance is on larger and half is on smaller particles, to represent lipoprotein distribution. 22 Our preliminary analyses indicated that HDL and LDL median diameters gave similar results to what we report in the present study but that their phenotypic correlations were not as strong as for HDL and LDL (r versus r 0.540, respectively; data not shown). Earlier investigations have noted that LDL and HDL sizes are significantly correlated, and the present study of 1100 subjects confirms that very strong correlation, accounting for 25% of the variance. These findings suggest a common pathway(s) that influences relative particle size distribution within the spectrum of possible sizes for LDLs and HDLs and for Lp(a) as well. 13 Our data also indicate that particle size phenotypes are under substantial genetic control, with both traits having heritabilities (h 2 )of 45%, consistent with previous estimates of 30% to 50% for LDL 30 and HDL 29,31 size variables. The novel finding of the present study is that genes account for a substantial proportion of the coordinated variation in lipoprotein size phenotypes. In fact, 31% of the genetic variance in the 2 traits is influenced by a set of genes with pleiotropic effects. This suggests that at least 1 of the pathways responsible for coordinated variation in lipoprotein sizes is under genetic control. What sort of genes might these be? On the basis of known metabolic pathways, a number of proteins might be invoked as candidates for the coordinated changes that we have detected in the present study. For example, cholesteryl ester transfer protein (CETP) facilitates exchange of neutral lipids among lipoproteins. Studies have demonstrated that CETP induces the net transfer of TGs from VLDL to LDL and HDL, which, coupled with lipase activity, is associated with the reduction of HDL and LDL sizes It is likely that this pathway would be influenced by TG levels, and recently, Pownall et al 35 have shown that changes in TG levels and the attendant changes in CETP are determinants of HDL and LDL composition and structure. A number of polymorphisms in the CETP gene are associated with variation in CETP levels and with HDL measures. 36 LIPC, which encodes hepatic lipase, is also a strong candidate for a lipoprotein size gene. Hepatic lipase activity is strongly associated with HDL size phenotype, and several studies have reported an association with LDL size as well However, when it was investigated in the same study, HDL but not LDL size was associated with the C 514T polymorphism in the hepatic lipase gene previously reported to show differences in enzyme activities. 44 Genetic studies have supported this association, with linkage of a locus affecting HDL size phenotype to the region on chromosome 15 containing the hepatic lipase gene 45 and suggestive evidence for linkage of a gene for LDL peak diameter in the same region. 46 Taken together, these data suggest hepatic lipase as a strong candidate for 1 of the genes globally affecting lipoprotein size distributions. Our data help to delineate the nature of the relationship among TGs and LDL and HDL size variation. We found strong correlations of TGs with LDL and HDL size, as has been observed in numerous studies already cited, and we also noted that a substantial proportion of the correlation was due to pleiotropic effects of the genes. This suggested the possibility that the genetic correlation between LDL and HDL size might be due to genetic effects mediated through TGs. However, when TG level was included in the model as a covariate, we found that the genetic correlation between the LDL and HDL size remained significant, albeit small. Taken together, these observations suggest distinct sets of genes that influence lipoprotein size: those that influence TGs, a critical component of lipoprotein particles, and those that affect lipoprotein size independent of TGs (although we cannot exclude the possibility that a more direct measure of TG-rich lipoproteins might absorb the TG-independent correlation). This observation extends our earlier study, which suggested a significant TG-independent phenotypic correlation between HDL and LDL sizes. 15 Thus, in addition to the factors uniquely influencing each trait, there appear to be genetic and nongenetic factors, TG-related and TG independent, that are responsible for covariation in lipoprotein size traits. Acknowledgments This work was supported by a grant from the National Institutes of Health (HL-45522). The authors are grateful to Perry H. Moore, Jr, Wendy R. Shelledy, and Jane F. VandeBerg for technical assistance and to Dr John L. VandeBerg for sharing his TG concentration data. References 1. Austin MA, Breslow JL, Hennekens CH, Buring JE, Willett WC, Krauss RM. Low-density lipoprotein subclass patterns and risk of myocardial infarction. JAMA. 1988;260: Gardner CD, Fortmann SP, Krauss RM. Association of small low-density lipoprotein particles with the incidence of coronary artery disease in men and women. JAMA. 1996;276: Stampfer MJ, Krauss RM, Ma J, Blanche PJ, Holl LG, Sacks FM, Hennekens CH. A prospective study of triglyceride level, low-density

