- Email: [email protected]
Studies on the relationship between the cholesterol content in total high density lipoprotein and its subfractions, HDL2 and HDL3 in normo- and hyperlipidemic subjects
Arherosclerosis, 61 (1986) 225-229 Elsevier Scientific Publishers Ireland,
4TH
225 Ltd.
03815
Studies on the Relationship between the Cholesterol Content in Total High Density Lipoprotein and its Subfractions, HDL, and HDL, in Normo- and Hyperlipidemic Subjects Pierre N.M. Demacker,
Hendrik
Baadenhuysen,
Paul M.J. Stuyt and Albert Van ‘t Laar
Depurtmenr of Medrcine, Du:r.rion of General Inrernal Medicine and Dioismn of Climcal Chemrstty, Unruersity of Nijmegen, NiJmegen (The Netherlands)
St. Radhoud Hosprrul,
(Received 16 January, 1986) (Revised, received 12 May, 1986) (Accepted 15 May, 1986)
Summary Total high density lipoprotein (HDL) cholesterol and cholesterol in its main subfractions, HDL, and HDL,, were determined in 160 normo- and 90 hyperlipidemic subjects by density gradient ultracentrifugation (range of HDL-cholesterol: 0.05-2.85 mmol/l). Both in the normolipidemics and in the combined group HDL,-cholesterol (HDL,-chol) as well as HDL,-cholesterol (HDL,-chol) showed a parabolic relationship with total HDL-chol. The results indicate, that at low total HDL-chol values almost all cholesterol is present in the HDL, fraction, which shows a linear increase with total HDL-chol from 0 to 0.75 mmol/l. At a further increase of total HDL-chol, cholesterol is increasingly isolated in the HDL,-fraction, especially when HDL,-chol has reached its maximum (about 1.25 mmol/l). Thus, the magnitude of the absolute intra-individual variation of either HDL,-chol or HDL,-chol (in mmol/l) is related to the total HDL-chol concentration. Given the strong correlation between cholesterol in both HDL-subfractions with total HDL-chol and the inverse relationship of total HDL-chol with the risk for coronary heart disease, a rise in HDL,-chol or in HDL,-chol may be equally favorable.
Key words:
Coronary heart disease - HDL,
- HDL,
~ Risk evaluation
Introduction This work was supported financially by the Netherlands Heart Foundation. Correspondence: Dr. Pierre N.M. Demacker, Department of Medicine, Division of General Internal Medicine, St. Radboud Hospital, University of Nijmegen, Geert Grooteplein Zuid 8, P.O. Box 9101, 6500 HB Nijmegen, The Netherlands. 0021-9150/86/$03.50
iij 1986 Elsevier Scientific
Publishers
Ireland,
Epidemiological studies have demonstrated an inverse correlation between the concentration of high density lipoprotein (HDL) cholesterol (HDLchol) and the occurrence of coronary heart disease Ltd
226 [1,2]. Total HDL is heterogenous and consists of HDL,, a minor component, HDL, and HDL, [3]. These subfractions can be isolated by various ultracentrifugation techniques [3,4]. Moreover, a simple dual precipitation method for selective determination of HDL,-chol and HDL,-chol has been described as suitable for large scale studies [5]. Most studies show that the factors which ultimately influence total HDL-chol concentration primarily act on HDL,-chol [6]; HDL2 has been shown to be the most variable subfraction [6,7], which also is most closely associated with the incidence of coronary heart disease [8,9]. For the estimation of the risk for coronary heart disease, the concentration of HDL,-chol is regarded to be of less importance [g-lo], although in one study considerably lower values of HDL, were found in patients with ischemic heart disease [ll]. In a previous study we observed a rise in HDL-chol mainly due to an increase of HDL,-chol in patients with type III and IV hyperlipoproteinemia during treatment with Acipimox; HDL,-chol did not change [12]. These results were difficult to interpret because in the literature insufficient information is available concerning the relationship of the concentrations of both HDL-subfractions with each other and with total HDL. We therefore report here our results concerning the interrelationships of HDL,-chol and HDL,-chol with total HDL-chol, as determined with a recently validated density gradient ultracentrifugation method [4] in 250 subjects with a wide range of total HDL-chol concentrations.