5 Rainwater et al Genetic Control of HDL and LDL Size 1833 lipoprotein particle diameter, and risk of myocardial infarction. JAMA. 1996;276: Lamarche B, Tchernof A, Moorjani S, Cantin B, Dagenais GR, Lupien PJ, Després J-P. Small, dense low-density lipoprotein particles as a predictor of the risk of ischemic heart disease in men: prospective results from the Quebec Cardiovascular Study. Circulation. 1997;95: Brewer HB Jr. Hypertriglyceridemia: changes in the plasma lipoproteins associated with an increased risk of cardiovascular disease. Am J Cardiol. 1999;83:3F 12F. 6. Barter PJ, Rye KA. High density lipoproteins and coronary heart disease. Atherosclerosis. 1996;121: Tailleux A, Fruchart JC. HDL heterogeneity and atherosclerosis. Crit Rev Clin Lab Sci. 1996;33: Sweetnam PM, Bolton CH, Yarnell JWG, Bainton D, Baker IA, Elwood PC, Miller NE. Associations of the HDL 2 and HDL 3 cholesterol subfractions with the development of ischemic heart disease in British men: the Caerphilly and Speedwell collaborative heart disease studies. Circulation. 1994;90: Atger V, Giral P, Simon A, Cambillau M, Levenson J, Gariepy J, Megnien JL, Moatti N, PCVMETRA Group. High-density lipoprotein subfractions as markers of early atherosclerosis. Am J Cardiol. 1995;75: Freedman DS, Otvos JD, Jeyarajah EJ, Barboriak JJ, Anderson AJ, Walker JA. Relation of lipoprotein subclasses as measured by proton nuclear magnetic resonance spectroscopy to coronary artery disease. Arterioscler Thromb Vasc Biol. 1998;18: Williams PT, Krauss RM, Vranizan KM, Stefanick ML, Wood PDS, Lindgren FT. Associations of lipoproteins and apolipoproteins with gradient gel electrophoresis estimates of high density lipoprotein subfractions in men and women. Arterioscler Thromb. 1992;12: Katzel LI, Krauss RM, Goldberg AP. Relations of plasma TG and HDL-C concentrations to body composition and plasma insulin levels are altered in men with small LDL particles. Arterioscler Thromb. 1994;14: Rainwater DL, Ludwig MJ, Haffner SM, VandeBerg JL. Lipid and lipoprotein factors associated with variation in Lp(a) density. Arterioscler Thromb Vasc Biol. 1995;15: Ohta T, Saku K, Takata K, Nagata N, Maung KK, Matsuda I. Fractional esterification rate of cholesterol in high density lipoprotein (HDL) can predict the particle size of low density lipoprotein and HDL in patients with coronary heart disease. Atherosclerosis. 1997;135: Rainwater DL. Lipoprotein correlates of LDL particle size. Atherosclerosis. 2000;148: MacCluer JW, Stern MP, Almasy L, Atwood LA, Blangero J, Comuzzie AG, Dyke B, Haffner SM, Henkel RD, Hixson JE, et al. Genetics of atherosclerosis risk factors in Mexican Americans. Nutr Rev. 1999;57: S59 S Diabetes mellitus: report of a WHO Study Group. World Health Organ Tech Rep Ser. 1985;727: Cheng M-L, Woodford SC, Hilburn JL, VandeBerg JL. A novel system for storage of sera frozen in small aliquots. J Biochem Biophys Methods. 1986;13: Warnick GR, Benderson J, Albers JJ. Dextran sulfate-mg 2 precipitation procedure for quantitation of high-density-lipoprotein cholesterol. Clin Chem. 1982;28: Demacker PN, Hijmans AG, Brenninkmeijer BJ, Jansen AP, Van t Laar A. Five methods for determining low-density lipoprotein cholesterol compared. Clin Chem. 1984;30: Demacker PNM, Toenhake-Dijkstra H, De Rijke YB, Stalenhoef AFH, Stuyt PMJ, Willems HL. On the presumed inaccuracy of the Friedewald formula in hypertriglyceridemic plasma: a role for imprecise analysis? Clin Chem. 1996;42: Rainwater DL, Moore PH Jr, Shelledy WR, Dyer TD, Slifer SH. Characterization of a composite gradient gel for the electrophoretic separation of lipoproteins. J Lipid Res. 1997;38: Rainwater DL. Electrophoretic separation of LDL and HDL subclasses. Methods Mol Biol. 1998;110: Singh ATK, Rainwater DL, Haffner SM, VandeBerg JL, Shelledy WR, Moore PH Jr, Dyer TD. Effect of diabetes on lipoprotein size. Arterioscler Thromb Vasc Biol. 1995;15: Williams PT, Krauss RM, Nichols AV, Vranizan KM, Wood PDS. Identifying the predominant peak diameter of high-density and lowdensity lipoproteins by electrophoresis. J Lipid Res. 1990;31: Lange K, Boehnke M. Extensions to pedigree analysis, IV: covariance components models for multivariate traits. Am J Med Genet. 