tions of the HDL,- and HDL,-fractions as could be concluded from the colored band distributions of the lipoproteins stained with Coomassie Brilliant Blue R prior to ultracentrifugation [4]. Also 160 subjects (104 men, 56 women, mean age 39 t- 13 years) with normal lipid values participated, including 81 subjects visiting the out-patient clinic of our department, and 79 hospital workers. In these normolipidemics blood was sampled fasting or within 2 h after a light breakfast. In normolipidemics, serum lipids and HDL-chol concentrations in blood samples taken within 2 h after a light breakfast were similar to fasting values [15]. Methods Serum was separated within 2 h after blood sampling, stored at 4°C and analysed for HDL,chol and HDL,-chol by density gradient ultracentrifugation [4] within 24 h. HDL, was isolated between d = 1.063 and 1.100 g/ml and HDL, between d = 1.100 and 1.185 g/ml. Cholesterol was determined enzymatically using the catalase method [16] by means of a test-kit (No. 124087, Boehringer, Mannheim, F.R.G.). Triglycerides were determined by semi-automated calorimetry [17]. The paired data of HDL,-chol and HDL,chol were analysed by linear regression and by variance analysis using a computer program which could generate linear and higher exponential functions to fit best with the paired data in the 160 or 250 samples. Results
Materials and methods Subjects Blood samples were obtained from 90 overnight fasted subjects (70 men, 20 women, age 50 + 12 years) with various types of hyperlipoproteinemia, visiting our out-patient clinic, including 3 patients with apo C-II deficiency [13], 15 patients with type II-A, 4 with type II-B, 18 with type III, 42 with type IV and 8 patients with type V hyperlipoproteinemia. The criteria used for phenotyping have been described elsewhere [14]. Approximately 40% of these subjects received lipidlowering drugs, mainly clofibrate or cholestyramine; which did not change the density distribu-
Generally, older subjects, men or hyperlipidemics have lower total HDL-chol concentrations than younger subjects, women or normolipidemics respectivley [1,2,6,8], but there is a considerable overlap in the HDL-chol concentrations of all these groups. When the results of all these groups are combined, the largest concentration range is obtained. In this combined group the relationship between HDL,-chol and HDL,-chol with total HDL-chol can better be studied than in a selected group. In the combined group of the subjects with a total HDL-chol concentration below 0.9 mmol/l 79% (38 men, 7 women) had hyperlipidemia, whereas women and younger people were overrep-
227 TABLE
1
STATISTICAL
ANALYSIS
In each comparison
HDL,/HDL
OF THE RELATIONSHIP
total HDL-chol
(in mmol/l)
BETWEEN
Polynomial
HDL,/HDL
analysis
WITH TOTAL
HDL-CHOL
= x.
Normolipidemics (n = 160)
All subjects (n = 250)
y= r = y= r =
0.57x - 0.34 0.87 0.19x’+0.01x+0.04 0.90
y= r = y= r =
0.50x - 0.22 0.89 0.20x 2 + 0.02x + 0.03 0.93
y = r = y = I =
0.36x + 0.38 0.76 -0.17x2 +0.85x+0.04 0.79
y= r = y= r =
0.50x + 0.22 0.88 - 0.20x’ +0.98x - 0.03 0.95
rotal
Linear regressIon Polynomial
analysis
HDL3-chol 2.0 A t t
0
0.5
1
1.5
2
2.5
3
HDL-chol HDLychol t
B
1.2
0.6
0
OR HDL,-CHOL
totul
Linear regression
1.8
HDL2-CHOL
0.5
1
1.5
2
2.5
HDL-chol
3
resented in the group with a total HDL-chol of 1.35 mmol/l or higher. The concentrations (mean + SD) of total HDL-chol, HDL,-chol and HDL,-chol in the 250 serum samples were: 1.12 * 0.41; 0.34 _t 0.23, and 0.78 -t 0.23 mmol/l (CV’s: 36.7; 68.7 and 29.6%, respectively). Thus, in agreement with [6 and 71, HDLz-chol shows the largest relative intra-individual variation. However, the absolute intra-individual variation of the cholesterol content in both HDL-subfractions (standard deviation) is nevertheless similar. The relationship between the cholesterol content in either the HDL, or the HDL, fractions with total HDL-chol was statistically analysed in both the normolipidemics (n = 160) and in the combined groups of normo- and hyperlipidemics (n = 250). In both groups the relations between HDL,-chol or HDL,-chol with total HDL-chol could best be fitted by a second order polynomial rather than by a linear curve (Table 1). This was clearest visible in the combined group of subjects (Fig. 1). The relation between HDL,-chol (y) and total HDL-chol (x) could be depicted best by the parabolic function y = -0.20x’ + 0.98x - 0.03 (Fig. 1). The relationship between the concentraFig. 1. Relationship of HDL,-chol (A) and HDL,-chol (B) with total HDL-chol (mmoI/l) in the 250 sera from normoand hyperlipidemic subjects.
228 tions of HDL,-chol (y) and total HDL-chol (x) could be depicted best by the parabolic function y = 0.20x2 + 0.02x + 0.03. These results indicate that at low total HDL-chol concentration (< 0.75 mmol/l) almost all cholesterol is present in the HDL,-fraction. When total HDL-chol rises from 0.75 to 1.25 mmol/l, the rise in HDL,-chol (from 0.16 to 0.37 mmol/l) becomes important, but that in HDL,-chol (from 0.60 to 0.88 mmol/l) is still larger in absolute amounts, as can be calculated with the above formulas. A rise at above-normal HDL-chol levels is caused almost exclusively by an increase of HDL,-chol. Discussion On the basis of these results, one could speculate that at a HDL-chol concentration near 0.75 mmol/l, saturation of the HDL,-particle prevents further accumulation of cholesterol and other lipids. The excess of these lipids is now integrated into a particle within the HDL,-fraction. The mechanisms underlying such a process are obscure. It has been proposed, that HDL, is converted into HDL, by uptake of surface components from triglyceride-rich lipoproteins when their triglyceride moiety is hydrolysed by lipoprotein lipase [18]. In this conversion lecithin-cholesteryl acyltransferase may be involved, as demonstrated in vitro [19]. The reverse process, namely the conversion of HDL, into HDL, seems to be mediated by hepatic lipase [20]. However, changes in the activity of these enzymes explain only a continuous spectrum of HDL-particles rather than the typical density distribution with 2 distinct fractions [19]. Recently, 2 classes of HDL-particles could be isolated by immunoaffinity chromatography: apo A-I containing particles with apo A-II and apo A-I containing particles without A-II [21]. The latter particles were preferentially isolated in the HDL,-fraction and the former in the HDL,fraction. This finding was corroborated by others [22]. Apparently, absence of apo A-II is associated with an enhanced ability to integrate cholesterol into the HDL particle. Thus, an altered rate of synthesis of A-II in relation to that of apo A-I [23], resulting in a change of the apo A-I/A-II ratio may be the underlying cause for a shift of HDL, to HDL, or vice versa.
Our results could explain findings in the literature concerning the effects of drugs, diet or physical activity on HDL,-chol and HDL,-chol concentrations. It is now comprehensible that Gofman et al. [8] observed 32% lower concentrations of HDL,-chol and 8% lower HDL,-chol values in patients with ischemic heart disease than in controls, because total HDL-concentrations in the patients were rather low. The effect of physical activity is usually studied in subjects with a normal or above-normal total HDL-chol concentration [6]. In these subjects HDL,-chol has almost reached its maximum so that an almost exclusive rise in HDL,-chol is obtained. It can now also be understood that Packard et al. [24] observed apparently different responses in normo- and hypertriglyceridemic subjects to nicotinic acid treatment. Treatment of patients with hypertriglyceridemia resulted in a rise of HDL, because the starting level of HDL was rather low. In the normotriglyceridemics an increase in HDL, was found, because these subjects had generally higher HDL-values. Assuming that the risk for coronary heart disease is linearly related to the inverse of the HDLchol concentration, our findings indicate, that a rise in HDL,-chol may be equally favorable as a rise in HDL,-chol. It may then be questioned whether subfractionation of HDL adds much to the risk estimation based on measurement of total HDL-chol only. Future epidemiological studies may shed light on this important question. However, at the moment the start of these studies is troubled by the imprecision and accuracy problems of HDL subclass measurement [25]. Acknowledgements We thank Mrs. Anneke Hijmans, Pieternel Van Heyst and Heidi Hak-Lemmers for excellent technical assistance, Dr. A.F.H. Stalenhoef for providing serum samples and Mrs. A. Verweijen and Mrs. Ans Ruesen for the preparation of the manuscript. References 1 Castelli,
W.P., Doyle, J.T., Gordon, T., HDL-cholesterol and other lipids in coronary heart disease, Circulation, 55 (1977) 161.
229 2 Miller. N.E.. Forde, O.H., Thelle, D.S. and Mjos, O.D., The Tromso study, high density lipprotein and coronary heart disease - A prospective case-control study, Lancet, i (1977) 965. 3 DeLalla. O.F., Elliot, H.A. and Gofman, J.W.. Ultracentrifugal studies of high density serum lipoproteins in clinically healthy adults. Amer. J. Physiol.. 179 (1954) 333. 4 Demacker. P.N.M., Van Sommeren-Zondag. D.F., Stalenhoef. A.F.H.. Stuyt, P.M. and Van ‘t Laar, A.. Ultracentrifugation in swinging-bucket and fixed-angle rotors evaluated for isolation and determination of high-density lipoprotein subfractions HDL, and HDL,, Clin. Chem., 29 (1983) 656. 5 Lewis, L.I., Miller. G.J., Burstein, M., Slagle, S. and Eder, H.A., Separation and quantitation of subclasses of human plasma high density lipoprotein by a simple precipitation procedure, J. Lipid. Res., 23 (1982) 1206. 6 Kraus. R.M., Regulation of high density lipoprotein levels, Med. Clin. N. Amer., 66 (1982) 403. I Anderson, D.W.. HDL-cholesterol - The variable components, Lancet, i (1978) 819. 8 Gofman, J.W., Young, W. and Tandy, R., Ischemic heart disease, atherosclerosis. and longevity, Circulation, 179 (1954) 333. Y Miller. N.E.. Hammett. S., Saltissi, S., Rao, S.. Van Zeller. H.. Coltart, J. and Lewis, B., Relation of angiographically defined coronary artery disease to plasma lipoprotein subfractions and apolipoproteins. Brit. Med. J.. 282 (1981) 1741. F.C.. Clark. R.S., Simpson, H.S. and Bal10 Ballantyne. lantyne. D.. High density and low density lipoprotein subfractions in survivors of myocardial infarction and in control SubJectS, Metabolism, 31 (1982) 433. E. and 11 Brook. J.G.. Aviram, M., Viener, A., Shilansky, Markiewicz, W.. High-density lipoprotein subfractions in normolipidemic patients with coronary atherosclerosis, Circulation, 66 (1982) 923. P.N.M. and 12 Stuyt. P.M.J., Stalenhoef, A.F.H., Demacker, Van ‘t Laar, A.. A comparative study of the effects of Acipimox and Clofibrate in type III and type IV hyperlipoproteinemia. Atherosclerosis. 55 (1985) 51. A.F.H.. Casparie. A.F., Demacker, P.N.M., 13 Stalenhoef, Stouten, J.T.J.. Lutterman, J.A. and Van ‘t Laar. A., Combined deficiency of apohpoprotein CII and lipoprotein lipase in familial hyperchylomicronemia, Metabolism, 30 (1981) 919. 14 Stuyt, P.M.J.. Demacker, P.N.M. and Van ‘t Laar. A., Serum lipids, lipoproteins and apohpoprotein E phenotypes
15
16
17
18
19
20
21
22
23
24
25
in relatives of patients with type III hyperlipoproteinemia, Europ. J. Clin. Invest., 14 (1984) 219. Demacker, P.N.M., Schade, R.W.B., Jansen, R.T.P. and Van ‘t Laar, A., Intra-individual variation of serum triglycerides and high-density lipoprotein cholesterol. cholesterol in normal humans, Atherosclerosis. 45 (1982) 259. Roschlau, P., Bernt. E. and Gruber. W.. Enzymatische Bestimmung des Gesamt-Cholesterins im Serum. J. Clin. Chem. Biochem., 12 (1974) 403. Demacker, P.N.M., Van Oppenraay, J.B., Baadenhuysen. H. and Jansen, A.P., An improved semi-automated method for the calorimetric determination of triglycerides m serum, Clin. Chim. Acta. 64 (1975) 45. Patsch, J.R., Gotto, Jr., A.M., Olivecrona, T. and Eisenberg, S.. Formation of high-density lipoprotein,-like particles during lipolysis of very low density lipoproteins in vitro, Proc. Nat. Acad. Sci. (USA), 75 (197X) 4519. Dieplinger, H.. Zechner, R. and Kostner, G.M.. The in vitro formation of HDLz during the action of LCAT The role of triglyceride-rich lipoproteins, J. Lipid. Res. 26 (1985) 273. Jansen, H.. Van Tol, A. and Hiilsmann. W.C., On the metabolic function of heparin releasable liver lipase, Biothem. Biophys. Res. Commun., 92 (1980) 53. Cheung. M.C. and Albers, J.J., Characterization of lipoprotein particles isolated by immunoaffinity chromatography - Particles containing A-I and A-II and particles containing A-I but not A-II, J. Biol. Chem.. 259 (1984) 12201. Atmeh. R.F.. Shepherd, J. and Packard, C.J., Subpopulations of apolipoprotein A-I in human high-density lipoproteins, their metabolic properties and response to drug therapy, Biochim. Biophys. Acta, 751 (1983) 175. Shepherd, J.. Packard, C.J., Patsch, J.R. and Gotto. Jr., A.M., Effects of nicotinic acid therapy on plasma high density lipoprotein subfraction distribution and composition and on apolipoprotein A metabolism. J. Clin. Invest.. 63 (1979) 858. Packard, C.J., Stewart. J.M.. Third, J.L., Morgan, H.G.. Lowry, T.D. and Shepherd, J., Effects of nicotinic acid therapy on high-density lipoprotein metabolism in type 111 and type IV hyperlipoproteinemia, Biochim. Biophys. Acta, 618 (1980) 53. Demacker, P.N.M., Hak-Lemmers, H.L.Mr, Hijmans. A.G.M. and Baadenhuysen, H., Evaluation of the dual-precipitation method for determination of cholesterol in highdensity lipoprotein subfractions HDL, and HDL? in serum, Clin. Chem.. 32 (1986) 819.
4TH
225 Ltd.
03815
Studies on the Relationship between the Cholesterol Content in Total High Density Lipoprotein and its Subfractions, HDL, and HDL, in Normo- and Hyperlipidemic Subjects Pierre N.M. Demacker,
Hendrik
Baadenhuysen,
Paul M.J. Stuyt and Albert Van ‘t Laar
Depurtmenr of Medrcine, Du:r.rion of General Inrernal Medicine and Dioismn of Climcal Chemrstty, Unruersity of Nijmegen, NiJmegen (The Netherlands)
St. Radhoud Hosprrul,
(Received 16 January, 1986) (Revised, received 12 May, 1986) (Accepted 15 May, 1986)
Summary Total high density lipoprotein (HDL) cholesterol and cholesterol in its main subfractions, HDL, and HDL,, were determined in 160 normo- and 90 hyperlipidemic subjects by density gradient ultracentrifugation (range of HDL-cholesterol: 0.05-2.85 mmol/l). Both in the normolipidemics and in the combined group HDL,-cholesterol (HDL,-chol) as well as HDL,-cholesterol (HDL,-chol) showed a parabolic relationship with total HDL-chol. The results indicate, that at low total HDL-chol values almost all cholesterol is present in the HDL, fraction, which shows a linear increase with total HDL-chol from 0 to 0.75 mmol/l. At a further increase of total HDL-chol, cholesterol is increasingly isolated in the HDL,-fraction, especially when HDL,-chol has reached its maximum (about 1.25 mmol/l). Thus, the magnitude of the absolute intra-individual variation of either HDL,-chol or HDL,-chol (in mmol/l) is related to the total HDL-chol concentration. Given the strong correlation between cholesterol in both HDL-subfractions with total HDL-chol and the inverse relationship of total HDL-chol with the risk for coronary heart disease, a rise in HDL,-chol or in HDL,-chol may be equally favorable.
Key words:
Coronary heart disease - HDL,
- HDL,
~ Risk evaluation
Introduction This work was supported financially by the Netherlands Heart Foundation. Correspondence: Dr. Pierre N.M. Demacker, Department of Medicine, Division of General Internal Medicine, St. Radboud Hospital, University of Nijmegen, Geert Grooteplein Zuid 8, P.O. Box 9101, 6500 HB Nijmegen, The Netherlands. 0021-9150/86/$03.50
iij 1986 Elsevier Scientific
Publishers
Ireland,
Epidemiological studies have demonstrated an inverse correlation between the concentration of high density lipoprotein (HDL) cholesterol (HDLchol) and the occurrence of coronary heart disease Ltd
226 [1,2]. Total HDL is heterogenous and consists of HDL,, a minor component, HDL, and HDL, [3]. These subfractions can be isolated by various ultracentrifugation techniques [3,4]. Moreover, a simple dual precipitation method for selective determination of HDL,-chol and HDL,-chol has been described as suitable for large scale studies [5]. Most studies show that the factors which ultimately influence total HDL-chol concentration primarily act on HDL,-chol [6]; HDL2 has been shown to be the most variable subfraction [6,7], which also is most closely associated with the incidence of coronary heart disease [8,9]. For the estimation of the risk for coronary heart disease, the concentration of HDL,-chol is regarded to be of less importance [g-lo], although in one study considerably lower values of HDL, were found in patients with ischemic heart disease [ll]. In a previous study we observed a rise in HDL-chol mainly due to an increase of HDL,-chol in patients with type III and IV hyperlipoproteinemia during treatment with Acipimox; HDL,-chol did not change [12]. These results were difficult to interpret because in the literature insufficient information is available concerning the relationship of the concentrations of both HDL-subfractions with each other and with total HDL. We therefore report here our results concerning the interrelationships of HDL,-chol and HDL,-chol with total HDL-chol, as determined with a recently validated density gradient ultracentrifugation method [4] in 250 subjects with a wide range of total HDL-chol concentrations.
tions of the HDL,- and HDL,-fractions as could be concluded from the colored band distributions of the lipoproteins stained with Coomassie Brilliant Blue R prior to ultracentrifugation [4]. Also 160 subjects (104 men, 56 women, mean age 39 t- 13 years) with normal lipid values participated, including 81 subjects visiting the out-patient clinic of our department, and 79 hospital workers. In these normolipidemics blood was sampled fasting or within 2 h after a light breakfast. In normolipidemics, serum lipids and HDL-chol concentrations in blood samples taken within 2 h after a light breakfast were similar to fasting values [15]. Methods Serum was separated within 2 h after blood sampling, stored at 4°C and analysed for HDL,chol and HDL,-chol by density gradient ultracentrifugation [4] within 24 h. HDL, was isolated between d = 1.063 and 1.100 g/ml and HDL, between d = 1.100 and 1.185 g/ml. Cholesterol was determined enzymatically using the catalase method [16] by means of a test-kit (No. 124087, Boehringer, Mannheim, F.R.G.). Triglycerides were determined by semi-automated calorimetry [17]. The paired data of HDL,-chol and HDL,chol were analysed by linear regression and by variance analysis using a computer program which could generate linear and higher exponential functions to fit best with the paired data in the 160 or 250 samples. Results
Materials and methods Subjects Blood samples were obtained from 90 overnight fasted subjects (70 men, 20 women, age 50 + 12 years) with various types of hyperlipoproteinemia, visiting our out-patient clinic, including 3 patients with apo C-II deficiency [13], 15 patients with type II-A, 4 with type II-B, 18 with type III, 42 with type IV and 8 patients with type V hyperlipoproteinemia. The criteria used for phenotyping have been described elsewhere [14]. Approximately 40% of these subjects received lipidlowering drugs, mainly clofibrate or cholestyramine; which did not change the density distribu-
Generally, older subjects, men or hyperlipidemics have lower total HDL-chol concentrations than younger subjects, women or normolipidemics respectivley [1,2,6,8], but there is a considerable overlap in the HDL-chol concentrations of all these groups. When the results of all these groups are combined, the largest concentration range is obtained. In this combined group the relationship between HDL,-chol and HDL,-chol with total HDL-chol can better be studied than in a selected group. In the combined group of the subjects with a total HDL-chol concentration below 0.9 mmol/l 79% (38 men, 7 women) had hyperlipidemia, whereas women and younger people were overrep-
227 TABLE
1
STATISTICAL
ANALYSIS
In each comparison
HDL,/HDL
OF THE RELATIONSHIP
total HDL-chol
(in mmol/l)
BETWEEN
Polynomial
HDL,/HDL
analysis
WITH TOTAL
HDL-CHOL
= x.
Normolipidemics (n = 160)
All subjects (n = 250)
y= r = y= r =
0.57x - 0.34 0.87 0.19x’+0.01x+0.04 0.90
y= r = y= r =
0.50x - 0.22 0.89 0.20x 2 + 0.02x + 0.03 0.93
y = r = y = I =
0.36x + 0.38 0.76 -0.17x2 +0.85x+0.04 0.79
y= r = y= r =
0.50x + 0.22 0.88 - 0.20x’ +0.98x - 0.03 0.95
rotal
Linear regressIon Polynomial
analysis
HDL3-chol 2.0 A t t
0
0.5
1
1.5
2
2.5
3
HDL-chol HDLychol t
B
1.2
0.6
0
OR HDL,-CHOL
totul
Linear regression
1.8
HDL2-CHOL
0.5
1
1.5
2
2.5
HDL-chol
3
resented in the group with a total HDL-chol of 1.35 mmol/l or higher. The concentrations (mean + SD) of total HDL-chol, HDL,-chol and HDL,-chol in the 250 serum samples were: 1.12 * 0.41; 0.34 _t 0.23, and 0.78 -t 0.23 mmol/l (CV’s: 36.7; 68.7 and 29.6%, respectively). Thus, in agreement with [6 and 71, HDLz-chol shows the largest relative intra-individual variation. However, the absolute intra-individual variation of the cholesterol content in both HDL-subfractions (standard deviation) is nevertheless similar. The relationship between the cholesterol content in either the HDL, or the HDL, fractions with total HDL-chol was statistically analysed in both the normolipidemics (n = 160) and in the combined groups of normo- and hyperlipidemics (n = 250). In both groups the relations between HDL,-chol or HDL,-chol with total HDL-chol could best be fitted by a second order polynomial rather than by a linear curve (Table 1). This was clearest visible in the combined group of subjects (Fig. 1). The relation between HDL,-chol (y) and total HDL-chol (x) could be depicted best by the parabolic function y = -0.20x’ + 0.98x - 0.03 (Fig. 1). The relationship between the concentraFig. 1. Relationship of HDL,-chol (A) and HDL,-chol (B) with total HDL-chol (mmoI/l) in the 250 sera from normoand hyperlipidemic subjects.
228 tions of HDL,-chol (y) and total HDL-chol (x) could be depicted best by the parabolic function y = 0.20x2 + 0.02x + 0.03. These results indicate that at low total HDL-chol concentration (< 0.75 mmol/l) almost all cholesterol is present in the HDL,-fraction. When total HDL-chol rises from 0.75 to 1.25 mmol/l, the rise in HDL,-chol (from 0.16 to 0.37 mmol/l) becomes important, but that in HDL,-chol (from 0.60 to 0.88 mmol/l) is still larger in absolute amounts, as can be calculated with the above formulas. A rise at above-normal HDL-chol levels is caused almost exclusively by an increase of HDL,-chol. Discussion On the basis of these results, one could speculate that at a HDL-chol concentration near 0.75 mmol/l, saturation of the HDL,-particle prevents further accumulation of cholesterol and other lipids. The excess of these lipids is now integrated into a particle within the HDL,-fraction. The mechanisms underlying such a process are obscure. It has been proposed, that HDL, is converted into HDL, by uptake of surface components from triglyceride-rich lipoproteins when their triglyceride moiety is hydrolysed by lipoprotein lipase [18]. In this conversion lecithin-cholesteryl acyltransferase may be involved, as demonstrated in vitro [19]. The reverse process, namely the conversion of HDL, into HDL, seems to be mediated by hepatic lipase [20]. However, changes in the activity of these enzymes explain only a continuous spectrum of HDL-particles rather than the typical density distribution with 2 distinct fractions [19]. Recently, 2 classes of HDL-particles could be isolated by immunoaffinity chromatography: apo A-I containing particles with apo A-II and apo A-I containing particles without A-II [21]. The latter particles were preferentially isolated in the HDL,-fraction and the former in the HDL,fraction. This finding was corroborated by others [22]. Apparently, absence of apo A-II is associated with an enhanced ability to integrate cholesterol into the HDL particle. Thus, an altered rate of synthesis of A-II in relation to that of apo A-I [23], resulting in a change of the apo A-I/A-II ratio may be the underlying cause for a shift of HDL, to HDL, or vice versa.
Our results could explain findings in the literature concerning the effects of drugs, diet or physical activity on HDL,-chol and HDL,-chol concentrations. It is now comprehensible that Gofman et al. [8] observed 32% lower concentrations of HDL,-chol and 8% lower HDL,-chol values in patients with ischemic heart disease than in controls, because total HDL-concentrations in the patients were rather low. The effect of physical activity is usually studied in subjects with a normal or above-normal total HDL-chol concentration [6]. In these subjects HDL,-chol has almost reached its maximum so that an almost exclusive rise in HDL,-chol is obtained. It can now also be understood that Packard et al. [24] observed apparently different responses in normo- and hypertriglyceridemic subjects to nicotinic acid treatment. Treatment of patients with hypertriglyceridemia resulted in a rise of HDL, because the starting level of HDL was rather low. In the normotriglyceridemics an increase in HDL, was found, because these subjects had generally higher HDL-values. Assuming that the risk for coronary heart disease is linearly related to the inverse of the HDLchol concentration, our findings indicate, that a rise in HDL,-chol may be equally favorable as a rise in HDL,-chol. It may then be questioned whether subfractionation of HDL adds much to the risk estimation based on measurement of total HDL-chol only. Future epidemiological studies may shed light on this important question. However, at the moment the start of these studies is troubled by the imprecision and accuracy problems of HDL subclass measurement [25]. Acknowledgements We thank Mrs. Anneke Hijmans, Pieternel Van Heyst and Heidi Hak-Lemmers for excellent technical assistance, Dr. A.F.H. Stalenhoef for providing serum samples and Mrs. A. Verweijen and Mrs. Ans Ruesen for the preparation of the manuscript. References 1 Castelli,
W.P., Doyle, J.T., Gordon, T., HDL-cholesterol and other lipids in coronary heart disease, Circulation, 55 (1977) 161.
229 2 Miller. N.E.. Forde, O.H., Thelle, D.S. and Mjos, O.D., The Tromso study, high density lipprotein and coronary heart disease - A prospective case-control study, Lancet, i (1977) 965. 3 DeLalla. O.F., Elliot, H.A. and Gofman, J.W.. Ultracentrifugal studies of high density serum lipoproteins in clinically healthy adults. Amer. J. Physiol.. 179 (1954) 333. 4 Demacker. P.N.M., Van Sommeren-Zondag. D.F., Stalenhoef. A.F.H.. Stuyt, P.M. and Van ‘t Laar, A.. Ultracentrifugation in swinging-bucket and fixed-angle rotors evaluated for isolation and determination of high-density lipoprotein subfractions HDL, and HDL,, Clin. Chem., 29 (1983) 656. 5 Lewis, L.I., Miller. G.J., Burstein, M., Slagle, S. and Eder, H.A., Separation and quantitation of subclasses of human plasma high density lipoprotein by a simple precipitation procedure, J. Lipid. Res., 23 (1982) 1206. 6 Kraus. R.M., Regulation of high density lipoprotein levels, Med. Clin. N. Amer., 66 (1982) 403. I Anderson, D.W.. HDL-cholesterol - The variable components, Lancet, i (1978) 819. 8 Gofman, J.W., Young, W. and Tandy, R., Ischemic heart disease, atherosclerosis. and longevity, Circulation, 179 (1954) 333. Y Miller. N.E.. Hammett. S., Saltissi, S., Rao, S.. Van Zeller. H.. Coltart, J. and Lewis, B., Relation of angiographically defined coronary artery disease to plasma lipoprotein subfractions and apolipoproteins. Brit. Med. J.. 282 (1981) 1741. F.C.. Clark. R.S., Simpson, H.S. and Bal10 Ballantyne. lantyne. D.. High density and low density lipoprotein subfractions in survivors of myocardial infarction and in control SubJectS, Metabolism, 31 (1982) 433. E. and 11 Brook. J.G.. Aviram, M., Viener, A., Shilansky, Markiewicz, W.. High-density lipoprotein subfractions in normolipidemic patients with coronary atherosclerosis, Circulation, 66 (1982) 923. P.N.M. and 12 Stuyt. P.M.J., Stalenhoef, A.F.H., Demacker, Van ‘t Laar, A.. A comparative study of the effects of Acipimox and Clofibrate in type III and type IV hyperlipoproteinemia. Atherosclerosis. 55 (1985) 51. A.F.H.. Casparie. A.F., Demacker, P.N.M., 13 Stalenhoef, Stouten, J.T.J.. Lutterman, J.A. and Van ‘t Laar. A., Combined deficiency of apohpoprotein CII and lipoprotein lipase in familial hyperchylomicronemia, Metabolism, 30 (1981) 919. 14 Stuyt, P.M.J.. Demacker, P.N.M. and Van ‘t Laar. A., Serum lipids, lipoproteins and apohpoprotein E phenotypes
15
16
17
18
19
20
21
22
23
24
25
in relatives of patients with type III hyperlipoproteinemia, Europ. J. Clin. Invest., 14 (1984) 219. Demacker, P.N.M., Schade, R.W.B., Jansen, R.T.P. and Van ‘t Laar, A., Intra-individual variation of serum triglycerides and high-density lipoprotein cholesterol. cholesterol in normal humans, Atherosclerosis. 45 (1982) 259. Roschlau, P., Bernt. E. and Gruber. W.. Enzymatische Bestimmung des Gesamt-Cholesterins im Serum. J. Clin. Chem. Biochem., 12 (1974) 403. Demacker, P.N.M., Van Oppenraay, J.B., Baadenhuysen. H. and Jansen, A.P., An improved semi-automated method for the calorimetric determination of triglycerides m serum, Clin. Chim. Acta. 64 (1975) 45. Patsch, J.R., Gotto, Jr., A.M., Olivecrona, T. and Eisenberg, S.. Formation of high-density lipoprotein,-like particles during lipolysis of very low density lipoproteins in vitro, Proc. Nat. Acad. Sci. (USA), 75 (197X) 4519. Dieplinger, H.. Zechner, R. and Kostner, G.M.. The in vitro formation of HDLz during the action of LCAT The role of triglyceride-rich lipoproteins, J. Lipid. Res. 26 (1985) 273. Jansen, H.. Van Tol, A. and Hiilsmann. W.C., On the metabolic function of heparin releasable liver lipase, Biothem. Biophys. Res. Commun., 92 (1980) 53. Cheung. M.C. and Albers, J.J., Characterization of lipoprotein particles isolated by immunoaffinity chromatography - Particles containing A-I and A-II and particles containing A-I but not A-II, J. Biol. Chem.. 259 (1984) 12201. Atmeh. R.F.. Shepherd, J. and Packard, C.J., Subpopulations of apolipoprotein A-I in human high-density lipoproteins, their metabolic properties and response to drug therapy, Biochim. Biophys. Acta, 751 (1983) 175. Shepherd, J.. Packard, C.J., Patsch, J.R. and Gotto. Jr., A.M., Effects of nicotinic acid therapy on plasma high density lipoprotein subfraction distribution and composition and on apolipoprotein A metabolism. J. Clin. Invest.. 63 (1979) 858. Packard, C.J., Stewart. J.M.. Third, J.L., Morgan, H.G.. Lowry, T.D. and Shepherd, J., Effects of nicotinic acid therapy on high-density lipoprotein metabolism in type 111 and type IV hyperlipoproteinemia, Biochim. Biophys. Acta, 618 (1980) 53. Demacker, P.N.M., Hak-Lemmers, H.L.Mr, Hijmans. A.G.M. and Baadenhuysen, H., Evaluation of the dual-precipitation method for determination of cholesterol in highdensity lipoprotein subfractions HDL, and HDL? in serum, Clin. Chem.. 32 (1986) 819.