1983;14: Boehnke M, Moll PP, Kottke BA, Weidman WH. Partitioning the variability of fasting plasma glucose levels in pedigrees. Am J Epidemiol. 1987;125: Comuzzie AG, Blangero J, Mahaney MC, Sharp RM, VandeBerg JL, Stern MP, MacCluer JW. Triiodothyronine exerts a major pleiotropic effect on reverse cholesterol transport phenotypes. Arterioscler Thromb Vasc Biol. 1996;16: Comuzzie AG, Rainwater DL, Blangero J, Mahaney MC, VandeBerg JL, MacCluer JW. Shared and unique genetic effects among seven HDL phenotypes. Arterioscler Thromb Vasc Biol. 1997;17: Austin MA. Genetic epidemiology of low-density lipoprotein subclass phenotypes. Ann Med. 1992;24: Williams PT, Vranizan KM, Austin MA, Krauss RM. Familial correlations of HDL subclasses based on gradient gel electrophoresis. Arterioscler Thromb. 1992;12: Clay MA, Newnham HH, Forte TM, Barter PI. Cholesteryl ester transfer protein and hepatic lipase activity promote shedding of apo A-I from HDL and subsequent formation of discoidal HDL. Biochim Biophys Acta. 1992;1124: Lagrost L, Gambert P, Lallemant C. Combined effects of lipid transfers and lipolysis on gradient gel patterns of human plasma LDL. Arterioscler Thromb. 1994;14: Chung BH, Segrest JP, Franklin F. In vitro production of -very low density lipoproteins and small, dense low density lipoproteins in mildly hypertriglyceridemic plasma: role of activities of lecithin:cholesterol acyltransferase, cholesterylester transfer proteins and lipoprotein lipase. Atherosclerosis. 1998;141: Pownall HJ, Brauchi D, Kilinç C, Osmundsen K, Pao Q, Payton-Ross C, Gotto AM Jr, Ballantyne CM. Correlation of serum triglyceride and its reduction by omega-3 fatty acids with lipid transfer activity and the neutral lipid compositions of high-density and low-density lipoproteins. Atherosclerosis. 1999;143: Bruce C, Chouinard RA Jr, Tall AR. Plasma lipid transfer proteins, high-density lipoproteins, and reverse cholesterol transport. Annu Rev Nutr. 1998;18: Kuusi T, Saarinen P, Nikkila EA. Evidence for the role of hepatic endothelial lipase in the metabolism of plasma high density lipoprotein2 in man. Atherosclerosis. 1980;36: Johansson J, Nilsson-Ehle P, Carlson LA, Hamsten A. The association of lipoprotein and hepatic lipase activities with high density lipoprotein subclass levels in men with myocardial infarction at a young age. Atherosclerosis. 1991;86: Katzel LI, Coon PJ, Busby MJ, Gottlieb SO, Krauss RM, Goldberg AP. Reduced HDL 2 cholesterol subspecies and elevated postheparin hepatic lipase activity in older men with abdominal obesity and asymptomatic myocardial ischemia. Arterioscler Thromb. 1992;12: Grundy SM, Vega GL, Otvos JD, Rainwater DL, Cohen JC. Hepatic lipase activity influences high density lipoprotein subclass distribution in normotriglyceridemic men: genetic and pharmacological evidence. J Lipid Res. 1999;40: Watson TDG, Caslake MJ, Freeman DJ, Griffin BA, Hinnie J, Packard CJ, Shepherd J. Determinants of LDL subfraction distribution and concentrations in young normolipidemic subjects. Arterioscler Thromb. 1994;14: Campos H, Dreon DM, Krauss RM. Associations of hepatic and lipoprotein lipase activities with changes in dietary composition and low density lipoprotein subclasses. J Lipid Res. 1995;36: Zambon A, Hokanson JE, Brown BG, Brunzell JD. Evidence for a new pathophysiological mechanism for coronary artery disease regression: hepatic lipase-mediated changes in LDL density. Circulation. 1999;99: Couture P, Otvos JD, Cupples LA, Lahoz C, Wilson PWF, Schaefer EJ, Ordovas JM. Association of the C-514T polymorphism in the hepatic lipase gene with variations in lipoprotein subclass profiles: the Framingham Offspring Study. Arterioscler Thromb Vasc Biol. 2000;20: Almasy L, Hixson JE, Rainwater DL, Cole S, Williams JT, Mahaney MC, VandeBerg JL, Stern MP, MacCluer JW, Blangero J. Human pedigree-based quantitative-trait-locus mapping: localization of two genes influencing HDL-cholesterol metabolism. Am J Hum Genet. 1999; 64: Allayee H, Dominguez KM, Aouizerat BE, Krauss RM, Rotter JI, Lu JY, Cantor RM, De Bruin TWA, Lusis AJ. Contribution of the hepatic lipase gene to the atherogenic lipoprotein phenotype in familial combined hyperlipidemia. J Lipid Res. 2000;41: